http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48881
Peter Åstrand <astrand at cendio dot se> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |UNCONFIRMED
Resolution|WORKSFORME |
--- Comment #6 from Peter Åstrand <astrand at cendio dot se> 2011-12-19
12:23:59 UTC ---
(In reply to comment #5)
> Well, I wouldn't say that this was rude. The issue here is most likely, that
> the runtime DLLs by gcc aren't in search-path when you are trying to execute
> your built application.
This is not how Windows works. If the DLLs aren't found, the application isn't
started at all. But in this case, the binaries do start, but immediately
crashes. Thus, this cannot be an installation problem.
It's of course possible that our *build* of GCC is different in some way that
triggers this bug. But something is clearly wrong here, and as I mentioned in
the original report, 4.4.3 worked fine. We are working a Subversion based build
environment, and our build of 4.5.2 is therefore done in exactly the same way
as 4.4.3, which worked. We did however upgrade binutils from 2.20 to 2.21 in
the same step. We also migrated our thread fixes. I will attach both patches.
Can you reproduce the problem if you are using the same settings as we?
