Mail Archives: cygwin/2009/10/01/16:22:33
[ Sorry, my first reply to this post went astray because the OP was
crossposted to both cygwin and cygwin-developers, and cygwin-developers set a
reply-to header, and my mailer merged the two duplicates into one. ]
Vincent R. wrote:
> > Are you sure the problem doesn't rely on gcc version release with latest
> > cygwin ?
As it happens, it's not related.
> > I get a segmentation fault inside ld ...
That should never happen, no matter how badly mangled the input LD is fed,
and is always a bug when it does.
> > Will I find answers about that kind of problem on this ML ?
It's always best to start with a report here, in case it's cygwin-specific,
and then we can escalate it to the binutils list if it turns out to come from
upstream.
In this case, you found a bug that has been around since like forever, but
only come to light now that we've started really using weak symbols in anger
on the PE platform. It was a trivial thinko in looking up the default value
for the weak symbol that only shows up if you have multiple object files in
the same link that all refer to the same undefined weak.
Fixed in binutils CVS HEAD by
http://sourceware.org/ml/binutils-cvs/2009-10/msg00010.html
and should be in 2.20 when it comes out. I verified that gettext now builds
all the way to completion, although I didn't run the tests. Thanks for
finding this!
[ Since first sending this I ran enough of the tests to see that they were
generally passing, so it looks OK. ]
cheers,
DaveK
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
- Raw text -