From: Junio C Hamano <gitster@pobox.com>
To: Ralf Thielow <ralf.thielow@gmail.com>
Cc: git@vger.kernel.org, larsxschneider@gmail.com, me@jnm2.com,
philipoakley@iee.org, john@keeping.me.uk
Subject: Re: [PATCH 2/2] help: make option --help open man pages only for Git commands
Date: Thu, 18 Aug 2016 12:51:35 -0700 [thread overview]
Message-ID: <xmqqy43t6ek8.fsf@gitster.mtv.corp.google.com> (raw)
In-Reply-To: <20160818185719.4909-3-ralf.thielow@gmail.com> (Ralf Thielow's message of "Thu, 18 Aug 2016 20:57:19 +0200")
Ralf Thielow <ralf.thielow@gmail.com> writes:
> If option --help is passed to a Git command, we try to open
> the man page of that command. However, we do it even for commands
> we don't know. Make sure it is a Git command.
What the patch does is correct, I think, but the explanation may
invite a false alarm. If you added a custom command git-who in your
$PATH, with an appropriate documentation for git-who(1), we would
still show its documentation, no?
The same comment applies to 1/2, too, in that the word "command"
will be interpreted differently by different people. For example,
"git co --help" and "git help co" would work, with or without 1/2 in
place when you have "[alias] co = checkout", so we are calling "Git
subcommands that we ship, custom commands 'git-$foo' the users have
in their $PATH, and aliases the users create" collectively "command".
As long as the reader understands that definition, both the log
messages of 1/2 and 2/2 _and_ the updated description for "git help"
we have in 1/2 are all very clear. I do not care too much about the
commit log message, but we may want to think about the documentation
a bit more.
Here is what 1/2 adds to "git help" documentation:
+Note that `git --help ...` is almost identical to `git help ...` because
+the former is internally converted into the latter with option --command-only
+being added.
To display the linkgit:git[1] man page, use `git help git`.
@@ -43,6 +44,10 @@ OPTIONS
Prints all the available commands on the standard output. This
option overrides any given command or guide name.
+-c::
+--command-only::
+ Display help information only for commands.
+
First, I do not think a short form is unnecessary; the users are not
expected to use that form, once they started typing "git help...".
If we flip the polarity and call it --exclude-guides or something,
would it make it less ambiguous?
> This breaks "git <concept> --help" while "git help <concept>" still works.
I wouldn't call that a breakage; "git everyday --help" shouldn't
have worked in the first place. It did something useful merely by
accident ;-).
> diff --git a/git.c b/git.c
> index 0f1937f..2cd2e06 100644
> --- a/git.c
> +++ b/git.c
> @@ -528,10 +528,23 @@ static void handle_builtin(int argc, const char **argv)
> strip_extension(argv);
> cmd = argv[0];
>
> - /* Turn "git cmd --help" into "git help cmd" */
> + /* Turn "git cmd --help" into "git help --command-only cmd" */
> if (argc > 1 && !strcmp(argv[1], "--help")) {
> + struct argv_array args;
> + int i;
> +
> argv[1] = argv[0];
> argv[0] = cmd = "help";
> +
> + argv_array_init(&args);
> + for (i = 0; i < argc; i++) {
> + argv_array_push(&args, argv[i]);
> + if (!i)
> + argv_array_push(&args, "--command-only");
> + }
> +
> + argc++;
> + argv = argv_array_detach(&args);
> }
>
> builtin = get_builtin(cmd);
The code does this after it:
if (builtin)
exit(run_builtin(...));
and returns. If we didn't get builtin, we risk leaking args.argv
here, but we assume argv[0] = cmd = "help" is always a builtin,
which I think is a safe assumption, so the code is OK. Static
checkers that are only half intelligent may yell at you for not
releasing the resources, though. I wonder if it is worth doing
static void handle_builtin(int argc, const char **argv)
{
struct argv_array args = ARGV_ARRAY_INIT;
...
if (argc > 1 && !strcmp(argv[1], "--help")) {
...
argv = args.argv;
}
builtin = get_builtin(cmd);
if (builtin)
exit(run_builtin(...));
argv_array_clear(&args);
}
to help unconfuse them.
By the way, I do not see these patches on gmane, public-inbox or
usual suspects. Perhaps vger is having a bad day or something?
next prev parent reply other threads:[~2016-08-19 0:50 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-12 2:00 `git stash --help` tries to pull up nonexistent file gitstack.html Joseph Musser
2016-08-12 15:48 ` Junio C Hamano
2016-08-12 16:03 ` Lars Schneider
2016-08-12 16:15 ` Joseph Musser
2016-08-12 16:25 ` Junio C Hamano
2016-08-12 18:14 ` Jacob Keller
2016-08-12 20:10 ` [PATCH] help: make option --help open man pages only for Git commands Ralf Thielow
2016-08-12 21:34 ` Junio C Hamano
2016-08-12 22:53 ` Junio C Hamano
2016-08-13 0:08 ` Philip Oakley
2016-08-13 15:31 ` Junio C Hamano
2016-08-15 5:36 ` [PATCH v2] " Ralf Thielow
2016-08-15 11:25 ` Philip Oakley
2016-08-15 17:57 ` Junio C Hamano
2016-08-15 20:40 ` Philip Oakley
2016-08-15 22:19 ` Junio C Hamano
2016-08-16 10:06 ` John Keeping
2016-08-16 16:20 ` [PATCH v3] " Ralf Thielow
2016-08-16 16:33 ` John Keeping
2016-08-16 16:39 ` Ralf Thielow
2016-08-16 17:27 ` Junio C Hamano
2016-08-16 17:57 ` Ralf Thielow
2016-08-16 19:06 ` Junio C Hamano
2016-08-18 18:57 ` [PATCH 0/2] " Ralf Thielow
2016-08-18 18:57 ` [PATCH 1/2] help: introduce option --command-only Ralf Thielow
2016-08-18 18:57 ` [PATCH 2/2] help: make option --help open man pages only for Git commands Ralf Thielow
2016-08-18 19:51 ` Junio C Hamano [this message]
2016-08-23 17:34 ` Ralf Thielow
2016-08-18 21:47 ` [PATCH 1/2] help: introduce option --command-only Philip Oakley
2016-08-19 8:32 ` Johannes Schindelin
2016-08-19 15:53 ` Junio C Hamano
2016-08-23 17:41 ` Ralf Thielow
2016-08-24 7:47 ` Johannes Schindelin
2016-08-19 8:39 ` Remi Galan Alfonso
2016-08-23 17:37 ` Ralf Thielow
2016-08-26 17:58 ` [PATCH v2 0/3] help: make option --help open man pages only for Git commands Ralf Thielow
2016-08-26 17:58 ` [PATCH v2 1/3] Revert "display HTML in default browser using Windows' shell API" Ralf Thielow
2016-08-26 17:58 ` [PATCH v2 2/3] help: introduce option --exclude-guides Ralf Thielow
2016-08-26 19:06 ` Junio C Hamano
2016-08-26 19:42 ` Junio C Hamano
2016-08-26 20:03 ` Ralf Thielow
2016-08-26 20:28 ` Junio C Hamano
2016-08-26 20:00 ` Ralf Thielow
2016-08-26 20:20 ` Junio C Hamano
2016-08-26 20:39 ` Ralf Thielow
2016-08-26 17:58 ` [PATCH v2 3/3] help: make option --help open man pages only for Git commands Ralf Thielow
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=xmqqy43t6ek8.fsf@gitster.mtv.corp.google.com \
--to=gitster@pobox.com \
--cc=git@vger.kernel.org \
--cc=john@keeping.me.uk \
--cc=larsxschneider@gmail.com \
--cc=me@jnm2.com \
--cc=philipoakley@iee.org \
--cc=ralf.thielow@gmail.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).