=?utf-8?Q?=C3=81lvaro?= Herrera <[email protected]> writes: > I think the best time to get it in is a few weeks after the feature > freeze. That way we minimize the impact on any fixes during the > stabilization period. I'm not sure at what point relative to the betas > though; is it better to do it before, so that we have a chance to fix > any possible problems before they hit users, or is it better to do after > beta1, so that patches that we discover need reverting for pg19 have > already had a chance to be so?
It seems unlikely to me that a pgindent change could reach to the point of causing user-visible problems. It could break code I guess, but we'd almost surely see compile failures if so. Also I rather imagine we'd have eyeballed all the code changes before committing. The other consideration is that we don't want to do it too close to when the next CF opens. It will likely break some pending patches, and people will want time to rebase those. Checking our commit history, it seems the last few intentional changes in pgindent's behavior were applied here: Author: Tom Lane <[email protected]> Branch: master Release: REL_18_BR [b27644bad] 2025-06-15 13:04:24 -0400 Sync typedefs.list with the buildfarm. Author: Tom Lane <[email protected]> Branch: master Release: REL_16_BR [0245f8db3] 2023-05-19 17:24:48 -0400 Pre-beta mechanical code beautification. Author: Tom Lane <[email protected]> Branch: master Release: REL_16_BR [064750af4] 2023-04-08 11:48:45 -0400 Improve indentation of multiline initialization expressions. Author: Tom Lane <[email protected]> Branch: master Release: REL_13_BR [fa27dd40d] 2020-05-16 11:54:51 -0400 Run pgindent with new pg_bsd_indent version 2.1.1. Author: Tom Lane <[email protected]> Branch: master Release: REL_12_BR [8255c7a5e] 2019-05-22 13:04:48 -0400 Phase 2 pgindent run for v12. So we haven't been totally consistent about it, but mid-May (a couple weeks before making the new branch) seems like the usual choice. regards, tom lane PS: this isn't an endorsement of the proposed patch; I've not looked at it.
