Tracking down a difference in test_results for 4.3.0-20070622

Tracking down a difference in test_results for 4.3.0-20070622

[Lee Rhodes]

It appears that Tim & I are the only ones to test the 4.3.0-20070622 on
Cygwin. Although I had only 12 more GCC failures than Tim had, I thought it
might be prudent (and instructive) for me to dig a little deeper.

Comparing our gcc.log files most of my additional failures had an (internal
compiler error) notation, which did not appear at all in Tim's log.  Making
the assumption that fixing one problem can often fix multiple test failures
I focused on the first appearance of this failure.  Cleaning out the
extraneous path information and duplicate lines, this first failure pattern
for our runs can be summarized as follows:

# Lee's gcc.log:
PASS: xgcc -O0 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c

PASS: xgcc -O1 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c

FAIL: xgcc -O2 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c
  gcc.c-torture/compile/20001226-1.c: In function 'foo': 
  gcc.c-torture/compile/20001226-1.c:23: internal compiler error:
Segmentation fault

FAIL: xgcc -O3 -w -fomit-frame-pointer -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c
  gcc.c-torture/compile/20001226-1.c: In function 'foo':
  gcc.c-torture/compile/20001226-1.c:23: internal compiler error:
Segmentation fault

FAIL: xgcc -O3 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c
  gcc.c-torture/compile/20001226-1.c: In function 'foo': 
  gcc.c-torture/compile/20001226-1.c:23: internal compiler error:
Segmentation fault

FAIL: xgcc -Os -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c
  gcc.c-torture/compile/20001226-1.c: In function 'foo': 
  gcc.c-torture/compile/20001226-1.c:23: internal compiler error:
Segmentation fault

# Tim's gcc.log:
PASS: xgcc -O0 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c

PASS: xgcc -O1 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c

PASS: xgcc -O2 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c

PASS: xgcc -O3 -w -fomit-frame-pointer -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c

FAIL: xgcc -O3 -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c
  /cygdrive/c/Temp/cc27F72c.s: Assembler messages:
  /cygdrive/c/Temp/cc27F72c.s:0: Warning: end of file not at end of a line;
newline inserted
  /cygdrive/c/Temp/cc27F72c.s:17: Warning: .stabs: missing comma
  (test for excess errors)

PASS: xgcc -Os -w -fno-show-column -o 20001226-1.o
gcc.c-torture/compile/20001226-1.c


The test file 20001226-1.c goes through 6 different optimization patterns
and I have the same failure on 4 of them. Tim had a different failure on one
of the -O3 patterns.  

I don't know how to proceed from here.  I have a hunch the difference might
be some difference in library versions, but I don't know where to look.
Anybody have any ideas?

Regards,

Lee.