This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: egcs 1.1 diffs
- To: Joe Buck <jbuck at synopsys dot com>
- Subject: Re: egcs 1.1 diffs
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 31 Aug 1998 22:41:59 -0600
- cc: burley at gnu dot org (Craig Burley), egcs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <199809010054.RAA02684@atrus.synopsys.com>you write:
>
> > The last thing we want is to spend the next few months collectively
> > discovering all the interesting failure modes resulting from people
> > applying the 1.1 patch against 1.0.3a but not applying the RESYNC
> > stuff afterwards!!
>
> It would appear that if the four files in RESYNC are not resync'd,
> the only possible errors are that the French and German language modes
> for texinfo may not work, and there may be a problem with "unprotoize".
> None of the compilers can be affected.
Right, but unprotoize might not build. Unfortunately I didn't get a
chance to check that today. I'm downloading egcs-1.0.3a from the ftp
site now so that I can get started from fresh trees.
> It appears that the problem with patch was that the original files did
> not end with a newline character.
That was the case for README.gcc
The unprotoize problem is that patch simply puts the #define in the
wrong place. It will work with unidiff format, but only the most
recent versions of patch will grok that format.
Similarly, the language support may keep things from building, which
I need to check. If we assumed gnu-patch, then we could make a useable
diff.
I seriously thought about not including the diffs (as Craig mentioned
they are huge). In fact, I really haven't made a decision either way
yet pending more feedback and information I get from my own tests
using the diff files to upgrade.
jeff