Anssi, Thanks. The comments were helpful and I've looked at the patch enough (again, I'm not it's author) that I think I can clean up the comments, break out the tests, and do a better job communicating exactly what it's up to with around a day's worth of effort. I'm gonna try to fit that into my schedule early next week... that is, unless someone else better prepared than me speaks up and wants to get involved.
Everything else you said is terrifying. Phill On Thursday, August 2, 2012 2:00:58 PM UTC-7, Anssi Kääriäinen wrote: > > On 2 elo, 19:17, Phill Tornroth <[email protected]> wrote: > > I just submitted a comment about some testing I did that led to > discovering > > that in 1.4.x, certain queries with null checks in them produce much > less > > efficient SQL than they did in prior versions. > > > > A patch that was originally submitted a couple years ago and revised > > recently by it's author fixes the original problem (submitted against > 1.1) > > and the new badness that came along with 1.4.x: > > > > https://code.djangoproject.com/ticket/10790 > > > > That said, the ticket is really old, so I wanted to see if there are > some > > non-obvious reasons for resisting moving forward with a merge, and > discuss > > the possibility of this landing in a 1.4 release as opposed to only on > > master. The regression that 1.4.x introduced makes it near impossible > for > > me to move my project past 1.3.x at the moment, and as I mentioned in > the > > ticket... this patch ought to speed up lots of queries for basically > anyone > > with a serious project. > > > > So... what can I do to help? > > I have added some comments to the ticket. > > Not related to this ticket in particular: the more general problem > with ORM fixes is that there seems to be nobody around who understand > the ORM completely. The code is complex, and it is hard to see what > happens, and especially why things are done in the way they are done. > This is the reason why many ORM tickets are stalled - it is really, > really hard to see if a patch is correct, as understanding the whole > logic in the ORM is so hard. > > - Anssi > -- You received this message because you are subscribed to the Google Groups "Django developers" group. To view this discussion on the web visit https://groups.google.com/d/msg/django-developers/-/A9BKpMHGzJgJ. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/django-developers?hl=en.
