Re: doc: add missing "id" attributes to extension packaging page

Brar Piening <brar@gmx.de>

From: Brar Piening <brar@gmx.de>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>, "Karl O. Pinc" <kop@karlpinc.com>
Cc: Dagfinn Ilmari Mannsåker <ilmari@ilmari.org>, Peter Eisentraut <peter.eisentraut@enterprisedb.com>, Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2025-11-15T10:53:34Z
Lists: pgsql-hackers
On 2025-11-14 at 11:59 Alvaro Herrera wrote:
> On 2023-Apr-13, Karl O. Pinc wrote:
>
>> Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> wrote:
>>> Another side note: I notice the links don't appear on <refsectN>
>>> elements (e.g.
>>> https://www.postgresql.org/docs/devel/sql-select.html#SQL-WITH),
>>> only <sectN>.
>> This we know.  Working with <refsectN> elements is a different
>> dive into the XSLT which was deliberately put off for future
>> work.
> Hi, has anybody been able to spend some time on the <refsectN> aspect of
> these link things?

Hi,

tbh, this has been in my backlog since I did the initial thing.

I haven't prioritized it since it is really a completely different area 
XSLT-wise that cannot build on the changes I have made and I wanted to 
see how well my initial work will prove its value and be accepted by the 
maintainers and the community before starting to work on this and 
introducing even more changes in this kind of ugly XSLT area that nobody 
really likes to touch.

I think, now that some  time has passed, everybody working on patches in 
the docs area should have some opinion on how annoying it is to be 
forced to provide ids for various elements and if the advantage of being 
able to refer to specific items in the HTML docs is really worth it.

Also, as we've learnt the last time, a patch that enforces the presence 
of ids in refsections would need to be applied in a quick and 
cooperative way once we agree on the general structure, since it needs 
to be rebased and enhanced whenever a new refsection without ids (or 
other element where it enforces ids) is added by some patch and will 
break all subsequent patches that try to add refsections without ids 
(forcing their developers/reviewers to rebase and add ids accordingly).

If we can build some consensus that we want this for the refsections as 
well, I'm willing to put some time aside and work on this (probably at 
some point next year since I'm currently drowning in work).

Regards,

Barr




Commits

  1. doc: Make HTML ids discoverable

  2. Add missing XML ID attribute

  3. postgres_fdw: Add support for parallel abort.

  4. Add missing XML ID attributes

  5. Add XML ID attributes to create_subscription.sgml.

  6. Stop recommending auto-download of DTD files, and indeed disable it.

  7. Doc: add XML ID attributes to <sectN> and <varlistentry> tags.

  8. Allow left join removals and unique joins on partitioned tables

  9. Improve GIN cost estimation