git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Andrew Sayers <andrew-git@pileofstuff.org>
Cc: Git Mailing List <git@vger.kernel.org>,
	solsTiCe d'Hiver <solstice.dhiver@gmail.com>
Subject: Re: [PATCH] bash-completion: Print a useful error when called in a non-bash shell
Date: Thu, 29 Jul 2010 09:25:32 -0700	[thread overview]
Message-ID: <7v4ofiuuf7.fsf@alter.siamese.dyndns.org> (raw)
In-Reply-To: <4C50B005.1030004@pileofstuff.org> (Andrew Sayers's message of "Wed\, 28 Jul 2010 23\:32\:37 +0100")

Andrew Sayers <andrew-git@pileofstuff.org> writes:

> Detect whether the shell supports process substitution with <()
> Shells that fail the test will not be able to load git-completion.bash
>
> If a bad shell is found, print a warning which gives the user as much debugging
> information as possible.
>
> This was added in response to a bug report on the git mailing list:
>   http://permalink.gmane.org/gmane.comp.version-control.git/151723
>
> Signed-off-by: Andrew Sayers <andrew-git@pileofstuff.org>
> ---

> +# Check in case we were called with something like "sh git-completion.bash":
> +$(
> +	exec 2>/dev/null
> +	$(exec < <( ))
> + )
> +if [[ 0 -ne $? ]]
> +then
> +	cat <<EOF
> +
> +ERROR: you don't seem to be running a full bash shell.
> +git-completion.bash is probably about to fail with a syntax error.
> +If you are sure that your system is calling git-completion.bash from a bash,
> +then please include the following in a bug report to git@vger.kernel.org:

I needed to read this twice to realize that majority of people who will
ever see this message on their screen are _not_ expected to send any bug
report to us.  Also if the user's "full bash" groks <() indirection but
still fails to run completion script correctly, this message does not help
them to find where to file a bug report at all.

I do agree that there need to be a way to find that information for the
end users, especially for those who just use binary-packaged git given by
their distros, but this codepath is _not_ the place to do it.

How about replacing these with something simple like:

    echo >&2 "You are not running full 'bash'; exiting." ; exit 127

> +	BASH_VERSION: {$BASH_VERSION}
> +	BASHOPTS: {$BASHOPTS}
> +	SHELLOPTS: {$SHELLOPTS}
> +	POSIXLY_CORRECT: {$POSIXLY_CORRECT}
> +EOF
> +	echo -n "	command line: {"
> +	tr '\0' ' ' < /proc/$$/cmdline

This looks like a Linux-ism to me.

  reply	other threads:[~2010-07-29 16:26 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-07-28 22:32 [PATCH] bash-completion: Print a useful error when called in a non-bash shell Andrew Sayers
2010-07-29 16:25 ` Junio C Hamano [this message]
2010-07-29 19:05   ` solsTiCe d'Hiver
2010-07-29 21:29   ` Andrew Sayers
2010-07-29 22:16     ` Ævar Arnfjörð Bjarmason
2010-07-29 23:38     ` Junio C Hamano
2010-07-30  6:02       ` Stephen Boyd
2010-08-06 21:31 ` Johannes Sixt

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=7v4ofiuuf7.fsf@alter.siamese.dyndns.org \
    --to=gitster@pobox.com \
    --cc=andrew-git@pileofstuff.org \
    --cc=git@vger.kernel.org \
    --cc=solstice.dhiver@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.
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 NNTP newsgroup(s).