git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Taylor Blau <me@ttaylorr.com>
To: Thomas Guyot-Sionnest <tguyot@gmail.com>
Cc: git@vger.kernel.org, dermoth@aei.ca
Subject: Re: [PATCH 2/2] Allow passing pipes for input pipes to diff --no-index
Date: Fri, 18 Sep 2020 10:36:47 -0400	[thread overview]
Message-ID: <20200918143647.GB1606445@nand.local> (raw)
In-Reply-To: <20200918113256.8699-3-tguyot@gmail.com>

Hi Thomas,

On Fri, Sep 18, 2020 at 07:32:56AM -0400, Thomas Guyot-Sionnest wrote:
> A very handy way to pass data to applications is to use the <() process
> substitution syntax in bash variants. It allow comparing files streamed
> from a remote server or doing on-the-fly stream processing to alter the
> diff. These are usually implemented as a symlink that points to a bogus
> name (ex "pipe:[209326419]") but opens as a pipe.

This is true in bash, but sh does not support process substitution with
<().

> Git normally tracks symlinks targets. This patch makes it detect such
> pipes in --no-index mode and read the file normally like it would do for
> stdin ("-"), so they can be compared directly.
>
> Signed-off-by: Thomas Guyot-Sionnest <tguyot@gmail.com>
> ---
>  diff-no-index.c          |  56 ++++++++++--
>  t/t4053-diff-no-index.sh | 189 +++++++++++++++++++++++++++++++++++++++
>  2 files changed, 238 insertions(+), 7 deletions(-)
>
> diff --git a/diff-no-index.c b/diff-no-index.c
> index 7814eabfe0..779c686d23 100644
> --- a/diff-no-index.c
> +++ b/diff-no-index.c
> @@ -41,6 +41,33 @@ static int read_directory_contents(const char *path, struct string_list *list)
>   */
>  static const char file_from_standard_input[] = "-";
>
> +/* Check that file is - (STDIN) or unnamed pipe - explicitly
> + * avoid on-disk named pipes which could block
> + */
> +static int ispipe(const char *name)
> +{
> +	struct stat st;
> +
> +	if (name == file_from_standard_input)
> +		return 1;  /* STDIN */
> +
> +	if (!lstat(name, &st)) {
> +		if (S_ISLNK(st.st_mode)) {

I had to read this a few times to make sure that I got it; you want to
stat the link itself, and then check that it links to a pipe.

I'm not sure why, though. Do you want to avoid handling named FIFOs in
the code below? Your comment that they "could block" makes me think you
do, but I don't know why that would be a problem.

> +			/* symlink - read it and check it doesn't exists
> +			 * as a file yet link to a pipe */
> +			struct strbuf sb = STRBUF_INIT;
> +			strbuf_realpath(&sb, name, 0);
> +			/* We're abusing strbuf_realpath here, it may append
> +			 * pipe:[NNNNNNNNN] to an abs path */
> +			if (!stat(sb.buf, &st))

Statting sb.buf is confusing to me (especially when followed up by
another stat right below. Could you explain?

> +test_expect_success 'diff --no-index can diff piped subshells' '
> +	echo 1 >non/git/c &&
> +	test_expect_code 0 git diff --no-index non/git/b <(cat non/git/c) &&
> +	test_expect_code 0 git diff --no-index <(cat non/git/b) non/git/c &&
> +	test_expect_code 0 git diff --no-index <(cat non/git/b) <(cat non/git/c) &&
> +	test_expect_code 0 cat non/git/b | git diff --no-index - non/git/c &&
> +	test_expect_code 0 cat non/git/c | git diff --no-index non/git/b - &&
> +	test_expect_code 0 cat non/git/b | git diff --no-index - <(cat non/git/c) &&
> +	test_expect_code 0 cat non/git/c | git diff --no-index <(cat non/git/b) -
> +'

Indeed this test fails (Git thinks that the HERE-DOC is broken, but I
suspect it's just getting confused by the '<()'). This test (like almost
all other tests in Git) use /bin/sh as its shebang. Does your /bin/sh
actually point to bash?

If you did want to test something like this, you'd need to source
t/lib-bash.sh instead of t/test-lib.sh.

Unrelated to the above comment, but there are a few small style nits
that I notice:

  - There is no need to run with 'test_expect_code 0' since the test is
    marked as 'test_expect_success' and the commands are all in an '&&'
    chain. (This does appear to be common style for others in t4053, so
    you may just be matching it--which is fine--but an additional
    clean-up on top to modernize would be appreciated, too).

  - The cat pipe is unnecessary, and is also violating a rule that we
    don't place 'git' on the right-hand side of a pipe (can you redirect
    the file at the end instead?).

Documentation/CodingGuidelines is a great place to look if you are ever
curious about whether something is in good style.

> +test_expect_success 'diff --no-index finds diff in piped subshells' '
> +	(
> +		set -- <(cat /dev/null) <(cat /dev/null)

Why is this necessary?

> +		cat <<-EOF >expect
> +			diff --git a$1 b$2
> +			--- a$1
> +			+++ b$2
> +			@@ -1 +1 @@
> +			-1
> +			+2
> +		EOF
> +	) &&
> +	test_expect_code 1 \
> +		git diff --no-index <(cat non/git/b) <(sed s/1/2/ non/git/c) >actual &&
> +	test_cmp expect actual
> +'

Thanks,
Taylor

  reply	other threads:[~2020-09-18 14:37 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-18 11:32 Allow passing pipes to diff --no-index + bugfix Thomas Guyot-Sionnest
2020-09-18 11:32 ` [PATCH 1/2] diff: Fix modified lines stats with --stat and --numstat Thomas Guyot-Sionnest
2020-09-18 14:46   ` Taylor Blau
2020-09-18 15:10     ` Thomas Guyot-Sionnest
2020-09-18 17:37       ` Jeff King
2020-09-18 18:00         ` Thomas Guyot-Sionnest
2020-09-20  4:53       ` Thomas Guyot
2020-09-18 17:27   ` Jeff King
2020-09-18 17:52     ` Thomas Guyot-Sionnest
2020-09-18 18:06       ` Junio C Hamano
2020-09-23 19:16         ` Johannes Schindelin
2020-09-23 19:23           ` Junio C Hamano
2020-09-23 20:44             ` Johannes Schindelin
2020-09-24  4:49               ` Thomas Guyot
2020-09-24  5:24                 ` [PATCH v3] " Thomas Guyot-Sionnest
2020-09-24  7:41                   ` [PATCH v4] " Thomas Guyot-Sionnest
2020-09-24  6:40                 ` [PATCH 1/2] " Junio C Hamano
2020-09-24  7:13                   ` Thomas Guyot
2020-09-24 17:19                     ` Junio C Hamano
2020-09-24 17:38                       ` Junio C Hamano
2020-09-23 15:05     ` Johannes Schindelin
2020-09-20 13:09   ` [PATCH v2] " Thomas Guyot-Sionnest
2020-09-20 15:39     ` Taylor Blau
2020-09-20 16:38       ` Thomas Guyot
2020-09-20 19:11       ` Junio C Hamano
2020-09-20 20:08         ` Junio C Hamano
2020-09-20 20:36         ` Junio C Hamano
2020-09-20 22:15           ` Junio C Hamano
2020-09-21 19:26         ` Jeff King
2020-09-21 21:51           ` Junio C Hamano
2020-09-21 22:20             ` Jeff King
2020-09-21 22:37               ` Junio C Hamano
2020-09-18 11:32 ` [PATCH 2/2] Allow passing pipes for input pipes to diff --no-index Thomas Guyot-Sionnest
2020-09-18 14:36   ` Taylor Blau [this message]
2020-09-18 16:34     ` Thomas Guyot-Sionnest
2020-09-18 17:19       ` Jeff King
2020-09-18 17:21         ` Jeff King
2020-09-18 17:39         ` Thomas Guyot-Sionnest
2020-09-18 17:48         ` Junio C Hamano
2020-09-18 18:02           ` Jeff King
2020-09-20 12:54             ` Thomas Guyot
2020-09-21 19:31               ` Jeff King
2020-09-21 20:14                 ` Junio C Hamano
2020-09-18 17:58       ` Taylor Blau
2020-09-18 18:05         ` Jeff King
2020-09-18 17:20     ` Jeff King
2020-09-18 18:00       ` Taylor Blau
2020-09-18 21:56   ` brian m. carlson
2020-09-18 17:51 ` Allow passing pipes to diff --no-index + bugfix Junio C Hamano
2020-09-18 18:24   ` Thomas Guyot-Sionnest

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=20200918143647.GB1606445@nand.local \
    --to=me@ttaylorr.com \
    --cc=dermoth@aei.ca \
    --cc=git@vger.kernel.org \
    --cc=tguyot@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).