Re: [PATCH,4.6/4.5.1,PR42776] Implement LTO for Windows (PE-COFF) targets.

[Dave Korn <dave dot korn dot cygwin at googlemail dot com>]

On 13/04/2010 07:16, Dave Korn wrote:

>   Bootstrapped on i686-pc-cygwin against r.158195, tests in progress; 

  "make check-gcc" results only posted:

http://gcc.gnu.org/ml/gcc-testresults/2010-04/msg01205.html

  The remaining LTO failures fall into two groups: they are caused either by
use of alias and visibility attributes not supported on Windows:

> compiler exited with status 1
> output is:
> /gnu/gcc/gcc/gcc/testsuite/gcc.dg/lto/20081222_1.c:14:21: error: 'EXT_x' aliased to undefined symbol 'INT_x'
> /gnu/gcc/gcc/gcc/testsuite/gcc.dg/lto/20081222_1.c:14:21: error: 'EXT_x' aliased to undefined symbol 'INT_x'
> 
> FAIL: gcc.dg/lto/20081222 c_lto_20081222_1.o assemble, -O0 -fwhopr

... or by use of ELF-specific assembler directives in inline asms:

> compiler exited with status 1
> output is:
> /win/c/DOCUME~1/ADMINI~1/LOCALS~1/Temp/ccgoMkGU.s: Assembler messages:
> /win/c/DOCUME~1/ADMINI~1/LOCALS~1/Temp/ccgoMkGU.s:42: Warning: .type pseudo-op used outside of .def/.endef ignored.
> /win/c/DOCUME~1/ADMINI~1/LOCALS~1/Temp/ccgoMkGU.s:42: Error: junk at end of line, first unrecognized character is `m'
> 
> FAIL: gcc.dg/lto/20090914-2 c_lto_20090914-2_0.o assemble, -O0 -fwhopr

  So, I don't think they reflect any problem with the patch.  Full results to
follow when check-target completes.

  NB, something seems to have gone wrong with one of the g++ parallel
sub-testruns.  Adding in the partial results as far as it got before it
seemingly crashed gives overall totals of:

		=== g++ Summary ===

# of expected passes		20494
# of unexpected failures	11
# of unexpected successes	1
# of expected failures		58
# of unsupported tests		278

as recorded in the overall g++.sum file.  I'll re-run them just to be safe.

    cheers,
      DaveK