Horton

Re: get rid of Pointer type, mostly

Robert Haas <robertmhaas@gmail.com>

View thread
From: Robert Haas <robertmhaas@gmail.com>
To: Jelte Fennema-Nio <postgres@jeltef.nl>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Peter Eisentraut <peter@eisentraut.org>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2025-11-24T17:35:44Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Remove no longer needed casts from Pointer

  2. Remove no longer needed casts to Pointer

  3. Change Pointer to void *

  4. Don't rely on pointer arithmetic with Pointer type

  5. Use more appropriate DatumGet* function

  6. Remove useless casts to Pointer

On Mon, Nov 24, 2025 at 12:30 PM Jelte Fennema-Nio <postgres@jeltef.nl> wrote:
> In this case, what we would accomplish is that no new developer to the project has to understand what some unclear typedef means, *unless* they touch GIN related code. Just from its name it's definitely not clear to me that Pointer means char * instead of void *. And this typedef is ven shorter than the thing it represents.

+1.

> Side annoyance: I think this is a falacy that hackers discussions end up in a lot. Someone suggesting that the partial improvements have (almost) no benefit and all cases need to be fixed in one go to before it should be committed. Then the patch author thinks that's too much work and then nothing ends up being improved at all.

This is definitely a thing that happens, but what also happens pretty
often is that people claim that we'll follow up on a partial
improvement with lots more work and then we never do, and then it
creates a big mess for somebody else to untangle later. I understand
the frustration with getting a partial solution blocked, because half
a loaf is better than none, but I've also done my share of cleaning up
changes that weren't so much half a loaf as half-baked.

-- 
Robert Haas
EDB: http://www.enterprisedb.com