With 135074 no regressions.
With 135087, I see the following regressions:
FAIL: ext/malloc_allocator/deallocate_local.cc (test for excess errors)
WARNING: ext/malloc_allocator/deallocate_local.cc compilation failed to produce
executable
FAIL: ext/mt_allocator/deallocate_local-2.cc (test for excess errors)
WARNING: ext/mt_allocator/deallocate_local-2.cc compilation failed to produce
executable
FAIL: ext/mt_allocator/deallocate_local-4.cc (test for excess errors)
WARNING: ext/mt_allocator/deallocate_local-4.cc compilation failed to produce
executable
FAIL: ext/mt_allocator/deallocate_local-6.cc (test for excess errors)
WARNING: ext/mt_allocator/deallocate_local-6.cc compilation failed to produce
executable
FAIL: ext/mt_allocator/deallocate_local-8.cc (test for excess errors)
WARNING: ext/mt_allocator/deallocate_local-8.cc compilation failed to produce
executable
FAIL: ext/new_allocator/deallocate_local.cc (test for excess errors)
WARNING: ext/new_allocator/deallocate_local.cc compilation failed to produce
executable
FAIL: ext/throw_allocator/deallocate_local.cc (test for excess errors)
WARNING: ext/throw_allocator/deallocate_local.cc compilation failed to produce
executable
with the .log file saying similar for all errors:
/tmp/ccDpHWIJ.s: Assembler messages:^M
/tmp/ccDpHWIJ.s:788: Error: can't resolve `.LC2' {.rodata.str1.2 section} -
`__ZNSbIcSt11char_traitsIcEN9__gnu_cxx16malloc_allocatorIcEEE4_Rep20_S_empty_rep_storageE'
{.bss._ZNSbIcSt11char_traitsIcEN9__gnu_cxx16malloc_allocatorIcEEE4_Rep20_S_empty_rep_storageE
section}^M
/tmp/ccDpHWIJ.s:788: Error: expression too complex^M
compiler exited with status 1
output is:
/tmp/ccDpHWIJ.s: Assembler messages:^M
/tmp/ccDpHWIJ.s:788: Error: can't resolve `.LC2' {.rodata.str1.2 section} -
`__ZNSbIcSt11char_traitsIcEN9__gnu_cxx16malloc_allocatorIcEEE4_Rep20_S_empty_rep_storageE'
{.bss._ZNSbIcSt11char_traitsIcEN9__gnu_cxx16malloc_allocatorIcEEE4_Rep20_S_empty_rep_storageE
section}^M
/tmp/ccDpHWIJ.s:788: Error: expression too complex^M
FAIL: ext/malloc_allocator/deallocate_local.cc (test for excess errors)
There's only the PR36090 simplify_plus_minus change between these two revisions
on the 4.3 branch. So, the CONST change now causes a wrap of a MINUS between
two symbols to _different_ sections, which must not happen. Curiously, the
same change on HEAD doesn't exhibit these regressions. Author of patch CC:ed.
--
Summary: [4.3 Regression] Fix for PR 36090 causes libstdc++
regressions
Product: gcc
Version: 4.3.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: rtl-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: hp at gcc dot gnu dot org
GCC host triplet: x86_64-unknown-linux-gnu
GCC target triplet: cris-axis-elf
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36182
