Celestin Apprentice 4/Apprentice-Release4.iso/Information/Digests/CSMP Digest/volume 3/csmp-digest-v3-063/doubleCR.1

Text File | 1995-12-31 | 84.1 KB | 2,142 lines

C.S.M.P. Digest             Thu, 06 Oct 94       Volume 3 : Issue 63

Today's Topics:

        <command>+? == PLAIN STUPID!!! (true)
        Apple Events Question
        File System Manager Software Development Kit
        GDGX Graphics Shell bus errors with GX Beta 3...
        Getting Started (Summary)
        Help: idle, shutdown, and MacTCP examples..
        Pascal Flavours for Macintosh
        Q: Script-Independant Command-Keys?
        XCMD memory residence



The Comp.Sys.Mac.Programmer Digest is moderated by Francois Pottier
([e-mail]).

The digest is a collection of article threads from the internet newsgroup
comp.sys.mac.programmer.  It is designed for people who read c.s.m.p. semi-
regularly and want an archive of the discussions.  If you don't know what a
newsgroup is, you probably don't have access to it.  Ask your systems
administrator(s) for details.  If you don't have access to news, you may
still be able to post messages to the group by using a mail server like
anon.penet.fi (mail [e-mail] for more information).

Each issue of the digest contains one or more sets of articles (called
threads), with each set corresponding to a 'discussion' of a particular
subject.  The articles are not edited; all articles included in this digest
are in their original posted form (as received by our news server at
nef.ens.fr).  Article threads are not added to the digest until the last
article added to the thread is at least two weeks old (this is to ensure that
the thread is dead before adding it to the digest).  Article threads that
consist of only one message are generally not included in the digest.

The digest is officially distributed by two means, by email and ftp.

If you want to receive the digest by mail, send email to [e-mail]
with no subject and one of the following commands as body:
    help                        Sends you a summary of commands
    subscribe csmp-digest Your Name    Adds you to the mailing list
    signoff csmp-digest            Removes you from the list
Once you have subscribed, you will automatically receive each new
issue as it is created.

The official ftp info is //ftp.dartmouth.edu/pub/csmp-digest.
Questions related to the ftp site should be directed to
[e-mail]. Currently no previous volumes of the CSMP
digest are available there.

Also, the digests are available to WAIS users.  To search back issues
with WAIS, use comp.sys.mac.programmer.src. With Mosaic, use
http://www.wais.com/wais-dbs/comp.sys.mac.programmer.html.


-------------------------------------------------------

>From [e-mail] (Anders Lundqvist)
Subject: <command>+? == PLAIN STUPID!!! (true)
Date: 19 Sep 1994 06:42:44 GMT
Organization: Dept. of E.E., Univ. of Linkoping

Greetings all programmers!

I wonder how many of you that implement keyboard shortcuts
like <command>+? realize that these DO NOT work on many
international keyboard layouts like the swedish????

Why? Because in order to produce a question mark I need to
use the shift key and since the command key disables the shift
key, this doesn't work. I have been playing around with the DA
Keyboard and there's simply no way around. Obiwan uses <command>+`
and this I can emulate by pressing <command>+<option>+<control>+'.
Sigh...

Of course, this wouldn't been so bad if there was an equivalent
meny command, but in many cases there isn't. Usually <command>+?
invokes some help mode and turns the pointer into a question mark
to allow the user to point at an object (e.g Disinfectant and
Mathematica).

IS THIS PROBLEM MENTIONED IN THE HUMAN INTERFACE GUIDELINES???


Thank you for any feedback!

Anders Lundqvist
Department of EE
Linkoping University
Sweden








+++++++++++++++++++++++++++

>From [e-mail] (Philip Machanick)
Date: 19 Sep 1994 08:16:47 GMT
Organization: Computer Science Dept, U of Witwatersrand

In article <35jbt4$[e-mail]>, [e-mail] wrote:

> Why? Because in order to produce a question mark I need to
> use the shift key and since the command key disables the shift
> key, this doesn't work.

This is also true on the English and US keyboards: "?" is shift-/. Where
did you get the idea that command disables shift? All the modifiers are
independent and can be sensed independently. If COMMAND-? is not working
for you there must be some other explanation.
--
Philip Machanick                   [e-mail]
Department of Computer Science, University of the Witwatersrand
2050 Wits, South Africa        (at University of Cape Town 4 July-7 Nov)
phone [phone]  fax [phone]

+++++++++++++++++++++++++++

>From [e-mail] (Pete Resnick)
Date: Mon, 19 Sep 1994 08:35:23 -0500
Organization: University of Illinois at Urbana-Champaign

In article <[e-mail]>,
[e-mail] (Philip Machanick) wrote:

> In article <35jbt4$[e-mail]>, [e-mail] wrote:
>
> > Why? Because in order to produce a question mark I need to
> > use the shift key and since the command key disables the shift
> > key, this doesn't work.
>
> This is also true on the English and US keyboards: "?" is shift-/. Where
> did you get the idea that command disables shift? All the modifiers are
> independent and can be sensed independently. If COMMAND-? is not working
> for you there must be some other explanation.

You're wrong Phil. Command does in fact disable the shift modifier. See
the tech note on International Cancelling. To properly handle the command
key in your application, you should always strip the command-key bit out
of the modifiers field and re-pass the keystroke through KeyTrans. Here's
some source code for doing it with the cancel key, but you should use
similar code for any command-key combination (like in MenuKey):

/* These both use Universal Headers. The second is for System 6 too. */

Boolean WasCancel(EventRecord *theEvent)
{
    short keyCode;
    long state = 0L, test;
    unsigned long keys;

    if(!(theEvent->modifiers & cmdKey))
        return(false);
    keyCode = (theEvent->modifiers & 0xFE00) |
              ((theEvent->message & keyCodeMask) >> 8);
    keys = KeyTranslate((Ptr)GetScriptManagerVariable(smKCHRCache),
                        keyCode, &state);
    return(((Byte)keys == '.') || ((Byte)(keys >>= 16) == '.'));
}

Boolean WasCancelSystem6or7(EventRecord *theEvent)
{
    short keyCode, rsrcID, scriptID;
    Ptr kchrPtr;
    Handle kchrRsrc = nil;
    long state = 0L;
    unsigned long keys;

    if(!(theEvent->modifiers & cmdKey))
        return(false);
    keyCode = (theEvent->modifiers & 0xFE00) |
              ((theEvent->message & keyCodeMask) >> 8);
    if((kchrPtr = (Ptr)GetScriptManagerVariable(smKCHRCache)) == nil) {
        scriptID = GetScriptManagerVariable(smKeyScript);
        rsrcID = GetScriptVariable(scriptID, smScriptKeys);
        kchrRsrc = GetResource('KCHR', rsrcID);
        if(kchrRsrc != nil)
            kchrPtr = *kchrRsrc;
    }
    if(kchrPtr != nil)
        keys = KeyTranslate(kchrPtr, keyCode, &state);
    if(kchrRsrc != nil)
        ReleaseResource(kchrRsrc);
    return(((Byte)keys == '.') || ((Byte)(keys >>= 16) == '.'));
}
--
Pete Resnick    (...so what is a mojo, and why would one be rising?)
Doctoral Student - Philosophy Department, Gregory Hall, UIUC
System manager - Cognitive Science Group, Beckman Institute, UIUC
Internet: [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Anders Lundqvist)
Date: 19 Sep 1994 13:31:07 GMT
Organization: Dept. of E.E., Univ. of Linkoping

In article [e-mail], [e-mail] (Philip Machanick) writes:
>In article <35jbt4$[e-mail]>, [e-mail] wrote:
>
>> Why? Because in order to produce a question mark I need to
>> use the shift key and since the command key disables the shift
>> key, this doesn't work.
>
>This is also true on the English and US keyboards: "?" is shift-/. Where
>did you get the idea that command disables shift? All the modifiers are
>independent and can be sensed independently. If COMMAND-? is not working
>for you there must be some other explanation.
>--
>Philip Machanick                   [e-mail]
>Department of Computer Science, University of the Witwatersrand
>2050 Wits, South Africa        (at University of Cape Town 4 July-7 Nov)
>phone [phone]  fax [phone]

I'll try again...

On my keybord the "?" is located next to the "0" (zero) key and is
<shift>+"+". If I use the DA Keyboard and press the shift key then all
letters turn into capital letters and the "+" sign turns into a question
mark. OK! If I now keep the shift key pressed and press the <command> key
all letter go back to normal and the "?" is transmogrified into a "+" again.

Does this happen to you? If it does and you can use <command>+"?" then I'm
completely lost...

Of course, everything works as a charm if I turn on the US keyboard layout.


Anders Lundqvist
Department of EE
Linkoping University
Sweden



+++++++++++++++++++++++++++

>From [e-mail] (Peter Castine)
Date: Mon, 19 Sep 1994 15:19:45 GMT
Organization: Process Control Center

In article <[e-mail]>,
[e-mail] (Philip Machanick) wrote:

> In article <35jbt4$[e-mail]>, [e-mail] wrote:
>
> > Why? Because in order to produce a question mark I need to
> > use the shift key and since the command key disables the shift
> > key, this doesn't work.
>
> This is also true on the English and US keyboards: "?" is shift-/. Where
> did you get the idea that command disables shift? All the modifiers are
> independent and can be sensed independently. If COMMAND-? is not working
> for you there must be some other explanation.

Anders is right--if the command-shift-anything is pressed, the ASCII code
of the unshifted character is returned in event record. If the programmer
wants to get the shifted character, she needs to reprocess the key code
with the cmdKey bit in the modifiers field reset. This was all discussed
years ago in a Tech Note called ``International Cancelling'', it gave code
for correctly detecting command-period on all keyboards.

Don't ask me where the code is in NIM, I don't know. You can also find the
technique in use in the TCL source.

The trick with Mathematica is that, although it tells you to use Cmd-?, it
really is looking for Cmd-/. There is a possibility that, upon finding '/'
in the event record, it checks if both the shift and command keys are
pressed, but that still won't help you if '?' is someplace different from
it's position on the American Sholes (QWERTY) keyboard layout.

Someday I'm gonna write an application that uses Command-‹ (that's
U-Umlaut if you're not reading on a Mac), just to let some people find out
what it feels like.

BTW, there is one work-around for Anders problem: switch to the U.S.
keyboard layout, at least when you want to get that nifty ?-cursor. This
is a pain in the neck, but there are a couple of short cuts you can
take--mail me if you want details.

--
-Peter [e-mail]--Process Control Center--TU Berlin-
*  A New York City judge ruled that if two women behind you at the movies   *
*  insist on discussing the probable outcome of the film, you have the      *
*  right to turn around and blow a Bronx cheer at them.                     *

+++++++++++++++++++++++++++

>From [e-mail] (Philip Machanick)
Date: 19 Sep 1994 15:01:50 GMT
Organization: Computer Science Dept, U of Witwatersrand

In article <35k3qr$[e-mail]>, [e-mail] wrote:

> On my keybord the "?" is located next to the "0" (zero) key and is
> <shift>+"+". If I use the DA Keyboard and press the shift key then all
> letters turn into capital letters and the "+" sign turns into a question
> mark. OK! If I now keep the shift key pressed and press the <command> key
> all letter go back to normal and the "?" is transmogrified into a "+" again.
>
> Does this happen to you? If it does and you can use <command>+"?" then I'm
> completely lost...

If I use the DA, the same thing happens yet I can use command-? in MS Word
etc. Maybe Pete Resnick's article throws some light on this but it's not
instantly clear why this should be different for a different keyboard
layout (on the US keyboard on my PB170 "?" is SHIFT-/).
--
Philip Machanick                   [e-mail]
Department of Computer Science, University of the Witwatersrand
2050 Wits, South Africa        (at University of Cape Town 4 July-7 Nov)
phone [phone]  fax [phone]

+++++++++++++++++++++++++++

>From [e-mail] (Stefan Kurth)
Date: 20 Sep 1994 01:36:04 +0200
Organization: none

In article <35jbt4$[e-mail]>, [e-mail] (Anders
Lundqvist) wrote:

> I wonder how many of you that implement keyboard shortcuts
> like <command>+? realize that these DO NOT work on many
> international keyboard layouts like the swedish????
>
> Why? Because in order to produce a question mark I need to
> use the shift key and since the command key disables the shift
> key, this doesn't work.
>
> IS THIS PROBLEM MENTIONED IN THE HUMAN INTERFACE GUIDELINES???

It is mentioned in NIM: Text, page C-23. And if all programs were
well-behaved and handled these things like Apple tells us, you would
have no problems.

The solution is to strip the cmdKey bit from theEvent.modifiers, and
then run it through KeyTrans() again before passing it to MenuKey().

(BTW, I just tried it in Disinfectant, and it works fine for me; and I
 have a German keyboard which I guess is similar to yours.)

Actually, some programs don't check for command-? at all, but instead
they check whether the shift key is down, and if it is, they check for
command-/ (because shift-/ yields ? on a US keyboard). Thus, you might
try command-shift-/ on your keyboard and see if it works.

Cheers,
  -Stefan

________________________________________________________________________
Stefan Kurth                 Berlin, Germany              [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Quinn "The Eskimo!")
Date: Tue, 20 Sep 1994 11:30:09 +0800
Organization: Department of Computer Science, The University of Western Australia

In article <35jbt4$[e-mail]>, [e-mail] wrote:

>Greetings all programmers!

Anders, you think *you've* got problems, us Dvorak keyboards users have to
put up with certain applications (like Photoshop 2.x and BBEdit 3.0 and
certain bits of HC 2.x) where it ignores the command-key table in the KCHR
and decides to use Dvorak command keys instead of QWERTY ones.  Do you
have any idea how annoying it is to hit command-X (to do a Cut) and have
the application quit.  *grrr*
--
Quinn "The Eskimo!"        "Kids, 3D and driving just don't mix."
  So many gumbies, so few ICBMs.
  Still using BBEdit 2.3.2.

+++++++++++++++++++++++++++

>From [e-mail] (Jon W{tte)
Date: Tue, 20 Sep 1994 10:36:02 +0200
Organization: Royal Institute of Something or other

In article <35jbt4$[e-mail]>,
[e-mail] (Anders Lundqvist) wrote:

>IS THIS PROBLEM MENTIONED IN THE HUMAN INTERFACE GUIDELINES???

Yes, it is. Applications are suggested only to use letters, and
where they use non-letters, to do the keyboard translation
themselves using KeyTrans with the command key stripped off
or something to that effect.

The easy thing for a user to do is to extract the Swedish
keyboard layout in ResEdit and change it to care about shift
with command down (that's just a simple re-binding)

However, you immediately get a new problem; cmd-shift-digit
means FKEY, and the international shifted numbers aren't the
same as the US ones (on my keyboard, from 1 through 0, it's:
bang quote hash sun percent ampersand slash leftparen
rightparen equal, or as you may, potion amulet sink unknown
food demon wawwnd tool armor ring :-)

Now, FoxPro for Macintosh uses cmd-F1 cmd-F2 etc for commands
that aren't edit commands, so it could be worse (I don't have
command keys on my keyboard) But then, FPfM says it supports
the full power of AppleEvents and AppleScripts, when the only
non-required AE they support is the DoScript one. Lame, and
misleading.

Cheers,

                    / h+


--
  Jon W‰tte ([e-mail]), Hagagatan 1, 113 48 Stockholm, Sweden
 V}ga v{gra nonkonformism!


+++++++++++++++++++++++++++

>From [e-mail] (Peter Castine)
Date: Tue, 20 Sep 1994 10:00:17 GMT
Organization: Process Control Center

In article <35l794$[e-mail]>, [e-mail] (Stefan
Kurth) wrote:

> Actually, some programs don't check for command-? at all, but instead
> they check whether the shift key is down, and if it is, they check for
> command-/ (because shift-/ yields ? on a US keyboard). Thus, you might
> try command-shift-/ on your keyboard and see if it works.

The problem is that this won't work if the slash (/) itself is a shift-key
combination as it is, for instance, on the German keyboard (/ == shift-7).

So, the only solution for Anders' problem (short of switching to the U.S.
keyboard layout) seems to be to use QuicKeys. Of course, if the command
isn't available in a menu somewhere, he won't have luck with that, either.

I sense a moral here: Don't implement command-key shortcuts that *aren't*
available through menus. Hey, if the cmd-key equivalents were for menu
commands, Anders could just tweak the appropriate resources. (Won't work
for MS products, but, hey, nothing works for MS products ;-} ).

Cheers

--
-Peter [e-mail]--Process Control Center--TU Berlin-
*  A New York City judge ruled that if two women behind you at the movies   *
*  insist on discussing the probable outcome of the film, you have the      *
*  right to turn around and blow a Bronx cheer at them.                     *

+++++++++++++++++++++++++++

>From [e-mail] (Keith Rollin)
Date: Wed, 21 Sep 1994 04:06:34 -0800
Organization: Apple ][ -> Mac -> Taligent -> Newton -> Windows?

In article <35jbt4$[e-mail]>, [e-mail] wrote:

>I wonder how many of you that implement keyboard shortcuts
>like <command>+? realize that these DO NOT work on many
>international keyboard layouts like the swedish????

FYI, the '?' key is shifted on US layouts, too.


>
>Why? Because in order to produce a question mark I need to
>use the shift key and since the command key disables the shift
>key, this doesn't work. I have been playing around with the DA
>Keyboard and there's simply no way around. Obiwan uses <command>+`
>and this I can emulate by pressing <command>+<option>+<control>+'.
>Sigh...
>
>Of course, this wouldn't been so bad if there was an equivalent
>meny command, but in many cases there isn't. Usually <command>+?
>invokes some help mode and turns the pointer into a question mark
>to allow the user to point at an object (e.g Disinfectant and
>Mathematica).
>
>IS THIS PROBLEM MENTIONED IN THE HUMAN INTERFACE GUIDELINES???

There's a technote called "International Canceling" that deals with this
issue. It focuses on the fact that '.' is sometimes a shifted character
(the technote cites the Italian keyboard as an example), and shows how to
get around the fact that Shift is ignored when Command is pressed. If
programmers care to get it right, the information is available.

- --------------------------------------------------------------------------
Keith Rollin --- Phantom Programmer --- Apple Computer, Inc. --- Team Newton

---------------------------

>From [e-mail] (Erik Anton Westwig)
Subject: Apple Events Question
Date: Wed, 21 Sep 1994 10:11:54 -0500
Organization: Cornell University

I'm trying to finally learn about Apple Events, but I'm having some
problems
understanding something quite fundamental:

Consider the first required AE: Open Application
  When the Finder starts my App, it calls the Process Manager to launch it
  and then sticks a kAEOpenApplication high-level event into my queue. I
  think I got that part, but because I want my app to be as friendly as
  possible to older systems (without AEs) I wouldn't want to stick ANY code
  in the OpenApp routine that is necessary to the starting of my program.
  I can't see what I would do differently in the OpenApp routine than I
  would do in the non AE environment --> so I guess what I'm saying is I
don't
  understand why this is an important event to support.
  Am I completely missing the whole point here, or what?

I think I see how the QuitApp event is different: it gives my App a chance
to
shut down itself gracefully (e.g. asking the user if they want to save open
documents, etc.)

Thanks,
ERIK

+++++++++++++++++++++++++++

>From [e-mail] (Jeremy Roschele)
Date: Wed, 21 Sep 1994 08:09:28 -0800
Organization: SimCalc Project

In article <westwig-210994101154@132.236.102.71>, [e-mail]
(Erik Anton Westwig) wrote:

>   I can't see what I would do differently in the OpenApp routine than I
>   would do in the non AE environment --> so I guess what I'm saying is I
> don't
>   understand why this is an important event to support.
>   Am I completely missing the whole point here, or what?
>

On the Open Application event you should create a blank document. The
Finder won't send OApp if it is printing, or the user launched with a
document.

jeremy

+++++++++++++++++++++++++++

>From [e-mail] (John W. Baxter)
Date: Wed, 21 Sep 1994 13:26:08 -0700
Organization: Internet for the Olympic Peninsula

In article <westwig-210994101154@132.236.102.71>, [e-mail]
(Erik Anton Westwig) wrote:

> I'm trying to finally learn about Apple Events, but I'm having some
> problems
> understanding something quite fundamental:
>
> Consider the first required AE: Open Application
>   When the Finder starts my App, it calls the Process Manager to launch it
>   and then sticks a kAEOpenApplication high-level event into my queue. I
>   think I got that part, but because I want my app to be as friendly as
>   possible to older systems (without AEs) I wouldn't want to stick ANY code
>   in the OpenApp routine that is necessary to the starting of my program.
>   I can't see what I would do differently in the OpenApp routine than I
>   would do in the non AE environment --> so I guess what I'm saying is I
> don't
>   understand why this is an important event to support.
>   Am I completely missing the whole point here, or what?

When your application starts, and realizes that Apple events are active
(all my apps realize that, because I don't write for System 6 any more),
it should do what needs to be done every time it starts up.  None of this
will be visible to a user.

When an Open Application event comes in, the app should do what it does
when it is double-clicked on in the Finder.  Often, that is to create and
open a new untitled document...sometimes it isn't.  At one time, Apple
suggested not doing that if an Open Application appears after the app is
happily doing things.  Recently, it seems that Apple may in the future
want the open a new untitled document behavior in that case, as an
"improvement" to the Mac look and feel.  So you probably should just go
ahead and handle Open Application whenever it shows up.

When an Open Document comes in before the Open Application, it means that
one or more docs were double-clicked rather than your application.
Normally, you would open them (assuming that's what your application
does).  Print Document may appear rather than Open Document...this becomes
more important with QuickDraw GX than it was with most users before, since
it's how you find out that a user has dragged one of your docs to a
desktop printer.

Another possibility is that an 'ascr'/'noop'  (I think 'noop' is right)
event appears.  Something else will happen later.  Most apps can simple
ignore this event, which says "I'll tell you later why you were started".
But some apps may need to know.

   --John

--
John Baxter    Port Ludlow, WA, USA  [West shore, Puget Sound]
   "Occasionally...astronomers add a second to either June 31 or
    December 31..."   IM: OS Utilities, p 4-12
   [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Erik Anton Westwig)
Date: Wed, 21 Sep 1994 18:35:46 -0500
Organization: Cornell University

In article <[e-mail]>,
[e-mail] (Jeremy Roschele) wrote:

> On the Open Application event you should create a blank document. The
> Finder won't send OApp if it is printing, or the user launched with a
> document.
>
> jeremy

yup, i get that part.  But what I don't see is what I need to do
differently
than before (without the AEs).  To be pre system 7 friendly, I need to be
able
to open my blank document without using the KAEOpenApplication event,
anyway.

ERIK

+++++++++++++++++++++++++++

>From [e-mail] (John W. Baxter)
Date: Wed, 21 Sep 1994 19:42:44 -0700
Organization: Internet for the Olympic Peninsula

In article <westwig-210994183547@132.236.102.67>, [e-mail]
(Erik Anton Westwig) wrote:

> In article <[e-mail]>,
> [e-mail] (Jeremy Roschele) wrote:
>
> > On the Open Application event you should create a blank document. The
> > Finder won't send OApp if it is printing, or the user launched with a
> > document.
> >
> > jeremy
>
> yup, i get that part.  But what I don't see is what I need to do
> differently
> than before (without the AEs).  To be pre system 7 friendly, I need to be
> able
> to open my blank document without using the KAEOpenApplication event,
> anyway.

The sample code in Inside Mac: IAC shows how to detect whether you should
wait for the 'oapp' because it's coming, or not.  I think the discussion
in Think Reference also does.  --John

--
John Baxter    Port Ludlow, WA, USA  [West shore, Puget Sound]
   "Occasionally...astronomers add a second to either June 31 or
    December 31..."   IM: OS Utilities, p 4-12
   [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Jon W{tte)
Date: Thu, 22 Sep 1994 16:11:18 +0200
Organization: Royal Institute of Something or other

In article <westwig-210994183547@132.236.102.67>,
[e-mail] (Erik Anton Westwig) wrote:

>yup, i get that part.  But what I don't see is what I need to do
>differently
>than before (without the AEs).  To be pre system 7 friendly, I need to be
>able
>to open my blank document without using the KAEOpenApplication event,
>anyway.

Well, the logic goes like this:

On startup:

if (hasn't got AppleEvents) then
    if (CountAppFiles) then
        Open or Print files
    else
        Make new document
else
    Install AppleEvent Handlers


on OpenApplication:
    Make new document

on OpenDocument:
    Open files

on PrintDocument:
    Print files


As you see, you should separate the actions (opening/printing
documents or making new documents) from their triggers
(starting up and/or choosing the menu items)

For instance, you wouldn't want your app to create a blank
document before receiving an open documents appleevent when the
user double-clicks a document in the Finder under system 7.

Cheers,

                / h+


--
  Jon W‰tte ([e-mail]), Hagagatan 1, 113 48 Stockholm, Sweden

"TextEdit does everything right."
    ã Jon W{tte


---------------------------

>From [e-mail] (Jump Long)
Subject: File System Manager Software Development Kit
Date: 15 Sep 1994 00:28:01 -0400
Organization: America Online, Inc. (1-800-827-6364)

Apple Developers:

The File System Manager Software Development Kit (SDK) version 1.2 is
available on Apple's ftp server at:

ftp.apple.com (IP 130.43.2.3)

/ftp/dts/mac/sys.soft/extensions/file-system-manager-sdk.hqx

The FSM SDK will be on Developer CDs and on AppleLink in the future.

Archive Contents
- --------------

* File System Manager -- A system extension which implements version 1.2
of FSM and the extended Disk Initialization Package. This extension is
compatible with System 7.0 and later. If in future system software, we
make changes to FSM, the File System Manager extension will disable
itself.

* FSMGlueLib.o -- The glue code for the File System Manager service
routines.

* FSM.h. FSM.p and FSM.a -- The Universal Interface files for the File
System Manager. This version of the interface files was built to work with
the Universal Interfaces 2.0a1 from ETO #15, MPW prerelease. However, for
68K development (which we expect all foreign file systems to be), FSM.h
works perfectly with the current release of Universal Interfaces on ETO
#15.

* Guide - File System Manager -- The documentation for the File System
Manager in Apple DocViewer format. The latest version of the Apple
DocViewer application can be found on any recent Developer CD.

Licensing Information
- -------------------

The File System Manager extension version 1.2 may be licensed for
distribution. Contact Apple Software Licensing at:

     Software Licensing
     Apple Computer, Inc.
     2420 Ridgepoint Drive, MS: 198-SWL
     Austin, TX  78754
     512 919-2645
     AppleLink: SW.LICENSE

The File System Manager Development and Developer Support Team

(I told ya we'd make it available when it was done - Jim Luther)

+++++++++++++++++++++++++++

>From [e-mail] (Quinn "The Eskimo!")
Date: Fri, 16 Sep 1994 10:57:22 +0800
Organization: Department of Computer Science, The University of Western Australia

In article <358igh$[e-mail]>, [e-mail] (Jump
Long) wrote:

>(I told ya we'd make it available when it was done - Jim Luther)

(:  Yay Jim!
--
Quinn "The Eskimo!"        "Scout in a can. Simple, cheap, easy
                            to use and it's expendable!"

+++++++++++++++++++++++++++

>From [e-mail] (Jump Long)
Date: 17 Sep 1994 14:57:02 -0400
Organization: America Online, Inc. (1-800-827-6364)

In article <[e-mail]>, [e-mail]
(Robert Mah) writes:

>Please pardon my ignorance, but I have one question.  What does it do?

Here's some meterial cut from the FSM docs which should answer this
question:

"The File System Manager is the part of the Macintosh Operating System
that manages the use of foreign file systems. The File System Manager
provides a general means by which foreign file systems can be installed,
identified, and interfaced to the Operating System.

The Operating System services provided by the File System Manager are
called FSM components and the interface mechanism between an foreign file
system and a particular FSM component is known as a FSM component
interface. At this time, two FSM component interface are defined for use
under the File System Manager

* the HFS Component Interface which allows a foreign file system to
process File Manager requests

* the Disk Initialization Package Component Interface which allows foreign
file systems to initialize foreign file system volumes on a Macintosh"


FSM also includes the File System Utility Routines which help you write a
foreign file system. Here's the description of those utilities:

"The File System Utility routines are intended for use by foreign file
systems to allow your foreign file system to

* access file control blocks (FCBs), volume control blocks (VCBs), working
directory control blocks (WDCBs), and drive queue elements (DrvQEl)

* set and get the default volume and working directory

* eject a volume

* validate and process parameters passed with Macintosh file system calls

* access the Macintosh file system's cache buffers and low-level I/O
services"


Writing a foreign file system is by no means simple, but we've tried to
make it easier to do a lot  of it.

- Jim Luther


+++++++++++++++++++++++++++

>From [e-mail] (Quinn "The Eskimo!")
Date: Sun, 18 Sep 1994 22:37:05 +0800
Organization: Department of Computer Science, The University of Western Australia

In article <35fe5u$[e-mail]>, [e-mail] (Jump
Long) wrote:

>Writing a foreign file system is by no means simple, but we've tried to
>make it easier to do a lot  of it.

So I spent the weekend reading the docs and now I have a question...

How do you write an asychronous file system based on a non-block device?
(eg AppleShare)  Presumably you should be using the cache routines because
of this comment...

"all calls to most device drivers should be made through the cache routines."

(:

However the cache routines (eg UTCacheReadIP) seem to imply that they
access the block device driver directly, because they call the
log2PhyProc.  Any ideas?
--
Quinn "The Eskimo!"        "Scout in a can. Simple, cheap, easy
                            to use and it's expendable!"

+++++++++++++++++++++++++++

>From [e-mail] (Jump Long)
Date: 18 Sep 1994 15:43:06 -0400
Organization: America Online, Inc. (1-800-827-6364)

In article <[e-mail]>,
[e-mail] (Quinn "The Eskimo!") writes:

>How do you write an asychronous file system based on a non-block device?
>(eg AppleShare)  Presumably you should be using the cache routines
because
>of this comment...
>
>"all calls to most device drivers should be made through the cache
routines."

Here's the answer I gave to a couple of developers during the seed to get
them around this:

- ---

So, how do you use the cache routines (which are designed to work with
block disk devices) with non-block devices?  You write a very simple block
disk driver - one that supports one writable block.  You'll need a driver
anyway to handle _Control calls with csCode 21 (return physical drive
icon) because the Finder uses that control call.

The single block driver is going to be your communications channel between
your FFS, which can only perform block I/O through the cache, and your
non-block driver.  As an example of how this trick works, let's say your
FFS needs to send a command to your WizzyNet driver.

First, your FFS needs to get a cache block and associate it with the block
on your simple block driver.  It does this by calling UTGetBlock passing
the volume's volume reference number for the refNum (so the cache code
knows which driver to call), NULL for the log2PhyProc, 0 for the block
number (the only block your driver has), and gbNoRead for the gbOption
(which tells the cache code not to bother reading the block from disk).
You'll get a pointer back to a 512-byte buffer in the cache.

Next, your FFS puts whatever data is wants to send to the block driver (to
send on to the WizzyNet driver) into the cache buffer.  Then, to tell the
cache that the block needs to be sent to the block driver, the FFS calls
UTMarkDirty with the pointer to the cache block. To tell the cache code to
write the block, the FFS calls UTFlushCache.  When the cache code is
called to flush the dirty block to the disk driver, it will call the block
driver to write the cache block.

The block driver can then call the WizzyNet driver asynchronously with a
completion routine and then RTS back to the Device Manager.  The Device
Manager will then either return control to the user code or will spin in
the SyncWait loop if the call was made synchronously.

When the WizzyNet driver completes the call and the Device Manager calls
the completion routine in the block driver, the block driver will jump to
jIODone, the Device Manager will call the completion routine in the File
Manager's cache code, and the cache code will pass control back through
FSM to the FFS that made the cache call.

Finally, when you're done making calls to your driver for this particular
File Manager request, you call UTReleaseBlock to release the block.

So, all your block driver needs to handle is _Open, _Close, _Control, and
_Write.  Since it'll never be getting _Read or _Status calls from your
FFS, you can simply return readErr to _Read and statusErr to _Status
requests.  The data your driver is asked to write is going to be the
command or data you want to send to your non-block driver.  You might want
to put a signature long-word in the cache blocks your write so that other
code that mistakenly writes to your driver can be ignored if the signature
isn't there.

- ---

Yes, that's a hack, but it works and it keeps your foreign file system
from deadlocking the system.  There are a few published foreign file
systems (I won't mention any names) that make synchronous calls to network
drivers which they don't completely own.  Every once in a while, those
file systems deadlock the system because: a) something interrupts while
the driver they call is busy, and b) that code makes an asynchronous File
Manager request and c) if the File Manager isn't busy, it passes the
request on to the foreign file system which d) makes a synchronous call to
the busy driver which causes the system to spin in SyncWait forever (or
until you reset the system or pull the plug).

- Jim Luther

+++++++++++++++++++++++++++

>From [e-mail] (Quinn "The Eskimo!")
Date: Mon, 19 Sep 1994 09:42:25 +0800
Organization: Department of Computer Science, The University of Western Australia

In article <35i58a$[e-mail]>, [e-mail] (Jump
Long) wrote:

>In article <[e-mail]>,
>[e-mail] (Quinn "The Eskimo!") writes:
>
>>How do you write an asychronous file system based on a non-block device?
>
>Yes, that's a hack, but it works [...]

Cool.  I figured that the hack would work something like that but it's
nice to know that I'm not missing something obvious.

Share and Enjoy.
--
Quinn "The Eskimo!"        "Scout in a can. Simple, cheap, easy
                            to use and it's expendable!"

+++++++++++++++++++++++++++

>From [e-mail] (Peter N Lewis)
Date: Tue, 20 Sep 1994 10:13:20 +0800
Organization: NCRPDA, Curtin University

>Cool.  I figured that the hack would work something like that but it's
>nice to know that I'm not missing something obvious.

It should be possible to write a generic driver to take the place of that
block driver, something where the block is formatted like:

<signature>
<address to call>
<generic parameter>

And have the block driver call the specified address, with a ptr to the
parameters.  Once we (someone) writes this and makes it available, that
should be one less chunk we all have to write.  That should work, right?
   Peter.
--
Peter N Lewis <[e-mail]> - Macintosh TCP fingerpainter
FTP my programs from redback.cs.uwa.edu.au:Others/PeterLewis/ or
amug.org:pub/peterlewis/ or nic.switch.ch:software/mac/peterlewis/

+++++++++++++++++++++++++++

>From [e-mail] (Jump Long)
Date: 20 Sep 1994 03:08:06 -0400
Organization: America Online, Inc. (1-800-827-6364)

In article <[e-mail]>,
[e-mail] (Peter N Lewis) writes:

>It should be possible to write a generic driver to take the place of
>that block driver, something where the block is formatted like:
>
><signature>
><address to call>
><generic parameter>
>
>And have the block driver call the specified address, with a ptr to the
>parameters.  Once we (someone) writes this and makes it available, that
>should be one less chunk we all have to write.  That should work, right?

That looks like you're planning on calling <address to call> synchronously
and that won't work. When the block driver gets a request (either
asynchronously or synchronously - this method works for both), the block
driver should:

1) Call the network driver asynchronously with a completion routine
2) RTS back to the Device Manager
3) When the completion routine is executed, jump to ioDone to finish up
the call.

The reason it should do it that way is that the network driver could be
busy. If that's the case, then the request will be queued for execution
later. The Device Manager will return control to the file system's cache
code (which called the block driver) and the file system will do one of
two things: 1) if the original File Manager request was synchronous, it
will spin in SyncWait until the device request's ioResult field is less
than or equal to noErr or 2) if the original File Manager request was
asynchronous, it will return control to the caller.  Either way, when the
network driver completes its request and the Device Manager calls the
block driver's completion routine and the block driver calls the File
Manager's completion routine (or the SyncWait loop code sees ioResult
change), then control will be returned to the Foreign File System that
used the cache.

So, if you're going to write some kind of generic block driver, then you
should only need to pass it a block with a structure like this:

enum
{
    kReadRequest = 1,
    kWriteRequest = 2,
    kControlRequest = 3,
    kStatusRequest = 4
};

struct CallDriverStruct
{
    OSType: mySignature; /* the creator type of my foreign file system */
                         /* to help stop unknown writers */
    short: callToMake;   /* pass kReadRequest, kWriteRequest, etc. to */
                         /* tell block driver which driver call to make.
*/
    ParamBlockRec: pb;   /* parameter block to make driver call with. */
};

All of that fits easily into a 512-byte cache block and you don't have to
worry about the parameter block moving or being reused since cache blocks
you get with UTGetBlock are yours until you release them (and they don't
move in memory - ever).

All of this is off the top of my somewhat sleepy head, so I hope it makes
sense. If you get something working and want a code review, let me know.
If I have time, I'll write it...

- Jim Luther


+++++++++++++++++++++++++++

>From [e-mail] (Benjamin Herrenschmidt)
Date: Wed, 21 Sep 94 23:00:53 +0100
Organization: (none)

Hi !

I was away from the usenet for some days (AppleExpo in Paris) and
i missed the beginning of this thread. The API to the File System
Manager is available somewhere ? I am interested....

BenH.

---------------------------

>From [e-mail]
Subject: GDGX Graphics Shell bus errors with GX Beta 3...
Date: 19 Sep 94 09:17:28 -0600
Organization: Fermi National Accelerator Lab

Hi all,
  I'm trying to teach myself some QDGX graphics this week. :)
  Short story is I built GDGX Graphics Shell with MW and I get a trashed
stack ("sc" in Macsbug says my stack pointer is odd!).  The precompiled version
on the develop CD bombs with a bus error as well.  Details next:
  I thought I would start by building the QDGX Graphics shell from Develop 15.
I pulled the source code from develop 19, and I'm using beta 3 of GX (I don't
system 7.5 yet, just 7.1).  The app on the disk crashed my mac with a bus error.
So, I decided to recompile everything.
  I'm using MW D/4.  First thing I have to do is #define ppcinterfaces in all
the source files (including the library files!).  Second, the shell sources
won't compile as is.  There is a routine GXQDGlobalToFixedLocal, for example,
that doesn't exist in any of the gx .h files.  I replaced it with
ConvertGXPoint, and used the parent view port of gcontentWindow to get the
local coords correct.  When creating the "runFeatures" for the styled layout
shape, the second feather (alternateblahblah) doesn't exist in gx anylonger.
Is that right?  For the time being, I've just removed it.
  So, what did I miss.  Do I have to do a #pramga parameters_in_a0 or something
like that around the gx interfaces?  Why isn't this done automagically, if I
do have to do it.
  I only just started last night, so I've not done a detailed step through
the program yet.  I just figured that others starting out with GX may have
seen the same thing, and save me some time.  Thought changing the to the
ConvertQDPoint taught me a bit about viewport coord systems!
  Any help would be great!  TIA.

    Cheers,
        Gordon.

+++++++++++++++++++++++++++

>From [e-mail] (Dave Johnson)
Date: Wed, 21 Sep 1994 17:56:24 GMT
Organization: Apple Computer

In article <[e-mail]>, [e-mail] wrote:

>   I thought I would start by building the QDGX Graphics shell from Develop 15.
> I pulled the source code from develop 19, and I'm using beta 3 of GX (I don't
> system 7.5 yet, just 7.1).  The app on the disk crashed my mac with a
bus error.

The QDGX Shell on the CDs was compiled for, and runs under, beta 1 of GX
(and without universal interfaces). It was never updated for beta 3,
unfortunately. The quick and (very) dirty way to get it up and running
would be to install beta 1 (it's on the issue 15 CD) and use old headers.
Not recommended.

> So, I decided to recompile everything.
>   I'm using MW D/4.  First thing I have to do is #define ppcinterfaces in all
> the source files (including the library files!).

It actually might be better to #define it in either a "prefix" file (if
you're not using precompiled headers) or in a .h file that you include
before any other headers.

> Second, the shell sources
> won't compile as is.  There is a routine GXQDGlobalToFixedLocal, for example,
> that doesn't exist in any of the gx .h files.  I replaced it with
> ConvertGXPoint, and used the parent view port of gcontentWindow to get the
> local coords correct.

Excellent exercise! :-) That routine went away between beta 1 and beta 3.

>   So, what did I miss.  Do I have to do a #pramga parameters_in_a0 or
something
> like that around the gx interfaces?

Here's the recommended "prefix" file for MW, most of which came from the GX SDK:

///

#define ppcinterfaces

#ifndef powerc
   // The following two typedef's are to work around a bug in the ETO #15
   // Universal headers -- the headers only typedef float_t and double_t
   // for "applec" and "powerc".  When this is eventually fixed, these
   // lines will generate "redefined type" errors.

   typedef long double float_t;
   typedef long double double_t;

   // The following is required if using Metrowerks 68K (CW4).
   // Otherwise, MetroWerks 68K expects (and uses what it thinks are)
   // A0 result values instead of D0 ones.

   #pragma pointers_in_D0     // Required for c-style toolbox glue.
#endif

///

On the November CDs (Bookmark 20 and Reference Library), the beta 1 GX
shell will probably be replaced with the "modern" version that comes with
the SDK (I still have to get buyoff for that, but I don't anticipate any
problems). It uses the release version of GX, and the code has been
cleaned up in a big way. Frankly I wouldn't recommend the old one any
more, except as an "exercise for the reader."

Have fun!

Dave Johnson
Technical Buckstopper
d e v e l o p

+++++++++++++++++++++++++++

>From [e-mail] (Paul Lalonde)
Date: Thu, 22 Sep 1994 11:24:47 GMT
Organization: Metrowerks

In article <[e-mail]>, [e-mail] wrote:

> Hi all,
>   I'm trying to teach myself some QDGX graphics this week. :)
>   Short story is I built GDGX Graphics Shell with MW and I get a trashed
> stack ("sc" in Macsbug says my stack pointer is odd!).  The precompiled
version
> on the develop CD bombs with a bus error as well.  Details next:
>   I thought I would start by building the QDGX Graphics shell from Develop 15.
> I pulled the source code from develop 19, and I'm using beta 3 of GX (I don't
> system 7.5 yet, just 7.1).  The app on the disk crashed my mac with a
bus error.
> So, I decided to recompile everything.
>   I'm using MW D/4.  First thing I have to do is #define ppcinterfaces in all
> the source files (including the library files!).  Second, the shell sources
> won't compile as is.  There is a routine GXQDGlobalToFixedLocal, for example,
> that doesn't exist in any of the gx .h files.  I replaced it with
> ConvertGXPoint, and used the parent view port of gcontentWindow to get the
> local coords correct.  When creating the "runFeatures" for the styled layout
> shape, the second feather (alternateblahblah) doesn't exist in gx anylonger.
> Is that right?  For the time being, I've just removed it.
>   So, what did I miss.  Do I have to do a #pramga parameters_in_a0 or
something
> like that around the gx interfaces?  Why isn't this done automagically, if I
> do have to do it.
>   I only just started last night, so I've not done a detailed step through
> the program yet.  I just figured that others starting out with GX may have
> seen the same thing, and save me some time.  Thought changing the to the
> ConvertQDPoint taught me a bit about viewport coord systems!
>   Any help would be great!  TIA.
>
>         Cheers,
>                 Gordon.

I've been through this problem, and what I did was make myself a precompiled
header whose source file included all of the GX headers.  Everything was then
wrapped between #pragma parameters_in_d0 and #pragma parameters_in_a0.  Then
things started working :)

You see, contrary to almost the whole Toolbox, the GX graphics routines use
the C calling convention.  Problem is, those conventions vary from one compiler
to the next.  Hence the need for the #pragma.

Also, if you're looking for those missing " alternateblahblah" declarations,
there's a demo on the develop 17 CD (*not* called "QDGX Graphics shell",
though) that puts up the same graphic as the shell on the develop 15 CD,
so you can just cut and paste.

Note:  The headers changed between B1 and B3.  Be sure you're using the right
headers for the version you're running.


Paul Lalonde
[e-mail]

---------------------------

>From [e-mail] (Robert)
Subject: Getting Started (Summary)
Date: 17 Sep 1994 23:45:40 GMT
Organization: kind of sloppy actually....

Thanks to everyone who sent responses to me.  Here's a compilation of some of
what I got when I asked about getting started with Macintosh programming for
around $200.  Hopefully this will prove helpful to others out there who are
toying with the idea of attempting to learn Mac programming.


>From Gary M. Greenberg <[e-mail]>

Part of what you do should depend on how much _other_ programming
background you have. Many newbies to the Mac are highly or fairly
experienced on other platforms. If that's you, then get the biggest,
baddest, _mostest_ package you can find (maybe even get SC++7.__ and CW
Gold).

On the other hand, if you're like me and you have no (zero) former
programming experience, _and_ you're teaching yourself rather than studying
CS at a university, then any ANSI C compiler and good books will be fine.

I've been teaching myself how to code for a few months but I've only had a
few actual weeks of time to spend working at it. Still, I've gotten through
7 chapters of the Waite Group's C Primer Plus, and I'm concurrently reading
K&R's The C Programming Language, 2nd Edition. I'm also chipping away at a
new book called Symantec C++ Programming for the Macintosh (with Diskette
for about $35.00). I bought SC++6.0 for $50.00 and got the free upgrades
from the SC FTP site. I'm sure that the OO Libraries are about 6-12 months
beyond my reach given the limited time I can invest in learning right now,
but I expect a geometric progression -- you should too.

I've spent more on books than any other resource. The books on CD are okay
if that is all money will permit, but you'll become hungry for the paper
soon, as too much screen time weakens one's concentration. I've used them
enough to know. And, I now subscribe to MacTech and Develop. I've written
some very small _useless_ programs/exercises (and had fun each minute
making them work) so my method is having a positive result. Books, source
code, following c.s.m.p. and a.s.m. (that's alt.sources.mac) and lots of
keystrokin' is what will bring it all together for you and the tool you
choose (the brand of programming environment) will be less important.
You'll see lots of passionate hooey about all the brands. Each has lovers
and haters (I personally can't wait til my skills become sharp enough that
I feel comfortable tackling Smalltalk) but they are only _tools_.

I firmly believe that after you are successful at learning a language,
you'll move around and experiment with many different types of tools. So,
$200 [hey, I made it back to your topic ;-) ] is a comfortable starting
point ...
yet, it's just that ... a starting point.


>From Tim Dixon  <[e-mail]>

If you're a student (or perhaps faculty, I don't know) CodeWarrior Gold is
available at the educational price of $99.  Call 1-800-377-5416 for details.

[Regarding having a friend who's a student order for you:]

That should work.  A friend and I placed our orders at the same time.  We
had to place separate orders, but they let us both use his credit card (I
don't have one so I paid him by check).  They'll also accept a check.


>From Peter S. Lau:

well, CodeWarrier is quiet good, and you can tell by the recent
created and active newsgroup comp.sys.mac.programmer.codewarrier.  I
think $99 to get the compiler (actually 3 compilers C, C++, and
Pascal), plus tools, plus class library is definitely a bargain.  I
have used the demo version in Thomson's book.  I feel quiet
comfortable with it.

A few third-party books are good, too.  "Macintosh Programming Secrets" 2nd
edition is good.  I found it a bit wordy, but the examples are practical, so
that's the good part.  You might want to check the library.


>From Robert B. Schmunk <[e-mail]>

The develop Bookmark CD no longer contains all the NIM books, but if you
request a copy of issue #17 (backissues cost about $10), I understand that
all that had been released before spring 1994 they're all on that one. To
get the newer ones, you'll need to subscribe, which is something like $37
for four issues. Supposedly, Addison-Wesley will be publishing a CD sometime
in the next couple months which contains nothing but a complete NIM (the
Bookmark CD contains lots more), and the street price will supposedly be
around $99.


>From Eric Drumbor <[e-mail]>

     I would check around for used books.  I purchased volumes 1-6, the
X-Reference, and a book on the Communications Toolbox (this was all from
the old series) for $30.  The CDs are nice to have, but it's much easier
to have a book to use as a reference.

     I don't know what you have as far as beginners programming books (if
you actually need them), but I would recommend either Learn C on the
Macintosh and The C Programming Primer from Dave Mark, or Think THINK C
by Dan Parks Sydow (Sydow has also written another "intermediate" book
that explains more about the Macintosh Toolbox commands).  Mark's books
are fairly good, but I think Sydow has a better understanding on how to
teach and what needs to be explained.

     What I've suggested goes over the $200 mark, but I'd recommend that
you get Code Warrior and some kind of C book before you get a Develop
subscription.  I've spent over $600 in books alone, and I'm still waiting
for my next paycheck to buy some more.  Hope this helps..good luck!



@@@@@ Robert Stone ([e-mail])
@@@@@ Some Sort of Computer-Related Person, UGA - Extension Dairy Sci.
@@@@@ I am not responsible for anything I might say.


+++++++++++++++++++++++++++

>From [e-mail] (Ken Long)
Date: Mon, 19 Sep 1994 04:04:17 GMT
Organization: NETCOM On-line Communication Services (408 261-4700 guest)

If someone wanted to program on the Mac and had minimal finances, and
could only afford a used Think C package, then what could they get to
learn more?

Kernighan on C - a big text (w.p.) document by Brian Kernigan on the C
language, SpInside Mac, and as much example source as could be found and
downloaded.  The books are all expensive and many only give you example
code and a book full of comments.  Some have useful info not found in
example code, but is it absolutely necessary for the price?

I've found few source examples in books (or on the extra disks you'd buy)
that did not have similar or identical example sources somewhere in
cyberspace.  And there are a LOT more examples in cyberspace that you'd
not see in any bought book.

I'm not against buying programming books - especially IM.  I've got
plenty of money tied up in them!  But if you don't have any money,
example source and SpInside Mac is your route.  As far as value in
programmer training, I'd say IM, example source and programmer message
forums are far superior than Mac programming books.  Think C 5's Standard
Libraries manual was a real close second to K&R's.

-Ken-

+++++++++++++++++++++++++++

>From Clay Thurmond <[e-mail]>
Date: 19 Sep 1994 04:45:00 GMT
Organization: TVMHS

In article <[e-mail]> Ken Long, [e-mail]
writes:
>Kernighan on C - a big text (w.p.) document by Brian Kernigan on the C
>language, SpInside Mac, and as much example source as could be found and
>downloaded.  The books are all expensive and many only give you example
>code and a book full of comments.  Some have useful info not found in
>example code, but is it absolutely necessary for the price?

I've heard of this Spinside Mac a number of times, but I've never been
able to find it.  Archie comes up dry.  I haven't been able to locate it
on ftp.apple.com either.  Any idea where it is exactly?  Also, is this
Kernighan document available on the net?

Thanks,
Clay

+++++++++++++++++++++++++++

>From [e-mail] (John W. Baxter)
Date: Mon, 19 Sep 1994 12:05:37 -0700
Organization: Internet for the Olympic Peninsula

In article <35j50c$[e-mail]>, Clay Thurmond <[e-mail]> wrote:

> In article <[e-mail]> Ken Long, [e-mail]
> writes:
> >Kernighan on C - a big text (w.p.) document by Brian Kernigan on the C
> >language, SpInside Mac, and as much example source as could be found and
> >downloaded.  The books are all expensive and many only give you example
> >code and a book full of comments.  Some have useful info not found in
> >example code, but is it absolutely necessary for the price?
>
> I've heard of this Spinside Mac a number of times, but I've never been
> able to find it.

SpInside Mac is a monster Hypercard stack (with bunches of supporting
files).  That in itself isn't so bad, but...SpInside Mac is also very
obsolete (roughly speaking, it is part way between volumes I through V and
volume VI, with some extras).

I've ignored it on the last couple of years of CDs it has come on, and
never looked for it on the net.  --John

--
John Baxter    Port Ludlow, WA, USA  [West shore, Puget Sound]
   "Occasionally...astronomers add a second to either June 31 or
    December 31..."   IM: OS Utilities, p 4-12
   [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Dylan the Hippy Wabbit)
Date: Wed, 21 Sep 1994 15:31:11 -0500
Organization: Imperial Cancer Research Fund

In article <35j50c$[e-mail]>, Clay Thurmond <[e-mail]>
wrote:


> I've heard of this Spinside Mac a number of times, but I've never been
> able to find it.  Archie comes up dry.  I haven't been able to locate it
> on ftp.apple.com either.  Any idea where it is exactly?  Also, is this
> Kernighan document available on the net?
>
> Thanks,
> Clay

Ftp to ftp.apple.com, I'm afraid I don't remember the directory name.
Alternatively use http://dts.apple.com if you have a web browser.  Either
way be warned about the time/disk space requirements.  There are eight
parts taking up 7 1/2 meg.  Being multipart you will need unstuffit rather
than stuffit expander.

Once you've got it you'll find it's like learning a language from a
French/English dictionary.  I've been scratching my head over the thing all
year.

Dave Spacey

--
Don't underestimate the abacus......it requires no power, can be made with
any materials you have to hand, and never goes bing in the middle of an
important piece of work.   (Many thanks to Douglas Adams.)

+++++++++++++++++++++++++++

>From [e-mail] (Simon Fraser)
Date: Wed, 21 Sep 94 19:52:09 BST
Organization: Centre for Population Biology, Imperial College, UK.

In article <35j50c$[e-mail]> Clay Thurmond, [e-mail]
writes:
>I've heard of this Spinside Mac a number of times, but I've never been
>able to find it.  Archie comes up dry.  I haven't been able to locate it
>on ftp.apple.com either.  Any idea where it is exactly?  Also, is this

OK, here it is!

ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt1.hqx
ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt2.hqx
ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt3.hqx
ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt4.hqx
ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt5.hqx
ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt6.hqx
ftp.apple.com:dts/mac/docs/stacks/spinside/spinside-mac-4-0-2-pt7.hqx

The whole lot uncompresses into a 10.8Mb folder.
I find it really useful (well, more so before I started to buy the
new Inside Macintosh books (you'll see them referred to as NIM: Text
for example).
Stretching out the Hypercard stack to the height of a 14" monitor
makes it a lot easier as well. I did this by a combination of
'set the userlevel to 5' in the message box, and having a new,
unprotected stack open at the same time, which enabled me to get
at the tool palette.

You might also find the technotes stack useful-- 5.9 Mb worth!

ftp.apple.com:dts/mac/docs/stacks/tnstack/tech-notes-stack-4-0-4-pt1.hqx
ftp.apple.com:dts/mac/docs/stacks/tnstack/tech-notes-stack-4-0-4-pt2.hqx
ftp.apple.com:dts/mac/docs/stacks/tnstack/tech-notes-stack-4-0-4-pt3.hqx
ftp.apple.com:dts/mac/docs/stacks/tnstack/tech-notes-stack-4-0-4-pt4.hqx
ftp.apple.com:dts/mac/docs/stacks/tnstack/tech-notes-stack-4-0-4-pt5.hqx

Hope this helps!

Simon
P.S. the best time to access files of this size of off-peak at a weekend.
It's dead fast early on a Sunday morning from the UK!
__________________________________________________________________________
Simon Fraser                            NERC Centre for Population Biology
[e-mail]                         Imperial College at Silwood Park
                                             Ascot, Berkshire, SL5 7PY. UK

---------------------------

>From [e-mail] (Brian S Eltoft)
Subject: Help: idle, shutdown, and MacTCP examples..
Date: 19 Sep 1994 13:31:21 GMT
Organization: Iowa Computer Aided Engineering Network


I'm looking for source code that shows how you would write a program to
sense that the mac has been idle for sometime... as well as an example
of a safe shutdown (as if I'm using the shutdown from the special menu).
I tried the example found in Inside Mac, but it gives me an error of
-609 when I try the AESend(..)  I was also looking about for some examples
of MacTCP...  something that would do a finger...

If anyone has any info on where to look for this code could you please mail
it to: [e-mail]

--
Brian Eltoft
Email:  [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Brook )
Date: 19 Sep 1994 23:28:43 GMT
Organization: The University of Western Australia

[e-mail] (Brian S Eltoft) writes:


>I'm looking for source code that shows how you would write a program to
>sense that the mac has been idle for sometime... as well as an example
>of a safe shutdown (as if I'm using the shutdown from the special menu).
>I tried the example found in Inside Mac, but it gives me an error of
>-609 when I try the AESend(..)  I was also looking about for some examples
>of MacTCP...  something that would do a finger...

 The following function is a correction to the one shown in IM:Processes
and works for me.
FUNCTION ShutDownSafely : OSErr
CONST
    kFinderSig = 'MACS'
VAR
    myErr:        OSErr;
    finderAddr:    AEDesc;
    myShutDown:    AppleEvent;
    nilReply:        AppleEvent;
    signature:    OSType;
BEGIN
    signature := kFinderSig;
    myErr := AECreateDesc ( typeApplSignature,
                            @signature,SizeOf(OSType),
                            finderAddr);
    IF myErr = noErr THEN
        myErr := AECreateAppleEvent ( kAEFinderEvents,
                    kAEShutDown,finderAddr,
                    kAutoGenerateReturnID,
                    AnyTransactionID,myShutDown);
    IF myErr = noErr THEN
        myErr := AESend ( myShutDown,nilReply,
                    kAENoreply + kAECanSwitchLayer
                    + kAEAlwaysInteract,
                    kAENormalPriority,
                    kAEDefaultTimeOut,nil,nil);
    ShutDownSafely := myErr;
END;

--
Brooklyn Waters (Senior Programmer)    email: [e-mail]
Centre for Computer Managed Training and Decision Support Systems
The University of Western Australia    Tel: [phone]
NEDLANDS WA 6009   AUSTRALIA           FAX: [phone]

+++++++++++++++++++++++++++

>From [e-mail] (David Seale)
Date: 21 Sep 1994 22:08:32 GMT
Organization: Magic Online Services Toronto Inc.

"I'm looking for source code that shows how you would write a program to
sense that the mac has been idle for sometime... as well as an example
of a safe shutdown (as if I'm using the shutdown from the special menu).
I tried the example found in Inside Mac, but it gives me an error of
-609 when I try the AESend(..)"

I seem to remember that the -609 error is because the example in IM:Processes
is wrong.
I think that the solution is to change the Finder signature to 'MACS',
instead of 'FNDR' or whatever they have.
If that doesn't work, use the process manager to look for the Finder, and
then use the ProcessSerialNumber version of the AE (That's what I did).

The idle thing is not as hard as it sounds. I used a timer task to check to
see if any of the keys were down on the keyboard (using GetKeys), and if the
mouse was still in the same position, works very well.

Hope this is helpful (and that I'm not confused again).

Dave.
[e-mail]


---------------------------

>From [e-mail] (Stuart Fehr)
Subject: Pascal Flavours for Macintosh
Date: 19 Sep 1994 19:16:07 GMT
Organization: University of Saskatchewan

Our campus bookstore sells two Pascal for Macintosh software packages.
They are:  Think Pascal by Symantec, and Pascal for the Macintosh, also
by Symantec.

What are the differences?  Or, maybe I should ask, what is Pascal for the
Macintosh?  I haven't read or heard anything about it at all.  I have
been reading the posts on Think Pascal, but have seen no mention of this
other one.

There is a significant price difference too -- Pascal for the Macintosh
is approx. $20.00 more than think Pascal.

Any suggestions or information would be appreciated.

+++++++++++++++++++++++++++

>From [e-mail] (Peter N Lewis)
Date: Wed, 21 Sep 1994 10:40:25 +0800
Organization: NCRPDA, Curtin University

In article <35ko1n$[e-mail]>, [e-mail] (Stuart
Fehr) wrote:

>Our campus bookstore sells two Pascal for Macintosh software packages.
>They are:  Think Pascal by Symantec, and Pascal for the Macintosh, also
>by Symantec.

No idea.  Symantec used to (and may still do) sell a Juse Enough Pascal
which was a education tool or something, I'm not sure.

Pascal Compilers:

THINK Pascal 4.0.2
Object Pascal.  68k only.  Very nice environment.  Basically bug free, but
the program is dead (it will likely never be updated) and is growing
incompatiblities with each new system version.

Metrowerks CodeWarrior Pascal
No Objects (scheduled for the end of the year).  Native & 68k, both are
still in development and are not bug free.  Quite nice environment, also
still in development.  Not quite production quality as yet, but this
program is getting being worked on full time by Metrowerks, and will
almost certainly be the Pascal compiler to choose after the end of the
year.

Language Systems Pascal
Object Pascal.  Native (& 68k?).  Still in beta, I dont know how solid it
is.  MPW evironment only, so not really suitable for education/lab use if
you want a friendly environment.

MPW Pascal
Object Pascal.  68K only.  Free from serious bugs.  Doesn't cut it for a
standalone project when compared to Think Pascal but useful when you need
to build something strange (INITs, DRVRs, etc).

Enjoy,
   Peter.
--
Peter N Lewis <[e-mail]> - Macintosh TCP fingerpainter
FTP my programs from redback.cs.uwa.edu.au:Others/PeterLewis/ or
amug.org:pub/peterlewis/ or nic.switch.ch:software/mac/peterlewis/

---------------------------

>From [e-mail] (Bob Bradley)
Subject: Q: Script-Independant Command-Keys?
Date: Sun, 18 Sep 1994 10:54:40 -0800
Organization: SS Software Inc.

I have a simple dialog for "Don't Save, "Cancel", "Save" and want to
handle commands keys for the dialog correctly. Right now, I'm just
checking for 'D', Escape or Command Period (I am doing international
canceling), and Return or Enter, respectively but, I know there's a better
way.

I'd like to make the code localizable and to also get rid of data (the raw
'D' letter, etc..) from the source code (so I don't have t re-compile to
change it).

How would I go about this? I assume it would all be thru the use of the
Script Manager but, I'm not sure on the implementation.

+++++++++++++++++++++++++++

>From [e-mail] (Pete Resnick)
Date: Mon, 19 Sep 1994 23:38:05 -0500
Organization: University of Illinois at Urbana-Champaign

In article <[e-mail]>, [e-mail]
(Bob Bradley) wrote:

> I have a simple dialog for "Don't Save, "Cancel", "Save" and want to
> handle commands keys for the dialog correctly. Right now, I'm just
> checking for 'D', Escape or Command Period (I am doing international
> canceling), and Return or Enter, respectively but, I know there's a better
> way.
>
> I'd like to make the code localizable and to also get rid of data (the raw
> 'D' letter, etc..) from the source code (so I don't have t re-compile to
> change it).
>
> How would I go about this? I assume it would all be thru the use of the
> Script Manager but, I'm not sure on the implementation.

For single-byte languages, this is easy: Just do a GetControlTitle() for
each of the controls in the dialog (doing a GetDialogItem() to get the
control handles) and then compare the keystroke with the first character
of the control title. Use the same method you do for command-period to
make sure that you get the correct ASCII character code, and then make
sure to use the CompareText() or IdenticalText() routines so that case and
diacriticals are handled properly.

For double-byte languages, you're going to have a problem since the
character codes for a single character are going to come through in two
events. In this case, you could get really cute and force the keyboard to
something that will produce single-bytes only (like Romanji in Japanese)
and then Transliterate() the names of the buttons and compare the first
character, but I think this gets a little absurd. Personally, if the
current font for the buttons (i.e. the system font or the window font of
your dialog if the buttons are useWFont) is a double-byte character font,
I would punt on the "command key for the first letter of the button"
method.

pr
--
Pete Resnick    (...so what is a mojo, and why would one be rising?)
Doctoral Student - Philosophy Department, Gregory Hall, UIUC
System manager - Cognitive Science Group, Beckman Institute, UIUC
Internet: [e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Bob Bradley)
Date: Sun, 18 Sep 1994 14:19:30 -0800
Organization: SS Software Inc.

In article <[e-mail]>,
[e-mail] (Pete Resnick) wrote:

> For single-byte languages, this is easy: Just do a GetControlTitle() for
> each of the controls in the dialog (doing a GetDialogItem() to get the
> control handles) and then compare the keystroke with the first character
> of the control title. Use the same method you do for command-period to
> make sure that you get the correct ASCII character code, and then make
> sure to use the CompareText() or IdenticalText() routines so that case and
> diacriticals are handled properly.
>
> For double-byte languages, you're going to have a problem since the
> character codes for a single character are going to come through in two
> events. In this case, you could get really cute and force the keyboard to
> something that will produce single-bytes only (like Romanji in Japanese)
> and then Transliterate() the names of the buttons and compare the first
> character, but I think this gets a little absurd. Personally, if the
> current font for the buttons (i.e. the system font or the window font of
> your dialog if the buttons are useWFont) is a double-byte character font,
> I would punt on the "command key for the first letter of the button"
> method.

I'd like to figure out a good method for storing the command-keys in a
resource which I could pull out later since a few of the button
command-keys I want to use are not the same as the first letter of the
button (ie. Add is usually command-G).

Is there anyway to tell if, when a double-byte script is installed the
user types the first half of a double-byte character? If so, I could
detect this, store the first half and if the second half comes thru in the
next event, do my compare as I would with a regular key.

I'd like to store the command-keys as a string resource with each key as
an individual string with the string id (index if in a STR# resource)
relating the particular button. Could I just compare the key entered with
the key stored in the string to see if they are equal or would I have to
use a special Script Manager comparison routine?

+++++++++++++++++++++++++++

>From [e-mail] (Ari I Halberstadt)
Date: Wed, 21 Sep 1994 04:19:44 GMT
Organization: The World Public Access UNIX, Brookline, MA

In article <[e-mail]>,
Pete Resnick <[e-mail]> wrote:
>In article <[e-mail]>, [e-mail]
>(Bob Bradley) wrote:
>
>> I have a simple dialog for "Don't Save, "Cancel", "Save" and want to
>> handle commands keys for the dialog correctly. Right now, I'm just
>> checking for 'D', Escape or Command Period (I am doing international
>> canceling), and Return or Enter, respectively but, I know there's a better
>> way.
>>...
>For single-byte languages, this is easy: Just do a GetControlTitle() for
>each of the controls in the dialog (doing a GetDialogItem() to get the
>control handles) and then compare the keystroke with the first character
>of the control title. Use the same method you do for command-period to
>make sure that you get the correct ASCII character code, and then make
>sure to use the CompareText() or IdenticalText() routines so that case and
>diacriticals are handled properly.

A resource that defines equivalence between buttons and command keys.
Use some special resource type, say "CMDK" or "DLCK" (dialog command
key), or whatever, and give it the same ID as the dialog or alert. Or
a <clover><char> syntax you tack onto the end of the button names,
then extract before displaying the dialog or alert. Other languages
might not have unique first-character button names, and you might not
want command key equivalence for some buttons, etc. so storing it in
the resource data is more flexible.
--
Ari Halberstadt                                 [e-mail]
One generation passes away, and another generation comes: but the
earth abides for ever. -- Ecclesiastes, 1:4.

+++++++++++++++++++++++++++

>From [e-mail] (Paul Lalonde)
Date: Thu, 22 Sep 1994 23:46:20 GMT
Organization: Metrowerks

In article <[e-mail]>, [e-mail]
(Bob Bradley) wrote:

> I have a simple dialog for "Don't Save, "Cancel", "Save" and want to
> handle commands keys for the dialog correctly. Right now, I'm just
> checking for 'D', Escape or Command Period (I am doing international
> canceling), and Return or Enter, respectively but, I know there's a better
> way.
>
> I'd like to make the code localizable and to also get rid of data (the raw
> 'D' letter, etc..) from the source code (so I don't have t re-compile to
> change it).
>
> How would I go about this? I assume it would all be thru the use of the
> Script Manager but, I'm not sure on the implementation.

You would basically need a modal dialog filter proc in which, every time
a key down was received with the command key down, you loop through all of
the items in the dialog's item list.  If an item is a button, you retrieve
its title and compare the first character with the character just pressed.

Of course, if you also want to deal with double-byte systems, you need to
store state information between calls to the filter proc so you will know
if a key corresponds to the first or second byte of a character.

Paul Lalonde
[e-mail]

+++++++++++++++++++++++++++

>From [e-mail] (Bob Bradley)
Date: Wed, 21 Sep 1994 16:21:17 -0800
Organization: SS Software Inc.

In article <[e-mail]>,
[e-mail] (Paul Lalonde) wrote:

> You would basically need a modal dialog filter proc in which, every time
> a key down was received with the command key down, you loop through all of
> the items in the dialog's item list.  If an item is a button, you retrieve
> its title and compare the first character with the character just pressed.
>
> Of course, if you also want to deal with double-byte systems, you need to
> store state information between calls to the filter proc so you will know
> if a key corresponds to the first or second byte of a character.

The way I'm doing it now is to create a key item command table when the
dialog is create based on a custom resource that contains the key (just a
2 byte short), the item it goes with, and which modifiers are required
(ie. command, option, control) then when I get a key down, in my proc to
handle key downs in dialogs (they are all modeless) I search the table for
the typed key and check whether or not the correct modifiers were down
when it was pressed (I also use the method in the Tech Note International
Canceling to work with keyboards that discharge the shift key when the
command key is down). I also made a simple Resorcerer template for
creating and editing the key item command table

It seems to work ok but, I haven't tested it with any non-roman scripts.

---------------------------

>From [e-mail] (Geoff Pointer)
Subject: XCMD memory residence
Date: 19 Sep 1994 10:50:46 GMT
Organization: Individual

   I am trying to manage an external window for Hypercard. It is geared
to handle several windows open at once and I have handled the fact that
all windows will be using the same copy of the XCMD code. What I am not
clear about, is what can happen to the code resource in between event
calls. I am aware that it can be moved in memory but can it ever be
purged? I am presently basing my logic on the code not being purged and
being copied from its original location in memory when it is moved rather
than recopied from the original resource. Is it possible to get definitive
answers??
   My problem involves how I use certain globals (via A4, I use Code
Warrior). I use RefCon stuff for the globals that go with each individual
window, but I need special action for the first window open and the last
window closed, which involves retaining knowledge across individual event
calls.

- --------- Cheers - Geoff %^> ----------

  Voice: 61 8 362 9890     63 Lambert Rd
  Fax:   [phone]     Joslin SA 5070
                           Australia

- ---------------------------------------


+++++++++++++++++++++++++++

>From [e-mail] (Eric Baumgartner)
Date: Mon, 19 Sep 1994 16:27:53 -0500
Organization: Northwestern University

In article <35jqe6$[e-mail]>,
[e-mail] (Geoff Pointer) wrote:

> all windows will be using the same copy of the XCMD code. What I am not
> clear about, is what can happen to the code resource in between event
> calls. I am aware that it can be moved in memory but can it ever be
> purged?

>    My problem involves how I use certain globals (via A4, I use Code
> Warrior). I use RefCon stuff for the globals that go with each individual
> window, but I need special action for the first window open and the last
> window closed, which involves retaining knowledge across individual event
> calls.

Yes, your xcmd can be purged in between event calls. If you want to use
"global globals" I know of two options:

1. Tell HC that you have interrupt code by calling XWHasInterruptCode.
Evil, evil, evil, but it'll keep your code resource resident and locked.
2. Use a HC global to store your globals. You could store the address of a
handle containing the globals, or you could store the globals themselves
in text form in the HC variable.

I'd be curious to hear if anyone has any other suggestions.

Eric Baumgartner                       *   Email: [e-mail]
Learning Sciences                      *   Phone: [phone]
School of Education and Social Policy  *
Northwestern University                *

+++++++++++++++++++++++++++

>From [e-mail] (Ari I Halberstadt)
Date: Tue, 20 Sep 1994 01:01:20 GMT
Organization: The World Public Access UNIX, Brookline, MA

In article <35jqe6$[e-mail]>,
Geoff Pointer <[e-mail]> wrote:
>   I am trying to manage an external window for Hypercard. It is geared
>to handle several windows open at once and I have handled the fact that
>all windows will be using the same copy of the XCMD code. What I am not
>clear about, is what can happen to the code resource in between event
>calls. I am aware that it can be moved in memory but can it ever be
>purged? I am presently basing my logic on the code not being purged and
>being copied from its original location in memory when it is moved rather
>than recopied from the original resource. Is it possible to get definitive
>answers??
>   My problem involves how I use certain globals (via A4, I use Code
>Warrior). I use RefCon stuff for the globals that go with each individual
>window, but I need special action for the first window open and the last
>window closed, which involves retaining knowledge across individual event
>calls.
>

You can mark the XCMD resource as non-purgeable. But that still
doesn't prevent HyperCard, or anyone else, from explicitely purging
it. Better to set a HyperCard global variable from within your XCMD.
There are a couple of call-backs for setting and retrieving a global
variable's value. When your XCMD exits, it sets the value of the
global to a handle to some structure containing the state information.
When the last window is disposed of, you clear the global (set it to
the empty string). You should give the global some unlikely name. For
instance, if the XCMD is called "MyXCMD", I'd call the global
"gMyXCMD". The "g" is just that Apple pseudo-standard for global
variables, and the "MyXCMD" part makes it clear to anyone looking at
the global in the debugger that it's part of your XCMD, and also
reduces the chance that someone will use the same name for their own
global variable.
--
Ari Halberstadt                                 [e-mail]
One generation passes away, and another generation comes: but the
earth abides for ever. -- Ecclesiastes, 1:4.

+++++++++++++++++++++++++++

>From [e-mail] (John Kevin Calhoun)
Date: 22 Sep 1994 16:45:21 -0700
Organization: Apple Computer Inc, Cupertino, CA


In article <35jqe6$[e-mail]>,
Geoff Pointer <[e-mail]> wrote:
>   I am trying to manage an external window for Hypercard. It is geared
>to handle several windows open at once and I have handled the fact that
>all windows will be using the same copy of the XCMD code. What I am not
>clear about, is what can happen to the code resource in between event
>calls. I am aware that it can be moved in memory but can it ever be
>purged? I am presently basing my logic on the code not being purged and
>being copied from its original location in memory when it is moved rather
>than recopied from the original resource. Is it possible to get definitive
>answers??

It's not possible to prevent an XCMD or XFCN resource from being purged
unless you also prevent it from being moved.  You can do this with
the callback XWHasInterruptCode, which, in effect, locks the resource
in memory.  However, leaving something locked in an unpredictable
location in HyperCard's heap is a very unfriendly thing to do, unless
it's absolutely necessary.

Furthermore, Apple does not guarantee that HyperCard's handling of
XCMD and XFCN resources will remain the same with each release of
HyperCard; in fact, it changed between 1.x and 2.x and might change
again.  Therefore I think it's not wise to rely on the way HyperCard
currently executes external code.

>   My problem involves how I use certain globals (via A4, I use Code
>Warrior). I use RefCon stuff for the globals that go with each individual
>window, but I need special action for the first window open and the last
>window closed, which involves retaining knowledge across individual event
>calls.

Instead of relying on A4 globals, you can store and update a count
of your windows in a HyperTalk global, which you can access via the
SetGlobal and GetGlobal callbacks.

With HyperCard 2.2, it's possible to count your windows on the fly,
as follows:

  function countWindowsOf externalName
    put 0 into count
    repeat with thisWindow = 1 to the number of windows
      if the owner of window thisWindow is externalName
      then add 1 to count
    end repeat
    return count
  end countWindowsOf

>From an external you could execute similar HyperTalk code by
using the RunHandler callback.

Kevin Calhoun
[e-mail]

---------------------------

End of C.S.M.P. Digest
**********************


ˇ