Hi all On Wed, Jul 06, 2011 at 09:00:42AM +0200, Salvatore Bonaccorso wrote: > Hi all > > On Mon, Jul 04, 2011 at 11:25:33AM +0200, Salvatore Bonaccorso wrote: > > Hi all > > > > On Sat, Jul 02, 2011 at 07:53:09PM +0300, Dimitrios Eftaxiopoulos wrote: > > > Στις Σαβ 02 Ιουλ 2011 18:41:18 Salvatore Bonaccorso γράψατε: > > > Hello Salvatore, > > > > > > > I'm cc'ing Dimitrios if he has an idea on this. mathgl should be in a > > > > much stable shape now as it was previously. > > > > > > After the last nmu's by Steve Langasek, mathgl should build much more smoothly > > > on all architectures. > > > > > > > Dimitrios do you have some > > > > insights on possible ABI break inbetween 1.10 -> 1.11? > > > > > > Sorry, I don't know. > > > > I now got down with snapshot.d.o packages, and indeed udav starts > > fine until 1.10.2.1-2 but with 1.11.0.1-1, udav has the first time > > > > udav: Symbol `mgls_base_cmd' has different size in shared object, consider re-linking > > > > So did indeed mathgl broke it's ABI there and needed to bump the > > soname? > > Julien, or someone else from the release team, could you add a > removal-hint for udav 0.6.3-2 in wheezy as it is due to this completly > non-functional (without rebuild) currently in testing? > > udav 0.7.1.2 is now in unstable, except on kfreebsd-amd64 due the > binutils problems [1]. > > [1] http://bugs.debian.org/632671 udav 0.7.1.2 is now in wheezy, so this request can be closed. Btw, from upstream I got the following: > Possible yes. I did some changes which may break ABI. > > I hope that new version (v.2) of MathGL will be much more stable in > ABI. It will be based on C-functions for most of plots, data handling > and so on. Regards Salvatore
Attachment:
signature.asc
Description: Digital signature