Celestin Apprentice 5/Apprentice-Release5.iso/Information/CSMP Digest/volume 3/csmp-digest-v3-119

Text File | 1995-12-31 | 56KB | 1,503 lines

C.S.M.P. Digest             Tue, 31 Oct 95       Volume 3 : Issue 119

Today's Topics:

        I've hit a brick wall!
        Implementing Command-. and Escape for cancel in dialogue
        PPC ProcPtr functions?
        Perl for Mac?
        Popup Menus in 9-pt geneva ?????
        Q: using OSAExecute to singlestep AppleScript
        Reading Notification Manager queue?



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

The digest is a collection of article threads from the internet newsgroups
comp.sys.mac.programmer.help, csmp.tools and csmp.misc. It is designed for
people who read news 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].

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

>From Keith Wiley <[e-mail]>
Subject: I've hit a brick wall!
Date: Mon, 2 Oct 1995 23:01:19 -0400
Organization: University of Maryland College Park

Ok, I can't continue until I learn how to segment my project because I've
hit the 32K limit.  I posted some questions earlier and got several
responses, thank you.  However, I still have some major problems.

Aside from making header files, (I gave up on that temporarily), I'm just
trying to put a full copy of all #define, #include, typedefs, and globals
at the top of each .c source file.  It seems to work except that I get
link errors complaining about multidefined global variables.  If I
eliminate the declarations from all but one file then I get undeclared
errors (of course).  I can't figure out how to get around this.  I'm
putting extern in all the declarations.  I've tried putting extern in all
files except the one with main().  I've tried everything.  What do I need
to do?

Also, what's up with making header files?

. . .. ... ..... ........ ............. .....................
.. ... ..... ....... ........... ............. .................
. .. .... ........ ................ ................................
Keith Wiley, Electrogenetic Engineer             *
University of Maryland at College Park            *   * *     *   *      *
email:  [e-mail]                      ***    **   * *    **     *
world wide web:  http://www.wam.umd.edu/~keithw        *     **   **    ***


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

>From [e-mail] (Kevin Tieskoetter)
Date: 3 Oct 1995 01:31:53 GMT
Organization: Albatross Productions

In article <[e-mail]>
Keith Wiley <[e-mail]> writes:

> Aside from making header files, (I gave up on that temporarily), I'm just
> trying to put a full copy of all #define, #include, typedefs, and globals
> at the top of each .c source file.  It seems to work except that I get
> link errors complaining about multidefined global variables.  If I
> eliminate the declarations from all but one file then I get undeclared
> errors (of course).  I can't figure out how to get around this.  I'm
> putting extern in all the declarations.  I've tried putting extern in all
> files except the one with main().  I've tried everything.  What do I need
> to do?

You can't give up on making header files - they're unavoidable.

A header file's purpose is to contain:

a) a prototype of all of the functions in the .c file that you want
other files to have access to
b) any #defines, typedefs that are related to the .c file, but you want
to have access to in any other files
c) >>>extern<<< declarations for globals that are in the .c file that
you want other files to have access to (I'll explain in a moment)

Notice that the common thread between the a, b, and c is that the
header file is for things that you want other files to have access to.
If you have a typedef structure in one .c file that you don't need to
use in any other file, don't put it into the header file (well, you
can, but it's good practice not to). Likewise, if you have a structure
that you need to use in more than one .c file, put it in a header file,
and include that header in each .c file that needs the structure.

If you're not familiar with prototypes, they are statements that tell
the compiler how a particular function is called. If you open up any
Macintosh header file (Dialogs.h for example), you'll find tons of
prototypes. Protypes are simple. If your function looks like this:

short MyFunction(short myInput)
{
 ....code.....
 return myShort;
}

Your prototype would look like this:

short MyFunction(short myInput);


Essentially, it's the declaration of the function, with a semicolon (;)
after it. Prototypes are a wonderful thing - they allow the compiler to
make sure you're calling your functions correctly and eliminate lots of
errors later in the process. I recommend that you change your compiler
settings to "Require Function Prototypes" and that you NEVER turn it
off.

Now, globals. You ONLY want to define globals inside .c files. As you
probably know, globals are defined like this:


short  gMyGlobalVariable;

When the compiler encounters this, it says "Oh, I need to allocate
space inside the code for this variable named "gMyGlobalVariable". If
you were to put this global inside a .h file, you would be telling the
compiler to allocate space for the variable every time it encountered
the declaration. Since this is obviously an error, the compiler won't
let you declare globals with the same name more than once (static
globals are an exception, but you don't need to worry about those yet).

But your other files may also need access to the same global. The
solution is to declare a sort of "prototype" for the global, which is
called an "extern". It is done like this:

extern short gMyGlobalVariable;

When the compiler encounters an extern global, it just says to itself
that "oh,  gMyGlobalVariable must be defined some other place; I'll let
the code compiler, and I'll figure it out later".

When you declare a global that needs to be used in more than one file,
you need to declare both the global itself (in the .c file) AND the
"extern" global (in the corresponding .h file).


Then, at the top of each of your .c files, figure out which of the .h
files you need to include to get it to work correctly, and add those
#include statements.


You CANNOT get by in C without using .h files!


Good luck!


-kevin

//---------------------------------------------------------------------
   Kevin Tieskoetter / Software Prestidigitator, Specular International
//---------------------------------------------------------------------

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

>From [e-mail] (Ken Long)
Date: Tue, 3 Oct 1995 07:24:49 GMT
Organization: NETCOM On-line Communication Services (408 261-4700 guest)

Keith Wiley ([e-mail]) wrote:
: Ok, I can't continue until I learn how to segment my project because I've
: hit the 32K limit.  I posted some questions earlier and got several
: responses, thank you.  However, I still have some major problems.

In Think C 5, if you had a segment that was too big, it would just tell
you as a link error.  In Think C 7 it asks if you'd like your project
automatically segmented.

: Aside from making header files, (I gave up on that temporarily), I'm just
: trying to put a full copy of all #define, #include, typedefs, and globals
: at the top of each .c source file.  It seems to work except that I get
: link errors complaining about multidefined global variables.

I try to simplify any projects I download, by putting all the source and
header data all in one file.  Strive for simplification anywhere possible.
If a set of headers and source files merged end up being less than about
90K it usually works fine.

All a header is, really, is a setup for that which is remaining.  Like a
newspaper headline is a header for the article.  The headboard of a bed
is nearest the head.  If you head your boat into the wind, the head (bow)
goes into the wind first.

Defines can go in one header file and be included by all files, as far as
I've seen, and give no errors.  Same with typedefs.  Global variables
MUST be delcared only once, in one file.  You do not need to "#include"
that file in any other file.  You merely make the declaration in the
other files that use the global variables *except* you put the keyword
"extern' infront of it.

Example:

In main.c, you say "short thisShort;" and, in events.c, *if* it uses that
variable, you say "extern short thisShort;" - it's that simple.

If you make a "MyProg.h" file and put "short thisShort;" in it and in both
main.c and event.c you say "#include "MyProg.h"" you will get a multiply
defined error on "thisShort"

: If I
: eliminate the declarations from all but one file then I get undeclared
: errors (of course).  I can't figure out how to get around this.  I'm
: putting extern in all the declarations.  I've tried putting extern in all
: files except the one with main().  I've tried everything.  What do I need
: to do?

If you put "extern" in front of *all* declarations, then it looks for the
declaration externally and doesn't find it because it's referred to an
external declaration every time.

The rule of thumb is:  "Declare the global in the first file it's used
in, then declare it "extern" in any *subsequent* files it's used in."

Like this:

// FileOne.c

short thisIsADeclaration;

TheFunctionAtTheJunction ()
{
    thisIsADeclaration = 1;

    DoBackFlips (thisIsADeclaration);
}
// end of file.

//FileTwo.c

extern short thisIsADeclaration;

FunctionThatsMunchin ()
{
    thisIsADeclaration = 4;

    DanceTheBooGaLoo (thisIsADeclaration);
}
//end of file.

This stuff is in the manual.  RTFM!  This is not stuff than can be, or
needs to be "figured out."  It needs to be read and understood and SEEN.
Look at example source project.  SEE how they are.  Read up on the layout
and operation of the programming environment you are using and the C
programming language if need be.

If you don't have a manual, write to the publisher and see about getting
one. All header files do is define things that would normally be defined
IN the source file where what's defined is *used*.

If you open a source file which has headers included and select
"Preprocess" from the appropriate menu, it will make a new file where all
the definitions and source are contained within it.  No header files
needed because everything that file uses is defined right in it. #defines
are "eaten" and what was #defined before is replaced by what it was
#defined as. In other words, the code is "washed" in preparation to
compiling.

Headers (predefinition/declaration before use) are needed, but header
files are not needed.  They are ONLY for convenience of the programmer.
They save him typing, copying and pasting and make his source file nice
and neat.

Open a project where a .c file has a #include...  Now open that header
and select all and copy, then close it.  Now select the line in the
source that said "#include "ThatHeader.h" and paste.  The include line
will be replaced by that which was included.  It's included internally,
now.  That's the only difference.  You don't have to say to include it
because it's already included.

-Ken-

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

>From [e-mail] (Carl Gustafson)
Date: 3 Oct 1995 11:57:59 GMT
Organization: Imaging and Computer Vision Center, Drexel University

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

> Ok, I can't continue until I learn how to segment my project because I've
> hit the 32K limit.  I posted some questions earlier and got several
> responses, thank you.  However, I still have some major problems.
>
> Aside from making header files, (I gave up on that temporarily), I'm just
> trying to put a full copy of all #define, #include, typedefs, and globals
> at the top of each .c source file.  It seems to work except that I get
> link errors complaining about multidefined global variables.  If I
> eliminate the declarations from all but one file then I get undeclared
> errors (of course).  I can't figure out how to get around this.  I'm
> putting extern in all the declarations.  I've tried putting extern in all
> files except the one with main().  I've tried everything.  What do I need
> to do?
>
> Also, what's up with making header files?

What you need to do is face Cupertino, genuflect, and repeat three times:
I'll try harder next time...

Actually, splitting your code into separate files to make separate
segments is implementation dependent. (Think C and it's relatives) In
other environments (MPW) you would use #pragma segment directives.

That said, you could also configure your compiler and linker to do a
"model far" or "big link" build, where you have branch islands in the
linked code, so that nothing is more than a 16-bit pc-relative jump away,
allowing virtually unlimited code size. PPC-native apps don't use segments
(or, as I say to my Intel-Inside buddies, segments are for worms.) Again,
environment dependent.

Header files should contain declarations (definitions? I never keep the
names straight) for only the externally visible entities in the associated
source code file. If your source file refers to a name defined [<= there,
that must be it!] in another source file, you #include the header file for
that source file.

--
Carl Gustafson
Imaging and Computer Vision Center
Drexel University, Philadelphia, Penna
- ----------------------------------------------------------
I don't speak for Drexel, and Drexel doesn't listen to me...

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

>From [e-mail] (Don Sample)
Date: Wed, 04 Oct 1995 15:53:31 -0400
Organization: Enterprise Solutions Ltd

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

> Ok, I can't continue until I learn how to segment my project because I've
> hit the 32K limit.  I posted some questions earlier and got several
> responses, thank you.  However, I still have some major problems.
>
> Aside from making header files, (I gave up on that temporarily), I'm just
> trying to put a full copy of all #define, #include, typedefs, and globals
> at the top of each .c source file.  It seems to work except that I get
> link errors complaining about multidefined global variables.  If I
> eliminate the declarations from all but one file then I get undeclared
> errors (of course).  I can't figure out how to get around this.  I'm
> putting extern in all the declarations.  I've tried putting extern in all
> files except the one with main().  I've tried everything.  What do I need
> to do?
>
> Also, what's up with making header files?
>
How you segment a program depends on what your development environment.
In MPW you use #pragma statements to tell the compiler/linker which
segment to put functions into.  In the Symantec and Metrowerks IDEs you
can move a file to a different segment by dragging it in the project
window.  To create a new segment simply drag it down below the bottom
'totals' line in the project window.


A header file is simply a text file.  You can put anything into it that
you can put into a C source file.  By convention header file names have
the sufix ".h" but they don't have to, you can use any valid file name.
Generally you put information into them which is shared between source
files (#defines, typedefs, function prototypes, etc.)

To include a header file in your source you use the #include statement.
For example if you have created a header file named "MyHeader.h" you would
include it using the statement:

    #include "MyHeader.h"

I suspect that you are trying to do something like "#include <MyHeader.h>"
and the compiler is telling you that it can't find the file.  The angle
brackets around a file name tell the compiler to look in a specific set of
directories for the header file.  Generally you only use them when
including files which came with the compiler.  For your own header files
use the quotes, which will search the directories containing your source
code for the header files.  (Exactly where the compiler looks for header
files depends on your development environment, but a good general rule of
thumb is that if you wrote the header file use "", and if it came with the
compiler use <>.)

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

>From [e-mail] (Richard Rostrom)
Date: 12 Oct 1995 04:05:46 GMT
Organization: Educational Computing Network

Don Sample ([e-mail]) wrote:
: In article <[e-mail]>,
: Keith Wiley <[e-mail]> wrote:

: > Aside from making header files, (I gave up on that temporarily), I'm just
: > trying to put a full copy of all #define, #include, typedefs, and globals
: > at the top of each .c source file.  It seems to work except that I get
: > link errors complaining about multidefined global variables.  If I
: > eliminate the declarations from all but one file then I get undeclared
: > errors (of course).  I can't figure out how to get around this.

declare the global variables in one file. Put extern declarations of
those variables in each of your other files.
To simplify your life, use an include file for the extern declarations,
so that all modules will use the same set of declarations.

This should be done for all typedefs and structs as well.

These should be two different files, BTW, as you don't want to include
the externs in the file where the globals are defined!

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

>From [e-mail] (Scott Kevill)
Date: Sat, 14 Oct 1995 11:35:52 +0800
Organization: The University of Western Australia

In article <45i46q$[e-mail]>, [e-mail]
(Richard Rostrom) wrote:

: Don Sample ([e-mail]) wrote:
: : In article <[e-mail]>,
: : Keith Wiley <[e-mail]> wrote:
:
: : > Aside from making header files, (I gave up on that temporarily), I'm just
: : > trying to put a full copy of all #define, #include, typedefs, and globals
: : > at the top of each .c source file.  It seems to work except that I get
: : > link errors complaining about multidefined global variables.  If I
: : > eliminate the declarations from all but one file then I get undeclared
: : > errors (of course).  I can't figure out how to get around this.
:
: declare the global variables in one file. Put extern declarations of
: those variables in each of your other files.
:
: To simplify your life, use an include file for the extern declarations,
: so that all modules will use the same set of declarations.
:
: This should be done for all typedefs and structs as well.
:
: These should be two different files, BTW, as you don't want to include
: the externs in the file where the globals are defined!

I don't know if it's good practice or not, but something I've done to make
this easier goes like this:

//-----------------------------
// Globals.h
#ifdef __MAIN__
#define extern
#endif

extern int   a, b, c;
extern char  d, e, f;

#undef extern
//-----------------------------
// Main.c
#define __MAIN__
#include "Globals.h"
....
//-----------------------------
// OtherFile.c
#include "Globals.h"
....

It's a bit of a kludge I guess, but it means only one copy of the globals
that I have to update each time. It could be bad that every file gets all
the globals instead of just the ones it needs.

Comments are welcome!

Scott Kevill.
[e-mail]

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

>From [e-mail] (Don Sample)
Date: Mon, 16 Oct 1995 13:52:21 -0400
Organization: Enterprise Solutions Ltd

In article <[e-mail]>,
[e-mail] (Scott Kevill) wrote:

>
> I don't know if it's good practice or not, but something I've done to make
> this easier goes like this:
>
> //-----------------------------
> // Globals.h
> #ifdef __MAIN__
> #define extern
> #endif
>
> extern int   a, b, c;
> extern char  d, e, f;
>
> #undef extern
> //-----------------------------
> // Main.c
> #define __MAIN__
> #include "Globals.h"
> ....
> //-----------------------------
> // OtherFile.c
> #include "Globals.h"
> ....
>
> It's a bit of a kludge I guess, but it means only one copy of the globals
> that I have to update each time. It could be bad that every file gets all
> the globals instead of just the ones it needs.
>
> Comments are welcome!
>

Except for 'extern' being a reserved word which I don't think many
compilers will let you redefine, I've seen this general technique used by
a lot of people, usually something like:

#ifndef __MAIN__
  #define EXTERN extern
#else
  #define EXTERN
#endif

EXTERN int a, b, c;

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

>From [e-mail] (Chris Watson)
Subject: Implementing Command-. and Escape for cancel in dialogue
Date: Tue, 03 Oct 1995 12:32:24 -0400
Organization: Communications Accessibles Montreal, Quebec Canada

I have a movable modal dialogue and would like to implement the Command-.
and escape keys to click the cancel button.  In modal dialogues, I use
SetDialogCancelItem () but this does not seem to work in movable modal
dialgues.

I know I have to check the keyDown event and all that, my question is
whether to check the charcode or the keycode and how specifically to do
that given a keyDown event.  I don't know what the values for the escape
and '.' are either.

Thanks for any help!


>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
  H H      |                   |                                       |
  | |      | Sean McBride      |                                       |
H-C-C-O-H  | [e-mail]   |   "Total destructive interference"    |
  | |      | Montreal, Canada  |                                       |
  H H      |                   |                                       |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

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

>From [e-mail] (Ronald N. Tjoelker)
Date: Tue, 03 Oct 1995 15:07:41 -0700
Organization: Seanet Online Services, Seattle WA

In article <[e-mail]>, [e-mail]
(Chris Watson) wrote:

> I have a movable modal dialogue and would like to implement the Command-.
> and escape keys to click the cancel button.  In modal dialogues, I use
> SetDialogCancelItem () but this does not seem to work in movable modal
> dialgues.
>
> I know I have to check the keyDown event and all that, my question is
> whether to check the charcode or the keycode and how specifically to do
> that given a keyDown event.  I don't know what the values for the escape
> and '.' are either.


if (event.what == keyDown)
{
     c = (event.message & charCodeMask);

     if (c == 13 || c == 3)                       /* return or enter */
     {
          done = true;
     }
     if ((c == '.' && event.modifiers & cmdKey) || c == 27) /* cmd . or esc */
     {
          done = true;
     }
}

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

>From [e-mail] (Erichsen)
Date: Tue, 03 Oct 1995 18:18:13 -0700
Organization: Disorganized

In article <[e-mail]>,
[e-mail] (Ronald N. Tjoelker) wrote:

> if ((c == '.' && event.modifiers & cmdKey) || c == 27) /* cmd . or esc */

When checking for a command-period, you should make sure it will also work
on keyboards that require a cmd-shift to get to the period. There's a
tech-note and some code that shows how to do it.

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

>From [e-mail] (Chris Watson)
Date: Tue, 03 Oct 1995 22:07:20 -0400
Organization: Communications Accessibles Montreal, Quebec Canada

In article <[e-mail]>,
[e-mail] (Ronald N. Tjoelker) wrote:

>> I have a movable modal dialogue and would like to implement the Command-.
>> and escape keys to click the cancel button.
>> I know I have to check the keyDown event and all that, my question is
>> whether to check the charcode or the keycode and how specifically to do
>> that given a keyDown event.
>
>     c = (event.message & charCodeMask);
>
>     if (c == 13 || c == 3)    /* return or enter */
>          done = true;
>     if ((c == '.' && event.modifiers & cmdKey) || c == 27)
>          done = true;

Perhaps my orignal post was not as clear as I had intended.  This isn't
really a programming problem as much as a question of Mac human interface
guidelines.

If I was to use the code you suggested a few issues are raised who's
answers I'm not sure of.

First of all, does the state of the command key matter when testing the
escape key?  Should Command-Escape also click cancel?

Next, if I use this code checking for a keypress with charCode == 27, then
pressing the 'clear' key on the keypad will also click cancel, since
escape and clear have the same char code.  Should I be checking the
keyCode instead of the charCode?

Next, how about the period key?  Should the one on the keypad also be
allowed?  If the user has set in the Numbers control panel that the
decimal character be a ',' (comma) instead of a period should pressing the
keypad 'period' also work?  Again, should I be checking for a keyCode or a
charCode?

Lastly, if I do need to check the keyCode, how do I do this?  (this is a
programming problem).  I know the charCode is event->message &
charCodeMask, how do I extract the keyCode from the event?  event->message
& keyCodeMask does not provide the correct value.

I hope this clears up what I mean... Thanks for any help!!

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
  H H      |                   |                                       |
  | |      | Chris Watson      |                                       |
H-C-C-O-H  | [e-mail]   |   "Total destructive interference"    |
  | |      | Montreal, Canada  |                                       |
  H H      |                   |                                       |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

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

>From [e-mail] (F. Degott)
Date: 4 Oct 1995 13:46:49 GMT
Organization: LMC-IMAG Grenoble France

In article <[e-mail]>, [e-mail]
(Chris Watson) wrote:

>[cut]
>Lastly, if I do need to check the keyCode, how do I do this?  (this is a
>programming problem).  I know the charCode is event->message &
>charCodeMask, how do I extract the keyCode from the event?  event->message
>& keyCodeMask does not provide the correct value.
>
>I hope this clears up what I mean... Thanks for any help!!
>

Hi Chris,

to get the charCode and the keyCode of the event.message field:

Pascal:
charCode:= char(BitAnd(event.message, charCodeMask));
keyCode:= BitAnd(event.message, keyCodeMask) div 256;

C:
charCode = (char)(event.message & charCodeMask);
keyCode = (event.message & keyCodeMask) / 256; {or >>8}

keyCode for <escape key> = $35 (53 decimal)
charCode for <escape key> = $1B (27 decimal)

HTH
Fr
___________________________________________________________________________
FR Degott ([e-mail])
LogiMath, Lab. LMC-IMAG - Grenoble - France

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

>From [e-mail] (Stefan Kurth)
Date: Wed, 04 Oct 1995 18:11:56 +0100
Organization: none

Chris Watson <[e-mail]> wrote:

> I have a movable modal dialogue and would like to implement the Command-.
> and escape keys to click the cancel button.  In modal dialogues, I use
> SetDialogCancelItem () but this does not seem to work in movable modal
> dialgues.

It does work in movable modal dialogs.  All you have to do is call the
standard filter proc.  (GetStdFilterProc)


--
Stefan Kurth
Berlin, Germany

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

>From [e-mail] (Richard Wells)
Date: 4 Oct 1995 23:14:00 GMT
Organization: 3M Company

[e-mail] (Chris Watson) wrote:
>First of all, does the state of the command key matter when testing the
>escape key?  Should Command-Escape also click cancel?
>
>Next, if I use this code checking for a keypress with charCode == 27, then
>pressing the 'clear' key on the keypad will also click cancel, since
>escape and clear have the same char code.  Should I be checking the
>keyCode instead of the charCode?
>
>Next, how about the period key?  Should the one on the keypad also be
>allowed?  If the user has set in the Numbers control panel that the
>decimal character be a ',' (comma) instead of a period should pressing the
>keypad 'period' also work?  Again, should I be checking for a keyCode or a
>charCode?
>
>Lastly, if I do need to check the keyCode, how do I do this?  (this is a
>programming problem).  I know the charCode is event->message &
>charCodeMask, how do I extract the keyCode from the event?  event->message
>& keyCodeMask does not provide the correct value.

One more example: try typing control-c when a dialog is displayed.  Most
applications treat this the same as the enter key, and act as if the
OK button was hit.  Of course, if your users are former DOS types, where
control-c meant cancel, this is rather counter-intuitive.

BTW:  (event->message & keyCodeMask) >> 8 will give you the key code.



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

>From [e-mail] (Sean McBride)
Date: Wed, 11 Oct 1995 19:40:19 -0400
Organization: Communications Accessibles Montreal, Quebec Canada

In article <44v4fo$[e-mail]>, [e-mail] (Richard Wells) wrote:

>One more example: try typing control-c when a dialog is displayed.  Most
>applications treat this the same as the enter key, and act as if the
>OK button was hit.  Of course, if your users are former DOS types, where
>control-c meant cancel, this is rather counter-intuitive.

That's an interesting point, but control-c produces the ascii character
for the enter key (keypad) and I've found every other application allows
this to me used.

Also, isn't it alt-c that cancels with dos, not control-c, since ascii is
used with dos also?

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
  H H      |                   |                                       |
  | |      | Sean McBride      |                                       |
H-C-C-O-H  | [e-mail]   |   "Total destructive interference"    |
  | |      | Montreal, Canada  |   "For Unlawful Carnal Knowledge"     |
  H H      |                   |                                       |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

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

>From [e-mail] (Richard Wells)
Date: 12 Oct 1995 19:09:16 GMT
Organization: 3M - St. Paul, MN  55144-1000 US

[e-mail] (Sean McBride) wrote:
>In article <44v4fo$[e-mail]>, [e-mail] (Richard Wells) wrote:
>
>>One more example: try typing control-c when a dialog is displayed.  Most
>>applications treat this the same as the enter key, and act as if the
>>OK button was hit.  Of course, if your users are former DOS types, where
>>control-c meant cancel, this is rather counter-intuitive.
>
>That's an interesting point, but control-c produces the ascii character
>for the enter key (keypad) and I've found every other application allows
>this to me used.
>
>Also, isn't it alt-c that cancels with dos, not control-c, since ascii is
>used with dos also?

No, its control-c, and therein lies the rub.  If a (reformed) DOS user
tries to use control-c to cancel a Mac dialog (old habits do die hard),
the program will almost always interpret that as a keypad enter, and act
as if the OK button was hit.

To be honest, I've never actually seen a user get bit by this.  The way
we found it was that we have some testing monkeys who, while hitting
random keys in an effort to produce a Shakespeare sonnet, hit control-c
while a dialog was displayed and were startled by the effect.

FWIW: control-c maps to ASCII 3, which is the ETX character.  I'm reasonably
sure that ETX stands for "end transmission".  This (mostly reformed)
programmer has to admit it makes more sense to interpret this as OK
rather than cancel, especially since control-x is the ASCII CAN character,
which I'm reasonably sure is short for "cancel".  But control-c had a
long tradition of meaning "cancel" (or perhaps, more correctly, "stop") in
old DEC systems (RSTS, VMS, and probably RT-11), and that is probably why
it was adopted for MS-DOS.  (Also, CP/M may have used control-c for that
purpose.  Its been awhile.  Have I given away my age yet?)



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

>From [e-mail] (Brian Quinlan)
Subject: PPC ProcPtr functions?
Date: 13 Oct 1995 20:52:36 GMT
Organization: Simon Fraser University

In my program I use, like everyone else, a handler in my TrackControl
call for my scrollbars. However, when I compile my program for PPC
this produces a crash. That makes sense but how do I fix it?

--

Brian Quinlan
[e-mail]

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

>From Darren Giles <[e-mail]>
Date: Sun, 15 Oct 1995 19:40:00 GMT
Organization: NETCOM On-line Communication Services (408 261-4700 guest)

In article <45mjik$[e-mail]> Brian Quinlan, [e-mail]
writes:
>In my program I use, like everyone else, a handler in my TrackControl
>call for my scrollbars. However, when I compile my program for PPC
>this produces a crash. That makes sense but how do I fix it?


The short answer: you need to create a Universal Procedure Pointer
(UPP) for any callback routines.  This is quite simple:

In your globals:
     ControlActionUPP     gMyTrackControlRoutineUPP;

In your initialization code:
     gMyTrackControlRoutineUPP= NewControlActionProc
(MyTrackControlRoutine)

When you call TrackControl:
    result= TrackControl (theControl, &thePoint,
gMyTrackControlRoutineUPP);


- Darren

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

>From [e-mail] (Richard Cardona)
Date: 15 Oct 1995 14:19:59 GMT
Organization: TradeWave Corporation (formerly EINet)

In article <45mjik$[e-mail]>, [e-mail] (Brian
Quinlan) wrote:

> In my program I use, like everyone else, a handler in my TrackControl
> call for my scrollbars. However, when I compile my program for PPC
> this produces a crash. That makes sense but how do I fix it?
>
> --
>
> Brian Quinlan
> [e-mail]

Try reading Inside Macintosh : PowerPC System Software.  It will explain
all about UniversalProcPtrs and how to fix your situation.  You could
probably soak up what you need in 1 bookstore visit or go to Apple's Web
site with Inside Mac online.

Rick

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

>From [e-mail] (Adam Lloyd)
Subject: Perl for Mac?
Date: Fri, 13 Oct 1995 18:32:00 +0100
Organization: a sluggish molehill.

Are there any good implementations of Perl for the Mac, other than the Perl
Tool for MPW? I'm thinking along the lines of something stand-alone...

Adam.



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

>From [e-mail] (Matthias Neeracher)
Date: 16 Oct 1995 16:41:52 GMT
Organization: Integrated Systems Laboratory, ETH, Zurich

In article <[e-mail]>, [e-mail] (Adam Lloyd) writes:
> Are there any good implementations of Perl for the Mac, other than the Perl
> Tool for MPW? I'm thinking along the lines of something stand-alone...

ftp://ftp.switch.ch/software/mac/perl/
ftp://ftp.share.com/pub/macperl/

Matthias

- ---
Matthias Neeracher <[e-mail]> http://err.ethz.ch/members/neeri.html
   "One fine day in my odd past..." -- Pixies, _Planet of Sound_

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

>From Gordon Tillman <[e-mail]>
Date: 15 Oct 1995 14:07:55 GMT
Organization: n/a

Hi Adam,

In article <[e-mail]> Adam Lloyd,
[e-mail] writes:
>Are there any good implementations of Perl for the Mac, other than the
Perl
>Tool for MPW? I'm thinking along the lines of something stand-alone...


You can use MacPerl.  It's been a while since I got it, so there is
probably a newer version around.  I have version 4.1.3, Patchlevel 36.

In the about box it said the Mac version was written by:

    Matthias Neeracher    [e-mail]
    Tim Endres            [e-mail]

You can probably find it at the usual Mac archives.  If you cannot, let
me know and I will see if I can help.  This version does not require MPW.

--gordon tillman
- ----------------------------------------------------------------------
      Internet: [e-mail]       |  Voice: 615-238-6506
                [e-mail]           |    Fax: [phone]
America Online: GordyT                   |US Mail: 7312 Flagstone Drive
    CompuServe: 74372,1071               |         Ooltewah, TN  37363
- ----------------------------------------------------------------------

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

>From [e-mail] (Ben Hellstrom)
Subject: Popup Menus in 9-pt geneva ?????
Date: Tue, 03 Oct 1995 20:27:46 -0400
Organization: jaguNET Access Services

Many of the utilities I have have nice little popups in 9-pt geneva.  They
dont have MDEFs (which means the MDEFs are either embedded in other code
resources or they are using the standard MDEF).

Can anyone give me a clue as to how to produce a menu in 9-point geneva?
Do I have to switch the system font?  If so, how?


TIA

--
Ben Hellstrom
[e-mail]

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

>From [e-mail] (Tim Dierks)
Date: Wed, 04 Oct 1995 00:35:56 -0700
Organization: Best Internet Communications

In article <[e-mail]>,
[e-mail] (Ben Hellstrom) wrote:

>Many of the utilities I have have nice little popups in 9-pt geneva.  They
>dont have MDEFs (which means the MDEFs are either embedded in other code
>resources or they are using the standard MDEF).
>
>Can anyone give me a clue as to how to produce a menu in 9-point geneva?
>Do I have to switch the system font?  If so, how?

Mmmm.... Code (including way-cool trap patches!)

Setting the width will need to be fixed to use a UPP for native
applications. Or, just hack it out. It's sick anyway. Unpatching traps?! Yuk.

- Tim.

Note: I tried to keep this from being wrapped.

// Call just like PopUpMenuSelect, only you can set the font, size, and
width of the popup menu.
//  Pass 0 for font, size, or width to use the defaults.
long
PopUpMenuSelectWithFontAndWidth(MenuHandle menu,short top,short left,short
popUpItem,short font,short size,short width)
{   long                menuChoice;                     // Stores the
chosen menu & item
    short               saveSysFontFam,saveSysFontSize; // Room for saving
the current system font and size
    GrafPtr             wMgrPort;                       // Pointer to the
window manager port
    CalcMenuSizeProc    oldCalcMenuSize;                // Address of the
old CalcMenuSize procedure

    // First, we'll clean up after buggy software.  A number of commercial
products
    // (word processors are most common) screw up the window manager port;
they leave the
    // txSize field set to 12.  It and the txFont field should always be
left set to 0;
    // this makes them the system font and size.  So first thing we do is
clean up
    // after these miscreants.

    GetWMgrPort(&wMgrPort);
    SetPort(wMgrPort);
    TextFont(0);
    TextSize(0);

    // Now, if we're going to change the font or the size, then we
remember the old
    // system font and size so they can be restored later

    if (font != 0 || size != 0)
    {   saveSysFontFam = *(short*)SysFontFam;
        saveSysFontSize = *(short*)SysFontSize;

        if (font != 0)
            *(short*)SysFontFam = font;
        if (size != 0)
            *(short*)SysFontSize = size;

        *(long*)LastSPExtra = -1;           // This forces the system to
recognize our changes
    }

    // Now we need to set the width.  Unfortunately, it's not as easy as
changing the
    // font and size; we need to patch CalcMenuSize so we can fake the
width of the menu
    if (width != 0)
    {   oldCalcMenuSize = (CalcMenuSizeProc)NGetTrapAddress(0x148,ToolTrap);

        gCMSParms.oldCMS = oldCalcMenuSize;
        gCMSParms.theMenu = menu;
        gCMSParms.width = width;

        NSetTrapAddress((long)CalcMenuSizePatch,0x148,ToolTrap);

        // Now we should zero out the menu's stored width and height to
ensure that
        // its size will be recalculated

        (**menu).menuWidth = 0;
        (**menu).menuHeight = 0;
    }

    menuChoice = PopUpMenuSelect(menu,top,left,popUpItem);

    // Now we restore the original CalcMenuSize if we changed it
    if (width != 0)
        NSetTrapAddress((long)oldCalcMenuSize,0x148,ToolTrap);

    // Now we restore the original font and size to their places
    // if we changed them
    if (font != 0 || size != 0)
    {   *(short*)SysFontFam = saveSysFontFam;
        *(short*)SysFontSize = saveSysFontSize;
        *(long*)LastSPExtra = -1;
    }

    // return the chosen menu & item
    return menuChoice;
}

// This is the patch to CalcMenuSize; first, we call the original
// to calculate the size.  If the menu being calculated is ours,
// we then change the width to our stored width.  Note that this
// is a tail patch; while still frowned upon, these are not as
// bad as they used to be, and it's really the best way to
// accomplish our task.
//
// Our use of a global to store the original CalcMenuSize and
// other data implies we'll be sure that A5 hasn't been changed;
// we can be fairly certain this is true in this particular
// case.

static pascal void
CalcMenuSizePatch(MenuHandle theMenu)
{
    gCMSParms.oldCMS(theMenu);          // Call the original CalcMenuSize();

    if (theMenu == gCMSParms.theMenu)   // If it's our menu, fix up the width
        (**theMenu).menuWidth = gCMSParms.width;
}

--
Tim Dierks - Software Haruspex - [e-mail]
If you can't lick 'em, stick 'em on with a big piece of tape. - Negativland

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

>From Richard Wesley <[e-mail]>
Date: 4 Oct 1995 14:42:41 GMT
Organization: Punch Deck Consulting

[e-mail] (Ben Hellstrom) wrote:
>Many of the utilities I have have nice little popups in 9-pt geneva.  They
>dont have MDEFs (which means the MDEFs are either embedded in other code
>resources or they are using the standard MDEF).
>
>Can anyone give me a clue as to how to produce a menu in 9-point geneva?
>Do I have to switch the system font?  If so, how?

If you are using PowerPlant and Constructor, just set the text traits of
the popup control to geneva-9.

If you are doing this from scratch, then what you need to do is:

1) make the current port the window that you want to add the control to;
2) set the font to geneva-9
3) make sure that the controlKind has kControlUsesOwningWindowsFontVariant;
4) instantiate the control.

No trap patching should be required.

- rmgw

http://www.punchdeck.com/hawkfish/PunchDeck.html

- --------------------------------------------------------------------------
Richard Wesley  [e-mail] | "What was that popping sound?"
Punch Deck Consulting [e-mail] | "A paradigm shifting without a
     Macintosh Software Development    |  clutch."  - Dilbert (Scott Adams)
- --------------------------------------------------------------------------



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

>From [e-mail] (Jeremy Roussak)
Date: 4 Oct 1995 19:45:07 GMT
Organization: Queen Mary & Westfield College, London, England

In article <[e-mail]>
[e-mail] (Ben Hellstrom) writes:

> Many of the utilities I have have nice little popups in 9-pt geneva.  They
> dont have MDEFs (which means the MDEFs are either embedded in other code
> resources or they are using the standard MDEF).
>
> Can anyone give me a clue as to how to produce a menu in 9-point geneva?
> Do I have to switch the system font?  If so, how?

Use the Popup CDEF with the "use window font" bit set and set the font
to 9 point Geneva.

Jeremy

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

>From [e-mail] (Joseph Bruni)
Date: Wed, 04 Oct 1995 23:21:58 -0700
Organization: Primenet

In article <[e-mail]>,
[e-mail] (Ben Hellstrom) wrote:

> Many of the utilities I have have nice little popups in 9-pt geneva.  They
> dont have MDEFs (which means the MDEFs are either embedded in other code
> resources or they are using the standard MDEF).
>
> Can anyone give me a clue as to how to produce a menu in 9-point geneva?
> Do I have to switch the system font?  If so, how?

Use the control manager to do the popup for you. Just make sure you use
the "useWFont" variant.

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

>From <[e-mail]>
Date: 6 Oct 1995 07:42:08 GMT
Organization: Spirit Networks

[e-mail] (Ben Hellstrom) wrote:

>Many of the utilities I have have nice little popups in 9-pt geneva.  >They dont have MDEFs (which means the MDEFs are either embed=
ded in
>other code resources or they are using the standard MDEF).
>
>Can anyone give me a clue as to how to produce a menu in 9-point
>geneva? Do I have to switch the system font?  If so, how?

Ben,

As you know, to create a control that uses the standard pop-up control
definition function, you specify the popupMenuProc constant in the
procID field of the resource description for the control.

If you add the constant useWFont to popupMenuProc, the pop-up title
and menu item text will be drawn in the graphics port's current font
rather than the system font.  So all you have to do then is set your
graphics port font and size to Geneva 9 point and you will
get what you want.

K. J. Bricknell
Canberra
Australia


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

>From [e-mail] (Ken Nicolson)
Date: Fri, 13 Oct 1995 10:22:33 GMT
Organization: Office Workstations Limited

[e-mail] (Ben Hellstrom) wrote:

>Many of the utilities I have have nice little popups in 9-pt geneva.  They
>dont have MDEFs (which means the MDEFs are either embedded in other code
>resources or they are using the standard MDEF).

>Can anyone give me a clue as to how to produce a menu in 9-point geneva?
>Do I have to switch the system font?  If so, how?

Here's the solution I came up with. Pass in the menu handle, position
and font number and size, and get returned the string of the menu
option, or -1 if no selection.

I assume Color QuickDraw, but it seems to work even if I switch my
monitor to monochrome.

The code was written after reading the TechNote TB-550, Menu Managers
Q&A. Apparently, this is how the Popup Control gadget does its stuff.

Maybe a routine like this should go into a FAQ somewhere, as it seems
a common thing to want to do, but how to do it is deeply buried on the
developer CDs.

int PopUpMenuSelectWithFont(MenuHandle hMenu,
                            int x, int y,
                            short txFont, short txSize,
                            char *szValue)
{
        SInt32          lPopUpResult;
        int             nPos;
        SInt16          saveFamily;
        SInt16          saveSize;
        SInt32          saveLastSP;
        CGrafPtr        wmPort;
        GrafPtr currentPort;

        // Save current viewport
        GetPort( ¤tPort );

        // Save current system default fonts
        saveFamily = LMGetSysFontFam();
        saveSize   = LMGetSysFontSize();
        saveLastSP = LMGetLastSPExtra();

        GetCWMgrPort( &wmPort );

        SetPort( (GrafPtr) wmPort );

        TextSize( txSize );

        // Install temporary font as system font
        LMSetSysFontFam( txFont );
        LMSetSysFontSize( txSize );
        LMSetLastSPExtra( (SInt32) -1 );

        // Process the pop-up menu
        lPopUpResult = PopUpMenuSelect( hMenu, y, x, 1 );

        // Restore port
        SetPort( currentPort );

        // Restore system default fonts
        LMSetSysFontFam( saveFamily );
        LMSetSysFontSize( saveSize );
        LMSetLastSPExtra( saveLastSP );

        if (HIWORD(lPopUpResult) != 0)
        {
                nPos = LOWORD( lPopUpResult );
                getitem( hMenu, nPos, pszValue );
        }
        else
                nPos = -1;

        return nPos;
}

The LMGet/LMSet functions are from CodeWarrior's LowMem.h to hide the
access to system variables.

Hope the types are correct - we have a set of #defines to hide a lot
of Mac specifics, so I've used the base definitions from memory.

Ken


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

>From [e-mail] (David Fagan)
Subject: Q: using OSAExecute to singlestep AppleScript
Date: Tue, 10 Oct 1995 01:35:41 -0400
Organization: BBN Educational Technologies

My application would like to execute one line of AppleScript at a time.
The lines originate as text, so I call OSACompileExecute. Unfortunately,
the AppleScript interpreter wants entire scripts or handlers at a time, so
if I pass a line such as,

tell application "Scriptable Text Editor"

I get back the error you'd expect if you tried to run the above out of the
script editor, namely, "Expecting end tell..."

Anyone have an idea to pass the interpreter one line at a time?

I suppose it has something to do with the contextID parameter to
OSACompileExecute, but nothing I've done with it helps, and IM Scripting
doesn't provide any clues that I can see.

Thanks

David Fagan
[e-mail]

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

>From [e-mail] (Jon Pugh)
Date: Fri, 13 Oct 1995 22:15:04 GMT
Organization: Will hack for food

David Fagan ([e-mail]) wrote:
> My application would like to execute one line of AppleScript at a time.
> The lines originate as text, so I call OSACompileExecute. Unfortunately,
> the AppleScript interpreter wants entire scripts or handlers at a time, so
> if I pass a line such as,

> tell application "Scriptable Text Editor"

> I get back the error you'd expect if you tried to run the above out of the
> script editor, namely, "Expecting end tell..."

> Anyone have an idea to pass the interpreter one line at a time?

You can't do this.  There is no AppleScript support for single stepping.

I spoke with the DTS folks helping you and we discussed a possible solution,
but not knowing what you are trying to do makes us unable to determine if
this will work for what you are trying to do.

Basically, you can do what the script debugging editors do.  That's to use
the OSASendProc to pause before and/or after each Apple event.  Given this
procedure, which is simply a wrapper for AESend which AppleScript calls
instead of calling AESend, you can decide when and/or if to actually send
the events.  As I stated before though, not knowing what you are trying
to accomplish means that I don't know if this will suffice.

Jon


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

>From [e-mail] (Ed Samuels)
Subject: Reading Notification Manager queue?
Date: 9 Oct 1995 22:54:23 -0400
Organization: The Pipeline

Is there any way to look into the queue of the Notification Manager? I
would like to be able to take a notification request that is generated by
another application off of the queue. IM only describes the NMInstall and
NMRemove functions, both of which require that you already have a pointer
to the notification request. It makes no reference to any other NM
routines, and it seems to imply that there are no others!

    Most of the other OS routines have some function that returns a pointer
to the bottom of the queue, but apparently the Notification Manager
doesn't. I can't even find a low-memory global that points to it! If there
isn't an easier way, I think I'll have to patch the routine to find out
what is going on. That is something that I DO NOT want to do--NMInstall
preserves most registers and does not move memory, so such a patch would be
very shaky, and definitely much too complicated for my purposes.

    If anyone can direct me to any info about the Notification Manager that
describes more than what is said in Inside Macintosh, thanks. :)
___________________________________
Richard Samuels
[e-mail]

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

>From [e-mail] (Brian Stern)
Date: 14 Oct 1995 03:37:16 GMT
Organization: The University of Texas at Austin, Austin, Texas

Richard,

The OS queue used by the Notification Manager is undocumented.  An earlier
version of Macsbug used to know about the lowmem global, but the current
version seems to have forgotten its name :-/  A quick look at NMInstall
shows a suspicious check to location 0x0DD5.

I have an INIT that patches NMInstall and keeps track of the records that
are posted.  As I was developing this I first walked the queue but in the
end decided that since it was undocumented I would do it another way.  You
can be sure things won't change before Copland anyway.  Not all requests
that are added to the queue are removed by the Notification manager.
There is a flags field in each NMRecord and apparently once a Notification
is shown a bit is set so that the notification isn't shown again.

Good luck,

Brian  }:-{)}

In article <45cn8v$[e-mail]>, [e-mail] (Ed
Samuels) wrote:

<Is there any way to look into the queue of the Notification Manager? I
<would like to be able to take a notification request that is generated by
<another application off of the queue. IM only describes the NMInstall and
<NMRemove functions, both of which require that you already have a pointer
<to the notification request. It makes no reference to any other NM
<routines, and it seems to imply that there are no others!
<
<    Most of the other OS routines have some function that returns a pointer
<to the bottom of the queue, but apparently the Notification Manager
<doesn't. I can't even find a low-memory global that points to it! If there
<isn't an easier way, I think I'll have to patch the routine to find out
<what is going on. That is something that I DO NOT want to do--NMInstall
<preserves most registers and does not move memory, so such a patch would be
<very shaky, and definitely much too complicated for my purposes.
<
<    If anyone can direct me to any info about the Notification Manager that
<describes more than what is said in Inside Macintosh, thanks. :)
<___________________________________
<Richard Samuels
<[e-mail]

____________________________________________________________________
Brian  Stern  {;-{)}                  Toolbox commando and Menu bard                  [e-mail]                          [e-mail]
INIT Writing FAQ etc. at <ftp://ftp.pbcomputing.com//Guests/BrianS/>

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

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