Tricks of the Mac Game Programming Gurus/TricksOfTheMacGameProgrammingGurus.iso/Information/CSMP Digest/volume 1/csmp-v1-199.txt

Text File | 1994-12-08 | 33.6 KB | 909 lines | [TEXT/R*ch]

C.S.M.P. Digest             Wed, 28 Oct 92       Volume 1 : Issue 199

Today's Topics:

    How to tell if NAN?
    ARA programatic interface
    Looking for fast sorting code for pascal
    Does anyone recognize _Jackson?
    Synchronizing AppleEvents



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] (Eric Bowman (bobo))
Subject: How to tell if NAN?
Date: 19 Sep 92 23:06:10 GMT
Organization: Reed College, Portland, OR

My apologies if you see this twice...I posted it once, but it never showed up
here, so I'm posting again.

I'm wondering the most general way to tell if a number is NAN.  I don't care
what kind of NAN it is; I only want to know that it is.  This occurs in a
replacement for printf, so actually catching the exception is not what I want
to do; I just want to print "NAN" or somesuch rather than have my code loop
endlessly wondering what the hell it's been passed.  (I'm sure this is in
the SANE manual, but some asshole appears to have stolen it.)

Thanks,
bobo
[e-mail]

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

From: [e-mail] (David E. Johnson)
Date: 20 Sep 92 01:51:24 GMT
Organization: Northwest Nexus Inc. [phone]


> ...

try
    FUNCTION ClassExtended(x: Extended): NumClass;
        (as in ClassExtended(theNum)<>NormalNum)
    or FUNCTION NAN(i: INTEGER): Extended;

    NAN's with extendeds & doubles are a pain in the neck
    Another hack is to convert it to a string and compare it to
    NAN.

David E Johnson
..sig not available. default enabled: [e-mail]

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

From: [e-mail] (Mike Carlton)
Date: 22 Sep 92 03:13:25 GMT
Organization: University of California, Berkeley

In article <[e-mail]> [e-mail] (Eric Bowman
(bobo)) writes:
- - My apologies if you see this twice...I posted it once, but it never showed up
- - here, so I'm posting again.
- -
- - I'm wondering the most general way to tell if a number is NAN.  I don't care
- - what kind of NAN it is; I only want to know that it is.  This occurs in a
- - replacement for printf, so actually catching the exception is not what I want
- - to do; I just want to print "NAN" or somesuch rather than have my code loop
- - endlessly wondering what the hell it's been passed.  (I'm sure this is in
- - the SANE manual, but some asshole appears to have stolen it.)
- -
- - Thanks,
- - bobo
- - [e-mail]

That's easy, by definition of a NaN:
    if (x != x)
        printf("x is a NaN\n");

For more info, see 'man 3m math' on a decent unix box (i.e. any box which
uses IEEE 754).

cheers,
- --mike ([e-mail])

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

From: [e-mail] (Bill Hofmann)
Subject: ARA programatic interface
Date: 21 Sep 92 16:08:47 GMT
Organization: Whole Earth 'Lectronic Link

I seem to recall hearing of a programatic interface (aka API) to AppleTalk
Remote Access.  Was I halucinating?  If not, where can I get docs?

- -Bill Hofmann

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

From: [e-mail] (Lawrence D'Oliveiro, Waikato University)
Date: 22 Sep 92 05:55:10 GMT
Organization: University of Waikato, Hamilton, New Zealand

In article <[e-mail]>, [e-mail] (Bill Hofmann) writes:
> I seem to recall hearing of a programatic interface (aka API) to AppleTalk
> Remote Access.  Was I halucinating?  If not, where can I get docs?

There are two products that might be what you're looking for. Both are
available from APDA:

* The AppleTalk Remote Access Developer's Toolkit v.1.0, code R0128LL/A, US$30.

This details the neat features of the ARA protocol, including SmartBuffering
compression and the "SecurityZone" feature for controlling inbound calls to
machines on your network. It also describes the services you can call
on, new AppleTalk Transition Queue events, how to coexist with the Serial Port
Arbitrator, and so on.

* The AppleTalk Remote Access Modem Toolkit v.1.0, code R0129LL/B, US$25.

This gives you instructions on how to create connection scripts to work with
particular modems or PBXs.

Lawrence D'Oliveiro                       fone: [phone]
Computer Services Dept                     fax: [phone]
University of Waikato            electric mail: [e-mail]
Hamilton, New Zealand    37^ 47' 26" S, 175^ 19' 7" E, GMT+12:00
If only people would complete their sentences.

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

From: [e-mail] (Matthew Xavier Mora)
Subject: Looking for fast sorting code for pascal
Date: 10 Sep 92 00:13:03 GMT
Organization: SRI International

Does anybody have code for sorting in pascal? I have a dismal version
of an exchange sort working that I ported from a basic program. I tried
porting
hayden's fastqsort.c code but it doesn't seem to work. ( I haven't tried
to find out what's wrong with it because it trashes my heap every time and
I have to reboot.)

I know this is text book stuff but maybe someone has already done
the work and even included some inline asm for the memswap code.

Thanks

Matt


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

From: [e-mail] (Andrew E Page)
Date: 17 Sep 92 12:00:17 GMT
Organization: The World Public Access UNIX, Brookline, MA


   As far as the heap collisions are concerned....  The QSort algorithms
that I've seen in the past depend on recursion.  This leads to the
posibility that your stack is somehow slamming into your heap,
leading to a Sys Err ID=28.  This can be due to a couple of things...
If the condition that would terminate the recursion is never evaluated
then the function will keep iterating until the stack zooms into the
heap.  Another possibility is that your local variables in the
recursion function are taking up too much room.  See if you can trim
them down, avoid Str255s like the plauge in recursion functions.


- --
Andrew E. Page CTO(Warrior Poet)|   Decision and Effort The Archer and Arrow
DSP Ironworks                   |     The difference between what we are
Macintosh and DSP Technology    |           and what we want to be.

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

From: [e-mail]
Date: Fri, 18 Sep 1992 00:49:16 GMT
Organization: Actrix Information Exchange

In article <[e-mail]> [e-mail] (Andrew E Page) writes:
>
>    As far as the heap collisions are concerned....  The QSort algorithms
> that I've seen in the past depend on recursion.  This leads to the
> posibility that your stack is somehow slamming into your heap,
> leading to a Sys Err ID=28.

Recursion is, of course, necessary in QuickSort, but you can easily
guarentee that it will go no deeper than log2(n) -- i.e. only 20
levels when sorting a million elements.

Here's my tried and tested Pascal sorting routine (written on the Mac
in 1987).  I've found it reliable and fast.  It tries to be reasonably
clever about avoiding bad partitioning, using the minimum of recursion
and using the best algorithm for the amount of data to be sorted.

- -- Bruce

- -------------
procedure sort(
   left, right:longint;
   function less_than(a,b:longint):boolean;
   procedure swap(a,b:longint)
);
var
   i, j, mid, median, small_pos: longint;
begin
   while (right-left) >= 15 do begin
      mid := (left+right) div 2;

      {find the Median of left, mid, right}
      if less_than(left,mid) then begin
         if less_than(mid,right) then median := mid
         else if less_than(left,right) then median := right
         else median := left
      end else begin
         if less_than(left,right) then median := left
         else if less_than(mid,right) then median := right
         else median := mid;
      end; {finding the median}

      {
         partition the region into three:

         left <= x <= j:  those smaller than median
         j < x < i:       those equal to median
         i <= x <= right: those greater than median
      }
      i := left;
      j := right;
      while i <= j do begin
         while (i<right) & less_than(i,median) do i := i + 1;
         while (j>left) & less_than(median,j) do j := j - 1;
         if i <= j then begin
            if i < j then begin
               {watch for median getting moved under us!}
               if median = i      then median := j
               else if median = j then median := i;
               swap(i,j);
            end;
            {no need to look at these two again}
            i := i + 1;
            j := j - 1;
         end;
      end;
      {skip over any items equal to the guess of the median}
      while (i<right) & not less_than(median,i) do i := i + 1;
      while (j>left)  & not less_than(j,median) do j := j - 1;

      {now sort the two halves}
      if (j-left) < (right-i) then begin
         {the left half is smaller -- sort it recursively first}
         sort(left,j,less_than,swap);
         left := i; {prepare for next iteration}
      end else begin
         {the right half is smaller -- sort it recursively first}
         sort(i,right,less_than,swap);
         right := j; {prepare for next iteration}
      end;
   end; {while more than xxx elements to sort}

   {now selection sort any remaining elements}
   for i := left to right-1 do begin
      small_pos := i;
      for j := i+1 to right do
         if less_than(j,small_pos) then small_pos := j;
      if small_pos <> i then swap(i,small_pos);
   end;

end; {sort}
- --
[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] (Lewis G. Pringle)
Date: 20 Sep 92 00:56:56 GMT
Organization: Sophist Solutions

In article <[e-mail]> [e-mail] writes:
>In article <[e-mail]> [e-mail] (Andrew E Page) writes:
>>
>>    As far as the heap collisions are concerned....  The QSort algorithms
>> that I've seen in the past depend on recursion.  This leads to the
>> posibility that your stack is somehow slamming into your heap,
>> leading to a Sys Err ID=28.
>
>Recursion is, of course, necessary in QuickSort, but you can easily
>guarentee that it will go no deeper than log2(n) -- i.e. only 20
>levels when sorting a million elements.

Actually recursion is NOT ever necessary in implementing any program. The
recursive algorithm can always be rewriten to keep an array (simulating
what the program stack would have done) of values that would have been
passed as parameters.

I would, OF COURSE, agree that QuickSort is most simply, and naturally
written recusively. But if you have some limit on how much stack space
you can use, its not hard to rewrite QuickSort non-recursively.


                Lewis.

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

From: [e-mail]
Date: 20 Sep 92 15:05:11 GMT
Organization: Actrix Information Exchange

In article <[e-mail]> [e-mail] (Lewis G. Pringle) writes:
> >Recursion is, of course, necessary in QuickSort, but you can easily
> >guarentee that it will go no deeper than log2(n) -- i.e. only 20
> >levels when sorting a million elements.
>
> Actually recursion is NOT ever necessary in implementing any program. The
> recursive algorithm can always be rewriten to keep an array (simulating
> what the program stack would have done) of values that would have been
> passed as parameters.

You're making a distinction without a difference.  QuickSort is a
recursive algorithm.  The computers we use are only Finite State
Machines, not Push Down Automata, so you have a choice:  either write
your program using recursive notation and have your compiler convert
it to an iterative form using (usually) the hardware supplied stack,
or else convert it to iterative form yourself and manage your own stack.



> But if you have some limit on how much stack space
> you can use, its not hard to rewrite QuickSort non-recursively.

How would that help?  Perhaps you would save a few less values on your
own stack than my procedure saves on its, but I could trivially
rewrite mine to use less stack space, if necessary, by changing some
of my automatic variables to static, global, or automatic variables in
an enclosing interface procedure, leaving only perhaps the parameters
left&right and the local variables i and j on the stack on each
recursive invocation, plus of course the program counter (and saved
frame pointer using most Mac compilers).

In any case, even in its present form, the posted routine will use
less than a thousand bytes of stack, even when sorting many millions
of elements.  It's hard to imagine an application in which you can
find space for millions of data elements, but can't find a K for a stack.

In the same situation, your "iterative" routine would still require a
half K for your explicitely managed stack, and I could come very close
to that with a slight rewrite (but still in recursive form) if I
thought there was a need.

- -- Bruce

- --
[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]
Date: 21 Sep 92 14:34:02 GMT
Organization: Actrix Information Exchange

In article <[e-mail]> [e-mail] (Jon Wdtte) writes:
>
> With QuickSort, you have a worst-case scenario where you use
> stack linearly with the number of items... However, usually
> you can code a little cleverness into it so that won't happen.
>
> Or have you never seen System Error 28 ? :-)

Only with a very naive my-first-quicksort implementation.  If you hit
the "[" key a couple of times in your trn you'll find that my posted
QuickSort code can't recurse more than log2(n) times because it
recurses on the *smaller* of the two partitions, and then iterates on
the larger one.  The worst case for sorting performance is the best
case for stack usage!
- --
[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] (Roger W. Brown)
Subject: Does anyone recognize _Jackson?
Date: 4 Sep 92 13:38:44 GMT
Organization: Dartmouth College, Hanover, NH

I frequently see this symbol in MacsBug when my Quadra crashes:
_Jackson. I haven't looked through the Tech Notes yet. I wondered if
anyone knows what it is. It does something with color icons. I just
want to know that it isn't a virus, and maybe that it is from an
extension that may be causing trouble.

- --------------------------------------------------------------------
Roger Brown                                [e-mail]
Software Development
Kiewit Computation Center, Dartmouth College, Hanover, NH

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

From: [e-mail] (Jamie R. McCarthy)
Date: 4 Sep 92 16:14:18 GMT
Organization: Kalamazoo College

[e-mail] (Roger W. Brown) writes:
>I frequently see this symbol in MacsBug when my Quadra crashes:
>_Jackson. I haven't looked through the Tech Notes yet.

You won't find it there.  :-)

>I wondered if
>anyone knows what it is. It does something with color icons.

It's an internal trap that QuickDraw uses for doing stuff with memory.
Those branches to _DisposCIcon+xxxx don't necessarily have anything to
do with color icons.  And if you ever hit the _Debugger trap in there,
something went wrong _before_ the trap was called.  Giving the app more
memory may or may not fix it;  it's worth a try.

And it gets its name from Jackson Pollock, the modern painter whose name
was given to early versions of Color QuickDraw.  (Or is it 32-bit
QuickDraw?  I've never been good at trivia...)

>I just
>want to know that it isn't a virus, and maybe that it is from an
>extension that may be causing trouble.

Nope, it's in your machine's ROM.  (It's hard to put a virus there.)

Note the followup;  hopefully csmp won't start up another flood of
_Jackson trivia...?
- --
 Jamie McCarthy      Internet: [e-mail]      AppleLink: j.mccarthy
 The essence of OOP:  "After some hacking, I finally got the program to
 work, but I'm still not sure why." - David Marcovitz ([e-mail])

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

From: [e-mail] (Marshall Clow)
Date: 4 Sep 1992 16:22:21 GMT
Organization: Hewlett Packard San Diego Printer Division

In article <[e-mail]>,
[e-mail] (Roger W. Brown) wrote:
>
> I frequently see this symbol in MacsBug when my Quadra crashes:
> _Jackson. I haven't looked through the Tech Notes yet. I wondered if
> anyone knows what it is. It does something with color icons. I just
> want to know that it isn't a virus, and maybe that it is from an
> extension that may be causing trouble.
>
> --------------------------------------------------------------------
> Roger Brown                                [e-mail]
> Software Development
> Kiewit Computation Center, Dartmouth College, Hanover, NH

It isn't a virus. Look closely, and you will notice that the address that
you are looking at is in ROM.

On my Quadra 900:
$4083143A: JACKSONPOLLOCK          _Debugger
$4083143C: RNEWHANDLE              _NewHandle, SYS, CLEAR
$4083143E: RNEWHANDLE + $0002      BNE.S $40831432
and so on.....

It's where QuickDraw jumps to if it gets an error that it can't recover
from (mostly when it runs out of memory). Never, never, do region
operations
where you could possibly not have enough memory!

Marshall Clow
San Diego Printer Division      Hewlett Packard
Internet: [e-mail]   AppleLink: HP.Marshall    AOL: MClow

P.S.    "Jackson Pollock" was the code name for 32-bit QuickDraw.

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

From: [e-mail] (Bernie Wieser)
Date: Fri, 04 Sep 92 20:34:57 GMT
Organization: The University of Calgary, Alberta

As the illustrious Tom from Apple says, this is most frequently
indicative of a segment loader problem.

Jackson, error 25, etc.

I most often get this with 32 bit segloader patch.
- --
Bernie Wieser
[e-mail]
[e-mail]

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

From: [e-mail] (Chuq Von Rospach)
Date: 4 Sep 92 19:31:38 GMT
Organization: I is a writur

[e-mail] (Roger W. Brown) writes:

>I frequently see this symbol in MacsBug when my Quadra crashes:
>_Jackson.

Yes, usually when an after dark module crashes.


- --
Chuq "IMHO" Von Rospach, ESD Support & Training (DAL/AUX) =+= Member, SFWA
  [e-mail] | GEnie: MAC.BIGOT | ALink:CHUQ =+= Editor, OtherRealms

    I believe there should be a constitutional amendment outlawing the

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

From: [e-mail] (Steve Anthony Mariotti)
Date: 14 Sep 1992 19:11:26 -0500
Organization: U Texas Dept of Computer Sciences, Austin TX

In article <[e-mail]> [e-mail] (Roger W. Brown) writes:
>I frequently see this symbol in MacsBug when my Quadra crashes:
>_Jackson. I haven't looked through the Tech Notes yet. I wondered if
>anyone knows what it is. It does something with color icons. I just
>want to know that it isn't a virus, and maybe that it is from an
>extension that may be causing trouble.

I'm guessing that this is the same trap that TMON Pro referred to as
_JacksonPollock.  I know that it's related to Color QuickDraw, but I
didn't bother reading through the code.

A similar question to yours:  what the hell is _Bass?  I was getting some
crashes when a program I was fixing thrashed thePort, and my stack crawl
led me through some code low in memory (very low, like $00001000 or so) that
called _Bass.

Does anyone know what _Bass is?  Can I expect to find _Trout, _Snapper,
_Haddock, _Cod, and more?

Steve Mariotti
[e-mail]

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

From: [e-mail] (Jon W{tte)
Date: 15 Sep 92 11:46:39 GMT
Organization: Royal Institute of Technology, Stockholm, Sweden

> [e-mail] (Steve Anthony Mariotti) writes:

   Does anyone know what _Bass is?  Can I expect to find _Trout, _Snapper,
   _Haddock, _Cod, and more?

Or maybe _Drums, _Guitar and _Cutup too ? :-)

- --
Jon W{tte, [e-mail], Sweden, Phone [phone]

Help eradicate FIDO-Net <-> Usenet gateways in our time!

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

From: [e-mail] (David W. Berry)
Date: 15 Sep 92 22:24:07 GMT
Organization: Greenwing Enterprises

In article <[e-mail]> Steve Anthony Mariotti,
[e-mail] writes:
>I'm guessing that this is the same trap that TMON Pro referred to as
>_JacksonPollock.  I know that it's related to Color QuickDraw, but I
>didn't bother reading through the code.
_JacksonPollock is a vectored trap that was added for 32-bit QuickDraw.

>
>A similar question to yours:  what the hell is _Bass?  I was getting some
>crashes when a program I was fixing thrashed thePort, and my stack crawl
>led me through some code low in memory (very low, like $00001000 or so)
that
>called _Bass.
Since Bass was the code name for the spline font project, I assume that
_Bass is an internal trap implementing the spline fonts.

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

From: [e-mail] (Steve Anthony Mariotti)
Date: 15 Sep 1992 20:21:10 -0500
Organization: U Texas Dept of Computer Sciences, Austin TX

In article <[e-mail]> [e-mail] (David W. Berry) writes:
>In article <[e-mail]> Steve Anthony Mariotti,
>[e-mail] writes:
>>A similar question to yours:  what the hell is _Bass?  I was getting some
>>crashes when a program I was fixing thrashed thePort, and my stack crawl
>>led me through some code low in memory (very low, like $00001000 or so)
>that
>>called _Bass.

>Since Bass was the code name for the spline font project, I assume that
>_Bass is an internal trap implementing the spline fonts.

Ok, so now the question is: Is it pronounced like the fish or the four-stringed
electric instrument?  BASE or BASS, I gotta know.

Steve Mariotti
[e-mail]



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

From: [e-mail] (Susan Margarit)
Organization: NMSU Computer Science
Date: Wed, 16 Sep 1992 03:01:45 GMT

My first impression of the meaning of _Jackson in conjunction with
crashes was that Jackson is slang for a certain portion of the male
anatomy. I kind of thought that when you get Jackson'ed, you've been
f***ed.

Jim Margarit

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

From: [e-mail] (Jamie R. McCarthy)
Organization: Kalamazoo College
Date: Wed, 16 Sep 1992 17:51:17 GMT

[e-mail] (Steve Anthony Mariotti) writes:
>...what the hell is _Bass?
>
>Ok, so now the question is: Is it pronounced like the fish or the four-stringed
>electric instrument?  BASE or BASS, I gotta know.

Like the fish.  Someone on the programming team enjoying fishing, I think.
- --
 Jamie McCarthy      Internet: [e-mail]      AppleLink: j.mccarthy
 "I don't think we should have to have them wandering the streets
  frightening women and people."                            - Pat Buchanan

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

From: [e-mail] (Owen Hartnett)
Date: 16 Sep 92 16:22:42 GMT
Organization: Camex Inc., Boston MA

The symbol JACKSON POLLOCK (note the OCK, not ACK) also appears in the
MPW ROM symbol maps, which is probably where the debuggers find it.

Note that the painter, Jackson Pollock, was famous for painting these
dripping types of modern art, which is reminiscent of certain Mac crashes,
which points out the poignancy of the label's proximity to the _Debugger
trap.

- -Owen

P.S. Thanks to my associates here, Kent Borg & Norm Speciner, who
were able to identify Mr. Pollock, who entered our conversation recently
from out of nowhere.

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

From: [e-mail] (Scott Boyd)
Date: 17 Sep 92 07:24:59 GMT
Organization: Apple Computer Inc.

In article <[e-mail]>, [e-mail] (Steve
Anthony Mariotti) wrote:
>
> In article <[e-mail]> [e-mail] (David W. Berry) writes:
> >In article <[e-mail]> Steve Anthony Mariotti,
> >[e-mail] writes:
> >>A similar question to yours:  what the hell is _Bass?  I was getting some
> >>crashes when a program I was fixing thrashed thePort, and my stack crawl
> >>led me through some code low in memory (very low, like $00001000 or so)
> >that
> >>called _Bass.
>
> >Since Bass was the code name for the spline font project, I assume that
> >_Bass is an internal trap implementing the spline fonts.
>
> Ok, so now the question is: Is it pronounced like the fish or the four-stringed
> electric instrument?  BASE or BASS, I gotta know.

_Bass is pronounced like the fish, and would rhyme with class.
Rather than looking for other fish names, you might think of
the name of other ales.  Not that you'll find any, but it might
give you a clue as to a possible origin of the name.

_Jackson is the place where QuickDraw goes when it simply cannot recover
from a serious error condition.

Hope you find this useful.

macintosh system software
[e-mail]

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

From: [e-mail] (Paul Campbell)
Date: 19 Sep 92 17:39:37 GMT
Organization: Taniwha Systems Design

In article <[e-mail]> [e-mail] (Owen Hartnett) writes:
>The symbol JACKSON POLLOCK (note the OCK, not ACK) also appears in the
>MPW ROM symbol maps, which is probably where the debuggers find it.

Let's put this one to death .... 'Jackson Pollock' was the code name for
32-bit Quickdraw, _Jackson is the entry point for the new 32-bit QD traps,
you get faults on it when you run code that uses it in systems that don't
have it. "Bass" was a code name for true-type, _Bass is the t-t trap for
the new t-t calls etc etc

A great way to snoop out new Apple code names is do dump out MacsBug's
trap symbol table and see what's in there that you don't recognize
(for example '_Egret') ....

    Paul

- --
Paul Campbell    UUCP: ..!mtxinu!taniwha!paul     AppleLink: CAMPBELL.P

           So far the Republican's economic policy is best described
             as "C'est la vie Capitalism"

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

From: [e-mail] (Raymond Suke Flournoy)
Subject: Synchronizing AppleEvents
Organization: Computer Science Department, Stanford University.
Date: Thu, 17 Sep 1992 22:17:22 GMT


  I am trying to use AppleEvents to facilitate communication between a
client and a background server application and I think that my lack of
success can be traced to a failure in sychronization between the two
processes.  Anyone out there who thinks they might be of some help?

  Right now, if the client simply launches, then kills the server, the
program works perfectly fine.  The only odd behavior is that periodic
checking of Ticks shows that the launch and kill messages are both
sent out before the server even begins.

  When a single message is added between the launch and kill, the
program no longer works as expected.  With the (Think C) debugger on,
the message will be sent and reply received as expected, but without
the debugger a system error and/or bus error occurs.

  One thought that I had was that the server needs to begin running
before the send is sent out, and that somehow the debugger slows down
the client enough that the server has time to begin, and thus
synchronization is taken care of.  But (if this is in fact the
problem) how can I force synchronization without the debugger?

  I tried inserting a WaitNextEvent statement after the launch, but
even with a sleep value of 6000 the Ticks show that only about 2-20
ticks pass during the execution of the wait.

  Does anyone have a clue what my problem could be, and/or how I need
to go about fixing it?

  Thanks for the help!


- --Raymond Flournoy
======================================================================
[e-mail]         "For reasons I cannot explain
Computer Science Dept.            There's some part of me wants to see
Stanford University, CA           Graceland"               -- P. Simon

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

From: [e-mail] (Brent)
Date: 18 Sep 92 05:36:21 GMT
Organization: Texas A&M Univ., Inc.

[e-mail] (Raymond Suke Flournoy) writes:
|
|The only odd behavior is that periodic
|checking of Ticks shows that the launch and kill messages are both
|sent out before the server even begins.
|
|  I tried inserting a WaitNextEvent statement after the launch, but
|even with a sleep value of 6000 the Ticks show that only about 2-20
|ticks pass during the execution of the wait.
|
|  Does anyone have a clue what my problem could be, and/or how I need
|to go about fixing it?

For background, I've done client-server programs using message queues
on UNIX systems (one server, n clients).  On this little server I wrote,
I started it, then started about 40 clients.  It hummed away, doing what
was asked.  Under UNIX, it's easy because processes will be blocked,
i.e., set into a non-running mode, by the operating system if no messages
are available for the program.

In the mac world, the program keeps ticking away, though.
One idea I had after I read your note would be fairly easy to do.  What
I was thinking was that you could have the client start up, and then
launch the server (was that what it was doing?).  Once the client does
that, have it wait for a "high sign."  Once the server gets going, have
it zap out an apple event to the client along the lines of a "Server
Startup" message.  This would clue the client that there is a server
out there, waiting for requests.  Then, the client may continue, sending
events to the server.

One note about the WaitNextEvent() call -- the sleep parameter, when used,
does not guarantee that the program will be idle for that number of
ticks.  It just means that the program *could* be idle that long, if other
applications needed the time.  I was caught by this, too.

Well, try my suggestion, if you want to, of course.  And whatever you
find that works, please let us know.  I, for one, am interested in
client-server operations.

- -Brent

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

From: [e-mail] (Martin Minow)
Date: 21 Sep 92 16:58:10 GMT
Organization: Apple Computer Inc., Cupertino, CA


Here's a simple recipe for doing server applications on the Mac. It
requires System 7.0 or later:

- -- the main application event loop calls WaitNextEvent with a very long
   sleep time. This will block it until something happens.
- -- the application posts asychronous read requests (PPCRead or ATPRead)
   with a completion routine.
- -- the read request should use a control block that has been extended to
   include all buffers and other variables that are needed for the request.
   The control block should also include the application A5 variable if you
   use AppleTalk.
- -- the completion routine refreshes A5 and enqueues the control block to an
   application-global "work queue." It then calls WakeUpProcess to pull the
   application out of WaitNextEvent. WakeUpProcess is only available under
   System 7.0 and later.
- -- as part of the application event loop process, the application checks
   for a non-empty work queue. If there is something, it dequeues it,
   processes it, and -- when complete -- re-posts the control block for
   another asychronouse read.

Actually, the sequence is probably:

- -- post asychronous read.
- -- when the read completes, process request.
- -- post asychronous write (reply).
- -- when the write completes, post asychronous read.

The above is a quick sketch. Any real server application will be more
complex. For example, if you're using PPCToolbox, your primary cycle
goes between PPCInform, PPCRead, PPCWrite, PPCEnd, and back to PPCInform.

One advantage of doing most of the work in the main event loop (rather
than in completion routines) is that you can debug most of your application
using a "real" debugger (Think Debugger).

Hope this helps.

Martin Minow
[e-mail]

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

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