Hi Andrew, On Wed, Jun 15, 2011 at 01:56:23PM +1000, Andrew Bartlett wrote: > > Quoting Steve Langasek (vor...@debian.org): > > > On Wed, Jun 08, 2011 at 06:39:06AM +0200, Christian PERRIER wrote: > > > > Quoting Steve Langasek (vor...@debian.org): > > > > > On Tue, Jun 07, 2011 at 07:32:17PM +0200, Christian PERRIER wrote: > > > > > > Do you guys ahave an advice about this bug report?
> > > > > Should be fixed by setting '#define _LARGEFILE64_SOURCE' and '#define > > > > > _FILE_OFFSET_BITS 64' in the header. > > > > <mode bubulle_is_very_dumb> > > > > You mean just dropping these two lines at top of libsmbclient.h? > > > Yep. > > > > Should this be done upstream? > > > It should be discussed upstream, but it's potentially an ABI change > > > upstream > > > for those platforms which were *not* previously building with LFS support > > > so > > > more caution is required there. > > So, the attached patch shold be the fix. Let's record that. > This seems a pretty dangerous define to put into a public header, > included in other software... So in any case where it's actually a problem, the affected software is already broken when building on Debian because we're *building* libsmbclient with these flags and therefore anything building against it without also getting the LFS types is mis-built. You're perfectly right that this is an imperfect solution. A correct upstream solution would be to define private "off_t" types like the ones in zlib, and export those as the interface instead of off_t itself. But in the meantime, I think this change is the closest we can get to correct behavior in the distro, and doesn't break anything not already broken. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ slanga...@ubuntu.com vor...@debian.org -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
