Tricks of the Mac Game Programming Gurus/TricksOfTheMacGameProgrammingGurus.iso/Information/CSMP Digest/volume 1/csmp-v1-204.txt

Text File | 1994-12-08 | 45.0 KB | 1,146 lines | [TEXT/R*ch]

C.S.M.P. Digest             Mon, 02 Nov 92       Volume 1 : Issue 204

Today's Topics:

    SCSI Disk Format & Initialization
    Launching from a document file
    Finding system folder in system 6
    Double Click in Window Content
    Bug fixes in MPW C 3.2.4
    Think Pascal and ResEdit questions



The Comp.Sys.Mac.Programmer Digest is moderated by Michael A. Kelly.

The digest is a collection of article threads from the internet newsgroup
comp.sys.mac.programmer.  It is designed for people who read c.s.m.p. semi-
regularly and want an archive of the discussions.  If you don't know what a
newsgroup is, you probably don't have access to it.  Ask your systems
administrator(s) for details.  (This means you can't post questions to the
digest.)

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
cs.uoregon.edu).  Article threads are not added to the digest until the last
article added to the thread is at least one month 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 entire digest is available for anonymous ftp from ftp.cs.uoregon.edu
[128.223.8.8] in the directory /pub/mac/csmp-digest.  Be sure to read the
file /pub/mac/csmp-digest/README before downloading any files.  The most
recent issues are available from sumex-aim.stanford.edu [36.44.0.6] in the
directory /info-mac/digest/csmp.  If you don't have ftp capability, the sumex
archive has a mail server; send a message with the text '$MACarch help' (no
quotes) to [e-mail] for more information.

The digest is also available via email.  Just send a note saying that you
want to be on the digest mailing list to [e-mail], and you will
automatically receive each new issue as it is created.  Sorry, back issues
are not available through the mailing list.

Send administrative mail to [e-mail].


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

From: [e-mail] (Jody Kravitz)
Subject: SCSI Disk Format & Initialization
Organization: The Foxtail Group
Date: Sun, 27 Sep 1992 17:16:11 GMT

Could someone explain to me (or point me to the appropriate CURRENT
reference) how it is decided if a SCSI disk is a valid Mac disk.

I have used a shareware progrogram SCSI Format&Init to initialize a 40 Meg
drive.  However I find that I must use the 'mountem' FKEY to get it to
mount after a reboot.  This leads me to wonder if the data structures
put on the disk by SCSI Format&Init are current.

Thanks

Jody

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

From: [e-mail] (Martin Minow)
Date: 28 Sep 92 17:43:30 GMT
Organization: Apple Computer Inc., Cupertino, CA

In article <[e-mail]> [e-mail]
(Jody Kravitz)  notes that a scsi disk will not automount.

The Macintosh SCSI implementation predates ANSI-SCSI by about four years.
Unfortunately, this means that there are a few incompatibilities between
the Mac and "standard" SCSI. Jody is probably being bitten by a requirement
(in "standard" SCSI) for devices to warn the host computer that they have
just been powered up or reset. The sequence is

1. System powers up.
2. Host computer resets the SCSI bus.
3. Host computer sends "test unit ready" to see if there's a device.
4. Device, having just been reset, sends "check condition" (warning) status.
5. Host computer sends "get extended sense" (give my the error status).
6. Device sends unit-attention "I've just been reset or powered up"
7. Host computer sends "test unit ready" again. This time it succeeds.

The Mac assumes that any "check condition" indicates an error and skips
the mount. You can work around this by using SCSI Probe (or similar
Control Panels) or an INIT to clear the unit attention condition and retry
the mount. Alternately, the disk manufacturer may choose to add Apple-specific
firmware that supresses power-up unit attention.

Hope this clarifies things slightly.

Martin Minow
[e-mail]
ps: the sequence above is more than slightly simplified. Read the Ansi
SCSI-II spec for the gory details.

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

From: [e-mail] (Paul Kohlmiller)
Subject: Launching from a document file
Date: 11 Sep 92 23:26:42 GMT

I am successfully launching applications via LaunchApplication.
The problem what I would really like to do is launch from a
document not from the application file itself. I want to be able
to do the equivalent of double-clicking a MSWord document and have it
come up with that document open. I think I need something like
LaunchApplicationwithOpenFile but I don't see this in IM Processes or
IM Files (on CD). Getting the application name from the document file
is probably not too hard but how do I get it to launch the app and open
the file, too?
thnx

- --
     // Paul H. Kohlmiller             //  "Cybers, Macs and Mips"    //
     // Control Data Systems, Inc.    // Internet: [e-mail]  //
     // Acronym under construction.  // America Online: Paul CDC    //
     // All comments are strictly my own. I won't lie but my       //
     // perception of reality is at best a personal hunch.        //

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

From: [e-mail] (Steve Orchard)
Date: 14 Sep 92 14:15:11 GMT
Organization: Martin Marietta

In article <[e-mail]>, [e-mail] (Paul Kohlmiller) writes:
>
> I am successfully launching applications via LaunchApplication.
> The problem what I would really like to do is launch from a
> document not from the application file itself. I want to be able
> to do the equivalent of double-clicking a MSWord document and have it
> come up with that document open. I think I need something like
> LaunchApplicationwithOpenFile but I don't see this in IM Processes or
> IM Files (on CD). Getting the application name from the document file
> is probably not too hard but how do I get it to launch the app and open
> the file, too?
> thnx

You must set the Finder Information first and then launch the Application.
Below is the code I use to do this for system 6.  For system 7 and beyond
use AppleEvents.  Check out the code snippet on ftp.apple.com.


typedef struct
{
    int            message;    /* Open = 0, Print = 1 */
    int            count;        /* Number of docs selected. */
    AppFile        findinfo;    /* AppFile structure */
} **FindrInfo;

void setfinder(thepath,num)
Str255    thepath;
int        num;
{
    OSErr                err;
    FindrInfo            finderstuff;
    FInfo                finfo;
    int                    i,strsize;
    AppFile                *temp;

    err = GetFInfo(thepath,0,&finfo);
    HUnlock(AppParmHandle);
    if(num==1)
        SetHandleSize(AppParmHandle,(Size)(sizeof(AppFile)+4));
    else
        SetHandleSize(AppParmHandle,(Size)(sizeof(AppFile))+(Size)(sizeof(AppParmHandle)));
    MoveHHi(AppParmHandle);
    HLock(AppParmHandle);
    finderstuff = (FindrInfo)AppParmHandle;
    (**finderstuff).message=0;
    (**finderstuff).count=num;
    temp = &(**finderstuff).findinfo;
    for(i=1;i!=num;i++)
    {
        strsize = *temp->fName + 9;
        if((strsize % 2) != 0) strsize=strsize+1;
        temp=(Ptr)temp + strsize;
    }
    temp->fType=finfo.fdType;
    temp->vRefNum=finfo.fdFldr;
    strcpy(temp->fName,thepath);
}

Good Luck
Steve Orchard

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

From: [e-mail] (Pablo Millan)
Date: 21 Sep 92 19:37:56 GMT
Organization: Pontificia Universidad Catolica de Chile

Paul Kohlmiller ([e-mail]) wrote:
: I am successfully launching applications via LaunchApplication.
: The problem what I would really like to do is launch from a
: document not from the application file itself. I want to be able
: to do the equivalent of double-clicking a MSWord document and have it
: come up with that document open. I think I need something like
: LaunchApplicationwithOpenFile but I don't see this in IM Processes or
: IM Files (on CD). Getting the application name from the document file
: is probably not too hard but how do I get it to launch the app and open
: the file, too?
: thnx
:


Hi.

Get the application name with Desktop Manager IM VI (I haven't see
the new IM). (i can't remember the name)

After you launch the application:

   Use AppleEvents (type odoc , Open Document) also documented in IM VI.

sorry, but i don't know the old way, maybe someone else could help?

- --
- ------------------------------------------------------------------------------
Pablo A. Millan L.        |  MIS OPINIONES SON MIAS (pero te las puedo vender)
- ------------------------------------------------------------------------------
email : [e-mail]    |  Seeds Limitada, Santiago, Chile
- ------------------------------------------------------------------------------
"Research is what I'm doing when I don't know what I'm doing" -- W. von Braun
- --
- ------------------------------------------------------------------------------
Pablo A. Millan L.        |  MIS OPINIONES SON MIAS (pero te las puedo vender)
- ------------------------------------------------------------------------------
email : [e-mail]    |  Seeds Limitada, Santiago, Chile
- ------------------------------------------------------------------------------
"Research is what I'm doing when I don't know what I'm doing" -- W. von Braun

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

From: [e-mail] (Steve Herman)
Date: 24 Sep 92 14:21:05 GMT
Organization: BCSS

In article <[e-mail]>, [e-mail]
(Pablo Millan) wrote:
>
> Paul Kohlmiller ([e-mail]) wrote:
> : I am successfully launching applications via LaunchApplication.
> : The problem what I would really like to do is launch from a
> : document not from the application file itself. I want to be able
> : to do the equivalent of double-clicking a MSWord document and have it
> : come up with that document open. I think I need something like
> : LaunchApplicationwithOpenFile but I don't see this in IM Processes or
> : IM Files (on CD). Getting the application name from the document file
> : is probably not too hard but how do I get it to launch the app and open
> : the file, too?
> : thnx
> :
>
>
> Hi.
>
> Get the application name with Desktop Manager IM VI (I haven't see
> the new IM). (i can't remember the name)
>
> After you launch the application:
>
>    Use AppleEvents (type odoc , Open Document) also documented in IM VI.
>
> sorry, but i don't know the old way, maybe someone else could help?
>

Another way to accomplish this (under System 7) is to send Finder an "Open
Selection" Apple Event.  You place the document you want opened in the
Apple Event and Finder figures out who the application to be launched is.
If the app is not running it gets launched with an 'odoc' event.  If it is
already running, it gets brought to the front and sent an 'odoc' event.

There is a code snippet on ETO which does this. I think it's called
SendFinderOpen or something like that.

I guess the only problem with this technique is that Finder is always going
to launch the "creator" application. (ie I don't think you could trick it
into opening a TeachText document with MS Word)

- ----------------------------------------------------
- - Steve Herman - BCSS
- - Boeing Computer Support Services
- - Huntsville, AL
- ----------------------------------------------------

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

From: [e-mail] (Daniel Morrow)
Subject: Finding system folder in system 6
Date: 28 Sep 92 02:45:24 GMT
Organization: Drexel University

Subject line says it all. I'm writing a program that will
have a preference file. I know how to get the Preferences folder
in System 7. I need to know how to get the System folder in
System 6. Forgive me if this is a FAQ...

Also, what is the best way to tell if system 6 or 7 is running?

Thanks in advance.
- -Dan.

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

From: [e-mail] (Grobbins)
Date: 28 Sep 92 06:29:04 GMT
Organization: Apple Computer Inc., Cupertino, CA

In article <[e-mail]> [e-mail] (Daniel Morrow) writes:
>Subject line says it all. I'm writing a program that will
>have a preference file. I know how to get the Preferences folder
>in System 7. I need to know how to get the System folder in
>System 6.

Call FindFolder and ask for the preferences folder (kPreferencesFolderType).
Under System 7, this will return the vRefNum and dirID of the
Preferences folder; under System 6, it will return the vRefNum and
dirID of the System folder (since that is where preferences should be
kept.)  FindFolder is documented on pages 9-42 and following of Inside
Mac VI.  Don't bother to call Gestalt to see if FindFolder is available;
if you're using Think or MPW C or Pascal, the FindFolder glue
will make certain that FindFolder succeeds on any HFS system.

>Also, what is the best way to tell if system 6 or 7 is running?

It's best to check if specific features are available, not what
system is running.  Gestalt is typically the way to go; read
chapter 3 of Inside Mac VI.

Grobbins          [e-mail]

Usual disclaimers apply.

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

From: [e-mail] (Michael Graf)
Date: Mon, 21 Sep 92 09:14:10 EST
Subject: Double Click in Window Content
Organization: Australian Programming Centre (IBMA)

  Can anyone please provide me with an algorithm, hint, pointer or advice
which will allow me to detect DOUBLE CLICKs within the content region of
a normal window ?

  I can easily detect single MouseDowns, but am not certain the best way
to detect a double click.

  Thanx, in advance, for the help...


**********************************************************************
Regards,
  Michael Graf                INTERNET:    ([e-mail])
                              BITNET:      (mgraf@vnet)
**********************************************************************

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

From: [e-mail] (Caleb Strockbine)
Date: 22 Sep 92 04:20:03 GMT
Organization: SUNY at Stony Brook Computer Science Dept.

In article <[e-mail]> [e-mail] (Michael Graf) writes:
     Can anyone please provide me with an algorithm, hint, pointer or advice
   which will allow me to detect DOUBLE CLICKs within the content region of
   a normal window ?

You can use the trap GetDblTime() to find the maximum number of ticks
allowed between the two clicks of a double click. (You can also use
the system global DoubleTime directly, if you prefer globals to traps.)

Compare the time between two clicks to see if it's less than the value
returned by GetDblTime(). If it is, and if the clicks are close together,
the click is actually a double click.

Caleb.
[e-mail]

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

From: [e-mail] (Eric Bowman (bobo))
Date: 22 Sep 92 03:41:12 GMT
Organization: Reed College, Portland, OR

In article <[e-mail]> [e-mail] (Caleb Strockbine) writes:
>In article <[e-mail]> [e-mail] (Michael Graf) writes:
>     Can anyone please provide me with an algorithm, hint, pointer or advice
>   which will allow me to detect DOUBLE CLICKs within the content region of
>   a normal window ?

>You can use the trap GetDblTime() to find the maximum number of ticks
>allowed between the two clicks of a double click. (You can also use
>the system global DoubleTime directly, if you prefer globals to traps.)

>Compare the time between two clicks to see if it's less than the value
>returned by GetDblTime(). If it is, and if the clicks are close together,
>the click is actually a double click.

This will work, but it's not quite right.  MPW 411 says this about GetDblTime:

GetDblTime returns the suggested maximum difference (in ticks) that should exist
between the times of a mouse-up event and a mouse-down event for those two mouse
clicks to be considered a double-click. The user can adjust this value by means of
the Control Panel desk accessory.

So you should catch mouse up events, and note store the time-of-event somewhere,
then check it in your mouse down event catcher.

bobo
[e-mail]

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

From: [e-mail] (Magnetic Technology Center)
Organization: dSoft (david's Software)
Date: Tue, 22 Sep 1992 13:46:53 GMT

[[e-mail] (Michael Graf) writes:]
[] Can anyone please provide me with an algorithm, hint, pointer or advice
[] which will allow me to detect DOUBLE CLICKs within the content region of
[] a normal window ?

[[e-mail] (Caleb Strockbine) writes:]
[] You can use the trap GetDblTime() to find the maximum number of ticks
[] ...
[] Compare the time between two clicks to see if it's less than the value
[] returned by GetDblTime(). If it is, and if the clicks are close together,
[] the click is actually a double click.

[[e-mail] (Eric Bowman (bobo)) writes:]
[] This will work, but it's not quite right.  MPW 411 says this about
[] GetDblTime: returns the suggested maximum difference (in ticks) that
[] should exist between the times of a mouse-up event and a mouse-down
[] event for those two mouse clicks to be considered a double-click.
[] ...
[] So you should catch mouse up events, and note store the time-of-event
[] somewhere, then check it in your mouse down event catcher.

MPW 411 (and IM - Event Manager) does not resolve an idiosyncratic
situation where a mousedown occurs followed by a lengthy pause with
the mouse not moving and then, in rapid succesion, a mouseup and
another mousedown.  This satisfies both GetDblTime() and mouse
movement constraints but generally isn't considered a double click
(since the two mousedowns are not close enough together in time).
This, apparently, is left open for the app writer to resolve.

  -dave-
[e-mail]

see also: StillDown(), WaitMouseUp() and IM - Event Manager

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

From: [e-mail] (Steven Dorner)
Date: 24 Sep 92 01:44:33 GMT
Organization: Qualcomm, Inc

Caleb Strockbine wrote:
> the Finder measures times between mousedown events
> a mouseup and a mousedown.

The Finder also uses the wrong paradigm for shift-click in its list
windows.

> Since I generally do a lot more double clicking
> in the Finder than anywhere else, and since I assume the average user
> probably does the same, I guess I'll make my futre apps follow the
> Finder's lead for the sake of consistancy.

The Finder is not a paragon of virtue in any sense.  If I had ATP
conformant code, I don't think I'd change it to match the Finder (but your
point about the Finder's prominence is well taken.)

>Not that anyone would ever notice...

Agreed.

> triple clicks, too. I also think nobody uses them because of the difficulty
> some folks have triple clicking.

I'm getting pretty used to triple-click as "select paragraph".  I certainly
wouldn't make it the ONLY way to do something, but it can be handy as an
accellerator.

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

From: [e-mail] (Caleb Strockbine)
Date: 25 Sep 92 09:56:19 GMT
Organization: SUNY at Stony Brook Computer Science Dept.

In article <[e-mail]> [e-mail] (Steven Dorner) writes:

   I'm getting pretty used to triple-click as "select paragraph".  I certainly
   wouldn't make it the ONLY way to do something, but it can be handy as an
   accellerator.

Absolutely. I don't have the guidelines handy, but I think there should
always be an alternative to double clicking, too. Double clicking, as well
as Option-, Shift-, or Command-clicking, is a shortcut that's not intuitive
for the novice user. If all of those are used, things can get really hairy
even for the experienced Mac user. I point to Illustrator as an example
of a very difficult-to-master user interface.

Caleb.

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

From: [e-mail]
Date: 25 Sep 92 05:28:54 GMT
Organization: Wright State University

In article <[e-mail]>, [e-mail] (Michael Graf) writes:
>   Can anyone please provide me with an algorithm, hint, pointer or advice
>   which will allow me to detect DOUBLE CLICKs within the content region of
>   a normal window ?

For the archives, here's code similar to what I sent Michael.

 var gLastMouseUp: EventRecord;

 procedure SaveLastMouseUp(event: EventRecord);
 begin
   gLastMouseUp := event
 end;

 function CheckDoubleClick (event: EventRecord): Boolean;
   var slop: Rect;
 begin
   with gLastMouseUp.where do
     SetRect(slop, h - 2, v - 2, h + 2, v + 2);
   CheckDoubleClick:= ((event.when - gLastMouseUp.when) <=
     GetDblTime) and (PtInRect(event.where, slop))
 end;

In the main event loop, call SaveLastMouseUp in response to mouseUp
events, and call CheckDoubleClick in response to mouseDown events. The
latter returns true when a double click (as defined in IM vol.I,
p.260) occurs.

o----------------------------------------------------------------------------o
| John B. Matthews, [e-mail], disclaimer:= myViews <> WSU |
|      "Whom the gods would destroy, they first invite to program in C"      |
o----------------------------------------------------------------------------o

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

From: [e-mail] (P. Hawthorne)
Date: 28 Sep 92 11:29:08 GMT
Organization: Reed College, Portland, OR


  [e-mail] (Caleb Strockbine) writes:
: Double clicking, as well as Option-, Shift-, or Command-clicking, is a
: shortcut that's not intuitive for the novice user. If all of those are
: used, things can get really hairy even for the experienced Mac user. I
: point to Illustrator as an example of a very difficult-to-master user
: interface.

  I point to Illustrator as the most effective user interface.  It permits
the user to become very expressive. Some short cuts are not intuitive, but
I can assure you that they come to be instinctive. Still, implying that
double clicking is too computery for people seems rather like saying that
if God wanted us to compute, he would have given us mice to fidget with.
People have to do something faintly like learning to communicate with a
computer. It can't be entirely one way. But beyond that...

  Difficult-to-master? Yes! Mastering Illustrator is not like mastering a
word processor or some trivial application. You wouldn't write Illustrator
in HyperCard. Lest the user-friendly trappings disguise it, Illustrator is
a Craft and a remarkably interesting subdiscipline of programming at that.

  I can't think of anything in Illustrator that one can do with the mouse
and the keyboard that you cannot find in the menus or the toolbox.  Every
short cut has it's representation in the user interface.


  Theus ([e-mail])

  P.S.
  Sorry for disappearing. I'm living... That might not be the right word.
I'm merely 'in' Utah. I've moved here. I'm not dead exactly, either, but...

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

From: [e-mail]
Subject: Bug fixes in MPW C 3.2.4
Date: 21 Sep 92 11:42:11 GMT
Organization: Actrix Information Exchange

I don't think I've seen anyone post details of just what exactly was wrong
with MPW C 3.2.3 on ETO #8.

Well, I got the special 3.2.4 update posting (dated 18/9/92) today --
kudos to the MPW guys for getting this out quickly.

Here is part of the readme file (note that much of it is changes from
3.2 to 3.2.3):

- -----

These notes describe changes from MPW C 3.2.3 to 3.2.4

The following bugs have been fixed:

- - The compiler will sometimes generate incorrect code for complicated
logical expressions containing constants.   The error will occur when
the logical expression is the controlling expression of a conditional
( ? : ) expression, or on the right side of an assignment statement.
The symptom of this bug is that the logical expression will return the
wrong value.

For example the following generates incorrect code:

      x = (a > 0) && 1 && (a < 2) ? 1 : 2;
                  or
      x = (a > 0) && 2 && (a < 2);

- - The compiler will generate object files with an incorrect value in
the FileID field of a Local Type record.  Depending on the contents of
the object file, the Linker (or Lib) may sometimes complain about the
invalid record.  If the link succeeds, the .SYM file will have
incorrect information; the application will be correct.

This bug occurs only when a program uses #pragma load and is compiled
and Linked (or Libd) with -sym on.

As a result of this fix, Load/Dump files created with earlier versions
of the C compiler should be rebuilt.

- - The compiler will incorrectly evaluate the following construct:
            aa < (bb << cc)
                     or
            aa < (bb >> cc)

when aa, bb, and cc are all constant expressions (i.e. either
constants themselves or expressions all of whose operands are
constants), aa is negative, and (bb << cc) or (bb >> cc) is greater
than 65536.


These notes describe changes from MPW C 3.2 to 3.2.3

- - An increase in the number of allowable static strings to above 32K.

- - Speed improvements ranging from 6% to 26% in compilation of large
files with the option "-sym on".

- - The option "-blksize n" has been added.  The parameter "n" is the
number of 512-byte blocks to be used for the compiler's input buffers.
The range of "n" is from 4 to 63.  You can use this option to tune the
performance of the compiler for your particular hardware
configuration.  The default value of 16 seems to be a reasonable
compromise for most classes of machines.

Bug fixes

- - The "-mc68020" option used to do an incorrect optimization for
multiplication by twelve.  This would often show up when calculating
an offset in a large array of "extended" values. The bug showed up
only when the value of the result required more than 15 bits.

- - The compiler emits ID numbers for use by the Linker and debuggers.
There used to be three ranges of these numbers. If there were enough
symbols in a compilation unit, the number ranges could overlap, giving
bogus symbolic information. The compiler now emits a single contiguous
range of ID numbers, allowing up to 32K symbols to exist in one
compilation unit.

- - A logical operator (AND, OR) with a constant operand would sometimes
cause a TEST instruction with an immediate operand to be emitted. This
would cause an "illegal instruction" error at runtime.

- - If a function allocated more than 32K of space for local variables,
the compiler would not clean up the stack correctly. The compiler now
detects the situation and emits a correct epilog for such functions.

- - If a conditional expression had a function call as one alternative
and a constant as the other, the result of the function call would get trashed.

- - The compiler used to create illegal "multiply" instructions that
used address registers as operands.

- - There were bugs in the algorithm for dead-code elimination. The
symptoms were that when using the "-b3" option with a function
containing string literals (as in "printf" statements), seemingly
random sections of the function would get trashed. This also happened,
even in the absence of the "-b3" option, when the function contained
floating point constants.

- - If a function contained a local static variable, compiling with
"-sym on" would cause random system crashes. Even if crashes did not
occur, the symbolic information for that variable would be garbage.

- - The compiler used to abort when a partial pathname was given in the
"-i<filename>" option.

- - Crashes occurred when source code contained floating point constants.

- - Object files created using "-opt full" but not using any "-bn"
option would fail to link when the source file contained a string literal.

- - When evaluating constructs of the form: "ptr++->field = somevalue;",
the address increment was calculated incorrectly in the high order
byte, causing a problem when running in 32-bit addressing mode.

- - Crashes sometimes occurred when compiling source using "-sym on"
when the source contained a struct for which some field was created by
a macro expansion.

- - Incorrect parameter values were passed when the option "-mc68881"
was used and the parameter was a constant that was explicitly cast to float.

- - Logical tests of expressions of the form:

     someBooleanExpression && (someVar = pointerArray[i].ptr)

failed because the generated code used an LEA instruction to assign
the value of  pointerArray[i].ptr despite the fact that LEA does not
set the condition code.

- - The condition codes were not properly tested in expressions of the form:
      if (x ? 1 : g() )
      if (x ? f() : g() )

- - The condition codes were not tested properly when the results of two
function calls were ANDed, e.g., func1() && func2().

- - A bitwise OR of two logical expressions failed in some instances
where one or both of the expressions was complex.

- - Under the option "-model far", a function pointer referring to a
function in the same segment was set to the code address instead of to
the jump table address.  This would cause the pointer to become
invalid after the code segment was unloaded.  This is now corrected,
provided that the Link version used is from MPW 3.2.2 ( or later).

- - A bug in the constant folding of Boolean expressions caused the
generation of the
illegal instruction TST.L #$0000 when compiling constructs like:

   someVar && (someExpression, 1)
   booleanVar = booleanVar && 1;

- - The same bug would cause generation of an incorrect test for
constructs like:
   if ((0 || (someVar && !someVar)) && 1)

- - A code generation error occurred when a function was declared to
return a pointer to a function.

- - A bus error would occur when compiling with "-sym on".
- --
[e-mail]   Twisted pair: [phone]
BIX: brucehoult                 Last Resort:  PO Box 4145 Wellington, NZ
"Cray's producing a 500 MIPS personal computer with 256MB RAM and 8 GB
hard disk that fits in your pocket!"   "Great!  Is it PC compatible?"

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

From: [e-mail] (Matthew Holiday)
Date: 21 Sep 92 21:51:14 GMT
Organization: Bell-Northern Research, Richardson, TX

In article <[e-mail]>, [e-mail] writes:
|> I don't think I've seen anyone post details of just what exactly was wrong
|> with MPW C 3.2.3 on ETO #8.
|>
|> Well, I got the special 3.2.4 update posting (dated 18/9/92) today --
|> kudos to the MPW guys for getting this out quickly.
|>

Where exactly is the updated compiler?  I'd like to get the 3.2.3 C/C++ update from APDA
on CD-ROM, but I won't order it until I know for sure that I'll get a version with the
fixes!  I still remember never getting the 2.0.1 bug fixes to MacApp, even though they
were promised to me when I got that CD in '90 :-(.

- --
Matt Holiday                                      #include <std/disclaimer>
[e-mail]
BNR Richardson, TX

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

From: [e-mail]
Date: 22 Sep 92 02:26:36 GMT
Organization: Actrix Information Exchange

In article <[e-mail]> [e-mail] writes:
> In article <[e-mail]>, [e-mail] writes:
> |> I don't think I've seen anyone post details of just what exactly was wrong
> |> with MPW C 3.2.3 on ETO #8.
> |>
> |> Well, I got the special 3.2.4 update posting (dated 18/9/92) today --
> |> kudos to the MPW guys for getting this out quickly.
> |>
>
> Where exactly is the updated compiler?  I'd like to get the 3.2.3 C/C++ update from APDA
> on CD-ROM, but I won't order it until I know for sure that I'll get a version with the
> fixes!  I still remember never getting the 2.0.1 bug fixes to MacApp, even though they
> were promised to me when I got that CD in '90 :-(.

Sorry I wasn't more specific: it arrived via TNT Skypak International
Express with a little green sticker saying "Priority Service".  It's
just a single floppy disk with compressed versions of C 3.2.4 and C
3.2.2, a new version of CMarker, and a CExamples folder.

I'm just a regular ETO customer, so you chaps in the US of A should
get yours soon :-)

- -- Bruce
(who can't quite beleive he seems to be getting stuff first out here in NZ)

- --
[e-mail]   Twisted pair: [phone]
BIX: brucehoult                 Last Resort:  PO Box 4145 Wellington, NZ
"Cray's producing a 500 MIPS personal computer with 256MB RAM and 8 GB
hard disk that fits in your pocket!"   "Great!  Is it PC compatible?"

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

From: [e-mail] (Keith Rollin)
Date: 23 Sep 92 19:59:48 GMT
Organization: Taligent

In article <[e-mail]>, [e-mail] (Matt Mora) wrote:
>
> In article <[e-mail]> [e-mail] writes:
>
> >- A bus error would occur when compiling with "-sym on".
>
> [Other 26 bug fixes deleted]
>
> Has THINK C or THINK Pascal ever had that many bugs? They charge you for that
> buggy compiler? :-) How do you MPW heads get anything done?
>

I got this file right here called "About THINK C 5.0.3." It details at
least 48 bugs fixed from THINK C 5.0 to 5.0.3. That doesn't count any of
the documentation errors or "gotchas" that don't quite qualify as bugs.

- -----
Keith Rollin
Phantom Programmer
Taligent, Inc.

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

From: [e-mail] (Paul Potts)
Date: 24 Sep 92 14:26:26 GMT
Organization: Instructional Technology Laboratory, University of Michigan

In article <[e-mail]> [e-mail] (Keith Rollin) writes:
>In article <[e-mail]>, [e-mail] (Matt Mora) wrote:
>>
>> In article <[e-mail]> [e-mail] writes:
>> Has THINK C or THINK Pascal ever had that many bugs? They charge you for that
>> buggy compiler? :-) How do you MPW heads get anything done?
>>
>
>I got this file right here called "About THINK C 5.0.3." It details at
>least 48 bugs fixed from THINK C 5.0 to 5.0.3. That doesn't count any of
>the documentation errors or "gotchas" that don't quite qualify as bugs.

All compilers have bugs. If you don't find them, it just means you aren't
doing anything difficult... :)


- --
"...remove protective cap. Hold atomizer with thumb at base and nozzle
between first and second fingers. Without tilting head, insert nozzle into
nostril. Fully depress rim with a firm, even stroke and sniff deeply."
- -advice for presidential candidates from Paul Potts - [e-mail]

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

From: [e-mail] (Victor Eijkhout)
Subject: Think Pascal and ResEdit questions
Date: 28 Sep 92 16:51:55 GMT
Organization: University of Tennessee, Knoxville - Department of Computer Science

Hello,

working with Think Pascal and ResEdit (and being mondo pissed that the
Pascal manual nowhere mentions that you need ResEdit for creating resources),
I had the following questions:

1/ using system 7 I would like to have both Pascal and ResEdit open, but
if I'm editing the resource, Pascal has trouble finding elements (menubar
for instance) in it. So I have to close the resource every time I want
to test run the program. Is there an explanation / work-around for this?

2/ I have programmed mouse-and-menu programs before on the Atari ST, and
the resource construction programs there gave you a possibility to
give symbolic names to items in the resource, which you then used in your
program. This way alterations in the resource do not necessitate alterations
in the program. Is such a thing possible with ResEdit, combined with Think
Pascal?

Thanks in advance.

Victor.

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

From: [e-mail] (Tom Emerson)
Date: 28 Sep 92 18:22:01 GMT
Organization: University of Vermont, EMBA Computer Facility

In article <[e-mail]> [e-mail] (Victor Eijkhout) writes:

   1/ using system 7 I would like to have both Pascal and ResEdit open, but
   if I'm editing the resource, Pascal has trouble finding elements (menubar
   for instance) in it. So I have to close the resource every time I want
   to test run the program. Is there an explanation / work-around for this?

There is no workaround. The resource file can *not* be open in two
applications simultaneously. One way to deal with that is to put your resource
file in the open special item in ResEdit's File menu.

   2/ I have programmed mouse-and-menu programs before on the Atari ST, and
   the resource construction programs there gave you a possibility to
   give symbolic names to items in the resource, which you then used in your
   program. This way alterations in the resource do not necessitate alterations
   in the program. Is such a thing possible with ResEdit, combined with Think
   Pascal?

Not as such. You can define a group of constants in your source file for
particular parts of your dialog or for resource id's. I'm not sure I
understand how the symbolic names of the Atari system differ from the resource
id's used with the Mac?

Tom
- --
   Tom Emerson                                    Technical Support
   Language Products Group                     Symantec Corporation
            Internet: [e-mail]

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

From: [e-mail] (Victor Eijkhout)
Date: 28 Sep 92 18:32:46 GMT
Organization: University of Tennessee, Knoxville - Department of Computer Science

In article <[e-mail]>, [e-mail] (Tom Emerson) writes:
|> In article <[e-mail]> [e-mail] (Victor Eijkhout) writes:
|>
|>    2/ I have programmed mouse-and-menu programs before on the Atari ST, and
|>    the resource construction programs there gave you a possibility to
|>    give symbolic names to items in the resource, which you then used in your
|>    program. This way alterations in the resource do not necessitate alterations
|>    in the program. Is such a thing possible with ResEdit, combined with Think
|>    Pascal?
|>
|> Not as such. You can define a group of constants in your source file for
|> particular parts of your dialog or for resource id's. I'm not sure I
|> understand how the symbolic names of the Atari system differ from the resource
|> id's used with the Mac?

On the Atari the resource construction program yielded a .RSC file
with the resource and a .RSD file with for C 'define' and for Pascal
'Const' statements equating for instance
 Const File_menu = 129,
where in the res. program you have told that 'File_menu' is the label
for the file menu.

This file is then loaded into your program, and you only have
to remember the name 'File_menu', and not its ID. I seem to remember
that there were reasons why you wanted to renumber in a resource every
once in a while, and then these names are very convenient.

If I've understood ResEdit correctly, then the number of an item in a
menu is simply the number of the line it appears on. This is awful:
If I insert a menu item higher up I have to change the item ID in
my program. Have I really understood this correctly?

|> Tom

Victor.

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

From: [e-mail] (Jamie R. McCarthy)
Date: 28 Sep 92 18:59:28 GMT
Organization: Kalamazoo College

[e-mail] (Victor Eijkhout) writes:
>[e-mail] (Tom Emerson) writes:
>|>[e-mail] (Victor Eijkhout) writes:
>|>
>|>    2/ I have programmed mouse-and-menu programs before on the Atari ST, and
>|>    the resource construction programs there gave you a possibility to
>|>    give symbolic names to items in the resource, which you then used in your
>|>    program.
>|>    ... Is such a thing possible with ResEdit, combined with Think
>|>    Pascal?
>|>
>|> Not as such.
>
>On the Atari the resource construction program yielded a .RSC file
>with the resource and a .RSD file with for C 'define' and for Pascal
>'Const' statements...
>
>This file is then loaded into your program, and you only have
>to remember the name 'File_menu', and not its ID.
>
>If I've understood ResEdit correctly, then the number of an item in a
>menu is simply the number of the line it appears on. This is awful:
>If I insert a menu item higher up I have to change the item ID in
>my program. Have I really understood this correctly?

Yes, you have.  A nice extension to Think C 4.0.x, "4Plus", provided
this and many other features.  (Not for menu items, though--it would
produce a .h file that did 'enum's attaching resource names to IDs.
So you could change menu numbers on the fly, but not menu items.)

This would be a way-cool tool, if someone wanted to write it--a
drag-n-drop app that accepted any file, and spelled out its resources in
some clean header file.  Preferences would include:  whether to write a
C or Pascal header;  the prefix for all the constants ("r", for
example);  whether to convert "two word" to "two_word" or "twoWord";
and so on.  The option of writing separate files for separate groups of
resources would be nice too--when I change a menu item, I don't want to
have to recompile all my C files that #include the enumerations of the
'snd ' resources.

Hmmm.  If I didn't already have six things on the back burners, I'd
consider tossing this one on too.  Or has it been done already?
- --
 Jamie McCarthy      Internet: [e-mail]      AppleLink: j.mccarthy
 "To flame someone on an open forum is, to me, the height of insensitivity."
   - P.J. Plauger

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

From: [e-mail] (J Eric Bush)
Date: 28 Sep 92 19:54:08 GMT
Organization: The Ohio State University

In article <[e-mail]> [e-mail] (Victor E
ijkhout) writes:
>Hello,
>
>working with Think Pascal and ResEdit (and being mondo pissed that the
>Pascal manual nowhere mentions that you need ResEdit for creating resources),
>I had the following questions:
>
>1/ using system 7 I would like to have both Pascal and ResEdit open, but
>if I'm editing the resource, Pascal has trouble finding elements (menubar
>for instance) in it. So I have to close the resource every time I want
>to test run the program. Is there an explanation / work-around for this?
>
>2/ I have programmed mouse-and-menu programs before on the Atari ST, and
>the resource construction programs there gave you a possibility to
>give symbolic names to items in the resource, which you then used in your
>program. This way alterations in the resource do not necessitate alterations
>in the program. Is such a thing possible with ResEdit, combined with Think
>Pascal?
>
>Thanks in advance.
>
>Victor.
>
>

1/ ResEdit will mark a file as being opened with the ability to read and
   write so another program can't have access to it at the same time.  The
   best work around is use the "Modify Menu" option in ResEdit so that
   command-1 opens the resource you want to edit.  Edit it, then close it
   when returning to Pascal, it can access everything just fine.  When you
   want to go back to ResEdit switch to it under mutilfinder and press
   command-1 and your back to editting your resource.  BTW, Think Pascal
   manuals do mention ResEdit all over the place.  Each of the examples
   are based on ResEdit and it comes with the program.

2/ If you want to use symbolic names, you have to use the ResEdit Get Info
   command and then type in a name for the resource.  Note: I have given
   names to resources, but do not use them in programs, so if this doesn't
   work, forget where you heard this.

Eric
- --
- ------------------------------------------------------------------------------
J. Eric Bush
Internet: [e-mail]

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

From: [e-mail] (Peter N Lewis)
Date: 29 Sep 92 03:30:55 GMT
Organization: NCRPDA, Curtin University

In article <[e-mail]>, [e-mail]
(Victor Eijkhout) wrote:

> 1/ using system 7 I would like to have both Pascal and ResEdit open, but
> if I'm editing the resource, Pascal has trouble finding elements (menubar
> for instance) in it. So I have to close the resource every time I want
> to test run the program. Is there an explanation / work-around for this?

No, there is no workaround, but once you've selected the resource file
once, ResEdit remembers where you looked, so reopening it is generally not
too difficult, and the suggestion of putting the resource file in the Open
Special menu in ResEdit was a good one (assuming you only have a few
projects at once :-)

> 2/ I have programmed mouse-and-menu programs before on the Atari ST, and
> the resource construction programs there gave you a possibility to
> give symbolic names to items in the resource, which you then used in your
> program. This way alterations in the resource do not necessitate alterations
> in the program. Is such a thing possible with ResEdit, combined with Think
> Pascal?

There is no app to do this to my knowledge.  What I did was implement
fmnu's ala Finder, which creates a mapping between an OSType (eg 'quit')
and menu's (eg thew 'Quit/Q' menu).  In my code I do things like:

SetFCommand('quit',@DoQuit)

And then when a menu is chosen, its OSType is found and the relevent
procedure is called.  It also associates an Enable procedure with each
menu, to grey/ungrey the menu as appropriate.

The code for this is in my Talk sources, which are available on
redback.cs.uwa.edu.au:/others/peterlewis (as well as on sumex and umich).
Thanks to Quinn and UWA Comp Sci for providing this site.

HTH,
   Peter.

_______________________________________________________________________
Peter N Lewis, NCRPDA, Curtin University       [e-mail]
GPO Box U1987, Perth WA 6001, AUSTRALIA             FAX: [phone]

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

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