git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Pratyush Yadav <me@yadavpratyush.com>
To: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Cc: git@vger.kernel.org
Subject: Re: [PATCH 0/1] git-gui: remove unused global declarations
Date: Wed, 30 Oct 2019 00:39:34 +0530	[thread overview]
Message-ID: <20191029190934.mr73g3ohwtgmndoo@yadavpratyush.com> (raw)
In-Reply-To: <nycvar.QRO.7.76.6.1910281412300.46@tvgsbejvaqbjf.bet>

On 28/10/19 02:15PM, Johannes Schindelin wrote:
> Hi Pratyush,
> 
> On Fri, 25 Oct 2019, Pratyush Yadav wrote:
> 
> > A lot of places have unused global variables declared. Remove them.
> >
> > It started as a couple of quick fixes and ended up in me writing an ugly
> > hacky script to catch all the instances of unused global declarations.
> > Lot more than I expected.
> >
> > The script can be found at [0].
> 
> Wouldn't it make more sense to integrate that script into a Makefile
> target `check`, rather than hiding it in a gist (that might become
> unavailable if GitHub goes away, as some of you seem to fear)?

The idea of the script was just a quick hack to list as many instances 
as possible, and then I could manually verify them. It turned out to be 
pretty accurate. I linked it here so other people can know what I used 
to find these, and can suggest some cases I missed, if any. That's why I 
linked to the script in the cover letter, and not in the commit message.

It would probably be a better idea to use a linter/static analyzer. This 
would provide much better coverage than a script written for one-off 
usage.

I looked up some linters. There doesn't seem to be a lot of options, and 
most of them are not very good. One option I found is [0], but it gives 
more useless warnings than useful ones. Tcl being a really difficult 
language to lint doesn't help. So unless someone knows a good Tcl 
linter, maybe the best compromise would be to clean up the script and 
add it to our build process. But I'd really prefer a "proper" linter.
 
> > The script at [0] will only catch the unused globals in 'proc's. But
> > git-gui also has 'method's and 'constructor's. Just change '^proc' to
> > '^constructor' or '^method'.
> 
> Why not use `grep -E -n -e '^(proc|method)'`?

No particular reason. I just wrote it for 'proc', and then realized I 
might as well check 'method' and 'constructor'. Like I said, the aim of 
the script was use-and-throw.
 
> > I manually checked each instance before removing just to be sure I'm not
> > doing something wrong. Still, a bit of testing would be appreciated.
> > Just apply this patch and go on about your business as usual. There
> > _should_ be no change in the behaviour. I tested some basic functions
> > like commit, push, etc, and they work fine for me.
> 
> I would, if I used Git GUI regularly ;-)

[0] https://github.com/Xilinx/XilinxTclStore/blob/master/support/linter/tcl_lint.tcl

-- 
Regards,
Pratyush Yadav

      reply	other threads:[~2019-10-29 19:09 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-25  1:32 Pratyush Yadav
2019-10-25  1:32 ` [PATCH 1/1] " Pratyush Yadav
2019-10-25  3:54   ` Junio C Hamano
2019-10-29 21:32     ` Pratyush Yadav
2019-10-28 13:15 ` [PATCH 0/1] " Johannes Schindelin
2019-10-29 19:09   ` Pratyush Yadav [this message]

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=20191029190934.mr73g3ohwtgmndoo@yadavpratyush.com \
    --to=me@yadavpratyush.com \
    --cc=Johannes.Schindelin@gmx.de \
    --cc=git@vger.kernel.org \
    --subject='Re: [PATCH 0/1] git-gui: remove unused global declarations' \
    /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

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).