Horton

Thread

Commits

  1. Add tab completion for data types after ALTER TABLE ADD [COLUMN] in psql

  2. Add more tab completion support for ALTER TABLE ADD in psql

  1. [PATCH] Tab completion for ALTER TABLE … ADD …

    Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> — 2021-08-03T11:48:38Z

    Hi Hackers,

The other day I noticed that there's no tab completion after ALTER TABLE
… ADD, so here's a patch.  In addition to COLUMN and all the table
constraint types, it also completes with the list of unique indexes on
the table after ALTER TABLE … ADD … USING INDEX.

- ilmari

  2. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> — 2021-08-07T22:25:44Z

    ilmari@ilmari.org (Dagfinn Ilmari Mannsåker) writes:

> Hi Hackers,
>
> The other day I noticed that there's no tab completion after ALTER TABLE
> … ADD, so here's a patch.  In addition to COLUMN and all the table
> constraint types, it also completes with the list of unique indexes on
> the table after ALTER TABLE … ADD … USING INDEX.

Added to the 2021-09 commitfest: https://commitfest.postgresql.org/34/3280/

- ilmari

  3. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Michael Paquier <michael@paquier.xyz> — 2021-08-27T04:04:00Z

    On Tue, Aug 03, 2021 at 12:48:38PM +0100, Dagfinn Ilmari Mannsåker wrote:
> The other day I noticed that there's no tab completion after ALTER TABLE
> … ADD, so here's a patch.  In addition to COLUMN and all the table
> constraint types, it also completes with the list of unique indexes on
> the table after ALTER TABLE … ADD … USING INDEX.

I was reading this patch (not actually tested), and that's a clear
improvement.  One extra thing that could be done here is to complete
with types for a ALTER TABLE ADD COLUMN foo.  We could as well have a
list of columns after UNIQUE or PRIMARY KEY, but that feels like extra
cream on top of the cake.  In short I am fine with what you have
here.
--
Michael

  4. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> — 2021-08-27T10:52:33Z

    Michael Paquier <michael@paquier.xyz> writes:

> On Tue, Aug 03, 2021 at 12:48:38PM +0100, Dagfinn Ilmari Mannsåker wrote:
>> The other day I noticed that there's no tab completion after ALTER TABLE
>> … ADD, so here's a patch.  In addition to COLUMN and all the table
>> constraint types, it also completes with the list of unique indexes on
>> the table after ALTER TABLE … ADD … USING INDEX.
>
> I was reading this patch (not actually tested), and that's a clear
> improvement.  One extra thing that could be done here is to complete
> with types for a ALTER TABLE ADD COLUMN foo.

That was easy enough to add (just a bit of extra fiddling to handle
COLUMN being optional), done in the attached v2 patch.

> We could as well have a list of columns after UNIQUE or PRIMARY KEY,
> but that feels like extra cream on top of the cake.

Doing a list of arbitrarily many comma-separated names is more
complicated, so that can be the subject for another patch.

>  In short I am fine with what you have here.

Thanks for the review.

- ilmari

  5. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Michael Paquier <michael@paquier.xyz> — 2021-08-30T00:55:53Z

    On Fri, Aug 27, 2021 at 11:52:33AM +0100, Dagfinn Ilmari Mannsåker wrote:
> That was easy enough to add (just a bit of extra fiddling to handle
> COLUMN being optional), done in the attached v2 patch.

This part was a bit misleading, as it would recommend a list of types
when specifying just ADD CONSTRAINT for example, so I have removed
it.  An extra thing that felt a bit overdoing is the addition of KEY
after PRIMARY/FOREIGN.

> Doing a list of arbitrarily many comma-separated names is more
> complicated, so that can be the subject for another patch.

No objections to that.  I have applied what we have now, as that's
already an improvement.
--
Michael

  6. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> — 2021-08-30T13:38:19Z

    Michael Paquier <michael@paquier.xyz> writes:

> On Fri, Aug 27, 2021 at 11:52:33AM +0100, Dagfinn Ilmari Mannsåker wrote:
>> That was easy enough to add (just a bit of extra fiddling to handle
>> COLUMN being optional), done in the attached v2 patch.
>
> This part was a bit misleading, as it would recommend a list of types
> when specifying just ADD CONSTRAINT for example, so I have removed
> it.

That was because I forgot to exclude all the other object types that can
come after ADD.  Attached is a patch that does that.  I also moved it
right next to the ALTER TABLE … ADD completion, and added a comment to
keep the two lists in sync.

> An extra thing that felt a bit overdoing is the addition of KEY after
> PRIMARY/FOREIGN.

Yeah, I guess people are unlikely to write out the whole PRIMARY or FOREIGN
and only then hit tab to complete the rest.

>> Doing a list of arbitrarily many comma-separated names is more
>> complicated, so that can be the subject for another patch.
>
> No objections to that.  I have applied what we have now, as that's
> already an improvement.

Thanks!

- ilmari

  7. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Michael Paquier <michael@paquier.xyz> — 2021-08-31T03:20:14Z

    On Mon, Aug 30, 2021 at 02:38:19PM +0100, Dagfinn Ilmari Mannsåker wrote:
> That was because I forgot to exclude all the other object types that can
> come after ADD.  Attached is a patch that does that.  I also moved it
> right next to the ALTER TABLE … ADD completion, and added a comment to
> keep the two lists in sync.

Looks fine to me, so applied while we are on it.
--
Michael

  8. Re: [PATCH] Tab completion for ALTER TABLE … ADD …

    Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> — 2021-08-31T08:36:48Z

    On Tue, 31 Aug 2021, at 04:20, Michael Paquier wrote:
> On Mon, Aug 30, 2021 at 02:38:19PM +0100, Dagfinn Ilmari Mannsåker wrote:
> > That was because I forgot to exclude all the other object types that can
> > come after ADD.  Attached is a patch that does that.  I also moved it
> > right next to the ALTER TABLE … ADD completion, and added a comment to
> > keep the two lists in sync.
> 
> Looks fine to me, so applied while we are on it.

Thanks!

- ilmari