Horton

Re: Potential ABI breakage in upcoming minor releases

Peter Geoghegan <pg@bowt.ie>

View thread
From: Peter Geoghegan <pg@bowt.ie>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Pavan Deolasee <pavan.deolasee@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, Noah Misch <noah@leadboat.com>, Tom Lane <tgl@sss.pgh.pa.us>
Date: 2024-11-14T16:35:25Z
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. Undo unintentional ABI break in struct ResultRelInfo.

  2. For inplace update durability, make heap_update() callers wait.

  3. Fix btmarkpos/btrestrpos array key wraparound bug.

  4. Fix calculation of which GENERATED columns need to be updated.

On Thu, Nov 14, 2024 at 11:29 AM Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> ISTM that we have spare bytes where we could place that boolean without
> breaking ABI.  In x86_64 there's a couple of 4-byte holes, but those
> won't be there on x86, so not great candidates.  Fortunately there are
> 3-byte and 7-byte holes also, which we can use safely.  We can move the
> new boolean to those location.

Wasn't this part of the official guidelines? I've been doing this all
along (e.g., in commit 3fa81b62e0).

> The holes are different in each branch unfortunately.

Yeah, that'd make it a bit more complicated, but still doable. It
would be necessary to place the same field in seemingly random
locations on each backbranch.

FWIW recent versions of clangd will show me information about field
padding in struct annotations. I don't even have to run pahole.


--
Peter Geoghegan