Thread: HAVING clause and 6.3.2 release
Attached is a list of bug reports for the HAVING clause. My question is, "Do we disable the HAVING clause for 6.3.2?" The bugs are serious and cause crashes. I have looked at the issues, and the basic problems are that the aggregate logic expects to be attached to an actual field in the target list, and the HAVING clause does not properly handle non-aggregate retrictions, nor does it prevent them. COUNT(*) uses the oid of the first FROM table, so that is a problem too. I have looked at the code, but don't have time to fix it before Friday, and holding up the release for that would be silly. I don't think there is one thing wrong, but several places that have to be change to get this working solidly. Do we disable it? ---------------------------------------------------------------------------
On Wed, 15 Apr 1998, Bruce Momjian wrote:
> Attached is a list of bug reports for the HAVING clause.
>
> My question is, "Do we disable the HAVING clause for 6.3.2?" The bugs
> are serious and cause crashes.
>
> I have looked at the issues, and the basic problems are that the
> aggregate logic expects to be attached to an actual field in the target
> list, and the HAVING clause does not properly handle non-aggregate
> retrictions, nor does it prevent them. COUNT(*) uses the oid of the
> first FROM table, so that is a problem too.
>
> I have looked at the code, but don't have time to fix it before Friday,
> and holding up the release for that would be silly. I don't think there
> is one thing wrong, but several places that have to be change to get
> this working solidly.
>
> Do we disable it?
Yes...but disabling means that it *will not* be available until
v6.4...no v6.3.3 :)
Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
On Thu, 16 Apr 1998, The Hermit Hacker wrote: > On Wed, 15 Apr 1998, Bruce Momjian wrote: > > > Attached is a list of bug reports for the HAVING clause. > > > > My question is, "Do we disable the HAVING clause for 6.3.2?" The bugs > > are serious and cause crashes. > > > > I have looked at the issues, and the basic problems are that the > > aggregate logic expects to be attached to an actual field in the target > > list, and the HAVING clause does not properly handle non-aggregate > > retrictions, nor does it prevent them. COUNT(*) uses the oid of the > > first FROM table, so that is a problem too. > > > > I have looked at the code, but don't have time to fix it before Friday, > > and holding up the release for that would be silly. I don't think there > > is one thing wrong, but several places that have to be change to get > > this working solidly. > > > > Do we disable it? > > Yes...but disabling means that it *will not* be available until > v6.4...no v6.3.3 :) > > What about including it as an optional feature by defining something like /* #define BUGGY_HAVING_CLAUSE */ Marc Zuckman marc@fallon.classyad.com _\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ _ Visit The Home and Condo MarketPlace _ _ http://www.ClassyAd.com _ _ _ _ FREE basic property listings/advertisements and searches. _ _ _ _ Try our premium, yet inexpensive services for a real _ _ selling or buying edge! _ _\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_
> > My question is, "Do we disable the HAVING clause for 6.3.2?" The
> > bugs are serious and cause crashes.
> > Do we disable it?
> Yes...but disabling means that it *will not* be available until
> v6.4...no v6.3.3 :)
Hmm. What is the downside to leaving it in with caveats or "stay away"
warnings in the release notes? Since it didn't exist as a feature
before, the only downside I see is somewhat increased traffic on the
questions list...
- Tom
On Wed, 15 Apr 1998, Bruce Momjian wrote:
> Attached is a list of bug reports for the HAVING clause.
>
> My question is, "Do we disable the HAVING clause for 6.3.2?" The bugs
> are serious and cause crashes.
>
> I have looked at the issues, and the basic problems are that the
> aggregate logic expects to be attached to an actual field in the target
> list, and the HAVING clause does not properly handle non-aggregate
> retrictions, nor does it prevent them. COUNT(*) uses the oid of the
> first FROM table, so that is a problem too.
>
> I have looked at the code, but don't have time to fix it before Friday,
> and holding up the release for that would be silly. I don't think there
> is one thing wrong, but several places that have to be change to get
> this working solidly.
>
> Do we disable it?
>
Don't do that. If you disable it, we can't help you to correct bugs ?
Jose'
On Thu, 16 Apr 1998, Thomas G. Lockhart wrote:
> > > My question is, "Do we disable the HAVING clause for 6.3.2?" The
> > > bugs are serious and cause crashes.
> > > Do we disable it?
> > Yes...but disabling means that it *will not* be available until
> > v6.4...no v6.3.3 :)
>
> Hmm. What is the downside to leaving it in with caveats or "stay away"
> warnings in the release notes? Since it didn't exist as a feature
> before, the only downside I see is somewhat increased traffic on the
> questions list...
I liked the one suggestion about having it as a compile time
option until its fixed...
> > > > My question is, "Do we disable the HAVING clause for 6.3.2?" The > > > bugs are serious and cause crashes. > > > Do we disable it? > > Yes...but disabling means that it *will not* be available until > > v6.4...no v6.3.3 :) > > Hmm. What is the downside to leaving it in with caveats or "stay away" > warnings in the release notes? Since it didn't exist as a feature > before, the only downside I see is somewhat increased traffic on the > questions list... We could do a elog(NOTICE,...) and have a small patch to fix all the issues once we have a final fix. -- Bruce Momjian | 830 Blythe Avenue maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026 + If your life is a hard drive, | (610) 353-9879(w) + Christ can be your backup. | (610) 853-3000(h)
> > On Thu, 16 Apr 1998, Thomas G. Lockhart wrote: > > > > > My question is, "Do we disable the HAVING clause for 6.3.2?" The > > > > bugs are serious and cause crashes. > > > > Do we disable it? > > > Yes...but disabling means that it *will not* be available until > > > v6.4...no v6.3.3 :) > > > > Hmm. What is the downside to leaving it in with caveats or "stay away" > > warnings in the release notes? Since it didn't exist as a feature > > before, the only downside I see is somewhat increased traffic on the > > questions list... > > I liked the one suggestion about having it as a compile time > option until its fixed... How about an elog(NOTICE,"...") so it runs, but they see the NOTICE every time. -- Bruce Momjian | 830 Blythe Avenue maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026 + If your life is a hard drive, | (610) 353-9879(w) + Christ can be your backup. | (610) 853-3000(h)
On Thu, 16 Apr 1998, Bruce Momjian wrote:
> >
> > On Thu, 16 Apr 1998, Thomas G. Lockhart wrote:
> >
> > > > > My question is, "Do we disable the HAVING clause for 6.3.2?" The
> > > > > bugs are serious and cause crashes.
> > > > > Do we disable it?
> > > > Yes...but disabling means that it *will not* be available until
> > > > v6.4...no v6.3.3 :)
> > >
> > > Hmm. What is the downside to leaving it in with caveats or "stay away"
> > > warnings in the release notes? Since it didn't exist as a feature
> > > before, the only downside I see is somewhat increased traffic on the
> > > questions list...
> >
> > I liked the one suggestion about having it as a compile time
> > option until its fixed...
>
> How about an elog(NOTICE,"...") so it runs, but they see the NOTICE
> every time.
That works too...but how does something like that work from within
a C program? Or Perl?
> > How about an elog(NOTICE,"...") so it runs, but they see the NOTICE > > every time. > > That works too...but how does something like that work from within > a C program? Or Perl? I have disabled HAVING completely, and removed it from the features list. I think we have enough bug reports on it that allowing people to use it is really not going to give us any additional bug-fixing information. We can always release a 6.3.2 patch that will enable it when we have it working 100%. -- Bruce Momjian | 830 Blythe Avenue maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026 + If your life is a hard drive, | (610) 353-9879(w) + Christ can be your backup. | (610) 853-3000(h)