git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Jeff King <peff@peff.net>
To: Emily Shaffer <emilyshaffer@google.com>
Cc: git@vger.kernel.org, jrnieder@gmail.com
Subject: Re: [PATCH] revision: remove stray whitespace when name empty
Date: Sun, 9 Jun 2019 09:00:04 -0400	[thread overview]
Message-ID: <20190609130004.GB23555@sigill.intra.peff.net> (raw)
In-Reply-To: <20190607225900.89299-1-emilyshaffer@google.com>

On Fri, Jun 07, 2019 at 03:59:00PM -0700, Emily Shaffer wrote:

> Teach show_object_with_name() to avoid writing a space before a name
> which is empty. Also teach tests for rev-list --objects --filter to not
> require a space between the object ID and name.
> [...]
> ---
> I don't see any reason _not_ to remove this stray whitespace at the end,
> since it seems like it just gets in the way of easy scripting. I also
> think this case will only present itself for root trees.

I'm a bit worried that this might break existing scripts. As ugly as
trailing whitespace is, it does tell you something here: that the object
is a root tree and not a commit.

So in the past I have done things like:

  git rev-list --objects --all | grep ' '

to get only the non-commits. I'm undecided on whether we're straying
into https://xkcd.com/1172/ territory here. I'd be more in favor if this
were making things significantly easier, but...

> show_object_with_name() inserts a space between an object's OID and name
> regardless of whether the name is empty or not. This causes 'git
> cat-file (--batch | --batch-check)' to fail to discover the type of root
> directories:
> 
>   git rev-list --objects --filter=tree:1 --max-count=1 HEAD \
>     | git cat-file --batch-check
>   git rev-parse HEAD: | xargs -I% git cat-file -t %
>   git rev-list --objects --filter=tree:1 --max-count=1 HEAD \
>     | xargs -I% echo "AA%AA"
>   git rev-list --objects --filter=tree:1 --max-count=1 HEAD \
>     | cut -f 1 -d ' ' | git cat-file --batch-check

Your patch only helps with this at all because you're using the "tree:1"
filter. It would not help:

  git rev-list --objects HEAD | git cat-file --batch-check

because there you'll have actual names which cat-file will choke on. So
it seems like this is helping only a very limited use case.

cat-file actually does know how to split on whitespace. Unfortunately it
does not do so by default, because that breaks some cases. See
97be04077f (cat-file: only split on whitespace when %(rest) is used,
2013-08-02).

So you _can_ do:

  git rev-list --objects HEAD |
  git cat-file --batch-check='%(objectname) %(objecttype) %(rest)'

But:

  1. That puts the %(rest) bits in your output, which you may not want.

  2. You have to actually specify the full format, so you might have to
     repeat batch-check's default format items.

I think it would be reasonable for cat-file to have an option to split
on whitespace (and if not given explicitly by the user, default to the
presence of %(rest) as we do now).

Alternatively, it would be reasonable to me to have an option for
"rev-list --objects" to have an option to suppress the filename (and
space) entirely.

I think in the longer run along those lines that "--objects" should
allow cat-file style pretty-print formats, which would eliminate the
need to pipe to cat-file in the first place. That makes this parsing
problem go away entirely, and it's way more efficient to boot (rev-list
already knows the types!).

-Peff

  parent reply	other threads:[~2019-06-09 13:00 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-07 22:59 [PATCH] revision: remove stray whitespace when name empty Emily Shaffer
2019-06-08  0:42 ` Eric Sunshine
2019-06-12 19:23   ` Emily Shaffer
2019-06-09 13:00 ` Jeff King [this message]
2019-06-10 16:29   ` Junio C Hamano
2019-06-12 19:37     ` Emily Shaffer
2019-06-13 15:20       ` Jeff King
2019-06-13 21:51 ` [PATCH v2] rev-list: teach --oid-only to enable piping Emily Shaffer
2019-06-14 16:07   ` Jeff King
2019-06-14 20:25     ` Junio C Hamano
2019-06-14 23:18       ` Emily Shaffer
2019-06-14 23:29     ` Emily Shaffer
2019-06-19 21:24       ` Jeff King
2019-06-14 23:48   ` [PATCH v3] rev-list: teach --no-object-names " Emily Shaffer
2019-06-17 22:32     ` Junio C Hamano
2019-06-18 22:08       ` Emily Shaffer
2019-06-18 22:29     ` [PATCH v4] " Emily Shaffer
2019-06-19 14:08       ` Junio C Hamano
2019-06-19 19:31         ` Emily Shaffer
2019-06-19 21:30           ` Jeff King
2019-06-19 20:56       ` [PATCH v5] " Emily Shaffer
2019-06-19 21:38         ` Jeff King

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=20190609130004.GB23555@sigill.intra.peff.net \
    --to=peff@peff.net \
    --cc=emilyshaffer@google.com \
    --cc=git@vger.kernel.org \
    --cc=jrnieder@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).