This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
x86 bootstrap failure
- From: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- To: gcc at gcc dot gnu dot org
- Date: Sun, 30 Dec 01 09:27:36 EST
- Subject: x86 bootstrap failure
The regression tester showed a bootstrap failure on x86 with my last checkin,
but claims it didn't bootstrap the last time either. I looked at it to
try to determine if indeed it was my change (since it didn't say if the
failure was the same the last time), but didn't get very far.
I was able to determine that the error is a miscompilation of the cmpiler
by the stage1 compiler since stage1/cc1 doesn't crash the same way.
However, I was unable to go further than that since the GDB I have doesn't
seem to support some feature that GCC now uses in Dwarf output. The two
x86 machines I usually use are both down and I can't find another GDB on
the one remaining machine. If this situation changes later in the day
I'll take a closer look.
However, I do see a warning for a variable in cfglayout.c that might be
referenced before it's defined. So it might be worth starting there.