From: j DOT aldrich6 AT genie DOT com
Message-Id: <199607230249.AA293040144@relay1.geis.com>
Date: Tue, 23 Jul 96 02:34:00 UTC 0000
To: djgpp AT delorie DOT com
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Subject: Re: Error compiling Emacs 19.3

Reply to message 5813537    from J DOT ALDRICH6 AT GE on 07/18/96  8:29PM


>>> gcc.exe: installation problem, cannot exec `c:/djgpp/bin\ld.exe': Not
enough
>>> memory (ENOMEM)

>fiddling with them anyway.  Actually, I deleted my emacs src directory
>to save space <g> - I'll have to extract and rebuild it in order to test
>things again.  When that's done I'll get back to you.  :)

Well, I ended up having to rebuild emacs anyway, due to an incorrectly
applied patch to libc.a, so I had a chance to examine this problem further.
It seems that the ld crash occurs *only* when there is not enough free
*physical* RAM.  It has nothing at all to do with EMM386, except as
regards the memory it uses.  Here are some round figures; I'll play with
them some more tonight to narrow it down some:

go32-v2 reports:  5572K physical, 11062K virtual     --- Result:  CRASH
go32-v2 reports:  5940K physical,  9704K virtual     --- Result:  CRASH
go32-v2 reports:  6988K physical,  9704K virtual     --- Result:  works ok

Again, this appears to be related to the amount of free physical RAM
remaining when ld.exe is called by gcc.exe.  I am using CWSDPMI v2,
but the same crash also occurred under Windows 3.1 DOS shell.

Well, folks?  ;)

John

P.S.:  I just upgraded to 24 MB of RAM  (Yay!!), so I don't think I'll be
having this particular problem again.  Nevertheless, it'd be nice to
know why it happened.  :)