Celestin Apprentice 5/Apprentice-Release5.iso/Information/THINK C Digest/1994/94-05

Text File | 1995-12-31 | 292KB | 8,169 lines

Archive THINK-C/MESSAGES, file 94-05.
Part 1/1, total size 290203 bytes:

------------------------------ Cut here ------------------------------
>From [e-mail] 1 May 94 02:08:46 GMT
Path: ucivax!gateway
From: [e-mail] ("Robert S. Mah")
Subject: Re: 'matching prototypes
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 38
Date: 1 May 94 02:08:46 GMT

At 11:58 PM 4/30/94 +0000, Clinton R Lefort wrote:
> but still get this message last argumetn to function GetNewWindow does
> not match prototype'
> [...]
> #define MOVE_TO_FRONT -1L
> [...]
> gHelloWindow = GetNewWindow( BASE_RES_ID , NIL_POINTER ,MOVE_TO_FRONT);
>
> The prototype seems to be
>
> pWindow := GetNewWindow(windowID,wStorage,behind);

That's not the prototype.  And looking at the header file, "windows.h"
shows that the prototype is...

pascal WindowPtr GetNewWindow(short windowID,void *wStorage,WindowPtr behind)

So...the last argument must have the type "WindowPtr", which if you look
further up the chain is the same as the type "GrafPtr" which is a struct.

Now, you have defined MOVE_TO_FRONT to be "-1L", which is of type "long".

Since "long" and "WindowPtr" (or "GrafPtr") are not the same type, the
compiler rightly complains that your call does not match the prototype.

Change the #define MOVE_TO_FRONT to...

  #define MOVE_TO_FRONT   ((GrafPtr)-1L)

..and all will be OK.

Cheers,
Rob

___________________________________________________________________________
Robert S. Mah  -=-  One Step Beyond  -=-  212-947-6507  -=-  [e-mail]


>From [e-mail] 1 May 94 03:03:55 GMT
Path: ucivax!gateway
From: [e-mail] ("Matthew E. Shafer")
Subject: Re: 'matching prototypes
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <Pine.3.07.9404302259.A20858-b100000@grog>
Approved: [e-mail]
Lines: 39
Date: 1 May 94 03:03:55 GMT

On 30 Apr 1994, Clinton R Lefort wrote:
>
>
>       gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER ,MOVE_TO_FRONT);
>       ShowWindow( gHelloWindow);
>       SetPort( gHelloWindow);
>       MoveTo(HORIZONTAL_PIXEL, VERTICAL_PIXEL );
>       DrawString("\pHello, World");
>       }
>
>
> The prototype seems to be
>
> pWindow := GetNewWindow(windowID,wStorage,behind);but the code that I have
>
> gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER ,MOVE_TO_FRONT);doesn't seem to take very well right now...
> any suggestiosn as to how to make this prototype work...the manual says that
> to move the window to the front you must use -1, Mark has -1L, I have tried both , but ot no avail>.
>
>
> Thanks
> Clinton R. LeFort
>

The last param to GetNewWindow needs to by typcast as a window pointer.
Change your macro to

#define MOVE_TO_FRONT (WindowPtr)-1L

-1 is a special value that can be passed to have the new window appear
in front of all other windows regardless. It needs to be cast as a
window pointer because you could pass any window pointer and the new
window would appear just in front of that one.

.Matt




>From [e-mail] 1 May 94 03:06:39 GMT
Path: ucivax!gateway
From: [e-mail] (Bare Bones Software)
Subject: BBEdit for Power Macintosh: Announcement
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL21]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 4273
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 91
Date: 1 May 94 03:06:39 GMT

April 30, 1994
For Immediate Release

Contact:
Rich Siegel
Bare Bones Software
1 Larkspur Way #4
Natick, MA 01760
Phone: [phone]
Fax: [phone]
Internet: [e-mail]
CIS: 73051,3255
AppleLink: BARE.BONES

Bare Bones Software announces BBEdit 3.0: High Performance Text Editor
for Power Macintosh(R)

Natick, MA -- Bare Bones Software, publisher of the premier Macintosh
text editor, BBEdit 2.5, is readying a new release of BBEdit with
extended functionality, including support for Power Macintosh(R) systems.

BBEdit 3.0 will take advantage of native PowerPC(TM) code to greatly
enhance the speed of common operations such as loading and saving files,
and text editing, searching and transformation.

In addition to support for the newest Macintosh line, BBEdit 3.0 will
feature a unique multi-file Compare facility, which can compare
individual files, folders, or project documents for differences. When
comparing folders or projects, BBEdit will display files which are
common to both folders or projects, as well as files unique to each
folder or project. Common files can be directly compared for
differences. As the result of an alliance with SciComp Software, BBEdit
3.0 will also feature integrated PopupFuncs(TM) technology, which allows
quick access to C, C++, or Pascal constructs defined in a source file
via a popup menu in that file's window.

Other significant enhancements include support for user scripting via
AppleScript, Frontier, or any other OSA-compliant scripting system, use
of Macintosh Drag and Drop in the user interface and for text editing,
and support for the AOCE "Standard Mail Package", so that users can
compose, send, and read PowerTalk(R) mail messages from within BBEdit.

BBEdit 3.0 also features a number of user-interface enhancements,
included in direct response to user comments and suggestions. In order
to smooth the transition from other text-editing products to BBEdit,
Bare Bones Software has included a number of new text transformations,
including line sorting, generation of C function prototypes, line
numbering, and organization of delimited words into rows and columns.

Response from beta testers has been enthusiastic. Tom Emerson, a
software engineer in Symantec's Macintosh Development Tools Group, said,
"The speed improvements on Power Macs are very impressive. Large files
open much faster than before, and search and replace operations,
particularly multi-file seaching, really fly. I'm able to get more work
done in less time, thanks to BBEdit."According to Leonard Rosenthol,
director of advanced technology at Aladdin Systems, "The new ability to
compare folders has proven itself invaluable. I'm able to quickly
compare two sets of source files and determine the differences between
them, and integrate those differences. Until BBEdit 3.0 came along, it
was a real chore." Chris Borton, senior software engineer at Storm
Technology, Inc., said that "BBEdit 3.0's combination of ease of use,
drag-and-drop text editing, XTND(TM) support, and integration with Symantec
C++ have made it my ideal development environment."

BBEdit 3.0 is slated to ship in early June. The suggested retail price
of BBEdit 3.0 will be $99.00.

Current BBEdit 2.5 owners can upgrade to the version 3.0 for $39.00
(U.S. shipping & handling included). Customers who purchase BBEdit 2.5
between May 1, 1994 and the shipping date of BBEdit 3.0 will be able to
upgrade for $5.00 (within the U.S.), or $10.00 for Canadian and overseas
customers (shipping and handling included).

Customers who are currently using a freeware version of BBEdit, or a
competing or complementary product, can upgrade to BBEdit 3.0 for $69.00
plus shipping and handling. (Contact Bare Bones Software for
cross-upgrade information.) Students will be able to purchase BBEdit 3.0
for $39.00.

For additional information, including quantity and site-license pricing,
contact Bare Bones Software at the address given above.

--

PowerPC is a trademark of International Business Machines Corporation.
Macintosh and PowerTalk are registered trademarks of Apple Computer, Inc.
PopupFuncs is a trademark of SciComp Software.
XTND is a trademark of Claris Corporation.

Other products mentioned are trademarked by their respective producers.

>From [e-mail] 1 May 94 07:16:31 GMT
Path: ucivax!gateway
From: [e-mail] ("Timothy J. Greear")
Subject: Vector Tutorial problem
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 28
Date: 1 May 94 07:16:31 GMT


Hi
Can someone help me I have a Problem with my Symantec C++
program. I am trying to run the Vector  Project Tutorial in the User Guide
It fails I get a link error and these code statements:

     undefined: operator <<(ostream&,vector<char>&)
     undefined: operator <<(ostream&,vector<myDate>&)
     undefined: operator <<(ostream&,vector<float>&)
     undefined: operator <<(ostream&,vector<int>&)



I have updated my program to version 7.0.  This is
the free upgrade not the full featured version from Symantec.

What I have tried to do the following:

Tried to look for (ostream&,vector) using the find function  *no go*.
Removed the objects after conversion of the IOstream
library and brought it back up to date *no go*.

I have a Mac llsi with 9Mb.

Thanks

Timothy j. Greear
[e-mail]
>From [e-mail] 1 May 94 08:03:18 GMT
Path: ucivax!gateway
From: [e-mail] (Clinton R Lefort)
Subject: prototypes & bus error
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 72
Date: 1 May 94 08:03:18 GMT


Hello,

I have tried the helpful suggestions of those who responded to my query,
unfortunately I tried to run this code with the'debugger on' and I stepped thru
the entire code , but whenit gets to the end of the code where it is asked to print "Hello world' I get the bus  error : 0x003FEFF6 at teh bottom of the debugger window....this occurs whether I place (WindowPtr)-1) (WindowPtr)-1 at rehdefines or at the function call to getHelloWindow:

gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER , (WindowPtr)MOVE_TO_FRONT);

Here is the code again, maybe someone else would like to try it, and see if they get teh same response...I am running THINKC 6.0...




#define BASE_RES_ID     400
#define NIL_POINTER 0L
#define MOVE_TO_FRONT   ((GrafPtr)-1L)
#define REMOVE_ALL_EVENTS 0
#define HORIZONTAL_PIXEL 30
#define VERTICAL_PIXEL 50

WindowPtr       gHelloWindow;

/*************************Main************/


main()

{




        ToolBoxInit();
        WindowInit();
        while ( !Button() );

}


/************************************ToolBoxInit */
ToolBoxInit()

{

        InitGraf( &thePort);
        InitFonts();
        FlushEvents( everyEvent, REMOVE_ALL_EVENTS );
        InitWindows();
        InitMenus();
        TEInit();
        InitDialogs(NIL_POINTER);
        InitCursor();

}

/***********************WindowInit***************************/
WindowInit()
{


        gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER , (WindowPtr)MOVE_TO_FRONT);
        ShowWindow( gHelloWindow);
        SetPort( gHelloWindow);
        MoveTo(HORIZONTAL_PIXEL, VERTICAL_PIXEL );
        DrawString("\pHello, World");
        }

        Thanks

Clinton R. LeFort

>From [e-mail] 1 May 94 14:53:51 GMT
Path: ucivax!gateway
From: [e-mail] ("Robert S. Mah")
Subject: Re: prototypes & bus error
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 35
Date: 1 May 94 14:53:51 GMT

At  8:03 AM 5/1/94 +0000, Clinton R Lefort wrote:
>
> I have tried the helpful suggestions of those who responded to my query,
> unfortunately I tried to run this code with the'debugger on' and I stepped
> thru the entire code , but whenit gets to the end of the code where it is
> asked to print "Hello world' I get the bus  error : 0x003FEFF6 at teh
> bottom of the debugger window....this occurs whether I place (WindowPtr)-1)
> (WindowPtr)-1 at rehdefines or at the function call to getHelloWindow:
>
> gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER ,
>                             (WindowPtr)MOVE_TO_FRONT);
> ShowWindow( gHelloWindow);
> SetPort( gHelloWindow);
> MoveTo(HORIZONTAL_PIXEL, VERTICAL_PIXEL );
> DrawString("\pHello, World");

The moveToFront param has nothing to do with your bus error.  Is the window
appearing?  If not, then you're not loading a window, probably because,
either
a) you don't have enough memory available (unlikely) or b) it can't find
the WIND resource.  If b), then this can be caused because 1) the resource
isn't there, 2) its been numbered incorrectly or 3) because you named your
resource file incorrectly.

In addition, you should ALWAYS check return values.  Your code is assuming
that the call to GetNewWindow will always succeed.  This is not a valid
assumption.

Cheers,
Rob

___________________________________________________________________________
Robert S. Mah  -=-  One Step Beyond  -=-  212-947-6507  -=-  [e-mail]


>From [e-mail] 1 May 94 15:17:49 GMT
Path: ucivax!gateway
From: [e-mail] ("Matthew E. Shafer")
Subject: Re: prototypes & bus error
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <Pine.3.07.9405011146.A10930-c100000@grog>
Approved: [e-mail]
Lines: 81
Date: 1 May 94 15:17:49 GMT

On 1 May 1994, Clinton R Lefort wrote:
> Hello,
>
> I have tried the helpful suggestions of those who responded to my query,
> unfortunately I tried to run this code with the'debugger on' and I stepped thru
> the entire code , but whenit gets to the end of the code where it is asked to print "Hello world' I get the bus  error : 0x003FEFF6 at teh bottom of the debugger window....this occurs whether I place (WindowPtr)-1) (WindowPtr)-1 at rehdefines or at the f
unction call to getHelloWindow:
>
> gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER , (WindowPtr)MOVE_TO_FRONT);
>
> Here is the code again, maybe someone else would like to try it, and see if they get teh same response...I am running THINKC 6.0...
>
>

[Buggy code deleted]

I noticed you had your BASE_RES_ID set to 400. GetNewWindow pulls a WIND
resource that has already been created. Make sure that you used ResEdit
to create the resource and that you set it's ID to 400. You should also
put in some error checking in case the program can't find the resource.
Try this code:

#define BASE_RES_ID       400   // assuming resource exists
#define MOVE_TO_FRONT     (WindowPtr)-1L
#define HORIZONTAL_PIXEL  30
#define VERTICAL_PIXEL    50

void  ToolBoxInit( void );
void  WindowInit( void );

/******** main **********/
void   main( void )
{
    ToolBoxInit();
    WindowInit();

    while ( !Button() );
}

/******* ToolBoxInit ********/
{
    InitGraf( &thePort );
    InitFonts();
    InitWindows();
    InitMenus();
    TEInit();
    InitDialogs( nil );
    InitCursor();
}

/******* WindowInit ********/
{
    WindowPtr  window;   /* use local variables if possible */

    window = GetNewWindow( BASE_RES_ID, nil, MOVE_TO_FRONT );

    if ( window == nil )
    {
        SysBeep( 10 );  /* can't find resource */
        ExitToShell();
    }

    ShowWindow( window );
    SetPort( window );

    MoveTo( HORIZONTAL_PIXEL, VERTICAL_PIXEL );
    DrawString( "\pHello, world!" );
}

A couple things:
(1) avoid using global variables. It was really no biggie in your
    original code, but it's bad style.
(2) If you use a Toolbox function that relies on resources (such as
    GetNewWindow), make sure that you have already created the resource.
    Name the resource file after your THINK project name. For example,
    My_project.pi.rsrc.
(3) Use error checking. It avoids some crashes.

.Matt


>From [e-mail] 1 May 94 18:39:04 GMT
Path: ucivax!gateway
From: [e-mail] (Clinton R Lefort)
Subject: PlaySound() and buggy code
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 119
Date: 1 May 94 18:39:04 GMT


Hello,

I was lucky to get my 'hello world' program running by placing my resources in the right place, now that that problem has been solved..I got bored , andso I havecreated another resource, but now that it isplaced inthe proper place, I cannot get the SndPlay(mySnd) to function.....

The code is all thesame, but I have added the sound.h and a function that plays
a resource  for me.. Wheni rn the debugger I get a syntax error at

SndChannel =&SndChannelPtr;



#define kAsync TRUE    //play is asynchronous
#define mySndlID 129  //resource ID of 'snd'
#define BASE_RES_ID     400
#define NIL_POINTER 0L
#define MOVE_TO_FRONT  ((WindowPtr)-1L)
#define REMOVE_ALL_EVENTS 0
#define HORIZONTAL_PIXEL 30
#define VERTICAL_PIXEL 50

WindowPtr       gHelloWindow;

/*************************Main************/

#include <Sound.h>




void CallSndPlay(void);

void callSndPlay()
{

Handle mySndHandle;
ResType snd;
short  rID;
rID = 129;
SndChannel = *SndChannelPtr;
SndChannelPtr mySndChan;
mySndChan = Chan1;
OSErr myErr;


//prototype for user defined error handling routine
void DoError(OSErr);

mySndChan= nil; //Initialize channel pointer fro error checking

//Read in 'snd' resource from resource file
mySndHandle = GetResource('snd', mySndID);

//check for a nil handle

        if(mySndHandle !=nil {
                myErr = SndPlay(mySndChan, mySndHandle,kAsync);
                if(myErr)
                DoError(myErr);
        }

}
main()

{

        //costants


        ToolBoxInit();
        WindowInit();
        while ( !Button() );

}


/************************************ToolBoxInit */
ToolBoxInit()

{

        int myFontNumber;
        int myFontSize;
        TextFont(myFontNumber);
        TextSize(myFontSize);
        InitGraf( &thePort);
        InitFonts();
        FlushEvents( everyEvent, REMOVE_ALL_EVENTS );

        InitWindows();
        InitMenus();
        TEInit();
        InitDialogs(NIL_POINTER);
        InitCursor();

}

/***********************WindowInit***************************/
WindowInit()
{


        gHelloWindow= GetNewWindow( BASE_RES_ID , NIL_POINTER , (WindowPtr)MOVE_TO_FRONT);
        ShowWindow( gHelloWindow);
        SetPort( gHelloWindow);
        MoveTo(HORIZONTAL_PIXEL, VERTICAL_PIXEL );
        TextFont(monaco);
        TextFace(bold);
        TextSize(24);
        DrawString("\pHello, World");
        }


Does anyone see the obvious that I don't see, other than that I can't program?

Thanks,

Clinton R. LeFort

>From [e-mail] 1 May 94 22:58:47 GMT
Path: ucivax!gateway
From: [e-mail] (Danny Thomas)
Subject: Re: PlaySound() and buggy code
X-Sender: [e-mail].
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 29
Date: 1 May 94 22:58:47 GMT

starting to use prototypes
> ...
>void CallSndPlay(void);
>

the suggestions to check return values for every system call is VERY
worthwhile, particularly when you're learning a system. It WILL save you
from programs mysteriously crashing - at least as long as you print out a
meaningful error message upon an unexpected return value. Of course
there'll be other mysterious crashes and problems...
    Another suggestion is to enable the "require prototypes" option in the
compiler. This is not directed at you, but it amazes me that many popular
programs I build for our unix system, eg sendmail, CAP, LIFE, NetBSD
itself, etc, do not (always) include all header files or prototypes for
local functions. Partly that's because turning on the the compiler warning
options gives so much it's hard to separate the useful from the
not-so-useful. Of course one pass could clear up the S/N to a large degree.
I'm sure many mac programmers follow the same sloppy practice and don't
give the compiler enough assistance to pick up their glaring errors. The
time it takes to include/write prototypes is a tiny fraction of the overall
programming time and they do save much more time overall. Unfortunately,
when THINK C (haven't run with 7 yet) reports a call/declaration prototype
mismatch it does not tell you what type it was given vs what type is
expected.

cheers,
Danny Thomas


>From [e-mail] 2 May 94 03:40:18 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: PlaySound() and buggy code
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Danny Thomas" at May 1, 94 10:58:47 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1415
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 27
Date: 2 May 94 03:40:18 GMT

> ...it amazes me that many popular programs I build for our unix
> system, eg sendmail, CAP, LIFE, NetBSD itself, etc, do not (always)
> include all header files or prototypes for local functions. Partly
> that's because turning on the the compiler warning options gives so
> much it's hard to separate the useful from the not-so-useful. Of
> course one pass could clear up the S/N to a large degree.

Also, many compilers have ways to control warnings to varying degrees.
Some support command line switches to control sets of warnings or
warning levels or individual warnings. Some have #pragmas which allow
you to control warnings in the code. (I like the former technique
better, because the settings are compiler-specific, and the command line
is compiler-specific anyway.)

There's no excuse for not responding to warnings, even if there are
no such controls over granularity. UNIX code gets away with it either
because it's old old code or because it's written by folks who learned
C before ANSI, and since they've been at it so long, they like to think
they know what they are doing better than the compiler does. Sometimes
their code can't even be compiled with prototypes without major changes.
I guess that's something of an excuse for old code, but not for new
code.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 Symantec C++ a "Bjarne Burner"? Yeah, I'd be pretty burned up, too!
>From [e-mail] 2 May 94 12:29:15 GMT
Path: ucivax!gateway
From: [e-mail] (Jens-Uwe Mager)
Subject: Re: Vector Tutorial problem
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date: 2 May 94 12:29:15 GMT


>Can someone help me I have a Problem with my Symantec C++
>program. I am trying to run the Vector  Project Tutorial in the User Guide
>It fails I get a link error and these code statements:
>
>     undefined: operator <<(ostream&,vector<char>&)
>     undefined: operator <<(ostream&,vector<myDate>&)
>     undefined: operator <<(ostream&,vector<float>&)
>     undefined: operator <<(ostream&,vector<int>&)

I did have the same problem, and found the answer in a c.s.p posting.
There is a difference in template instantiation in 7.0, you have to add
lines like the following to each of the instantion .cp files:

#pragma template operator <<(ostream&, vector<char>&)

The above line would go into the vector <char>.cp file. You will have
to do that for all the vector <xxx>.cp files.
>From [e-mail] 2 May 94 13:24:22 GMT
Path: ucivax!gateway
From: [e-mail] (Joshua Ball /ADVISOR Baggs)
Subject: Hypercard 2.2
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 23
Date: 2 May 94 13:24:22 GMT

Anyone out there have Hypercard 2.2 they could lend me for a
couple of days.

I have been trying to write a program to display sounds and pictures
and do some timing for a psych experiment and have failed miserably.

I am attempting to graduate here on saturday and it is due
Thursday so I am reverting back to Hypercard and need to knock it
out.
Sometimes grades are more important than learning=> go figure :-(

I have 2.1 but have a book for 2.2.

Help a student in desparate (ie graduation depends on it) NEED

Thanks,

Josh    [e-mail]

P.S. If you ar one of those who are getting all huffy about stealing
software, believe me, I have no desire to keep this program, just
use it temporarily to run an experiment. I promise I will use my "copy"
of disk doctor to erase it after I am done :-)
>From [e-mail] 2 May 94 14:31:04 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: PlaySound() and buggy code
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 12
Date: 2 May 94 14:31:04 GMT

Two Things:

First, it's a 'snd ', not a 'snd' resource, probably why you are crashing -
check the validity of mySndHandle befor calling sndplay.

Second, you can use the kasync parameter, but the sound will only play
syncronously unless you create a 'real' sndChannel.

Good Luck!
Chris

(no dorky poems to follow)
>From [e-mail] 2 May 94 15:41:22 GMT
Path: ucivax!gateway
From: [e-mail] (Lance Helsten)
Subject: Re: PlaySound() and buggy code
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 30
Date: 2 May 94 15:41:22 GMT

> the suggestions to check return values for every system call is VERY
> worthwhile, particularly when you're learning a system. It WILL save you
> from programs mysteriously crashing - at least as long as you print out a
> meaningful error message upon an unexpected return value. Of course
> there'll be other mysterious crashes and problems...
>     Another suggestion is to enable the "require prototypes" option in the
> compiler. This is not directed at you, but it amazes me that many popular
> programs I build for our unix system, eg sendmail, CAP, LIFE, NetBSD
> itself, etc, do not (always) include all header files or prototypes for
> local functions. Partly that's because turning on the the compiler warning
> options gives so much it's hard to separate the useful from the
> not-so-useful. Of course one pass could clear up the S/N to a large degree.
> I'm sure many mac programmers follow the same sloppy practice and don't
> give the compiler enough assistance to pick up their glaring errors. The
> time it takes to include/write prototypes is a tiny fraction of the overall
> programming time and they do save much more time overall. Unfortunately,
> when THINK C (haven't run with 7 yet) reports a call/declaration prototype
> mismatch it does not tell you what type it was given vs what type is
> expected.
>

I agree with Danny. If you want clean code turn on all the warnings your
compiler can give you. Those warnings are common places that programmers
make mistakes with the language. Another hint is to use assert in your
development code, it will help track down bogus pointers and other values
out of range while the system is running and not after the system has crashed.

Lance


>From [e-mail] 2 May 94 15:54:15 GMT
Path: ucivax!gateway
From: [e-mail] ("Griffin, Ed E")
Subject: Re: PalySound() and buggy code
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 13
Date: 2 May 94 15:54:15 GMT

Dear Clinton,

     I second and "third" the sentiments of DT and PG on checking returns and
using prototypes.  Turning on the "Require Prototypes" switch in the compiler
has saved me from silly errors countless times.  Prototype checking even
helped me find a small but deadly typo in the first printing of the Mac
Programming Primer vol. 2!
     Furthermore, given that Symantec supplies you with the useful and simple
to use _Prototype Helper_ (in the Utilities folder), there's no excuse for
not using current coding standards even on your older projects.

                              Ed Griffin
                              [e-mail]
>From [e-mail] 2 May 94 19:00:17 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: #define's for Fonts Folder
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 24
Date: 2 May 94 19:00:17 GMT

I am using TC5 and can't find header files which contain definitions for
the constants kFontsFolderType, kFontsFolderAliasType, and
fontsFolderIconResource.

I assume that they are not defined in any of the TC5 header files.  Please
let me know if they are and where.  If they are not, I need to define them
myself.  I make the following assumptions from the 'icl8' and 'fld#'
resources of the System file.

        #define kFontsFolderType          'font'
        #define fontsFolderIconResource   -3968

I'm not sure if these definitions are correct and I still need a definition
for kFontsFolderAliasType.

        #define kFontsFolderAliasType     ??

~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 if (atFirstYouDontSucceed)
     Try(Internet);
 /* Brent Taylor */
~~~~~~~~~~~~~~~~~~~~~~~~~~~~


>From [e-mail] 2 May 94 21:39:59 GMT
Path: ucivax!gateway
From: [e-mail] (David A Hausladen)
Subject: building code resource
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 19
Date: 2 May 94 21:39:59 GMT

I am part of an academic team trying to build an XFCN for HyperCard.  We
are using Symantec C++ v.6.  The C++ code compiles without error, but
when we try to build a code resource out of it, we get a link failure
with the following solitary message:
    undefined:  _XTOI
when we check link, it tells us that the undefined identifier occurs in
our source file, i.e. not in the libraries in the project.  In our
source file, we have the default file prefix <MacHeaders++> included, as
well as <HyperXCmd.h> and <stdlib.h>.  The libraries included in the
project are HyperXLib, MacTraps, and the source file. We have also tried
including the ANSI-A4++ library.

Where is _XTOI defined?  What library should be included in the project
to get a successful link?  Can we skirt the issue entirely?

Thanks,
Dave Hausladen
[e-mail]

>From [e-mail] 2 May 94 23:30:04 GMT
Path: ucivax!gateway
From: [e-mail] (David A Hausladen)
Subject: building code resources
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 22
Date: 2 May 94 23:30:04 GMT


I am part of an academic team trying to build an XFCN for HyperCard.  We
are using Symantec C++ v.6.  The C++ code compiles without error, but
when we try to build a code resource out of it, we get a link failure
with the following solitary message:
    undefined:  _XTOI
when we check link, it tells us that the undefined identifier occurs in
our source file, i.e. not in the libraries in the project.  In our
source file, we have the default file prefix <MacHeaders++> included, as
well as <HyperXCmd.h> and <stdlib.h>.  The libraries included in the
project are HyperXLib, MacTraps, and the source file. We have also tried
including the ANSI-A4++ library.

Where is _XTOI defined?  What library should be included in the project
to get a successful link?  Can we skirt the issue entirely?

Thanks,
Dave Hausladen
[e-mail]



>From [e-mail] 3 May 94 03:36:12 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: building code resources
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "David A Hausladen" at May 2, 94 11:30:04 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 400
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 10
Date: 3 May 94 03:36:12 GMT

> Where is _XTOI defined? What library should be included in the project
> to get a successful link? Can we skirt the issue entirely?

Try "CPlusLib-A4". It will probably require "ANSI-A4++", too.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "The human race will decree from time to time: 'There is something at
 which it is absolutely forbidden to laugh.'" -- Nietzche on C++
>From [e-mail] 3 May 94 15:30:32 GMT
Path: ucivax!gateway
From: [e-mail] (Jannice Caufield)
Subject: Floppies vs the Desktop Manager
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 36
Date: 3 May 94 15:30:32 GMT
Fro: [e-mail] (Brent Taylor)

As I understand IM6, the Finder does NOT maintain a desktop database for
volumes with capacity less than 2 MB (eg. floppies).  Therefore, I can't
use the Desktop Manager routines on these volumes.  Doesnw how to
retrieve Desktop information (icons, comments, etc) from thume?

Relevant portions of IM6 text follow:

/******************************* IM6, p9.45-46 ********************************/

"The Finder maintains a desktop database for each volume with a capacity
greater than 2 MB.  For most volumes, such as hard disks, the database is
stored on the volume itself.  For read-only volumes--such as some compact
dtha don't contain their own desktop database, the Desktop Manager
creates it and stores it in the System Folder of the boot drive....

"For compatibility with older versions of system software, the Finder keeps
the information for ejectable volumes with a capacity smaller than 2 MB in
a resource file instead of a database....

"In earlier versions of system software, Finder information for each volume
was stored in the volume's Desktop file, a resource file created and used
by the Finder and invisible to the user.  This strategy meets the needs of
a single-user system with reasonably small volumes.  The Desktop file is
still used on ejectable volumes with a capacity less than 2 MB so that
these floppy disks can be shared with Macintosh computers running earlier
versions of system software.  (Note, however, that resources can't be
shared.  Since the Finder is always running in system software version 7.0,
it keeps each floppy disk's Desktop file open, so your application can't
read or write it.)"
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 if (atFirstYouDontSucceed)
     Try(Internet);
 /* Brent Taylor */
~~~~~~~~~~~~~~~~~~~~~~~~~~~~


>From [e-mail] 3 May 94 15:39:11 GMT
Path: ucivax!gateway
From: [e-mail] (John Grimes)
Subject: Show a Moving Region
Message-ID: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 23
Date: 3 May 94 15:39:11 GMT


        I was wondering if anyone could help me on my latest difficulty.
Anyway I have picts (usually defines as controls but sometimes just drawn
on the screen) and I would like to show them moving from point A to B.
Sometimes I can use draggray region to do this (when working with the
mouse but thats not always the case).  Also this only shows the gray not
the pict.  But more importantly I would like to be able to move the pict
from point A to B and make it possible for the user to see the movement.
Not just having the pict show up at a new place.  I appreciate any help.
                                                Thanks,
                                                John Grimes

******************************************************************************
        This was no time for play
        This was no time for fun
        This was no time for games
        There was work to be done.
                        Cat in the Hat Comes Back-Dr. Seuss
******************************************************************************
                                        John Grimes
                                        reaper@jhunix
                                        HAC Consultant

>From [e-mail] 3 May 94 17:19:05 GMT
Path: ucivax!gateway
From: [e-mail] (Kevin O'Malley)
Subject: SCSIGet() crashes system
Message-ID: <[e-mail]>
Content-Type: TEXT/plain; charset=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 3 May 94 17:19:05 GMT


Under Symantec 7.0 (as well as 5.0/6.0.1) the call to SCSIGet() crashes our
machine. Symantec tech. support, after trying it, also crashed their machine.
They said to contact Apple for more documentation on the SCSI Manager. Has
anyone seen this problem, or know a solution or workaround?

Kevin - [e-mail]




>From [e-mail] 3 May 94 18:03:47 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: Error ID -410
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 13
Date: 3 May 94 18:03:47 GMT

I just got error -410 while trying to obtain desktop database information
from a CD volume.  My documentation (IM1-6) does not contain a description
for this error ID.

Does anybody know what it is?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 if (atFirstYouDontSucceed)
     Try(Internet);
 /* Brent Taylor */
~~~~~~~~~~~~~~~~~~~~~~~~~~~~


>From [e-mail] 3 May 94 19:59:28 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Floppies vs the Desktop Manager
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Jannice Caufield" at May 3, 94 03:30:32 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 842
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 17
Date: 3 May 94 19:59:28 GMT

> As I understand IM6, the Finder does NOT maintain a desktop database
> for volumes with capacity less than 2 MB (eg. floppies). Therefore, I
> can't use the Desktop Manager routines on these volumes. Doesnw how to
> retrieve Desktop information (icons, comments, etc) from thume?

Finder does maintain a desktop database, but it's not in the same
format. It's a System 6 era resource file called "Desktop" (big
surprise). Its format was never documented, and those who wanted to read
it had to reverse engineer it. On the good side of the undocumentedness,
though, is that it's probable the format will never again change in a
way that would break code which reads a file generated by a System 6
Finder.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 Symantec C++ a "Bjarne Burner"? Yeah, I'd be pretty burned up, too!
>From [e-mail] 3 May 94 21:00:44 GMT
Path: ucivax!gateway
From: [e-mail] (Wes Covalt)
Subject: Mini-spreadsheet
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 9
Date: 3 May 94 21:00:44 GMT

I want to build a mini-spreadsheet for entering data for statistical
analysis.  Can anyone suggest a book that gives sample C code
for the procedures needed for such an application?  (I know I
_should_ learn C++, but I want to be secure in the basics first.)

Thanks
Wes
Northern Illinois University
[e-mail]
>From [e-mail] 4 May 94 00:20:50 GMT
Path: ucivax!gateway
From: [e-mail] (Joe Kalash)
Subject: Symantec c and applescript?
Message-ID: <[e-mail]>
X-Mailer: Mail*Link SMTP/QM 3.0.0b6
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 29
Date: 4 May 94 00:20:50 GMT

                       Subject:                               Time:3:50 PM
  OFFICE MEMO          Symantec c and applescript?            Date:5/3/94

I am trying to automate the build of several Think-C projects (libraries,
applications, code resources) from within MPW. What I would like to do is ask
the Project Manager to check a project (a Make with "use disk"), and then (if
neccesary) build the resulting object.

I can open and build a think project w/o problem from an applescript
controlling the think project manager.  My dilemma is determining whether the
build failed or succeeded.  Has anyone figured out a way?  I do not seem to
be able to look for the presence of the 'compile errors' window or anything
similar.  Perhaps I could 'poke' some think c structure (undocumented) to
determine whether a project needs building or not but I'm unsure as to where
to look.  Any assistance would be greatly appreciated.

While I'm at it, has anyone figured a way of passing a document name into an
applescript (other than by drag and dropping)?

Is there some source for think project manager controlling scripts?

Anyone w/experience writing an mpw tool to call applescript?

Thanks for the help.

Joe Kalash
StarNine Technologies, Inc.
[e-mail]

>From [e-mail] 4 May 94 12:49:38 GMT
Path: ucivax!gateway
From: [e-mail] (barnett aaron)
Subject: getting desk pattern
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 7
Date: 4 May 94 12:49:38 GMT


how can i get a handle to the desktop pattern for use in an application?


\.
aaron <[e-mail]>
"It's a safety mechanism, the mind shuts off."
>From [e-mail] 4 May 94 17:09:17 GMT
Path: ucivax!gateway
From: [e-mail] (Chris Metcalfe)
Subject: Re:  Mini-spreadsheet
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 10
Date: 4 May 94 17:09:17 GMT

The book "Easy Object Programming for the Macintosh" has an exellent
example of a spreadsheet, annotation, and plotting package called
"Ensemble."  It compiles on Think C 5.0, and uses AppMaker 1.5 to
edit the UI.  However, since all the source files are given for each
stage of the development, you don't really need AppMaker to hack on
the final version of the program.  (Although it is a worthwhile
purchase.)

-- Chris Metcalfe

>From [e-mail] 4 May 94 19:16:54 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: Re: Error ID -410
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 20
Date: 4 May 94 19:16:54 GMT

In a previous message, I said:
>I just got error -410 while trying to obtain desktop database information from
>>a CD volume.  My documentation (IM1-6) does not contain a description for
>this >error ID.
>
>Does anybody know what it is?

To be more specific, the error resulted from a PBDTGetPath() call on a CD.
This call seems to work with most CD's, however, it failed on a particular
CD.  I'm guessing that perhaps the CD lacks it's own desktop database, but
I can't really be certain until I learn what error -410 really means.  I
would appreciate any help on this one.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 if (atFirstYouDontSucceed)
     Try(Internet);
 /* Brent Taylor */
~~~~~~~~~~~~~~~~~~~~~~~~~~~~


>From [e-mail] 4 May 94 21:53:31 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: Floppies vs the Desktop Manager
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 20
Date: 4 May 94 21:53:31 GMT

>> As I understand IM6, the Finder does NOT maintain a desktop database
>> for volumes with capacity less than 2 MB (eg. floppies). Therefore, I
>> can't use the Desktop Manager routines on these volumes. Doesnw how to
>> retrieve Desktop information (icons, comments, etc) from thume?

>Finder does maintain a desktop database, but it's not in the same
>format. It's a System 6 era resource file called "Desktop" (big
>surprise). Its format was never documentedx

The desktop file and the desktop database are two different things. The
desktop FILE (called "Desktop" in sys. 6 and I believe "Desktop DF" in sys.
7) has been a part of the Macintosh at lease since HFS has been around. The
desktop DATABASE is different. Its sole purpose is to allow the volume to be
shared. Since the desktop file cannot be shared (because it's a resource
file), there is a separate file called the "Desktop DB" which stores the same
information as the file, is updated when necessary, but is only read when the
volume is mounted. The desktop database is also available to applications.

Greg Scherrer
[e-mail]-
>From [e-mail] 5 May 94 03:01:27 GMT
Path: ucivax!gateway
From: [e-mail] (Jerry Gatlin)
Subject: Re: getting desk pattern
X-Sender: [e-mail] (Unverified)
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <9405050244.AA10470@hipark>
Approved: [e-mail]
Lines: 40
Date: 5 May 94 03:01:27 GMT

>how can i get a handle to the desktop pattern for use in an application?
>

This is a real quick and dirty but here goes. I haven't done any error
checking, etc. so please keep this in mind when really coding it. First, if
you want the B&W Desktop pattern use the first example. If you have color,
use the second. Make sure you include LowMem.h. These routines worked fine
for me although this was my first time trying to do this so I don't claim
to be an expert. If anyone else has a better way of doing this speak up.

Jerry


Pattern                 thePat;
Rect                    theRect;
PixPatHandle            thePPat;

/* B&W example */

LMGetDeskPattern(&thePat);
SetRect( &theRect, 100,100, 200,200 );
FrameRect( &theRect);                   /* frame the area */
InsetRect( &theRect, 1,1);              /* don't overwrite the frame */
FillRect( &theRect, &thePat);           /* fill with B&W desktop pattern */

/* Color example */

thePPat = LMGetDeskCPat();
SetRect( &theRect, 200,200, 300,300 );
FrameRect( &theRect);                   /* frame the area */
InsetRect( &theRect, 1,1);              /* don't overwrite the frame */
FillCRect( &theRect, thePPat);          /* fill with color desktop pattern */


____________________________________________________________________
Jerry Gatlin                       [e-mail]
Hill Country Softworks (my other life)                    Austin, TX
____________________________________________________________________


>From [e-mail] 5 May 94 09:58:05 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: Getting a filename in a fil
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 38
Date: 5 May 94 09:58:05 GMT


>From Pindar user Ian McCall

Subject: _Getting a filename in a file filter
I'm writing a CustomGetFile filter that depends on the file's name
for its result. This filter takes the ParamBlkPtr it's passed and
then does a PBGetFInfo. Despite the fact the ParmBlkPtr was passed
over by CustomGetFile, I then get an OSErr of -43 (fnfErr).

Can anyone tell me what I'm doing wrong? The code's below.



pascal Boolean ListFileFilter(ParmBlkPtr fileDetails, void
*dummyData)
{
   Str255   fileName;
   OSErr    theResult;


   fileDetails->fileParam.ioCompletion = NULL;
   fileDetails->fileParam.ioFDirIndex  = 0;
   fileDetails->fileParam.ioFVersNum   = 0;
   fileDetails->fileParam.ioNamePtr    = fileName;
   theResult = PBGetFInfo(fileDetails, FALSE);

   if(theResult == noErr)
      return(IsListFile(fileName));
   else
      return(TRUE);
}



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 5 May 94 12:10:23 GMT
Path: ucivax!gateway
From: [e-mail] ("Mark Davies, Caddetc, (44) 532 305005" )
Subject: Help
Via: uk.ac.leeds.icf.leva; Thu, 5 May 1994 12:22:01 +0100
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 38
Date: 5 May 94 12:10:23 GMT

Help,

I know this is not anything to do with Mac programming, but my Mac is dying.

Last night it stopped responding, I was debugging some code and it stopped.
I rebooted it, but it has not come alive again.

What happens.

Press Reset.
Smiley Mac comes on
Runs INITS
Loads Finder
Then it stops, before it can display the desktop, the cursor alternates between
the arrow and the watch, this is far as it gets.

Things I've tried

Leaving it of overnight

Rebooting with Disk Utilities :- This runs into the same problem as above

Rebuilding Desk top :- This runs into the same problem

I have reinstalled System 7.1 and after restart I just get the same problem.

It would seem the hard drive is knackered, but I'm at the end of my meagre
mac skills.

Any help would be appreciated thanks,

BTW

Its a Quadra 800 running System 7.1, it has 28 meg of ram and 1/2 G of HD.

Mark Davies
Caddetc
UK
>From [e-mail] 5 May 94 17:19:40 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Getting a filename in a fil
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Pindar Infotek Ltd" at May 5, 94 09:58:05 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 2969
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 67
Date: 5 May 94 17:19:40 GMT

Attn: Ian McCall

> I'm writing a CustomGetFile filter that depends on the file's name
> for its result. This filter takes the ParamBlkPtr it's passed and
> then does a PBGetFInfo. Despite the fact the ParmBlkPtr was passed
> over by CustomGetFile, I then get an OSErr of -43 (fnfErr).
> Can anyone tell me what I'm doing wrong? The code's below.

I can try to tell you what you are doing wrong, but my first reaction
is that you are doing something you don't need to do. The parameter
block to which you are passed a pointer has already been filled out on
your behalf by the Standard File package. You don't need to fill it out
again. 'fileDetails->fileParam.ioNamePtr' already points to the filename
in which you are interested. Your entire filter function could look
like this:

pascal Boolean ListFileFilter (ParmBlkPtr fileDetails, void *dummyData)
{
        return IsListFile (fileDetails->fileParam.ioNamePtr);
}

However, it's possible to assume that you edited your code for posting
and it no longer does what you really want it to do, so I will attempt
to look at it anyway.

> pascal Boolean ListFileFilter(ParmBlkPtr fileDetails, void *dummyData)
> {
>    Str255   fileName;
>    OSErr    theResult;
>
>    fileDetails->fileParam.ioCompletion = NULL;
>    fileDetails->fileParam.ioFDirIndex  = 0;
>    fileDetails->fileParam.ioFVersNum   = 0;
>    fileDetails->fileParam.ioNamePtr    = fileName;
>    theResult = PBGetFInfo(fileDetails, FALSE);
>
>    if(theResult == noErr)
>       return(IsListFile(fileName));
>    else
>       return(TRUE);
> }

The data you are passing to 'PBGetFInfo' has not been fully initialized.
When you put 0 in 'fileDetails->fileParam.ioFDirIndex', you are telling
'PBGetFInfo' that it will be able to find a pointer to the name of the
file in which you are interested in 'fileDetails->fileParam.ioNamePtr'.
Although you have set 'fileDetails->fileParam.ioNamePtr' to point
to a sufficiently (indeed, maximally) sized buffer, you have not
put a filename into that buffer. 'PBGetFInfo' thus has insufficient
information to get info on the file in which you are interested.
In fact, if you are attempting to obtain the filename, setting
'fileDetails->fileParam.ioFDirIndex' to 0 will not help you, because you
then need the filename in order to obtain the filename! :-)

Another note: you are getting a "file not found" error as opposed to a
"volume not found" error because the Standard File package, as I said
above, has already filled in the parameter block on your behalf. Your
code forgets to set 'fileDetails->fileParam.ioVRefNum', but this field
has already been set by the Standard File package, so you get "file
not found" instead of the "volume not found" message which would be
likely if 'fileDetails->fileParam.ioVRefNum' contained the usual stack
or newly-allocated-heap garbage data.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 Symantec C++ a "Bjarne Burner"? Yeah, I'd be pretty burned up, too!
>From [e-mail] 5 May 94 17:32:24 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: Help
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 7
Date: 5 May 94 17:32:24 GMT

Try zapping your p-ram (Hold dowm CMD - OPT - P - R on sys 7 machines) on
restart.  When you hear it restart you can let go of the keys.  You can also
try trashing your finder prefs.  If that does not work, try booting with
extentions off.  If it boots without extentions, it is some kind of init
problem.

-GomollD
>From [e-mail] 5 May 94 19:23:08 GMT
Path: ucivax!gateway
From: [e-mail] (Jerry Gatlin)
Subject: Re: getting desk pattern
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <9405051913.AA12776@hipark>
Approved: [e-mail]
Lines: 35
Date: 5 May 94 19:23:08 GMT

>Jerry, do you know something I don't?  I can't find (in THINK Reference
>anyway) any record of LMGetDeskPattern() or LMGetDeskCPat().  Are these
>utility functions you wrote, am I just missing them, or do you access to
>*secrets*.  :-)
>
>                                                just mike

Mike,

Funny you should ask. Actually, LowMem.h is in Apple's new Universal
Headers. These routines are now the preferred way of accessing any of
Apple's low memory globals. I guess I'm so used to them now, I forget that
a lot of folks aren't using them yet. I'm coding exclusively in CodeWarrior
now. These replace having to use the globals defined in SysEqu.h. And no,
they aren't defined in Think Reference. ARGH!

I looked around in SysEqu for the LowMem equivalants and only came up with:

 DeskPattern = 0xA3C,                   /*[GLOBAL VAR] Pattern with which
desktop is painted (8 bytes)

which I assume corresponds to the B&W DeskPat only. I don't know how you
might access the color PixPatHandle of the color desktop pattern. Any clues
anyone??

Thanks for catching my oversight!

Jerry

____________________________________________________________________
Jerry Gatlin                       [e-mail]
Hill Country Softworks (my other life)                    Austin, TX
____________________________________________________________________


>From [e-mail] 5 May 94 19:50:41 GMT
Path: ucivax!gateway
From: [e-mail] (Tim Schaffer)
Subject: File Structures
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 14
Date: 5 May 94 19:50:41 GMT

Does anyone have any advice (even very basic - I'm in poland and
it's rather difficult for me to get any mac programming books here)
on how to store record and field type information in a file?  Do
I basically have to scan through the whole file to get some basic
identifying info on each record, even if I don't want to read in
the whole record?  Is there any way to encode jumping points for each
record at the beginning of each file?  (I know that this is somewhat
of an advanced question, but although I've programmed fairly significantly
(though not in C), I haven't really done much actual writing to files
(the file intensive programs I've written have been languages that
were specifically designed to accesss database files) and have no
real idea of how different sorts of files are stored on the mac.
Thanks,
Tim
>From [e-mail] 6 May 94 03:40:35 GMT
Path: ucivax!gateway
From: [e-mail] ("Robert S. Mah")
Subject: Re: File Structures
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 42
Date: 6 May 94 03:40:35 GMT

At  7:50 PM 5/5/94 +0000, Tim Schaffer wrote:
> [...]
> on how to store record and field type information in a file?  Do
> I basically have to scan through the whole file to get some basic
> identifying info on each record, even if I don't want to read in
> the whole record?  Is there any way to encode jumping points for each
> record at the beginning of each file?  (I know that this is somewhat
> of an advanced question, but although I've programmed fairly significantly
> (though not in C), I haven't really done much actual writing to files
> (the file intensive programs I've written have been languages that
> were specifically designed to accesss database files) and have no
> real idea of how different sorts of files are stored on the mac.

Think of it this way, the software/languages you used to do database
work were probably written in C.  Therefore, with some sweat, you'll
be able to duplicate anything you've seen before :-)  Of course, it
is easier to build on someone else's work.

What I'm trying to say is that you'll either have to:
   1. Find a record oriented or B-Tree database package for C.
or
   2. Write your own.

I suggest looking for an appropriate database package on the net.  There
are dozens of different ones out there, ranging from simple flat file,
fixed length record access to ones that go whole hog with b-trees and
schema builders.  Since these usually come with complete source code,
you'll be able to modify them to suit.

BTW, you could also use Resource Manager as your database as long as
you understand its extreme limitations.  I know, I know, "the Resource
Manager isn't a database", but it is quick and easy so no nasty
messages please.  If you're aware of its limitations and you might
be OK.

Cheers,
Rob

___________________________________________________________________________
Robert S. Mah  -=-  One Step Beyond  -=-  212-947-6507  -=-  [e-mail]


>From [e-mail] 6 May 94 14:28:14 GMT
Path: ucivax!gateway
From: [e-mail] ("Conrad P. Masshardt")
Subject: Help with includes...
X-Sun-Charset: US-ASCII
Message-ID: <[e-mail]>
content-length: 281
Newsgroups: fa.think-c
X-Message-ID: <9405061358.AA16889@yoda>
Approved: [e-mail]
Lines: 10
Date: 6 May 94 14:28:14 GMT

Meistros,

I am relatively new to C programming and I am wondering what the difference
between #include <thisfile.h> and #include "thisfile.h" is.  How do I get
the compiler to understand the first?  Do I have to precompile?

Thanks in advance.

Conrad.
[e-mail]
>From [e-mail] 6 May 94 22:01:41 GMT
Path: ucivax!gateway
From: [e-mail] (Igor Mikolic-Torreira)
Subject: Re: Help with includes...
NNTP-Posting-Host: alumni.caltech.edu
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Newsreader: NN version 6.5.0 #4 (NOV)
X-Message-ID: <2qeene$[e-mail]>
Approved: [e-mail]
Organization: California Institute of Technology, Pasadena
Lines: 21
Date: 6 May 94 22:01:41 GMT
References: <[e-mail]>

"Conrad P. Masshardt" <[e-mail]> writes:

>I am relatively new to C programming and I am wondering what the difference
>between #include <thisfile.h> and #include "thisfile.h" is.  How do I get
>the compiler to understand the first?  Do I have to precompile?

If you have #include <thisfile.h> the compiler will only look for thisfile.h
in the system libraries.  For THINK C that means libaries in the THINK C (or
better said, the PRoject Manager) folder or sub-folders of that folder.

If you have #include "thisfile.h", the compiler first looks in your local
work area and only if it fails to find it there does it look in the
system libraries.  For THINK C, the local work area is the folder your
project file is in or sub-folders of it.

The THINK manual has very detailed explanations of exactly where
THINK will search for included files.

--
   Igor Mikolic-Torreira
   [e-mail]                (finger for PGP public keys)
>From [e-mail] 6 May 94 23:23:32 GMT
Path: ucivax!gateway
From: [e-mail] (Canna Software Development)
Subject: Think C 7 and BBEdit...
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 17
Date: 6 May 94 23:23:32 GMT

I apologize if this has already been covered, but I am having problems
getting Think C and BBEdit to talk.  Is this a bug that has been solved?
Or do I have something ugly I need to look into?

Thanks in advance...


----------------------------+------------------------------------
Lloyd Sargent               | "Podia est gamia" - Sherlock Holmes
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 7 May 94 15:34:08 GMT
Path: ucivax!gateway
From: [e-mail] (Dana S Emery)
Subject: Re: Help with includes...
Message-ID: <[e-mail]>
In-Reply-To: Your message <[e-mail]> of 6 May 94
 14:28:14 GMT
Content-Type: TEXT/plain; charset=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 38
Date: 7 May 94 15:34:08 GMT

The difference is the search path the compiler uses to loacte the
included file.

for <foo.h> the compiler assumes a system header context and looks
in the "compiler" area for a matching header.

for "foo.h" the compiler assumes a project context and looks in the
project area, but if it fails to find anything there it will then
also look in the system area.

If all includes for some particular file in a library package
(say TCL's CApplication.h) were always included using "", then a
user could implment her own project-specific version of that file
and be sure that it would be used within a given project, but not
outside that project.

The TCL follows that convention thruout, in spite of the fact that it
would be faster to take advantage of its placement inside the compiler
tree by using <>, but by using "" you are not trapped into the TCL's
own version of its headers, thus you can fix bugs or make other
modifications to the TCL source as you need to (with the usual caveats,
always use a copy, always mark up your "improvements", consider using
subclasses in preference to re-writes...).

I have deliberatly simplified the scheme a little to give you the
general idea clearly.  Actually, searches are not always done, but
can be forced (thru the make... dialog); but when done the first
place searched is the parent folder of the .c file being worked on.
You should RTFM to get the whole story, especially the bit about the
search path excluding folders whose names are wraped in parens:
(foo folder) is never searched.  This allows project folders to
include ancilliary and alternate files, excluding them speeds things
up and perhaps avoids confusion as well.

hope that helps
--
dana s emery <[e-mail]>

>From [e-mail] 7 May 94 19:12:40 GMT
Path: ucivax!gateway
From: [e-mail] (John Grimes)
Subject: Alert Memory Problems
Message-ID: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 39
Date: 7 May 94 19:12:40 GMT



        I have two questions I would really appreciate any help on.

        1)  I have several alert boxes that I'm calling in sequential
order.  When I get to the last one it doesn't show the pict thats inside
it.  Just the alert box.  Anyway I found this strange since all the
others work fine.  So I placed that Alert box first in the calling
order.  Now what happens is that the second last box isn't shown but this
the last box is.  Well so I figure its memory problems.  I think I need
to purge the PICTs from the other alerts out of memory.  I tried using
GetPicture to these picts and then RemoveResource but that doesn't seem
to work.  Does anyone have any ideas.  This is probably cake for most of you.

        2)  This one I really don't understand.  I have a dialog box
which in which I would like to put that extra rectangle around the OK
Box.  So straight out of IM is this code.

        GetDItem (PlayerCharacter,dOK, &iType, &iHandle,&iRect);
        PenSize(3,3);
        InsetRect( &iRect, -4,-4);
        FrameRoundRect(&iRect,16,16);

        This draws the rectangle above and to the left of the OK button.
Since I use the Dialog Item number of the OK button in my ModalDialog
function I know thats correct.  Well I'd appreciate yourhelp, thanks.


******************************************************************************
        This was no time for play
        This was no time for fun
        This was no time for games
        There was work to be done.
                        Cat in the Hat Comes Back-Dr. Seuss
******************************************************************************
                                        John Grimes
                                        reaper@jhunix
                                        HAC Consultant

>From [e-mail] 7 May 94 19:52:48 GMT
Path: ucivax!gateway
From: [e-mail] (Joshua Kuperman)
Subject: trivial Sym 7 C++ ? 'bout visual arch & beeper demo
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL21]
Content-Type: text
Content-Length: 791
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 21
Date: 7 May 94 19:52:48 GMT

Subject: trivial Sym 7 C++ ? 'bout visual arch & beeper demo

Okay I got symantec C++ 7.0 for the ed price and saved enought to
get another 8Meg of ram.

I worked through the Beeper app demo using Visual Architect.  I decided
it would be nice to make sure it couldn't beep more than 10 times.
I used Visual Architects features to limit the range in the field, but
it only checks it when I hit OK (a command that was inherited from
someplace) and it ignores the edit when I hit the button for beep (the
one that goes to my code).  Why didn't it inherit something to update
the fields before it called my function.  What do I have to do to mayk
it.



--
Josh Kuperman                  The favorite doesn't always win,
[e-mail]      no matter what the odds. -I. Gershwin


>From [e-mail] 9 May 94 14:31:00 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: Preferences Files
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date: 9 May 94 14:31:00 GMT

I wish to create a preferences file that the user is unable to open or
print from the Finder.  IM6 p9.20-23 suggests:

"Register a signature ... that is different from the signature of your
application, and set this signature as the creator of files that you don't
want your users to open.  This ensures that the Finder displays your
message ['STR ' resource id -16397] instead of launching your application
when the user double-clicks these documents."

Is this the recommended approach?  Is there any way other than custom icon
resources to give such a preferences file its own icon?

_____________________________________________________________________________
Brent Taylor, Computer Science / Math Teacher            Phone: [phone]
Elkford Secondary School                                   Fax: [phone]
Box 910, Elkford, BC, Canada  V0B 1H0
[e-mail]

>From [e-mail] 9 May 94 14:35:09 GMT
Path: ucivax!gateway
From: [e-mail] (Hong Shi)
Subject: Help: Think Reference
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 9 May 94 14:35:09 GMT

I am new to Macintosh programming, please help me to start, thanks ahead.

As my understanding of manual for Symantec c++, the Think Reference
should include reference for ToolBox procedures, but I cannot find it in
my packgae (Version 6.0). And, somewhere in the manual it says I can
just click on a procedure  name and it will find the manual page for me.
I tried but it doesn't work, should I set something?

I am using a Centris 650, System 7.1

Please reply to [e-mail].
>From [e-mail] 9 May 94 15:48:14 GMT
Path: ucivax!gateway
From: [e-mail] ("Mark Davies, Caddetc, (44) 532 305005" )
Subject: Thanks
Via: uk.ac.leeds.icf.leva; Mon, 9 May 1994 15:24:03 +0100
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 13
Date: 9 May 94 15:48:14 GMT

Thanks,

 to all of you that helped. I tried all your suggestions to no avail.

I then got serious and looked in the box, I pressed the SIMM's down ( more
in hope then anything!) and the machine came back to life, I then
had to rebuild the deskstop and reboot without finder re-loading windows.

Everythings fine now,

Thanks for the help.

Mark Davies
>From [e-mail] 9 May 94 16:52:03 GMT
Path: ucivax!gateway
From: [e-mail] (Brian Bezanson)
Subject: RE>Help: Think Reference
Message-ID: <[e-mail]>
X-Mailer: InterCon Dispatcher/SMTP for QuickMail
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
X-Priority: 4
Lines: 28
Date: 9 May 94 16:52:03 GMT

Date    5/9/94
Subject RE>Help- Think Reference
From    Brian Bezanson
To      Hong Shi, think-c

        Reply to:   RE>Help: Think Reference
>I am new to Macintosh programming, please help me to start, thanks ahead.
>
>As my understanding of manual for Symantec c++, the Think Reference
>should include reference for ToolBox procedures, but I cannot find it in
>my packgae (Version 6.0). And, somewhere in the manual it says I can
>just click on a procedure  name and it will find the manual page for me.
>I tried but it doesn't work, should I set something?

The version of THINK Reference is the full version of the application, but it
only includes the ANSI C and C++ Streams databases. To get the databases for
the Mac Toolbox calls you need to purchase the seperate package.

====
Brian R Bezanson                      Phone: [phone]
Management Graphics, Inc.             Fax:   [phone]
1401 E 79th Street                    internet: [e-mail]
Minneapolis, MN 55425                 AppleLink: BEZANSON





>From [e-mail] 9 May 94 17:25:37 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Preferences Files
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Brent Taylor" at May 9, 94 02:31:00 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 993
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 20
Date: 9 May 94 17:25:37 GMT

> Is this the recommended approach? Is there any way other than custom
> icon resources to give such a preferences file its own icon?

Set its bundle bit and go through the whole ritual necessary to create
a BNDL and all other appropriate resources? That's uglier than a custom
resource, but at least the user can't inadvertently change it.

Also consider simply setting the file type and creator to that of one of
the system preferences files. The icon will automatically be the little
radio button charicature we all know and love. With an appropriate
filename, I can't imagine any user confusion, especially since it's
already been done several times.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "...someone not acquainted with the Christian mythology of the
 Crucifixion might consider a Crucifix to be a particularly sadistic
 piece of erotica... We of the ACLU will continue to defend your right to
 worship such objects if it pleases you." -- Gregory J. Wageman
>From [e-mail] 9 May 94 17:30:58 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Help: Think Reference
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Hong Shi" at May 9, 94 02:35:09 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 987
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 19
Date: 9 May 94 17:30:58 GMT

> As my understanding of manual for Symantec c++, the Think Reference
> should include reference for ToolBox procedures, but I cannot find it
> in my packgae (Version 6.0). And, somewhere in the manual it says I
> can just click on a procedure name and it will find the manual page
> for me. I tried but it doesn't work, should I set something?

>From page 29: "If you own THINK Reference... you can get... online
information concerning toolbox calls..." Note the big 'if'. What the
manual means there is *if* you own THINK Reference as a separate
product, you can do the lookups you want to do. THINK Reference *does*
work, and it *is* useful, but the only database you have for use with it
is the database for the ANSI libraries ('fopen' and friends). To get the
other databases, including those for the Mac Toolbox, you have to buy a
separate package.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 You thought Obfuscated C was confusing? Wait for Obfuscated C++!
>From Anderson#m#[e-mail] 9 May 94 18:32:58 GMT
Path: ucivax!gateway
From: Anderson#m#[e-mail] ("Anderson, Tim")
Subject: address to subscribe?
Message-ID: <[e-mail]>
Return-receipt-to: "Anderson, Tim" <Anderson#m#[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 7
Date: 9 May 94 18:32:58 GMT

Could someone please help me by sending the e-mail address where I can
subscribe to the think c mailing list. I was receiving mail, but then my local
mail account died, and since then, I have not received any more (and I lost the
address.)

Thanks in advance,
Tim
>From [e-mail] 9 May 94 20:45:00 GMT
Path: ucivax!gateway
From: [e-mail] ("jim (j.) blakeman" )
Subject: Unsubscribe
X400-Originator: /dd.id=1625593/g=jim/i=j/s=blakeman/@bnr.ca
Content-Identifier: Unsubscribe
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X400-Received:
 by mta bnr.ca in /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed; Mon, 9 May 1994 16:43:49 -0400
X400-Received:
 by /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed; Mon, 9 May 1994 16:42:41 -0400
X400-Received:
 by /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed; Mon, 9 May 1994 16:40:00 -0400
X-Message-ID: <"23383 Mon May  9 16:42:50 1994"@bnr.ca>
Approved: [e-mail]
X400-Content-Type: P2-1984 (2)
Lines: 5
Date: 9 May 94 20:45:00 GMT
X400-MTS-Identifier:
 [/PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/;bcars735.b.374:09.04.94.20.42.41]

unsubscribe

Please unsibscribe me from this list.  Thank-you.

[e-mail]
>From [e-mail] 9 May 94 23:13:47 GMT
Path: ucivax!gateway
From: [e-mail] ("Richard K. Wolf [phone]" )
Subject: AEInstallHandler Question.
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 40
Date: 9 May 94 23:13:47 GMT

Could anyone clue me in on why the following:

/* function declarations ... */

void         InitEventHandlers( void );
pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
                         long refCon);

..

void main (void)
{
   InitToolbox();
   InitEventHandlers();
   ...
}

void InitEventHandlers( void )
{

 OSErr err;

 err = AEInstallEventHandler( kCoreEventClass, kAEOpenApplication,
          DoOpen, 0L, false);
 ...
}

pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
          long refCon)
{
 return noErr;
}

fails to compile claiming the third argument in my AEInstallEventHandler
was not expected from the prototype?

Yours,
C Idiot Extroadinaire


>From [e-mail] 10 May 94 06:44:23 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: re: File Structures
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 21
Date: 10 May 94 06:44:23 GMT

>Does anyone have any advice (even very basic - I'm in poland and
>it's rather difficult for me to get any mac programming books here)
>on how to store record and field type information in a file?
>Do I basically have to scan through the whole file to get some basic
>identifying info on each record, even if I don't want to read in
>the whole record?  Is there any way to encode jumping points for each
>record at the beginning of each file?

Try using a header for each file that contains information about the file,
including a jump table to access specific fields or records.  Set aside a
specific amount at the beginning of the file that contains all the
information you want, including the number of fields, the length of each
field, etc.  The next portion of the file could then contain a jump table to
indicate the location of the records (assuming you want variable length
records).  You'll have to maintain the jump table as the file is edited, or
simply set it up at write time.  If you have wildly divergent field types,
you can also set aside a space in the file for field descriptors.

Good luck,

duff caldewey
>From [e-mail] 10 May 94 06:55:57 GMT
Path: ucivax!gateway
From: [e-mail] (John Grimes)
Subject: OpenDeskAcc Problems
Message-ID: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 21
Date: 10 May 94 06:55:57 GMT



        For a while I had the OpenDeskAcc function working properly, then
suddenly it stopped working.  Now the desk accesories aren't opened until
after I quit the application.  Or to explain, they are still given the
open command yet control of the computer is not given to them until after
I quit my application.  Since none of my recent changes have anything to
do with this function I'm kind of stumped.  I'd appreciate any help that
you can give me.  Thank you.

******************************************************************************
        This was no time for play
        This was no time for fun
        This was no time for games
        There was work to be done.
                        Cat in the Hat Comes Back-Dr. Seuss
******************************************************************************
                                        John Grimes
                                        reaper@jhunix
                                        HAC Consultant

>From [e-mail] 10 May 94 11:42:41 GMT
Path: ucivax!gateway
From: [e-mail] (John Lawrie)
Subject: unsubscribe
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 58
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 3
Date: 10 May 94 11:42:41 GMT

please reomve me from this list.
Thank You
[e-mail]
>From [e-mail] 10 May 94 12:18:22 GMT
Path: ucivax!gateway
From: [e-mail] ("Jeffrey K. Carney")
Subject: Re: AEInstallHandler Question.
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 6
Date: 10 May 94 12:18:22 GMT

Did you try passing the ADDRESS of the function instead of just the name?
Strict prototyping might require a function pointer.

Jeff


>From [e-mail] 10 May 94 12:44:23 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: (none)
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 1
Date: 10 May 94 12:44:23 GMT

help
>From [e-mail] 10 May 94 13:19:04 GMT
Path: ucivax!gateway
From: [e-mail] (Dave Sugar)
Subject: Reading Serial Port?
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 48
Date: 10 May 94 13:19:04 GMT


  I am working on a project where I need to read and write data from the
serial port.  I have the basic functions for acessing the serial port
don.  But, while writing some more advanced stuff I'm having some problems.

  What I am trying to do is to read a line (deliminated by a new line or
return) from the serial port.  The code that I have used to read from the
serial port is as follows:

long DCSerial::Read (char data[], long length)
{
   IOParam   param;
   OSErr     err = 0;
   long      avail;

// checks to make sure port open

// checks for characters avail, store in avail

   memset (¶m, 0, sizeof (IOParam));
   param.ioRefNum = in_refNum;          // read from .AIn
   param.ioBuffer = data;               // read into data
   param.ioReqCount = avail;            // read as many chars as available
   param.ioPosMode = fsAtMark;          // read from mark

   err = PBRead ((ParmBlkPtr) ¶m, FALSE);

Now, this code works just fine and it reads everything that is in the port
buffer into the variable data with no problem.  What I would like to do is
to read only up to the first new-line character.  Looking in Think Ref it
seems that this would be possible.  By chaning the line param.ioPosMode etc.
to something like:

  param.ioPosMode = fsAtMark | 0x80 | (256 * '\n');

  But when I do this I don't get any differecne in the data that is read in.
It will still read in as much data as it can (size of avail).  The mode
that I am trying to set with in the above like seems to be a PBRead thing
not specific to serial ports.  I have not yet tried it with a file to see
if I can get it to work that way.

  Any help would be greatly appreciated..
  Thanks
  Dave Sugar
  [e-mail]
  [e-mail]


>From [e-mail] 10 May 94 14:53:34 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: File Structures
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "[e-mail]" at May 10, 94 06:44:23 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1877
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 34
Date: 10 May 94 14:53:34 GMT

> Try using a header for each file that contains information about the
> file, including a jump table to access specific fields or records. Set
> aside a specific amount at the beginning of the file that contains all
> the information you want, including the number of fields, the length
> of each field, etc. The next portion of the file could then contain a
> jump table to indicate the location of the records (assuming you want
> variable length records). You'll have to maintain the jump table as
> the file is edited, or simply set it up at write time. If you have
> wildly divergent field types, you can also set aside a space in the
> file for field descriptors.

One trick worth mentioning is a basic technique for allocating
variable-length chunks. At some point, one needs to be able to read a
constant-length block and decide what to do with it. Since sometimes
even the index to a file can be variable-length, you don't want to put
it at the head of the file. If you did, you'd constantly be shuffling
other data out of the way in order to grow the index.

The trick: store only a 32-bit unsigned offset at the beginning of the
file. This offset points you to the index, which lives at the end of the
file. You can expand the end of the file all you like without shuffling
data out of the way. If you keep the index in memory while modifying
the file (common practice to enhance performance), you can even append
non-index data to the file, overwriting the disk representation of the
index, and then rewrite the index onto the (new) end of the file again.

Of course, you could always use the resource fork for the index and rely
on the Resource Manager to shuffle chunks on your behalf, but that's not
a C technique but a Macintosh technique.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "...where they burn books, people are next." -- XTC
>From [e-mail] 10 May 94 17:54:32 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: AEInstallHandler Question.
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Jeffrey K. Carney" at May 10, 94 12:18:22 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 459
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 10 May 94 17:54:32 GMT

> Did you try passing the ADDRESS of the function instead of just the
> name? Strict prototyping might require a function pointer.

The name of a function is a synonym for a pointer to that function,
just like the name of an array is a synonym for a pointer to the first
element of the array. I'd wager this is not his problem.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "Reality is 50 million polygons per second." -- Alvy Ray Smith
>From [e-mail] 10 May 94 17:54:34 GMT
Path: ucivax!gateway
From: [e-mail] (Dana S Emery)
Subject: Re: File Structures
Message-ID: <[e-mail]>
In-Reply-To: Your message <[e-mail]> of 10 May 94
 06:44:23 GMT
Content-Type: TEXT/plain; charset=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 65
Date: 10 May 94 17:54:34 GMT

>Does anyone have any advice (even very basic - I'm in poland and
>it's rather difficult for me to get any mac programming books here)
>on how to store record and field type information in a file?

this isnt really a mac problem, the mac file system imposes no
structure on the content of a data fork, this means you get to
determine whatever structure is used there.

There are a number of conventional structures in common use, the
file type code will be set to reflect their usage: EPSF, TIFF,
TEXT, PICT...  Most applications have unique binary file content
with appropriate file type codes, yours will presumably fit into
this category.  Apple has a registration program for your
applications signature, it may also have one for file types, Im
not sure on that tho.

There are a number of clever ways to structure complex binary
information in a file, files on tape have the difficulty that random
access usage is very expensive timewise, and rewriting parts of a
tape can be problematical, so structures used for tape are often
different from structures used for disk, luckily you are unlikely
to have to deal with tape, so no more need be said about that.

One common approach on disk is to have a preliminary record much
like a table of contents in a book.  Commonly it is a structure
with fields each of which gives the value of the mark needed to
read a given structure, and perhaps also the length of that
structure (which may be variable).

Future growth of your code would normally entail modifications
to this structure, so knowledge of it should be limited to the
units that deal with file I/O.  I like to work with objects, I
would use an object to encaposulate that code in a way that frees
the rest of your code from specific details.  It real hard to
make detailed recomendations without in-depth knowledge, and
really that should not be necessary, you have to be able to
cope with the debugging (yes, there will be problems..), so
you need to do enough of the design work that you are comfortable
with the design.

I would start with some primitive utilitys, you probably need a
central routine which understands how to put an arbitrary byte
field to the file, it should be fed by routines which have
progressivly more knowledge of the types you are actually using.

Higher level routines would ensure that the "table of contents"
gets filled in properly, and that it gets written out properly.

Reading in your data has to be feasible, and supplying an initial,
empty version of the file is also necessary, get these 3 coded and
tested and you will have proven your design.

The Table of Contents should contain some declaration of version
level to allow for expansion, and, finally, you should give some
thought to invalid states and how your code should react to them
(corrupt files).

It is commonly said that code flows easily from a good data
specification, so if you find the code flowing slowly rethink how
you are dealing with your data.

good luck.
--
dana s emery <[e-mail]>

>From [e-mail] 10 May 94 18:42:26 GMT
Path: ucivax!gateway
From: [e-mail] (Robert Story)
Subject: Re: Reading Serial Port?
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Dave Sugar" at May 10, 94 01:19:04 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 785
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 14
Date: 10 May 94 18:42:26 GMT

>   But when I do this I don't get any differecne in the data that is read in.
> It will still read in as much data as it can (size of avail).  The mode
> that I am trying to set with in the above like seems to be a PBRead thing
> not specific to serial ports.  I have not yet tried it with a file to see
> if I can get it to work that way.
>
Why not just have two read routines - one that reads all available chars
into a buffer, then a second that reads up to the first newline in the
buffer (calling the first routine if the buffer gets empty).  You have to
make sure all your read routines use the buffer, though.  But it can be done.

--
|   Robert Story    |  Software Engineer  |#include <std/disclaimer>|
|  [e-mail]   |  Atlanta, Georgia   |   War Damn Eagle! '90   |
>From [e-mail] 10 May 94 21:38:28 GMT
Path: ucivax!gateway
From: [e-mail] (WaldorfM)
Subject: subscribe request
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 3
Date: 10 May 94 21:38:28 GMT

help
subscribe Markus Waldorf

>From [e-mail] 10 May 94 21:53:59 GMT
Path: ucivax!gateway
From: [e-mail] (Danny Thomas)
Subject: Re: Reading Serial Port?
X-Sender: [e-mail].
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 13
Date: 10 May 94 21:53:59 GMT

>  I am working on a project where I need to read and write data from the
>serial port.  I have the basic functions for acessing the serial port
>don.  But, while writing some more advanced stuff I'm having some problems.
when faced with serial port I/O I took the route of using the Comms Toolbox
instead of low-level drivers or accessing hardware. Your needs mightn't be
met by this, but it was easier than I had hoped. And that's speaking as a
very part-time mac programmer who was glad that he didn't have to write a
dialog box a simple as the serial tool's

cheers,
Danny Thomas


>From [e-mail] 10 May 94 23:01:44 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: C2PStr
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 5
Date: 10 May 94 23:01:44 GMT

Where the Heck is the library that contains the C2PStr and the P2CStr
routines? The headers are defined in "pascal.h" but it won't link without the
libraries(of coarse). Some suggested the MacTraps2 library but that one isn't
it, or if it is I have something else wrong like link order or code in wrong
segments if that's possible.
>From [e-mail] 10 May 94 23:41:55 GMT
Path: ucivax!gateway
From: [e-mail] (Canna Software Development)
Subject: Re: AEInstallHandler Question.
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 104
Date: 10 May 94 23:41:55 GMT

>Could anyone clue me in on why the following:
>
>/* function declarations ... */
>
>void         InitEventHandlers( void );
>pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
>                         long refCon);
>
>...
>
>void main (void)
>{
>   InitToolbox();
>   InitEventHandlers();
>   ...
>}
>
>void InitEventHandlers( void )
>{
>
> OSErr err;
>
> err = AEInstallEventHandler( kCoreEventClass, kAEOpenApplication,
>          DoOpen, 0L, false);
> ...
>}
>
>pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
>          long refCon)
>{
> return noErr;
>}
>
>fails to compile claiming the third argument in my AEInstallEventHandler
>was not expected from the prototype?
>
>Yours,
>C Idiot Extroadinaire


Well, I took the above code (having messed with AppleEvents quite heavily
as of late) and compared it with my code. Nope. Couldn't see anything
obvious.  So I took the above code and stuffed it into a file named main.c.
Below is the code:

#include <AppleEvents.h>

/* function declarations ... */

void         InitEventHandlers( void );
pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
                         long refCon);


void main (void)
{
//   InitToolbox();
   InitEventHandlers();
}

void InitEventHandlers( void )
{

 OSErr err;

 err = AEInstallEventHandler( kCoreEventClass, kAEOpenApplication,
          DoOpen, 0L, false);
}

pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
          long refCon)
{
 return noErr;
}


I then set the compiler to the require prototypes (explicit) and check
pointer types (just to be on the save side).  Guess what?  It compiled.  In
fact I couldn't get it to so much as WHINE about DoOpen UNLESS i removed
the #include <AppleEvents.h>.  Then it gets really huffy and complains
about a syntax error on the prototype (since AppleEvent is no longer
defined, it is to be expected).

So what is the solution?  Well, it is obvious that either there is
something else going on (that was deleted for brevity) or your compiler is
way out of whack.  I suggest you copy the above code and see if you can
compile it.  Let me know what you get, I am curious now as to how it worked
just fine for me and seems to blow up on you!

Good luck,

Lloyd


----------------------------+------------------------------------
Lloyd Sargent               | "Podia est gamia" - Sherlock Holmes
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 11 May 94 03:24:37 GMT
Path: ucivax!gateway
From: [e-mail] (Dave Sugar)
Subject: Re: Reading Serial Port?
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 23
Date: 11 May 94 03:24:37 GMT


  Hmm, ok, seeing some of the responces maybe I can clarify my self just
a little bit more.  When I was trying to ioPosMode I did try both to
deliminate with a  '\r' and '\n' character.  They both seemed to work the
same in returning everything that was in the buffer.

  I have thought about making a second buffer that would just read in the
data and then I would read from that buffer to give the program what it
needs.  This seems like the easiest method right now, though I would like
to see if I can get the ioPosMode to work correctly.  It seems a little
easier.

  As for using the CommToolbox, again I thought of that, but the code that
I am writing is for an interface for a very specific machine.  There is
only going to be communication at a max of 9600 baud and of text data
only.  I don't think it would be worth delving into CommToolbox for this,
though I may be wrong, I should look into the book..

  Thanks for all the responces, more ary welcome.
  Dave Sugar
  [e-mail]
  [e-mail]

>From [e-mail] 11 May 94 14:13:34 GMT
Path: ucivax!gateway
From: [e-mail] (Canna Software Development)
Subject: Re: AEInstallHandler Question.
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 104
Date: 11 May 94 14:13:34 GMT

>Could anyone clue me in on why the following:
>
>/* function declarations ... */
>
>void         InitEventHandlers( void );
>pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
>                         long refCon);
>
>...
>
>void main (void)
>{
>   InitToolbox();
>   InitEventHandlers();
>   ...
>}
>
>void InitEventHandlers( void )
>{
>
> OSErr err;
>
> err = AEInstallEventHandler( kCoreEventClass, kAEOpenApplication,
>          DoOpen, 0L, false);
> ...
>}
>>
>pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
>          long refCon)
>{
> return noErr;
>}
>
>fails to compile claiming the third argument in my AEInstallEventHandler
>was not expected from the prototype?
>
>Yours,
>C Idiot Extroadinaire


Well, I took the above code (having messed with AppleEvents quite heavily
as of late) and compared it with my code. Nope. Couldn't see anything
obvious.  So I took the above code and stuffed it into a file named main.c.
Below is the code:

#include <AppleEvents.h>

/* function declarations ... */

void         InitEventHandlers( void );
pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
                         long refCon);


void main (void)
{
//   InitToolbox();
    InitEventHandlers();
}

void InitEventHandlers( void )
{

    OSErr err;

    err = AEInstallEventHandler( kCoreEventClass, kAEOpenApplication,
          DoOpen, 0L, false);
}

pascal OSErr DoOpen( AppleEvent *theAppleEvent, AppleEvent *reply,
          long refCon)
{
    return noErr;
}


I then set the compiler to the require prototypes (explicit) and check
pointer types (just to be on the save side).  Guess what?  It compiled.  In
fact I couldn't get it to so much as WHINE about DoOpen UNLESS i removed
the #include <AppleEvents.h>.  Then it gets really huffy and complains
about a syntax error on the prototype (since AppleEvent is no longer
defined, it is to be expected).

So what is the solution?  Well, it is obvious that either there is
something else going on (that was deleted for brevity) or your compiler is
way out of whack.  I suggest you copy the above code and see if you can
compile it.  Let me know what you get, I am curious now as to how it worked
just fine for me and seems to blow up on you!

Good luck,

Lloyd


----------------------------+------------------------------------
Lloyd Sargent               | "Podia est gamia" - Sherlock Holmes
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 11 May 94 17:22:37 GMT
Path: ucivax!gateway
From: [e-mail] (Bernie Bernstein)
Subject: anyone use ASLM?
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 16
Date: 11 May 94 17:22:37 GMT

I have been doing some work with the Apple Shared Library Manager, and now
I'm just looking for anyone else who has any experience with it. I haven't
tried compiling any Think C code with it yet, but it is supposed to be
possible. Is there anyone out there who has tried it?

Is it more of a hassle than just using it in MPW?

Thanks,


    o,  ,,   ,      | Bernie Bernstein                    | ,    ,,
    L>O/  `,/ `    ,| University of Colorado at Boulder   |/ `,,/  `
   O./  '  / . `, / | home: 303-444-3168                  |     / ` `  ,.
  ,/   /  ,      '  | email: [e-mail]      | /        ''  `


>From [e-mail] 11 May 94 18:29:04 GMT
Path: ucivax!gateway
From: [e-mail] ("jim (j.) blakeman" )
Subject: unsubscribe request
X400-Originator: /dd.id=1625593/g=jim/i=j/s=blakeman/@bnr.ca
Content-Identifier: unsubscribe r...
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X400-Received:
 by mta bnr.ca in /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed; Wed, 11 May 1994 14:28:34 -0400
X400-Received:
 by /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed; Wed, 11 May 1994 14:28:21 -0400
X400-Received:
 by /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed; Wed, 11 May 1994 14:26:00 -0400
X-Message-ID: <"2615 Wed May 11 14:28:25 1994"@bnr.ca>
Approved: [e-mail]
X400-Content-Type: P2-1984 (2)
Lines: 2
Date: 11 May 94 18:29:04 GMT
X400-MTS-Identifier:
 [/PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/;bcars735.b.614:11.04.94.18.28.21]

unsubscribe [e-mail]

>From [e-mail] 11 May 94 19:08:14 GMT
Path: ucivax!gateway
From: [e-mail] (Jean-Marie Laborde)
Subject: Bug in atan2()?
X-Sender: laborde@isis
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 58
Date: 11 May 94 19:08:14 GMT

It "seems" that atan2, which has many advantages over atan(),
gives an erroneous answer when its second argument is exactly 0:

/****************************/
#include <math.h>
#define PI             3.14159265358979323846264
main()
{
    double  x=.0001,
            y=1.0,
            atp0001, atm0001, atp0, atm0,
            fact=PI/180;                  /* forcing result in degree */

     atp0001=atan2(y,x)/fact;
     atm0001=atan2(y,-x)/fact;

     x=.0;
     atp0=atan2(y,x)/fact;
     atm0=atan2(y,-x)/fact;

     printf(" %G %G %G %G", atp0001,atm0001,atp0,atm0);
}
/* output is:
   89.9943 90.0057 90 -90
     making atan2 a non continous function...

   89.9943 90.0057 90 90  would be expected!
*/
/****************************/


/* Note: the source for atan2 in math.c ( copyright Symantec) is
double atan2(double y, double x)
{
        double z;

        DomainCheck(x == 0 && y == 0, 0);
        z = _atan(y / x);
        if (x < 0) {
                if (y < 0)
                        z -= Pi;
                else
                        z += Pi;
        }
        return(z);
}
It is no wonder that atan2(1,-0) will be evaluated as -PI/2 and not PI/2  */


****************************************************************************
Jean-Marie Laborde                                       [e-mail]
Responsable du Projet IMAG Cabri-geometre
Laboratoire de Structures Discretes et de Didactique - IMAG
Universite Joseph Fourier - CNRS          BP 53 38041 Grenoble cedex 9 FRANCE
(33) 76 51 46 10   (sec 76 51 46 16)      (fax [phone])
****************************************************************************


>From [e-mail] 11 May 94 20:25:21 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: Handling Save As... Command
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 169
Date: 11 May 94 20:25:21 GMT

I am struggling in my efforts to develop an appropriate algorithm for
implementing the Save As... command from the File menu.  My source code
(portions omitted) is included below.

Some of my particular concerns are as follows:

1. If after closing the previous file (assuming it exists), what do I do if
my attempt to create or open the new file fails?  What name should appear
in the window (old fsSpec.name, new fsSpec.name, "\pUntitled")?

2. How do I handle errors from FSClose() when closing the previous file.
It seems that I should proceed to open the new file and save the window's
data, otherwise it can't be saved.  However, in what state does this leave
the old file?

3. I would appreciate any comments on my error handling approach.  My
approach is to abort the process, clean up the mess, and report the error
at the first error encountered.


/************************** My source ***************************/

OSErr HandleSaveCommand(void);
OSErr HandleSaveAsCommand(void);
OSErr SaveDataFile(DataRecHandle myData);
OSErr SaveTempFile(FSSpecPtr tempSpec, DataRecHandle myData);
OSErr GetTempSpec(FSSpecPtr tempSpec);
OSErr SaveMyData(short fRefNum, DataRecHandle myData, long *fileLength);


/*
    HandleSaveCommand() handles the Save command from the File menu.
*/
OSErr HandleSaveCommand()
{
    DataRecHandle  myData;
    OSErr          errNo;

    myData = (DataRecHandle) GetWRefCon(FrontWindow());
    if ((**myData).fRefNum == 0)
        errNo = HandleSaveAsCommand();
    else
        errNo = SaveDataFile(myData);
    return(errNo);
}


/*
        HandleSaveAsCommand() handles the Save As... command from the File menu.
*/
OSErr HandleSaveAsCommand()
{
    WindowPtr          myWindow;
    DataRecHandle      myData;
    Str255             promptStr;
    Str255             nameStr;
    StandardFileReply  sfReply;
    OSErr              errNo;

    myWindow = FrontWindow();
    myData = (DataRecHandle) GetWRefCon(myWindow);
    GetIndString(promptStr, idMiscStr, iSaveAsStr);
    CopyString(nameStr, (**myData).dataSpec.name);
    if ((errNo = ResError()) == noErr) {
        StandardPutFile(promptStr, nameStr, &sfReply);
        if (sfReply.sfGood) {
            /* close the current file */
            if ((**myData).fRefNum != 0) {
                FSClose((**myData).fRefNum);
                (**myData).fRefNum = 0;
            }

            /* open the new file */
            if (sfReply.sfReplacing)
                errNo = FSpCreate(&sfReply.sfFile, 'ROCK', 'Rock',
                                          smSystemScript);
            if (errNo == noErr)
                if ((errNo = FSpOpenDF(&sfReply.sfFile,
                                          fsRdWrPerm, &fRefNum)) == noErr) {
                    (**myData).fRefNum = fRefNum;
                    (**myData).dataSpec = sfReply.sfFile;
                    SetWTitle(myWindow, sfReply.sfFile.name);
                    errNo = SaveDataFile(myData);
                }
        } else
            errNo = userCanceledErr;
    }
    return(errNo);
}


/*
    SaveDataFile() saves the data file using the safe save method.
*/
OSErr SaveDataFile(DataRecHandle myData)
{
    FSSpec  dataSpec;
    FSSpec  tempSpec;
    OSErr   errNo;

    dataSpec = (**myData).dataSpec;
    if ((errNo = GetTempSpec(&tempSpec)) == noErr)
      if ((errNo = FSpCreate(&tempSpec, 'ROCK', 'Rock',
                                                smSystemScript)) == noErr) {
        if ((errNo = SaveTempFile(&tempSpec, myData)) == noErr)
          if ((errNo = FSpExchangeFiles(&tempSpec, &dataSpec)) == noErr)
            return(FSpDelete(&tempSpec));
        FSpDelete(&tempSpec);
      }
    return(errNo);
}


/*
    SaveTempFile() saves the data to the temporary file.
*/
OSErr SaveTempFile(FSSpecPtr tempSpec, DataRecHandle myData)
{
    long    fileLength;
    short   fRefNum;
    OSErr   errNo;

    fileLength = 0;
    if ((errNo = FSpOpenDF(tempSpec, fsRdWrPerm, &fRefNum)) == noErr) {
      if ((errNo = SetFPos(fRefNum, fsFromStart, 0)) == noErr)
        if ((errNo = SaveMyData(fRefNum, myData, &fileLength)) == noErr)
          if ((errNo = SetEOF(fRefNum, fileLength)) == noErr)
            if ((errNo = GetVRefNum(fRefNum)) == noErr)
              if ((errNo = FlushVol(fRefNum)) == noErr) {
                if ((errNo = FSClose(fRefNum)) == noErr)
                  (**myData).dirty = FALSE;
                return(errNo);
              }
      FSClose(fRefNum);
    }
    return(errNo);
}


/*
    GetTempSpec() returns the file specification of a unique temporary file.
*/
OSErr GetTempSpec(FSSpecPtr tempSpec)
{
    long    time;
    short   vRefNum;
    long    parID;
    Str255  name;
    OSErr   errNo;

    if ((errNo = FindFolder(fsSpec.vRefNum, kTemporaryFolderType,
                                kCreateFolder, vRefNum, parID)) == noErr) {
        do {
            GetDateTime(&time);
            NumToString(time, name);
            errNo = FSMakeFSSpec(vRefNum, parID, name, tempSpec);
        } while (errNo == noErr);
        if (errNo == fnfErr)
            errNo = noErr;
    }
    return(errNo);
}


__________________________________________________________________________
Brent Taylor, Computer Science / Math Teacher         Phone: [phone]
Elkford Secondary School                                Fax: [phone]
Box 910, Elkford, BC, Canada  V0B 1H0               [e-mail]

>From [e-mail] 11 May 94 22:40:22 GMT
Path: ucivax!gateway
From: [e-mail] (Stan Warford)
Subject: Printing PostScript
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.3 PL11]
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 21
Date: 11 May 94 22:40:22 GMT

This may not be quite the right forum for this question, but does
anyone know of a Mac utility that will print a postscript file that
I recieved over the net.  I want to print it to my LaserWriter.
The beginning of the file looks like this:

--1073741863-633813925-758409756:#4036
Content-Type: APPLICATION/PostScript; name="po.ps"
Content-Transfer-Encoding: BASE64
Content-ID: <[e-mail]>
Content-Description: po.ps attached (56541 byte postscript file)

JSENCgQNCiUlQ29tbWVudDogVGhlIEN0cmwtRCBhYm92ZSBpcyB0byBlbnN1
cmUgam9icyB3b24ndCBqYW0NCiUlICBpbnNpZGUgdGhlIHByaW50ZXIgYW5k
IG1lc3MgdXAgdGhlIFZNIG1hbmFnZW1lbnQuDQovRHZpVG9QUyBzYXZlIGRl
Zg0KJSVDcmVhdGlvbkRhdGU6IFdlZCBKYW4gMTIgMTE6MjI6MTYgMTk5NA0K

and so on for about 1200 more lines.

Thanks
Stan Warford
[e-mail]
>From [e-mail] 12 May 94 14:47:23 GMT
Path: ucivax!gateway
From: [e-mail] (Clinton R Lefort)
Subject: compiling 4.0 to 6.0 errors
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 41
Date: 12 May 94 14:47:23 GMT


Hello think-c'ers


Hello,   In a program that I am compiling  in THINK C 6.0, originally written in 4.0, I get these error messages during compilation time:


1. File tc_files.c; Line 15
Error:   pointer types do not match

static  char *dSizeStr = "\psample-size";

2.File aiff.c; Line 120
Error:   second argument to function write does not match prototype
          || write(fd, &comm1,sizCommChunk1) != sizCommChunk1

3. File tc_cursor.c; Line 31
Error:   argument to function HNoPurge does not match prototype
        HNoPurge(&waitCursHand);

4.File tc_misc.h; Line 2
Error:   invalid redefinition of macro TRUE
#define TRUE    (!FALSE)

5. File tc_get_path.c; Line 57
Error:   invalid redeclaration of getHFSPath
OSErr getHFSPath(path,fname,vrefnum)

6.File winmac.c; Line 149
Error:   third argument to function NewWindow does not match prototype

        csWindow = NewWindow(0L, &windowBounds, title, true, documentProc, -1L, true, 0);

There were over 16 bugs during compilation; I was able to fix 10 of them with my limited programming skills, and these are the remaining one.  I would appreciate any insights and helpful corrections.

Thank you

CRL

Please feel free to comment on any  of theabove as time warrants you!!!

>From [e-mail] 12 May 94 16:41:03 GMT
Path: ucivax!gateway
From: [e-mail] ("Griffin, Ed E")
Subject: RE:Printing Postscript
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 15
Date: 12 May 94 16:41:03 GMT

Stan,

     Two things:

1) It appears to me that the file has been encoded in some way.  PostScript,
by definition is a human readable language.  Your file may have been BinHexed
(did it have an extension like .hqx when you got it?)  If this is the case,
then you need to decode it before sending to the printer.

2) You can get a great and free PS download program from Adobe!  It is called
PSTool and is available from ftp.adobe.com.  Its operation is self evident.

                           Hope this helps,
                           Ed Griffin
                           [e-mail]
>From [e-mail] 12 May 94 16:46:55 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Handling Save As... Command
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Brent Taylor" at May 11, 94 08:25:21 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 3192
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 63
Date: 12 May 94 16:46:55 GMT

Let me start by saying I am very pleased to hear someone in the universe
is taking error checking seriously.

> 1. If after closing the previous file (assuming it exists), what
> do I do if my attempt to create or open the new file fails? What
> name should appear in the window (old fsSpec.name, new fsSpec.name,
> "\pUntitled")?

Here's the tricky part: don't close the old file. You should design (or
rearrange) your routines so that they don't care which file they write
to. Then you can attempt to create, open, write, and close the new file
without disturbing the old one. After the new file is successfully
saved, *then* close the old one and change the window title. (In a
perfect world, you'd be able to [easily] check whether there was enough
memory to change the window title.) If an error occurs writing the new
file, you can just delete the new file, pretend the whole thing never
happened, and leave the document the way it was.

> 2. How do I handle errors from FSClose() when closing the previous
> file. It seems that I should proceed to open the new file and save the
> window's data, otherwise it can't be saved. However, in what state
> does this leave the old file?

Well, I think the previous answer handles most of these concerns.
However, there is the issue of what to do about a file that won't close.
If there were no buffering involved, I'd say to simply call DebugStr
if ever you get an error closing a file. Why? Because closing a file
*should* be a trivial operation that shouldn't fail unless your program
has a bug. However, closing a file in reality is not as trivial an
operation as one might like, because buffers have to be flushed, etc.
Does anybody out there have a good strategy?

> 3. I would appreciate any comments on my error handling approach. My
> approach is to abort the process, clean up the mess, and report the
> error at the first error encountered.

This seems reasonable in the general case; the tricky part is doing
things in the right sequence (as described above) so that you can safely
employ this tactic.

> OSErr GetTempSpec(FSSpecPtr tempSpec)

There is a problem in this routine. If it's called more than once in a
second, it will produce a duplicate filename. Also, if more than one
program uses this algorithm and a context switch occurs during that
second, it's possible two different programs could be competing for
access to the same file.

Rather than forcing this routine to last at least a second, which slows
down the system, and rather than using TickCount instead of GetDateTime,
since a 60th of a second can also be a surprisingly long time on a fast
machine, allow me to suggest you rely on another algorithm, possibly
one which empirically verifies the filename is unique by creating the
file with progressive filenames in some sort of loop until a file is
successfully created.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "...someone not acquainted with the Christian mythology of the
 Crucifixion might consider a Crucifix to be a particularly sadistic
 piece of erotica... We of the ACLU will continue to defend your right to
 worship such objects if it pleases you." -- Gregory J. Wageman
>From [e-mail] 12 May 94 16:59:20 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: compiling 4.0 to 6.0 errors
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Clinton R Lefort" at May 12, 94 02:47:23 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 2253
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 64
Date: 12 May 94 16:59:20 GMT

> 1. File tc_files.c; Line 15
> Error:   pointer types do not match
>
> static        char *dSizeStr = "\psample-size";

Pascal strings (note the "\p") are unsigned. The declaration (and
others like it) should read:

        static unsigned char *dSizeStr = "\psample-size";

> 2.File aiff.c; Line 120
> Error:   second argument to function write does not match prototype
>         || write(fd, &comm1,sizCommChunk1) != sizCommChunk1

Probably can't help you there; not enough context. However, I can guess
that the problem is that 'comm1' is not a (signed) character, which is
what 'write' expects.

> 3. File tc_cursor.c; Line 31
> Error:   argument to function HNoPurge does not match prototype
>       HNoPurge(&waitCursHand);

If 'waitCursHand' is a 'CursHandle', then you've just found a bug
lurking in this program which has been there for a while. If it's
a 'Handle', then the problem is the same but the fix is different.

CusrHandle  solution: 'HNoPurge ((Handle) waitCursHand);'
Handle      solution: 'HNoPurge (waitCursHand);'

If 'waitCursHand' is neither of the above things, you'll have to post
more context.

> 4.File tc_misc.h; Line 2
> Error:   invalid redefinition of macro TRUE
> #define TRUE  (!FALSE)

Just comment this buzzard out. It's a valid macro, but TRUE is already
defined in the standard headers.

> 5. File tc_get_path.c; Line 57
> Error:   invalid redeclaration of getHFSPath
> OSErr getHFSPath(path,fname,vrefnum)

That looks like an old-style declaration conflicting with a previously
declared prototype. Update the declaration to include type names and see
what happens.

> 6.File winmac.c; Line 149
> Error:   third argument to function NewWindow does not match prototype
>
>       csWindow = NewWindow(0L, &windowBounds, title, true, documentProc, -1L, true, 0);

Probably 'title' is a (signed) pointer to character. The best thing to
do would be to fix the declaration of 'title', but the program will
probably compile if you simply do:

   csWindow = NewWindow (0,&windowBounds,(StringPtr)title,true,
      documentProc,(WindowPtr)-1,true,0);

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "Bedrock does support a form of drag and drop. Development dragged on
 for about two years, and then was dropped." -- Brian Clark
>From [e-mail] 12 May 94 17:42:18 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: DA's in System 7
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 29
Date: 12 May 94 17:42:18 GMT

In System 7, is it possible for a desk accessory to be launched in the
application heap?  I seem to recall reading somewhere that a DA can be
launched into the application heap with the Command or the Option key or
something.

The real question:  Do I require the following code in a System 7 dependent
application?


/*
        HandleCloseCommand() handles the Close command from the File menu.
*/
OSErr HandleCloseCommand(Boolean quitting)
{
        WindowPtr       theWindow;
        OSErr           errNo;

        theWindow = FrontWindow();
        if ((refNum = ((WindowPeek) theWindow)->windowKind) < 0)
                CloseDeskAcc(refNum);
        else...
}


__________________________________________________________________________
Brent Taylor, Computer Science / Math Teacher         Phone: [phone]
Elkford Secondary School                                Fax: [phone]
Box 910, Elkford, BC, Canada  V0B 1H0               [e-mail]

>From [e-mail] 12 May 94 22:43:32 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: DA's in System 7
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Brent Taylor" at May 12, 94 05:42:18 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 2958
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 63
Date: 12 May 94 22:43:32 GMT

> In System 7, is it possible for a desk accessory to be launched in the
> application heap?  I seem to recall reading somewhere that a DA can be
> launched into the application heap with the Command or the Option key or
> something.

Yes and no. The option key trick no longer works. However, Desk
Accessories can still be installed into the app's resource fork and the
system will launch them into the application heap. You could assume
that anybody who installs a desk accessory into your app's resource
fork should get whatever punishment occurs when your programs fails to
accomodate the DA, but there are other considerations.

> The real question: Do I require the following code in a System 7
> dependent application?
>
> /*
>         HandleCloseCommand() handles the Close command from the File menu.
> */
> OSErr HandleCloseCommand(Boolean quitting)
> {
>         WindowPtr       theWindow;
>         OSErr           errNo;
>
>         theWindow = FrontWindow();
>         if ((refNum = ((WindowPeek) theWindow)->windowKind) < 0)
>                 CloseDeskAcc(refNum);
>         else...
> }

This is a good question. Let's assume that you want to punish people who
modify your app's resource fork by insalling a DA into it. There are
worse crimes in the world, but what the heck, you only live once. Under
this assumption, I'd say you could skip this business of checking the
windowKind field.

One the other hand, remember that a DA is a driver. DAs are allowed to
open windows not because they are DAs but because they are drivers --
drivers can open windows. Drivers are the super-set of DAs. Now, you
might think, why on earth would a driver that wasn't a DA want to open a
window? Well, consider a program like MultiClip. This is a clipboard and
scrapbook management utility. It's implemented partially in a driver,
not a DA, which opens a window. Its window shows up with a windowKind
less than 0. I can't remember precisely the details, but let's say it's
a hot-key sequence which causes MultiClip's driver to open its window.
No Apple Menu access needed.

Now you have this driver window in your app's window list, and it would
seem that the only thing to do is to call CloseDeskAcc against it.
But it's not a DA! Does CloseDeskAcc handle this situation? A cursory
disassmbly of CloseDeskAcc (Q800/Sys7.1) seems to indicate that it does
not handle this situation; it simply calls Close.

This is a good question. I had never thought about it before. In any
case, I don't think it's safe quite yet to completely ignore the
possibility that there will be a driver window in your window list.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 'It seems the firm contracted by Intel to produce the famed "flying
 Pentium" ads could not complete the 3-D modeling on a PC before
 deadline. So in the best know-your-enemy tradition, they chose the next
 best thing -- a Quadra 840AV.'                 -- Mac The Knife 4/18/94
>From [e-mail] 13 May 94 01:50:26 GMT
Path: ucivax!gateway
From: [e-mail] (Robert Pellerin)
Subject: File name
Message-ID: <[e-mail]>
X-Mailer: Eudora F1.4.2
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 24
Date: 13 May 94 01:50:26 GMT

I am learning C and working on my first program. So far I love C much
better than BASIC but I have to put a file mane in a DLOG static field.

I have a StandardFileReply (gSourceFile) in a global and need to put the
file name in a Str255.

The gSourceFile.sfFile.name is a Str63. How do I convert that in a Str255
so I'll be able to use: SetIText( iHandle, theFileName )?

Typecast to a Str255 doesn't work, and simply passing one pointer to the
ohter one give stranges results

It might look obvious for somebody, but mot for me.

Thanks!


 -------------------------------------------------------------
 * Robert Pellerin - Montreal, Qc Canada ([e-mail]) *
 -------------------------------------------------------------
 *     res.: [phone]    *     BBS:  [phone]    *
 *     bur.:                   *     Programmation sur Mac   *
 *     fax.:                   *     ouvert 24 hres / jour   *

>From [e-mail] 13 May 94 02:06:04 GMT
Path: ucivax!gateway
From: [e-mail] (Reuben Gellman)
Subject: Address of hypercard newsgroup
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL22]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 245
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 8
Date: 13 May 94 02:06:04 GMT

Hello:
Anybody out there know the **address** of the (a?) hypercard discussion
group? (I mean e-mail address, not just usenet comp.sys.....hypercard.)
Feel free to save bandwidth & e-mail me directly (see below).

Reuben Gellman
[e-mail]

>From [e-mail] 13 May 94 08:14:52 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Symantec C++ RISC Compiler
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 5
X-Mts: smtp
Date: 13 May 94 08:14:52 GMT

Does anybody know something about the Symantec C++ compiler for Power-Mac's ?


                                Lukas

>From [e-mail] 13 May 94 11:51:39 GMT
Path: ucivax!gateway
From: [e-mail] (Thomas Bridgwater)
Subject: Machine differences
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1182
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 22
Date: 13 May 94 11:51:39 GMT

Hello --

I wrote a little program called 'Quiet Start' which shuts the volume to
0 just before shut down and back to it's previous setting during the
INIT load. This forces some newer macs not to make the startup 'bong'.
Problem is, it doesn't have the desired effects on all macs. I realize
that there's nothing I can really do about it as the code that makes the
sound is executed from ROM before the INITs load, but I fugure I could
be nice and recognize when it's not going to work. Is there some sort of
list available that shows such details of the behaviour of various mac
models? Or, is there someone I can get free info from at Apple who can
tell me which macs do what? Another solution... I'm sure that the among
the readers of this list there are many different mac models available.
Would you mind checking what YOUR mac does by setting the volume to 0
and restarting. I'd like to know if the bong happens, and if so at what
volume level (i.e. some or all Quadras sound at level 7 if set to 0)

In addition, where can I get a current list of Mac models and the
appropriate machine specific Gestalt values?

     Thanks for your troubles --
           -- Tom Bridgwater
>From [e-mail] 13 May 94 17:29:49 GMT
Path: ucivax!gateway
From: [e-mail] (Chris Metcalfe)
Subject: Re:  Symantec C++ RISC Compiler
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 22
Date: 13 May 94 17:29:49 GMT

In the upgrade announcement that I received, Symantec offered a
"Power Macintosh Cross Development Kit":

        With this new add-in translator for Symentec C++ for
        Macintosh, you can compile your 68K Macintosh applications
        for Apple's new Power Macintosh.  Complete with FREE tools
        from Apple, the Cross Development kit provides the easiest
        migration of your programs to the Power Macintosh.

        Just $100*

        *Requires Symantec C++ 7.0 for Macintosh

So for a hundred dollars you get some free software!!  No mention
is made of any API extensions, so I assume the add-in module is
a script that simply translates a few application framework
conventions and toolbox calling conventions.  This means that
your code will be executing in the 68K emulation mode of the
PowerPC.

-- Chris Metcalfe

>From [e-mail] 13 May 94 19:39:09 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: File name
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Robert Pellerin" at May 13, 94 01:50:26 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1926
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 35
Date: 13 May 94 19:39:09 GMT

> The gSourceFile.sfFile.name is a Str63. How do I convert that in a
> Str255 so I'll be able to use: SetIText( iHandle, theFileName )?

The quick answer is that you don't. Both a Str63 and a Str255 are
generally Pascal strings. This means that the first byte holds the
number of subsequent bytes. So, for a Pascal string "\pfoo", the first
byte will be a 3. This definition of Pascal string is the only thing
that 'SetIText' cares about. You can pass 'gSourceFile.sfFile.name'
directly to 'SetIText'. From a code maintenance standpoint, it's
probably not the best thing to do, but at this stage of the game it's
probably not in your interest to worry overmuch about that.

There are several issues involved in why the compiler allows you to do
this. Note that 'Str63' and 'Str255' and the other StrXXX types are
defined simply as unsigned arrays of character. The trick is that when
it comes to function definitions and prototypes and calls, arrays don't
really exist. They're treated basically as pointers. When you desribe
or call a function interface which takes an array as a parameter,
think of the compiler passing only a pointer to the first element of
the array. In the situation we're discussing, that would be a pointer
to an unsigned character in the cases of both 'Str63' and 'Str255'.
So for function calls they're pretty much interchangeable as far as
the compiler is concerned. Any good C book covers this in far more
exhaustive detail than I have here.

You do have to know whether the function you are calling is assuming
that all 255 bytes are there. 'SetIText' only reads as many bytes as are
numbered by the first byte of the array, so it's safe to pass any Pascal
string, even a literal like "\pfoo".

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "The human race will decree from time to time: 'There is something at
 which it is absolutely forbidden to laugh.'" -- Nietzche on C++
>From [e-mail] 13 May 94 20:24:48 GMT
Path: ucivax!gateway
From: [e-mail] (Matthew Xavier Mora)
Subject: Re:  Symantec C++ RISC Compiler
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 40
Date: 13 May 94 20:24:48 GMT

>In the upgrade announcement that I received, Symantec offered a
>"Power Macintosh Cross Development Kit":
>
>        With this new add-in translator for Symentec C++ for
>        Macintosh, you can compile your 68K Macintosh applications
>        for Apple's new Power Macintosh.  Complete with FREE tools
>        from Apple, the Cross Development kit provides the easiest
>        migration of your programs to the Power Macintosh.
>
>        Just $100*
>
>        *Requires Symantec C++ 7.0 for Macintosh
>
>So for a hundred dollars you get some free software!!  No mention
>is made of any API extensions, so I assume the add-in module is
>a script that simply translates a few application framework
>conventions and toolbox calling conventions.  This means that
>your code will be executing in the 68K emulation mode of the
>PowerPC.


Well I've been playing with it all morning and it does indeed create
native PowerPC. It works quite well amazingly enough. I'm trying to get a
program to work that uses stdio and its not cooperating but other programs
that I compiled worked without hassle. Its not as fast as MetroWorks but its
usable.

Matt


__________________________________________________________________
Matthew Xavier Mora  (cybernaut)            The keeper of the UMPG
SRI International                              [e-mail]
[Sent using Eudora 2.0.2       ]              [e-mail]
[Using my brand new 8100/80av !]
 "Thanks to Apple Computers, it's Bitchin what these things can
  do." Tori Amos
__________________________________________________________________


>From [e-mail] 13 May 94 20:40:05 GMT
Path: ucivax!gateway
From: [e-mail] (Rich Siegel)
Subject: Re: Symantec C++ RISC Compiler
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Chris Metcalfe" at May 13, 94 05:29:49 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1679
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 40
Date: 13 May 94 20:40:05 GMT

>
> In the upgrade announcement that I received, Symantec offered a
> "Power Macintosh Cross Development Kit":
>
>       With this new add-in translator for Symentec C++ for
>       Macintosh, you can compile your 68K Macintosh applications
>       for Apple's new Power Macintosh.  Complete with FREE tools
>       from Apple, the Cross Development kit provides the easiest
>       migration of your programs to the Power Macintosh.
>
>       Just $100*
>
>       *Requires Symantec C++ 7.0 for Macintosh
>
> So for a hundred dollars you get some free software!!  No mention
> is made of any API extensions, so I assume the add-in module is
> a script that simply translates a few application framework
> conventions and toolbox calling conventions.  This means that
> your code will be executing in the 68K emulation mode of the
> PowerPC.

With all due respect, you're wrong.

The add-in translator is, in fact, a translator, just like the C, C++,
and Rez compilers. It's a C++ compiler which uses the same front-end
code as the 7.0 C++ compiler, and whose back end emits PowerPC code,
which is packaged in the XCOFF format. The "free tools" that are
supplied are the necessary Apple tools which are necessary to build a
working application from XCOFF files: PPCLink, MakePEF, MakeSYM, and
others. (A couple of other tools are supplied as well, but aren't
relevant to this discussion.)

Applications built with the cross-development kit are real live
PowerPC executables, not emulated. The CDK also includes universal
headers (though they may already be a part of the paid 7.0 upgrade - I
don't recall for certain), and native versions of the standard C and
C++ libraries and Mac OS and Toolbox glue.

R.

>From [e-mail] 13 May 94 20:55:47 GMT
Path: ucivax!gateway
From: [e-mail] (Brian Bezanson)
Subject: RE>Re:  Symantec C++ RISC Compiler
Message-ID: <[e-mail]>
X-Mailer: InterCon Dispatcher/SMTP for QuickMail
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
X-Priority: 4
Lines: 25
Date: 13 May 94 20:55:47 GMT

Date    5/13/94
Subject RE>Re-  Symantec C++ RISC C
From    Brian Bezanson
To      Chris Metcalfe, think-c

        Reply to:   RE>Re:  Symantec C++ RISC Compiler
I have the PowerPC Symantec Kit. The "FREE" stuff is the Apple Linker and
Debugger for PowerPC. You uses a different translator, instead of "Symantec C++
" you use something like "Symantec PowerPC C++" to compile your code.

Nothing runs in Emulation unless you call something or use a code fragment
that is emulated.

Brian

====
Brian R Bezanson                      Phone: [phone]
Management Graphics, Inc.             Fax:   [phone]
1401 E 79th Street                    internet: [e-mail]
Minneapolis, MN 55425                 AppleLink: BEZANSON





>From [e-mail] 13 May 94 22:29:44 GMT
Path: ucivax!gateway
From: [e-mail] (Will Glass-Husain)
Subject: what's the deal with version 7.0
Message-ID: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
content-length: 0
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 4
Date: 13 May 94 22:29:44 GMT


Has anyone heard anything about version 7.0 of Think C?

-->Will Glass-Husain
>From [e-mail] 14 May 94 04:41:46 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: what's the deal with version 7.0
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Will Glass-Husain" at May 13, 94 10:29:44 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 385
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 14 May 94 04:41:46 GMT

> Has anyone heard anything about version 7.0 of Think C?

It's still the same old reliable Kahl compiler from the 5 era. Still
great, if you're satisfied with Object C. :-)

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 C   credo: "Trust the programmer."
 C++ credo: "Put the muzzle of a shotgun in your mouth and
             give the programmer the trigger..."
>From [e-mail] 14 May 94 20:54:50 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: I is outta here
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
MIME-version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Organization: Indiana University of Pennsylvania
X-VMS-To: NETMAIL%"[e-mail]"
Approved: [e-mail]
Lines: 6
Date: 14 May 94 20:54:50 GMT
X-Envelope-to: [e-mail]


Please remove me from this list.

Sorry if this is the wrong way to do this.

Jeff Bertrand
>From [e-mail] 15 May 94 02:35:28 GMT
Path: ucivax!gateway
From: [e-mail] (Mauricio Tavares)
Subject: Reading data from a file
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.3 PL2]
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 9
Date: 15 May 94 02:35:28 GMT

        For a code I am writing, some configuration parameters are in a
text file that is read by the program.  What I would be doing is reading
some doubles and integers and strings into a struct.  I know how to do that
using the standard C routines (fopen(), fscanf(), and so on), but how can I
do that using the Mac file routines?

If someone wants me to be more precise, I can upload my standard C version
of the code.

>From [e-mail] 15 May 94 15:54:22 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Reading data from a file
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Mauricio Tavares" at May 15, 94 02:35:28 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 2465
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 44
Date: 15 May 94 15:54:22 GMT

> reading some doubles and integers and strings into a struct.

This turns out to be non-trivial. You wouldn't think so to look at the
facilities supported by most languages, but there's a reason almost
every language has those facilities. Text files are actually pretty
tedious (if not difficult) to read in a robust manner.

   (Standard C file i/o totally fails in the robsutness department --
   try reading a float into an integer and see what happens. I wouldn't
   use those calls unless I could guarantee that a human would never
   touch the files. And if that's the case, a lot of the reasons for
   text files go out the window.)

The basic idea is that you read everything in as strings first, then
convert the strings to numbers. You start with strings because you
(probably) want to eliminate the white space around each number. What
I have done in the past is write a routine which will read the file up
to a maximum number of bytes, and I pass it a maximum of 30 because I
know 30 is bigger than the number of digits in any numeral (as opposed
to number) I will want to read. This routine is called by one of a
set of routines which expect to read a short or a long or a float
or whatever. The second routine analyzes what it got from the first
routine to eliminate the white space, and resets the file pointer to one
position beyond the end of the numeral. (After reading 30 bytes, the
file pointer is probably positioned well beyond the end of the numeral.)
Then it calls another routine to convert the string to a number.
For simple cases, StringToNum will work (but is slow, since it is a
trap). There are also some SANE (library only?) calls which should
help you. Optimally, you should be WorldScript-savvy when you do these
conversions, but that may be beyond the scope of your program.

This is basically a slow method, but there are any number of ways to
enhance its performance. Double-buffering your file i/o helps a lot.
Reading one byte at a time until you encounter a non-digit is definitely
slower. Combining the two techniques is faster than the technique I
outlined above, but slower than simply adding double-buffering to the
technique I outlined above.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 '"We believe the customer goes to retail stores to buy Microsoft
 software," said Ed Belleba, director of retail strategy at Microsoft.'
                                                     -- MacWEEK 4/18/94
>From [e-mail] 16 May 94 09:32:57 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: Choosing a folder with Cust
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 93
Date: 16 May 94 09:32:57 GMT


>From Pindar user Ian McCall

Subject: _Choosing a folder with CustomGetFile
I've put together a couple of routines wich allow the user to select
a folder using a GetFile dialogue box. However, weird effects are
being experienced on certain Macintoshes.

What's happening is that should a user click on any folder in the
list, the selection bar resets itself to the top of the list again -
in effect give a 'springloaded' look to the whole thing.

This bug is intermittent on some Macintoshes, and non-existent on the
others (the majority). Unfortunately, my development Mac is one where
the problem never occurrs so I'm finding it very difficult to track
the problem down.

Since the file filter I've passed to CustomGetFile works (only
directories are shown), and since CustomGetFile is simply a Toolbox
call, I believe the problem would be in my modal dialogue event
filter.

Here's the code. The values 'ChooseCurrentFolder' and
'ChooseSelectedFolder' are contants referring to dialogue items.


pascal short SelectOnlyFoldersDialogueHook(short whichItem,
                      DialogPtr theDialogue, void *theData)
{
   static Boolean   selectButtonState = TRUE;
   Handle           itemHandle;
   short            itemType;
   Rect             itemRect;


   if(GetWRefCon(theDialogue) == sfMainDialogRefCon)
   {
      switch(whichItem)
      {
         case sfHookNullEvent:
            if((EqualString(theData, "\p", FALSE, FALSE))
             && selectButtonState)
            {
               GetDItem(theDialogue, ChooseSelectedFolder,
                               &itemType, &itemHandle, &itemRect);
               HiliteControl((ControlHandle)itemHandle, 255);
               selectButtonState = FALSE;
            }
            else if((!(EqualString(theData, "\p", FALSE, FALSE)))
                                        && (!(selectButtonState)))
            {
               GetDItem(theDialogue, ChooseSelectedFolder,
                               &itemType, &itemHandle, &itemRect);
               HiliteControl((ControlHandle)itemHandle, 0);
               selectButtonState = TRUE;
            }
         break;

         case ChooseCurrentFolder:
            fsCurrentFolderSelected = TRUE;
            return(sfItemOpenButton);
         break;

         case ChooseSelectedFolder:
            return(sfItemOpenButton);
         break;

         default:
            return(whichItem);
         break;
      }
   }
}


Sorry about the length of that - the original's stuffer with comments
too which makes it even longer.

Now - as far as I can see I am at no stage interfering with the
actual selection, which means I'm at a loss to explain the bug. Could
it be a Macintosh Toolbox problem, or is it (more likely) my fault
after all?


Cheers,
      Ian



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 16 May 94 13:32:26 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: strncmp
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 28
Date: 16 May 94 13:32:26 GMT


>From Pindar user Ian McCall

Subject: _strncmp
Has anyone had any trouble with 'strncmp' in THINK C7? I don't know
about earlier versions, but under 7 the call

   strncmp((char *)"Jefferson", (char *)"Jef", (size_t)3)

(taken from the THINK Reference example) returns a result of 1. From
empirical testing, strncmp seems to be returning whether one string
is larger than another, and not correctly comparing characters.

I may not be the world's greatest Mac wizard, but I -am- aware of how
to use calls such as strncmp!

Any clues?


Cheers,
       Ian



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 16 May 94 17:51:51 GMT
Path: ucivax!gateway
From: [e-mail] (Mark Nagel)
Subject: ADMIN: bounced mail messages
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 10
Date: 16 May 94 17:51:51 GMT

If anyone on the list receives a bounce message because a member
delivery fails, please send a message to
[e-mail].  The problem is usually due to an
incorrectly configured mail system at the delivery site, and I have
no way of knowing unless I'm told by one of you.  I already know
about the problems at powershare.itd.umich.edu, so no need to alert
me on that one, but if you receive any others, please let me know...

Thanks,
Mark
>From [e-mail] 16 May 94 20:28:39 GMT
Path: ucivax!gateway
From: [e-mail] (Stan Warford)
Subject: Thanx for PS tip
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.3 PL11]
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 17
Date: 16 May 94 20:28:39 GMT

Thanks to:

Danny Thomas, Don Gillaspie, Alex Knight, Matthew E. Shafer
Vince Paragano, L Sargent, Jim Wintermyre, [e-mail],
Donald R. Fredkin, Alan D., Greg Vaughn, Barry_Wolman,
Dana S Emery, Robert S. Mah, Griffin, Ed E, Joshua Kuperman

for helping me solve my postscript printing problem.  As several
pointed out, the file was attached to an email message that required
a MIME compatible mail reader.  I discovered our local pine mailer,
which is MIME compatible and was able to decode the BASE64 attached
file, and print it with drop*PS, a utility that several suggested
I could find on sumex.  Worked great.  Thanks for the tips.

Stan Warford
[e-mail]

>From [e-mail] 17 May 94 01:02:34 GMT
Path: ucivax!gateway
From: [e-mail] (barnett aaron)
Subject: #include problems
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 31
Date: 17 May 94 01:02:34 GMT

im getting a syntax error when i include a certain
header file more than once.
ive got all the #ifndef stuff in.
the error happens in a typdef statement.
the header is sort of like this:

 #ifndef HEADER_H
 #define HEADER_H

 #include <systemFoo.h>
 #include "localFoo.h"

 typedef struct
 {
   int stuff;               <-syntax error
   void *moreStuff;
 } aThing;

 extern int anInteger;

 void aProtoType(void);

 #endif /* HEADER_H */

so...
what do you think?


\.
aaron <[e-mail]>
"It's a safety mechanism, the mind shuts off."
>From [e-mail] 17 May 94 01:42:37 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: Reading data from a file.
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 25
Date: 17 May 94 01:42:37 GMT


>What I would be doing is reading
>some doubles and integers and strings into a struct.  I know how >to do that
>using the standard C routines (fopen(), fscanf(), and so on), >but how can I
>do that using the Mac file routines?

You can do this, I do it myself.  The UNIX-style routines (fopen(), etc) need
to be replaced with the Mac-style calls.  I'd suggest the FspXXX calls, but
the FSXXX calls will work fine too.

The Mac calls will read and write data in the same way the UNIX-style calls
will. In fact, on the Mac, the UNIX-style calls are implemented with the Mac
calls.  You'll use FSRead(), and FSWrite() to read and write data and
corresponding calls to open and close.  See the File Manager chapters of
Inside Mac for more details on the calls.

For configuration parameters you may want to consider a resource.  This is
usually a natural thing to write or read a struct from.  Information on doing
this is in the Resource Manager chapter of Inside Mac.

Malcolm Teas
"Relax, it's only ones and zeros..."
Teas Technology - Custom software for the Mac and Newton.
Internet: "[e-mail]" AppleLink or AOL: "mhteas"

>From [e-mail] 17 May 94 09:53:39 GMT
Path: ucivax!gateway
From: [e-mail] (Jan Hoeglund - ABB Corporate Research Sweden)
Subject: C programming style
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 17 May 94 09:53:39 GMT

>From my experience, disagreements over programming style have been
a primary obstacle to teams attempting to work closely together.
One problem is indentation and brace placement because it greatly
changes the appearance of the code. Careful visual layout of the
code is important. It makes the difference between a program that
anyone can read and one that everyone want to rewrite.
Are there any C programming style guidelines out there? Any books?

Jan Hoeglund
[e-mail]
>From [e-mail] 17 May 94 11:03:42 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: List Manager
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 25
Date: 17 May 94 11:03:42 GMT


>From Pindar user Ian McCall

Subject: _List Manager
How do I stop the List Manager from compressing text when displaying
a list?

I've got data which is sometimes slightly too long to fit in the cell
size I've given, and the List Manager is trying to be helpful by
changing the character spacing and making it fit. Trouble is, I'm
trying to make things line up consistently and this helpfulness is
getting in the way...

Thanks in advance for any information.


Cheers,
       Ian



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 17 May 94 12:47:39 GMT
Path: ucivax!gateway
From: [e-mail] (Jens-Uwe Mager)
Subject: Re: List Manager
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 47
Date: 17 May 94 12:47:39 GMT


The only help I have found is to shorten the strings myself. I do put
en ellepsis and the end of the string to say it is truncated. The
following is an excerpt from an LDEF I am using:

/*
 * Draw string in width pixels. If the string is longer truncate on the
 * right side and show an ellepsis.
 */
void rAbbrDraw(s, width)
Str255 s;
int width;
{
        register int curWidth;

        width -= CharWidth(' ');
        if ((curWidth = StringWidth(s)) < width)
                DrawString(s);
        else {
                register int len = s[0];
                register u_char *p = s+len;
                curWidth += CharWidth('\311'); /* option-. (...) */
                while (len && curWidth > width) {
                        len--;
                        curWidth -= CharWidth(*p);
                        p--;
                }
                DrawText(s+1, 0, len);
                DrawChar('\311');
        }
}

I have also a variant that truncates from the left, e.g. if your focus is
on the end of the string.

Best Regards,
Jens-Uwe Mager

HELIOS Software GmbH
Lavesstr. 80
30159 Hannover
Germany

Phone:          [phone]
FAX:            [phone]
AppleLink:      helios.de       Attn: Jens-Uwe Mager
Internet:       [e-mail]
>From [e-mail] 17 May 94 13:45:13 GMT
Path: ucivax!gateway
From: [e-mail] (Olof Morell)
Subject: Re:  C programming style
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Charset: ASCII
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 8
Date: 17 May 94 13:45:13 GMT
X-Char-Esc: 29

There is a nice book:
C elements of style by Steve Oualline (C & C++),
1992, ISBN 0-13-148222-X, M&T Publishing

Some of it is just cosmetics, but there are many great tips
on how to avoid evil traps with #include, etc.

Olle
>From [e-mail] 17 May 94 15:49:13 GMT
Path: ucivax!gateway
From: [e-mail] (Chris Webster)
Subject: Re: C programming style
X-Sender: [e-mail] (Unverified)
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date-warning: Date header was inserted by AC.DAL.CA
Date: 17 May 94 15:49:13 GMT

>Are there any C programming style guidelines out there?


ftp from cs.utoronto.ca:doc/programming

>ihstyle is the "Indian Hill C Style and Coding Standards" from Bell
>Labs, "as amended for U of T Zoology" by Henry Spencer.

>pikestyle contains Rob Pike's notes on Programming in C, some comments on
>programming style for C.

>ten-commandments contains Henry's Ten Commandments for writing C code.



--
[e-mail]

>From [e-mail] 17 May 94 16:41:32 GMT
Path: ucivax!gateway
From: [e-mail] (WalrathW)
Subject: RE: List Manager
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 35
Date: 17 May 94 16:41:32 GMT


>
_______________________________________________________________________________
> From: Pindar Infotek Ltd on Tue, May 17, 1994 14:24
> Subject: List Manager
> To: [e-mail]
>
>
> From Pindar user Ian McCall
>
> Subject: _List Manager
> How do I stop the List Manager from compressing text when displaying
> a list?
>
> I've got data which is sometimes slightly too long to fit in the cell
> size I've given, and the List Manager is trying to be helpful by
> changing the character spacing and making it fit. Trouble is, I'm
> trying to make things line up consistently and this helpfulness is
> getting in the way...
>
> Thanks in advance for any information.

You could try using different columns to line up the data, or you might try
using a mono-spaced font. Neither of these suggestions addresses the problem of
the text getting squashed. This might not happen with a mono-spaced font, but
you might try making the cell width longer than the list width. If this works,
it would mean that some of your text would get chopped in a rather unsightly
manner.

The ultimate solution to your problem is to write a custom LDEF. Sample code
abounds on the Internet and other places.

-wayne

[e-mail]
>From [e-mail] 17 May 94 17:46:56 GMT
Path: ucivax!gateway
From: [e-mail] (Lance Helsten)
Subject: Re: C programming style
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date: 17 May 94 17:46:56 GMT

> >From my experience, disagreements over programming style have been
> a primary obstacle to teams attempting to work closely together.
> One problem is indentation and brace placement because it greatly
> changes the appearance of the code. Careful visual layout of the
> code is important. It makes the difference between a program that
> anyone can read and one that everyone want to rewrite.
>
> Are there any C programming style guidelines out there? Any books?
>

"Writing solid code" is a fairly good book on programming style, though it
does not cover {} and indents. That is really a personal or team choice.
There are other issues which are more critical to solid code than code
formatting.

Lance


>From [e-mail] 17 May 94 17:48:31 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: List Manager
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Pindar Infotek Ltd" at May 17, 94 11:03:42 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1011
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 21
Date: 17 May 94 17:48:31 GMT

Attn: Ian McCall

> How do I stop the List Manager from compressing text when displaying
> a list? I've got data which is sometimes slightly too long to fit in
> the cell size I've given, and the List Manager is trying to be helpful
> by changing the character spacing and making it fit. Trouble is, I'm
> trying to make things line up consistently and this helpfulness is
> getting in the way...

In your call to 'LNew', try setting 'cellSize.h' to something larger
than 'rView.right - rView.left'. THINK Reference documents the method
used by the List Manager to determine the (height and) width if you pass
0,0 for 'cellSize'; you could augment that algorithm.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 'It seems the firm contracted by Intel to produce the famed "flying
 Pentium" ads could not complete the 3-D modeling on a PC before
 deadline. So in the best know-your-enemy tradition, they chose the next
 best thing -- a Quadra 840AV.'                 -- Mac The Knife 4/18/94
>From [e-mail] 17 May 94 20:16:08 GMT
Path: ucivax!gateway
From: [e-mail] (Will Glass-Husain)
Subject: Think C version 7.0
Message-ID: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
content-length: 187
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 8
Date: 17 May 94 20:16:08 GMT


Does anyone knoe when version 7.0 is going to come out?  I want to
upgrade (I have version 5.0).  Should I upgrade to 6.0 or wait for 7.0?

Thanks,

Will Glass-Husain
[e-mail]
>From [e-mail] 18 May 94 16:35:55 GMT
Path: ucivax!gateway
From: [e-mail] (Francesco Stiffoni)
Subject: Strange things after debugging
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 16
Date: 18 May 94 16:35:55 GMT

I'm using THINK C++ 7.0 recently upgraded. Trying some TCL examples with
THINK debugger on something strange happens. Randomly, but more often when
I put breakpoints, after exiting debugging, the call of traps _loadSeg and
_addDrive (probably patched by debugger) activates MacsBug (if loaded, else
TRAP not found error). Inspecting the code reveals _Debugger traps put
before and around the PC. The same thing happened with version 6.0, but
after version 6.0.1 disappeared. All the matter is annoying, please give me
some help.


Francesco Stiffoni
University of Venice
Linguistic Department
Laboratory of Computational Linguistic


>From [e-mail] 18 May 94 17:21:38 GMT
Path: ucivax!gateway
From: [e-mail] (Chris Metcalfe)
Subject: Re: Symantec C++ RISC Compiler
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 8
Date: 18 May 94 17:21:38 GMT

OK --

The Symantec PowerPC cross-development kit does output native
PowerPC executables.  Thanks for the many corrections.  Now
I know first-hand what happens when one 'assumes' ...

-- Chris Metcalfe

>From [e-mail] 18 May 94 22:06:51 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Programing Partner in D.C.
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 9
Date: 18 May 94 22:06:51 GMT

I live in Washington D.C. area and was looking for someone one was interested
in collaborating on writing some games for publication, either shareware or
though a distributor. I have started on a pinochole program and an
ultima/king's bouty type game.

If anyone is intrested e-mail me directly

                                                        Don

>From [e-mail] 18 May 94 22:51:46 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: List Manager
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 33
Date: 18 May 94 22:51:46 GMT

>
______________________________________________________________________________
_
> From: Pindar Infotek Ltd on Tue, May 17, 1994 14:24
> Subject: List Manager
> To: [e-mail]
>
>
> From Pindar user Ian McCall
>
> Subject: _List Manager
> How do I stop the List Manager from compressing text when displaying
> a list?
>
> I've got data which is sometimes slightly too long to fit in the cell
> size I've given, and the List Manager is trying to be helpful by
> changing the character spacing and making it fit. Trouble is, I'm
> trying to make things line up consistently and this helpfulness is
> getting in the way...
>
> Thanks in advance for any information.

Sounds to me like you have one of the utility programs installed which
automatically compresses the file names in open and close dialog boxes.  I
believe this is a feature of NOW Utilities and others.

If the file names in the open and close dialog boxes also compress when you
have long file names when you are outside of the C programming enviroment -
this is probably the cause.

Dennis Jones
[e-mail]

>From [e-mail] 19 May 94 02:50:37 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Problem w/ TCL C++ DEMO
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 16
Date: 19 May 94 02:50:37 GMT

Hello,
When I tried to run TCL Demo(any demo that is writte in C++), I got this
error:

Error: Prefix "Symantec C++"; line 1
Different configuration for precompiled header

The line contains: #include <TCLHeaders++>

I had no problem if I compile/run TCL demo, C object, but if I try to
compile/run  any TCL demo C++, then I get this error.
Does anyone know what the problem is?
Thanks in advance.

[e-mail]

>From [e-mail] 19 May 94 08:32:49 GMT
Path: ucivax!gateway
From: [e-mail] (Craig Hollabaugh)
Subject: templates and overloading bug
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 35
Date: 19 May 94 08:32:49 GMT

Folks,

I looked around for information concerning my problem but found none. Is
there a published bug list for the 7.0 version of C++ compiler? Anyway ...

I'm trying to overload the "<<" operator while using templates. After some
fooling around for a couple hours, I figured out that the C++ compiler has
a problem generating code for this situation (the linker then can't find my
overloaded function). Looking for code examples, I stumbled on the Symantec
Vector Demo, which is exactly the same as the project I'm working on.

If I use the 6.01 version of C++ everything works fine, but if I use 7.0
and recompile the vector demo, the linker can't find the overloaded "<<"
functions. Here are the 7.0 linker errors from the vector demo.

undefined: operator <<(ostream&,vector<char>&) (main.cp)
undefined: operator <<(ostream&,vector<myDate>&) (main.cp)
undefined: operator <<(ostream&,vector<float>&) (main.cp)
undefined: operator <<(ostream&,vector<int>&) (main.cp)


The 7.0 disassembler doesn't produce code for this line in vector.cp
(actually vector<XXX>.cp).

template<class T> ostream& operator<<(ostream& s,  vector<T>& v)




Can someone verify this problem and possibly offer a work around?

Thanks,
craig


>From [e-mail] 19 May 94 11:31:12 GMT
Path: ucivax!gateway
From: [e-mail] (Robert Ward)
Subject: ResEdit STR templates
Via: uk.ac.cambridge.mrc-applied-psychology; Thu, 19 May 1994 12:00:07 +0100
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 14
Date: 19 May 94 11:31:12 GMT


I want to have some easily editable STR# resources in which the STR's
themselves may contain TAB characters. Unfortunately the STR editor in
ResEdit uses TAB, shift-TAB, option-TAB, etc to move between fields.
Putting "\t" within the STR seems to print literally, "\t". I can insert
tabs with the hex editor, but I was hoping there was a more user-friendly
solution.

Robert Ward
MRC Applied Psychology Unit           phone:  [phone] ext 602
15 Chaucer Road                       fax:    [phone]
Cambridge CB2 2EF, UK                 e-mail: [e-mail]


>From [e-mail] 19 May 94 14:05:31 GMT
Path: ucivax!gateway
From: [e-mail] (barnett aaron)
Subject: small icons in lists
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 5
Date: 19 May 94 14:05:31 GMT

can anyone point me to a ldef that does the small icons along with text.

\.
aaron <[e-mail]>
"It's a safety mechanism, the mind shuts off."
>From [e-mail] 19 May 94 16:01:52 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: ResEdit STR templates
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Robert Ward" at May 19, 94 11:31:12 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 920
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 19
Date: 19 May 94 16:01:52 GMT

> I want to have some easily editable STR# resources in which the STR's
> themselves may contain TAB characters. Unfortunately the STR editor in
> ResEdit uses TAB, shift-TAB, option-TAB, etc to move between fields.
> Putting "\t" within the STR seems to print literally, "\t". I can
> insert tabs with the hex editor, but I was hoping there was a more
> user-friendly solution.

I'd say write a wrapper routine for GetIndString which translates "\t"
into <tab> and "\\" into '\'. While you're at it, you could have it
process the other C-style string escape sequences, too.

Resorcerer doesn't do *anything* with tabs in this situation (the UI is
completely different), so I can't recommend it like I wanted to. :-(

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "The human race will decree from time to time: 'There is something at
 which it is absolutely forbidden to laugh.'" -- Nietzche on C++
>From [e-mail] 19 May 94 16:05:08 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: List Manager
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "[e-mail]" at May 18, 94 10:51:46 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1259
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 23
Date: 19 May 94 16:05:08 GMT

> Sounds to me like you have one of the utility programs installed which
> automatically compresses the file names in open and close dialog
> boxes. I believe this is a feature of NOW Utilities and others. If
> the file names in the open and close dialog boxes also compress when
> you have long file names when you are outside of the C programming
> enviroment - this is probably the cause.

This compression is a feature of many lists under System 7. LDEF 0
does it and so does the LDEF used by Standard File. Unfortunately, I
don't have time right now to verify this for certain regarding LDEF 0,
but that's what I remember. It doesn't surprise me that someone would
encounter this while writing a new program, because normally you would
want to truncate a compressed string or make the list/cell wider instead
of relying on the compression, so finished programs should not exhibit
too much of this behavior.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 'It seems the firm contracted by Intel to produce the famed "flying
 Pentium" ads could not complete the 3-D modeling on a PC before
 deadline. So in the best know-your-enemy tradition, they chose the next
 best thing -- a Quadra 840AV.'                 -- Mac The Knife 4/18/94
>From [e-mail] 19 May 94 16:49:20 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: At Ease AppleEvents
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 7
Date: 19 May 94 16:49:20 GMT

Does anyone know where I can get a list of them?

Specificaly - I need to send an 'odoc' event to the finder.  This works fine
when there IS a Finder... What happens if At Ease quits the finder?

[e-mail]

>From [e-mail] 19 May 94 18:33:23 GMT
Path: ucivax!gateway
From: [e-mail] (WalrathW)
Subject: RE: small icons in lists
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 40
Date: 19 May 94 18:33:23 GMT



 Macintosh Programming Secrets by Knaster & Rollin contains a simple custom
LDEF using small icons and  text. I recommend buying the book for all the other
interesting info in it, but the sources were generously uploaded to the nets by
the authors. Look in the source directories (at umich or sumex) for MPS or
something similar. There's also code for zooming windows, a custom WDEF, fancy
dialogs, and more.

-wayne
_______________________________________________________________________________
From: barnett aaron on Thu, May 19, 1994 16:34
Subject: small icons in lists
To: [e-mail]

can anyone point me to a ldef that does the small icons along with text.

\.
aaron <[e-mail]>
"It's a safety mechanism, the mind shuts off."

------------------ RFC822 Header Follows ------------------
Received: by mumcbm.rferl.org with SMTP;19 May 1994 16:34:34 U
Received: from moose.cs.indiana.edu by MUVAXA.RFERL.ORG (PMDF V4.2-12 #5010) id
 <[e-mail]>; Thu, 19 May 1994 16:34:26 CET
Received: from binky.ics.uci.edu by moose.cs.indiana.edu (5.65c/9.4jsm) id
 AA13515; Thu, 19 May 1994 09:34:18 -0500
Received: from ics.uci.edu by binky.ics.uci.edu id aa12281; 19 May 94 7:08 PDT
Received: from ics.uci.edu by binky.ics.uci.edu id aa12277; 19 May 94 7:05 PDT
Received: from USENET by q2.ics.uci.edu id aa08387; 19 May 94 7:05 PDT
Date: Thu, 19 May 1994 14:05:31 +0000 (GMT)
From: barnett aaron <[e-mail]>
Subject: small icons in lists
To: [e-mail]
Message-id: <[e-mail]>
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Content-transfer-encoding: 7BIT
Newsgroups: fa.think-c
X-Message-Id: <[e-mail]>
Approved: [e-mail]
>From [e-mail] 19 May 94 19:48:43 GMT
Path: ucivax!gateway
From: [e-mail] (Marco Sandri)
Subject: XCMDs and CopyBits...
Message-ID: <[e-mail]>
Return-Receipt-To: [e-mail]
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 894
MIME-Version: 1.0
Newsgroups: fa.think-c
Reply-To: Marco Sandri <[e-mail]>
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 25
Date: 19 May 94 19:48:43 GMT

Hello!

I'm developing a XCMD that should read and write PICTs.
I do the read/write work with the quickdraw bottlenecks, but it works only
in applications, and not in code resources like XCMDs.

My code works as follows:
- installation of my bottleneck routine (getpict)
- GetPictInfo to obtain info about the pict image (size, depth, ...)
- restore of the old bottlenecks
- creation of the GWorld of the right depth and size (and SetGWorld)
- installation of quickdraw bottlenecks
- DrawPict
- restore of the prev. bottlenecks.

I'm doing the work this way, because I would allow the user to load an image
without requiring a large amount of free RAM (eg: I'd like to load a 1Mb PICT
with 1.2Mb of free mem). I couldn't figure out how to do this without the
bottlenecks...

Of course, I've the same problem creating the PICT (with putpictproc bottleneck)

Any ideas?

Many thanks in advance.
>From [e-mail] 19 May 94 20:47:15 GMT
Path: ucivax!gateway
From: [e-mail] (John Wiggins (LS TP/414 508-262-1021) )
Subject: Cross Development Kit & Compiling on Power Mac
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 30
Date: 19 May 94 20:47:15 GMT

Folks:

I'm new to the group, so you may have covered this ground already this
month.  I have looked in the archives for an answers, but didn't this
covered there...

I have two questions about the Cross Development Kit* that Symantec is
selling for $100, plus S&H&tax.  I have called both their sales, and
Tech support lines, but neither could give definitive answers.  Here
are the questions:

        1) I have upgraded by C++ 6.0 to 7.0 using the (free) upgrade
           kit. Can I get by using just the "Cross Development Kit"
           alone, with just losing the neat extra functionality of introduced
           in version 7? (Must I spend another ~$170, with S&H&tax, for the
           "C++ 7.0 for Macintosh" package they offer with it?)
        2) Are there any problems in running the "Cross Development Kit"
           directly on a Power Mac to generate the Power Mac code?  It
           would be less useful to have to go to a 68K machine to compile...

Thanks,
_John_

email: [e-mail]                 Dr. John Wiggins, TP/414
 tele: 508-262-1021                             1100 Technology Park Drive
  FAX: [phone]                             Billerica, MA 01821

_______
*  This is the kit to cross-compile to PPC 601 code for Power Mac

>From [e-mail] 20 May 94 01:58:01 GMT
Path: ucivax!gateway
From: [e-mail] (Joshua Kuperman)
Subject: VA problems and questions
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL21]
Content-Type: text
Content-Length: 1457
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 32
Date: 20 May 94 01:58:01 GMT

I am trying to get a grip on using Virtual Architect and
would appreciate any help (including being told to rtfm if
you include page numbers).

1.  I did the Beeper example. I then tried to change it to
use the range editing that could be set with VA. I changed it
to only allow 1 to 10 beeps.When I run the app - the field is checked
for valid values when I try to quit. However when I hit the "beep"
button it is ignored.  What do I have to do to force all the edited
fields to be checked. I'm assuming that all the other buttons are
inheriting a method, that I lost when I put in the stuff to do the
beep.

2.  I went through the Process Monitor. I then tried to copy some
buttons, icons, and other resources into my VisualArchitect.rsrc to
see if I could use them with an app.  I couldn't get so much as one
button to show up.  What did I miss.

3.  What would I need to do to copy a video player window (the window
with the sound control on the left, a play/button, and then a farily
normal scroll plane on the bottom) like the one used by MovieEditor,
VideoFusion Recorder, and SimplyPlayer.  How could I then save so that
it would always be available for any new Visual Architect project.

TIA, but be warned - if I ever get VA figured out I'll probably just
start posting lots of "real coding" questions.

--
Josh Kuperman                  The favorite doesn't always win,
[e-mail]      no matter what the odds. -I. Gershwin


>From [e-mail] 20 May 94 14:41:45 GMT
Path: ucivax!gateway
From: [e-mail] (Lance Helsten)
Subject: Re: List Manager
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 24
Date: 20 May 94 14:41:45 GMT

> > I've got data which is sometimes slightly too long to fit in the cell
> > size I've given, and the List Manager is trying to be helpful by
> > changing the character spacing and making it fit. Trouble is, I'm
> > trying to make things line up consistently and this helpfulness is
> > getting in the way...
> >
> > Thanks in advance for any information.
>
> Sounds to me like you have one of the utility programs installed which
> automatically compresses the file names in open and close dialog boxes.  I
> believe this is a feature of NOW Utilities and others.
>
> If the file names in the open and close dialog boxes also compress when you
> have long file names when you are outside of the C programming enviroment -
> this is probably the cause.
>

Actually, System 7.x does the compression automatically and the only thing
that can be done about it is to write your own LDEF. For the how to on
that see New Inside Mac More Toolbox (I believe that is the name).

Lance


>From [e-mail] 20 May 94 19:44:56 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: Alert Positions (SFP) in S7
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 10
Date: 20 May 94 19:44:56 GMT

Does anyone know the specific algorithm System 7 uses to compute alert
positions?  Specifically, does anyone know how S7 centers Standard File
Package dialogs?  I wish to align my app window with the SFP windows if
possible.

__________________________________________________________________________
Brent Taylor, Computer Science / Math Teacher         Phone: [phone]
Elkford Secondary School                                Fax: [phone]
Box 910, Elkford, BC, Canada  V0B 1H0               [e-mail]

>From [e-mail] 22 May 94 18:43:40 GMT
Path: ucivax!gateway
From: [e-mail] (Markus Waldorf)
Subject: How to play with windows in an INIT?
Content-transfer-encoding: 7BIT
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 103
Date-warning: Date header was inserted by AdminA.RFERL.Org
Date: 22 May 94 18:43:40 GMT
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-type: text/plain; charset="us-ascii"
MIME-version: 1.0
Newsgroups: fa.think-c

I'm currently creating an INIT (just for fun!) and have the
following problem:

At some point I want the INIT to display a PICT or text in a
plain window during the startup process. The window should
discard automatically before loading all the other INITs.
I thought I may try using Quickdraw and the Window Manager,
but yet have not much success. I've checked out other source
code, but haven't found what I need. Perhaps someone can help me.
Thanks!

How can I draw a window, and how to dispose it?


Here is some of the coding I have so far:


struct qdStruc {
        char            privates[76];
        long            randSeed;
        BitMap          screenBits;
        Cursor          arrow;
        Pattern         dkGray;
        Pattern         ltGray;
        Pattern         gray;
        Pattern         black;
        Pattern         white;
        GrafPtr         thePort;
        long            end;
};

typedef struct qdStruc qdStruc;


void main( void )
{

        RememberA0();               // set up A4 to access resource globals
        SetUpA4();

        StartupStatus();            // might call DrawWindow()

        RestoreA4();

} // Main()



void DrawWindow( void )
{
        WindowPtr   window;
        qdStruc     quickDraw;
        GrafPort    localPort;
        long        our_a5;
        long        old_a5;
        long        old_CurrentA5;


        asm 68000                        // Prepare A5 for InitGraf call
            {
            move.l          CurrentA5,old_CurrentA5
            move.l          a5,old_a5
            lea             our_a5,a5
            move.l          a5,CurrentA5
            }


        InitGraf( & (quickDraw.thePort ) );     // InitGraf
        OpenPort( &localPort );                 // open our GrafPort

        SysBeep( 10 );

        window = GetNewWindow( 128, nil, (WindowPtr)-1L );

        SysBeep( 10 );

        /* AFTER THE 1st BEEP THE COMPUTER TOTATLY STOPS!
           Perhaps I cannot use the GetNewWindow function here.
           It's a small window! */

        /* ... */


        ClosePort( &localPort );


        asm 68000                     // Reset A5 and CurrentA5
            {
            movea.l         old_a5,a5
            move.l          old_CurrentA5,CurrentA5
            }

} //  DrawWindow()

(eom)

Internet: [e-mail]






>From [e-mail] 23 May 94 02:40:08 GMT
Path: ucivax!gateway
From: [e-mail] (Mauricio Tavares)
Subject: Re: Reading data from a file
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.3 PL2]
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 37
Date: 23 May 94 02:40:08 GMT


        Ok, I think some of you are still kind of confused about what i
want *and* I am confused about your explanations to my problem.  So, let us
go to shock treatment.  Here is my original code (uses stdio.h):

/**********************************************************************
 *
 * Function ReadData.c
 *      Reads data from input file
 *
 **********************************************************************/

#include "Astro.h"

void ReadData(file, data)

FILE* file;
inputDataFormat *data;
{
        double time, dnu, mu;
        double r1[3], r2[3];

        fscanf(file, "%lf\n", &(data->dt));
        fscanf(file, "%lf\n", &(data->dnu));
        fscanf(file, "%le", &(data->mu));

        fscanf(file,"%le %le %le", &(data->r1[0]), &(data->r1[1]),
                                   &(data->r1[2]));       /* Get vector r1 */
        fscanf(file,"%le %le %le", &(data->r2[0]), &(data->r2[1]),
                                   &(data->r2[2]));       /* Get vector r2 */

}

What should I do to make it use the Mac routines such as FSRead()?  I read
IM over and over but did not find anything that could help me with my
problem.  As usual, suggestions are welcome.

>From [e-mail] 23 May 94 06:48:33 GMT
Path: ucivax!gateway
From: [e-mail] (Rodolfo Cardarelli)
Subject: Tools Plus Info
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 6
Date: 23 May 94 06:48:33 GMT

I downloaded the shareware version of Tools Plus, and I took a look at the
manual. Did anibody tried the package before? It looks like once you've
started using it, you can't mix it with "tandard" Think C programming, or
even TCL.
I'd appreciate any opinion. Thank you
Rodolfo Cardarelli
>From [e-mail] 23 May 94 08:11:35 GMT
Path: ucivax!gateway
From: [e-mail] (Big Nose)
Subject: Drag & Drop. How?
X400-Originator: [e-mail]
Content-Identifier: 1185BA471100
Message-ID: <[e-mail]>
X400-Recipients: [e-mail]
Sensitivity: Company-Confidential
Newsgroups: fa.think-c
X-Message-ID: <2121100923051994/A02109/IAPE/1185BA471100*@MHS>
X400-Received: by mta sun2.mhs-relay.ac.uk in /PRMD=uk.ac/ADMD= /C=gb/; Relayed; Mon, 23 May 1994 09:10:42 +0100
X400-Received: by mta arcs.afrc.ac.uk in /PRMD=UK.AC/ADMD= /C=GB/; Relayed; Mon, 23 May 1994 09:12:17 +0100
X400-Received: by /PRMD=UK.AC/ADMD= /C=GB/; converted (ia5 text (2)); Relayed; Mon, 23 May 1994 10:10:26 +0100
Approved: [e-mail]
X400-Content-Type: P2-1988 (22)
Lines: 15
Date: 23 May 94 08:11:35 GMT
X400-MTS-Identifier: [/PRMD=UK.AC/ADMD= /C=GB/;2121100923051994/A02109/IAPE]


How the hell do I get the Finder to give my App Drag & Drop startup.
I have an app that receives AppleEvents and I want to let the user
drag a number of files onto its icon and just sit back while it
processes them. I assume that this is acheived through the 'odoc'
Event. However, the Finder steadfastly refuses to let me launch with
D&D. I have the acceptHighLevelEvents bit set in the SIZE -1 resource.

What else do I have to do?

Thanks,

Andy Law

( Lawa @ afrc.ac.uk                         Big Nose in Edinburgh )
>From [e-mail] 23 May 94 12:29:12 GMT
Path: ucivax!gateway
From: [e-mail] ("Robert S. Mah")
Subject: Re: Drag & Drop. How?
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 20
Date: 23 May 94 12:29:12 GMT

At  8:11 AM 5/23/94 +0000, Big Nose wrote:
> How the hell do I get the Finder to give my App Drag & Drop startup.
> I have an app that receives AppleEvents and I want to let the user
> drag a number of files onto its icon and just sit back while it
> processes them. I assume that this is acheived through the 'odoc'
> Event. However, the Finder steadfastly refuses to let me launch with
> D&D. I have the acceptHighLevelEvents bit set in the SIZE -1 resource.

Double check your BNDL information.  And then if that's OK, rebuild
the desktop.  The Finder searches the desktop for the BNDL info, so
you have to rebuild it every time you change the BNDL if you want
Finder to know about it.

Cheers,
Rob

___________________________________________________________________________
Robert S. Mah  -=-  One Step Beyond  -=-  212-947-6507  -=-  [e-mail]


>From [e-mail] 23 May 94 14:10:58 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: Re: Drag & Drop. How?
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 73
Date: 23 May 94 14:10:58 GMT

>How the hell do I get the Finder to give my App Drag & Drop startup.
>I have an app that receives AppleEvents and I want to let the user
>drag a number of files onto its icon and just sit back while it
>processes them. I assume that this is acheived through the 'odoc'
>Event. However, the Finder steadfastly refuses to let me launch with
>D&D. I have the acceptHighLevelEvents bit set in the SIZE -1 resource.
>
>What else do I have to do?
>
>Thanks,
>
>Andy Law
>
>( Lawa @ afrc.ac.uk                         Big Nose in Edinburgh )

More. Lots more.

On page 9-17 of IM 6, you will find a good explanation of what you have to
do to add drag and drop. Basically what you do is add another item to your
FREF resource and BNDL resource (I prefer to use ResEdit to do this. Edit
the BNDL resource, then select the Extended View under the BNDL menu).
This resource details what can be dropped on your application.  For
example, let us say that this is what ResEdit displays to you (roughly)

  FREF                 FINDER ICONS
 --------------------+-----------------------------------------------------
 local  ResID  Type  |  local  ResID  ICN#  icl4  icl8  ics#  ics4  ics8
 --------------------+-----------------------------------------------------
   0     128   APPL  |    0     128   [icon] []    []    []    []     []
 --------------------------------------------------------------------------

Okay, lets say you want to drop items of type TEXT.  You would add the
following:

 --------------------+-----------------------------------------------------
   1     129   TEXT  |    1     129     []   []    []    []    []     []
 --------------------------------------------------------------------------

Now you can drop anything that has the type 'TEXT' (Teachtext documents for
example!).  For a disk (hard and floppy), it is 'disk'. For folders it is
'fold'.  Using '****' allows you to drop everything EXCEPT disks or
folders. If you want to drop any possible thing, then adjust BNDL (and
FREF) to look like the following:

  FREF                 FINDER ICONS
 --------------------+-----------------------------------------------------
 local  ResID  Type  |  local  ResID  ICN#  icl4  icl8  ics#  ics4  ics8
 --------------------+-----------------------------------------------------
   0     128   APPL  |    0     128   [icon] []    []    []    []     []
 --------------------------------------------------------------------------
   1     129   ****  |    1     129   [icon] []    []    []    []     []
 --------------------------------------------------------------------------
   2     130   fold  |    2     130   [icon] []    []    []    []     []
 --------------------------------------------------------------------------
   3     131   disk  |    3     131   [icon] []    []    []    []     []
 --------------------------------------------------------------------------

As far as I can remember, this is ALL you have to do!

Good luck,

Lloyd

----------------------------+-----------------------------------------
Lloyd Sargent               | "On Internet no one can hear you scream"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 23 May 94 14:18:34 GMT
Path: ucivax!gateway
From: [e-mail] ("B. Kirsch")
Subject: Mixing C and C++
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 23 May 94 14:18:34 GMT


I am writing a C program and one of my modules probably could be better
programmed with C Objects using C++.  If I want to use those classes in
other modules in my project, do I also have to rename those files with
the .cpp extension to use the Sym C++ translator?  If I have only one .cpp
file in the project, do all of the other ones in the project have to be .cpp?


Thanks,
Barry Kirsch
[e-mail]
>From [e-mail] 23 May 94 15:56:13 GMT
Path: ucivax!gateway
From: [e-mail] (Steve Trottier)
Subject: Re: Drag & Drop. How?
Message-ID: <[e-mail]>
X-Mailer: WordPerfect Office 4.1
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 8
Date: 23 May 94 15:56:13 GMT

 The finder checks an applications 'FREF' resources to determine if a
 file can be dropped onto it.  When you create your 'BNDL' resource
 correctly, ResEdit should automatically generate the needed 'FREF'
 resources.

 Steve Trottier <[e-mail]>


>From [e-mail] 23 May 94 23:28:02 GMT
Path: ucivax!gateway
From: [e-mail] (Danny Thomas)
Subject: Re: Reading data from a file
X-Sender: [e-mail].
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 32
Date: 23 May 94 23:28:02 GMT

>        Ok, I think some of you are still kind of confused about what i
>want *and* I am confused about your explanations to my problem.  So, let us
>go to shock treatment.  Here is my original code (uses stdio.h):
>
>/**********************************************************************
> *
> * Function ReadData.c
> *      Reads data from input file
> *
> **********************************************************************/
>
>#include "Astro.h"
>
>void ReadData(file, data)
>
>FILE* file;
>inputDataFormat *data;
>{
>        double time, dnu, mu;
>        double r1[3], r2[3];
>
>        fscanf(file, "%lf\n", &(data->dt));
>        fscanf(file, "%lf\n", &(data->dnu));
this code is fragile if you're not checking for errors, ie that the number
of successful conversions equals the number of values you're trying to
read.


cheers,
Danny Thomas  ([e-mail])


>From [e-mail] 24 May 94 19:02:18 GMT
Path: ucivax!gateway
From: [e-mail] (Jean-Marie Laborde)
Subject: Adressing memory screen
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 25
Date: 24 May 94 19:02:18 GMT

It is not recommanded to write directly in the memory screen but some
time we need to do that.

We managed for some graphics intensive application to use the
correct baseAddr of the screen bit(pix)map related to most
of the standard (Apple) monitors used for main device as
        screenBits.baseAddr.

Unfortunately on some (especially non Apple) monitors connected
(especially on not 32 bits machines) this does not work.
Does anyone knows about writing outside of the 24-bit world on
a 24 bit machine?
(In such cases tricks using  SwapMMUMode() does not work...)



****************************************************************************
Jean-Marie Laborde                                       [e-mail]
Responsable du Projet IMAG Cabri-geometre
Laboratoire de Structures Discretes et de Didactique - IMAG
Universite Joseph Fourier - CNRS          BP 53 38041 Grenoble cedex 9 FRANCE
(33) 76 51 46 10   (sec 76 51 46 16)      (fax [phone])
****************************************************************************


>From [e-mail] 24 May 94 21:45:09 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: Swapping out the Switchboard in TCLibs
X-Sender: [e-mail] (Unverified)
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date: 24 May 94 21:45:09 GMT

I am attempting to swap out the Switchboard by subclassing my own.
However, I seem to be having little luck. Is there anyone out there who has
done this and can give me a few pointers?

TIA,

Lloyd

----------------------------+-----------------------------------------
Lloyd Sargent               | "On Internet no one can hear you scream"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 25 May 94 00:50:08 GMT
Path: ucivax!gateway
From: [e-mail] (Kirk Crawford)
Subject: Debugger LockUp
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 12
Date: 25 May 94 00:50:08 GMT

Subject:  Debugger LockUp
I recently upgraded my TC6 to TC7.  It worked for a few days, but now when I
turn on the Debugger and run my project the Think C locks up when it looks like
it is launching the debugger.  I don't even get the source and data windows,
just a blank menu bar.  I replaced the debugger with my backup and that failed
too.  Any ideas?

Kirk Crawford
[e-mail]



>From [e-mail] 25 May 94 03:48:19 GMT
Path: ucivax!gateway
From: [e-mail] (Francesco Stiffoni)
Subject: TCL 2.0.1
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 12
Date: 25 May 94 03:48:19 GMT

Where can I get TCL 2.0.1. I know that Symantec distributes it with CDK
because TCL 2.0 doesn't compile well on PowerPC. Even if it has no real
anhacements perhaps TCL 2.0.1 corrects some bugs of TCL 2.0.

Thanks in advance for information

Francesco Stiffoni
University of Venice
Linguistic Department
Laboratory of Computational Linguistic


>From [e-mail] 25 May 94 03:53:29 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: Reading data from a file.
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 40
Date: 25 May 94 03:53:29 GMT

Marcio says:

> Ok, I think some of you are still kind of confused about what i
>want *and* I am confused about your explanations to my problem.  >So, let us
>go to shock treatment.  Here is my original code (uses stdio.h):

I'm not sure we're the one's confused, *you* came here with a problem, not
the other way around :)

The answer is quite simple. I'd aready emailed it privately to marcio
earlier.

Don't use fscanf().  This reads from UNIX-style (stdio-style) files.  You
want to use FSRead(), so use it!  Read the data into an in-memory buffer.
 Then use another ANSI library call, sscanf() to do the same thing fscanf()
does, but reads the data from memory instead of from a file.  This is a
standard buffering scheme.

Moral: Know your libraries.  Also, get decent reference books.

A side effect of using FSRead() (or the more appropriate call FSpRead()) is
that you can control the amount of data you read at a time.  The Mac (like
most OS's) has a fair amount of overhead for reading one byte off of the
disk, but spread this relatively fixed amount of overhead across reading
hundreds of bytes at a time and your application speeds up.  It may not be an
issue in this app, but it's a nice thing to know about.

Using the ANSI calls to read and write data does two things: It adds a layer
of calls over the existing FSxxx() calls that actually read the data.  And it
removes the control over buffering from you to the library.  I've not looked
at the recent library's sources, but in Think 5, it was relatively
ineffecient.

Hope that's detailed enough :)

Malcolm Teas
"Relax, it's only ones and zeros..."
Teas Technology - Custom software for the Mac and Newton.
Internet: "[e-mail]" AppleLink or AOL: "mhteas"

>From [e-mail] 25 May 94 03:53:42 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: Tools Plus Info
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 8
Date: 25 May 94 03:53:42 GMT

What's Tools Plus do?  Can you tell us a little about it's purpose and
features?

Malcolm Teas
"Relax, it's only ones and zeros..."
Teas Technology - Custom software for the Mac and Newton.
Internet: "[e-mail]" AppleLink or AOL: "mhteas"

>From [e-mail] 25 May 94 05:06:35 GMT
Path: ucivax!gateway
From: [e-mail] (Monty Solomon)
Subject: MPW Tools
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 16
Date: 25 May 94 05:06:35 GMT

The Symantec C++ box states that it supports MPW tools.  I can't seem to find
any information about this in the manuals.

What is the procedure for writing a MPW tool so that it can be used in the
Symantec C++/Think C environment?

Thanks.

Monty


---
# Monty Solomon / PO Box 2486 / Framingham, MA  01701-0405
# [e-mail]


>From [e-mail] 25 May 94 15:01:39 GMT
Path: ucivax!gateway
From: [e-mail] (Lance Helsten)
Subject: Re: Reading data from a file.
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date: 25 May 94 15:01:39 GMT

> Don't use fscanf().  This reads from UNIX-style (stdio-style) files.  You
> want to use FSRead(), so use it!  Read the data into an in-memory buffer.
>  Then use another ANSI library call, sscanf() to do the same thing fscanf()
> does, but reads the data from memory instead of from a file.  This is a
> standard buffering scheme.
>

I agree with you on FSRead/FSWrite when you are writing mac only files
because writing to binary files is 40 to 80 times faster than fscanf. But,
if you need the file to go from platform to platform Text files are
usually the best (no byte swapping--big/little endian problems) though
MUCH slower. So a mixture of both of these are needed in any application.

Lance
Chief Engineer
dBL Software


>From [e-mail] 25 May 94 16:59:06 GMT
Path: ucivax!gateway
From: [e-mail] (Kirk Crawford)
Subject: Debugger LockUp Summary
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 21
Date: 25 May 94 16:59:06 GMT

Subject:  Debugger LockUp Summary
Thanks to the quick help from the kind people on this list I have solved my
previous problem of the Debugger locking up when I tried to run with the
debugger.  I replaced the Think Project Manager and Think Debugger files with
the original ones from my virgin 6.0 disks, and then patched them.  It works
fine now.

Thanks especially to Marco Sandri  ([e-mail])

Some thought the project file might have been corrupt, but the bug showed on
ALL my old projects and I made a new project with a Hello World program and
that failed too.

BTW:
There is a bad address in the list.  I got an automated response from
[e-mail] it seems his address needs to be changed to
[e-mail].

-Kirk


>From [e-mail] 25 May 94 17:27:13 GMT
Path: ucivax!gateway
From: [e-mail] (Brian Bezanson)
Subject: RE>MPW Tools
Message-ID: <[e-mail]>
X-Mailer: InterCon Dispatcher/SMTP for QuickMail
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
X-Priority: 4
Lines: 60
Date: 25 May 94 17:27:13 GMT

Date    5/25/94
Subject RE>MPW Tools
From    Brian Bezanson
To      Monty Solomon, think-c

        Reply to:   RE>MPW Tools
It can't dirtectly use MPW Tools or even create them. It can run then using
ToolServer or SourceServer--this is/should be documented in the manuals.

Brian

--------------------------------------
Date: 5/25/94 2:47 AM
To: Brian Bezanson
From: Monty Solomon
Resent-Date: Wed, 25 May 1994 02:44:02 CDT
Resent-From: <[e-mail]>
Resent-To: <BEZANSON@MACSRV>
Received: from binky.ics.uci.edu by VIXVAX.MGI.COM (MX V3.3 VAX) with SMTP;
          Wed, 25 May 1994 02:43:58 CDT
Received: from ics.uci.edu by binky.ics.uci.edu id aa08663; 24 May 94 22:08
PDT
Received: from ics.uci.edu by binky.ics.uci.edu id aa08659; 24 May 94 22:06
PDT
Received: from USENET by q2.ics.uci.edu id aa11522; 24 May 94 22:06 PDT
From: Monty Solomon <[e-mail]>
Subject: MPW Tools
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Date: 25 May 94 05:06:35 GMT
To: [e-mail]


The Symantec C++ box states that it supports MPW tools.  I can't seem to find

any information about this in the manuals.

What is the procedure for writing a MPW tool so that it can be used in the
Symantec C++/Think C environment?

Thanks.

Monty


---
# Monty Solomon / PO Box 2486 / Framingham, MA  01701-0405
# [e-mail]








>From [e-mail] 25 May 94 18:33:42 GMT
Path: ucivax!gateway
From: [e-mail] (Kirk Swenson)
Subject: Re: Swapping out the Switchboard in TCLibs
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 14
Date: 25 May 94 18:33:42 GMT

>I am attempting to swap out the Switchboard by subclassing my own.
>However, I seem to be having little luck. Is there anyone out there who has
>done this and can give me a few pointers?

Override CApplication::MakeSwitchboard() to create your specific subclass
of CSwitchboard and set CApplication's itsSwitchboard instance variable.
Works for me. :-)

Kirk Swenson
Smith-Kettlewell Eye Research Institute
San Francisco, CA
[e-mail]


>From [e-mail] 25 May 94 18:52:23 GMT
Path: ucivax!gateway
From: [e-mail] (Brent Taylor)
Subject: FSpExchangeFiles()
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 41
Date: 25 May 94 18:52:23 GMT

I'm having difficulty implementing the "safe" save described in
NIM-"Files", p1.23-27.  SaveDataFile() receives a file specification for
the currently opened file.  GetTempSpec() returns a file specification for
a unique temporary file.  WriteTempFile() opens, writes, and closes the
temporary file.  FSpExchangeFiles() exchanges the opened file with the
temporary file.  IM6 and NIM-"Files" both state that FSpExchangeFiles()
will work on both opened and closed files.

My problem is that, after the exchange, I get fBsyErr (-47 file busy) on my
call to FSpDelete().  It would seem that the file referenced by tempSpec is
open, I suppose, as a result of my call to FSpExchangeFiles().

Can anyone elaborate on what specifically happens when FSpExchangeFiles()
exchanges an opened file with a closed file and, perhaps, recommend a
possible solution?


/******************************** SOURCE CODE ********************************/

OSErr SaveDataFile(FSSpecPtr fsSpec)
{
    FSSpec  tempSpec;
    OSErr   errNo;

    if ((errNo = GetTempSpec(fsSpec->vRefNum, &tempSpec)) == noErr)
        if ((errNo = FSpCreate(&tempSpec, 'ROCK', 'Rock',
                                        smSystemScript)) == noErr) {
            if ((errNo = WriteTempFile(&tempSpec)) == noErr)
                if ((errNo = FSpExchangeFiles(&tempSpec, fsSpec)) == noErr)
                    return(FSpDelete(&tempSpec)); /* returns fBsyErr */
            FSpDelete(&tempSpec);
        }
    return(errNo);
}


__________________________________________________________________________
Brent Taylor, Computer Science / Math Teacher         Phone: [phone]
Elkford Secondary School                                Fax: [phone]
Box 910, Elkford, BC, Canada  V0B 1H0               [e-mail]

>From [e-mail] 26 May 94 07:08:06 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Using the Standard Libraries.
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 18
Date: 26 May 94 07:08:06 GMT
X-Envelope-to: [e-mail]

I'm converting some C code to the mac, which uses
a lot of Standard Libraries routines.
The code compiles ok, but unfortunately the program
crashes at some points, or does not work correct.

for instance:

I can not write to stderr using fprintf(stderr,".......

fgets(...) returns NULL, while fgetc or fscanf on the same stream
works well.

vprintf(.....) gives unpredictable results.

I use Think_c 5.0.4.
Any idea whats wrong?

Gerrit.
>From [e-mail] 26 May 94 07:20:43 GMT
Path: ucivax!gateway
From: [e-mail] ("Jan R. M. Rohman")
Subject: Text and color
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Reply-To: [e-mail]
Approved: [e-mail]
Lines: 32
Date: 26 May 94 07:20:43 GMT


Hello,

I'm writing a nice application with 3-D buttons, input field (in
dialogs) with a 3-D look etc. Now I have some problems left before I
can distribute my program.

1.) I made my input field light gray to have a nice look. But I
thought I should die when the prompter (text cursor) in the fields got
yellow. No normal human eyes can see the yellow prompter on the gray
background. How do I change the color of the prompter???

2.) I plot functions and solutions to differential equations in a
window. Is there any free code, or have any body out there code to
make logarithmic coordinate axes.

3.) At last I want to know how to draw (input) text at any position
in a window. When I have a plot of a function in my plotting window
I want a cursor to place at any position in the window for writing
text. Is there any body out there ("Pink Floyd") who have a snippet
to draw text???

        Please HELP Me.....

----------------------------------------------------------------------
 From:  [e-mail]  Jan R. M. Rohman                          !
 Phone: [phone]    ABB Corporate Research Center Dept. KF    !
 Fax:   [phone]    Department of Computational Methods,      !
                           Electromagnetic Fields and Fluid Dynamics !
                           S-721 78 Vasteras, SWEDEN                 !
 Home:   Vasagatan 35, 722 15 Vasteras, Phone:  [phone]      !
----------------------------------------------------------------------
>From [e-mail] 26 May 94 14:27:52 GMT
Path: ucivax!gateway
From: [e-mail] (Dana S Emery)
Subject: Re: Swapping out the Switchboard in TCLibs
Message-ID: <[e-mail]>
In-Reply-To: Your message <[e-mail]> of 25 May 94
 18:33:42 GMT
Content-Type: TEXT/plain; charset=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 41
Date: 26 May 94 14:27:52 GMT

>I am attempting to swap out the Switchboard by subclassing my own.
>However, I seem to be having little luck. Is there anyone out
>there who has
>done this and can give me a few pointers?

The switchboard is a mission critical part of the TCL.  Before tweaking it one
should do a thorough analysis of its behaviour and interactions (and potential
interactions) with the entirety of the TCL.  Hopefully this was obvious, but
still it needs to be said.

If the overhead of dispatching overloaded calls proves expensive
you will want to use a rewriten version of the TCl file, this is
easy enough (assuming you havent precompiled the switchboard),
simply remove the TCL version of CSwitchboard.c from your project,
then add your own rewritten version.  Many of the project files
will then need recompilation.

But if that overhead is affordable you can work more cleanly,
simply author a subclass of CSwitchboard to define overrides of
old methods and/or new methods which will accomplish your
desired behaviour.

v1.x TCL (and probably later versions as well) CApplication
provides a hook which makes it easy to integrate your subclass
of the switchboard, when CApplication wants to have the
switchboard object created it sends itself a message "MakeClipboard".

The CApplication subclass you _have_ to create anyway can override
this method in order to create and initialize the CSwitchboard
subclass you will be writing, an examination of the code inherited
CApplication::Makeclipboard method should reveal that you dont need
to invoke it, but you can take it for a model.  TCL 2.0 may have
decided to use templates for this, in which case you are on your
own, I don mess with them.

The "MakeXxx" concept is employed in several places in the TCL, it
is a common way to allow subclasses of yourself to provide subclasses
of the Xxx.
--
dana s emery <[e-mail]>

>From [e-mail] 26 May 94 18:55:48 GMT
Path: ucivax!gateway
From: [e-mail] (Hong Shi)
Subject: Need Help: Mac Resource
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 17
Date: 26 May 94 18:55:48 GMT

I am new to Macintosh programming, and I found the concept of resource
somewhat difficult. In particular, I have the following questions:

1. What's the exact role of resource in a program? In the think C
project manager, the list does not include resource file.
2. How do I know I need some kinds of resource for my program?
3. How to define and use resource? I read the part of ResEdit in Think C
manual, but it still kind of mystery to me, for the use, in Macintosh
Revealed, it described several ways to open and use resource files, but
I didn't find similar examples in the Think C package, although those
projects have their resource files. So it seems to me resource is
automatically been used. And if something is wrong with the resource, do
I have anyway to locate it?

Any help will be appreciated.

Hong Shi
>From [e-mail] 27 May 94 00:59:51 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: Re- FSpExchangeFiles()
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 23
Date: 27 May 94 00:59:51 GMT


>From Pindar user Ian McCall

Subject: _Re: FSpExchangeFiles()
> My problem is that, after the exchange, I get fBsyErr (-47 file
> busy) on my call to FSpDelete().  It would seem that the file
> referenced by tempSpec is open, I suppose, as a result of my
> call to FSpExchangeFiles().

I had this, and it took me a while to see what was up. What's
happening is that although you've physically exchanged the files, any
fileRef you've got going from opening this file hasn't been swapped.

I found the only reliable method was to close both files. I'm sure I
took the easy route out and there's another way of doing it, but this
way certainly works.



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 27 May 94 02:22:56 GMT
Path: ucivax!gateway
From: [e-mail] ("Mark D. Nagel")
Subject: ARCHIVE: Symantec C++ 7.0.2 update
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <29048.770004351@barracuda>
Reply-To: [e-mail]
Approved: [e-mail]
Organization: Relational Data Systems, Irvine, CA
Lines: 12
Date: 27 May 94 02:22:56 GMT
Phone: [phone]

I have placed the update for THINK C/Symantec C++ 7.0.2 on
ftp.ics.uci.edu.  Please note that this update is _only_ for those
who updated from 6.0.x with publically available updaters.  I will
try to get the other updater (for people who bought 7.0) as soon as
possible.

[archived as: EDS_7.0.2_Update.sea.bin; 312K]
--------
Mark D. Nagel <[e-mail]>      Relational Data Systems
                                             30 Executive Park, Suite 260
Eat right.  Exercise.  Die anyway.           Irvine, CA 92714
                                             [phone]
>From [e-mail] 27 May 94 03:51:38 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: ICON
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 15
Date: 27 May 94 03:51:38 GMT

Hello,
I think this is a simple question for you guys.  Could anybody please tell me
how to create a custom icon for my application?  After build my application,
it always gives simple generic icon for the application.  I used ResEdit and
create an icon and made BNDL resource that has 'APPL' and the icon I created.
 And then I assinged arbitrary signature in the signature field(4 characters
long).  But still, I don't get the icon for my application.  Do I have to add
some code into my source code to do this?

Thanks...

[e-mail]



>From [e-mail] 27 May 94 08:50:35 GMT
Path: ucivax!gateway
From: [e-mail] (Lindsay Davies)
Subject: Chooser name - where is it?
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Reply-To: Lindsay Davies <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 27 May 94 08:50:35 GMT

This is something that has been niggling me for a while, and I have
never found out the answer.

How do I get the Chooser/Owner name that's set from the file sharing extension?
I have scoured IM and Think Ref, but can't seem to hit on the answer.

TIA

Lindsay


>From [e-mail] 27 May 94 10:58:33 GMT
Path: ucivax!gateway
From: [e-mail] (WalrathW)
Subject: RE: ICON
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 49
Date: 27 May 94 10:58:33 GMT

Try rebuilding your desktop (hold down cmd-option keys while rebooting your
mac). If that doesn't work than you must have screwed up somthing with the
resources. Make sure you also have an FREF resource (I think it gets created
when you make the BNDL, though). Go into another application's resources and
compare against theirs, making certain that yours looks more or less the same.

cheers,
  -wayne
_______________________________________________________________________________
From: [e-mail] on Fri, May 27, 1994 10:48
Subject: ICON
To: [e-mail]

Hello,
I think this is a simple question for you guys.  Could anybody please tell me
how to create a custom icon for my application?  After build my application,
it always gives simple generic icon for the application.  I used ResEdit and
create an icon and made BNDL resource that has 'APPL' and the icon I created.
 And then I assinged arbitrary signature in the signature field(4 characters
long).  But still, I don't get the icon for my application.  Do I have to add
some code into my source code to do this?

Thanks...

[e-mail]




------------------ RFC822 Header Follows ------------------
Received: by mumcbm.rferl.org with SMTP;27 May 1994 10:46:48 U
Received: from moose.cs.indiana.edu by MUVAXA.RFERL.ORG (PMDF V4.2-12 #5010) id
 <[e-mail]>; Fri, 27 May 1994 10:46:26 CET
Received: from binky.ics.uci.edu by moose.cs.indiana.edu (5.65c/9.4jsm) id
 AA10253; Fri, 27 May 1994 02:23:22 -0500
Received: from ics.uci.edu by binky.ics.uci.edu id aa27422; 26 May 94 20:54 PDT
Received: from ics.uci.edu by binky.ics.uci.edu id aa27418; 26 May 94 20:51 PDT
Received: from USENET by q2.ics.uci.edu id aa07291; 26 May 94 20:51 PDT
Date: Fri, 27 May 1994 03:51:38 +0000 (GMT)
From: [e-mail]
Subject: ICON
To: [e-mail]
Message-id: <[e-mail]>
X-Mailer: America Online Mailer
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Content-transfer-encoding: 7BIT
Newsgroups: fa.think-c
X-Message-Id: <[e-mail]>
Approved: [e-mail]
>From [e-mail] 27 May 94 16:23:16 GMT
Path: ucivax!gateway
From: [e-mail] ("Mark D. Nagel")
Subject: Re: patch
Message-ID: <[e-mail]>
In-reply-to: Craig Hollabaugh's message of Fri, 27 May 1994 01:54:37 MDT.
        <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <13317.770055098@rdatasys>
Reply-To: [e-mail]
Approved: [e-mail]
Organization: Relational Data Systems, Irvine, CA
Lines: 33
Date: 27 May 94 16:23:16 GMT
Phone: [phone]

On Fri, 27 May 1994 01:54:37 MDT Craig Hollabaugh wrote:

    Saw the patch to 7.0.2 and grabbed it.

    Got a "verify resource CNFG 0" error when going from 7.0.1

    So I converted a 6.0.1 to 7.0.1 then to 7.0.2, got the same
    error.

    Unfortunately, I'm on the road without my 6.0 disks, so I can't
    start with a "clean" version.

    Does the patch work for you? (Stupid question since you posted it) Are
    other people having the same problem?

I applied all of the patch modules without error except for the
Symantec C++ 7.0.2 patch.  It apparently requires a 7.0.1 version of
the translator, even though the AppleLink message containing the
update kit says it is for those who used the previously published
update.

I had not attempted the patch prior to uploading the archive, since
I retrieved the kit directly from AppleLink and (mistakenly) assumed
it had been tested first.  I'll upload an update as soon as it is
made available.

Mark

--------
Mark D. Nagel <[e-mail]>      Relational Data Systems
                                             30 Executive Park, Suite 260
Eat right.  Exercise.  Die anyway.           Irvine, CA 92714
                                             [phone]
>From [e-mail] 27 May 94 16:55:34 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: Reading data from a file.
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 29
Date: 27 May 94 16:55:34 GMT

Lance says,

>I agree with you on FSRead/FSWrite when you are writing mac only files
>because writing to binary files is 40 to 80 times faster than fscanf. But,
>if you need the file to go from platform to platform Text files are
>usually the best (no byte swapping--big/little endian problems) though
>MUCH slower. So a mixture of both of these are needed in any application.

So use text and FSxxx() calls.  Like the beer commercials say: No need to
argue, you can have both.

FSxxx() calls just read and write bits.  Use the FSxxx() calls in a
platform-dependent layer to read and write buffers.  Use the sscanf() and
sprintf() calls to read and write the data in the buffers.

I've used this technique several times and it *does* work.  As Lance points
out, it's a lot faster too.

BTW, I've noticed that there's a large difference between writing 1-5 bytes
at a time to writing 100 bytes at a time.  The curve continues, 1000 bytes is
better than 100 and so on.  The curve flattens out between 32K and 64K bytes
(my tests were at sizes handy for me).  There was also an article in MacTech
magazine last year I believe that discusses block sizes and disk I/O.

Malcolm Teas
"Relax, it's only ones and zeros..."
Teas Technology - Custom software for the Mac and Newton.
Internet: "[e-mail]" AppleLink or AOL: "mhteas"

>From [e-mail] 27 May 94 16:58:06 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: MPW Tools
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 14
Date: 27 May 94 16:58:06 GMT

>The Symantec C++ box states that it supports MPW tools.

Does it support the ToolServer?  I believe that Symantec 6 was supposed to do
this.  At least their earlier publicity for 6 said this.  My guess is that
Symantec finally got TS in.

If it *is* TS, then you'd write the tool just like you'd write it for MPW.
 After all, TS is just MPW.

Malcolm Teas
"Relax, it's only ones and zeros..."
Teas Technology - Custom software for the Mac and Newton.
Internet: "[e-mail]" AppleLink or AOL: "mhteas"

>From [e-mail] 27 May 94 17:05:25 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: strncmp returns!
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 69
Date: 27 May 94 17:05:25 GMT


>From Pindar user Ian McCall

Subject: _strncmp returns!
A while back, I posted code saying I was having trouble with the
library function 'strncmp'. A couple of people posted back saying
they weren't having this problem, so I put it down to something I was
doing and worked around it.

However, I've now had call to use this routine again, and this time
there's a definite problem. Here is an excerpt from a piece of
conversion code I'm writing.


Boolean GetTagClassForIdentifier(char *theIdentifier)
/* This function return the type of tag indicated by the
   identifier string 'theIdentifier'.
*/
{
   TagType   theType;
   int       test;

       /*   It's just a glorified if statement. We can't use case
since
       the data we're testing doesn't have a discrete value. Ho
       hum...
   */
   if((test = strncmp(theIdentifier, kPatentNumberString,
                               strlen(kPatentNumberString))) == 0)
   {
      theType = kPatentNumberTag;
   }
   else if(strcmp(theIdentifier, kGrantTitleString) == 0)
      theType = kGrantTitleTag;
   else if(strcmp(theIdentifier, kIntCl5String) == 0)
         etc.
   else
      theType = kUnknownTag;

   /*   Return the result   */
   return(theType);
}

The test data is

   theIdentifier       = "(54)Grant Title"
   kPatentNumberString = "(11)"


Now - under THINK, whatever I pass into this routine, strncmp says is
equal. ANYTHING. These means that everything I pass in gets
classified as being a kPatentNumberTag.

Compiling precisely the same code under MPW and using precisely the
same test data I get the correct result - the identifier is classed
as kGrantTitleTag.

Any clues?


Cheers,
      Ian



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 27 May 94 18:06:01 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Re: ICON
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset="us-ascii"
MIME-version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Date-warning: Date header was inserted by AdminA.RFERL.Org
Lines: 48
Date: 27 May 94 18:06:01 GMT

>From:   IN%"[e-mail]" 27-MAY-1994 10:27:22.38
>To:     IN%"[e-mail]"
>CC:
>Subj:   ICON
>
>
>Hello,
>I think this is a simple question for you guys.  Could anybody please tell me
>how to create a custom icon for my application?  After build my application,
>it always gives simple generic icon for the application.  I used ResEdit and
>create an icon and made BNDL resource that has 'APPL' and the icon I created.
> And then I assinged arbitrary signature in the signature field(4 characters
>long).  But still, I don't get the icon for my application.  Do I have to add
>some code into my source code to do this?
>
>Thanks...
>
>[e-mail]

If you have previously given your application a custom Icon before (Get
Info cut&paste),
make sure you have NOT selected the same one in your BNDL resource (ID -
16455).
And just in case, use Resedit (File Get file/folder info) etc. and check if
your application
still has the custom Icon flag - if so, turn it off.

Try rebuilding the desktop. I would copy the application to a floppy disk,
then re-insert the
disk with Command and Option key pressed down. However, sometimes this is
not completely
effective because the original desktop file was corrupted, so rebuilding it
only yields an
updated desktop file which is still damaged. There are free tools that can
help like
DestopReset or Techtool, but you can also use Resedit. Change the file
attribute of the
Desktop DF and Desktop DB file to visible, put the two files in the trash
and restart.


Best Regards,

Markus Waldorf
Information Technology Support Consultant
RFE/RL Inc, Munich, Germany


>From [e-mail] 27 May 94 19:25:12 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: Chooser name - where is it?
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Lindsay Davies" at May 27, 94 08:50:35 am
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1242
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 38
Date: 27 May 94 19:25:12 GMT

> This is something that has been niggling me for a while, and I have
> never found out the answer. How do I get the Chooser/Owner name that's
> set from the file sharing extension? I have scoured IM and Think Ref,
> but can't seem to hit on the answer.

Here is the code I use to get the Chooser name. The names and tabs have
been changed to protect the innocent and the narrow mailers. The 'STR '
resource lives in the System file -- this is why you don't see me
releasing it or changing its handle state to purgeable or some other
method of conserving the memory associated with it. I put no warrantees
on this code; use at your own risk.

pascal OSErr GetChooserName (Str31 chooserName)
{
        OSErr oe = noErr;

        StringHandle chooserNameH =
                (StringHandle) GetResource ('STR ',-16096);

        if (!(oe = ResError ( )))
        if (!chooserNameH) oe = resNotFound;
        else
        {
                unsigned char length = **chooserNameH;

                if (length > 31) length = 31;
                BlockMoveData ((*chooserNameH) + 1, chooserName + 1, length);
                *chooserName = length;
        }

        return oe;
}

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "Bedrock does support a form of drag and drop. Development dragged on
 for about two years, and then was dropped." -- Brian Clark
>From [e-mail] 27 May 94 20:25:07 GMT
Path: ucivax!gateway
From: [e-mail] (Kirk Swenson)
Subject: Apparent CodeGen bug in TC 7
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 63
Date: 27 May 94 20:25:07 GMT

While recompiling some code originally compiled with TC5 in TC7, I tracked
down an apparent problem with the optimizer's code generation.  It seems
that the Code Motion optimization doesn't handle volatile pointer variables
in the expected manner.  With code motion enabled, the reference to the
volatile variable is moved outside of the loop, and a local copy is checked
inside the loop.  The following code demonstrates this in its simplest
form.  Is this a bug or am I missing something?  If the former, is it a
known bug?

Note that the code that was optimized with code motion is less efficient in
time and space, in addition to being incorrect.  The compiler was upgraded
to version 7 using the online updater from the THINK C archives.  (For the
curious, the motivation for this code was to check the status of a global
flag that would be set by a VBL routine.)

main()
{
    /*
     * Declare ptr as a pointer to a volatile short.
     */
    volatile        short           *ptr;

    /*
     * Initialize volatile value to 0.
     * Loop until volatile value is set by VBL routine.
     */
    for( *ptr=0; *ptr == 0; );
}


#if 0

// With Code Motion
main:
00000000: 4E56 FFFE    LINK      A6,#$FFFE
00000004: 2F0C         MOVE.L    A4,-(A7)
00000006: 4254         CLR.W     (A4)
00000008: 3D54 FFFE    MOVE.W    (A4),$FFFE(A6) <-- Incorrectly moves volatile
0000000C: 4A6E FFFE    TST.W     $FFFE(A6)          value to local variable
00000010: 67FA         BEQ.S     *-$0004        ; 0000000C
00000012: 285F         MOVEA.L   (A7)+,A4
00000014: 4E5E         UNLK      A6
00000016: 4E75         RTS
00000018

// Without Code Motion
main:
00000000: 2F0C         MOVE.L    A4,-(A7)
00000002: 4254         CLR.W     (A4)
00000004: 4A54         TST.W     (A4)           <-- Leaves volatile var alone
00000006: 67FC         BEQ.S     *-$0002        ; 00000004
00000008: 285F         MOVEA.L   (A7)+,A4
0000000A: 4E75         RTS
0000000C

#endif


Kirk Swenson
UCB School of Optometry
[e-mail]


>From [e-mail] 27 May 94 20:53:10 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: What update did QuickdrawText.h come in?
X-Sender: [e-mail] (Unverified)
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 25
Date: 27 May 94 20:53:10 GMT

I just downloaded the 7.0.2 update and am attempting to precompile the
headers (as indicated in the instructions). But my attempt to do so bails
on the following:

        // QuickdrawText
        #if 1
                #include <QuickdrawText.h>    <<<<====== NO CAN FIND!
        #endif

I have looked in the 6.0.1 update and in the 7.0 update (NOT to mention the
original 6.0 disks) and can NOT find this file! Am I just blind?

Sign me frustrated in Texas....


----------------------------+-----------------------------------------
Lloyd Sargent               | "It's binary - it either works or it doesn't"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 27 May 94 21:13:51 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: Re: patch
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 52
Date: 27 May 94 21:13:51 GMT

>On Fri, 27 May 1994 01:54:37 MDT Craig Hollabaugh wrote:
>
>    Saw the patch to 7.0.2 and grabbed it.
>
>    Got a "verify resource CNFG 0" error when going from 7.0.1
>
>    So I converted a 6.0.1 to 7.0.1 then to 7.0.2, got the same
>    error.
>
>    Unfortunately, I'm on the road without my 6.0 disks, so I can't
>    start with a "clean" version.
>
>    Does the patch work for you? (Stupid question since you posted it) Are
>    other people having the same problem?
>
>I applied all of the patch modules without error except for the
>Symantec C++ 7.0.2 patch.  It apparently requires a 7.0.1 version of
>the translator, even though the AppleLink message containing the
>update kit says it is for those who used the previously published
>update.
>
>I had not attempted the patch prior to uploading the archive, since
>I retrieved the kit directly from AppleLink and (mistakenly) assumed
>it had been tested first.  I'll upload an update as soon as it is
>made available.
>
>Mark
>
>--------
>Mark D. Nagel <[e-mail]>      Relational Data Systems
>                                             30 Executive Park, Suite 260
>Eat right.  Exercise.  Die anyway.           Irvine, CA 92714
>                                             [phone]


Well, I had to go back and get the bloody original off of 6.0, apply the
6.0->7.0 patch and THEN the 7.0.2 patch (which worked like a charm).
Apparently this resource gets altered and the patch doesn't take that into
account (computer error ;)  I've lost three hours of time already on this
patch (does my frustration show?).... GRRRRRRR....


----------------------------+-----------------------------------------
Lloyd Sargent               | "It's binary - it either works or it doesn't"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 27 May 94 23:40:31 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: SymC++ 7.0.2 updater requires homage to Rube Goldberg?
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 901
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 23
Date: 27 May 94 23:40:31 GMT

In order to apply the SymC++ 7.0.2 patches, it looks like I am going to
have to:

        1) install 6.0 from floppy
        2) download the 6.0.1 updater package
        3) apply the 6.0.1 updater package
        4) download the 7.0 updater package
        5) apply the 7.0 updater package
        6) apply the 7.0.2 updater package

I believe this to be the case because the updater READ ME file suggests
the updater will not work on anything but a virgin installation of
7.0 (or, presumably, 7.0.1, which of course I do not possess). I have
verified this empirically (read: tried updating used files anyway). Past
experience indicates that I am not going to be able to coax the updater
programs to do their stuff against files which have been used.

Does anybody have any suggestions on how to avoid the above sequence?

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "...where they burn books, people are next." -- XTC
>From [e-mail] 28 May 94 02:31:55 GMT
Path: ucivax!gateway
From: [e-mail] (Mauricio Tavares)
Subject: Re: Reading data from a file
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.3 PL2]
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 56
Date: 28 May 94 02:31:55 GMT

> If you figured out how to use iErr = FSRead(fRefNum,&inOutCount,bufPtr);,
> you can use the stdio.h function numargs = sscanf(string,format,); to
> parse your bufPtr string, instead of fscanf.
>
> Hope this helps,

        I think it should, now that I have time to do some programming.
Based on that, I though on doing the following:

1. Find the size of the data fork of the file

I thought on reading the entire file into memory since it is more
convenient for me and I do not expect it to be more than 20K long. Now, how to
find the size of the file?  I thought on using something like GetEOF(), but
will it return the size of the data fork only?

2. Allocate memory
char *fileBuffPtr  = (char*)calloc(fileSize,sizeof(char));

3. Read in the entire file into the pointer
iErr = FSRead(fRefNum,fileSize,fileBuffPtr);

4. Read the data out of the buffer

After some thought, I decided to have my file with the following format:

VARNAME: N

where VARNAME: is the name of the variable and N is its value.  For
instance, I might have something like:

NAME: Moose
DT: 2500
DNU:  .0537
MU: 123e-4
R1: 1e4  .234e4  -1.344e4

This way, I can simply look first for the VARNAME: string and then I read
in the data associated with VARNAME:.  If one of these variables is not
available, I can always return an error code... or even use the default
data if I think the user will be too scared if he sees my dialog. B)

I know that is not as easy as I am making it to be -- it should in fact be
quite complicated since it needs to search around for my magic string and
ignore any leading space -- but I think it worth the effort.  For the time
being, I will instead use sscanf() instead of fscanf(), with small modifcations to allow for testing whether the data was read or not:

if(!sscanf(fileBuffPtr, "%lf\n", &(data->dt));
{
        Complain("dt");
}
[...]

BTW, Have you ever done something like what I want to do in the end
(VARNAME: and such)?

>From [e-mail] 28 May 94 12:26:58 GMT
Path: ucivax!gateway
From: [e-mail] (Markus Waldorf)
Subject: HLock/HUnlock ?
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset="us-ascii"
MIME-version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Date-warning: Date header was inserted by AdminA.RFERL.Org
Lines: 27
Date: 28 May 94 12:26:58 GMT

Hi,


Remember, handles are pointers to pointers. If you ever want to singly
dereference a handle, make sure you lock it first!

                           "Macintosh C Programming Primer, Vol. 1"


For instance: HLock( (Handle)iconHandle );
              HUnlock( (Handle)iconHandle );


English is not my native language, perhaps that's why I don't know
what "singly dereference" means. Could someone please give me a further
explanation? Many Thanks!


--------------------------------------------------------------------------
_/ _/   _/_/_/ _/_/_/   Markus Waldorf                      _/ _/   _/
_/  _/  _/     _/       Information Technology Consultant   _/  _/  _/
_/_/    _/_/   _/_/     RFE/RL Inc, Munich, Germany         _/_/    _/
_/  _/  _/     _/       Compu$erve: 100022,2373             _/  _/  _/
_/  _/  _/     _/_/_/   Internet: [e-mail]        _/  _/  _/_/_/
--------------------------------------------------------------------------


>From [e-mail] 28 May 94 15:13:13 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: Probs with 7.0 code generation...
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 28
Date: 28 May 94 15:13:13 GMT

I have been trying to get this thing to spit out the right code for the
last 4 days. I have consistantly fail.  The code is simple and looks like
the following:

Session->paramblock.startParam.ioCompletion = (ProcPtr) OpenSessionCallback;

Fortunately, MOST of the code seems to be correct. It comes out as:

        LEA OpenSessionCallBack,A0

What is weird is that it generates the correct offset (EA), but tries to
address it off of A5 instead of the PC. Has anyone else had this prob and
do they have a solution? This is really annoying.

Thanks!



----------------------------+-----------------------------------------
Lloyd Sargent               | "It's binary - it either works or it doesn't"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 28 May 94 15:52:25 GMT
Path: ucivax!gateway
From: [e-mail] (Steve Madancy)
Subject: Re: SymC++ 7.0.2 updater requires homage to Rube Goldberg?
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 27
Date: 28 May 94 15:52:25 GMT

On 27 May 1994, Pete Gontier wrote:

> In order to apply the SymC++ 7.0.2 patches, it looks like I am going to
> have to:
>
>       1) install 6.0 from floppy
>       2) download the 6.0.1 updater package
>       3) apply the 6.0.1 updater package
>       4) download the 7.0 updater package
>       5) apply the 7.0 updater package
>       6) apply the 7.0.2 updater package
>
> I believe this to be the case because the updater READ ME file suggests
> the updater will not work on anything but a virgin installation of
> 7.0 (or, presumably, 7.0.1, which of course I do not possess). I have
> verified this empirically (read: tried updating used files anyway). Past
> experience indicates that I am not going to be able to coax the updater
> programs to do their stuff against files which have been used.
>
MY question is, IS it possible to run the upgrader on a virgin 7.0.1
installation? How about some help from Symantec on this one?


----------------------------
Steve Madancy
[e-mail]

>From [e-mail] 28 May 94 16:47:40 GMT
Path: ucivax!gateway
From: [e-mail] (WalrathW)
Subject: RE: What update did QuickdrawText.h come in?
Message-ID: <[e-mail]>
Content-transfer-encoding: 7BIT
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 27
Date: 28 May 94 16:47:40 GMT


>
_______________________________________________________________________________
> From: Lloyd Sargent on Sat, May 28, 1994 5:15
> Subject: What update did QuickdrawText.h come in?
> To: [e-mail]
>
> I just downloaded the 7.0.2 update and am attempting to precompile the
> headers (as indicated in the instructions). But my attempt to do so
> bails
> on the following:
>
>         // QuickdrawText
>         #if 1
>                 #include <QuickdrawText.h>    <<<<====== NO CAN FIND!
>         #endif
>
> I have looked in the 6.0.1 update and in the 7.0 update (NOT to
> mention the
> original 6.0 disks) and can NOT find this file! Am I just blind?
>
> Sign me frustrated in Texas....

Perhaps the best advice is never to apply a Symantec patch when you have
deadlines to meet! I've been afraid of even trying to update my version.

Could it be a Universal header?
>From [e-mail] 28 May 94 17:04:04 GMT
Path: ucivax!gateway
From: [e-mail] (Jean Berthomieu)
Subject: EDS-7.0.2 - CNFG problem, the clue(?)
Message-ID: <[e-mail]>
Content-Type: text
Content-Length: 743
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 20
Date: 28 May 94 17:04:04 GMT


I too had to reinstall the C++ translator from the original disks to
allow the Symantec C++ 7.0.2 patch to be applied. That is: install 6.0,
update to 7.0, then update to 7.0.2 :-(

Then I used ResCompare between the 7.0 version I had previously used and
a copy of the new one on which the patch program did work.

Now, that's the clue: The CNFG resources both had a name, and they should
not (to be patched).

So instead of reinstalling the whole stuff, you may use Resedit and clear
the resources names for CNFG 0 and 1 (in the Get Info dialog).
Then you can apply the patch!

Use a copy of the C++ translator to do this, and since you may not change
its name, do it in an other place than the Translators folder, e.g. on the
desktop ;-)


>From [e-mail] 28 May 94 17:44:38 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: SymC++ 7.0.2 updater requires homage to Rube Goldberg?
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Pete Gontier" at May 27, 94 11:40:31 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1483
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 35
Date: 28 May 94 17:44:38 GMT

> In order to apply the SymC++ 7.0.2 patches, it looks like I am going to
> have to:
>
>       1) install 6.0 from floppy
>       2) download the 6.0.1 updater package
>       3) apply the 6.0.1 updater package
>       4) download the 7.0 updater package
>       5) apply the 7.0 updater package
>       6) apply the 7.0.2 updater package

Well, it turns out that the first patch I was trying to apply was the
C++ translator patch. The TPM patch and the Debugger patch work fine.

The C++ translator patcher apparently balks because a CNFG resource in
the 7.0 translator has a name and it expects that resource to not have
a name. So, based on a hint from Usenet, I nuked the name and the
patcher app worked.

The THINK C translator patcher I still can't apply because even though
the patcher fails with a similar complaint, the resources in question
don't have names, so there are no names to delete. Maybe I should *add*
names. Hmmmm.

All of this hassle (both for us and for Symantec, who has to listen to
us whine) leads me to wonder: since the translators are useless without
TPM, why not just distribute whole translator files? Are they that
large? <checking> Hmmm. Well, sort of. 440K for the C++ translator. I'd
be happier waiting a little longer for that download than going through
the 1 thru 6 sequence above. Undoubtedly there are people who would
complain about it.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 You thought Obfuscated C was confusing? Wait for Obfuscated C++!
>From [e-mail] 28 May 94 18:03:23 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: HLock/HUnlock ?
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Markus Waldorf" at May 28, 94 12:26:58 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 3020
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 61
Date: 28 May 94 18:03:23 GMT

> "Macintosh C Programming Primer, Vol. 1" writes:
>
>> Remember, handles are pointers to pointers. If you ever want to singly
>> dereference a handle, make sure you lock it first!
>
> English is not my native language, perhaps that's why I don't know
> what "singly dereference" means. Could someone please give me a further
> explanation? Many Thanks!

Firstly, if that's all the book says, it's wrong. More on that later.
For now, here is an example of "half dereferencing" a handle. (I don't
know where the book got "singly dereference" -- it's not jargon I've
seen before, though I can parse it.)

        Handle  h                       = NewHandle (1);
        Ptr     halfDereferenced        = *h;
        char    fullyDereferenced       = **h;

Since the Memory Manager sometimes moves the memory block around to
reduce fragmentation, the 'Ptr' can become stale. Describing the
"sometimes" is what your book appears to neglect explaining.

See, lots of system calls rely on the Memory Manager. Lots of system
calls rely on other system calls which rely on the Memory Manager. What
happens is that a routine that looks totally innocent might cause memory
to move just because it relies on some other innocent-looking routine
which relies on still other routines which ultimately rely on the Memory
Manager. My policy, rather than memorizing or looking up which calls
move memory, is to assume any system call moves memory. (There isn't
a list of calls which do/n't move memory that Apple hasn't disowned
anyway.)

This is an especially useful assumption when you consider there are lots
of badly written system extensions out there which cause memory to move
during system calls which are not supposed to move memory. And there's
always the possibility that Apple will change a system call to move
memory; 'SysBeep', for example, used to be safe, but then it was changed
to support digitized sound resources and suddenly it needed to move
memory.

Another thing to watch out for is subroutine calls across code segments.
The Segment Loader can move memory while it loads in a code resource
which is not yet present in memory. (Loading and locking all your 'CODE'
resources at startup can prevent this.)

However, in the code sample above, there is no risk, and that's why I
think the explanation in your book is suspect. See, a C compiler is
never going to generate a function which itself moves memory. It doesn't
know how; it's beyond the scope of the language. So you can do whatever
you want with a half-dereferenced handle as long as you don't make a
function call -- and sometimes you can even make a function call if you
can guarantee the called function lives in the same segment -- but that
gets to be a maintenance headache.

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 "...someone not acquainted with the Christian mythology of the
 Crucifixion might consider a Crucifix to be a particularly sadistic
 piece of erotica... We of the ACLU will continue to defend your right to
 worship such objects if it pleases you." -- Gregory J. Wageman
>From [e-mail] 28 May 94 22:32:30 GMT
Path: ucivax!gateway
From: [e-mail] (Marco Sandri)
Subject: Re: SymC++ 7.0.2 updater requires...
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1263
MIME-Version: 1.0
Newsgroups: fa.think-c
Reply-To: Marco Sandri <[e-mail]>
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 34
Date: 28 May 94 22:32:30 GMT

Hi Pete,

> In order to apply the SymC++ 7.0.2 patches, it looks like I am going to
> have to:
>
>       1) install 6.0 from floppy
>       2) download the 6.0.1 updater package
>       3) apply the 6.0.1 updater package
>       4) download the 7.0 updater package
>       5) apply the 7.0 updater package
>       6) apply the 7.0.2 updater package
>
> I believe this to be the case because the updater READ ME file suggests
> the updater will not work on anything but a virgin installation of
> 7.0 (or, presumably, 7.0.1, which of course I do not possess). I have
> verified this empirically (read: tried updating used files anyway). Past
> experience indicates that I am not going to be able to coax the updater
> programs to do their stuff against files which have been used.
>
> Does anybody have any suggestions on how to avoid the above sequence?

I simply stored a perfectly clean installation (never used) on a cartridge;
every time I get a patch, I apply this patch to a copy of this installation
and, if all completes successfully, I replace the previous copy with
the newer one.

Bye,
Marco

============================================================
  Marco Sandri
  [e-mail]
  [e-mail]
============================================================
>From [e-mail] 29 May 94 01:47:28 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: RE: What update did QuickdrawText.h come in?
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 52
Date: 29 May 94 01:47:28 GMT

>>
>_______________________________________________________________________________
>> From: Lloyd Sargent on Sat, May 28, 1994 5:15
>> Subject: What update did QuickdrawText.h come in?
>> To: [e-mail]
>>
>> I just downloaded the 7.0.2 update and am attempting to precompile the
>> headers (as indicated in the instructions). But my attempt to do so
>> bails
>> on the following:
>>
>>         // QuickdrawText
>>         #if 1
>>                 #include <QuickdrawText.h>    <<<<====== NO CAN FIND!
>>         #endif
>>
>> I have looked in the 6.0.1 update and in the 7.0 update (NOT to
>> mention the
>> original 6.0 disks) and can NOT find this file! Am I just blind?
>>
>> Sign me frustrated in Texas....
>
>Perhaps the best advice is never to apply a Symantec patch when you have
>deadlines to meet! I've been afraid of even trying to update my version.
>
>Could it be a Universal header?
               ^^^^^^^^^^^^^^^^

Heh, heh, apparently it doesn't exist in MY limited universe!

These are the three include files that I had to #if out  in order to
precompile everything...

                #include <QuickdrawText.h>
                #include <TextServices.h>
                #include <TextUtils.h>

Now, my question is, did Symantec FAIL to include these in the update? If
so, then how do I get them?



----------------------------+-----------------------------------------
Lloyd Sargent               | "It's binary - it either works or it doesn't"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 29 May 94 18:20:52 GMT
Path: ucivax!gateway
From: [e-mail] (Will Glass-Husain)
Subject: how to turn on program linking without chooser
Message-ID: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
content-length: 146
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 5
Date: 29 May 94 18:20:52 GMT


Does anyone know how to turn on program linking in sys 7 without using
the chooser?  I can't seem to find anything in Think Reference.

-->Will
>From [e-mail] 29 May 94 21:32:22 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: Mailing list
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 2
Date: 29 May 94 21:32:22 GMT

Please remove from mailing list.

>From [e-mail] 29 May 94 22:52:36 GMT
Path: ucivax!gateway
From: [e-mail] (Marco Sandri)
Subject: RE: What update did QuickdrawText.h
Message-ID: <[e-mail]>
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 1823
MIME-Version: 1.0
Newsgroups: fa.think-c
Reply-To: Marco Sandri <[e-mail]>
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 55
Date: 29 May 94 22:52:36 GMT

Hi,

>>_______________________________________________________________________________
>>> From: Lloyd Sargent on Sat, May 28, 1994 5:15
>>> Subject: What update did QuickdrawText.h come in?
>>> To: [e-mail]
>>>
>>> I just downloaded the 7.0.2 update and am attempting to precompile the
>>> headers (as indicated in the instructions). But my attempt to do so
>>> bails
>>> on the following:
>>>
>>>         // QuickdrawText
>>>         #if 1
>>>                 #include <QuickdrawText.h>    <<<<====== NO CAN FIND!
>>>         #endif
>>>
>>> I have looked in the 6.0.1 update and in the 7.0 update (NOT to
>>> mention the
>>> original 6.0 disks) and can NOT find this file! Am I just blind?
>>>
>>> Sign me frustrated in Texas....
>>
>>Perhaps the best advice is never to apply a Symantec patch when you have
>>deadlines to meet! I've been afraid of even trying to update my version.
>>
>>Could it be a Universal header?
>               ^^^^^^^^^^^^^^^^
>
>Heh, heh, apparently it doesn't exist in MY limited universe!
>
>These are the three include files that I had to #if out  in order to
>precompile everything...
>
>                #include <QuickdrawText.h>
>                #include <TextServices.h>
>                #include <TextUtils.h>
>
>Now, my question is, did Symantec FAIL to include these in the update? If
>so, then how do I get them?

I think Symantec includes them only in 7.0 original package, along with
PPC translator...
I found them on the Develop CD issue 16. There was a pre-release of
Universal Headers; probabily the issue 17 contains a release.. I don't know,
I haven't seen it yet.

Bye,
Marco

============================================================
  Marco Sandri
  [e-mail]
  [e-mail]
============================================================
>From [e-mail] 30 May 94 03:51:31 GMT
Path: ucivax!gateway
From: [e-mail]
Subject: *** Custom ICON summary ***
Message-ID: <[e-mail]>
X-Mailer: America Online Mailer
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 148
Date: 30 May 94 03:51:31 GMT

Dear netters:

I really appreciate everyone who helped me...
This is what I asked:

Q: -----------------------------------
Could anybody please tell me
how to create a custom icon for my application?  After build my application,
it always gives simple generic icon for the application.  I used ResEdit and
create an icon and made BNDL resource that has 'APPL' and the icon I created.
 And then I assinged arbitrary signature in the signature field(4 characters
long).  But still, I don't get the icon for my application.  Do I have to add
some code into my source code to do this?

A:------------------------------------
1]==========================================
If you have previously given your application a custom Icon before (Get
Info cut&paste),
make sure you have NOT selected the same one in your BNDL resource (ID -
16455).
And just in case, use Resedit (File Get file/folder info) etc. and check if
your application
still has the custom Icon flag - if so, turn it off.

Try rebuilding the desktop. I would copy the application to a floppy disk,
then re-insert the
disk with Command and Option key pressed down. However, sometimes this is
not completely
effective because the original desktop file was corrupted, so rebuilding it
only yields an
updated desktop file which is still damaged. There are free tools that can
help like
DestopReset or Techtool, but you can also use Resedit. Change the file
attribute of the
Desktop DF and Desktop DB file to visible, put the two files in the trash
and restart.


Best Regards,

Markus Waldorf
Information Technology Support Consultant
RFE/RL Inc, Munich, Germany

2]==============================================
Rebuild the desktop or set the init finder flag to false and restart. The
BNDL needs to be added to the Desktop database.

Lance

3]==============================================
Hey,

Assuming system 7, this is quite easy. After drawing your icon in whatever
drawing program you like, simply copy your icon to
the clipboard, do a "get info" [command - I] on your app,
 and paste it onto your app's icon in the "get info" box of your app.

You can also get icons from other files/folders by copying them from
their respective 'get info' boxes.  Have fun.

        -Lar
4]==============================================
Try rebuilding your desktop (hold down cmd-option keys while rebooting your
mac). If that doesn't work than you must have screwed up somthing with the
resources. Make sure you also have an FREF resource (I think it gets created
when you make the BNDL, though). Go into another application's resources and
compare against theirs, making certain that yours looks more or less the
same.

cheers,
  -wayne

5]==============================================
Make sure you've set the BNDL bit in your application's file info.
You may also have to rebuild your desktop. Ie hold down ctrl and cmd
during boot (just before the finder starts).
--
Amrik Thethi.   Tel. [phone] Fax. [phone]
Setanta Software Ltd.  Internet: [e-mail]
Cambridge, UK.

6]==============================================
Hi,

You should do two things:
1) Set the hasCustomIcon bit with ResEdit's "Get File Info..."  menu (or with
   utilities like FileTyper, or FileBuddy, if you have them...).
2) Include in your application's resource fork an icon family (ics, icl and
   ICON resources) with ID=... hmmm... I don't remember the ID, nor have
   access to IM volumes, at this time...
   You can obtain the right ID by pasting an icon onto a file on the Finder,
   and looking into their resource fork. It should be -16xxx.

Hope this helps.

Bye,
Marco

7]==============================================
Well, given that you have set the BNDL correctly, you will probably find that
your custom icon will only appear once you have rebuilt the desktop.

(To bebuild your desktop, restart, hold the command and option keys down
until
your machine asks you if you want to rebuild the desktop).

Lindsay

8]==============================================
There should also be an FREF resource. You have to rebuild the desktop
before getting the icon. When you create the file you chould also
check or uncheck the box 'initted', I always forget which. But, I believe
you should uncheck it, because I THINK that this indicates whether it has
been processed for the desktop file or not. But I'm not sure about the
last thing.

  Good Luck,
     Olle
---------------------------------------------------
- #1,3,6 method is same.  This method works fine, but the icon can be removed
by pressing delete key after select the icon from Get Info menu of System 7.
 I guess my question was unclear.  I wanted to have a custom icon that cannot
be deleted by pressing the delete key.
- I found what I did wrong.  After I compiled my program, the compiler
assigned "????" for the "Creator".  After I assigned  as same signature to
the "Creator" field as BNDL resource's signature, it worked.

*for the people who is a beginner like me*
1) by using ResEdit, select Get Info from the menu, and choose your
application.
2) set Creator signature as you like.  (4 characters long.  ex. 'Test')
3) open your application by ResEdit
4) create all the icons.  (icl, ics...etc)
5) Create BNDL resource. ('FREF' resource creates automatically)
6) from BNDL, you'll see this and set it as:
Signature [Test] <--- Same signature as "Creator"
Type | Finder Icons
-------------
APPL| <Select Icon  you created >

I guess I made a stupid mistake, and asked for help with unclear question.
 Anyways, I really appreciate everyone who has answered.

Sincerely,
[e-mail]


>From [e-mail] 30 May 94 17:01:20 GMT
Path: ucivax!gateway
From: [e-mail] (Lloyd Sargent)
Subject: Address of Function Problem (for callback)...
Message-ID: <[e-mail]>
Content-Type: text/plain; charset="us-ascii"
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 26
Date: 30 May 94 17:01:20 GMT

I have been trying to get this thing to spit out the right code for the
last 4 days. I have consistantly fail.  The code is simple and looks like
the following:

Session->paramblock.startParam.ioCompletion = (ProcPtr) OpenSessionCallback;

Fortunately, MOST of the code seems to be correct. It comes out as:

        LEA OpenSessionCallBack,A0

What is weird is that it generates the correct offset (EA), but tries to
address it off of A5 instead of the PC. Has anyone else had this prob and
do they have a solution? This is really annoying.

Thanks!

----------------------------+-----------------------------------------
Lloyd Sargent               | "It's binary - it either works or it doesn't"
Head Honcho                 | Internet:  [e-mail]
Canna Software Development  | AOL:       [e-mail]
14308-B FM 812              | Phone:     [phone]
Del Valle, Texas 78617      | Fax:       [phone]  (8-5 CST)

    "My company believes in First Admendment rights. So there!"


>From [e-mail] 31 May 94 00:45:50 GMT
Path: ucivax!gateway
From: [e-mail] (Joel Henderson)
Subject: Where's the FAQ & archives?
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Organization: EOS / Educational Online Sources <[e-mail]>
Lines: 36
Date: 31 May 94 00:45:50 GMT

Trivial question, sorry to bother you: where's the FAQ & archives?
Better question: how would I have found out for myself, instead of posting?

Real question: I'm helping a school set up a "mini-server" of hypercard
stacks (yes, this will relate to TCL) accessible via Mosaic. Kids browse
html descriptions of projects, click on one they want to see, and Mosaic
downloads the stack and pipes it into hypercard.

I'd like for hypercard to be able to "talk back" to Mosaic, so that
buttons can contain hrefs. I assume I'll use XCMD-->AppleEvents, unless
someone suggests something better. So I'm looking for TCL classes for
AppleEvents, ideally with a small demo program, so that I can learn
them. I'm also looking for classes to handle display & formatting
of HTML, or possibly a more general tokenizer class.

One person has suggested skipping Mosaic and using hypercard as the
browser, and sending AppleScript commands to Fetch to ftp the stacks.
How does TCL work with AppleScript? I was under the impression that
A.S. was merely an easy way to write AppleEvents.

The AppleEvents are also necessary to write a small app which gives each
teacher a small edittext window which talks to a centralized admin/tech
person. I know a zillon things like this must have been written, but they
want it incorporated into an existing TCL project and tweaked (security
in particular, meaning that if someone has a DES or Kerberos class I'd be
thrilled).

Thanks in advance for help.

- Joel

--------------------------------------------------------------------------
[e-mail]           Web--> http://home.eos.brown.edu/eos1/
EOS: Educational Online Sources    FTP/Gopher--> home.eos.brown.edu
Box 2325 / Providence, RI 02906    Send mail to [e-mail]
[e-mail] / 1-800-ask-eos1    and somewhere in it write: help!
>From [e-mail] 31 May 94 07:38:34 GMT
Path: ucivax!gateway
From: [e-mail] (Big Nose)
Subject: Symantec 7.0.2 updater homage requirements
X400-Originator: [e-mail]
Content-Identifier: 1185FA200200
Message-ID: <[e-mail]>
X400-Recipients: [e-mail]
Sensitivity: Company-Confidential
Newsgroups: fa.think-c
X-Message-ID: <2716340831051994/A17043/IAPE/1185FA200200*@MHS>
X400-Received: by mta sun.mhs-relay.ac.uk in /PRMD=uk.ac/ADMD= /C=gb/; Relayed; Tue, 31 May 1994 08:35:24 +0100
X400-Received: by mta arcs.afrc.ac.uk in /PRMD=UK.AC/ADMD= /C=GB/; Relayed; Tue, 31 May 1994 08:36:31 +0100
X400-Received: by /PRMD=UK.AC/ADMD= /C=GB/; converted (ia5 text (2)); Relayed; Tue, 31 May 1994 09:34:24 +0100
Approved: [e-mail]
X400-Content-Type: P2-1988 (22)
Lines: 14
Date: 31 May 94 07:38:34 GMT
X400-MTS-Identifier: [/PRMD=UK.AC/ADMD= /C=GB/;2716340831051994/A17043/IAPE]


Pete Gontier wrote:

[... stuff about rigmarole of updating deleted ...]

>Does anybody have any suggestions how to avoid the above sequence?

(or words to that effect)

Buy CodeWarrior? :)


Andy Law
( Lawa @ afrc.ac.uk                             Big Nose in Edinburgh )
>From [e-mail] 31 May 94 12:09:30 GMT
Path: ucivax!gateway
From: [e-mail] (Pindar Infotek Ltd)
Subject: sfHookNullItem constant
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
Reply-To: [e-mail]
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 22
Date: 31 May 94 12:09:30 GMT


>From Pindar user Ian McCall

Subject: _sfHookNullItem constant
I'm trying to put together a dialogue hook for CustemGetFile that
does some processing on the sfHookNullEvent. At the end of this
processing, according to THINK Reference and Inside Mac VI 26-12 I'm
supposed to return the constant sfHookNullItem.

The trouble is - no such constant is defined! Anywhere, in any
header. At the moment I'm returning zero, but I've no idea if this is
what the constant should be.

Anyone ever found this constant? Or does anyone know what I should
return after a sfHookNullEvent?



---
Please make 'Attn:<space>Ian McCall' the first line of your reply
(not the subject) to ensure automatic forwarding to the correct
Pindar user.
>From [e-mail] 31 May 94 15:34:23 GMT
Path: ucivax!gateway
From: [e-mail] (Diego Akerman)
Subject: RE: how to turn on program linking without chooser
Message-ID: <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 19
Date: 31 May 94 15:34:23 GMT

Hi Will!
>Does anyone know how to turn on program linking in sys 7 without using
>the chooser?  I can't seem to find anything in Think Reference.
>
>-->Will

I think you are mistaken, program linking is not turned on from the Chooser, you
toggle it from the Sharing Setup control panel. On the other hand, there are a
couple of utilities that turn on/off program linking without the control panel
intervention. I think they are from Ambrosia Software.

Hope this helps. Doglan//
________________________________________________________________________________
  _/_/_/_/ _/_/_/_/ _/_/_/_/ _/_/  _/_/ _/  | DiegO GabrieL AkermaN
 _/       _/    _/ _/    _/ _/  _/  _/ _/   |
_/_/_/_/ _/_/_/_/ _/_/_/_/ _/  _/  _/ _/    | Av. Jorge Newbery 3800
     _/ _/       _/    _/ _/      _/        | 1427 Buenos Aires, Argentina
_/_/_/ _/       _/    _/ _/      _/ _/      |
                                            | E-Mail: [e-mail]
>From [e-mail] 31 May 94 16:21:53 GMT
Path: ucivax!gateway
From: [e-mail] ("Mark D. Nagel")
Subject: ADMIN: list future (was Re: Symantec 7.0.2 updater...)
Message-ID: <[e-mail]>
In-reply-to: Big Nose's message of 31 May 1994 07:38:34 GMT.
        <[e-mail]>
Newsgroups: fa.think-c
X-Message-ID: <27106.770400594@rdatasys>
Reply-To: [e-mail]
Approved: [e-mail]
Organization: Relational Data Systems, Irvine, CA
Lines: 43
Date: 31 May 94 16:21:53 GMT
Phone: [phone]

On 31 May 1994 07:38:34 GMT Big Nose wrote:

    Pete Gontier wrote:

    [... stuff about rigmarole of updating deleted ...]

    >Does anybody have any suggestions how to avoid the above sequence?

    (or words to that effect)

    Buy CodeWarrior? :)

Good answer!  This actually is an interesting point that has some
bearing on the future of this list.  Currently, this list is
chartered to discuss THINK C and Symantec C++ compilers and
associated issues.  It should certainly continue to do so, but with
the introduction of CodeWarrior, I wonder whether it would be a good
idea to expand the charter to discuss that as well.  I know that my
next compiler purchase will be from Metrowerks and I imagine that is
true for many others on this list.  On the other hand, there probably
already is a list for CodeWarrior, so it would be redundant to
overlap with that lists discussions.

What do others think about this?

On another administrative note, I want everyone to be aware that
there will soon be a major shift in the way the list works as I move
its base of operations from UC Irvine to Relational Data Systems.  I
am currently testing the ListProcessor mailing list package and I'm
pleased with it so far.  It will allow for archive browsing and
retrieval, automatic list subscription/deletion, options such as
digesting of list mail, and many other cool features.  I will
attempt to make the transition as smooth as possible, but there are
bound to be _some_ problems.  Consider this advance warning.  I will
make another announcement prior to the switchover and I intend to
resolve all problems as quickly as possible.

Mark
--------
Mark D. Nagel <[e-mail]>      Relational Data Systems
                                             30 Executive Park, Suite 260
Eat right.  Exercise.  Die anyway.           Irvine, CA 92714
                                             [phone]
>From [e-mail] 31 May 94 19:51:43 GMT
Path: ucivax!gateway
From: [e-mail] (Pete Gontier)
Subject: Re: EDS-7.0.2 - CNFG problem, the clue(?)
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Jean Berthomieu" at May 28, 94 05:04:04 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Content-Length: 754
MIME-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 15
Date: 31 May 94 19:51:43 GMT

> Now, that's the clue: The CNFG resources both had a name, and they
> should not (to be patched). So instead of reinstalling the whole
> stuff, you may use Resedit and clear the resources names for CNFG 0
> and 1 (in the Get Info dialog). Then you can apply the patch!

I managed to do this, but I also had a similar problem with the C
translator. Unfortunately, the CNFG resources in the C translator have
no names, so they cannot be deleted! :-) Anybody have clues on this?

--
 Pete Gontier, CTO, Integer Poet Software; [e-mail]

 '"We believe the customer goes to retail stores to buy Microsoft
 software," said Ed Belleba, director of retail strategy at Microsoft.'
                                                     -- MacWEEK 4/18/94
>From [e-mail] 31 May 94 21:19:04 GMT
Path: ucivax!gateway
From: [e-mail] (Jean Berthomieu)
Subject: Re: EDS-7.0.2 - CNFG problem, the clue(?)
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]> from "Pete Gontier" at May 31, 94 12:51:28 pm
Content-Type: text
Content-Length: 487
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 11
Date: 31 May 94 21:19:04 GMT

This is a pure guess, but I imagine the CNFG resources keep your preferences
settings (?) If you reset them to "Factory settings",
..maybe the resource is reset to the right value for the patch to apply ?
I had no problem with the C translator, (and I wouldn't try to redo the whole
process just to know if I'm true :-)), in any case it should not be a big
work to try it while you still have the 7.0 version.

Best regards.

   Jean Berthomieu <[e-mail]>

>From [e-mail] 31 May 94 22:09:00 GMT
Path: ucivax!gateway
From: [e-mail] (Markus Waldorf)
Subject: using Gestalt to get system version
Content-transfer-encoding: 7BIT
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 15
Date-warning: Date header was inserted by AdminA.RFERL.Org
Date: 31 May 94 22:09:00 GMT
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-type: text/plain; charset="us-ascii"
MIME-version: 1.0
Newsgroups: fa.think-c

Hi,

        I assume a Gestalt call is a good way to obtain the version
of the current system software. Can someone please provide me some
info? Thanks!

--------------------------------------------------------------------------
    _/_/    _/_/_/ _/_/_/    Markus Waldorf                     _/_/    _/
   _/  _/  _/     _/        Information Technology Consultant  _/  _/  _/
  _/_/    _/_/   _/_/      RFE/RL Inc, Munich, Germany        _/_/    _/
 _/  _/  _/     _/        Compu$erve: 100022,2373            _/  _/  _/
_/  _/  _/     _/_/_/    Internet: [e-mail]       _/  _/  _/_/_/
--------------------------------------------------------------------------


>From [e-mail] 31 May 94 22:19:19 GMT
Path: ucivax!gateway
From: [e-mail] (Christopher Klaus Larson)
Subject: Re: sfHookNullItem constant
Message-ID: <[e-mail]>
In-Reply-To: <[e-mail]>
Content-Type: TEXT/PLAIN; charset=US-ASCII
Mime-Version: 1.0
Newsgroups: fa.think-c
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 15
Date: 31 May 94 22:19:19 GMT

On 31 May 1994, Pindar Infotek Ltd wrote:

> Anyone ever found this constant? Or does anyone know what I should
> return after a sfHookNullEvent?

If memory serves, you should just return sfHookNullEvent.

--Chris
_______________________________________________________________________________
Chris Larson -- Amateur Macintosh Geek,
Davis Institute of Slowly and Painfully Working Out the Surprisingly Obvious

(Insert disclaimer here)
Internet: [e-mail]

>From [e-mail] 31 May 94 22:29:33 GMT
Path: ucivax!gateway
From: [e-mail] (Markus Waldorf)
Subject: KeyMap and codes
Content-transfer-encoding: 7BIT
X-Message-ID: <[e-mail]>
Approved: [e-mail]
Lines: 32
Date-warning: Date header was inserted by AdminA.RFERL.Org
Date: 31 May 94 22:29:33 GMT
X-Sender: [e-mail]
Message-ID: <[e-mail]>
Content-type: text/plain; charset="us-ascii"
MIME-version: 1.0
Newsgroups: fa.think-c

So far it's clear, but one thing that I don understand (not sure) is,
why the Shift key is Hex 38. According to ASCII, this is "8".


KeyMap  keys;                   // defined in <Events.h>

/*
        typedef long KeyMap[4]<->16 byte binary map of key states.
        The 16 keys bytes are in left to right order but the bits are
        in right to left order in each byte.

        The Shift key is 0x38 which is decimal 56 and dividing by 8
        (56/8=7) means it's bit is in the 7th byte which is the rightmost
        byte of keys[1] and is the rightmost bit in that byte
        (56/8=7 remainder 0).
*/


        GetKeys( keys );

        if ( Button() || ( 1 & keys[1] ) )
        ....

--------------------------------------------------------------------------
    _/_/    _/_/_/ _/_/_/    Markus Waldorf                     _/_/    _/
   _/  _/  _/     _/        Information Technology Consultant  _/  _/  _/
  _/_/    _/_/   _/_/      RFE/RL Inc, Munich, Germany        _/_/    _/
 _/  _/  _/     _/        Compu$erve: 100022,2373            _/  _/  _/
_/  _/  _/     _/_/_/    Internet: [e-mail]       _/  _/  _/_/_/
--------------------------------------------------------------------------


------------------------------ Cut here ------------------------------