git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: "Lénaïc Huard" <lenaic@lhuard.fr>
To: Eric Sunshine <sunshine@sunshineco.com>
Cc: "Git List" <git@vger.kernel.org>,
	"Junio C Hamano" <gitster@pobox.com>,
	"Derrick Stolee" <dstolee@microsoft.com>,
	"Đoàn Trần Công Danh" <congdanhqx@gmail.com>,
	"Felipe Contreras" <felipe.contreras@gmail.com>,
	"Phillip Wood" <phillip.wood123@gmail.com>,
	"Martin Ågren" <martin.agren@gmail.com>,
	"Ævar Arnfjörð Bjarmason" <avarab@gmail.com>,
	"Bagas Sanjaya" <bagasdotme@gmail.com>,
	"brian m . carlson" <sandals@crustytoothpaste.net>,
	"Johannes Schindelin" <Johannes.Schindelin@gmx.de>,
	"Jeff King" <peff@peff.net>
Subject: Re: [PATCH v6 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>`
Date: Fri, 02 Jul 2021 17:04:10 +0200	[thread overview]
Message-ID: <2080051.vtJqlmqO2r@coruscant.lhuard.fr> (raw)
In-Reply-To: <CAPig+cSLi7aN=6ahrHwy4fO-7JMBN3pmzfpWe5ZXOcC9j4+e+g@mail.gmail.com>

Le lundi 14 juin 2021, 06:36:09 CEST Eric Sunshine a écrit :

> Thanks. Unfortunately, I haven't been following this series too
> closely since I reviewed v1, so I set aside time to review v6, which I
> have now done.
> …
> I do, though, have one question (below) about is_crontab_available()
> for which I could not figure out the answer.

Thank you very much for your review.
I’ve just submitted a new re-roll that should address the points you raised 
and in particular the `is_crontab_available` unexpected behavior.

…
> > +static int is_launchctl_available(void)
> > +{
> > +       const char *cmd = "launchctl";
> > +       int is_available;
> > +       if (get_schedule_cmd(&cmd, &is_available))
> > +               return is_available;
> > +
> > +#ifdef __APPLE__
> > +       return 1;
> > +#else
> > +       return 0;
> > +#endif
> > +}
> 
> On this project, we usually frown upon #if conditionals within
> functions since the code often can become unreadable. The usage in
> this function doesn't suffer from that problem, however,
> resolve_auto_scheduler() is somewhat ugly. An alternative would be to
> set up these values outside of all functions, perhaps like this:
> 
>     #ifdef __APPLE__
>     #define MAINT_SCHEDULER SCHEDULER_LAUNCHCTL
>     #elif GIT_WINDOWS_NATIVE
>     #define MAINT_SCHEDULER SCHEDULER_SCHTASKS
>     #else
>     #define MAINT_SCHEDULER SCHEDULER_CRON
>     #endif
> 
> and then:
> 
>     static int is_launchctl_available(void)
>     {
>         if (get_schedule_cmd(...))
>             return is_available;
>         return MAINT_SCHEDULER == SCHEDULER_LAUNCHCTL;
>     }
> 
>     static void resolve_auto_scheduler(enum scheduler *scheduler)
>     {
>         if (*scheduler == SCHEDULER_AUTO)
>             *scheduler = MAINT_SCHEDULER;
>     }
> 

This approach would unfortunately work only for the second patch of this 
series where a single scheduler is available on each platform.
With the third patch of this series, `resolve_auto_scheduler` doesn’t return a 
value that is fully determined at compilation time anymore.
On Linux, both `crontab` and `systemd-timers` are susceptible to be available 
and this is checked at runtime.
So, with the third patch of this series, it wouldn’t be possible anymore to 
define a single value for `MAINT_SCHEDULER` and to base 
`resolve_auto_scheduler` on it.

…
> > +                       PARSE_OPT_NONEG, maintenance_opt_scheduler),
> > +               OPT_END()
> > +       };
> > diff --git a/t/t7900-maintenance.sh b/t/t7900-maintenance.sh
> > @@ -494,8 +494,21 @@ test_expect_success !MINGW 'register and unregister
> > with regex metacharacters' ' +test_expect_success 'start
> > --scheduler=<scheduler>' '
> > +       test_expect_code 129 git maintenance start --scheduler=foo 2>err
> > &&
> > +       test_i18ngrep "unrecognized --scheduler argument" err &&
> > +
> > +       test_expect_code 129 git maintenance start --no-scheduler 2>err &&
> > +       test_i18ngrep "unknown option" err &&
> > +
> > +       test_expect_code 128 \
> > +               env
> > GIT_TEST_MAINT_SCHEDULER="launchctl:true,schtasks:true" \ +              
> > git maintenance start --scheduler=crontab 2>err && +       test_i18ngrep
> > "fatal: crontab scheduler is not available" err +'
> 
> Why does this test care about the exact exit codes rather than simply
> using test_must_fail() as is typically done elsewhere in the test
> suite, especially since we're also checking the error message itself?
> Am I missing some non-obvious property of the error codes?

I have no strong opinion on this.
I only mimicked the `help text` test that is at the top of the `t7900-
maintenance.sh` file as it was also testing invalid commands and checking the 
resulting error message.

> I don't see `auto` being tested anywhere. Do we want such a test? (It
> seems like it should be doable, though perhaps the complexity is too
> high -- I haven't thought it through fully.)

My main problem with the `auto` is that a big part of its logic is determined 
at compilation time with `#if` statements based on the platform.
And it seems that the tests are designed so far to test the same things on all 
platforms.
A solution could be to completely get rid of the platform `#if` statements and 
to turn all the detection logic as runtime tests.
But it would mean that, for ex., the git binary for Linux would systematically 
check if the MacOS and Windows specific schedulers are available.

Cheers,
Lénaïc.



  parent reply	other threads:[~2021-07-02 15:04 UTC|newest]

Thread overview: 138+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-01 14:52 [PATCH] maintenance: use systemd timers on Linux Lénaïc Huard
2021-05-01 20:02 ` brian m. carlson
2021-05-02  5:28 ` Bagas Sanjaya
2021-05-02  6:49   ` Eric Sunshine
2021-05-02  6:45 ` Eric Sunshine
2021-05-02 14:10   ` Phillip Wood
2021-05-05 12:19     ` Đoàn Trần Công Danh
2021-05-05 14:57       ` Phillip Wood
2021-05-05 12:01   ` Ævar Arnfjörð Bjarmason
2021-05-09 22:34     ` Lénaïc Huard
2021-05-10 13:03       ` Ævar Arnfjörð Bjarmason
2021-05-02 11:12 ` Bagas Sanjaya
2021-05-03 12:04 ` Derrick Stolee
2021-05-09 21:32 ` [PATCH v2 0/1] " Lénaïc Huard
2021-05-09 21:32   ` [PATCH v2 1/1] " Lénaïc Huard
2021-05-10  1:20     ` Đoàn Trần Công Danh
2021-05-10  2:48       ` Eric Sunshine
2021-05-10  6:25         ` Junio C Hamano
2021-05-12  0:29           ` Đoàn Trần Công Danh
2021-05-12  6:59             ` Felipe Contreras
2021-05-12 13:26               ` Phillip Wood
2021-05-12 13:38             ` Phillip Wood
2021-05-12 15:41               ` Đoàn Trần Công Danh
2021-05-10 18:03     ` Phillip Wood
2021-05-10 18:25       ` Eric Sunshine
2021-05-10 20:09         ` Phillip Wood
2021-05-10 20:52           ` Eric Sunshine
2021-06-08 14:55       ` Lénaïc Huard
2021-05-10 19:15     ` Martin Ågren
2021-05-11 14:50     ` Phillip Wood
2021-05-11 17:31     ` Derrick Stolee
2021-05-20 22:13   ` [PATCH v3 0/4] " Lénaïc Huard
2021-05-20 22:13     ` [PATCH v3 1/4] cache.h: rename "xdg_config_home" to "xdg_config_home_git" Lénaïc Huard
2021-05-20 23:44       ` Đoàn Trần Công Danh
2021-05-20 22:13     ` [PATCH v3 2/4] maintenance: introduce ENABLE/DISABLE for code clarity Lénaïc Huard
2021-05-20 22:13     ` [PATCH v3 3/4] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-05-21  9:52       ` Bagas Sanjaya
2021-05-20 22:13     ` [PATCH v3 4/4] maintenance: optionally use systemd timers on Linux Lénaïc Huard
2021-05-21  9:59       ` Bagas Sanjaya
2021-05-21 16:59         ` Derrick Stolee
2021-05-22  6:57           ` Johannes Schindelin
2021-05-23 18:36             ` Felipe Contreras
2021-05-23 23:27               ` brian m. carlson
2021-05-24  1:18                 ` Felipe Contreras
2021-05-24  7:03                 ` Ævar Arnfjörð Bjarmason
2021-05-24 15:51                   ` Junio C Hamano
2021-05-25  1:50                     ` Johannes Schindelin
2021-05-25 11:13                       ` Felipe Contreras
2021-05-26 10:29                       ` CoC, inclusivity etc. (was "Re: [...] systemd timers on Linux") Ævar Arnfjörð Bjarmason
2021-05-26 16:05                         ` Felipe Contreras
2021-05-27 14:24                         ` Jeff King
2021-05-27 17:30                           ` Felipe Contreras
2021-05-27 23:58                           ` Junio C Hamano
2021-05-28 14:44                           ` Phillip Susi
2021-05-30 21:58                             ` Jeff King
2021-05-24 17:52                   ` [PATCH v3 4/4] maintenance: optionally use systemd timers on Linux Felipe Contreras
2021-05-24  7:15     ` [PATCH v4 0/4] add support for " Lénaïc Huard
2021-05-24  7:15       ` [PATCH v4 1/4] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-05-24  9:33         ` Phillip Wood
2021-05-24 12:23           ` Đoàn Trần Công Danh
2021-05-24  7:15       ` [PATCH v4 2/4] maintenance: introduce ENABLE/DISABLE for code clarity Lénaïc Huard
2021-05-24  9:41         ` Phillip Wood
2021-05-24 12:36           ` Đoàn Trần Công Danh
2021-05-25  7:18             ` Lénaïc Huard
2021-05-25  8:02               ` Junio C Hamano
2021-05-24  9:47         ` Ævar Arnfjörð Bjarmason
2021-05-24  7:15       ` [PATCH v4 3/4] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-05-24 10:12         ` Phillip Wood
2021-05-30  6:39           ` Lénaïc Huard
2021-05-30 10:16             ` Phillip Wood
2021-05-24  7:15       ` [PATCH v4 4/4] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-05-24  9:55         ` Ævar Arnfjörð Bjarmason
2021-05-24 16:39           ` Eric Sunshine
2021-05-24 18:08         ` Felipe Contreras
2021-05-26 10:26         ` Phillip Wood
2021-05-24  9:04       ` [PATCH v4 0/4] " Junio C Hamano
2021-06-08 13:39       ` [PATCH v5 0/3] " Lénaïc Huard
2021-06-08 13:39         ` [PATCH v5 1/3] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-06-08 13:39         ` [PATCH v5 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-06-08 13:40         ` [PATCH v5 3/3] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-06-09  9:34           ` Jeff King
2021-06-09 15:01           ` Phillip Wood
2021-06-09  0:21         ` [PATCH v5 0/3] " Junio C Hamano
2021-06-09 14:54         ` Phillip Wood
2021-06-12 16:50         ` [PATCH v6 0/3] maintenance: " Lénaïc Huard
2021-06-12 16:50           ` [PATCH v6 1/3] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-06-12 16:50           ` [PATCH v6 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-06-14  4:36             ` Eric Sunshine
2021-06-16 18:12               ` Derrick Stolee
2021-06-17  4:11                 ` Eric Sunshine
2021-06-17 14:26               ` Phillip Wood
2021-07-02 15:04               ` Lénaïc Huard [this message]
2021-06-12 16:50           ` [PATCH v6 3/3] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-07-02 14:25           ` [PATCH v7 0/3] " Lénaïc Huard
2021-07-02 14:25             ` [PATCH v7 1/3] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-07-02 14:25             ` [PATCH v7 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-07-06 19:56               ` Ævar Arnfjörð Bjarmason
2021-07-06 20:52                 ` Junio C Hamano
2021-07-13  0:15                   ` Jeff King
2021-07-13  2:22                     ` Eric Sunshine
2021-07-13  3:56                       ` Jeff King
2021-07-13  5:17                         ` Eric Sunshine
2021-07-13  7:04                       ` Bagas Sanjaya
2021-07-06 21:18                 ` Felipe Contreras
2021-08-23 20:06                 ` Lénaïc Huard
2021-08-23 22:30                   ` Junio C Hamano
2021-07-02 14:25             ` [PATCH v7 3/3] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-07-06 20:03               ` Ævar Arnfjörð Bjarmason
2021-07-02 18:18             ` [PATCH v7 0/3] " Junio C Hamano
2021-07-06 13:18             ` Phillip Wood
2021-08-23 20:40             ` [PATCH v8 " Lénaïc Huard
2021-08-23 20:40               ` [PATCH v8 1/3] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-08-23 20:40               ` [PATCH v8 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-08-24 17:45                 ` Derrick Stolee
2021-08-23 20:40               ` [PATCH v8 3/3] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-08-24 17:45                 ` Derrick Stolee
2021-08-24 17:47               ` [PATCH v8 0/3] " Derrick Stolee
2021-08-27 21:02               ` [PATCH v9 " Lénaïc Huard
2021-08-27 21:02                 ` [PATCH v9 1/3] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-08-27 21:02                 ` [PATCH v9 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-08-27 23:54                   ` Ramsay Jones
2021-08-27 21:02                 ` [PATCH v9 3/3] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-09-04 20:54                 ` [PATCH v10 0/3] " Lénaïc Huard
2021-09-04 20:54                   ` [PATCH v10 1/3] cache.h: Introduce a generic "xdg_config_home_for(…)" function Lénaïc Huard
2021-09-04 20:54                   ` [PATCH v10 2/3] maintenance: `git maintenance run` learned `--scheduler=<scheduler>` Lénaïc Huard
2021-09-04 20:55                   ` [PATCH v10 3/3] maintenance: add support for systemd timers on Linux Lénaïc Huard
2021-09-07 16:48                   ` [PATCH v10 0/3] " Derrick Stolee
2021-09-08 11:44                     ` Derrick Stolee
2021-09-09  5:52                       ` Lénaïc Huard
2021-09-09 19:55                         ` Derrick Stolee
2021-09-27 12:50                           ` Ævar Arnfjörð Bjarmason
2021-09-27 21:44                             ` Lénaïc Huard
2021-08-17 17:22         ` [PATCH v5 0/3] " Derrick Stolee
2021-08-17 19:43           ` Phillip Wood
2021-08-17 20:29             ` Derrick Stolee
2021-08-18  5:56           ` Lénaïc Huard
2021-08-18 13:28             ` Derrick Stolee
2021-08-18 18:23               ` Junio C Hamano

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=2080051.vtJqlmqO2r@coruscant.lhuard.fr \
    --to=lenaic@lhuard.fr \
    --cc=Johannes.Schindelin@gmx.de \
    --cc=avarab@gmail.com \
    --cc=bagasdotme@gmail.com \
    --cc=congdanhqx@gmail.com \
    --cc=dstolee@microsoft.com \
    --cc=felipe.contreras@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=martin.agren@gmail.com \
    --cc=peff@peff.net \
    --cc=phillip.wood123@gmail.com \
    --cc=sandals@crustytoothpaste.net \
    --cc=sunshine@sunshineco.com \
    /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).