MacTech 1 to 12/MacTech-vol-1-12.toast/Reference/the cmsp digests ('94-'97)/csmp digest Vol 4 No 015

Text File | 1996-10-27 | 138KB | 3,813 lines

C.S.M.P. Digest             Fri, 25 Oct 96       Volume 4 : Issue 15

Today's Topics:

        Access fault exception
        Angle brackets as command keys?
        Another weird sound problem
        Converting file handles
        Dialogs and TE commands - bus error?
        Displaying Status of Background Task
        Help! How do a draw a picture offscreen?
        Icons on the desktop?
        Is it MacOS 4.1 or System 4.1?
        Localizing application on the Mac
        Looking for Apple Remote Access API
        Low memory situations & Toolbox
        MacOS Java problems (Was: [Java]:  Refresh (paint()) in MW Java v. Netscape)
        Pascal compiler??
        Sound Manager delays!?
        Strange Pop-Up Problem...
        VIP-C and CodeWarrior 9
        Where is TMON Pro?
        [ANN] CodeWarrior IDE 1.7.1  is available for downloading
        [Q] Reliable notification from completion routine
        is MacOS multiplatform?
        parent class method called instead of overloaded one!



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

The digest is a collection of article threads from the internet
newsgroups comp.sys.mac.programmer.help, csmp.tools, csmp.misc and
csmp.games. 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
ee.mcgill.ca).  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 digests can be obtained by email, ftp or through the World Wide Web.

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                      Adds you to the mailing list
    unsubscribe csmp                    Removes you from the list

Once you have subscribed, you will automatically receive each new
issue as it is created.

Back issues are available by ftp from Info-Mac mirror sites in the
per/csmp subdirectory, e.g.

  ftp://sumex-aim.stanford.edu/info-mac/per/csmp/

The contents of all back issues can be searched by accessing the
following URL, courtesy of Andrew Barry ([e-mail]):

    http://marvin.stattech.com.au/search.html

They can also be searched through the following URLs, thanks to
Tim Tuck ([e-mail]):

    http://wais.sensei.com.au/searchform.html
    wais://wais.sensei.com.au:210/csmp?

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

From [e-mail] (Sean Crist)
Subject: Access fault exception
Date: 9 Oct 1996 23:52:50 GMT
Organization: University of Pennsylvania

I'm just learning C and the CodeWarrior environment after years of Think
Pascal.  Can someone knowledgeable with CW please help me with this problem?

While using the debugger to step through my code, I get an alert "Access
fault exception".  I searched several of the online manuals and could find
no explanation of this message.  Anyway, after I dismiss the alert, the
debugger resets the execution of the program.  Then, shortly after, the
machine crashes and I have to reboot (cursor freezes, sometimes with a few
prelimary jumps into Macsbug).

This first happened deep in my code at a call to GetNewDialog(), so I tried
putting an Alert() as the very first thing my program does; Alert() causes
the same problem.  The problem is 100% replicable.

What does this error mean, and what is the fix?

Thank you.

  \/ __ __    _\_     --Sean Crist  ([e-mail])
 ---  |  |    \ /     For a free copy of the Bill of Rights, finger
  _| ,| ,|   -----    this account.  It's also available through
  _| ,| ,|    [_]     my homepage:
   |  |  |    [_]     http://www.ling.upenn.edu/~kurisuto/


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

From Patrick Bakx <[e-mail]>
Date: Thu, 10 Oct 1996 08:53:12 +0200
Organization: Tractebel

Hi Sean,
Are you initializing the Mac toolbox ? At the start of any Mac program
you must initialize the various managers. Any Mac example program will
show you how to do this.

Cheers,
Patrick

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

From [e-mail] (Wiesmann)
Date: 12 Oct 1996 09:18:16 GMT
Organization: University of Karlsruhe, Germany

You are using non-valid pointers, you try to access memory somewhere in nirwana, etc.

Check your pointers!! May be that a function returned NULL instead of a valid address.

Hartwig Wiesmann



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

From [e-mail] (Matthew Mucker)
Subject: Angle brackets as command keys?
Date: Wed, 25 Sep 1996 00:43:34 +0500
Organization: Internet America

I would like to use the greater than and less than symbols as command key
equivalents for 'next' and 'previous' in my app.

It, uh, doesn't work.  Anybody ever tried this?  I know that it means
sacrificing the command-period for cancel, but that's okay in my app.

Thanks to all,

   -Matt

--
Life is 10% what happens to you, and 90% how you respond to what happens to you.

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

From [e-mail] (EANewcombe)
Date: 28 Sep 1996 06:41:44 -0400
Organization: America Online, Inc. (1-800-827-6364)

In "Angle brackets as command keys?" [e-mail]  asks:

I would like to use the greater than and less than symbols as command key
equivalents for 'next' and 'previous' in my app.

It, uh, doesn't work.  Anybody ever tried this?  I know that it means
sacrificing the command-period for cancel, but that's okay in my app.

Thanks to all,

   -Matt

The cause of your problem is that the shift key seems to
be ignored when the command key is pressed, command >
becomes command period and command < becomes command comma. One way around
this is to check the
key pressed  and converting it before calling MenuKey.

example:
key = (char) ( event.message & charCodeMask );
if (key == '.')    // ***
    key = '>';      // ***
if (key == ',')   // ***
    key = '<';     // ***
result = MenuKey( key );

where event is the EventRecord returned from
Get/WaitNextEvent.

Of course now you've really sacrificed command period
but your user might prefer this (2 keys to press
simultaneously rather than 3). You have a problem
if the user's keyboard does not happen to have >
above period and < above comma so this is not
a global solution.

hope this helps, Alex

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

From [e-mail] (Matthew Mucker)
Date: Sat, 05 Oct 1996 22:56:30 +0500
Organization: Internet America

In article <52ivd8$[e-mail]>, [e-mail]
(EANewcombe) wrote:

> I would like to use the greater than and less than symbols as command key
> equivalents for 'next' and 'previous' in my app.

> The cause of your problem is that the shift key seems to
> be ignored when the command key is pressed, command >
> becomes command period and command < becomes command comma. One way around
> this is to check the
> key pressed  and converting it before calling MenuKey.
>
>
> Of course now you've really sacrificed command period
> but your user might prefer this (2 keys to press
> simultaneously rather than 3). You have a problem
> if the user's keyboard does not happen to have >
> above period and < above comma so this is not
> a global solution.

That last point is a good one.  I don't like the idea of displaying < and
> as keyboard equivalents on the menu, and then checking for comma and
period internally.  But that seems to be the only solution to the
problem.  Your last point is a good one, though.  I'll have to ponder that
one...

Thanks for the input.

-Matt

--
Life is 10% what happens to you, and 90% how you respond to what happens to you.

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

From Maf Vosburgh <[e-mail]>
Date: 11 Oct 1996 15:10:02 GMT
Organization: MultiMedia Corporation

In article <[e-mail]> Matthew
Mucker, [e-mail] writes:
>That last point is a good one.  I don't like the idea of displaying < and
>> as keyboard equivalents on the menu, and then checking for comma and
>period internally.  But that seems to be the only solution to the
>problem.  Your last point is a good one, though.  I'll have to ponder that
>one...
>

What I did was store "." and "," in the menu resource, and then I wrote a
custom MDEF that looks for those chars and draws them as '<' and '>'.
This technique is also handly for command +.


Maf


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

From [e-mail] (John Stiles)
Subject: Another weird sound problem
Date: Sat, 05 Oct 1996 23:50:43 -0800
Organization: Data Communication and Newtorking Services

OK, thanks to the numerous replies of the members of c.s.m.p.g, I've got
my sound code working great!

Almost. It crashes PowerBook 5300's consistently. For some reason, it
works on all desktop Power Macs but I've had lots of users saying their
5300 dies instantly when they open up the application.

Here's my current sound code. Much abbreviated of course, because the
algorithm isn't important in this case (is it?). FWIW, this code worked
when I used the SndPlayDoubleBuffer technique, but I don't know if it
worked quite as fluidly.

void InitSound( void )
{
   SndCommand command;
   SndCallBackUPP callBackUPP = NewSndCallBackProc( SoundCallBack );

   SndNewChannel( &sndChannel, sampledSynth, initMono, callBackUPP );

   sndHeader = (SoundHeaderPtr) NewPtrClear( sizeof( SoundHeader ) +
SOUND_BUFSIZE );
   sndHeader->samplePtr    = nil;
   sndHeader->length       = SOUND_BUFSIZE;
   sndHeader->sampleRate      = SoundRate << 16;
   sndHeader->baseFrequency   = kMiddleC;

   command.cmd    = callBackCmd;
   command.param1 = 0;
   command.param2 = (long)LMGetCurrentA5();

   SoundCallBack( sndChannel, &command ); // start the ball rolling
}

pascal void SoundCallBack( SndChannelPtr chan, SndCommand *command )
{
   int i;
   SndCommand myCommand;
   long myA5 = SetA5( command->param2 ); // set up A5 world

   if( soundIsOn )
   {
     // sndHeader->sampleArea[i] = stuff;
     // fill up the sound buffer here
   }
    else
    {
      memset( sndHeader->sampleArea, 0x80, SOUND_BUFSIZE );
    }

    sndHeader->sampleRate = SoundRate << 16; // <-- the soundrate can
change dynamically

   myCommand.cmd = bufferCmd;
   myCommand.param1 = 0;
   myCommand.param2 = (long) sndHeader;
   SndDoImmediate( chan, &myCommand );  // buffer the sound
   SndDoCommand( chan, command, true ); // callback again

    SetA5( myA5 ); // close up A5 world
}

Note the SndDoCommand a few lines up. It is calling itself back once
again. Therefore the sound should loop continuously.

Any ideas, folks? Did I make any obviously dumb blunders?

*Stiles

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

From Jackson Software Development <[e-mail]>
Date: 8 Oct 1996 14:20:25 GMT
Organization: Jackson Software

In article <[e-mail]> John
Stiles, [e-mail] writes:
>Note the SndDoCommand a few lines up. It is calling itself back once
>again. Therefore the sound should loop continuously.

I had a similar problem that only showed up when Quicktime was installed.
I fixed it by not actually calling SndDoCommand in the callback routine,
but
instead setting a variable in the callback and then checking the variable
in
my event loop and restarting the sound there. As long as the loop was
greater than about 50 milliseconds it worked without clicking.

I have read that SndDoCommand can be called at interrupt time depending
on the command, and that the bufferCmd was one of those commands, but
it didn't work for me when Quicktime was installed.

Darrin

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

From [e-mail] (Norman Franke)
Date: Tue, 08 Oct 1996 10:22:41 -0700
Organization: Lawrence Livermore National Laboratory

In article <[e-mail]>,
[e-mail] (John Stiles) wrote:

> Almost. It crashes PowerBook 5300's consistently. For some reason, it
> works on all desktop Power Macs but I've had lots of users saying their
> 5300 dies instantly when they open up the application.
> pascal void SoundCallBack( SndChannelPtr chan, SndCommand *command )

> {
>     sndHeader->sampleRate = SoundRate << 16; // <-- the soundrate can
> change dynamically
> }

I think this is your problem. You can only call SndDoImmediate or
SndDoCommand with commands that do not move memory. Changing the sampling
rate can cause memory movement and allocation as it opens components to
perform sample rate conversion. If you want to change any parameters, I'd
recommend some other method.

--
Norman Franke
[e-mail]

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

From [e-mail] (Jim Reekes)
Date: Tue, 08 Oct 1996 19:30:57 -0700
Organization: Apple Computer, Inc.

In article <[e-mail]>,
[e-mail] (John Stiles) wrote:

> In article <53dnv9$[e-mail]>, Jackson Software Development
> <[e-mail]> wrote:
>
> > In article <[e-mail]> John
> > Stiles, [e-mail] writes:
> > >Note the SndDoCommand a few lines up. It is calling itself back once
> > >again. Therefore the sound should loop continuously.
> >
> > I had a similar problem that only showed up when Quicktime was installed.
> > I fixed it by not actually calling SndDoCommand in the callback routine,
> > but
> > instead setting a variable in the callback and then checking the variable
> > in
> > my event loop and restarting the sound there. As long as the loop was
> > greater than about 50 milliseconds it worked without clicking.
> >
> > I have read that SndDoCommand can be called at interrupt time depending
> > on the command, and that the bufferCmd was one of those commands, but
> > it didn't work for me when Quicktime was installed.
>
> Oh... great!! This is news to me. Oh well. I am not sure how well this
> technique will work in my code--I am currently generating about 10ms of
> sound at a time--but I can hack it out.

I use the same basic code as you. Mine has never crashed. QuickTime uses
the same idea as well. It doesn't crash either. The only thing I noticed
was that you never initialized your sndChannel to nil (and don't allocate
your own).

> Thanks. I guess THIS is why SndPlayDoubleBuffer is important...!

No, SndPlayDoubleBuffer() is only important for one reason. It was
designed for SndStartFilePlay(). Other than that, the bufferCmd and
callBackCmd sequence are the same. The later provides more control and
flexibility. Both internally use the same code to play the sou
+++++++++++++++++++++++++++

From [e-mail] (Jim Reekes)
Date: Tue, 08 Oct 1996 19:30:57 -0700
Organization: Apple Computer, Inc.

In article <[e-mail]>,
[e-mail] (John Stiles) wrote:

> In article <53dnv9$[e-mail]>, Jackson Software Development
> <[e-mail]> wrote:
>
> > In article <[e-mail]> John
> > Stiles, [e-mail] writes:
> > >Note the SndDoCommand a few lines up. It is calling itself back once
> > >again. Therefore the sound should loop continuously.
> >
> > I had a similar problem that only showed up when Quicktime was installed.
> > I fixed it by not actually calling SndDoCommand in the callback routine,
> > but
> > instead setting a variable in the callback and then checking the variable
> > in
> > my event loop and restarting the sound there. As long as the loop was
> > greater than about 50 milliseconds it worked without clicking.
> >
> > I have read that SndDoCommand can be called at interrupt time depending
> > on the command, and that the bufferCmd was one of those commands, but
> > it didn't work for me when Quicktime was installed.
>
> Oh... great!! This is news to me. Oh well. I am not sure how well this
> technique will work in my code--I am currently generating about 10ms of
> sound at a time--but I can hack it out.

I use the same basic code as you. Mine has never crashed. QuickTime uses
the same idea as well. It doesn't crash either. The only thing I noticed
was that you never initialized your sndChannel to nil (and don't allocate
your own).

> Thanks. I guess THIS is why SndPlayDoubleBuffer is important...!

No, SndPlayDoubleBuffer() is only important for one reason. It was
designed for SndStartFilePlay(). Other than that, the bufferCmd and
callBackCmd sequence are the same. The later provides more control and
flexibility. Both internally use the same code to play the sound, and
therefore neither are more effecient.

--
Jim Reekes, Polterzeitgeist |       QuickTime Products R&D
                            |        Sound Manager Expert
Apple Computer, Inc.        | "All opinions expressed are mine, and
2 Infinite Loop  MS 302-3KS |  do not necessarily represent those
Cupertino, CA 95014         |  of my employer, Apple Computer Inc."

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

From [e-mail] (John Stiles)
Date: Tue, 08 Oct 1996 16:19:44 -0800
Organization: University of California, Berkeley

In article <53dnv9$[e-mail]>, Jackson Software Development
<[e-mail]> wrote:

> In article <[e-mail]> John
> Stiles, [e-mail] writes:
> >Note the SndDoCommand a few lines up. It is calling itself back once
> >again. Therefore the sound should loop continuously.
>
> I had a similar problem that only showed up when Quicktime was installed.
> I fixed it by not actually calling SndDoCommand in the callback routine,
> but
> instead setting a variable in the callback and then checking the variable
> in
> my event loop and restarting the sound there. As long as the loop was
> greater than about 50 milliseconds it worked without clicking.
>
> I have read that SndDoCommand can be called at interrupt time depending
> on the command, and that the bufferCmd was one of those commands, but
> it didn't work for me when Quicktime was installed.

Oh... great!! This is news to me. Oh well. I am not sure how well this
technique will work in my code--I am currently generating about 10ms of
sound at a time--but I can hack it out.

Thanks. I guess THIS is why SndPlayDoubleBuffer is important...!

*Stiles
.
.
.
.
.
.
.
.
.
.
.
.
...... trying to confound his stupid NNTP server

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

From [e-mail] (John Stiles)
Date: Tue, 08 Oct 1996 21:38:12 -0800
Organization: University of California, Berkeley

First off... wow!! Jim Reekes! I am talking to God. Jim, the new Sound
Manager is awesome, just so you know. Never thought I'd get a reply from
the author of the Sound Manager himself...!

> > Oh... great!! This is news to me. Oh well. I am not sure how well this
> > technique will work in my code--I am currently generating about 10ms of
> > sound at a time--but I can hack it out.
> I use the same basic code as you. Mine has never crashed. QuickTime uses
> the same idea as well. It doesn't crash either. The only thing I noticed
> was that you never initialized your sndChannel to nil (and don't allocate
> your own).

Actually, I was doing this when I was defining the sndChannel
(SndChannelPtr sndChannel = nil;). But that didn't seem to be the cause at
all.

One thing I noticed, but I am unsure if it is the problem, was that I was
defining the callback function as:
pascal void SoundCallBack( SndChannelPtr chan, SndCommand *command )

Shouldn't it be:
pascal void SoundCallBack( SndChannelPtr chan, SndCommand command )
...?

However, I did rewrite it this way, but it didn't seem to have any major
effect on the program on my Mac at all. This might fix the PowerBook
problem, but I'm not sure about that, because I don't have a test
PowerBook to work with. (Isn't that just a pain?!)

> > Thanks. I guess THIS is why SndPlayDoubleBuffer is important...!
> No, SndPlayDoubleBuffer() is only important for one reason. It was
> designed for SndStartFilePlay(). Other than that, the bufferCmd and
> callBackCmd sequence are the same. The later provides more control and
> flexibility. Both internally use the same code to play the sound, and
> therefore neither are more effecient.

Oops. Guess I did a little extra work then, because I just re-implemented
my old SndPlayDoubleBuffer solution. I mean, it worked, right? And now, to
handle a possible change in SoundRate, the code simply disposes the sound
channel whenever the rate changes and makes a new one. Simple, and
inefficient, but the SoundRate will never change unless the user opens up
a customization dialog anyway (thereby doing something like
SndDisposeChannel/SndNewChannel won't cause a speed hit).

Say, though, I have one big Sound Manager question for you. If I turn on
Virtual Memory, the sound code I have listed above starts to sound awful!
The sound latency is quite unacceptable. It is as if it were buffering
several times in a row rather than just once. You see, the code is
actually a square wave generator, and the frequency and volume variables
change over time. But the changes are choppy and rough when Virtual Memory
is on, yet clean and smooth when VM is off. I have done a million
different things to avoid it, so I have to conclude that it's a Sound
Manager bug (or hidden feature). Help!

(As a side note, the entire reason I ever bothered to implement this
gosh-darn callback code which I just ripped out was to try to avoid this
bug/hidden feature, but it didn't fix the problem at all!)

*Stiles

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

From [e-mail] (Jim Reekes)
Date: Wed, 09 Oct 1996 16:19:27 -0700
Organization: Apple Computer, Inc.

In article <[e-mail]>,
[e-mail] (John Stiles) wrote:

> Say, though, I have one big Sound Manager question for you. If I turn on
> Virtual Memory, the sound code I have listed above starts to sound awful!
> The sound latency is quite unacceptable. It is as if it were buffering
> several times in a row rather than just once. You see, the code is
> actually a square wave generator, and the frequency and volume variables
> change over time. But the changes are choppy and rough when Virtual Memory
> is on, yet clean and smooth when VM is off. I have done a million
> different things to avoid it, so I have to conclude that it's a Sound
> Manager bug (or hidden feature). Help!

Uh huh. So don't turn on VM. It's not a Sound Mgr bug. It's just a
conflict of interest between real time services and VM.

> (As a side note, the entire reason I ever bothered to implement this
> gosh-darn callback code which I just ripped out was to try to avoid this
> bug/hidden feature, but it didn't fix the problem at all!)

Yeah, life as a programmer!

Jim

--
Jim Reekes, Polterzeitgeist |       QuickTime Products R&D
                            |        Sound Manager Expert
Apple Computer, Inc.        | "All opinions expressed are mine, and
2 Infinite Loop  MS 302-3KS |  do not necessarily represent those
Cupertino, CA 95014         |  of my employer, Apple Computer Inc."

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

From [e-mail] (Jim Reekes)
Date: Thu, 10 Oct 1996 15:50:44 -0700
Organization: Apple Computer, Inc.

In article <[e-mail]>,
[e-mail] (John Stiles) wrote:

> In article <[e-mail]>, [e-mail]
> (Jim Reekes) wrote:
>
> > Uh huh. So don't turn on VM. It's not a Sound Mgr bug. It's just a
> > conflict of interest between real time services and VM.
>
> Oh, come on, great answer!! I don't use VM as a rule--32MB is quite
> enough, thanks--but I have many users that do use VM. There's got to be a
> way to make it smoother than this! And if you try to tell me that's a
> feature, I'll laugh *really* hard.
>
> True, I know, it's hard to call a routine 100 times per second when you're
> paging in and out from the hard disk. But I don't think that's an
> acceptable excuse for this Sound Manager behaviour. And FWIW, I can run
> the program with 33MB of virtual memory (on a 32MB system) with system
> memory not even 50% full, and the same problem surfaces.
>
> I'd really like to be able to run my system at a 33MB virtual memory
> setting, actually, so I can take advantage of the reduced application
> footprint this provides. But this is the main reason that I can't!

I'm not going into this one, again. I'll give you this simple clue. You
cannot page during a sound hardware interrupt. Therefore I must defer the
actual function until a time that paging is safe, using the call
DeferUserFn() to VM. VM then owns this function and will call it when it
decides it's safe. This happens at some time later, when all SCSI activity
is clear. At this point in time the hardware interrupt handler calls into
the Sound Mgr for more data to satisfy the hardware request. It's exactly
like I had said, "It's not a Sound Mgr bug. It's just a conflict of
interest between real time services and VM." If you'll have noticed, you
cannot use QuickTime and VM either. For one thing, the SCSI bus is really
busy while trying to play a movie and therefore sound is held off for
longer periods of time.

Jim

--
Jim Reekes, Polterzeitgeist |       QuickTime Products R&D
                            |        Sound Manager Expert
Apple Computer, Inc.        | "All opinions expressed are mine, and
2 Infinite Loop  MS 302-3KS |  do not necessarily represent those
Cupertino, CA 95014         |  of my employer, Apple Computer Inc."

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

From [e-mail] (Al Evans)
Date: 11 Oct 1996 07:42:19 -0500
Organization: Powertools, Austin, Texas

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

>It's exactly
>like I had said, "It's not a Sound Mgr bug. It's just a conflict of
>interest between real time services and VM."

No, Jim. It's a conflict of interest between real time services and VM
*IN THE MACOS*. And the same kind of comment goes for your response in
another thread, which was along the lines of "your problem is caused
by the sound-hardware buffer size, about which I can do nothing."

A counter-example: I can crank up enough movies on my BeBox to start
VM thrashing. Within a couple of seconds, it will settle down to play
them ALL, with reasonable performance, hitting the disk continuously.

The conflict of interest is not inherent, it is due to the failure of
Apple to redesign their operating system. And -- perhaps unfortunately --
since you're a well-known Apple representative, your implication that
these deficiencies are just "facts of life" will be widely perceived
as indicating Apple's continued unwillingness or inability to do
anything about them.

It seems to me that you guys should at least cop a line from Metrowerks:
"Apple Computer apologizes for the inconvenience."

Not quite ready to start expecting less from Apple....

                    --Al Evans--
--
|||| Al Evans -- [e-mail] -- proud of Graphic Elements Release 3 ||||
====  A new standard for high-performance interactive Macintosh graphics  ====
====         Available from mac.archive.umich.edu and mirrors in          ====
||||        /mac/development/libraries/graphicelements3.0.sit.hqx         ||||

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

From [e-mail] (John Stiles)
Date: Wed, 09 Oct 1996 22:23:26 -0800
Organization: University of California, Berkeley

In article <[e-mail]>, [e-mail]
(Jim Reekes) wrote:

> In article <[e-mail]>,
> [e-mail] (John Stiles) wrote:
>
> > Say, though, I have one big Sound Manager question for you. If I turn on
> > Virtual Memory, the sound code I have listed above starts to sound awful!
> > The sound latency is quite unacceptable. It is as if it were buffering
> > several times in a row rather than just once. You see, the code is
> > actually a square wave generator, and the frequency and volume variables
> > change over time. But the changes are choppy and rough when Virtual Memory
> > is on, yet clean and smooth when VM is off. I have done a million
> > different things to avoid it, so I have to conclude that it's a Sound
> > Manager bug (or hidden feature). Help!
> Uh huh. So don't turn on VM. It's not a Sound Mgr bug. It's just a
> conflict of interest between real time services and VM.

Oh, come on, great answer!! I don't use VM as a rule--32MB is quite
enough, thanks--but I have many users that do use VM. There's got to be a
way to make it smoother than this! And if you try to tell me that's a
feature, I'll laugh *really* hard.

True, I know, it's hard to call a routine 100 times per second when you're
paging in and out from the hard disk. But I don't think that's an
acceptable excuse for this Sound Manager behaviour. And FWIW, I can run
the program with 33MB of virtual memory (on a 32MB system) with system
memory not even 50% full, and the same problem surfaces.

I'd really like to be able to run my system at a 33MB virtual memory
setting, actually, so I can take advantage of the reduced application
footprint this provides. But this is the main reason that I can't!

> > (As a side note, the entire reason I ever bothered to implement this
> > gosh-darn callback code which I just ripped out was to try to avoid this
> > bug/hidden feature, but it didn't fix the problem at all!)
> Yeah, life as a programmer!

OK, OK, I know it sucks. But there must be a way around this blatant flaw,
and if anyone knows it, it's you. So really, what's the secret?!?

*Stiles

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

From [e-mail] (John Stiles)
Date: Thu, 10 Oct 1996 19:27:01 -0800
Organization: University of California, Berkeley

> I'm not going into this one, again. I'll give you this simple clue. You
> cannot page during a sound hardware interrupt. Therefore I must defer the
> actual function until a time that paging is safe, using the call
> DeferUserFn() to VM. VM then owns this function and will call it when it
> decides it's safe. This happens at some time later, when all SCSI activity
> is clear. At this point in time the hardware interrupt handler calls into
> the Sound Mgr for more data to satisfy the hardware request. It's exactly
> like I had said, "It's not a Sound Mgr bug. It's just a conflict of
> interest between real time services and VM." If you'll have noticed, you
> cannot use QuickTime and VM either. For one thing, the SCSI bus is really
> busy while trying to play a movie and therefore sound is held off for
> longer periods of time.

OK, I wasn't fully clear on the issue. I guess nothing can be done without
rewriting Virtual Memory. Well, hopefully this doesn't mean Copland will
maul my sound code. Thanks for replying, at least, and clarifying the
issue.

Now if only there was a way to get VM to call the function more often...
but that's not your department and I know it!...

*Stiles
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..... who hates the campus NNTP server that forces him to waste this space


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

From [e-mail] (Mauricio Tavares)
Subject: Converting file handles
Date: 9 Oct 1996 15:26:21 GMT
Organization: University of Florida

    Let's say that I used the Mac file routines to find where a file
is (drive/directory).  How can I change that into something fread() can
use?

--
 ===========================+==========================================
| Mauricio Tavares          | "We will attack...                       |
| [e-mail] |  ...under the cover of daylight!" Rimmer |
 ===========================+==========================================

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

From [e-mail] (John Stiles)
Date: Thu, 10 Oct 1996 15:27:52 -0800
Organization: University of California, Berkeley

>         Let's say that I used the Mac file routines to find where a file
> is (drive/directory).  How can I change that into something fread() can
> use?

Convert the FSSpec you get to an absolute pathname. I know it's lame, but
I haven't found a better way yet. Here's a routine that has always worked
for that purpose; I didn't write it, but it's useful:

short FullPathname( FSSpec *in, Str255 out )
{
   Str255      dirName;
   DirInfo     dirInfo;
   OSErr    err;

   out[0] = dirName[0] = 0;

   dirInfo.ioNamePtr = dirName;
   dirInfo.ioDrParID = in->parID;

   do
   {
      dirInfo.ioVRefNum = in->vRefNum;
      dirInfo.ioFDirIndex = -1;           // -1 means use ioDrDirIDä
      dirInfo.ioDrDirID = dirInfo.ioDrParID;

      err = PBGetCatInfoSync( (CInfoPBPtr) &dirInfo );

      if( err == noErr )
      {
         dirName[ ++dirName[0] ] = ':';

         if( dirName[0] + out[0] > 255 )
         {
            err = bdNamErr;            // too big to eat!
         }
         else
         {
            LPstrcat( dirName, out );
            LPstrcpy( out, dirName );
         }
      }
   }
   while( (dirInfo.ioDrDirID != fsRtDirID) && (err == noErr) );

   if( dirName[0] + in->name[0] > 255 )
   {
      err = bdNamErr;
   }
   else
   {
      LPstrcat( out, in->name );
   }

   return err;
}

*Stiles

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

From [e-mail] (Michael Schuerig)
Date: Fri, 11 Oct 1996 16:11:38 +0200
Organization: RHRZ - University of Bonn (Germany)

John Stiles <[e-mail]> wrote:

> >         Let's say that I used the Mac file routines to find where a file
> > is (drive/directory).  How can I change that into something fread() can
> > use?
>
> Convert the FSSpec you get to an absolute pathname. I know it's lame, but
> I haven't found a better way yet. Here's a routine that has always worked
> for that purpose; I didn't write it, but it's useful:

DON'T! There is a better way. If you're using CodeWarrior look at

:Metrowerks CodeWarrior:Metrowerks Standard Library:MSL C:MSL
Mac:FSp_fopen:

where you will find the definition for the following function

FILE * FSp_fopen(ConstFSSpecPtr spec, const char * open_mode)

If you're using some other compiler/library use the following function
from the Mac Programming FAQ maintained by Chris K. Thompson.


4.22) Q: How do I use fopen or other ANSI C I/O functions with files the
user
has selected using StandardGetFile?

A: The "standard" ANSI C I/O routines are not well suited for the
Macintosh way of
doing things because they expect either a simple file name or partial
pathname from the
current default volume/directory, or a full pathname. However, if you
are doing a port
for your own enjoyment and benefit (or maybe for in-house work) you can
use the
SetDefault and RestoreDefault functions in MoreFiles (see 4.2) as a
wrapper around
ANSI C I/O routines where the location of the file is implied to be the
default volume
and directory. The following function is an example of how you might
pass an FSSpec to
fopen using the SetDefault and RestoreDefault functions:

FILE *fopen_mac(const FSSpec *spec, char *mode)
{
    FILE *ret = NULL;
    short oldVRefNum;
    long oldDirID;

    /* set the default volume/directory */
    if ( SetDefault(spec->vRefNum, spec->parID, &oldVRefNum, &oldDirID)
== noErr )
    {
        /* call ANSI fopen */
        ret = fopen(fileName, mode);

        /* restore the default volume/directory */
        (void) RestoreDefault(oldVRefNum, oldDirID);
    }
    return ret;
}


Note: In older versions of MPW, this technique didn't work because the
MPW libraries
used GetVol and explicitly used that value as the default location.



Michael

- -
Michael Schuerig
mailto:[e-mail]
http://www.rhrz.uni-bonn.de/~uzs90z/


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

From [e-mail] (Matt Bishop)
Subject: Dialogs and TE commands - bus error?
Date: Wed, 09 Oct 1996 02:48:23 -0500
Organization: Zippo

I am running a modeless dialog and I want to call some TextEdit commands
to do things like set the outline selection flage (TEFeatureFlag) and send
the editText field a keypress (TEKey).  However, I can get none of them to
work.  For example:

GetDialogItem(theDialog, theItem, &iType, &fieldHandle, &iRect);
TEKey (theChar, (TEHandle) fieldHandle);  //bus error here

suggestions?  please email back.

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

From [e-mail] (Dave Overton)
Date: 9 Oct 1996 18:37:45 GMT
Organization: NETCOM Network Operations

Matt,
Your problem here is that the Handle returned from GetDItem is not a text
edit handle. All the text edit items in a dialog share a single text edit
field.
You have to insure that that item is selected ( with the cursor ) and look
in the DialogPeek record to get the text edit handle for the dialog.
Dave Overton




In article <[e-mail]>,
[e-mail] (Matt Bishop) wrote:

> I am running a modeless dialog and I want to call some TextEdit commands
> to do things like set the outline selection flage (TEFeatureFlag) and send
> the editText field a keypress (TEKey).  However, I can get none of them to
> work.  For example:
>
> GetDialogItem(theDialog, theItem, &iType, &fieldHandle, &iRect);
> TEKey (theChar, (TEHandle) fieldHandle);  //bus error here
>
> suggestions?  please email back.


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

From Steve Collins <[e-mail]>
Subject: Displaying Status of Background Task
Date: Mon, 07 Oct 1996 09:56:55 +0000
Organization: Netcom

Hi,

Here's a user interface question:

My "task" (currently implemented as an application) wants to
display connection state information - details not relevant, so I
won't go into them.

Some software vendors (Global Village) add a graphic item to the
menu bar (thermometer type) to accomplish this.  This looks good,
but is not "standard".

Apple uses a control strip module to display timers for MacTCP.

Another approach would be a status window for the app, which I
mention only in passing, I really don't care for it.

I favor a control strip module for my product, but wonder if it's
the right approach.  Is it acceptable to users? Do most users use
it?  Does it have a future?

Any thoughts out there?

Thanks

Steve

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

From [e-mail] (Jochen Lippert)
Date: Mon, 07 Oct 1996 19:32:53 +0200
Organization: University of  Bremen

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

> I favor a control strip module for my product, but wonder if it's
> the right approach.  Is it acceptable to users? Do most users use
> it?  Does it have a future?

You could also use a control panel, just like the "File Sharing Monitor".
I don't know much about this control strip, is it shipping with all macs
today? I heard it was only for Powerbooks, but times change :)

Jochen Lippert

[e-mail]

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

From [e-mail] (Alexander McHugh)
Date: Wed, 09 Oct 1996 13:37:48 +1000
Organization: The Zipsters

In article <[e-mail]>,
[e-mail] (Jochen Lippert) wrote:
:> I favor a control strip module for my product, but wonder if it's
:> the right approach.  Is it acceptable to users? Do most users use
:> it?  Does it have a future?

i don't know how popular they are but every mac user i have shown the
control strip to has fallen in love with it

yet another thing apple had underpromoted.

i would recommend developing for extensions strip as well if you can. It
was the first shareware i ever registered. SO useful. A pity the authour
seems to have dropped interest in it.

:You could also use a control panel, just like the "File Sharing Monitor".
:I don't know much about this control strip, is it shipping with all macs
:today? I heard it was only for Powerbooks, but times change :)

it's for all macs as of system 7.5.3

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

From [e-mail] (John Stiles)
Date: Thu, 10 Oct 1996 15:33:49 -0800
Organization: University of California, Berkeley

In article <[e-mail]>,
[e-mail] (Alexander McHugh) wrote:

> In article <[e-mail]>,
> [e-mail] (Jochen Lippert) wrote:
> :> I favor a control strip module for my product, but wonder if it's
> :> the right approach.  Is it acceptable to users? Do most users use
> :> it?  Does it have a future?
>
> i don't know how popular they are but every mac user i have shown the
> control strip to has fallen in love with it
> yet another thing apple had underpromoted.

Honestly, I don't like Control Strip at all. In theory it's fantastic, but
it gets in the way!! Especially if you open a game or something that wants
the whole screen, it just takes this ugly little triangular chunk in the
corner of the screen and won't let it go. Which of course, means that
games that CopyBits slow down because the destination region isn't a plain
rectangle, but that's another point...

I would WORSHIP the control strip if it were an icon the menu bar between
the Keyboard and Apple-Guide menus, and it dropped down like a menu and
rolled back up when I was done with it. That sounds like a much better
idea to me (no desktop space eaten constantly). Of course, Windows 95
users got used to a fixed toolbar at the bottom of the screen, so I guess
I could too... (of course, at least in Win95, windows don't appear with
the resize boxes hidden right under the start bar!)

BTW what is extensions strip...?

*Stiles

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

From Madmax <[e-mail]>
Subject: Help! How do a draw a picture offscreen?
Date: Sun, 29 Sep 1996 14:44:54 -0800
Organization: "oronet, Penn Valley, CA"

How do I draw a picture off screen so I can have it appeear instantly
on screen?
Cory Loken

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

From "Tedd F. Sperling" <[e-mail]>
Date: Mon, 30 Sep 1996 09:02:49 +0000
Organization: Sperling Corporation

Madmax wrote:
>
> How do I draw a picture off screen so I can have it appeear instantly
> on screen?
> Cory Loken


What language?

If you're in FutureBasic, I'll send you the code.

tedd


--
   ____
  |[  ]| mac programmer                             tedd f. sperling
  |[__]|                               mailto: [e-mail]
  |___-|                http://www.sojourn.com/~sperling/resume.html
- -

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

From [e-mail] (Mauricio Tavares)
Date: 9 Oct 1996 14:23:11 GMT
Organization: University of Florida

Tedd F. Sperling ([e-mail]) wrote:
: Madmax wrote:
: >
: > How do I draw a picture off screen so I can have it appeear instantly
: > on screen?
: > Cory Loken
:
: What language?
:
: If you're in FutureBasic, I'll send you the code.
:
    I have some code to do that in C.

--
 ===========================+==========================================
| Mauricio Tavares          | "We will attack...                       |
| [e-mail] |  ...under the cover of daylight!" Rimmer |
 ===========================+==========================================

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

From [e-mail] (Jim Wintermyre)
Date: 11 Oct 1996 15:05:41 GMT
Organization: Rome Laboratory

In article <53gcgf$[e-mail]>
[e-mail] (Mauricio Tavares) writes:

> Tedd F. Sperling ([e-mail]) wrote:
> : Madmax wrote:
> : >
> : > How do I draw a picture off screen so I can have it appeear instantly
> : > on screen?
> : > Cory Loken
> :
> : What language?
> :
> : If you're in FutureBasic, I'll send you the code.
> :
>         I have some code to do that in C.

There's some Pascal code that does this in the develop 22 source code.
Look in the "Custom Color Search Procedures" code for the files
SP_PICTs.p and SP_GWorlds.p.

Jim

[e-mail]
[e-mail]


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

From [e-mail] (Duncan Mullier)
Subject: Icons on the desktop?
Date: Wed, 02 Oct 96 12:38:05 GMT
Organization: Digital Amusement Ltd

How do I get an icon to be associated with my application? I've drawn one,
and included it in my project but to no avail.

Dunc

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

From [e-mail] (Simon Fraser)
Date: Thu, 03 Oct 1996 23:10:54 -0700
Organization: Santa Fe Institute

In article <531qo9$[e-mail]>, "Thomas L. Ferrell"
<[e-mail]> wrote:

>[e-mail] (Duncan Mullier) wrote:
>>How do I get an icon to be associated with my application? I've drawn
>one, and included it in my project but to no avail.
>>
>>Dunc
>
>You can do it the proper way with ResEdit and code, or you can do it the
easy way. Just select your icon, go to File--Get Info, sele=
>ct the icon in the GetInfo window, copy it to the clipboard, then do
File--Get Info on your app, select its icon in the GetInfo wind=
>ow, and paste your custom icon from the clipboard.

No, this is not what he wants to do. This method inserts some icon
resources in the file, with ID -16455, but does not associate
documents with their creating application in any way.

Duncan, you need to read a ResEdit primer, and some of Inside Mac
to get the full description, so that you understand how it all work.
But, in ResEdit, the basic procedure is this.

In the resource file for your app, creature a new resource of
type 'BNDL' (leave the default ID at 128). In the window that
opens, type your app's signature (a 4-letter code that identified
your app. You should use either all uppercase, or a mixture of
upper- and lowercase letters; Apple reserver the lowercase ones).
For example, NewsWatcher's is 'NNTP'.

Now create a new file type by hitting Command-k while the BNDL
window is still open. First, you should add the type 'APPL',
which provides icons for you application; click in the left
column, and type APPL as the type. Then double-click on the
icon boxes on the RHS, and draw yourself a nice icon for the app.

Now add another type, which will be one of your document types.
For example, if this app handles text docs, and you want to
make TEXT files with an icon from your app, do Command-k again,
type 'TEXT' as the type, and draw another icon for text file.

Continue, until you have types for each of the file types that
you want this app to own.

Now, when you close the BNDL window, ResEdit will show that you
have created in addition a bunch of FREF resources (one for each
type), some icon resources, and the owner resource, which has the
type you have the app.

Then go on and compile this resource file into your app.
There is one final stage to get the icons to show up. You need
to get them into the desktop database, so either do a desktop
rebuild on restart (command-option), or use a utility like
Save-a-BNDL.

Of course, when your app creates a file, the creator code of
the file needs to be set to that of the app, and the type code
should be one of the ones you entered into the resource file.

Simon

--
________________________________________________________________
Simon Fraser                                  Santa Fe Institute
[e-mail]                             1399 Hyde Park Road
http://www.santafe.edu/~smfr/                 Santa Fe, NM 87501

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

From "Thomas L. Ferrell" <[e-mail]>
Date: 4 Oct 1996 01:54:17 GMT
Organization: Oak Ridge National Lab

[e-mail] (Duncan Mullier) wrote:
>How do I get an icon to be associated with my application? I've drawn >one, and included it in my project but to no avail.
>
>Dunc

You can do it the proper way with ResEdit and code, or you can do it the easy way. Just select your icon, go to File--Get Info, sele=
ct the icon in the GetInfo window, copy it to the clipboard, then do File--Get Info on your app, select its icon in the GetInfo wind=
ow, and paste your custom icon from the clipboard.
tom


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

From [e-mail] ( nick.c @MT )
Date: Fri, 11 Oct 1996 17:54:54 -0700
Organization: MacTech Magazine



[e-mail] (Simon Fraser) wrote:

>Duncan, you need to read a ResEdit primer, and some of Inside Mac
>to get the full description, so that you understand how it all work.
>But, in ResEdit, the basic procedure is this.
>
>In the resource file for your app, creature a new resource of
>type 'BNDL' (leave the default ID at 128). In the window that
>opens, type your app's signature (a 4-letter code that identified
>your app. You should use either all uppercase, or a mixture of
>upper- and lowercase letters; Apple reserver the lowercase ones).
>For example, NewsWatcher's is 'NNTP'.

...

>Then go on and compile this resource file into your app.
>There is one final stage to get the icons to show up. You need
>to get them into the desktop database, so either do a desktop
>rebuild on restart (command-option), or use a utility like
>Save-a-BNDL.


     One more step: you need to tell your compiler to set the
       'creator' code of your appliction to the same four letter
       code (ie. 'NNTP').  In CodeWarrior (for example) open
       your project, go to the edit menu and open the preference
       item.  Select the project item in the dialog that appears
       and look for a box that has the word "creator" next to
       it and "????" in the box.  Change the "????" to your
       creator code "NNTP".  Do this before compiling your
       app (otherwise the bundle will be entered into the
       desktop database, but the finder won't know to associate
       the correct icon with your applications and you'll
       still get the default).




____Nicholas C. DeMello, Ph.D.___________________________________________
"MacTech Online"--MacTech Magazine, for Mac OS Programmers and Developers
     http://www.MacTech.com/
                                        _/   _/  _/  _/_/_/   _/   _/
   Chemistry: [e-mail]       _/_/ _/  _/  _/   _/  _/_/_/
     MacTech: [e-mail]      _/ _/_/  _/  _/       _/ _/
        http://www.chem.ucla.edu/~nick/   _/  _/   _/_/_/  _/   _/

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

From "Seth Hill" <[e-mail]>
Subject: Is it MacOS 4.1 or System 4.1?
Date: 7 Oct 96 10:25:57 -0700
Organization: ProAxis Communications, Inc.

I=B9m writting some documentation for my upcomming program, and I have
to refer to an old version of system software, and I was wondering
what the official way of describing it is. Back when we were all using
System 6.0.2, we called it "System 6.0.2." But now, we have "MacOS
7.5" and "MacOS 8." I figure that Copland and 7.5 qualify as a "MacOS"
and not a "System," but what about the older system software? Are we
supposed to call it "System 7" or "MacOS 7?"

Just trying to solve another daunting Macintosh mystery.

--
Seth Hill -- [e-mail]
http://www.proaxis.com/~ferris/
Don't drink and park because accidents can cause people...



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

From [e-mail] (Jude Giampaolo)
Date: 7 Oct 1996 19:42:47 GMT
Organization: CyberDrugs

In article <AE7E8B41-AA70@206.163.142.175>, "Seth Hill"
<[e-mail]> wrote:

> I=B9m writting some documentation for my upcomming program, and I have
> to refer to an old version of system software, and I was wondering
> what the official way of describing it is. Back when we were all using
> System 6.0.2, we called it "System 6.0.2." But now, we have "MacOS
> 7.5" and "MacOS 8." I figure that Copland and 7.5 qualify as a "MacOS"
> and not a "System," but what about the older system software? Are we
> supposed to call it "System 7" or "MacOS 7?"

I'm pretty sure it was "System x.x.x" before the release of 7.5.3 I think.
THey renamed the whole contraption to "MacOS" at that point.

--
Jude Charles Giampaolo        'I was lined up for glory, but the
[e-mail]                 tickets sold out in advance' -Rush
[e-mail]      http://prozac.cwru.edu/jude/JudeHome.html

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

From "John W. Baxter" <[e-mail]>
Date: Mon, 07 Oct 1996 18:51:55 -0700
Organization: Internet for the Olympic Peninsula

Seth Hill wrote:
>
> I1m writting some documentation for my upcomming program, and I have
> to refer to an old version of system software, and I was wondering
> what the official way of describing it is. Back when we were all using
> System 6.0.2, we called it "System 6.0.2." But now, we have "MacOS
> 7.5" and "MacOS 8." I figure that Copland and 7.5 qualify as a "MacOS"
> and not a "System," but what about the older system software? Are we
> supposed to call it "System 7" or "MacOS 7?"
>
> Just trying to solve another daunting Macintosh mystery.

No one ever thought of 4.1 as MacOS 4.1, since the term MacOS wasn't
invented (and the necessary trademarking, etc done) until long after
that time.  [It was done to get the word "Macintosh" out of software
which runs on other machines in addition to Macintosh.  Running on other
hardware by "obtaining" a Mac ROM didn't count.]  It's very highly
unlikely that 4.1 will run on any of the machines whose existence
triggered the MacOS name.

If you call it System 4.1 "MacOS" now, you'll just confuse things.  And
might be violating trademark.

   --John

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

From [e-mail] (Laurent Gasser)
Date: 8 Oct 1996 12:09:37 GMT
Organization: Swiss Meterological Institute

The old Mac had Systems (up to 7.1). The MacOS way to call it appeared with
the new licensing policy.  System 7.5.1 could no more be called so, because
it would run on computers others than Apple's.

--
Laurent Gasser ([e-mail])
Computers do not solve problems, they execute solutions.



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

From [e-mail] ( nick.c @MT )
Date: Tue, 08 Oct 1996 09:47:54 -0700
Organization: MacTech Magazine



   I'd call it "System 4.1", however if you want to emphasize that
     you're refering to the Macintosh operating system you could
     say "Macintosh System 4.1".  I think calling it MacOS 4.1
     would just confuse folks (JMO).

____Nicholas C. DeMello, Ph.D.________________________________________

   Online from MacTech Magazine, for Mac OS Programmers and Developers
     http://www.MacTech.com/
                                       _/   _/  _/  _/_/_/   _/   _/
  Chemistry: [e-mail]       _/_/ _/  _/  _/   _/  _/_/_/
    MacTech: [e-mail]      _/ _/_/  _/  _/       _/ _/
       http://www.chem.ucla.edu/~nick/   _/  _/   _/_/_/  _/   _/

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

From [e-mail] (Jay Riley)
Date: Wed, 09 Oct 1996 13:40:49 -0500
Organization: DATAMAGIK Ä Systems, Software & Design Engineering

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

>Seth Hill wrote:
>>
>> I1m writting some documentation for my upcomming program, and I have
>> to refer to an old version of system software, and I was wondering
>> what the official way of describing it is. Back when we were all using
>> System 6.0.2, we called it "System 6.0.2." But now, we have "MacOS
>> 7.5" and "MacOS 8." I figure that Copland and 7.5 qualify as a "MacOS"
>> and not a "System," but what about the older system software? Are we
>> supposed to call it "System 7" or "MacOS 7?"
>>
>> Just trying to solve another daunting Macintosh mystery.
>
>No one ever thought of 4.1 as MacOS 4.1, since the term MacOS wasn't
>invented (and the necessary trademarking, etc done) until long after
>that time.  [It was done to get the word "Macintosh" out of software
>which runs on other machines in addition to Macintosh.  Running on other
>hardware by "obtaining" a Mac ROM didn't count.]  It's very highly
>unlikely that 4.1 will run on any of the machines whose existence
>triggered the MacOS name.
>
>If you call it System 4.1 "MacOS" now, you'll just confuse things.  And
>might be violating trademark.

Hmmm...I doubt it would count as trademark infringement since MacOS is an
Apple trademark. I've tried to start refering to all incarnations of the
OS as MacOS for consistency. Besides the word "System" is fairly generic.

Seth I personally wouldn't be confused by either reference, or even a
mixture of the two, but some newcomers to the Mac might find "MacOS 4.1"
clearer and besides it gives you a sense of history (MacOS has been around
since 1983). I believe the version that shipped in 1984 was MacOS 1.1g.

 ====================================================================
  DATAMAGIK == 107 Ranch Rd. 620 S. #10F == Austin, Texas  USA 78734
  <mailto:[e-mail]> or visit <http://www.flash.net/~jriley/>
 ====================================================================
  Currently routed by FlashNet Communications <http://www.flash.net>

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

From [e-mail] (Steven Ellis)
Subject: Localizing application on the Mac
Date: Wed, 09 Oct 1996 14:46:33 +0100
Organization: MicroAPL Ltd


We have developed an application in English which now need to be localised
for France, Germany, and Spain. We have also followed all of Apple's
guidelines regarding the placement of strings in resources etc.

We are currently using AppleGlot to insert the Localised strings into the
various versions of the application, but have a problem with dialog sizes.

The application was devloped in C++ using Metrowerks PowerPlant, and as
such we have a mix of PPobs and standard Mac dialogs. Some of these require
resizing for the localised strings to fit.

The problem with AppleGlot, is it loses the resized values when you update
the resulting application. As this is part of an ongoing project, we are
looking for a more automated solution..

Any ideas out there? Is there an FAQ on the subject, and why is Apple's
documentation so bad when they are so keen for use to localise
applications.

Thanks

Steven Ellis

- --
At work so usually only lurking ....   [e-mail].
- --
Eric Clapton....  May the hand of God play guitar forever.

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

From [e-mail] (Kent Johnson)
Date: Thu, 10 Oct 1996 13:19:21 GMT
Organization: Transparent Language, Inc.

In article <[e-mail]>,
[e-mail] (Steven Ellis) wrote:

> We have developed an application in English which now need to be localised
> for France, Germany, and Spain.
>
> We are currently using AppleGlot to insert the Localised strings into the
> various versions of the application, but have a problem with dialog sizes.
>
> The application was devloped in C++ using Metrowerks PowerPlant, and as
> such we have a mix of PPobs and standard Mac dialogs. Some of these require
> resizing for the localised strings to fit.
>
> The problem with AppleGlot, is it loses the resized values when you update
> the resulting application.

I am using AppleGlot and I don't have this problem with PPob's (I don't
use any DLOG's so I don't know about that). I don't know what you are
doing differently, but for me one of the nice features of AppleGlot is
that it DOES preserve my UI tweaks across releases.

What version of AppleGlot are you using? I use 2.4b2.

Kent Johnson                                [e-mail]
Transparent Language, Inc.                http://www.transparent.com

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

From [e-mail] (Sak Wathanasin)
Date: Fri, 11 Oct 1996 12:22:13 GMT
Organization: Network Analysis Ltd

In article <[e-mail]>,
[e-mail] (Steven Ellis) wrote:

> We have developed an application in English which now need to be localised
> for France, Germany, and Spain. We have also followed all of Apple's
> guidelines regarding the placement of strings in resources etc.

I've done something similar using MacApp rather than PowerPlant. I agonized
over whether to just replace the strings or entire views/panes/dialogues.
In the end, I opted for the latter since we had to support Japanese as
well, and it was just too awkward to try to fit Japanese into the same
space (it needs a different font, min size of 12 pt to be legible and so
on). You may find this easier than trying to use the same dlgs/panes with
different sets of text strings.

I certainly found, after the initial translation, that it was easier to
make incremental changes to the other language versions than to regenerate
them all from (say) the English version. YMMV. Also, I found it easier to
work with the derezed form of the resource. That way I could use
comparefiles to merge 2 different versions of the resource.

> and why is Apple's
> documentation so bad when they are so keen for use to localise
> applications.

Hah! I take it you haven't fought the 2-byte stuff yet (:-).

--
Sak Wathanasin
Network Analysis Limited
178 Wainbody Ave South, Coventry CV3 6BX, UK

Internet: [e-mail]
uucp:     ...!britain.eu.net!nan!sw
Phone: [phone]                              Fax: [phone]


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

From [e-mail] (Michael Thies)
Subject: Looking for Apple Remote Access API
Date: 8 Oct 1996 06:56:30 GMT
Organization: University of Paderborn, Germany

Is the Apple Remote Access API available somewhere for download?
I have only found it mentioned as part of the Apple Developer Mailing
on the SDK CD, which would be overkill for the problem at hand.

I am interested in controling the answer call status of the ARA Personal
Server on a Macintosh running under At Ease. So ARA Commander and other
solutions which simulate the user clicking the Answer Calls check box in
the ARA Setup control panel won't work.

The intention is to allow printing to a locally connected printer, while
ARA PS server is setup to answer incomming calls (and no call is currently
being answered). The printer shares the modem port with the modem used for
ARA via a mechanical switchbox (printer port is used for local AppleTalk
network).

Interestingly, FaxExpress has no difficulties sending out faxes over the
same modem with Answer Calls on, but the Apple LaserWriter LS driver
complains about the port being in use. Perhaps this check could just be
patched out and Serial Port Arbitrator would take care of the rest?

Thanks in advance for any info.
--
Michael Thies  ([e-mail])



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

From [e-mail] (Jay Farrell)
Date: Tue, 08 Oct 1996 12:11:18 -0400
Organization: Net Access - Philadelphia's Original ISP

In article <53ctuu$[e-mail]>, [e-mail] wrote:

> Is the Apple Remote Access API available somewhere for download?
> I have only found it mentioned as part of the Apple Developer Mailing
> on the SDK CD, which would be overkill for the problem at hand.
>
> I am interested in controling the answer call status of the ARA Personal
> Server on a Macintosh running under At Ease. So ARA Commander and other
> solutions which simulate the user clicking the Answer Calls check box in
> the ARA Setup control panel won't work.
>
> The intention is to allow printing to a locally connected printer, while
> ARA PS server is setup to answer incomming calls (and no call is currently
> being answered). The printer shares the modem port with the modem used for
> ARA via a mechanical switchbox (printer port is used for local AppleTalk
> network).
>
> Interestingly, FaxExpress has no difficulties sending out faxes over the
> same modem with Answer Calls on, but the Apple LaserWriter LS driver
> complains about the port being in use. Perhaps this check could just be
> patched out and Serial Port Arbitrator would take care of the rest?

I don't know if this will help or not.  The chooser extension Apple
recommends for the Laserwriter LS is Laserwriter 300, as opposed to the
original printer-specific extension.  Why not try that?

There is also an Apple Remote Access OSAX available (see info-mac), but the
last version I tried was still buggy, and I can't recall if it does what
you want:

<URL:ftp://mirrors.aol.com/pub/info-mac/comm/ara/ara-scripting-addition.hqx>

Cheers,
Jayfar

   ////~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~////
  ////   The Mops Page    <URL:http://www.netaxs.com/~jayfar/mops.html>    ////
 ////~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~////
////   Mops is Mike Hore's Freeware Forth/Smalltalk hybrid for Macintosh ////
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  Jay Farrell     [e-mail]        Philadelphia, Pennsylvania, USA

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

From [e-mail] (Jump Long)
Date: 10 Oct 1996 00:59:14 -0400
Organization: America Online, Inc. (1-800-827-6364)

Michael Thies wrote:

>Is the Apple Remote Access API available somewhere for download?
>I have only found it mentioned as part of the Apple Developer
>Mailing on the SDK CD, which would be overkill for the problem
>at hand.

Many of Apple's SDKs are available for download from
<ftp://ftpdev.info.apple.com/Developer_Services/Development_Kits>, but I
don't see the ARA SDK there.

- Jim Luther

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

From David Thomas Mcwherter <[e-mail]>
Subject: Low memory situations & Toolbox
Date: Tue,  8 Oct 1996 09:56:42 -0400
Organization: Carnegie Mellon, Pittsburgh, PA

    I'm attempting to add the ability to my application to handle
low-memory situations a little gracefully.  I have an alert, which
is defined in a 'DITL' and 'ALRT' resource.  I would like to show
this alert whenever I run into a low memory situation, offering
the user the ability to quit the program.  I load these two resources,
using GetResource() at the very beginning of my application.  I
mark the two handles I get non-purgeable, so they stay in the heap
pretty much forever.  I then generate a low memory situation in
which I have less than 4K left in my heap (by calling NewPtr() many
times).  I have two problems beginning here.

1)  When I try to call GetNewWindow(), the call itself crashes when
there is 1K or less in my heap.  This occurs even if I preload the
'WIND' resource that I'm trying to use.

2)  When I try to call StopAlert() when there is 4K or less of available
RAM in my heap, the call crashes my computer.  (the call is trying
to load the alert that I tried to preallocate above)

Is there a way to detect when these crashes will occur?  Or is there
a way to prevent them from occuring?

In case it's useful, or if I have found a system bug, I've only tested
this on a PowerMac 7500 with System 7.5.3, virtual memory turned off,
with 24MB of real RAM available, and 1.5MB given to my application.

Thanks,
David McWherter
[e-mail]


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

From [e-mail] (Tom Emerson)
Date: Tue, 08 Oct 1996 11:22:46 -0400
Organization: Apple Computer, Inc.

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

[...]
>1)  When I try to call GetNewWindow(), the call itself crashes when
>there is 1K or less in my heap.  This occurs even if I preload the
>'WIND' resource that I'm trying to use.

You should never let your heap get to the point that there is only 1K left:
I'm not surprised the app crashes at that point.

>2)  When I try to call StopAlert() when there is 4K or less of available
>RAM in my heap, the call crashes my computer.  (the call is trying
>to load the alert that I tried to preallocate above)

Preloading the DITL and ALRT won't guarantee that calling StopAlert() will
succeed, and while you have at most 4K total available, how much contiguous
memory do you have?

>Is there a way to detect when these crashes will occur?  Or is there
>a way to prevent them from occuring?

I would suggest allocating some amount of "emergency" memory at startup
(16K? 40K? Whatever it takes to allow your app to notify the user that they
may be hosed and that they should save whatever they're doing) which you
deallocate in your GrowZone proc when the need arises. Set a flag in your
GrowZone proc saying that you're low on memory, and the next time through
your even loop call StopAlert() (or whatever).

I presume you are not calling the Toolbox from your grow zone?

    -tre

--
Tom Emerson                                                 Cambridge R&D
Senior Software Engineer                             Apple Computer, Inc.
<mailto:[e-mail]>                  <http://www.tiac.net/users/tree>

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

From [e-mail] (Don Sample)
Date: Tue, 08 Oct 1996 14:01:52 -0400
Organization: DASoftware

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

>    I'm attempting to add the ability to my application to handle
>low-memory situations a little gracefully.  I have an alert, which
>is defined in a 'DITL' and 'ALRT' resource.  I would like to show
>this alert whenever I run into a low memory situation, offering
>the user the ability to quit the program.  I load these two resources,
>using GetResource() at the very beginning of my application.  I
>mark the two handles I get non-purgeable, so they stay in the heap
>pretty much forever.  I then generate a low memory situation in
>which I have less than 4K left in my heap (by calling NewPtr() many
>times).  I have two problems beginning here.
>
...

>Thanks,
>David McWherter
>[e-mail]

Preloading the ALRT, DITL and WIND resources doesn't help much, because
these are just templates used to build the data structures required to
create the windows.  When you call StopAlert() or GetNewWindow() a lot
more memory gets allocated.

--
Don Sample                          |
[e-mail]                     |    Quando Omni Flunkus Moritati
[e-mail]                 |
http://www.synapse.net/~dsample/    |

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

From [e-mail] (Michael L. Richardson)
Date: Thu, 10 Oct 1996 16:40:29 -0700
Organization: University of Washington

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

>     I'm attempting to add the ability to my application to handle
> low-memory situations a little gracefully.
[...]

Read IM:Memory, and learn about emergency memory reserves and GrowZoneProcs.

At a bare minimum, wrap each call to NewPtr/NewHandle with a check for the
amount of free memory, and refuse the request when there is about 30 - 50K
left.

-thurman gillespy


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

From [e-mail] (Stephen Jonke)
Subject: MacOS Java problems (Was: [Java]:  Refresh (paint()) in MW Java v. Netscape)
Date: Wed, 09 Oct 1996 11:05:56 -0500
Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA

NOTE: I'm posting this response to two newsgroups as I felt it was
appropriate for more then just the .codewarrior group where the original
article was posted.  I note this so that people are aware of it and may
remove one or the other as appropriate in their responses.

In article <53eion$[e-mail]>, [e-mail]
(Mike Rogers) wrote:

>Loathe as I am to admit it, I do not see the same problem in Windoze '95,
>i.e., the screen does get properly repainted().  Has anybody else seen
>this as a bug in Netscape 3.0 (Gold) for the Mac?

Unfortunately just about everything Java is done much better in Windows
95/NT then it is on the Mac.  Only the most basic Java applications/applets
run properly on the Mac and virtually all applets/applications runs great
under Windows 95/NT.  It's quite depressing.  :(

One particularly unexpected/bizarre Java problem on the Mac is the MacOS
file name limit of 31 characters!  It is common to come across stuff that
doesn't work on the Mac because some of the class names are too long and
their associated class file names are getting truncated (which breaks
Java).  It's easy to say "well then the names are too long", but aside from
the pointlessness of this argument since class names are being written that
are too long, it's hard to argue that a class name like
"ArrayIndexOutOfBoundsException" (don't forget to add ".java" to the end
for a total of 35 characters in the filename) defining a particular
Exception is innappropriate and that truncating/removing some of the words
is a good idea.

And that particular example is a standard Java Exception that Sun defines -
it sets a precedant for how they should be named and there are many others
in standard Java that are similarly too long for the MacOS.  Only the MacOS
and Windows 3.1 choke on it without a kludge (Metrowerks et. al. were
forced to kludge solutions to the problem just to be able to provide
development tools.)  It's a miracle that anything runs at all under Windows
3.1, but they are no worse off then the MacOS, sad to say.

All the Mac Java VM's have many problems.  Suns, Metrowerk's, Natural
Intelligence's, Symantecs....  I hope that Apple realizes this problem and
is making it a top priority to help these folks out as well as updating the
MacOS to handle Java better.

Steve

--
Seen in computer peripheral ad: "User-friendly dip switches!"

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

From [e-mail] (Adam Winer)
Date: Wed, 09 Oct 1996 21:42:33 -0800
Organization: Oracle Corporation

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

> In article <[e-mail]>,
> [e-mail] (Stephen Jonke) wrote:
>
> [snip out all the interesting stuff]
>
> > All the Mac Java VM's have many problems.  Suns, Metrowerk's, Natural
> > Intelligence's, Symantecs....  I hope that Apple realizes this problem and
> > is making it a top priority to help these folks out as well as updating the
> > MacOS to handle Java better.
>
> I read in several places that Apple is working hard on integrating a Java
> VM into the MacOS so applications can use Java or call applets, etc.
> There's a beta or alpha on one of their developer web sites that allows
> java applets to be embedded in OpenDoc containers (hey, I thought it was
> cool).  The web site seemed to imply that it would be available as a
> shared library and a part, but I know I read something about Apple wanting
> to integrate it into the MacOS.  Hard to know for sure these days what
> Apple intends to do with the MacOS, tho.  I also recall reading about a
> deal between Natural Intelligence and Apple over NI's Java VM, perhaps
> someone else here know the details.

It's an alpha, and a very early alpha at that.  It's also
pretty irrelevant because:
  a) It uses the Sun VM/AWT implementation, not the NI JIT VM.
  b) It doesn't include any of the API they'll expose in the future
     to let the rest of us expose Java in our apps.
  c) It's somewhat slow and buggy as hell.

I've heard rumors that the MacOS Java integration (i.e.,
the official release of the MacOS Java Runtime) comes
with 7.7, which should show in July, but nothing official.

-- Adam Winer
[e-mail]

--
Adam Winer
[e-mail]
Technical Staff, Oracle Corp.

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

From [e-mail] (Stephen Jonke)
Date: Thu, 10 Oct 1996 12:30:51 -0500
Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA

I think the most important thing Apple needs to solve is the 31 character
filename limit.  Natural Intelligence/Metrowerks/Symantec/Netsape can't do
this for Apple.  Java classes with names longer then 26 characters (+
".java" = 31 characters) are not going to go away by themselves, especially
since some are part of the Java spec!  Apple needs to bring the MacOS up to
snuff when it comes to file names and they need to do it now.

I hope that that Sytem 7.6 (not 7.7 or 8.0 or 9.0 or ...) will increase the
max file name length.  Then, of course, apps will have to be updated to not
presume a file name of <= 31 characters.  It's a major problem and it needs
to be solved lest the Mac be left in the dust in the world of Java.

Steve

--
Seen in computer peripheral ad: "User-friendly dip switches!"

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

From [e-mail] (Ben Trumbull)
Date: Wed, 09 Oct 1996 15:08:54 -0400
Organization: Yale University

In article <[e-mail]>,
[e-mail] (Stephen Jonke) wrote:

[snip out all the interesting stuff]

> All the Mac Java VM's have many problems.  Suns, Metrowerk's, Natural
> Intelligence's, Symantecs....  I hope that Apple realizes this problem and
> is making it a top priority to help these folks out as well as updating the
> MacOS to handle Java better.

I read in several places that Apple is working hard on integrating a Java
VM into the MacOS so applications can use Java or call applets, etc.
There's a beta or alpha on one of their developer web sites that allows
java applets to be embedded in OpenDoc containers (hey, I thought it was
cool).  The web site seemed to imply that it would be available as a
shared library and a part, but I know I read something about Apple wanting
to integrate it into the MacOS.  Hard to know for sure these days what
Apple intends to do with the MacOS, tho.  I also recall reading about a
deal between Natural Intelligence and Apple over NI's Java VM, perhaps
someone else here know the details.

In any event, with Apple and Sun's new alliance, you can expect Java to
receive considerable attention and effort from Apple.

terminally curious,

Ben
___________________________________________________________________
Benjamin Trumbull
[e-mail]
Yale University
       You can't be in hell; you can still read your e-mail

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

From [e-mail] (Lars Farm)
Date: Fri, 11 Oct 1996 14:41:08 +0200
Organization: pv

Stephen Jonke <[e-mail]> wrote:

> Then, of course, apps will have to be updated to not
> presume a file name of <= 31 characters.

This is probably a problem (according to Murphy anything that can go
wrong will), but it *should* not be a problem because files are
currently (since 91) located with FSSpec's and they store 63 chars. Of
course replacing a 31 char limit with a 63 char limit is an improvement,
but when one does that one has to ask: "why stop at Str64?".


--
Lars Farm, [e-mail]

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

From Glenn R. Howes <[e-mail]>
Date: 11 Oct 1996 00:04:21 GMT
Organization: Generally Helpful Software

In article <[e-mail]>
Stephen Jonke, [e-mail] writes:
>I think the most important thing Apple needs to solve is the 31 character
>filename limit.  Natural Intelligence/Metrowerks/Symantec/Netsape can't do
>this for Apple.  Java classes with names longer then 26 characters (+
>".java" = 31 characters) are not going to go away by themselves, especially
>since some are part of the Java spec!

When a standard universal class archive is set, this problem will pretty
much go away. Then all the class names will be in the archive, not the
file system, like you can do with .zip files under Netscape now.

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

From "John W. Baxter" <[e-mail]>
Date: Fri, 11 Oct 1996 19:41:24 -0700
Organization: Internet for the Olympic Peninsula

Lars Farm wrote:
>
> Stephen Jonke <[e-mail]> wrote:
>
> > Then, of course, apps will have to be updated to not
> > presume a file name of <= 31 characters.
>
> This is probably a problem (according to Murphy anything that can go
> wrong will), but it *should* not be a problem because files are
> currently (since 91) located with FSSpec's and they store 63 chars. Of
> course replacing a 31 char limit with a 63 char limit is an improvement,
> but when one does that one has to ask: "why stop at Str64?".
>
The 63 character FFSpec name was a compromise between not eating the
whole stack, and supporting real-world MFS floppies (where the actual
file name limit was 253 (given a 1 character volume name...the whole
path with ':' had to be 255 or less)).  File names were often longer
than 31...officially it was 127 each side of the ':' in MFS.  In
hindsight, an overall FSSpec length which occupies a multiple of
processor cache line size would have been nice...probably meaning a
55-character (oops...byte) name length.

The first released System 7 Finder could be crashed very easily:  insert
an MFS floppy containing a file name longer than 63 characters.
Somewhere in Finder, an FSSpec was written beyond, and bang.  I don't
know whether the current Finder has had the problem fixed, or defined
out of existence [if it hurts, don't do it] (I don't have such a
floppy).

  --John

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

From [e-mail] (Francois Pottier)
Date: 11 Oct 1996 16:13:15 GMT
Organization: INRIA Rocquencourt, BP 105, 78153 Le Chesnay Cedex, France

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

>currently (since 91) located with FSSpec's and they store 63 chars. Of

FSSpecs have 63 chars but the limit is still documented to be 31 chars,
and some apps use this knowledge to save space (for instance, when saving
filenames to disk). So, some apps would still break.

--
Francois Pottier
[e-mail]
http://pauillac.inria.fr/~fpottier/

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

From [e-mail] (Stephen Jonke)
Date: Fri, 11 Oct 1996 13:38:59 -0500
Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA

NOTE: I've moved this to only comp.sys.mac.programmer.misc as I don't think
that it applies specifically to Codewarrior any more.

In article <53k2u5$[e-mail]>, Glenn R. Howes
<[e-mail]> wrote:

>When a standard universal class archive is set, this problem will pretty
>much go away. Then all the class names will be in the archive, not the
>file system, like you can do with .zip files under Netscape now.

That would work, although it only applies to end users, not developers.
If, like I, you are working in a mixed PC, Mac and Unix environment and
your colleague creates a new exception class (or whatever) source file with
a file name too long for the Mac, you still have a problem.  And don't
forget about getting java classes from the net (in the form of source
files), a lot of which are being produced by PC and Unix users who are
unaware of this problem or could care less.

I suppose if the development tool completely bypasses the Mac file system
and works directly with these archives containing source .java files (which
I'm not sure is even the intent) that would work, but you would not be able
to take that file and open it with some other Mac tool unless it also
explicitely supported opening source "files" directly from an archive.  And
in any case this is a hack that makes it at best akward to use a Mac for
Java development.  The only real solution is for the MacOS to inherently
support longer file names.

There's something particularly humiliating/dissatisfying about having to
ask a Windows NT user to shorten their file names because they are too long
for the Mac!  :(

Steve

--
Seen in computer peripheral ad: "User-friendly dip switches!"


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

From "kleinemans" <[e-mail]>
Subject: Pascal compiler??
Date: 29 Sep 1996 17:28:16 GMT
Organization: EuroNet Internet


--Cyberdog-AltBoundary-0002F7FB
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

I'm a biginning macintosh programmer, and I'm looking for a cheap
pascal compiler.
(Don't begin about learning C pleasse)
Are there any shareware compilers? I'm now using borlands turbo
pascal('87)
Help me!


- -------------------------------------------------
This message was created and sent using the Cyberdog Mail System
- -------------------------------------------------



--Cyberdog-AltBoundary-0002F7FB
Content-Type: multipart/mixed; boundary="Cyberdog-MixedBoundary-0002F7FD"
Content-Transfer-Encoding: 7bit


--Cyberdog-MixedBoundary-0002F7FD
Content-Type: text/enriched; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<SMALLER><X-FONTSIZE><PARAM>10</PARAM><FONTFAMILY><PARAM>Geneva</PARAM>=
I'm a biginning macintosh programmer, and I'm looking for a cheap
pascal compiler.

(Don't begin about learning C pleasse)

Are there any shareware compilers? I'm now using borlands turbo
pascal('87)

Help me!



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

This message was created and sent using the Cyberdog Mail System

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

</FONTFAMILY></X-FONTSIZE></SMALLER>
--Cyberdog-MixedBoundary-0002F7FD--

--Cyberdog-AltBoundary-0002F7FB--


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

From Jason J Mullins <[e-mail]>
Date: Mon, 30 Sep 1996 11:30:14 +0100
Organization: McQueen

kleinemans wrote:
>
> I'm a biginning macintosh programmer, and I'm looking for a cheap
> pascal compiler.
> (Don't begin about learning C pleasse)
> Are there any shareware compilers? I'm now using borlands turbo
> pascal('87)

I don't know of a shareware compiler, but MetroWerks do a 'Bronze'
version of CodeWarrier, which has Pascal, C, C++, etc, etc, and costs
much less than the full version, and does 68k only. If you are a student
then the prices on Academic CW are even better.

The best bit is that with all options you get lots of electronic
manuals, teaching books, the CW API and updates.

=0)

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

From [e-mail] ( nick.c @MT )
Date: Wed, 09 Oct 1996 15:34:07 -0700
Organization: MacTech Magazine



[e-mail] (Mauricio Tavares) wrote:
>Jason J Mullins ([e-mail]) wrote:

>: I don't know of a shareware compiler, but MetroWerks do a 'Bronze'
>: version of CodeWarrier, which has Pascal, C, C++, etc, etc, and costs
>: much less than the full version, and does 68k only. If you are a student
>: then the prices on Academic CW are even better.
>:
>: The best bit is that with all options you get lots of electronic
>: manuals, teaching books, the CW API and updates.
>:
>        Or he can get Symantec's Think Pascal...


      TP is a very kool tool--but it's worth noting you can't
        generate PPC binaries with it.  If you can get the CW
        academic, it's probably a better deal.




____Nicholas C. DeMello, Ph.D.___________________________________________
"MacTech Online"--MacTech Magazine, for Mac OS Programmers and Developers
     http://www.MacTech.com/
                                        _/   _/  _/  _/_/_/   _/   _/
   Chemistry: [e-mail]       _/_/ _/  _/  _/   _/  _/_/_/
     MacTech: [e-mail]      _/ _/_/  _/  _/       _/ _/
        http://www.chem.ucla.edu/~nick/   _/  _/   _/_/_/  _/   _/

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

From [e-mail] (Mauricio Tavares)
Date: 9 Oct 1996 17:54:20 GMT
Organization: University of Florida

Jason J Mullins ([e-mail]) wrote:
: kleinemans wrote:
: >
: > I'm a biginning macintosh programmer, and I'm looking for a cheap
: > pascal compiler.
: > (Don't begin about learning C pleasse)
: > Are there any shareware compilers? I'm now using borlands turbo
: > pascal('87)
:
: I don't know of a shareware compiler, but MetroWerks do a 'Bronze'
: version of CodeWarrier, which has Pascal, C, C++, etc, etc, and costs
: much less than the full version, and does 68k only. If you are a student
: then the prices on Academic CW are even better.
:
: The best bit is that with all options you get lots of electronic
: manuals, teaching books, the CW API and updates.
:
    Or he can get Symantec's Think Pascal...

--
 ===========================+==========================================
| Mauricio Tavares          | "We will attack...                       |
| [e-mail] |  ...under the cover of daylight!" Rimmer |
 ===========================+==========================================

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

From [e-mail] (Emil Tin)
Subject: Sound Manager delays!?
Date: 9 Oct 1996 11:17:55 GMT
Organization: Elektrodelia

Hi all,

I'm playing a sound, with real time control over volume, filters etc. But
I'm having troubles with the delays apparently build into Sound Manager.

I'm using the double buffer technique, where I can control the buffer size
myself.

BUT:
It appears that the SM internally uses its own package size, independetly
of the size I use. (This is not very strange, since SM services many
applications.)
By some experimentation, I found that this size is actually 1056 sample points.

This is a problem, since it means that all real time control will have
MINIMUM DELAY OF 1056 SAMPLE POINTS! This is actually more that 20ms,
which is too long for serius musical use.


And this is the best case. I meassued the time interval between the SM
calling my sound routine (by using the Microseconds() call). I found that
if I used a buffer size of excactly 1056, I would get almost even
intervals like:
23000
23000
24000
23000
24000
etc.

If i used smaller buffer sizes, i would get something like:
600
23000
600
600
22000
600
23000
600
23000
600
600
600
23000

This shows that if I set my buffer size to a smaller value than 1056, SM
will not call my sound callback routine at a regular (quite fast)
interval, but rather wait 1056 sample points, and then call my routine a
number of times very rapidly, as to fill up the internal big (1056) sample
point buffer. Thus I have a delay thats still at least 1056 sample points,
and therefore win nothing.

Is there any way to get faster real time control over the sound output???????

Small question:Is there any way to meassure, at a given time, the
corresponding currentyl playing sample point of the currently playing
sound package?? (needed when I must wrap a change to the next sound
package.)


Thanks in advance,
   Emil Tin

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

From [e-mail] (Emil Tin)
Date: 10 Oct 1996 10:33:09 GMT
Organization: Elektrodelia

Thanks a lot for the answers. A couple of things..

> > It appears that the SM internally uses its own package size, independetly
> > of the size I use. (This is not very strange, since SM services many
> > applications.)
> > By some experimentation, I found that this size is actually 1056 sample
> points.
>
> This is NOT the Sound Mgr, but the hardware's internal buffer. The Sound
> Mgr has no control over this. You'd need to change the hardware design
> and/or software support (e.g. low level drivers).

Does this mean, that I could theoretically write my own sound driver and
get better respons times?


> > Small question:Is there any way to meassure, at a given time, the
> > corresponding currentyl playing sample point of the currently playing
> > sound package?? (needed when I must wrap a change to the next sound
> > package.)
>
> I think you're asking for level metering of output, like what's already
> supported with input. I'd like that too. It's one of the feature requests
> on my list. I'd rather have this supported in hardware, but I'll see what
> I can do.

No, let me be more excact:

The hardware starts playing back its internal buffer (1056 sample pts) at
t=0 time units.
Lets say this buffer takes 100 time units to play back.

If a real time controller appears at t=30 time units, it can obviously not
have any effect on the currently played back buffer, since that buffer has
already been calculated.

Now there's two possibilities:
A     It can have effect from the first sample point in the next package.
(at t=100)
B     It can have effect at the time it appeared + 100 time units (at t=130)


A is not very good, since it introduces a variable (random) delay ranging
from 0 to 100 time units.
B is better, since it always introduces the same delay (100 time units)

But in order to use possibility B, there must be a way to know which time
(or at which sample point) the controller appeared, relative to time the
buffer started playing. (Toolbox call...?!)

Alternativly, it would suffice to know the excact time at which the buffer
started playing back the first sample point of the current package.


Thanks,
   Emil Tin

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

From Sean sharkey <[e-mail]>
Date: Thu, 10 Oct 96 10:36:12 -0000
Organization: EXP Ltd


In article <[e-mail]>, Emil Tin
writes:

>
> Hi all,
>
> I'm playing a sound, with real time control over volume, filters
etc. But
> I'm having troubles with the delays apparently build into Sound
Manager.
>
> I'm using the double buffer technique, where I can control the
buffer size
> myself.

Sound manager doesn't set any buffer size. It's the built-in hardware
for your Mac that's got the buffer. Even worse perhaps is that the
size of the buffer varies from Mac to Mac. For example my LC475 has a
hardware buffer of 512 byte and my PowerPC has 1024.

You can get the size of the hardware buffer by calling,

osErr = SPBGetDeviceInfo (inRefNum, siDeviceBufferInfo, InfoData)

See Inside Mac "Sound Input Manager," for more details.

 This shows that if I set my buffer size to a smaller value than 1056,
SM
> will not call my sound callback routine at a regular (quite fast)
> interval, but rather wait 1056 sample points, and then call my
routine a
> number of times very rapidly, as to fill up the internal big (1056)
sample
> point buffer. Thus I have a delay thats still at least 1056 sample
points,
> and therefore win nothing.
>
> Is there any way to get faster real time control over the sound
output???????

If i were you I would switch to using an Interrupt Routine then
during
the interrupt  move the sample data from the hardware buffer into your
apps buffer.


>
> Small question:Is there any way to meassure, at a given time, the
> corresponding currentyl playing sample point of the currently
playing
> sound package?? (needed when I must wrap a change to the next sound
> package.)

Yes you can get the meter level by calling,

osErr = SPBGetDeviceInfo (inRefnum, siLevelMeterOnOff, infoData)

once again see Inside Mac for the details.

Hope this helps,


Sean.

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

From [e-mail] (Jim Reekes)
Date: Wed, 09 Oct 1996 16:26:28 -0700
Organization: Apple Computer, Inc.

In article <[e-mail]>, [e-mail]
(Emil Tin) wrote:

> I'm using the double buffer technique, where I can control the buffer size
> myself.
>
> BUT:
> It appears that the SM internally uses its own package size, independetly
> of the size I use. (This is not very strange, since SM services many
> applications.)
> By some experimentation, I found that this size is actually 1056 sample
points.

This is NOT the Sound Mgr, but the hardware's internal buffer. The Sound
Mgr has no control over this. You'd need to change the hardware design
and/or software support (e.g. low level drivers).

> This is a problem, since it means that all real time control will have
> MINIMUM DELAY OF 1056 SAMPLE POINTS! This is actually more that 20ms,
> which is too long for serius musical use.

I agree, but what is acceptable for "serius [sic] musical use?"

> And this is the best case. I meassued the time interval between the SM
> calling my sound routine (by using the Microseconds() call). I found that
> if I used a buffer size of excactly 1056, I would get almost even
...
...
...
> This shows that if I set my buffer size to a smaller value than 1056, SM
> will not call my sound callback routine at a regular (quite fast)
> interval, but rather wait 1056 sample points, and then call my routine a
> number of times very rapidly, as to fill up the internal big (1056) sample
> point buffer. Thus I have a delay thats still at least 1056 sample points,
> and therefore win nothing.

You could have just used the siHardwareFormat selector and found out the
internal hardware's format, which includes the buffer size. You should try
to make your buffers a multiple of the hardware's format. (By the way, I
have NOTHING to do with the hardware design. Do you think I'd choose
1056?)

> Is there any way to get faster real time control over the sound output???????

Not at the present time.

> Small question:Is there any way to meassure, at a given time, the
> corresponding currentyl playing sample point of the currently playing
> sound package?? (needed when I must wrap a change to the next sound
> package.)

I think you're asking for level metering of output, like what's already
supported with input. I'd like that too. It's one of the feature requests
on my list. I'd rather have this supported in hardware, but I'll see what
I can do.

Jim

--
Jim Reekes, Polterzeitgeist |       QuickTime Products R&D
                            |        Sound Manager Expert
Apple Computer, Inc.        | "All opinions expressed are mine, and
2 Infinite Loop  MS 302-3KS |  do not necessarily represent those
Cupertino, CA 95014         |  of my employer, Apple Computer Inc."

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

From [e-mail] (Mark Coniglio)
Date: Fri, 11 Oct 1996 09:17:30 -0400
Organization: Troika Ranch

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

> You can get the size of the hardware buffer by calling,
>
> osErr = SPBGetDeviceInfo (inRefNum, siDeviceBufferInfo, InfoData)
>
> See Inside Mac "Sound Input Manager," for more details.

This is the _input_ buffer, no? (the "si" is for sound input.)

-- Mark

--
mail: [e-mail]
web:  http://www.art.net/~troika

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

From Murat Konar <[e-mail]>
Date: Fri, 11 Oct 1996 12:24:42 -0700
Organization: Macromedia

Sean sharkey wrote:
>
> In article <[e-mail]>, Emil Tin
> writes:
[deletia]
> > Small question:Is there any way to meassure, at a given time, the
corresponding currentyl playing sample point of the currently
playing sound package?? (needed when I must wrap a change to the next
sound  package.)
>
> Yes you can get the meter level by calling,
>
> osErr = SPBGetDeviceInfo (inRefnum, siLevelMeterOnOff, infoData)
>

This will give you input level metering.  To get output level metering,
we made a sound output component whose sole purpose was to observe the
samples as they went by on their way to the hardware.  Still some
latency using this technique, but not bad.


--
_______________________________________________________________
Murat N. Konar
SoundEdit 16 Guy
[e-mail]

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

From [e-mail] (LOTTSIM)
Subject: Strange Pop-Up Problem...
Date: 7 Oct 1996 21:03:56 -0400
Organization: America Online, Inc. (1-800-827-6364)

I have a pop up menu (and a defined 'CNTL') that is Appended when the
Dialog is first opened.  However, after that, I can't seem to call
SetCtlValue to set the pop-up to a value greater than how it comes
originally, with 4 items.

It just sets it so that a "-" item appears, and nothing is checked.  Why
would this be happening?

I've checked (and added error checking) so that I can see what number menu
item it is trying to set the menu to, and it IS valid, but it was
appended.  What am I doing wrong?

Thanks!

Alex Rampell

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

From [e-mail] (Roger Bilewicz)
Date: 8 Oct 96 12:37:19 GMT
Organization: Freie Universitaet Berlin

[e-mail] (LOTTSIM) writes:

>I have a pop up menu (and a defined 'CNTL') that is Appended when the
>Dialog is first opened.  However, after that, I can't seem to call
>SetCtlValue to set the pop-up to a value greater than how it comes
>originally, with 4 items.

>It just sets it so that a "-" item appears, and nothing is checked.  Why
>would this be happening?

>I've checked (and added error checking) so that I can see what number menu
>item it is trying to set the menu to, and it IS valid, but it was
>appended.  What am I doing wrong?

>Thanks!

>Alex Rampell

Hi Alex,

1. check if your DITL, CNTL, and MENU resources are ok;
2. give us a snippet of your source code for closer investigation.

So long,
Roger


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

From [e-mail] (Kevin Parichan)
Date: Tue, 08 Oct 1996 10:57:10 -0800
Organization: Best Internet Communications

In article <53c99s$[e-mail]>, [e-mail] (LOTTSIM) wrote:

>I have a pop up menu (and a defined 'CNTL') that is Appended when the
>Dialog is first opened.  However, after that, I can't seem to call
>SetCtlValue to set the pop-up to a value greater than how it comes
>originally, with 4 items.

Try calling SetControlMaximum() where the max is the new total of items in
the menu.

-Kevin Parichan

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

From Murat Konar <[e-mail]>
Date: Tue, 08 Oct 1996 11:23:40 -0700
Organization: Macromedia

[e-mail] (LOTTSIM) writes:

>I have a pop up menu (and a defined 'CNTL') that is Appended when the
>Dialog is first opened.  However, after that, I can't seem to call
>SetCtlValue to set the pop-up to a value greater than how it comes
>originally, with 4 items.

Presumabaly you are appending menu items to the popup?  If you don't
call SetCtlMax() to increase the control's maximum value before calling
SetCtlValue(), your new value will get "clipped" to the control's
current maximum value.



--
_______________________________________________________________
Murat N. Konar
SoundEdit 16 Guy
[e-mail]

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

From [e-mail] (LOTTSIM)
Date: 9 Oct 1996 13:47:19 -0400
Organization: America Online, Inc. (1-800-827-6364)

Everything works great as long as I'm setting the control (a pop up menu)
somewhere between 1 and 3 (4 is a "-" item).

However, when I set it to 5, 6, or anything greater (the pop-up is
generated on the fly, so it could have 100 items, theoretically).

Here's the code segment I use for setting the item:

GetDItem( myDialog, kMyMenu, &itemType, &itemHandle, &twoRect );


SetCtlValue( (ControlHandle)itemHandle,
    (*prefsHandle)->FunFile[current-kF1].refNumber+kMenuChoice);
                    //Add over four existing numbers

This gets executed only if it does belong in the bottom half of the menu
(kMenuChoice, equal to 4, is added to the refNum which is between 1 and
Count1Resources of a certain resource).

I've even checked the value of FunFile[current-kF1].refNumber+kMenuChoice,
but doing a for loop, beeping "refNumber+kMenuChoice" number of times.

FunFile is a struct I defined, within my prefs resource.

Here's how I make the menu:
myMenu = GetMenu(-4040);
    if ( myMenu )
    {
        HLock( (Handle)myMenu );
        SetResLoad(false);
        for ( x = 1; x <= Count1Resources('pRoC'); x++ )
        {
            itemHandle = Get1IndResource('pRoC', x);
            GetResInfo(itemHandle,&y,&rType,theString);
            AppendMenu(myMenu, theString);
        }
        HUnlock( (Handle)myMenu );
    }

I can view and select all the choices in this menu...I just can't have my
program set them to what *I* want.

Also, on another control related problem, I can't get SetDItem to work to
move a control (static text).  It just stays in the original position.
MoveControl does the same thing (nothing!)

Thanks,
Alex Rampell

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

From [e-mail] (Roger Bilewicz)
Date: 10 Oct 96 14:03:24 GMT
Organization: Freie Universitaet Berlin

[e-mail] (LOTTSIM) writes:

>Everything works great as long as I'm setting the control (a pop up menu)
>somewhere between 1 and 3 (4 is a "-" item).

>However, when I set it to 5, 6, or anything greater (the pop-up is
>generated on the fly, so it could have 100 items, theoretically).

>Here's the code segment I use for setting the item:
-- snip --

From this, my suggestion is to fix the control's maximum value first,
as Murat and Kevin already proposed. The contrlMax field of the Control
Record is set when the popup's MENU resource is loaded. After this, when
you append more menu items, the control record cannot be updated
automatically. You must do this by a call to SetCtlMax.
SetCtlValue checks the new value against the contrlMax field and
therefore behaves not as expected if the menu item count and the
control's maximum setting are not equal.

>Also, on another control related problem, I can't get SetDItem to work to
>move a control (static text).  It just stays in the original position.
>MoveControl does the same thing (nothing!)

A similar situation? The dialog manager calls DrawControl to draw the
controls, but DrawControl doesn't know about dialog item rects.
Therefore you should take care to have both the item rect and the
control's rect identical. MoveControl cannot set a dialog item rect, and
SetDItem cannot change a control's rect. Therefore we have to change
both rects to move a control in a dialog.

However, static text isn't a control. I think you confuse something
here. To change the place where static text is drawn, SetDItem should work.

Hope it helps,
Roger



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

From [e-mail] (Kjell Erling Hagen)
Subject: VIP-C and CodeWarrior 9
Date: Tue, 08 Oct 1996 17:10:00 +0200
Organization: Hiÿ

Just installed VIP-C 2.0 (payed for it!) on my PowerMac 8200/120 and went
through the tutorials. I have also CodeWarrior 9, and tried to compile
with it through the Project-menu in VIP-C.

The following operation went OK:
Set project type: Application, Self contained.

But the Update project operation gave this message:
"Could not update the project".
The error in CodeWarrior was "illegal precompiled header version",
reffering to the line "#include "MicroCalc.h" (a headerfile made by
VIP-C). And it tells me that the header file is old or defective. In the
CodeWarrior preferences Language Settings C/C++ Language the Prefix File
is "VIP_C_Headers_PPC".

I tried the same thing on a Macintosh IIfx, and the same thing happened
(another Prefix File).

What is the solution?

Mail me a copy of your answer/suggestion.

--
Kjell Erling Hagen, Borgergata 38, N-1767 Halden, Norway.
[e-mail]
http://www-ia.hiof.no/~kjeller/

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

From [e-mail] (MW Ron)
Date: Wed, 09 Oct 1996 23:50:53 -0400
Organization: Metrowerks

In article <[e-mail]>,
[e-mail] (Kjell Erling Hagen) wrote:

>Just installed VIP-C 2.0 (payed for it!) on my PowerMac 8200/120 and went
>through the tutorials. I have also CodeWarrior 9, and tried to compile
>with it through the Project-menu in VIP-C.
>
>The following operation went OK:
>Set project type: Application, Self contained.
>
>But the Update project operation gave this message:
>"Could not update the project".
>The error in CodeWarrior was "illegal precompiled header version",
>reffering to the line "#include "MicroCalc.h" (a headerfile made by
>VIP-C). And it tells me that the header file is old or defective. In the
>CodeWarrior preferences Language Settings C/C++ Language the Prefix File
>is "VIP_C_Headers_PPC".
>
>I tried the same thing on a Macintosh IIfx, and the same thing happened
>(another Prefix File).
>
>What is the solution?

I haven't used VIP-C but the fix generally for an error of this sort is to
create a new precompiled header of using the same items as were in the
original precompiled header.  You should be able to determine this from
the VIP-C items.  for example in Metrowerks MacHeaders.h is the result of
opening the file MacHeaders.pch  ( this file sets the nameing convention
and it includes MacHeaders.c  the true file to be precompiled)  so you
would do the equivalent of opening MacHeaders.pch and selecting precompile
from the project menu.

This is probably better documented in either the IDE User's guide or the
Targeting MacOS user's guide

Ron

--
METROWERKS                   Ron Liechty
"Software at Work"    [e-mail]
http://www.metrowerks.com/about/people/rogues.html#mwron

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

From [e-mail] (Paul J. Voogt)
Date: 11 Oct 1996 20:01:37 GMT
Organization: Euronet

In article <[e-mail]>,
[e-mail] (Kjell Erling Hagen) wrote:

>Just installed VIP-C 2.0 (payed for it!) on my PowerMac 8200/120 and went
>through the tutorials. I have also CodeWarrior 9, and tried to compile
>with it through the Project-menu in VIP-C.
>
>The following operation went OK:
>Set project type: Application, Self contained.
>
>But the Update project operation gave this message:
>"Could not update the project".
>The error in CodeWarrior was "illegal precompiled header version",
>reffering to the line "#include "MicroCalc.h" (a headerfile made by
>VIP-C). And it tells me that the header file is old or defective. In the
>CodeWarrior preferences Language Settings C/C++ Language the Prefix File
>is "VIP_C_Headers_PPC".
>
>I tried the same thing on a Macintosh IIfx, and the same thing happened
>(another Prefix File).
>
>What is the solution?
>
>Mail me a copy of your answer/suggestion.

You have to create new precompiled headers. To do this look inside the
"VIP" folder that is located in the CodeWarrior folder:

Two Metrowerks project files are provided to generate the VIP-C 2.0
precompiled header files automatically: "VIP_C_Headers.micro68K" and
"VIP_C_Headers.microPPC" ('micro' is the micro symbol which looks like a
'u'). A minor modification must be made to a Universal Header File in order
to generate the new files. This modification doesn't create any conflict
and is necessary to generate the files for the VIP-C 2.0 environment.
Proceed as follows for the update:
1) Open the "OCEStandardMail.h" file which is located in the "Universal
Headers" folder in the CW9 environment.
2) Search for the word "pageNumber" in this file, and replace it by
"pageNum".
3) Close and save the updated "OCEStandardMail.h" file.
4) Open the "VIP_C_Headers.micro68K" project file, select the file in
the project and select the "Precompile..." item from the "Project" menu.
5) Repeat step #4 for the "VIP_C_Headers.microPPC" project file.
6) That's it!

--
Paul J. Voogt

http://www.euronet.nl/users/pjvoogt/               mailto:[e-mail]
- --------------------------------------------------------------------------
I have seen the future and it is just like the present, only longer.


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

From [e-mail]
Subject: Where is TMON Pro?
Date: Mon, 30 Sep 1996 16:52:25 -0800
Organization: NETCOM Network Operations


Does anyone know what happened to TMON (this time).  I paid my fees when
MindVision took over but have not heard a thing since.

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

From [e-mail] ( nick.c @MT )
Date: Mon, 30 Sep 1996 18:58:32 -0800
Organization: MacTech Magazine



[e-mail] wrote:

>Does anyone know what happened to TMON (this time).  I paid my fees when
>MindVision took over but have not heard a thing since.


   Dunno, but I have some contact info for them.  I'd suggest
     giving 'em a call:

         [phone] (phone)
         [phone] (fax)

         <http://www.mindvision.com/>
         <mailto:[e-mail]>

____Nicholas C. DeMello, Ph.D.________________________________________

   Online from MacTech Magazine, for Mac OS Programmers and Developers
     http://www.MacTech.com/
                                       _/   _/  _/  _/_/_/   _/   _/
  Chemistry: [e-mail]       _/_/ _/  _/  _/   _/  _/_/_/
    MacTech: [e-mail]      _/ _/_/  _/  _/       _/ _/
       http://www.chem.ucla.edu/~nick/   _/  _/   _/_/_/  _/   _/

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

From [e-mail] (Jonathan Oei)
Date: Fri, 04 Oct 1996 12:30:32 -0400
Organization: HOPS, International

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

>Does anyone know what happened to TMON (this time).  I paid my fees when
>MindVision took over but have not heard a thing since.

TMON is currently at revision 4.0b2 and you can purchase the beta
(or, if you currently own an older TMON, purchase a subscription).

Long live TMON!

--
Jonathan Oei                                      [e-mail]

"Christianity has not been tried and found wanting; it has been
 found difficult and not tried" -- G.K. Chesterton

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

From [e-mail]
Date: Mon, 07 Oct 1996 08:43:56 -0800
Organization: NETCOM Network Operations



I did purchase it back when MindVision took over and got the 4.0b1 version
which was unusable.  I have not seen anything since.


In article <[e-mail]>,
[e-mail] (Jonathan Oei) wrote:
>
> TMON is currently at revision 4.0b2 and you can purchase the beta
> (or, if you currently own an older TMON, purchase a subscription).
>
> Long live TMON!
>
> --
> Jonathan Oei                                      [e-mail]
>
> "Christianity has not been tried and found wanting; it has been
>  found difficult and not tried" -- G.K. Chesterton

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

From Bob Meyer <[e-mail]>
Date: Tue, 08 Oct 1996 08:52:49 -0400
Organization: Meyer Software

> > TMON is currently at revision 4.0b2 and you can purchase the beta
> > (or, if you currently own an older TMON, purchase a subscription).

could someone tell me where/how to get TMON??

old faithful programmer that misses tmon.

(aka bob meyer)

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

From [e-mail] (Terry Gritton)
Date: Wed, 09 Oct 1996 09:53:56 -0700
Organization: ^self -> (CompSci/MolBiol)

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

>> > TMON is currently at revision 4.0b2 and you can purchase the beta
>> > (or, if you currently own an older TMON, purchase a subscription).
>
>could someone tell me where/how to get TMON??
>
http://www.mindvision.com/mvweb/Products/TMON/
--
--
Terry Gritton        "Glycobiology - the new frontier of biosemiotics"
[e-mail]

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

From [e-mail] (MW Ron)
Subject: [ANN] CodeWarrior IDE 1.7.1  is available for downloading
Date: Fri, 04 Oct 1996 23:46:52 -0400
Organization: Metrowerks

CodeWarrior IDE 1.7.1  is available for downloading

now or soon from the following URLS

http://www.metrowerks.com/tools/software/updates/current
or
ftp://ftp.metrowerks.com/pub/updates/CW10/
European Mirror Site
ftp://ftp.promo.de/pub/Metrowerks/mirror/updates/CW10/


=============================
CodeWarrior IDE Release Notes
=============================

Version: 1.7.1
Date:    October 1, 1996

This is a minor update to the 1.7 IDE on the CW10 CodeWarrior CD. In
response to some of the early feedback we've received, this update
modifies the editor window to remove the gray border on either side of
the text and slightly reduces the height of the button bar. We expect
to release a more comprehensive netborne update in late October.

Please address any questions, feedback, or bug reports to
[e-mail].

--
METROWERKS                   Ron Liechty
"Software at Work"    [e-mail]
http://www.metrowerks.com/about/people/rogues.html#mwron

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

From [e-mail] (MW Ron)
Date: Sat, 05 Oct 1996 10:44:51 -0400
Organization: Metrowerks

In article <[e-mail]>,
[e-mail] (MW Ron) wrote:

>CodeWarrior IDE 1.7.1  is available for downloading

I should mention this is a purely cosmetic update and if you are happy
with the look now there is no need to download this.

Ron

--
METROWERKS                   Ron Liechty
"Software at Work"    [e-mail]
http://www.metrowerks.com/about/people/rogues.html#mwron

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

From [e-mail] (Mike Cohen)
Date: Sat, 05 Oct 1996 11:54:28 -0600
Organization: ISIS International

In article <[e-mail]>,
[e-mail] (MW Ron) wrote:

>This is a minor update to the 1.7 IDE on the CW10 CodeWarrior CD. In
>response to some of the early feedback we've received, this update
>modifies the editor window to remove the gray border on either side of
>the text and slightly reduces the height of the button bar. We expect
>to release a more comprehensive netborne update in late October.

ABSOLUTELY PERFECT! I downloaded it yesterday and it's WONDERFUL. The
windows look & act exactly the way they're supposed to and no longer waste
space for a useless decorative border.

Thank you!


[ Mike Cohen | [e-mail] | http://pobox.com/~djembe (personal) ]
[            | [e-mail]   | http://www.isis-intl.com/ (company) ]
[     Harry Browne for President - http://www.harrybrowne96.org       ]


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

From [e-mail] (Andrew)
Date: 9 Oct 1996 23:34:31 GMT
Organization: Apple Computer, Inc.

In article <[e-mail]>, [e-mail] (Brett
Wayne) wrote:

> The patch to the debugger has been available since last Thursday which
> means you guys had to have known about this for awhile - yet there
> has been no mention of it in the newsgroup.

I read such a message late last week (friday or saturday) about
the bugs and the first patch.  They said a more complete patch
was due soon.

Perhaps you missed it amongest all the posts complaining about
the new IDE look?

>
> I'm angry because I just wasted the better part of an afternoon
> assuming I had an INIT conflict, or possibly a corrupted system. I
> finally gave up and deinstalled CW10 only to find out that my project
> files were now toast and had to be recovered off the backup disks.
> All a _huge_ waste of my time.

CW10 saves your old project files.  It keeps the old project and
appends the word 'copy' to the file name.  I've often used this
feature to go back and forth using shared source yet having a
CW9 and CW10 project.



>
> You can be damned sure I won't be so eager to install CW10 again,
> or for that matter, any future CW releases. I am also seriously
> considering not renewing my licenses again until I see a substantive
> change in the product.


Bugs are a fact of life and I don't think I've ever seen a
devlopemnt environment updated as quickly as MW has done
with CW.  Metrowerks is to be praised for their reactions to
customer complaints.

Oh well...

--
Andrew Carol               "Could be worse.  Could be raining."
[e-mail]         [e-mail]

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

From [e-mail] (Jon Guyer)
Date: Wed, 09 Oct 1996 12:51:16 -0500
Organization: Northwestern University, Evanston, IL

In article <[e-mail]>, [e-mail] (Brett
Wayne) wrote:

> More importantly...how about acknowledging that there was some serious
> problems with the debugger and announcing that a patch exists for that
> as well?

You mean like article <[e-mail]>,
"[ANN ]MWDebug/MacOS 1.6.3 is available for downloading", which was posted
Fri, 04 Oct 1996 23:47:35 -0400... forty-three seconds after the article
you just replied to?

> The patch to the debugger has been available since last Thursday which
> means you guys had to have known about this for awhile - yet there
> has been no mention of it in the newsgroup.

What newsgroup would that be? comp.sys.mac.programmer had threads running
back at least a week prior to the patch release, discussing problems with
the debugger, the fact that MW knew about it, and that they'd get the fix
out as soon as possible. A beta was available a few days before if you
contacted support directly.

Other than not having the bug in the first place, I don't see how MW could
have handled this any better. If you really didn't see any of this, your
beef should be with your ISP, not MW.

--
Jonathan E. Guyer

Join LMESOTENPAMSNFZ

("Let's make everything south of the 89th parallel
  a Microsoft no-fly-zone" AKA "Give Bill the Pole!")

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

From [e-mail] (Mike Cohen)
Date: Wed, 09 Oct 1996 11:13:18 -0600
Organization: ISIS International

In article <[e-mail]>,
[e-mail] (Brett Wayne) wrote:

>The patch to the debugger has been available since last Thursday which
>means you guys had to have known about this for awhile - yet there
>has been no mention of it in the newsgroup.

It *was* mentioned here, with a complete description of the bugs it fixed.
You probably missed that one because it didn't inspire the flurry of
follow-ups this message did.


[Mike Cohen | [e-mail] | http://pobox.com/~djembe (personal)]
[           | [e-mail]   | http://www.isis-intl.com/ (company)]
[     Harry Browne for President - http://www.harrybrowne96.org     ]

*** JUNK MAIL NOT ACCEPTED - I CHANGED MY RETURN ADDRESS SO SPAMMERS
*** CAN'T GET MY EMAIL ADDRESS FROM MY POSTS. IF YOU WISH TO REPY,
*** COPY MY EMAIL ADDRESS FROM MY SIGNATURE

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

From [e-mail] (Bruce Toback)
Subject: [Q] Reliable notification from completion routine
Date: Sun, 6 Oct 1996 03:36:20 GMT
Organization: NETCOM On-line Communication Services (408 261-4700 guest)

Hi all,
Try as I might, I can't seem to find the answer to this:

I'm using Device Manager calls -- MacTCP in this case -- with asynchronous
completion routines. According to the documentation, these routines
execute at interrupt time, so they can't do anything that might
move memory, nor can they rely on access to the application's globals.
Fair enough.

But how do I reliably notify the application of an IO completion and
schedule execution? If I call WakeUpProcess(), I might call just
after the application has checked for completions but before it's
gone to sleep. Posting an event would work, but the docs say
not to post anything except mouse-up, mouse-down, key-up, key-down
and disc-inserted events. And there appears to be no way for the
application to do a "check-this-and-go-to-sleep" as an atomic operation.
The only thing I can think of is posting a key-up event with an illegal
parameter, but this seems like a real kludge.

Is there some "official" way to do this? Or do I kludge up something with
PostEvent?

Please email replies if possible, since my news access is sporadic.

Thanks,
-- Bruce Toback
[e-mail]



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

From [e-mail] (Chris Newman)
Date: Sun, 06 Oct 1996 01:18:43 -0700
Organization: Altopia Corp. - Affordable Usenet Access - http://www.alt.net

In article <[e-mail]>, [e-mail] (Bruce
Toback) wrote:
>I'm using Device Manager calls -- MacTCP in this case -- with asynchronous
>completion routines. According to the documentation, these routines
>execute at interrupt time, so they can't do anything that might
>move memory, nor can they rely on access to the application's globals.
>Fair enough.
>
>But how do I reliably notify the application of an IO completion and
>schedule execution? If I call WakeUpProcess(), I might call just
>after the application has checked for completions but before it's
>gone to sleep. Posting an event would work, but the docs say
>not to post anything except mouse-up, mouse-down, key-up, key-down
>and disc-inserted events. And there appears to be no way for the
>application to do a "check-this-and-go-to-sleep" as an atomic operation.
>The only thing I can think of is posting a key-up event with an illegal
>parameter, but this seems like a real kludge.
>
>Is there some "official" way to do this? Or do I kludge up something with
>PostEvent?

The best way that I know of to do this is to have a single flag (e.g.
iotaskready) which you check right before your WaitNextEvent() call, and if
set, you set the timeout to immediate.  You can access this flag through
the userDataPtr to avoid globals.  Granted this leaves a tiny window
between the flag check and the WNE(), but there's no way around this and it
doesn't seem to be a problem in practice.  One friend of mine once asked
about this through Apple's offical tech support and they never understood
the question properly.

Feel free to check out the source code for NiftyTelnet at:
      <http://andrew2.andrew.cmu.edu/dist/niftytelnet.html>

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

From [e-mail] (Ben Trumbull)
Date: Sun, 06 Oct 1996 21:39:26 -0400
Organization: Yale University

In article <[e-mail]>, [e-mail] (Bruce
Toback) wrote:


> But how do I reliably notify the application of an IO completion and
> schedule execution? If I call WakeUpProcess(), I might call just
> after the application has checked for completions but before it's
> gone to sleep. Posting an event would work, but the docs say
> not to post anything except mouse-up, mouse-down, key-up, key-down
> and disc-inserted events. And there appears to be no way for the
> application to do a "check-this-and-go-to-sleep" as an atomic operation.
> The only thing I can think of is posting a key-up event with an illegal
> parameter, but this seems like a real kludge.
>
> Is there some "official" way to do this? Or do I kludge up something with
> PostEvent?

There is an example with sample code for using the Thread Manager (and
threads ;-) to do just this.  If you don't want to use threads, you could
just have your app loop around a WaitNextEvent with a long sleep
paramater.  Inside the loop your app could check if your completion
routine changed the value of some global variable or whatever.

terminally curious,

Ben
___________________________________________________________________
Benjamin Trumbull
[e-mail]
Yale University
       You can't be in hell; you can still read your e-mail

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

From [e-mail] (Jump Long)
Date: 10 Oct 1996 01:07:13 -0400
Organization: America Online, Inc. (1-800-827-6364)

Bruce Toback wrote:
>>I'm using Device Manager calls -- MacTCP in this case -- with
asynchronous
>>completion routines. According to the documentation, these routines
>>execute at interrupt time, so they can't do anything that might
>>move memory, nor can they rely on access to the application's globals.
>>Fair enough.
>>
>>But how do I reliably notify the application of an IO completion and
>>schedule execution? If I call WakeUpProcess(), I might call just
>>after the application has checked for completions but before it's
>>gone to sleep. Posting an event would work, but the docs say
>>not to post anything except mouse-up, mouse-down, key-up, key-down
>>and disc-inserted events. And there appears to be no way for the
>>application to do a "check-this-and-go-to-sleep" as an atomic operation.
>>The only thing I can think of is posting a key-up event with an illegal
>>parameter, but this seems like a real kludge.
>>
>>Is there some "official" way to do this? Or do I kludge up something
with
>>PostEvent?

Chris Newman wrote:
>The best way that I know of to do this is to have a single flag (e.g.
>iotaskready) which you check right before your WaitNextEvent() call, and
if
>set, you set the timeout to immediate.  You can access this flag through
>the userDataPtr to avoid globals.  Granted this leaves a tiny window
>between the flag check and the WNE(), but there's no way around this and
it
>doesn't seem to be a problem in practice.  One friend of mine once asked
>about this through Apple's offical tech support and they never understood
>the question properly.

They obviously didn't ask the right question, because I answered this
question correctly many times during the time I worked in DTS. Here's what
I told developers who asked:

Before System 7.1.2, WakeUpProcess doesn't do anything if the process
being targeted is the active process and is executing the code in its
event loop when WakeUpProcess is called. There are a three things you can
do to work around this problem.

First, in your interrupt time code, change value of the sleep parameter
variable your application passes to WaitNextEvent to 0. That way, your
next call to WaitNextEvent will give up as little time as possible.

Second, check the sleep parameter variable just before you call
WaitNextEvent and if it is zero (meaning your interrupt time code has
changed it), recheck the data you're expecting from the interrupt time
code. That should help to prevent WaitNextEvent from being called when you
don't need it to be called.

Third, for the sleep value you normally pass to WaitNextEvent while you
have a connection open, use a value that's small enough (for example 15 -
30 ticks) so that even if you hit the very small race window left by the
first two suggestions above, your overall performance won't be affected
very much.

With System 7.1.2 and later, the WakeUpProcess request is noted by the
Process Manager and the next time your application calls WaitNextEvent,
the Process Manager ignores the sleep value and handles the request as if
you'd passed 0.

- Jim Luther (former Apple Developer Technical Support engineer and proud
of it!)


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

From [e-mail]
Subject: is MacOS multiplatform?
Date: Wed, 9 Oct 1996 22:10:42 GMT
Organization: MIND LINK! - British Columbia, Canada

I'm wondering if any of the versions of MacOS are multiplatform?
If so, can any of the multi-platform versions of MacOS run on an Intel CPU?
thanx

/-----------------------------------------------------------------------\
| "a Union should be a way to store more than one memory type in the    |
| same merory location, not a way for filthy employees to squeeze every |
| last penny out of their employer"                                     |
+-----------------------------------------------------------------------+
| Michael Celli                                |
|            email: [e-mail]            |
\-----------------------------------------------------------------------/

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

From [e-mail] ( nick.c @MT )
Date: Thu, 10 Oct 1996 15:24:34 -0700
Organization: MacTech Magazine



[e-mail] wrote:

>I'm wondering if any of the versions of MacOS are multiplatform?
>If so, can any of the multi-platform versions of MacOS run on an Intel CPU?


   Not really.  But there is a port that allows you to run a MacOS
     (sort of) on a UNIX box.  'Course there are Window's emulators
     for the Mac, so you can run Windows apps on a Mac.




____Nicholas C. DeMello, Ph.D.___________________________________________
"MacTech Online"--MacTech Magazine, for Mac OS Programmers and Developers
     http://www.MacTech.com/
                                        _/   _/  _/  _/_/_/   _/   _/
   Chemistry: [e-mail]       _/_/ _/  _/  _/   _/  _/_/_/
     MacTech: [e-mail]      _/ _/_/  _/  _/       _/ _/
        http://www.chem.ucla.edu/~nick/   _/  _/   _/_/_/  _/   _/

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

From [e-mail] (Jordan Zimmerman)
Date: Thu, 10 Oct 1996 19:08:34 -0800
Organization: Altura Software, Inc.

In article <[e-mail]>, [e-mail] (
nick.c @MT ) wrote:

> [e-mail] wrote:
>
> >I'm wondering if any of the versions of MacOS are multiplatform?
> >If so, can any of the multi-platform versions of MacOS run on an Intel CPU?

Ardi software (http://www.ardi.com) has something called "Executor" that
allows some Mac binaries to run on Windows.

--
Jordan Zimmerman
Altura Software, Inc.
home page: http://www.altura.com/jordanz

Harry Browne for President!
http://www.harrybrowne96.org   1 [phone]
Stop the Browne Out! Let Harry Browne debate!     http://www.twr.com/STBO

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

From [e-mail] (Tommy K. Hwang)
Date: 12 Oct 1996 22:30:35 GMT
Organization: Internet Solutions Inc.

In article <jordanz-1010961908490001@204.147.232.52>,
Jordan Zimmerman <[e-mail]> wrote:
>> >I'm wondering if any of the versions of MacOS are multiplatform?
>> >If so, can any of the multi-platform versions of MacOS run on an Intel CPU?
>
>Ardi software (http://www.ardi.com) has something called "Executor" that
>allows some Mac binaries to run on Windows.

    I personally had more problems with Executor than any other emulations
I ran (not many, I can assure you).  If you are not in a hurry, then a PPC
system that is PPCP compatible will be your best bet.  Unfortunately, those
systems are not out yet.  Good and bad since by the time the PPCP systems
are out, your Intel box will be 'out of date' anyway.  MacOS, as it exists
today, is NOT really multi-platform eventhough there are rumors of a beta
version of the PPCP MacOS floating around.

                                -ME

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

From [e-mail] (A. Bertrand)
Subject: parent class method called instead of overloaded one!
Date: 9 Oct 1996 19:31:37 GMT
Organization: Francenet -- Paris, France

I have the parent class defined as below :
class TDocument
{ public :
   TDocument();
   virtual void DoAutoKey();
   virtual void DoButton();
   virtual void DoChgInField();
   virtual void DoCmd(short theCmd);
   virtual void DoClick();
....
}
a derived class defined as below :
class TBezierSectionDocument:public TDocument
{  public:
   TBezierSectionDocument();
   virtual void DoCmd(short theCmd);
   virtual void SetBasicData(void *Ptr);
....
   protected:
some data

}
But when I try to call the overloaded DoCmd(short theCmd) I always get the
parent class mÈthod. I really can't find out where I'm wrong.
The  begginnning of the constructor is like that :
TBezierSectionDocument::TBezierSectionDocument():TDocument()
{  Str255 WindowTitle;
   GetIndString(WindowTitle,rCurveDialStringID,rCurveDialTitle);
   p2cstr(WindowTitle);
   fMainWindow = gMyApp->DoNewWindow(tCurveDialWind,(char*)WindowTitle,this);
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^®
And I do get the right window so the good constructor is called.
I'm new to C++ but not to object programming. I'm pretty sure that with
object pascal I would get the overloaded method!!!
I triied to trace with the debugger and it seems that the descendant class
is damaged by the compiler because though I initialize  its extra data in
the constructor, they appear randomly modified. This modification seems to
be quite fast because if I do the following:

TBezierSectionDocument *theTBSDocument;
    switch (theDocType)
     {  case tSectionBezierDoc   :
         {  theTBSDocument=  new  TBezierSectionDocument();
            theTBSDocument->DoCmd(0);
I do get the parent method .
I have other class build and inherited a la object pascal and they do work well.
I use ToolsPlus. Already heard of such problems incurring from it?

Any help would be _greatly_ appreciated


TIA

--
Alain Bertrand
[e-mail]

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

From [e-mail] (Bretton Wade)
Date: Thu, 10 Oct 1996 11:49:26 -0400
Organization: qualia, inc.

In article <[e-mail]>,
[e-mail] (A. Bertrand) wrote:

# But when I try to call the overloaded DoCmd(short theCmd) I always get the
# parent class mÈthod. I really can't find out where I'm wrong.

Sounds like you are calling DoCmd from the base class constructor. If so,
the behavior you are getting is correct C++. While the base class is
constructing, the object type is that of the base class, thus any virtual
functions called will belong to the base class, and not any inherited
classes. After construction is complete, then any call to the virtual
function will yield the inherited class version.

If you really need to call DoCmd during initialization, the usual practice
is to add a member function called 'initialize', which you call after the
object is constructed.

--
Bretton Wade ([e-mail])
http://www.qualia.com/~bwade/

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

From [e-mail] (A. Bertrand)
Date: 10 Oct 1996 23:02:19 GMT
Organization: Francenet -- Paris, France

In article (Dans l'article) <[e-mail]>,
[e-mail] (Bretton Wade) wrote (Ècrivait)†:

> In article <[e-mail]>,
> [e-mail] (A. Bertrand) wrote:
>
> # But when I try to call the overloaded DoCmd(short theCmd) I always get the
> # parent class mÈthod. I really can't find out where I'm wrong.
>
> Sounds like you are calling DoCmd from the base class constructor. If so,
> the behavior you are getting is correct C++. While the base class is
> constructing, the object type is that of the base class, thus any virtual
> functions called will belong to the base class, and not any inherited
> classes.

Hi,

Thanks a lot for your help, but the problem seems to come from elsewhere.

After constructing my object by
theTBSDocument= new TBezierSectionDocument();
I can call
theTBSDocument->Initialize() which is a method that doesn't belong to the
parent class. I get it.
If I call
theTBSDocument-> DoCmd( theCmd) I get the parent class method.

It seems like the jump table for this object is wrongly built. So why this
one and not the other ? I changed the name of the class but no succes.

Regards,

--
Alain Bertrand
[e-mail]
Ce matin est venu mon medecin. Il s'appelle Werner mais c'est un Russe.Qu'y a-t-il d'etonnant a cela? Je connaissais un Ivanov qui etait allemand. (Lermontov)

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

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