Horton

Re: fix for BUG #3720: wrong results at using ltree

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Nikita Glukhov <n.gluhov@postgrespro.ru>
Cc: Tomas Vondra <tomas.vondra@2ndquadrant.com>, Alexander Korotkov <a.korotkov@postgrespro.ru>, Pgsql Hackers <pgsql-hackers@postgresql.org>, Oleg Bartunov <obartunov@postgrespro.ru>, Thomas Munro <thomas.munro@gmail.com>, Filip Rembiałkowski <filip.rembialkowski@gmail.com>, Teodor Sigaev <teodor@sigaev.ru>
Date: 2020-03-28T23:26:52Z
Lists: pgsql-hackers

Attachments

Nikita Glukhov <n.gluhov@postgrespro.ru> writes:
> On 24.01.2020 21:29, Tomas Vondra wrote:
>> Unfortunately, the current code is somewhat undercommented :-(

> The main problem is that no one really understands how it works now.

Indeed.  I was disturbed to realize that lquery_op.c, despite being
far from trivial code, contained NOT ONE SINGLE COMMENT before today,
other than the content-free file header and a commented-out (visibly
unsafe, too) debugging printing function.  This is a long way south
of minimally acceptable, in my book.

Anyway, I concur that Nikita's two patches are bug fixes, so I pushed
them.  Nonetheless, he *did* hijack this thread, so in hopes of restoring
attention to the original topic, here's a rebased version of the original
patch.

My main complaint about it remains the same, that it changes a
disturbingly large number of existing regression-test results,
suggesting that there's not a meeting of the minds about what
this logic is supposed to do.  Maybe it's okay or maybe it's
not, but who's going to decide?

Also, now that I've looked at it a bit more, I'd be inclined to
strip out the parts of the patch that remove setting up the
LQUERY_HASNOT flag.  Even if we're not using that right now,
we might want it again someday, and we're not saving much of
anything by introducing a minor on-disk incompatibility.

			regards, tom lane

Commits

  1. Back-patch addition of stack overflow and interrupt checks for lquery.

  2. Fix lquery's NOT handling, and add ability to quantify non-'*' items.

  3. Improve error messages in ltree_in and lquery_in.

  4. Fix lquery's behavior for consecutive '*' items.

  5. Protect against overflow of ltree.numlevel and lquery.numlevel.