Re: [PATCH v2 2/2] log: add log.showSignature configuration variable
On Sat, Jun 18, 2016 at 8:25 AM, Mehul Jain <mehul.jain2029@gmail.com> wrote:
Show 16 quoted lines
> Users may want to always use "--show-signature" while using git-log and
> related commands.
>
> When log.showSignature is set to true, git-log and related commands will
> behave as if "--show-signature" was given to them.
>
> Note that this config variable is meant to affect git-log, git-show,
> git-whatchanged and git-reflog. Other commands like git-format-patch,
> git-rev-list are not to be affected by this config variable.
>
> Signed-off-by: Mehul Jain <mehul.jain2029@gmail.com>
> ---
> diff --git a/t/t4202-log.sh b/t/t4202-log.sh
> @@ -900,6 +900,31 @@ test_expect_success GPG '--no-show-signature overrides --show-signature' '
> +test_expect_success GPG 'log.showsignature=true behaves like --show-signature' '
> + git checkout -b test_sign master &&
It appears that you copied the bulk of this test from the 'log --graph --show-signature' test and changed it to create a new branch named 'test_sign' rather than 'signed', however...
Show 5 quoted lines
> + echo foo >foo &&
> + git add foo &&
> + git commit -S -m signed_commit &&
> + test_config log.showsignature true &&
> + git log -1 signed >actual &&
... you're incorrectly testing against the 'signed' branch rather than the 'test_sign' created specifically for this test.
Show 12 quoted lines
> + grep "gpg: Signature made" actual &&
> + grep "gpg: Good signature" actual
> +'
> +
> +test_expect_success GPG '--no-show-signature overrides log.showsignature=true' '
> + test_config log.showsignature true &&
> + git log -1 --no-show-signature signed >actual &&
> + ! grep "^gpg:" actual
> +'
> +
> +test_expect_success GPG '--show-signature overrides log.showsignature=false' '
> + test_when_finished "git reset --hard && git checkout master" &&
So, in the first of these three new tests, you're setting up some state by creating and checking out a new branch named 'test_sign', and leaving that branch checked out while these three tests run, and finally use test_when_finished() in the last of the three tests to restore sanity (by returning to the 'master' branch) when that test exits.
This is ugly and couples these three tests too tightly. It would be better to make each test more self-contained, not relying upon state left over from previous tests.
Show 5 quoted lines
> + test_config log.showsignature false &&
> + git log -1 --show-signature signed >actual &&
> + grep "gpg: Signature made" actual &&
> + grep "gpg: Good signature" actual
> +'
In fact, the original 'log --graph --show-signature' test which created the 'signed' branch, the new --no-show-signature test added in patch 1/2, and the three new tests here could all just work against the same 'signed' branch. There is no need to create a new 'test_sign' branch for these three tests, or a 'nosign' branch for the patch 1/2 test.
Therefore, it probably would make more sense to add a new distinct 'setup signed' test (or just enhance the existing 'setup' test) which creates the 'signed' branch and update the original 'log --graph --show-signature' to take advantage of it, as well as each of the new tests introduced by this patch series. And since each test would mention 'signed' explicitly in its git-log invocation, there's no need to leave that branch checked out, so the setup test itself only needs test_when_finished() to ensure that the current branch is restored to 'master'.
Mehul Jain