Thread

1. citext_1.out, citext.out confusing comment

   jian he <jian.universality@gmail.com> — 2025-12-09T05:07:29Z

   hi.
   
   While working on CAST DEFAULT, I found out citext_1.out, citext.out comments are
   confusing.
   
   contrib/citext/expected/citext_1.out:
   SELECT 'a'::citext >= 'B'::text AS t;  -- text wins.
    t
   ---
    f
   (1 row)
   
   
   contrib/citext/expected/citext.out:
   SELECT 'a'::citext >= 'B'::text AS t;  -- text wins.
    t
   ---
    t
   (1 row)
   
   same input, return different result,  The comment "-- text wins."
   seems confusing.
   (this happens in other places within citext.out, citext_1.out).
   
   
   
   

2. Re: citext_1.out, citext.out confusing comment

   Kirill Reshke <reshkekirill@gmail.com> — 2025-12-09T07:30:51Z

   On Tue, 9 Dec 2025, 10:08 jian he, <jian.universality@gmail.com> wrote:
   
   > hi.
   >
   > While working on CAST DEFAULT, I found out citext_1.out, citext.out
   > comments are
   > confusing.
   >
   > contrib/citext/expected/citext_1.out:
   > SELECT 'a'::citext >= 'B'::text AS t;  -- text wins.
   >  t
   > ---
   >  f
   > (1 row)
   >
   >
   > contrib/citext/expected/citext.out:
   > SELECT 'a'::citext >= 'B'::text AS t;  -- text wins.
   >  t
   > ---
   >  t
   > (1 row)
   >
   > same input, return different result,  The comment "-- text wins."
   > seems confusing.
   > (this happens in other places within citext.out, citext_1.out).
   >
   
   
   Well, that's because there is single sql test file, so no variation here is
   possible. We can simply delete comment, or we can rephrase it to reflect,
   under which conditions text wins
   
   >
   

3. Re: citext_1.out, citext.out confusing comment

   Julien Rouhaud <rjuju123@gmail.com> — 2025-12-09T09:55:16Z

   On Tue, Dec 09, 2025 at 12:30:51PM +0500, Kirill Reshke wrote:
   > On Tue, 9 Dec 2025, 10:08 jian he, <jian.universality@gmail.com> wrote:
   >
   > > hi.
   > >
   > > While working on CAST DEFAULT, I found out citext_1.out, citext.out
   > > comments are
   > > confusing.
   > >
   > > contrib/citext/expected/citext_1.out:
   > > SELECT 'a'::citext >= 'B'::text AS t;  -- text wins.
   > >  t
   > > ---
   > >  f
   > > (1 row)
   > >
   > >
   > > contrib/citext/expected/citext.out:
   > > SELECT 'a'::citext >= 'B'::text AS t;  -- text wins.
   > >  t
   > > ---
   > >  t
   > > (1 row)
   > >
   > > same input, return different result,  The comment "-- text wins."
   > > seems confusing.
   > > (this happens in other places within citext.out, citext_1.out).
   > >
   >
   > Well, that's because there is single sql test file, so no variation here is
   > possible. We can simply delete comment, or we can rephrase it to reflect,
   > under which conditions text wins
   
   text always wins, which is why the alternative exists: the result depends on
   the default collation sort order under which that test is run.