atrick Bernard wrote: >> Hello Patrick, >> >> On Sunday 18 April 2010, 05:49, Patrick Bernard wrote: >> > If unpublished interfaces can be changed at any time in any release, is >> > it advisable to use an unpublished interface in released code ? >> > >> > Besides, is it possible, when compiling the code, to be warned that >> > unpublished interfaces are used ? >> >> this kind of questions have been discussed several times on this (and, if >> IRRC, others OOo) mailing lists, you can search the archives. >> This is the kind of topic where the thread ends up open, with no clear >> answer, because different OOo developers have different opinions about what >> should be understood under "unpublished" (may be from "this is unstable API >> I plan to change it any time" to "I just plan to enhance it by adding >> further functionality, so I leave this unpublished"). >> So a warning when using unpublished API does not sound very realistic. > > > In any case, using an unpublished API means that the code could stop working > with a newer version of OpenOffice.org (maybe leading to crashes), so it > would be useful to detect it at compile time or at runtime.
A tool to check used APIs for unpublished ones sounds reasonable, but it's not really doable. At least usage of unpublished or deprecated "old style" services can't be checked as no code is generated for them and thus nothing is visible from them in the client code (no includes etc.). All in all I think it's better to rely on the developers ability to know what (s)he is doing. As Ariel stated, developers have different opinions about what "deprecated" or "unpublished" means. Removing deprecated APIs is something that I would do only for a major release. Changing "unpublished" APIs should not be done in a micro release. But that's only my personal POV. Regards, Mathias -- Mathias Bauer (mba) - Project Lead OpenOffice.org Writer OpenOffice.org Engineering at Sun: http://blogs.sun.com/GullFOSS Please don't reply to "[email protected]". I use it for the OOo lists and only rarely read other mails sent to it. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
