Re: contrib/xml2 and PG_MODULE_MAGIC
От
Andrew Dunstan
Тема
Re: contrib/xml2 and PG_MODULE_MAGIC
Дата
Msg-id
450583CE.8060607@dunslane.net
Ответ на
Re: contrib/xml2 and PG_MODULE_MAGIC (Tom Lane)
Список
Дерево обсуждения
contrib/xml2 and PG_MODULE_MAGIC Michael Fuhr <mike@fuhr.org>
Re: contrib/xml2 and PG_MODULE_MAGIC Tom Lane <tgl@sss.pgh.pa.us>
Re: contrib/xml2 and PG_MODULE_MAGIC Peter Eisentraut <peter_e@gmx.net>
Re: contrib/xml2 and PG_MODULE_MAGIC Bruce Momjian <bruce@momjian.us>
Re: contrib/xml2 and PG_MODULE_MAGIC Andrew Dunstan <andrew@dunslane.net>
Tom Lane wrote: > Michael Fuhr writes: > >> contrib/xml2 currently has PG_MODULE_MAGIC in xslt_proc.c, which >> results in a runtime error on systems that built the module without >> support for libxslt per the comments in the Makefile. Should >> PG_MODULE_MAGIC be in xpath.c instead? >> > > [ examines xml2/Makefile ... ] Yeah, sure looks that way. Will fix. > > Even more interesting would be to fix things so that xml2 gets built as > part of the regular contrib build, but I'm not sure if we're ready to > add stuff to the configure script for the sole benefit of a contrib > module. I'd be more willing to do it if there were a long-term plan for > putting libxml-dependent code into the core. Do we have a roadmap yet > for XML development? I thought there was a session at the anniversary > conference devoted to sketching one, but if anything came out of that > I missed it. > > Me too. Part of the trouble is that support for the standard will require support in the parser, so we need to decide if we want to go down that road. Personally I think we should, but it would make a significant change from what we have available now. Peter seemed a bit more ambivalent about it, though. cheers andrew
В списке pgsql-hackers по дате отправления