This is the mail archive of the
mailing list for the GCC project.
Re: [patch] bootstrap failure in obj-c++
- From: fjahanian <fjahanian at apple dot com>
- To: Andrew Pinski <pinskia at physics dot uc dot edu>
- Cc: mrs at apple dot com (Mike Stump), toa at pop dot agri dot ch (Andreas Tobler), gcc-patches at gcc dot gnu dot org (GCC Patches), geoffk at apple dot com (Geoffrey Keating), pinskia at gmail dot com (Andrew Pinski)
- Date: Fri, 3 Nov 2006 07:59:35 -0800
- Subject: Re: [patch] bootstrap failure in obj-c++
- References: <200611030035.kA30ZhAX013805@localhost.localdomain>
On Nov 2, 2006, at 4:35 PM, Andrew Pinski wrote:
This is obvious as cp_finish_file is named
and is called via LANG_HOOKS_WRITE_GLOBALS.
Well, actually, they are both wrong, or at least incomplete. The
effect is to move metadata writing up earlier, thus breaking things.
This is wrong. Geoff, did you build obj-c++? Did you test it?
Fariborz tells me that the testsuite should catch such things.
I did run the testsuite on powerpc-darwin after this patch and I got
the same results as before the other patch except for one small
testsuite bug which I was going to commit when I got home tonight.
Mainline test suite does not have the objective-c++ specific tests
which expose the problem. I sent a patch a while back (one month ago
with an objective-c++ test case). Try that test case.