dash.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Harald van Dijk <harald@gigawatt.nl>
To: Martijn Dekker <martijn@inlv.org>, dash@vger.kernel.org
Subject: Re: [PATCH] quote arguments in xtrace output
Date: Mon, 27 Feb 2017 22:34:31 +0100	[thread overview]
Message-ID: <0ef1b102-a04c-f3ac-609e-5285d59904d0@gigawatt.nl> (raw)
In-Reply-To: <6e8434e8-87e3-db2a-9d80-147f064c4793@inlv.org>

On 27/02/2017 21:08, Martijn Dekker wrote:
> Op 27-02-17 om 20:24 schreef Harald van Dijk:
>> If the modification to single_quote can handle all cases (perhaps
>> quoting unnecessarily if it's unclear), then the conditional
>> parameter could be removed and applied unconditionally: as far as I
>> can tell, there are no existing calls that require the single quotes
>> for simple words.
>
> True, but this would change the output of something like:
>
> $ dash -c 'alias stuff=xyz; trap true EXIT; alias; trap'
>
> from:
>
> stuff='xyz'
> trap -- 'true' EXIT
>
> to:
>
> stuff=xyz
> trap -- true EXIT
>
> While there is technically nothing wrong with this, I think it's nicer
> to leave those quoted unconditionally, because that facilitates editing
> the output of those commands to feed it back to the shell.

Fair point.

Looking at other shells:

             | bash | ksh93 | zsh | posh | dash | dash (patched)
    alias    | 2    | 1     | 1   | -    | 2    | 2
     trap    | 2    | 1     | 1   | 1    | 2    | 2
   export -p | 2    | 1     | 1   | 0    | 2    | 2
readonly -p | 2    | 1     | 1   | 0    | 2    | 2
      set    | 1    | 1     | 1   | -    | 2    | 2
      set -x | 1    | 1     | 1   | 0    | 0    | 1

Here, 0 means "never quoted", 1 means "quoted when containing special 
characters", and 2 means "always quoted".

Although there is a bit of variation in other shells, your approach 
brings dash closer to bash, which I believe is considered a goal for 
dash, all else being equal. My suggestion would bring dash closer to 
bash in one command, but further from bash in four others, so unless 
there is a good reason for it, I think I agree that that shouldn't be done.

Cheers,
Harald van Dijk

  reply	other threads:[~2017-02-27 21:34 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-02-27  6:22 [PATCH] quote arguments in xtrace output Martijn Dekker
2017-02-27 19:24 ` Harald van Dijk
2017-02-27 20:08   ` Martijn Dekker
2017-02-27 21:34     ` Harald van Dijk [this message]
2017-02-27 23:17     ` Martijn Dekker
2017-02-28  3:39       ` Martijn Dekker
2017-03-18  4:31         ` Martijn Dekker

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=0ef1b102-a04c-f3ac-609e-5285d59904d0@gigawatt.nl \
    --to=harald@gigawatt.nl \
    --cc=dash@vger.kernel.org \
    --cc=martijn@inlv.org \
    /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).