Mark, On Sat, Nov 18, 2006 at 11:52:55AM +0000, Mark Purcell wrote: > On Thursday 16 November 2006 06:03, Steve Langasek wrote: > [...] > > I don't know yet why this define isn't propagating; it seems to be part of > > debian/patches/libgsm.patch though, and there's probably a race condition > > in the autotools handling that's preventing it from getting where it needs > > to be. In a rebuild test on amd64, config.log looks right to me, yet it's > > not *consistently* ending up in the generated src/twinkle_config.h, because > > there's a race condition with whether src/twinkle_config.h gets regenerated > > after autoheader is run.
> I think I have tracked it down. > debian/patches/libgsm.patch was only applied to configure.in, the necessary > changes to configure wern't in the patch. > As a result, if autoconf was called and configure was regenerated it would > dynamically link libgsm, but if configure wasn't regenerated twinkle would > try and both dynamically and static link libgsm, hence the race condition. In my testing, configure was always regenerated but it wasn't guaranteed that *autoheader* would be called at the right time to regenerate the twinkle_config.h.in file. To completely fix this race condition, I think you need to handle both the autoheader-generated and the autoconf-generated pieces at patch time before calling configure. Cheers, -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]