Newsgroups: comp.sys.mac.programmer
Path: utzoo!utgpu!cunews!bnrgate!bwdls61!usenet
From: aruigrok@bnr.ca (Adrian C Ruigrok)
Subject: Re: MacApp C++ load/dump problem
Message-ID: <1991Feb27.210002.24332@bwdls61.bnr.ca>
Sender: usenet@bwdls61.bnr.ca (Use Net)
Organization: Bell Northern Research
References: <1991Feb21.202709.16319@bwdls61.bnr.ca> <49509@apple.Apple.COM>
Date: Wed, 27 Feb 1991 21:00:02 GMT

In article <49509@apple.Apple.COM> marc@Apple.COM (Mark Dawson) writes:
>>The problem occurs when I build a MacApp program with different options
(other
>>than -debug -sym) it does not build a new dump file.  But it still knows
enough
>>to tell C++ to use the dump file on the compile.  The end result is that the
>>first file it tries to compile it gets an error because it can't find the
dump
>>file it forget to create:
>>MABuildTool - v. 2.0.1 Release 10/25/90          Start: 3:27:03 PM 2/21/91
>>
>>Target Folder: "Macintosh HD:MPW Folder:MailTool:.NoDebug Sym:"
>>Making:        MailTool.MAMake
>>Compiling:     UMailObjects.cp
>>can't open: Macintosh HD:MPW Folder:Load Files:.NoDebug Sym:UMacApp.h.dump
>>### MPW Shell - Execution of Macintosh HD:MPW Folder:MailTool:.NoDebug
>>Sym:MailTool.MakeIt terminated.
>>### MPW Shell - Execution of mabuild terminated.
>>
>>I dug about and the MakeIt file has no reference to building the dump file as
>>it should.  I delete all references to the dump file, and recompiling with
>>-debug -sym builds the dump file like a charm.  -sym alone, or anything else
>>does not.  So somewhere it is getting in a state and I don't know how to get
>>around it.  For now I gave up on load/dump facility.
>>If you have seen this but never figured it out, drop me a line.  I will let
you
>>know what I learned.
>>Thanks,
>>Adrian
>>
>>---------------------------------
>>Adrian C Ruigrok
>>Bell-Northern Research
>>aruigrok@bnr.ca
>
>Did you use the "-cplusload" MaBuild option?  When I've recompiled my system,
>using it always rebuilt the MacApp dump file.
>
>Mark
>
Ah... ya.  Remember it works for one case (-debug -sym), but in any other case
it refuses to build the dump file.
Others have suggested -Autobuild.  That rebuilt all of MacApp, as you would
expect, but still did not build the dump file!
Some other nice chap had it happen that the load dump commands were actually
generated later in the build sequence.  So I went on a hunt, and sure as hell,
there they were.  I copied them to the front of the .Makeit file and tried to
run the .MakeIt file manually.  Alas it was not to be.  I got:
':.NoDebug Sym:'mailtool.Makeit
C++ Load/Dump: UMacApp.h.dump
File "Macintosh HD:MacApp 2.0.1:Interfaces:CIncludes:UMacAppUtilities.h"; line
334 # error:  SHRT_MAX undefined
File "Macintosh HD:MacApp 2.0.1:Interfaces:CIncludes:UMacAppUtilities.h"; line
334 # error:  - of void*
File "Macintosh HD:MacApp 2.0.1:Interfaces:CIncludes:UMacApp.h"; line 473 #
error:  LONG_MAX undefined
Compiling:     UNetwork.cp
can't open: Macintosh HD:MPW Folder:Load Files:UMacApp.h.dump
### MPW Shell - Execution of :.NoDebug Sym:mailtool.Makeit terminated.

Something is still wrong.  Because of the errors it still won't generate the
thing.

I think something fundamental is going on here.  Now SetRect calls fail to set
a rect and function return calls are failing on the return.  I am sure it is
all unrelated to confuse me, but it is always nice to lump all your problems
into one to make people feel sorry for you! :)
Thanks for the help so far folks...
Ideas are still welcome on the load/dump problem.
Adrian
