Re: [patch] fix spurious SIGSEGV faults under Cygwin

[Brian Dessent <brian at dessent dot net>]

"Gautero, L. (Luca)" wrote:
> 
> Dear Mr
> 
> After giving chase to a segmentation fault error, on the internet one thread to which you contributed came out:
> 
> http://www.mail-archive.com/cygwin-patches@cygwin.com/msg03531.html
> 

Please keep discussions on the mailing list.  Do not send email directly
to me.

The issue you are referring to has been fixed for quite some time.  The
patches I posted are completely redundant and useless now.  Pretend they
do not exist.

The patch that implments this on the gdb side is
<http://sourceware.org/ml/gdb-patches/2006-02/msg00352.html>.

The patch that implements the start of the fix on the cygwin side is
<http://cygwin.com/ml/cygwin-cvs/2006-q1/msg00148.html>.

So, if you run a recent Cygwin snapshot and a build of gdb after
2006-02-20 you will not experience the problem.

> You propose some patches to solve the problem, but from what I understood these are not for end user of cygwin since they act on sources which after a normal installation are no longer there, is there an update to solve this problem?

Sorry, but that is completely bogus logic.  The issue only affects
people using GDB.  End users don't typically use GDB.  If you are
debugging something with GDB then you are probably familiar enough with
source code and how to build patch and build it.  If not, then just wait
for an updated binary of the affected programs to be released.  If you
are talking about a situation where a SEGV fault occurs outside of GDB
then this whole issue is irrelevant, because you are talking about
something unrelated.

> 
> p.s. still related (mailing list)
> http://www.nabble.com/Thread-support-in-cygwin%21-t1442219.html#a3899853

This has absolutely nothing to do with the issue you first mentioned.

> http://www.cygwin.com/ml/cygwin/2005-01/msg00527.html

This too has absolutely nothing to do with the above issue.

> The program that gives out this error uses fltk, and on the forum of the site someone quote a possible solution in changing the POSIX thread of
> WIN32 from internal to a code available online, unfortunately this is not possible under cygwin, either way probably the error would have stayed
> there...
> http://www.fltk.org/newsgroups.php?s1+gfltk.bugs+v1+T0+Qpthread

This is talking about detecting the thread library during "configure"
execution.  Again, this has absolutely nothing to do with anything you
have mentioned previously in this email, especially not the GDB issue. 
If you are building a native WIN32 application (i.e. non-Cygwin) then
using -lpthread is probably wrong.  If you are building a Cygwin
application then you must use Cygwin's pthread implmentation and you
don't need any kind of -lpthread flag at all.

Brian

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/