Laurenz Albe writes:
> I came up with the attached patch set.
I did some more work on the comments, adjusted a couple of places that
could be simplified, and pushed it.
> I used two separate patches for clarity and ease of review, but both
> should get backpatched.
I didn't really love the "fix
On Sat, Sep 13, 2025 at 05:00:20PM -0400, Tom Lane wrote:
> Laurenz Albe writes:
> > I came up with the attached patch set.
>
> I did some more work on the comments, adjusted a couple of places that
> could be simplified, and pushed it.
My apologies for the silence here.
The timing of the event
On Fri, 2025-09-12 at 10:07 -0400, Tom Lane wrote:
> Dominique Devienne writes:
> > > This DOES look like a bug, no? I've done regexes for a long time,
> > > and these two forms should be equivalent IMHO. --DD
>
> Yeah, I agree it's busted. You can use EXPLAIN VERBOSE to see the
> translated-to-
On Fri, Sep 12, 2025 at 9:35 PM Dominique Devienne wrote:
>
> On Fri, Sep 12, 2025 at 3:29 PM Dominique Devienne
> wrote:
> > On Fri, Sep 12, 2025 at 3:24 PM Dominique Devienne
> > wrote:
> > > On Fri, Sep 12, 2025 at 3:11 PM Dominique Devienne
> > > wrote:
> >
> > > So I don't see how my `.
On Sat, 2025-09-13 at 17:00 -0400, Tom Lane wrote:
> Laurenz Albe writes:
> > I came up with the attached patch set.
>
> I did some more work on the comments, adjusted a couple of places that
> could be simplified, and pushed it.
Thank you!
> > I used two separate patches for clarity and ease o
On Fri, 2025-09-12 at 21:53 -0400, Tom Lane wrote:
>
> Ah, got it. But this logic definitely deserves more comments.
> What do you think of something like
>
> if (pchar == ']' && charclass_start > 2)
> {
> /* found the real end of a bracket pair */
>
On Fri, 2025-09-12 at 20:12 -0400, Tom Lane wrote:
> I had not particularly studied the new charclass-parsing logic.
> Looking at it now, this bit further down (lines 871ff) looks
> fishy:
>
> if (pchar == ']' && charclass_start > 2)
> charclass_depth--;
> e
Laurenz Albe writes:
> On Fri, 2025-09-12 at 20:12 -0400, Tom Lane wrote:
>> Should not we be setting charclass_start to 1 after incrementing
>> charclass_depth?
> What I call "charclass depth" is misleading, I am afraid.
> Really, it should be "bracket depth". Only the outermost pair of bracket
Laurenz Albe writes:
> On Fri, 2025-09-12 at 10:07 -0400, Tom Lane wrote:
>> The _ and % are not getting converted to their POSIX equivalents
>> ("." and ".*").
> Indeed, and I have to take the blame for introducing a bug in a minor
> release :^(
> The attached patch should fix the problem.
I ha
On Fri, Sep 12, 2025 at 4:07 PM Tom Lane wrote:
> Dominique Devienne writes:
> >> This DOES look like a bug, no? I've done regexes for a long time,
> >> and these two forms should be equivalent IMHO. --DD
>
> Yeah, I agree it's busted. You can use EXPLAIN VERBOSE to see the
> translated-to-POSIX
On 2025-Sep-12, Tom Lane wrote:
> Dominique Devienne writes:
> >> This DOES look like a bug, no? I've done regexes for a long time,
> >> and these two forms should be equivalent IMHO. --DD
>
> Yeah, I agree it's busted. You can use EXPLAIN VERBOSE to see the
> translated-to-POSIX pattern, and i
Dominique Devienne writes:
>> This DOES look like a bug, no? I've done regexes for a long time,
>> and these two forms should be equivalent IMHO. --DD
Yeah, I agree it's busted. You can use EXPLAIN VERBOSE to see the
translated-to-POSIX pattern, and it's wrong:
regression=# explain verbose with
On Fri, Sep 12, 2025 at 3:54 PM jian he wrote:
> > > select v from t where v similar to 'foo[\d\w]_%';
> > > select v from t where v similar to 'foo[[[:digit:]][[:word:]]]_%';
> The above two examples are the same, per
> (Table 9.21. Regular Expression Class-Shorthand Escapes)
Of course.
> my g
On Thu, Sep 4, 2025 at 5:03 PM Dominique Devienne wrote:
> OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> KO with 16.10 and 17.6 (and beta3 too, released at the same time)
I've tracked down the regression to this particular query, FWIW:
select rolname, rolsuper, rolinherit, rolcreater
On Fri, Sep 12, 2025 at 3:29 PM Dominique Devienne wrote:
> On Fri, Sep 12, 2025 at 3:24 PM Dominique Devienne
> wrote:
> > On Fri, Sep 12, 2025 at 3:11 PM Dominique Devienne
> > wrote:
>
> > So I don't see how my `... where v similar to 'foo[\d\w]_%'` is incorrect.
> > So again, is this a bug
On Fri, Sep 12, 2025 at 2:45 PM Laurenz Albe wrote:
> On Fri, 2025-09-12 at 14:22 +0200, Dominique Devienne wrote:
> > On Thu, Sep 4, 2025 at 5:03 PM Dominique Devienne
> > wrote:
> > > OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> > > KO with 16.10 and 17.6 (and beta3 too, released
On Fri, Sep 12, 2025 at 3:24 PM Dominique Devienne wrote:
> On Fri, Sep 12, 2025 at 3:11 PM Dominique Devienne
> wrote:
> So I don't see how my `... where v similar to 'foo[\d\w]_%'` is incorrect.
> So again, is this a bug / regression or not? Thanks, --DD
If I use (x|y) instead of [xy] it see
On Fri, Sep 12, 2025 at 3:11 PM Dominique Devienne wrote:
> On Fri, Sep 12, 2025 at 2:45 PM Laurenz Albe wrote:
> > You don't show us that data that match the pattern in 17.5, but
> > not in 17.6. Unless you show us a counterexample, I'd say that
> > the behavior in 17.6 is correct.
I've reread
On Fri, 2025-09-12 at 14:22 +0200, Dominique Devienne wrote:
> On Thu, Sep 4, 2025 at 5:03 PM Dominique Devienne wrote:
> > OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> > KO with 16.10 and 17.6 (and beta3 too, released at the same time)
>
> I've tracked down the regression to this pa
On Thu, 2025-09-04 at 17:28 +0200, Dominique Devienne wrote:
> But this is a patch. Patches are not supposed to break things.
Right. But the PostgreSQL developers aren't perfect either.
If you are interested in getting this fixed, make the problem
reproducible for us.
Yours,
Laurenz Albe
On Thu, Sep 4, 2025 at 5:18 PM Laurenz Albe wrote:
> On Thu, 2025-09-04 at 17:03 +0200, Dominique Devienne wrote:
> > OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> > KO with 16.10 and 17.6 (and beta3 too, released at the same time)
> >
> > Which for a minor patch is a surprise to us, n
On Thu, Sep 4, 2025 at 5:12 PM Adrian Klaver wrote:
> On 9/4/25 08:03, Dominique Devienne wrote:
> > But now, we've ascertain that a particular test is:
> > OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> > KO with 16.10 and 17.6 (and beta3 too, released at the same time)
> How did you
On Thu, 2025-09-04 at 17:28 +0200, Dominique Devienne wrote:
> But this is a patch. Patches are not supposed to break things.
Right. But the PostgreSQL developers aren't perfect either.
If you are interested in getting this fixed, make the problem
reproducible for us.
Yours,
Laurenz Albe
On Fri, Sep 5, 2025 at 1:10 AM Ron Johnson wrote:
> Dumb question: did you read the release notes?
Well, now I have. Carefully. Nothing stands out.
There's RLS mentioned, but for indexing only.
So no smoking gun. We'll have to investigate... :(
> In our case, 17.6 "pg_dump --schema-only" broke i
On Thu, Sep 4, 2025 at 5:31 PM Dominique Devienne wrote:
> On Thu, Sep 4, 2025 at 5:12 PM Adrian Klaver
> wrote:
> > On 9/4/25 08:03, Dominique Devienne wrote:
> > > But now, we've ascertain that a particular test is:
> > > OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> > > KO with 16
On 9/4/25 08:03, Dominique Devienne wrote:
In the same vein as v18betas breaking our unit-tests...
(we still don't understand, haven't looked much though).
OK, above it's about a major upgrade. So that's one thing.
But now, we've ascertain that a particular test is:
OK with 16.9 and 17.5 (we c
On Thu, 2025-09-04 at 17:03 +0200, Dominique Devienne wrote:
> OK with 16.9 and 17.5 (we cannot test on beta2 anymore)
> KO with 16.10 and 17.6 (and beta3 too, released at the same time)
>
> Which for a minor patch is a surprise to us, not in a good way.
>
> So... Did anything change around RLS i
27 matches
Mail list logo
