On 19/04/2010 13:05, Simon McVittie wrote: > On Sun, 18 Apr 2010 at 15:36:20 +0200, Dario Freddi wrote: > >> unit tests. As you all probably know, setting up a test environment for >> Telepathy is everything but trivial, and that's also why tp-qt4 has a >> dedicated test library. >> > The test library is really not as big as all that; it's also not API-stable > (if it was, it'd be part of telepathy-qt4!). I wouldn't recommend relying on > it externally. > Yes, the test library itself is not - the problem is the whole glib suite, and the setup part (fake dbus daemon, etc), which definitely requires some magic at buildtime. Given that tp-kde has many different components, having a shared library (either shipped with Tp-Qt4 or residing in KDE's svn) is quite a strict requirement for us - it's not very feasible to copy the whole code over each component.
About API stability, of course I'd try to reach a somehow stable API (that's also the reason why I added a GlibTest class), that's a strict requirement even if it would end up in KDE svn. At this stage, my main concern is the following: if I managed to reach some quality requirements, would this be material for Tp-Qt4 or should I move this code to KDE and create a test library over there? P.S.: I would probably create a small library in KDE anyway, as I have to sandbox nepomuk as well, but my main concern is avoiding code duplication. > S > _______________________________________________ > telepathy mailing list > telepathy@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/telepathy > -- ------------------- Dario Freddi Software Engineer at Collabora Ltd. Email: dario.fre...@collabora.co.uk Mobile: +39 3337131075 GPG Key Signature: 511A9A3B
signature.asc
Description: OpenPGP digital signature
_______________________________________________ telepathy mailing list telepathy@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/telepathy