Andrew Sullivan <ajs@xxxxxxxxxxxxxxx> writes: > On Wed, Nov 14, 2007 at 02:03:50PM -0500, Tom Lane wrote: >> At some point it might get integrated, but right now it seems to need >> its own release schedule. We put the core hooks in for 8.2 but the >> thing didn't actually get published for many months after that. > What I'd prefer, in any case, is what (Dunstan? I think?) proposed, which is > something like CPAN for PostgreSQL packages. Things that _have_ to be in > the core -- hooks and such like -- would be shipped with the basic code, but > everything else could be pulled from pgfoundry. Something like that could work well for modules that are moderately independent of the core database version. I'm afraid that won't be true for the plpgsql debugger, unfortunately --- it needs to know all about plpgsql's internal data structures, and thus for example will need an update every time we invent a new plpgsql statement or feature. I think ease of maintenance will dictate pulling it into the core "blob", once it's settled enough for that to be a reasonable thing to do. Otherwise the debugger will soon become a maze of twisty little #ifdefs as it tries to support multiple PG releases. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
