Re: plpgsql_check_function - rebase for 9.3
Petr Jelinek <pjmodos@pjmodos.net>
From: "Petr Jelinek" <pjmodos@pjmodos.net>
To: "'Tom Lane'" <tgl@sss.pgh.pa.us>
Cc: "'Pavel Stehule'" <pavel.stehule@gmail.com>, "'PostgreSQL Hackers'" <pgsql-hackers@postgresql.org>
Date: 2013-01-26T23:41:40Z
Lists: pgsql-hackers
> -----Original Message----- > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Sent: 26 January 2013 20:12 > Subject: Re: [HACKERS] plpgsql_check_function - rebase for 9.3 > > I wrote: > > [ pokes around... ] Hm, it appears that that does work on Linux, > > because for some reason we're specifying RTLD_GLOBAL to dlopen(). > > TBH that seems like a truly horrid idea that we should reconsider. > > A bit of research in the archives revealed that we're using it because back in > 2001, the lame hack that then passed for a shared-library version of python > required it: > http://www.postgresql.org/message-id/Pine.LNX.4.30.0105121914200.757- > 100000@peter.localdomain > > There was subsequent discussion of removing it, because reportedly now > (a) that's no longer the case, and (b) we need to get rid of it to allow > plpython2 and plpython3 to coexist in one session. See for instance: > http://www.postgresql.org/message- > id/1277506674.5356.27.camel@vanquo.pezone.net > > Nothing's been done about that yet, but I think that assuming that we'll be > using RTLD_GLOBAL forever would be foolish. > Ok then it was a bad idea after all. Regards Petr