On Tuesday 25 September 2007, Tom Albers wrote: > > > - https://bugzilla.redhat.com/show_bug.cgi?id=284171 (short version: > > > lazy symbol resolving in glibc is not threadsafe and can deadlock) This > > > bug can cause spurious deadlocks. If at all possible I'd like to see us > > > requiring this > > > glibc version (the fix was posted on Tue, 18 Sep 2007 so perhaps > > > there's a bit of dreaming involved on my side...) > > > > Anyone who can comment on this? > > The debian-team thinks we should not 'depend' on it, only make a note in > the requirements as in 'because of bug 897 we advise glibc version', where > 897 is the bugreport from glibc. I suggest we do so.
Sounds good, perhaps we should, in addition, put a workaround inside
phonon-xine and setenv("LD_BIND_NOW", "1", 1) if compiling against an older
glibc? But probably I should ask this on a technical list?
--
________________________________________________________
Matthias Kretz (Germany) <><
http://Vir.homelinux.org/
[EMAIL PROTECTED], [EMAIL PROTECTED],
[EMAIL PROTECTED]
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ release-team mailing list [email protected] https://mail.kde.org/mailman/listinfo/release-team
