Coin, Sorry for being late, i'm way too busy these days...
Guus Sliepen <[EMAIL PROTECTED]> writes: > I disabled OPCODE support in the ode 0.5 package, because the code is > broken on 64 bit architectures. Ode 0.6 is out already, and it has a > working OPCODE for 64 bit architectures, however it is not compatible > with 0.5, and upstream did not bump the soname. My idea is to build a > static library of 0.6 only. However, currently all packages in Debian > depending on ODE seem to use 0.5, and they will become unbuildable if I > upload 0.6. We should really have upstream do some more efforts to handle the soname bump correctly. Can't you bump it yourself and ask him to patch 0.6, which is to be considered buggy ? > Marc, I Cc'ed this to you, I'd like your comments. Is python-soya's > upstream aware of the issues with OPCODE in 0.5, and/or are they > planning to move to 0.6? Is it possible for python-soya to link with a > static library? No, upstream people are unaware of it. Soya was recently fixed by a bugreporter, so some bugs may still remain. Upstream authors are quite willing to accept patches, and the latest is already included, but there is absolutly no work done on 64bits architectures, as noone has such a machine to test it. I'm asking them for the 0.6 move, but the ODE-master of Soya seems to be away on holidays, so we'll have to be patient. In the past, linking both with static and dynamic librarie to create a new library led to horrible results, so i'd like to avoid this situation. Remember i was the one who strongly asked for a dynamic libode in the past. Stay tuned... -- Marc Dequènes (Duck)
pgpbhsjRBCVEQ.pgp
Description: PGP signature