On Thu, Jun 27, 2024 at 11:31 PM Peter Eisentraut <peter@eisentraut.org> wrote:
> Here is an updated patch for this.
I took a look at this. I added some tests and found a few that give
the wrong result (I believe). The new tests are included in the
attached patch, along with the results I expect. Here are the
failures:
-- inner %% matches b then zero:
SELECT U&'cb\0061\0308' LIKE U&'c%%\00E4' COLLATE ignore_accents;
?column?
----------
- t
+ f
(1 row)
-- trailing _ matches two codepoints that form one char:
SELECT U&'cb\0061\0308' LIKE U&'cb_' COLLATE ignore_accents;
?column?
----------
- t
+ f
(1 row)
-- leading % matches zero:
SELECT U&'\0061\0308bc' LIKE U&'%\00E4bc' COLLATE ignore_accents;
?column?
----------
- t
+ f
(1 row)
-- leading % matches zero (with later %):
SELECT U&'\0061\0308bc' LIKE U&'%\00E4%c' COLLATE ignore_accents;
?column?
----------
- t
+ f
(1 row)
I think the 1st, 3rd, and 4th failures are all from % not backtracking
to match zero chars.
The 2nd failure I'm not sure about. Maybe my expectation is wrong, but
then why does the same test pass with __ leading not trailing? Surely
they should be consistent.
> I have added some more documentation based on the discussions, including
> some examples taken directly from the emails here.
This looks good to me.
> One thing I have been struggling with a bit is the correct use of
> LIKE_FALSE versus LIKE_ABORT in the MatchText() code. I have made some
> small tweaks about this in this version that I think are more correct,
> but it could use another look. Maybe also some more tests to verify
> this one way or the other.
I haven't looked at this yet.
Yours,
--
Paul ~{:-)
pj@illuminatedcomputing.com