[PATCH v5] commit: add a commit.verbose config variable

Add commit.verbose configuration variable as a convenience for those
who always prefer --verbose.

Signed-off-by: Pranit Bauva <redacted>
Mentored-by: Eric Sunshine [off-list ref]

---
The previous versions of this patch are:
 - [v4] $gmane/288652
 - [v3] $gmane/288634
 - [v2] $gmane/288569
 - [v1] $gmane/287540

The changes with respect to the last version are :
 - Use the in built check-for-diff "editor" instead of cat and breaking
   the code unnecessarily.
 - Add test to check whether the verbose option did not break status.
---
 Documentation/config.txt     |  4 ++++
 Documentation/git-commit.txt |  3 ++-
 builtin/commit.c             |  4 ++++
 t/t7507-commit-verbose.sh    | 18 ++++++++++++++++++
 4 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/Documentation/config.txt b/Documentation/config.txt
index 01cca0a..9b93f6c 100644
--- a/Documentation/config.txt
+++ b/Documentation/config.txt

@@ -1110,6 +1110,10 @@ commit.template::
 	"`~/`" is expanded to the value of `$HOME` and "`~user/`" to the
 	specified user's home directory.
 
+commit.verbose::
+	A boolean to specify whether to always include the verbose option
+	with `git commit`. See linkgit:git-commit[1].
+
 credential.helper::
 	Specify an external helper to be called when a username or
 	password credential is needed; the helper may consult external

diff --git a/Documentation/git-commit.txt b/Documentation/git-commit.txt
index 9ec6b3c..d474226 100644
--- a/Documentation/git-commit.txt
+++ b/Documentation/git-commit.txt

@@ -290,7 +290,8 @@ configuration variable documented in linkgit:git-config[1].
 	what changes the commit has.
 	Note that this diff output doesn't have its
 	lines prefixed with '#'. This diff will not be a part
-	of the commit message.
+	of the commit message. See the `commit.verbose` configuration
+	variable in linkgit:git-config[1].
 +
 If specified twice, show in addition the unified diff between
 what would be committed and the worktree files, i.e. the unstaged

diff --git a/builtin/commit.c b/builtin/commit.c
index b3bd2d4..e0b96231 100644
--- a/builtin/commit.c
+++ b/builtin/commit.c

@@ -1505,6 +1505,10 @@ static int git_commit_config(const char *k, const char *v, void *cb)
 		sign_commit = git_config_bool(k, v) ? "" : NULL;
 		return 0;
 	}
+	if (!strcmp(k, "commit.verbose")) {
+		verbose = git_config_bool(k, v);
+		return 0;
+	}
 
 	status = git_gpg_config(k, v, NULL);
 	if (status)

diff --git a/t/t7507-commit-verbose.sh b/t/t7507-commit-verbose.sh
index 2ddf28c..bb7ce7c 100755
--- a/t/t7507-commit-verbose.sh
+++ b/t/t7507-commit-verbose.sh

@@ -96,4 +96,22 @@ test_expect_success 'verbose diff is stripped out with set core.commentChar' '
 	test_i18ngrep "Aborting commit due to empty commit message." err
 '
 
+test_expect_success 'commit.verbose true and --verbose omitted' '
+	test_config commit.verbose true &&
+	! git status | grep "*diff --git" &&
+	git commit --amend
+'
+
+test_expect_success 'commit.verbose true and --no-verbose' '
+	test_must_fail git -c commit.verbose=true commit --amend --no-verbose
+'
+
+test_expect_success 'commit.verbose false and --verbose' '
+	git -c commit.verbose=false commit --amend --verbose
+'
+
+test_expect_success 'commit.verbose false and --verbose omitted' '
+	test_must_fail git -c commit.verbose=false commit --amend
+'
+
 test_done

--

https://github.com/git/git/pull/205

On Sat, Mar 12, 2016 at 1:41 AM, Pranit Bauva [off-list ref] wrote:

Add commit.verbose configuration variable as a convenience for those
who always prefer --verbose.

Signed-off-by: Pranit Bauva <redacted>
Mentored-by: Eric Sunshine [off-list ref]

More typical would be to say Helped-by: rather than Mentored-by:.

Also, place your sign-off last.

This is a bogus regular expression; some greps may barf on it
outright, while others will treat "*" as a literal. Either way, it
could never match "diff --git" so the test would succeed by accident
even if git-status did erroneously respect "commit.verbose".

+       git commit --amend
+'

I realize that it's possible to (mis)read Junio's recommendation[1]
about also testing git-status as a hint to combine the git-commit and
git-status checks into a single test, but that's not what was
intended. These are conceptually distinct checks, thus they should
each have a separate test. I'd suggest restoring the
"commit.verbose=true and --verbose omitted" test as it was shown in
[2], and then adding a new test in this script for git-status which
looks something like this:

    test_expect_success 'status ignores commit.verbose=true' '
        git status >actual &&
        ! grep "^diff --git" actual
    '

[1]: http://article.gmane.org/gmane.comp.version-control.git/288648
[2]: http://article.gmane.org/gmane.comp.version-control.git/288674

+test_expect_success 'commit.verbose true and --no-verbose' '
+       test_must_fail git -c commit.verbose=true commit --amend --no-verbose
+'
+
+test_expect_success 'commit.verbose false and --verbose' '
+       git -c commit.verbose=false commit --amend --verbose
+'
+
+test_expect_success 'commit.verbose false and --verbose omitted' '
+       test_must_fail git -c commit.verbose=false commit --amend
+'

Not at all mandatory, but it wouldn't hurt to add a couple additional tests:

* commit.verbose=true and --verbose
* commit.verbose=false and --no-verbose

 test_done

On Mon, Mar 14, 2016 at 1:54 PM, Eric Sunshine [off-list ref] wrote:

On Sat, Mar 12, 2016 at 1:41 AM, Pranit Bauva [off-list ref] wrote:

quoted

Add commit.verbose configuration variable as a convenience for those
who always prefer --verbose.

Signed-off-by: Pranit Bauva <redacted>
Mentored-by: Eric Sunshine [off-list ref]

More typical would be to say Helped-by: rather than Mentored-by:.

Also, place your sign-off last.

Sure!

This is a bogus regular expression; some greps may barf on it
outright, while others will treat "*" as a literal. Either way, it
could never match "diff --git" so the test would succeed by accident
even if git-status did erroneously respect "commit.verbose".

I was unaware about this. Thanks for pointing it out

I realize that it's possible to (mis)read Junio's recommendation[1]
about also testing git-status as a hint to combine the git-commit and
git-status checks into a single test, but that's not what was
intended. These are conceptually distinct checks, thus they should
each have a separate test. I'd suggest restoring the
"commit.verbose=true and --verbose omitted" test as it was shown in
[2], and then adding a new test in this script for git-status which
looks something like this:

    test_expect_success 'status ignores commit.verbose=true' '
        git status >actual &&
        ! grep "^diff --git" actual
    '

I misread Junio's recommendation. Will change this.

Not at all mandatory, but it wouldn't hurt to add a couple additional tests:

* commit.verbose=true and --verbose
* commit.verbose=false and --no-verbose

I was thinking of putting these tests as when I was debugging (simply
by printing verbose variable), I found that when commit.verbose=true
and --verbose the value of the variable `verbose` is 2. But then I
thought it wouldn't be that useful. But since you have pointed it out
now, I will definitely include them.

Regards,
Pranit Bauva

On Mon, Mar 14, 2016 at 3:09 PM, Pranit Bauva [off-list ref] wrote:

On Mon, Mar 14, 2016 at 1:54 PM, Eric Sunshine [off-list ref] wrote:

quoted

Not at all mandatory, but it wouldn't hurt to add a couple additional tests:

* commit.verbose=true and --verbose
* commit.verbose=false and --no-verbose

I was thinking of putting these tests as when I was debugging (simply
by printing verbose variable), I found that when commit.verbose=true
and --verbose the value of the variable `verbose` is 2. But then I
thought it wouldn't be that useful. But since you have pointed it out
now, I will definitely include them.

The '2' is a result of OPT_VERBOSE() supporting multiple levels of
verbosity (that is, "--verbose --verbose" could give more noisy output
than "--verbose") for commands for which verbosity levels make sense.

In this case, I wasn't suggesting these news tests for that reason,
but merely to prove that the combinations behave in the expected way.
By the way, the just-submitted v6 seems to be lacking these new tests
(though, as I said, they are not by any means mandatory).

On Mon, Mar 14, 2016 at 5:29 PM, Eric Sunshine [off-list ref] wrote:

By the way, the just-submitted v6 seems to be lacking these new tests
(though, as I said, they are not by any means mandatory).

Ignore me. They are present in v6, yet I somehow overlooked them.