From: Glen Choo <chooglen@google.com>
To: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
Cc: Junio C Hamano <gitster@pobox.com>,
Glen Choo via GitGitGadget <gitgitgadget@gmail.com>,
git@vger.kernel.org, Jonathan Tan <jonathantanmy@google.com>,
Emily Shaffer <nasamuffin@google.com>, Jeff King <peff@peff.net>,
Derrick Stolee <derrickstolee@github.com>
Subject: Re: [PATCH 0/6] [RFC] config.c: use struct for config reading state
Date: Wed, 08 Mar 2023 15:18:25 -0800 [thread overview]
Message-ID: <kl6lmt4mhlxa.fsf@chooglen-macbookpro.roam.corp.google.com> (raw)
In-Reply-To: <230308.867cvrziac.gmgdl@evledraar.gmail.com>
Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:
>> Yes, exactly. Having a config_set on the repository makes sense, but I
>> don't see a good reason to have the reader on the repository.
>
> [...]
>
> I hadn't looked closely at this aspect of it, and just took it as a
> given that we needed to persist this data outside of the configset
> lifetime.
There really isn't a need to persist the "config_reader" state, we'd
only need it while reading the file, and as you've hinted, once we've
cached the info in the configset, we'd just use that instead.
> If that's not the case then we don't need it in the file scope, nor a
> "struct repository" or whatever, and could just have it materialized by
> git_config_check_init(), no? I.e. when we create the configset we'd
> create it, and throw it away after the configset is created?
Yes, except that I'm proposing that we should do it even lower in the
chain, like do_config_from().
> The catch (aka the reason I stopped halfway through) is that I
> couldn't find a way to expose "struct config_reader" state
> without some fairly big changes, complexity-wise or LoC-wise[...]
>
> I didn't look into exactly why config_fn_t would need your new "reader",
Ah, this is what I was referencing in the CL here:
I believe that only config callbacks are accessing this [config
reader] state, e.g. because they use the low-level information (like
builtin/config.c printing the filename and scope of the value) or for
error reporting (like git_parse_int() reporting the filename and line
number of the value it failed to parse)
> but if you accept that we could stick such a thing into "the_repository"
> then there's no catch or need for the churn to change all those
> callbacks.
>
> Of course something that wants to use the API as a "real" library would
> need to use some alternate mechanism, as you reference in adding a new
> "config_state_fn_t". You note:
>
> but I couldn't find a great way to do this kind of 'switching
> between config callback types' elegantly.
>
> So, I don't know, but I was suggesting looking into doing that based on
> "the_repository" in play...
And yes, since the primary purpose is to make git_config_from_file()
usable by a library caller (and secondarily, prepare for a future where
reading config is thread-safer because of less global state, as Jonathan
discussed [1]), I'd prefer not to use the_repository.
1. https://lore.kernel.org/git/20230306195756.3399115-1-jonathantanmy@google.com
next prev parent reply other threads:[~2023-03-08 23:18 UTC|newest]
Thread overview: 72+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-01 0:38 [PATCH 0/6] [RFC] config.c: use struct for config reading state Glen Choo via GitGitGadget
2023-03-01 0:38 ` [PATCH 1/6] config.c: plumb config_source through static fns Glen Choo via GitGitGadget
2023-03-03 18:02 ` Junio C Hamano
2023-03-01 0:38 ` [PATCH 2/6] config.c: don't assign to "cf" directly Glen Choo via GitGitGadget
2023-03-01 0:38 ` [PATCH 3/6] config.c: create config_reader and the_reader Glen Choo via GitGitGadget
2023-03-03 18:05 ` Junio C Hamano
2023-03-01 0:38 ` [PATCH 4/6] config.c: plumb the_reader through callbacks Glen Choo via GitGitGadget
2023-03-08 9:54 ` Ævar Arnfjörð Bjarmason
2023-03-08 18:00 ` Glen Choo
2023-03-08 18:07 ` Junio C Hamano
2023-03-01 0:38 ` [PATCH 5/6] config.c: remove current_config_kvi Glen Choo via GitGitGadget
2023-03-06 20:12 ` Calvin Wan
2023-03-01 0:38 ` [PATCH 6/6] config.c: remove current_parsing_scope Glen Choo via GitGitGadget
2023-03-06 19:57 ` [PATCH 0/6] [RFC] config.c: use struct for config reading state Jonathan Tan
2023-03-06 21:45 ` Glen Choo
2023-03-06 22:38 ` Jonathan Tan
2023-03-08 10:32 ` Ævar Arnfjörð Bjarmason
2023-03-08 23:09 ` Glen Choo
2023-03-07 11:57 ` Ævar Arnfjörð Bjarmason
2023-03-07 18:22 ` Glen Choo
2023-03-07 18:36 ` Ævar Arnfjörð Bjarmason
2023-03-07 19:36 ` Junio C Hamano
2023-03-07 22:53 ` Glen Choo
2023-03-08 9:17 ` Ævar Arnfjörð Bjarmason
2023-03-08 23:18 ` Glen Choo [this message]
2023-03-16 0:11 ` [PATCH v2 0/8] " Glen Choo via GitGitGadget
2023-03-16 0:11 ` [PATCH v2 1/8] config.c: plumb config_source through static fns Glen Choo via GitGitGadget
2023-03-16 21:16 ` Jonathan Tan
2023-03-16 0:11 ` [PATCH v2 2/8] config.c: don't assign to "cf_global" directly Glen Choo via GitGitGadget
2023-03-16 21:18 ` Jonathan Tan
2023-03-16 21:31 ` Junio C Hamano
2023-03-16 22:56 ` Glen Choo
2023-03-16 0:11 ` [PATCH v2 3/8] config.c: create config_reader and the_reader Glen Choo via GitGitGadget
2023-03-16 21:22 ` Jonathan Tan
2023-03-16 0:11 ` [PATCH v2 4/8] config.c: plumb the_reader through callbacks Glen Choo via GitGitGadget
2023-03-16 0:11 ` [PATCH v2 5/8] config.c: remove current_config_kvi Glen Choo via GitGitGadget
2023-03-16 0:11 ` [PATCH v2 6/8] config.c: remove current_parsing_scope Glen Choo via GitGitGadget
2023-03-16 0:11 ` [PATCH v2 7/8] config: report cached filenames in die_bad_number() Glen Choo via GitGitGadget
2023-03-16 22:22 ` Jonathan Tan
2023-03-16 23:05 ` Glen Choo
2023-03-16 0:11 ` [PATCH v2 8/8] config.c: rename "struct config_source cf" Glen Choo via GitGitGadget
2023-03-16 0:15 ` [PATCH v2 0/8] config.c: use struct for config reading state Glen Choo
2023-03-16 22:29 ` Jonathan Tan
2023-03-17 5:01 ` [RFC PATCH 0/5] bypass config.c global state with configset Ævar Arnfjörð Bjarmason
2023-03-17 5:01 ` [RFC PATCH 1/5] config.h: move up "struct key_value_info" Ævar Arnfjörð Bjarmason
2023-03-17 5:01 ` [RFC PATCH 2/5] config.c: use "enum config_origin_type", not "int" Ævar Arnfjörð Bjarmason
2023-03-17 5:01 ` [RFC PATCH 3/5] config API: add a config_origin_type_name() helper Ævar Arnfjörð Bjarmason
2023-03-17 5:01 ` [RFC PATCH 4/5] config.c: refactor configset_iter() Ævar Arnfjörð Bjarmason
2023-03-17 5:01 ` [RFC PATCH 5/5] config API: add and use a repo_config_kvi() Ævar Arnfjörð Bjarmason
2023-03-17 17:17 ` Junio C Hamano
2023-03-17 20:59 ` Jonathan Tan
2023-03-17 16:21 ` [RFC PATCH 0/5] bypass config.c global state with configset Junio C Hamano
2023-03-17 16:28 ` Glen Choo
2023-03-17 19:20 ` Glen Choo
2023-03-17 23:32 ` Glen Choo
2023-03-29 11:53 ` Ævar Arnfjörð Bjarmason
2023-03-28 17:51 ` [PATCH v3 0/8] config.c: use struct for config reading state Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 1/8] config.c: plumb config_source through static fns Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 2/8] config.c: don't assign to "cf_global" directly Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 3/8] config.c: create config_reader and the_reader Glen Choo via GitGitGadget
2023-03-29 10:41 ` Ævar Arnfjörð Bjarmason
2023-03-29 18:57 ` Junio C Hamano
2023-03-29 20:02 ` Glen Choo
2023-03-30 17:51 ` Glen Choo
2023-03-28 17:51 ` [PATCH v3 4/8] config.c: plumb the_reader through callbacks Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 5/8] config.c: remove current_config_kvi Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 6/8] config.c: remove current_parsing_scope Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 7/8] config: report cached filenames in die_bad_number() Glen Choo via GitGitGadget
2023-03-28 17:51 ` [PATCH v3 8/8] config.c: rename "struct config_source cf" Glen Choo via GitGitGadget
2023-03-28 18:00 ` [PATCH v3 0/8] config.c: use struct for config reading state Glen Choo
2023-03-28 20:14 ` Junio C Hamano
2023-03-28 21:24 ` Glen Choo
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=kl6lmt4mhlxa.fsf@chooglen-macbookpro.roam.corp.google.com \
--to=chooglen@google.com \
--cc=avarab@gmail.com \
--cc=derrickstolee@github.com \
--cc=git@vger.kernel.org \
--cc=gitgitgadget@gmail.com \
--cc=gitster@pobox.com \
--cc=jonathantanmy@google.com \
--cc=nasamuffin@google.com \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).