From: "Wolfgang Faust" <contrib-git@wolfgangfaust•com>
To: "Johannes Sixt" <j6t@kdbg•org>, "Junio C Hamano" <gitster@pobox•com>
Cc: "Birger Skogeng Pedersen" <birger.sp@gmail•com>,
"Pratyush Yadav" <me@yadavpratyush•com>,
"Marc Branchaud" <marcnarc@xiplink•com>,
git@vger•kernel.org
Subject: Re: [PATCH 0/4] run auto maintenance in git-gui
Date: Tue, 10 Mar 2026 21:48:36 -0700 [thread overview]
Message-ID: <876fd32d-3965-4587-b567-399787741247@app.fastmail.com> (raw)
In-Reply-To: <40ccd060-e6f7-4130-a25e-3c2f65df8eb7@kdbg.org>
On Sat, Mar 7, 2026, at 2:37 PM, Johannes Sixt wrote:
> Am 07.03.26 um 23:01 schrieb Junio C Hamano:
>> Johannes Sixt <j6t@kdbg•org> writes:
>>> So, users that are hurt by this new change are those where all these
>>> conditions are true:
>>> ...
>>> How many could this be? Not many, I guess. The conservative safe
>>> approach would be to treat gui.gcwarning=false as an indication that
>>> automatic cleanup is not desired.
>
> The option does not only control whether or not a warning appears, but
> also whether garbage collection happens or not. When it is set to false,
> then in addition to squelching the warning, garbage collection does
> *not* happen. The option is on by default, so if we find it off, the
> user must have set it explicitly, a clear sign (IMO) that Git GUI should
> not do the garbage collection.
I see your point about not garbage-collecting for people who haven't
been expecting it, but on the other hand, checking `gui.gcwarning`
would not help the people who were achieving the same result by
clicking "No" every time.
However, it seems to me that the conditions you outlined are very
unlikely. In particular:
> - They configure maintenance.* to do more cleanups than the default
> confituration (which is the same as `git gc --auto`, I think).
>
> - They never use `git maintenance run` through some other facility.
Are these not more or less mutually exclusive? Why would someone set
up maintenance to do things, and then never run maintenance?
Given the above, I see two options:
1. Assume that in practice everybody wants auto maintenance and
`gui.gcwarning` is set because they were annoyed by the dialog.
2. Assume that some people do *not* want auto maintenance, and if the
`gui.gcmaintenance` option is unset then show some kind of dialog
that tries to explain the situation to the user and encourages them
to set `gui.gcmaintenance=auto`.
#2 is the safer option, but is rather complicated and involves
bothering every user for the sake of strict compatibility. I assume
that git core itself had to make a similar tradeoff when deciding to
enable automatic garbage collection, but I don't know where I'd find
that discussion and perhaps it's far enough in the past that the
rationales are no longer relevant anyway.
next prev parent reply other threads:[~2026-03-11 4:49 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-26 17:31 git-gui: disable the "loose objects popup" dialog? Birger Skogeng Pedersen
2019-09-26 18:54 ` Johannes Sixt
2019-09-26 19:13 ` Birger Skogeng Pedersen
2019-09-26 19:15 ` Pratyush Yadav
2019-09-26 21:12 ` Birger Skogeng Pedersen
2019-09-26 21:13 ` Johannes Sixt
2019-10-01 18:00 ` Pratyush Yadav
2019-10-02 7:12 ` Birger Skogeng Pedersen
2019-10-02 18:48 ` Johannes Sixt
2026-03-06 5:15 ` [PATCH 0/4] run auto maintenance in git-gui Wolfgang Faust
2026-03-06 5:26 ` [PATCH 1/4] git-gui: run auto maintenance on commit Wolfgang Faust
2026-03-06 5:30 ` [PATCH 2/4] git-gui: remove hint_gc dialog Wolfgang Faust
2026-03-06 5:32 ` [PATCH 3/4] git-gui: remove "Compress Database" feature Wolfgang Faust
2026-03-06 5:32 ` [PATCH 4/4] scalar: remove obsolete gui.GCWarning setting Wolfgang Faust
2026-03-07 11:32 ` [PATCH 0/4] run auto maintenance in git-gui Johannes Sixt
2026-03-07 22:01 ` Junio C Hamano
2026-03-07 22:37 ` Johannes Sixt
2026-03-08 6:02 ` Junio C Hamano
2026-03-11 4:48 ` Wolfgang Faust [this message]
2026-03-13 12:38 ` Johannes Sixt
2019-10-02 20:41 ` git-gui: disable the "loose objects popup" dialog? Marc Branchaud
2019-09-26 21:14 ` Marc Branchaud
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=876fd32d-3965-4587-b567-399787741247@app.fastmail.com \
--to=contrib-git@wolfgangfaust$(echo .)com \
--cc=birger.sp@gmail$(echo .)com \
--cc=git@vger$(echo .)kernel.org \
--cc=gitster@pobox$(echo .)com \
--cc=j6t@kdbg$(echo .)org \
--cc=marcnarc@xiplink$(echo .)com \
--cc=me@yadavpratyush$(echo .)com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox