Re: gold work on mingw/mingw64 support

[Vladimir Simonov <sv at sw dot ru>]

On 04/12/2011 10:06 PM, Ian Lance Taylor wrote:
> Vladimir Simonov<sv@sw.ru>  writes:
>
> > On 01/04/2011 08:35 PM, Ian Lance Taylor wrote:
> > > Vladimir Simonov<sv@sw.ru>   writes:
> > >
> > > > IMO mremap.c, pread.c, etc. are a kind of such library.
> > > > Analysis of target OS capabilities is overkill here, we
> > > > know in advance - mingw doesn't have mmap and most probably
> > > > won't have it.
> > >
> > > Sure, but there are other systems that also don't have mmap.  We don't
> > > want to test for all of them.
> > >
> > > > If __MINGW32__ looks too OS specific it may be
> > > > replaced with HAVE_NO_MMAP, HAVE_NO_CLOEXEC defined
> > > > in config/mh-mingw. Really __MINGW32__ looks quite strange
> > > > in the middle of descriptors.cc but I don't know how to do it
> > > > better - fake function in mremap.h would be also strange,
> > > > new fcntl.c - the same...
> > >
> > > I'm complaining in part about the __MINGW32__ and in part about the
> > > #ifdef.  Apart from ENABLE_THREADS, the only #ifdefs in gold are very
> > > small ones at the top of a few .cc files.  Let's keep it that way.
> > >
> > > Ian
> >
> > OK. Here is a version without __MINGW32__, with sys/mman.h and mmap
> > function presence detection and new mmap.h/mmap.c.
> > It is larger but should satisfy you requirements.
> > Could you comment?
>
> Thanks.
>
> I decided to go a different route.  It seems that mmap could in general
> fail for a file which can nevertheless be read.  So I decided to go
> ahead and change gold to handle the case of mmap failing.  This then
> turns the mingw32 code into a matter of supporting systems which don't
> provide mmap by providing a stub mmap which always fails.
>
> I committed the appended patch.  Let me know if you find problems with
> it.  This patch also cleans up mremap a bit even on systems which
> provide mmap.
>
> Ian

Thank you. I'll check the patch in couple days and report.

The only I can say from the first glance - chunk from top-level
config/mh-mingw about CXXFLAGS += -D__USE_MINGW_ACCESS is missed.
It is below.
Cause gold is the only c++ program in binutils
I see no reasons against it.

Best regards
Vladimir

--- binutils-2.21.orig/config/mh-mingw	2008-11-27 20:22:09.000000000 +0300
+++ binutils-2.21/config/mh-mingw	2011-01-05 14:01:25.012653500 +0300
@@ -2,5 +2,6 @@
 # Vista (see PR33281 for details).
 BOOT_CFLAGS += -D__USE_MINGW_ACCESS -Wno-pedantic-ms-format
 CFLAGS += -D__USE_MINGW_ACCESS
+CXXFLAGS += -D__USE_MINGW_ACCESS
 # Increase stack limit to same as Linux default.
 LDFLAGS += -Wl,--stack,8388608