This is the mail archive of the
mailing list for the Java project.
Re: Patch: Override auto-guessed max cmd line length in MingW
- From: Mohan Embar <gnustuff at thisiscool dot com>
- To: java-patches at gcc dot gnu dot org
- Date: Wed, 12 Mar 2003 00:13:49 -0600
- Subject: Re: Patch: Override auto-guessed max cmd line length in MingW
- Reply-to: gnustuff at thisiscool dot com
Hi Ranjit & Andrew & Everyone,
Mohan>> +# This compensates for the fact that ltconfig chooses too high a length
Mohan>> +# for the command line.
Mohan>> +case $host_os in
Mohan>> + mingw*)
Mohan>> + lt_cv_sys_max_cmd_len=30000
Mohan>> + ;;
Ranjit> Should this not be "build_os" based? Or are they the same for
Ranjit> ltconfig? BTW, have you tried with recent MSYS 1.0.9 snapshots
I've verified that build = host = <library target> (which is i386-pc-mingw32)
for the MingW native, as well as BOTH the Linux cross and crossed-native compilers.
This has the consequence that for both the cross and crossed-native compilers,
we would set the max. command line length too conservatively low.
As such, even though my patch would work, I'd like to put it on hold until
I can figure out the right thing to do:
- fix one or more of the libtool scripts?
- fix MSYS so that the test isn't broken?
- go ahead and do my patch anyway?
I'm confused as to what the most correct / most expedient choice is.
What do you think, Ranjit? Fix MSYS?