[PATCH] Document ls-files -t as obsolete.

[PATCH] Document ls-files -t as obsolete.

[Matthieu Moy]

The behavior of "git ls-files -t" is very misleading (see
http://thread.gmane.org/gmane.comp.version-control.git/126516 and
http://thread.gmane.org/gmane.comp.version-control.git/144394/focus=144397
for examples of mislead users) and badly documented, hence we point the
users to superior alternatives.

The feature is marked as "obsolete" but not "scheduled for removal" since
it's a plumbing command, scripts might use it, and Git testsuite already
uses it to test the state of the index.

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
---
 Documentation/git-ls-files.txt |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/Documentation/git-ls-files.txt b/Documentation/git-ls-files.txt
index 3521637..7b86bb5 100644
--- a/Documentation/git-ls-files.txt
+++ b/Documentation/git-ls-files.txt
@@ -106,6 +106,11 @@ OPTIONS
 	with `-s` or `-u` options does not make any sense.
 
 -t::
+	This feature is deprecated. Authors of scripts should use
+	linkgit:git-status[1] `--porcelain`, and users can look at
+	linkgit:git-status[1] `--short` or linkgit:git-diff[1]
+	`--name-status` for alternatives.
+
 	Identify the file status with the following tags (followed by
 	a space) at the start of each line:
 	H::	cached
-- 
1.7.0.3.299.gb22d9.dirty

Re: [PATCH] Document ls-files -t as obsolete.

[Junio C Hamano]

Matthieu Moy <Matthieu.Moy@imag.fr> writes:

> The behavior of "git ls-files -t" is very misleading (see
> http://thread.gmane.org/gmane.comp.version-control.git/126516 and
> http://thread.gmane.org/gmane.comp.version-control.git/144394/focus=144397
> for examples of mislead users) and badly documented, hence we point the
> users to superior alternatives.

I am no entirely happy with this patch.

If all your scripted Porcelain wants to see is "what are the paths I might
want to run 'git add' on?", you do not want to run diff twice (which is
what "status" does); you would say "ls-files -t -m -o --exclude-standard",
instead. The alternative is not even superiour in this case.

Having said that, I personally don't think of any other combination of
flags with which "ls-files -t" is useful.

If there are parts of the documentation that are misleading, perhaps that
is what your patch should be fixing instead, no?

Re: [PATCH] Document ls-files -t as obsolete.

[Tor Arntsen]

Sorry for being stupid and a bit off-topic, but..

>  -t::
> +       This feature is deprecated. Authors of scripts should use
> +       linkgit:git-status[1] `--porcelain`, [...]

I've always understood 'porcelain' as for users, and 'plumbing' for
scripts.. so, if the option is meant for scripting (as opposed to
--short), why wasn't it called --plumbing? There's obviously something
I'm not getting here! -:)

-Tor

Re: [PATCH] Document ls-files -t as obsolete.

[Daniel Grace]

On Wed, Apr 14, 2010 at 9:39 AM, Junio C Hamano <gitster@pobox.com> wrote:

> If all your scripted Porcelain wants to see is "what are the paths I might
> want to run 'git add' on?", you do not want to run diff twice (which is
> what "status" does); you would say "ls-files -t -m -o --exclude-standard",
> instead. The alternative is not even superiour in this case.

Why -m? I have a script that attempts to ask just this question and I
don't use -m. Reading the documentation, I'm not sure why I would. In
what cases would I get wrong information without -m? For the record, I
also don't use -t, though using -t in this case does make sense to me,
I just apparently didn't notice that bit of the documentation when
writing that feature.

Daniel
http://www.doomstick.com

Re: [PATCH] Document ls-files -t as obsolete.

[Matthieu Moy]

Junio C Hamano <gitster@pobox.com> writes:

> Matthieu Moy <Matthieu.Moy@imag.fr> writes:
>
>> The behavior of "git ls-files -t" is very misleading (see
>> http://thread.gmane.org/gmane.comp.version-control.git/126516 and
>> http://thread.gmane.org/gmane.comp.version-control.git/144394/focus=144397
>> for examples of mislead users) and badly documented, hence we point the
>> users to superior alternatives.
>
> I am no entirely happy with this patch.
>
> If all your scripted Porcelain wants to see is "what are the paths I might
> want to run 'git add' on?", you do not want to run diff twice (which is
> what "status" does); you would say "ls-files -t -m -o --exclude-standard",
> instead. The alternative is not even superiour in this case.
>
> Having said that, I personally don't think of any other combination of
> flags with which "ls-files -t" is useful.
>
> If there are parts of the documentation that are misleading, perhaps that
> is what your patch should be fixing instead, no?

After having tried to understand how "git ls-files -t" was supposed to
work
( http://thread.gmane.org/gmane.comp.version-control.git/126516/focus=126520 ),
I admit that I have no idea how to document such weird behavior.

If power users want to use "git ls-files -t", let them do (I'm not
saying the feature should be removed), but I don't see a point trying
to help people to use it.

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/

Re: [PATCH] Document ls-files -t as obsolete.

[Matthieu Moy]

Tor Arntsen <tor@spacetec.no> writes:

> Sorry for being stupid and a bit off-topic, but..
>
>>  -t::
>> +       This feature is deprecated. Authors of scripts should use
>> +       linkgit:git-status[1] `--porcelain`, [...]
>
> I've always understood 'porcelain' as for users, and 'plumbing' for
> scripts.. so, if the option is meant for scripting (as opposed to
> --short), why wasn't it called --plumbing? There's obviously something
> I'm not getting here! -:)

There's prior art ("git annotate --porcelain" at least). I disagree
with the rationale, but --porcelain means "to be used by porcelain
scripts", hence "behave like plubing".

--
Matthieu Moy
http://www-verimag.imag.fr/~moy/