Thread: Re: [pgsql-advocacy] Increased company involvement
> -----Original Message----- > From: Marc G. Fournier [mailto:scrappy@postgresql.org] > Sent: 06 May 2005 16:04 > To: Dave Page > Cc: Marc G. Fournier; Magnus Hagander; Robert Treat; Tom > Lane; Josh Berkus; pgsql-hackers@postgresql.org > Subject: Re: [pgsql-advocacy] [HACKERS] Increased company involvement > > On Fri, 6 May 2005, Dave Page wrote: > > > - Who/how will the release processes for all these seperate > projects be > > coordinated? > > Who does now? As far as I know, PLs or contrib files > *aren't* tested by > the regression tests, so, at best, they are getting 'spotty > testing' right > now when we release ... we know they build, that's it. And > that won't > change before/after ... Andrew is looking to add PL testing > to the build > farm, something that isn't done now ... Yes, but isn't the point of the so-called WTKS to pull in other projects like PL/R, libpqxx and a range of other external projects from places like Gborg? We have precisely zero control over their quality. > > - How will users be sure that the external projects are of > the quality > > we expect of PostgreSQL? > > Again, how are they sure now? > > If you are referring to my thought to include stuff like JDBC > or libpqxx, > then as I've already stated, it will be their responsibility > to work with > us if they want to be included ... if we are ready to release > 8.1, and > they don't set down a 8.1 tag that we can 'pull from', they won't be > included in that release ... we're not going to chase them down ... So we could easily end up with one package being included in one release, but not the next, then included in the next two, then dropped from a couple? Users will go loopy trying to figure that out! Regards, Dave.
On Fri, 6 May 2005, Dave Page wrote: > Yes, but isn't the point of the so-called WTKS to pull in other projects > like PL/R, libpqxx and a range of other external projects from places > like Gborg? We have precisely zero control over their quality. Course we have control over it ... if it isn't up to snuff, we just don't include it ... its not much different then if we were to pull them all into our core CVS, but nobody ever tests them when we release ... other then 'ensuring it builds', unless someone actively tested libpqxx, or JDBC, or ODBC when they were in our "core CVS", those went out with the "presumption" of the "same quality as the rest of the code", but they could have had the biggest bugs in them that nobody would know about ... > So we could easily end up with one package being included in one > release, but not the next, then included in the next two, then dropped > from a couple? Users will go loopy trying to figure that out! True ... there has to be some sort of decision process for what to include other then just arbitrarily adding things ... personally, I'm only currently looking at our current core CVS, with 'external stuff' being something we *could* do ... some stuff, I'm fairly confident could be easily/safely done, like JDBC, since those folks are active on these lists ... I don't know if libpqxx folks are around here, but if they were, one would expect them to make their voice heard ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664