tag:blogger.com,1999:blog-3748796162921493670.post6273403797286553508..comments2019-08-13T20:03:36.926+01:00Comments on Thom's memory leak: Deprecated features in PostgreSQL - Past to presentUnknownnoreply@blogger.comBlogger12125tag:blogger.com,1999:blog-3748796162921493670.post-90343163576228194932012-04-16T22:31:40.533+01:002012-04-16T22:31:40.533+01:00Aha! Perfect! I shall sleep soundly tonight. ;)
A...Aha! Perfect! I shall sleep soundly tonight. ;)<br /><br />And, um, get round to uploading the fixes to those env and i18n functions at some point.<br /><br />Oh, and figuring out what is actually up with overloading operators on custom domain types, and when it broke, and why. Currently have to use ||| rather than ||, which is really fiddly and annoying.<br /><br />Interesting to see the => operator being moved towards reserved status. I actually went for ==> in the end, to avoid conflicts with hstore.Rowan Tomminshttps://www.blogger.com/profile/11501625256059457499noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-41890458747481987132012-04-16T22:25:05.181+01:002012-04-16T22:25:05.181+01:00It's not what you think. Variables prefixed w...It's not what you think. Variables prefixed with a class name will no longer be validated against such a list, so you no longer have to worry about defining your class beforehand.<br /><br />So from 9.2, feel free to use:<br /><br />SET random_undefined_class.gibberish_setting = 'meow';Thomhttps://www.blogger.com/profile/02563102260614838138noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-61026383546615962682012-04-16T22:05:47.725+01:002012-04-16T22:05:47.725+01:00I was reading through this, thinking "hm"...I was reading through this, thinking "hm" and "heh" and "OK, not even sure what that means" and then I got to this at the very end:<br /><br />custom_variable_classes<br />Considered better to remove it as it only causes more maintenance with minimal benefit.<br /><br />NOOOOOOOOOOO!<br /><br />Those are essential to my pg_env functions (as in http://rwec.co.uk/pg-atom/env.html but the non-performance-sucking versions that I haven't posted online yet but are in use on production sites 8/)<br /><br />Please tell me there's some other way to make a custom session-persistent variable? And no, I won't count pl/perl and its weird global hash, whose name I can't remember this minute :P<br /><br />Or am I completely missing the point, and there's a permanently available namespace I can dump my variables in now?<br /><br />If not, I guess I'll have to get down and write a proper extension to reimplement them as "environment variables", if no-one's done so yet...<br /><br /><br />RowanRowan Tomminshttps://www.blogger.com/profile/11501625256059457499noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-54008085271340060422012-04-06T09:38:55.718+01:002012-04-06T09:38:55.718+01:00Thanks Bruce. I could have been much more extensiv...Thanks Bruce. I could have been much more extensive and covered<br />- pginterface<br />- zap_ltv<br />- datetime (as a module that provides functions)<br />- apache_logging<br />- int8 (yes, first time 64-bit integers were allowed, but relegated to a contrib module)<br />- sequence contrib module (sequence functions)<br />- "misc utils" contrib module<br />- retep (Peter Mount's own utilities... whatever happened to Peter?)<br />- erServer module (replication tools)<br /><br />... and loads more. :)<br /><br />The history of Postgres is quite interesting, especially to see how it evolved technically over time. For example, I hadn't noticed that Postgres95 was ported to Windows NT (or at least attempted). And thankfully the git repo logs contain *everything*...<br /><br />Postgres95 Beta 0.01 (Mon May 1 19:03:10 PDT 1995)<br /><br />:D<br /><br />And Bruce, I see your first patch submission back in July 1996. And a presentation you write back in 2006 about its history: http://momjian.us/main/writings/pgsql/decade.pdf<br /><br />Speaking of which, were there thoughts for Postgres's 20th anniversary in 3 years?Thomhttps://www.blogger.com/profile/02563102260614838138noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-45304857013464907762012-04-06T03:22:41.302+01:002012-04-06T03:22:41.302+01:00Wow, what a great trip back through time --- thank...Wow, what a great trip back through time --- thanks. That had to take a lot of work.Bruce Momjianhttps://www.blogger.com/profile/07875088787463864011noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-46039888873973662602012-04-04T10:14:15.770+01:002012-04-04T10:14:15.770+01:00Thanks Josh. The garbage you're referring to ...Thanks Josh. The garbage you're referring to is a custom CSS block I put in place to lay out the page into an easier to read format. But I neglected to use commented CDATA markup, which I've now put in place. I'll have to remember that next time. :)<br /><br />But as per your request, I've added it to the wiki: http://wiki.postgresql.org/wiki/Deprecated_Features<br /><br />Feel free to tidy it up or add anything.Thomhttps://www.blogger.com/profile/02563102260614838138noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-80531033126405564192012-04-04T02:31:09.584+01:002012-04-04T02:31:09.584+01:00Nice, this is a terrific resource. Thanks for com...Nice, this is a terrific resource. Thanks for compiling it! Do you think you could copy it to the PostgreSQL wiki?<br /><br />You might want to check your RSS feed, though; it's full of HTML garbage which is making your blog unreadable on Planet Postgres.Josh Berkushttps://www.blogger.com/profile/09671139717468724246noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-21171956765950331642012-04-03T09:03:39.438+01:002012-04-03T09:03:39.438+01:00@Greg
I take it you're referring to processin...@Greg<br /><br />I take it you're referring to processing literal language names case-sensitively? Yes, good point. I had forgot about that. I'll add it in. Thanks.Thomhttps://www.blogger.com/profile/02563102260614838138noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-33259546926151191622012-04-03T00:48:26.568+01:002012-04-03T00:48:26.568+01:00There's a 'new one' coming in 9.2, la...There's a 'new one' coming in 9.2, language parameter you pass on to function definition is now not going to work as it did 9.1 and previous.Greg Jaskiewiczhttps://www.blogger.com/profile/01582040055358451515noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-11787492654784626632012-04-02T22:43:29.330+01:002012-04-02T22:43:29.330+01:00There you go, updated it with time travel. :DThere you go, updated it with time travel. :DThomhttps://www.blogger.com/profile/02563102260614838138noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-47236747633530026772012-04-02T22:36:13.867+01:002012-04-02T22:36:13.867+01:00Yes, there will be things I haven't mentioned....Yes, there will be things I haven't mentioned. Time travel is ancient, in that it was dropped way back in 6.2 when we used punch cards and had to replace valves. ;)Thomhttps://www.blogger.com/profile/02563102260614838138noreply@blogger.comtag:blogger.com,1999:blog-3748796162921493670.post-64599032904885894322012-04-02T22:32:17.783+01:002012-04-02T22:32:17.783+01:00Hi Thom. Good post, but I would have liked to have...Hi Thom. Good post, but I would have liked to have seen "time travel" mentioned, if only as a historic curiosity.Peter Geogheganhttps://www.blogger.com/profile/02874568372191778321noreply@blogger.com