Celestin Apprentice 5/Apprentice-Release5.iso/Information/CSMP Digest/volume 1/csmp-v1-193.txt

Text File | 1992-12-31 | 48.5 KB | 1,288 lines | [TEXT/R*ch]

C.S.M.P. Digest             Tue, 20 Oct 92       Volume 1 : Issue 193

Today's Topics:

    MacsBug symbols from assembler code?
    # of files
    more on sublaunching
    animating palettes



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] (Peter Andersen)
Subject: MacsBug symbols from assembler code?
Date: 20 Aug 92 15:57:33 GMT
Organization: DAIMI: Computer Science Department, Aarhus University, Denmark

How do I make a (global) symbol known to MacsBug?
I have tried adding

   DC.B $8C, 'MySymbol'
   DC.W $0000

after the last RTS in the code of the procedure, as described in
the MacsBug manual (Appendix G, p.118, MacsBug version 6.1 manual),
but the symbol is not known to MacsBug, when I try to debug the code.

I am using MPW 3.2.2, System 7.0.1 (US), MacsBug version 6.2b2 on a
Macintosh IIx.

I have also tried {MPW}:Examples:AExamples:Sample.a, but although it
tries to add MacsBug symbols with the macro BdgInfo, again these symbols
are not known to my MacsBug.

Am I missing something?

Peter Andersen, e-mail: [e-mail]

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

From: [e-mail] (Quinn)
Date: 21 Aug 92 03:09:41 GMT
Organization: The University of Western Australia

In article <[e-mail]> Peter Andersen,
[e-mail] writes:
>Am I missing something?

Hmm.  I had great fun getting MacsBug symbols working consistently
in my latest Asm project.  Try using the following macro (with
apologies to those people whose systems don't expand tabs
properly)...

- ----------------------------------------------
    macro
    MacsBug    &Name,&rts:int
    lclc    &oldstr
    gbla    &debug

    if &debug then

    if &rts then
    rts                ; force rts in specific cases
    endif

    dc.b    &len(&name)+$80        ; length of string + $80 marks symbol
&oldstr    setc    &setting('string')
    string    AsIs
    dc.b    '&upcase(&Name)'    ; define the string AsIs
    string    &oldstr
    align                ; pad to word boundary
    dc.w    0            ; no literals

    endif

    endm
- ----------------------------------------------

... which you invoke as.
        MacsBug         Fred
        MacsBug         Jim,1
at the end of routine Fred that ends in an RTS (or JMP (A0) or
RTD) or routine Jim that doesn't end in one of the above.

However theres still the problem of whether MacsBug will actually
see the symbol.  Sometimes it does, sometimes it doesn't.  I'm not
actually sure what it depends on but I think it has something to do
with what code has been executed recently.  One thing that I found
that helps is doing an 'hx' into the heap that the symbol is in.
But if all else fails you can at least search for the symbol string
(-:

Quinn "The Eskimo!"      <[e-mail]>     "Support HAVOC!"
Department of Computer Science, The University of Western Australia
  -- MacsBug:  Don't leave home without it.
  -- This macro is for Pascal; converting it to C is left as an
     exercise to the reader.  (-:

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

From: [e-mail] (Charlie Reiman)
Date: 21 Aug 92 17:19:11 GMT
Organization: Apple Computer Inc., Cupertino, CA

[e-mail] (Quinn) writes:

>In article <[e-mail]> Peter Andersen,
>[e-mail] writes:
>>Am I missing something?

>Hmm.  I had great fun getting MacsBug symbols working consistently
>in my latest Asm project.  Try using the following macro (with
>apologies to those people whose systems don't expand tabs
>properly)...

>----------------------------------------------
>    macro
>    MacsBug    &Name,&rts:int
>    lclc    &oldstr
>    gbla    &debug

>    if &debug then
>
>    if &rts then
>    rts                ; force rts in specific cases
>    endif
>
>    dc.b    &len(&name)+$80        ; length of string + $80 marks symbol
>&oldstr    setc    &setting('string')
>    string    AsIs
>    dc.b    '&upcase(&Name)'    ; define the string AsIs
>    string    &oldstr
>    align                ; pad to word boundary
>    dc.w    0            ; no literals
>
>    endif
>
>    endm

This is almost correct :-) If you have the Macsbug manual, the definition
of a procedure is on p. 408. Follow the bouncing ball if you have a copy,
otherwise just mumble through the quote below.

Begin quote:

A procedure is defined as follows:
 o LINK A6 - This instruction is optional; if it is missing, the start
   of the proche procedure is assumed to be immediately after the preceding
   procedure, or at the start of the heap block.
 o Procedure code
 o RTS or JMP (a0) or RTD
 o Procedure name
 o Procedcure constants

End quote.

So the most likely reason you aren't always seeing your symbols is because
you don't have LINK/UNLK pairs. (I usually don't use them since most asm
code I write is just glue. Stack frames are overkill for me. I'm making
the assumption you don't use them much either.)

Second problem: You macro only covers one of the 4 styles of symbols.
To paraphrase The Book:

Valid symbol characters: [a-zA-Z0-9_%. ] Space is only allowed to pad out
fixed length symbols

8  character names: First byte is $20-$7f or $a0-$ff. If the high bit is
set, ignore it. The next character must have its high bit clear.

16 character names: First byte is $20-$7f or $a0-$ff, as for 8 byte
symbols. The second byte will have its high bit set, however. This symbol
style is used for Object Pascal. The first 8 bytes are the method, the
second 8 the class.

Still with me? Good. Now we get to the styles you probably want to use:

Short variable length symbol: First byte is $81-$9F. This is the length
with the high bit set. This is immediately followed by the name itself.

Long variable length symbol: First byte is $80, followed by a length byte,
followed by the actual name.

For both variable length symbols, its a good idea to place the constant
data marker after the symbol. The Book sez: "The first word after the
name specified how many bytes of constant data are present. If there
are no  constants, a length of 0 must be given."

In short, a good symbol would be defined like this:

MyFunkyProc   PROC
          link a6,#0
          unlk a6
          rts
          dc.b 11+128,'MyFunkyProc'
          dc.w 0000
          ENDP

You might be able to use only the long variable length style in your macro
to make life easier on yourself. I haven't tried this but its worth a shot.

Have f
Bus Err while trying to execute from $50FFC001.


- --
Charlie Reiman - Speaking as an individual, not for Apple Computer.
[e-mail]
        "NEW! Posting Lite! 98% fact free!"

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

From: [e-mail] (Dave Ely)
Date: 22 Aug 92 02:47:24 GMT
Organization: Symantec / Peter Norton

[e-mail] (Peter Andersen) writes:

 > How do I make a (global) symbol known to MacsBug? I have tried
 > adding
 >
 >    DC.B $8C, 'MySymbol'
 >    DC.W $0000
 >
 > after the last RTS in the code of the procedure, as described in the
 > MacsBug manual (Appendix G, p.118, MacsBug version 6.1 manual), but
 > the symbol is not known to MacsBug, when I try to debug the code.
 >
 > I am using MPW 3.2.2, System 7.0.1 (US), MacsBug version 6.2b2 on a
 > Macintosh IIx.

You need something like this:

   DC.B $8C, 'MySymbol'
   dc.b $0
   DC.W $0000

The MacsBug symbol has to be padded to an even word length so that the
data length offset (DC.W $0000) is on an even address. There is a
macro which comes with the samples for for the MPW assembler which
shows how to create MacsBug strings. I think it's called DbgStr.

- --
 ______________________________________________________________
    Dave Ely          \   Internet:  [e-mail]
    Symantec Corp.     \             [e-mail]
    Peter Norton Group  \ AppleLink: Ely.D

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

From: [e-mail] (Peter Andersen)
Date: 24 Aug 92 08:15:54 GMT
Organization: DAIMI: Computer Science Department, Aarhus University, Denmark

In <[e-mail]> I wrote:

>How do I make a (global) symbol known to MacsBug?
>I have tried adding

>   DC.B $8C, 'MySymbol'
>   DC.W $0000

>after the last RTS in the code of the procedure, as described in
>the MacsBug manual (Appendix G, p.118, MacsBug version 6.1 manual),
>but the symbol is not known to MacsBug, when I try to debug the code.

>I am using MPW 3.2.2, System 7.0.1 (US), MacsBug version 6.2b2 on a
>Macintosh IIx.

To sum up: The length byte should be $80 + <length of symbol>.
Here it should have been $88.

HOWEVER, the MPW assembler has default STRING pascal, meaning that it
inserts the length byte itself.
Thus one should use $80 in the first byte, mening that the length is in
next byte, and then let the assmbler calculate the length:

   DC.B $80, 'MySymbol'
   DC.W $0000

This works for me.
Someone also mentioned that the DC.W should be word aligned, meaning that
a padding $0 should be inserted:

   DC.B $80, 'MySymbol'
   DC.B $00
   DC.W $0000

Peter Andersen

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

From: [e-mail] (Jim Reekes)
Date: 24 Aug 92 19:22:58 GMT
Organization: Apple Computer, Inc.

In article <[e-mail]>, [e-mail]
(Quinn) wrote:
>
> In article <[e-mail]> Peter Andersen,
> [e-mail] writes:
> >Am I missing something?
>
> Hmm.  I had great fun getting MacsBug symbols working consistently
> in my latest Asm project.  Try using the following macro (with
> apologies to those people whose systems don't expand tabs
> properly)...
>
> ----------------------------------------------
>     macro
>     MacsBug    &Name,&rts:int
>     lclc    &oldstr
>     gbla    &debug
>
>     if &debug then
>
>     if &rts then
>     rts                ; force rts in specific cases
>     endif
>
>     dc.b    &len(&name)+$80        ; length of string + $80 marks symbol
> &oldstr    setc    &setting('string')
>     string    AsIs
>     dc.b    '&upcase(&Name)'    ; define the string AsIs
>     string    &oldstr
>     align                ; pad to word boundary
>     dc.w    0            ; no literals
>
>     endif
>
>     endm
> ----------------------------------------------

Here's the macro I currently use.



   macro
    symbol      &routineName=&SysMod,&size=0
    lclc        &name, &oldString

    IF &TYPE('DEBUG') <> 'UNDEFINED' THEN
        rts
        if &routineName[1:1] = '''' then
            &name: setc &eval(&routineName)
        else
            &name: setc &routineName
        endif

        &oldString: setc &setting('string')
        string asis
        if &len(&name) < 32 then
            dc.b &len(&name)+$80,'&name'
        else
            dc.b $80,&len(&name),'&name'
        endif
        string &oldString
        align
        dc.w &size
    endif

    endm


Just becareful where you add this.

It needs the RTS because sometimes you write Asm routines without a
Link/Unlink.  The usage should be like this.

SomeRoutine   proc
              ...
              ...
              ...
              symbol "SomeRoutine"
              endp


- -----------------------------------------------------------------------
Jim Reekes, Polterzeitgeist  |     Macintosh Toolbox Engineering
                             |          Sound Manager Expert
Apple Computer, Inc.         | RAll opinions expressed are mine, and do
20525 Mariani Ave. MS: 81-KS |   not necessarily represent those of my
Cupertino, CA 95014          |       employer, Apple Computer Inc.S

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

From: [e-mail] (Kelly Schwarzhoff)
Subject: # of files
Date: 24 Aug 92 05:47:45 GMT
Organization: HoloNet (BBS: 510-704-1058)

Does anyone know the maximum # of files that can exist in a folder/disk/
computer (I'm interested in all 3)?  In addition, is there anything that
will crash before this limit is hit so that there is a smaller practical
limit (for example, if the operating system works with 2^32 files, but
Microsoft word crashes with more than 2^16 files).


- --
- -------------------------------------------------------------------------
Kelly Schwarzhoff
Internet: [e-mail]   Fidonet: 1:161/445.0
- -------------------------------------------------------------------------

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

From: [e-mail] (Frank Probul)
Date: 24 Aug 92 11:14:56 GMT
Organization: Technische Universitaet Muenchen, Germany


In article <[e-mail]>, [e-mail] (Kelly Schwarzhoff) writes:
|> Does anyone know the maximum # of files that can exist in a folder/disk/
|> computer (I'm interested in all 3)?  In addition, is there anything that
|> will crash before this limit is hit so that there is a smaller practical
|> limit (for example, if the operating system works with 2^32 files, but
|> Microsoft word crashes with more than 2^16 files).
|>
|>
There are some limits in the hierarchial file system:
(The old file System "MFS" on system 3.2 and prior will not be discussed)

- --------------
V O L U M E S :
- --------------
- - all these limits are for volumes and partitions. So if you have a 120 MB
  hard disk partitioned to 40 MB and 80 MB, it's the same as using
  physically different HDs with 40 and 80 MB.

- - each volume is partitioned to a maximum of 2^16 = 65536 blocks.
  The blocksize is a multiple of 512 Bytes.
  The result: on a 30 MB-hard disk each block is 512 bytes, on larger volumes
  the blocksize will be increased. On a 200 MB hard disk the blocksize will be
  3.5 KBytes, on a CD-ROM or 650 MB hard disk the size is about 10 KBytes.

- - for each fork of a file a total block is used! If you put a 200 Bytes
  TeachText document on a CD-ROM, it occupies 10 KBytes disk space.
  (A MPW document has a resource fork AND a data fork, so 20 KBytes of
  disk space are used)

- - The number of forks on a volume is thus limited to 2^16 = 65536.
  (if each fork is smaller than the blocksize, 65536 files can be there).
  Some files have only a data fork, some files have only a resource fork,
  some have both.

- - the max block size is 2^32 Bytes = 4 GigaBytes! The theoretical limit of
  each partition is thus 256 TeraBytes. This is the suported size of the
  File Manager.

- - There seems to be a bug in the calulating of sizes of the
  Finder, so when using volumes larger than 2 GB some unusual results occur.
  Sometimes the free space will be negative, or the finder won't copy files,
  because of rounding errors.
      (If the hard disk is 4 GBytes and 50 KBytes, the 4 GigaBytes will be
      rounded away resulting in a volume size of 50 kBytes. Now th Finder
      won't copy a 200 KBytes-File, because he thinks that there isn't
      enough memory...)

- - There are some limits of the SCSI-protocol, but I don't know the exact values.

- - The max size of an AppleShare Volume
  (AppleShare 2.0, 3.0 and Personal FileSharing) is limited to 4 GB.

- - The max size of an Novell file server mounted on Macs seems to be 2 GB.
  The Novell-server can handle larger volumes, but the chooser extensions
  won't mount such volumes.

- - The maximum number of files servers which are mounted at the same time
  is limited by the machine type:
  (here I'm not sure, the limits are documented in Inside Mac 5:File Manager)
  Mac Plus: 3 Servers
  Mac SE: 8 Servers
  other Macs: more...

- - The number of volumes mounted on each server is unlimited.

- - THE CONCLUSION:
  1) Don't use partitions larger than 2 GBytes!
  2) Partition your volumes if you don't want to waste disk space.

- ---------------
F O L D E R S :
- ---------------
- - The max number of files in a folder is 32768 but the recommended limit
  is about 1000 files, because opening such a volume takes a lot of time.
  Perhaps there can be problems using the standard open/close-dialog because
  the list manager data structure is limited to 32 KBytes. If buffering the
  file names with 30 characters each the limit will be reached when using
  about 1000 files.

- - The max number of items in the Apple Menu folder is limited to 32.000, but
  only the alphabetically 52 first items will be displayed in the Apple Menu.

- -----------------
M A C H I N E S :
- -----------------
_ the limits of machines are the limits of their volumes and folders.
  The number of partitions is unlimited, but the number of open files is
  limited. So if this limit is reaches, no more desktop files can be opened
  an then thye volume can not be mounted.
  I think System 7 handles the number of FCBs (File Control Blocks)
  dynamically.


Regarding you concerns about limitations caused by applications:
  (:-Be careful using MicroSoft-Applications... :-)




Frank Probul
Emanuelstr. 17, D-8000 Munich 40, Germany

AppleLink: Probul.F
internet:  [e-mail]

Munich University of Technology
Department of Computer Science
Germany
- ---------------------------------------------

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

From: [e-mail] (Howard Cheng)
Date: 25 Aug 92 00:33:07 GMT
Organization: Auspex Systems, Santa Clara

In article <[e-mail]> [e-mail] (Frank Probul) writes:
>- There seems to be a bug in the calulating of sizes of the
>  Finder, so when using volumes larger than 2 GB some unusual results occur.
>  Sometimes the free space will be negative, or the finder won't copy files,
>  because of rounding errors.
>      (If the hard disk is 4 GBytes and 50 KBytes, the 4 GigaBytes will be
>      rounded away resulting in a volume size of 50 kBytes. Now th Finder
>      won't copy a 200 KBytes-File, because he thinks that there isn't
>      enough memory...)

Actually, we just tried this a little while ago.  We made a VLP (Very Large
Partition, > 4GB) on the UNIX server and then mounted that volume onto the
desktop using NFS/Share (v. 1.1.2) and the Mac was able to read the correct
size of the volume (some ungodly large number).  Other information: Mac IIci,
System 7.0.1, Tuner 1.1.1.

Howard


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

From: [e-mail] (Kelly Schwarzhoff)
Organization: HoloNet (BBS: 510-704-1058)
Date: Wed, 26 Aug 1992 17:56:01 GMT

> - each volume is partitioned to a maximum of 2^16 = 65536 blocks.
>   The blocksize is a multiple of 512 Bytes.
>   The result: on a 30 MB-hard disk each block is 512 bytes, on larger
> volumes
>   the blocksize will be increased. On a 200 MB hard disk the
> blocksize will be
>   3.5 KBytes, on a CD-ROM or 650 MB hard disk the size is about 10
> KBytes.

Is there anyway to get around this limit? I have a 600MB drive and have it
formatted so that each block is 10k and can't find a way to have smaller
blocks.  Is the limit of 2^16 due to the operating software or just the
software that formats the drive.

The reason I'm asking this is that we're setting up a large FirstClass BBS
software and it stores every message as a file and you cannot have the
messages in seperate volumes / partitions.  I'm running System 7.0 on a IIfx.


- --
- -------------------------------------------------------------------------
Kelly Schwarzhoff
Internet: [e-mail]   Fidonet: 1:161/445.0
- -------------------------------------------------------------------------

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

From: [e-mail] (Keith Rollin)
Organization: Taligent
Date: Wed, 26 Aug 1992 20:18:19 GMT

In article <[e-mail]>, [e-mail] (Kelly
Schwarzhoff) writes:
>
> > - each volume is partitioned to a maximum of 2^16 = 65536 blocks.
> >   The blocksize is a multiple of 512 Bytes.
> >   The result: on a 30 MB-hard disk each block is 512 bytes, on larger
> > volumes
> >   the blocksize will be increased. On a 200 MB hard disk the
> > blocksize will be
> >   3.5 KBytes, on a CD-ROM or 650 MB hard disk the size is about 10
> > KBytes.
>
> Is there anyway to get around this limit? I have a 600MB drive and have it
> formatted so that each block is 10k and can't find a way to have smaller
> blocks.  Is the limit of 2^16 due to the operating software or just the
> software that formats the drive.
>
> The reason I'm asking this is that we're setting up a large FirstClass BBS
> software and it stores every message as a file and you cannot have the
> messages in seperate volumes / partitions.  I'm running System 7.0 on a IIfx.

The limit of 2^16 allocation blocks is due to the HFS format. All the data
structures used to keep track of where files are located on the volume use
2-byte values for the block number.

The only way to get more room from your 600MB drive is to partition it into two
or more smaller logical volumes. Or rewrite the File Manager.

- --
Keith Rollin
Phantom Programmer
Taligent, Inc.


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

From: [e-mail] (Bill York)
Date: 26 Aug 92 21:02:43 GMT
Organization: Lucid, Inc.

In article <[e-mail]> [e-mail] (Kelly Schwarzhoff) writes:

   From: [e-mail] (Kelly Schwarzhoff)

   Is there anyway to get around this limit? I have a 600MB drive and have it
   formatted so that each block is 10k and can't find a way to have smaller
   blocks.  Is the limit of 2^16 due to the operating software or just the
   software that formats the drive.

   The reason I'm asking this is that we're setting up a large FirstClass BBS
   software and it stores every message as a file and you cannot have the
   messages in seperate volumes / partitions.  I'm running System 7.0 on a IIfx.

Can the software store messages in sub-folders of the main
message-area folder?  You could then make these sub-folders be aliases
pointing to folders on different partitions.  Or does the BBS software
circumvent your access to the filesystem completely?

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

From: [e-mail] (Mark Lanett)
Date: 27 Aug 92 04:16:19 GMT
Organization: Apple Computer Inc., Cupertino, CA

[e-mail] (Kelly Schwarzhoff) writes:

>Is there anyway to get around this limit? I have a 600MB drive and have it
>formatted so that each block is 10k and can't find a way to have smaller
>blocks.  Is the limit of 2^16 due to the operating software or just the
>software that formats the drive.

>The reason I'm asking this is that we're setting up a large FirstClass BBS
>software and it stores every message as a file and you cannot have the
>messages in seperate volumes / partitions.  I'm running System 7.0 on a IIfx.

Not that I'm aware of, but you can always partition the drive anyway and stick
the non-messages on the other partition. That should drop it to 5K or so.
- --
Have a bajillion brilliant Jobsian lithium licks.
Mark Lanett, NOT speaking for anyone. Personal opinion only.

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

From: [e-mail] (Kent Borg)
Date: 27 Aug 92 18:38:19 GMT
Organization: Camex Inc., Boston MA

In article <[e-mail]> [e-mail] writes:
>In article <[e-mail]> [e-mail] (Kelly Schwarzhoff) writes:
>
>   From: [e-mail] (Kelly Schwarzhoff)
>
>   Is there anyway to get around this limit? I have a 600MB drive and have it
>   formatted so that each block is 10k and can't find a way to have smaller
>   blocks.  Is the limit of 2^16 due to the operating software or just the
>   software that formats the drive.
>
>   The reason I'm asking this is that we're setting up a large FirstClass BBS
>   software and it stores every message as a file and you cannot have the
>   messages in seperate volumes / partitions.  I'm running System 7.0 on a IIfx.
>
>Can the software store messages in sub-folders of the main
>message-area folder?  You could then make these sub-folders be aliases
>pointing to folders on different partitions.  Or does the BBS software
>circumvent your access to the filesystem completely?
>


The 2^16 limit is an HFS limit.

No, Firstclass does not circumvent the file system, aliases are not
implemented by the file system--which is what annoys the makers of
Firstclass.  I think they wish they were on a Unix box.  To use
aliases would require extra irksome work on their part.

As a result Firstclass does not know about aliases and isn't likely to
soon--at least judging from a phone conversation I once had with
someone fairly technical there.  It seems that they are getting burned
by the Mac's lack of preemptive multitasking and generally starting to
fight with their server platform choice, not cooperate with it.  I am
worried.  (Notice sometime how the server's screen--1.0 at least--can
get painted and scroll *while* a menu is down.  Horrors.)

Disclaimer: I have not yet installed Firstclass version 2, my only
experiance is with the first version.

As for how to make partitioning choices, put your system folder and
any non-FirstClass stuff on a small partition (or separate disk) and
give the the Firstclass database (I think just the "FirstClass Post
Office" folder) as much space as you can.  Your limit is likely going
to be the 64K-blocks wall.  A big allocation size will keep the larger
files from gobbling up so many of those blocks.  Yes, seeing 10K
vanish with each little message hurts, but total bytes is not the
first limit you are going to hit with a little-file-laden BBS on a big
disk.  Yes, using two 300 Meg partitions would get you farther, but
only if you are allowed.  A 600 partition is always going to hold at
least as much as a 300--and usually far more.

Weasel Words: I am sure I have the term "allocation size" slightly
wrong, but it shouldn't demolish my underlying points.

Coward's Words: Please be gentle when it does become obvious that I
have no valid underlying points.


- --
Kent Borg                                   [e-mail] or [e-mail]
                                            H:[phone]  W:[phone]
As always, things look better when some costs are left out.
                              -Economist 3-28-92 p. 94

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

From: [e-mail] (Steve Dorner)
Date: 27 Aug 92 22:07:40 GMT
Organization: Qualcomm, Inc., San Diego, CA

[e-mail] (Kent Borg) writes:
>No, Firstclass does not circumvent the file system, aliases are not
>implemented by the file system--which is what annoys the makers of
>Firstclass.  I think they wish they were on a Unix box.  To use
>aliases would require extra irksome work on their part.

Sounds like they have an Attitude.  Of course, they're not the
only ones.  I can think of one particular development system that ALSO
doesn't have a clue about aliases.  It's called MPW, and the company
is some fruit name.  Ah, well, they probably weren't Apple Partners, and
didn't get any warning of what was going to happen in System 7. :-)

Aliases are implemented in an unpleasant way for any app that doesn't
just use standard file, and it's no particular shame that aliases
aren't supported in such apps.  Of course, as the years wear on, one's
patience wears thinner.
- --
Steve Dorner, Qualcomm, Inc.
Yes, I'm still working on Eudora and it's still free.

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

From: [e-mail] (Harry Morris)
Subject: more on sublaunching
Date: 15 Sep 92 16:59:55
Organization: Thinking Machines Corporation, Cambridge MA, USA


thanks to everyone who sent suggestion son sub-lanuching.  I think I've
gotten sub-launching pretty much working now!

However, I've got a document I need to launch, I know it's name, type, and
creator.  I've set up the AppParmHandle to point to my file, and that works
fine.  Two problems remain though.  I'd appreciate any help I can get.

1) I need to figure out which app to launch, given a type and creator.
This is probably normally done by the finder.  I think the necessary info
is in the desktop file, but TN 29 says not to rely on that.  Is there any
kosher way to do this?

2) When I launch an app (say, MSWord) with a doc, it opens the doc, and
becomes the frontmost app.  But then if I switch back to my app, and try to
launch another doc in the same app, the app becomes frontmost, but the
second do is not loaded.  Any idea how do this?  I've seen OnLocation do
it.

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

From: [e-mail] (Steve Orchard)
Date: 16 Sep 92 13:37:20 GMT
Organization: Martin Marietta

In article <[e-mail]>, [e-mail] (Harry Morris) writes:
>
> 1) I need to figure out which app to launch, given a type and creator.
> This is probably normally done by the finder.  I think the necessary info
> is in the desktop file, but TN 29 says not to rely on that.  Is there any
> kosher way to do this?

To do this open the resource fork of the deskTop file, then load the APPL
resource.  After that then search for the creator, once the creator is found
it tells you the applications name and Dir ID.  You can use this information to
launch the appliction.  If system 7 is running then you should use AppleEvents
to send the finder a open doc message.

> 2) When I launch an app (say, MSWord) with a doc, it opens the doc, and
> becomes the frontmost app.  But then if I switch back to my app, and try to
> launch another doc in the same app, the app becomes frontmost, but the
> second do is not loaded.  Any idea how do this?  I've seen OnLocation do
> it.

This is harder what the finder does is fakes a menu hit in the open... under
the file menu, then it bypasses the file picker.  If you use AppleEvents under
system 7 you won't have this problem.

Good Luck
Steve Orchard
[e-mail]


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

From: [e-mail] (Grobbins)
Date: 17 Sep 92 05:44:58 GMT
Organization: Apple Computer Inc., Cupertino, CA

In article <[e-mail]> [e-mail] (Steve Orchard) writes:
>In article <[e-mail]>, [e-mail] (Harry Morris) writes:
>> 1) I need to figure out which app to launch, given a type and creator.
>> This is probably normally done by the finder.  I think the necessary info
>> is in the desktop file, but TN 29 says not to rely on that.  Is there any
>> kosher way to do this?
>To do this open the resource fork of the deskTop file, then load the APPL
>resource.  After that then search for the creator, once the creator is found
>it tells you the applications name and Dir ID.  You can use this information to
>launch the appliction.  If system 7 is running then you should use AppleEvents
>to send the finder a open doc message.

Opening the resource fork of the desktop file is a bad idea since
the resource map may not be accurate.

Sending an "open doc" event to the Finder doesn't do anything; you can
send an Open Selection Finder event, though the current Finder does not
give any feedback to indicate whether the event succeeded, so this is
not a great approach.

Under System 6, there is no good, documented, reliable way to find
the creator of a document short of searching the disk.  Under System 7,
the Desktop Manager makes it easy.  Pasted below is some code which
shows both the "safe" (Desktop Manager) and "unsafe" (Desktop file)
ways of finding the application for a document.

Launching an application should be done by calling LaunchApplication
(shown in the Process Management chapter of Inside Mac VI, or by
calling _Launch under System 6, as documented in the Sublaunch Tech Note.)
To open a document with an application, pass LaunchApplication a coerced
'odoc' event in the appParameters field; see the snippet "LaunchWithDoc"
(available at ftp.apple.com) for an example.

>> 2) When I launch an app (say, MSWord) with a doc, it opens the doc, and
>> becomes the frontmost app.  But then if I switch back to my app, and try to
>> launch another doc in the same app, the app becomes frontmost, but the
>> second do is not loaded.
>This is harder what the finder does is fakes a menu hit in the open... under
>the file menu, then it bypasses the file picker.  If you use AppleEvents under
>system 7 you won't have this problem.

Under System 7, the system will pull "puppet strings" for you by
simulating an Open menu selection if your application sends an 'odoc'
event but the target is not high-level event aware.  So go ahead and
send 'odoc's to everyone (except when launching; then, coerce the
odoc and include it with the LaunchApplication call.)

Grobbins         [e-mail]

Usual disclaimers apply.

- ---


  FUNCTION FindAppFromDoc(docFSSpec: FSSpec; VAR appFSSpec: FSSpec): OSErr;
  { first find an app on the same volume; next try local volumes;
    next try remote volumes }

  { to find applications on volumes without a Desktop Database, we rely on
    the Desktop file.  This is bad, bad, as explained below. }

  LABEL 1;
  TYPE
    volumePass = (document, local, remote, done);

    { record in APPL resource of Desktop file }
    APPLTypePtr = ^APPLType;
    APPLType = RECORD
                   creator: OSType;
                   dirID: LongInt;
                   name: Str63;    { actually, name size is variable }
                 END;
  VAR
    myDTPBRec: DTPBRec;
    docFInfo, dummyFInfo: FInfo;
    retCode: OSErr;
    theVolumePass: volumePass;
    theVolumeCounter: INTEGER;

    myHParamBlockRec: HParamBlockRec;
    myGetVolParmsInfoBuffer: GetVolParmsInfoBuffer;
    remoteFlag, foundFlag: BOOLEAN;

    resRefNum: INTEGER;
    resHandle: Handle;
    theAPPLType: APPLType;
    theAPPLTypePtr: APPLTypePtr;
    offset: LONGINT;
  BEGIN

    theVolumePass := document;
    foundFlag := FALSE;

    retCode := FSpGetFInfo(docFSSpec, docFInfo);
    IF retCode <> noErr THEN GOTO 1;

    REPEAT  { for each kind of pass }

      theVolumeCounter := 0;

      REPEAT  { for each volume, except on first pass }

        { first, find the vRefNum of the volume to be checked }

        IF theVolumePass = document THEN { use doc's own volume }
          myDTPBRec.ioVRefNum := docFSSpec.vRefNum
        ELSE
          BEGIN
            theVolumeCounter := theVolumeCounter + 1;

            { loop through all drives }
            myHParamBlockRec.ioCompletion := NIL;
            myHParamBlockRec.ioNamePtr := NIL;
            myHParamBlockRec.ioVRefNum := 0;
            myHParamBlockRec.ioVolIndex := theVolumeCounter;
            retCode := PBHGetVInfoSync(@myHParamBlockRec);

            IF retCode = nsvErr THEN Leave; { checked all drives }
            IF retCode <> noErr THEN GOTO 1;

            { is this one remote? }
            myHParamBlockRec.ioBuffer := @myGetVolParmsInfoBuffer;
            myHParamBlockRec.ioReqCount :=

              Sizeof(myGetVolParmsInfoBuffer);
            retCode := PBHGetVolParmsSync(@myHParamBlockRec);
            IF (retCode <> noErr) THEN GOTO 1;

            { volume should be remote only on remote pass, else skip it }
            IF (myGetVolParmsInfoBuffer.vMServerAdr <> 0)
              <> (theVolumePass = remote) THEN Cycle;

            { found a volume of interest }
            myDTPBRec.ioVRefNum := myHParamBlockRec.ioVRefNum;

          END; { ELSE }

        myDTPBRec.ioNamePtr := NIL;

        { Get the application name and directory from the desktop database, if
          the volume has one.  Otherwise, check the desktop file. }

        retCode := PBDTGetPath(@myDTPBRec);
        IF (retCode = noErr) AND (myDTPBRec.ioResult = noErr) AND
          (myDTPBRec.ioDTRefNum <> 0) THEN
          BEGIN
            { get application name and directory }
            myDTPBRec.ioNamePtr := @appFSSpec.name; { change the file spec name}
            myDTPBRec.ioIndex := 0;
            myDTPBRec.ioFileCreator := docFInfo.fdCreator;
            myDTPBRec.ioCompletion := NIL;

            retCode := PBDTGetAPPLSync(@myDTPBRec);
            IF retCode = noErr THEN
              BEGIN
                appFSSpec.parID := myDTPBRec.ioAPPLParID;
                appFSSpec.vRefNum := myDTPBRec.ioVRefNum;
                foundFlag := TRUE;
              END;
          END
        ELSE
          { check Desktop file if there is one on the volume }
          { WARNING:  this next section is a bad idea; leave
            it out unless you absolutely have to find the
            application.  The steps shown are officially
            undocumented and discouraged.   }

          { This is risky, because the Desktop file is
            probably already open with write priveleges by
            the Finder, and the Macintosh OS does not guarantee
            that the resource map is valid }

          BEGIN

            { root directory's dirID is always 2 }
            { application directory info is kept in APPL resource }
            { HOpenResFile with read permission should give a unique
              resource reference number, making it safe to call
              CloseResFile later. }

            resRefNum := HOpenResFile(myDTPBRec.ioVRefNum, 2,
              'Desktop', fsRdPerm);
            retCode := ResError;
            IF retCode = noErr THEN
              BEGIN
                resHandle := Get1IndResource('APPL', 1);
                retCode := ResError;

                IF (retCode = noErr) AND (resHandle <> NIL) THEN
                { just checking resHandle <> NIL is enough }
                  BEGIN
                    HLock(resHandle);
                    offset := 0;

                    { loop through all application information in the APPL }

                    WHILE (offset < SizeResource(resHandle)) AND NOT foundFlag
                      DO
                      BEGIN
                        theAPPLTypePtr :=
                           APPLTypePtr(Pointer(offset+ORD4(resHandle^)));
                        theAPPLType := theAPPLTypePtr^;

                        IF (theAPPLType.creator = docFInfo.fdCreator) AND
                          { desktop database sometimes points to phantom apps }
                          (HGetFInfo(myDTPBRec.ioVRefNum, theAPPLType.dirID,
                            theAPPLType.name,
                            dummyFInfo) = noErr) THEN { a file really is there }
                          BEGIN
                            { found the app }
                            appFSSpec.parID := theAPPLType.dirID;
                            appFSSpec.vRefNum := myDTPBRec.ioVRefNum;
                            appFSSpec.name := theAPPLType.name;
                            { really should call FSMakeFSSpec instead }
                            foundFlag := TRUE;
                          END;
                        { to find next, skip over creator code, dirID, name,
                          pad byte }
                        offset := offset + 4 + 4 + 1 + Length(theAPPLType.name);

                        { of course, MPW Pascal compiles this down to
                          ADDQ.L     #$4,D0
                          ADDQ.L     #$4,D0
                          ADDQ.L     #$1,D0
                          CLR.W      D1
                          MOVE.B     -$01AA(A6),D1
                          EXT.L      D1
                          MOVE.L     D1,D4
                          ADD.L      D0,D4                            }

                        IF (offset MOD 2) = 1 THEN offset := offset + 1;
                      END; { WHILE }
                    HUnlock(resHandle);

                    CloseResFile(resRefNum);
                  END; { IF }
              END; { IF }
          END; { ELSE }

      UNTIL foundFlag OR (theVolumePass = document);
        { OR ran out of drives, due to Leave statement }

      theVolumePass := SUCC(theVolumePass)

    UNTIL foundFlag OR (theVolumePass = done);

    IF NOT foundFlag THEN retCode := fnfErr;
  1:
    FindAppFromDoc := retCode;
  END;


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

From: [e-mail] (George Economou)
Subject: animating palettes
Date: 29 Aug 92 02:45:20 GMT
Organization: FidoNet node 1:129/120.4 - Mac For The Mind, Pittsburgh PA

I am trying to animate a palette but a few colors refuse to animate.  How can
I make it so that all colors animate?
- --
George Economou via cmhGate - Net 226 fido<=>uucp gateway Col, OH
UUCP: ...!n8emr!bluemoon!cmhgate!129!120.4!George.Economou
INET: [e-mail]

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

From: [e-mail] (Daniel Norman Johnson)
Organization: Rensselaer Polytechnic Institute, Troy, NY.
Date: Thu, 3 Sep 1992 12:17:26 GMT

In article <[e-mail]>, [e-mail] (George Economou) writes:
|> I am trying to animate a palette but a few colors refuse to animate.  How can
|> I make it so that all colors animate?

Are the colors that wont animate Black and White? If not, I have no clue
what is wrong. If so, the Palette Manager does that on purpose (so
windows, menus, etc can be drawn) and you should probably let it do so. If
you just gotta change those colors, I expect the Color Manager can do it,
but I dont have details. (I have never tried to do this)
- --
            - Dan Johnson
And God said "Jeeze, this is dull"... and it *WAS* dull. Genesis 0:0

These opinions have had all identifiying marks removed, and are untraceable.
You'll never know whose they are.

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

From: [e-mail] (George Economou)
Date: 14 Sep 92 00:13:50 GMT
Organization: FidoNet node 1:129/120.4 - Mac For The Mind, Pittsburgh PA

>We need more to go on than this.  What colors aren't animating?  Some
>code would be helpful.
>
>If you want to animate the first and last color in the color table,
>black and white...well, you just can't do it with the Palette Manager.
>You'll have to use the Color Manager instead (though I recommend a long
>hard think before deciding to do this).

I have no troubles dealing with the black and white (I don't use them for the
drawing).  What I am doing, is first getting a palette from a resource and
assigning it to a windowptr with SetPalette().  Two questions: what would the
ideal usage be, pmanimated or pmanimated+pmtolerant? and should I set the
cupdates field of SetPalette() to true or false?
I then call Activatepalette() for the windowptr.
Then, I draw into an offscreen pixmap with the same palette SetPalette()'ed to
it using rgbforecolor, not pmforecolor (I abandoned pmforecolor 'cause I
thought it might be a problem cause and I can do better effects without it).
Then I copybits() it to the windowptr previously mentioned and it works.  Then
I call animatepalette() and only about half of the colors animate, the other
half just sitting there doing nothing.  It is the same colors every time, and
they are a few reds and a few greens, mixed in with all the others. :P

Hope that helps.... thanks for helping out!
*8)
- -G     e    o   r  g eE c  o   n    o     m      o       u
:):):):)

- --
George Economou via cmhGate - Net 226 fido<=>uucp gateway Col, OH
UUCP: ...!n8emr!bluemoon!cmhgate!129!120.4!George.Economou
INET: [e-mail]

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

From: [e-mail] (Jamie R. McCarthy)
Date: 16 Sep 92 21:40:11 GMT
Organization: Kalamazoo College

[e-mail] (George Economou) writes:
>What I am doing, is first getting a palette from a resource and
>assigning it to a windowptr with SetPalette().  Two questions: what would the
>ideal usage be, pmanimated or pmanimated+pmtolerant? and should I set the
>cupdates field of SetPalette() to true or false?

Well, since there's no such usage category as pmAnimated+pmTolerant, I
guess that narrows it down.  :-)  pmAnimated will do you just fine;
there wouldn't be any point to making it tolerant or courteous, since
that only matters when two windows share a palette entry, and pmAnimated
colors are never shared between windows.

And the updates field should be TRUE, generally speaking.  If you pass
FALSE, you'll never get update events when the color environment
changes--for example, when your app goes into the background and you
lose all those nice pmAnimated colors.  (You might pass FALSE if your
app will never be backgrounded, or, say, if you're the ViewColors DA,
which just draws the color environment pmExplicitly anyway.  But
generally, pass TRUE.)

>I then call Activatepalette() for the windowptr.

That's unnecessary;  SetPalette() takes care of that for you.  (I don't
think it messes anything up, but I could be wrong.  Anyone else know?)

>Then, I draw into an offscreen pixmap with the same palette SetPalette()'ed to
>it using rgbforecolor, not pmforecolor (I abandoned pmforecolor 'cause I
>thought it might be a problem cause and I can do better effects without it).

OK.  In the Winter 92 develop, Forrest Tanaka has a little 2-page
article about the Palette Manager and off-screen pixmaps, in which he
explains that pmCourteous, pmWhite, and pmBlack usages work, but that
pmTolerant, pmAnimated, and pmExplicit do not.

I'm not sure what it would mean to SetPalette() a palette to a pixmap.
I'm presuming you have indeed set up your offscreen CGrafPort properly,
and you're passing its pointer to SetPalette().

RGBForeColor will, I believe, assign the color index based on the
current _GDevice_, so that should work if your window's already been
created, because the color environment on the _screen_ is the same as
the one offscreen.  But this will fail if things get iffy;  for example,
if your window isn't frontmost at the time.

>Then I copybits() it to the windowptr previously mentioned and it works.

Two possibilities here;  which one is the case, depends on how you made
your offscreen pixmap.  Either (1) its color table's seed is identical
to the screen's, and CopyBits is blindly putting the colors up, or (2)
the seeds are different, and CopyBits, doggedly refusing to use those
pmAnimated colors, is instead giving you very close matches to the
colors you want, so close that you don't immediately notice.

(Or, (3), you've already done what I'm about to say, and just didn't
mention it.  :-)

There's two ways to CopyBits an offscreen pixmap into pmAnimated colors:
the fast, not-so-easy way, and the slow, a-little-easier way.

The fast way is (1) above:  make _sure_ your offscreen color table and
the onscreen color table agree about the important colors.  Then, set
your offscreen pixmap's color table's ctSeed value to the screen's
pixmap's color table's ctSeed, and blast those bits to the screen.
CopyBits, when it sees equal seeds, just copies the pixels.  This is a
little harder if you want to draw in the offscreen pixmap, because
normal QuickDraw commands don't care whether the destination ctSeed is
equal to the screen's (in fact, aren't even aware of the screen at all).

The slow way is to set bit 14 of your offscreen pixmap's color table,
and set all the ColorSpec.value entries to the _palette_ entries you
want.  CopyBits will, for each source pixel, look its value up in the
table, convert that palette entry to a clut entry, and write that to the
screen.  It's probably two or three times slower (I've never tested it).
But it's easier, if you're using QuickDraw to go offscreen, because
there's none of this chicanery with ctSeeds...

>Then
>I call animatepalette() and only about half of the colors animate, the other
>half just sitting there doing nothing.

Hmmm.  Strange behavior.  You've stumped me.  :-)  If I had to guess,
I'd say the seeds were equal, and the P.M. had coincidentally given half
the colors the same clut index as palette index, so that only these were
animating...
- --
 Jamie McCarthy      Internet: [e-mail]      AppleLink: j.mccarthy
 Memo to myself:
 Do The Dumb Things I Gotta Do.
 Touch The Puppet Head.

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

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