Hi Joel,
On Wed, 4 Apr 2018, Joel Teichroeb wrote:
> In preparation for converting the stash command incrementally to
> a builtin command, this patch improves test coverage of the option
> parsing. Both for having too many parameters, or too few.
Very good.
> Signed-off-by: Joel Teichroeb <joel@xxxxxxxxxxxxx>
> ---
> t/t3903-stash.sh | 35 +++++++++++++++++++++++++++++++++++
> 1 file changed, 35 insertions(+)
>
> diff --git a/t/t3903-stash.sh b/t/t3903-stash.sh
> index aefde7b172..4eaa4cae9a 100755
> --- a/t/t3903-stash.sh
> +++ b/t/t3903-stash.sh
> @@ -444,6 +444,36 @@ test_expect_failure 'stash file to directory' '
> test foo = "$(cat file/file)"
> '
>
> +test_expect_success 'giving too many ref agruments does not modify files' '
Quick, before Eric beats me to it! A typo! s/agruments/arguments/
> + git stash clear &&
> + test_when_finished "git reset --hard HEAD" &&
> + echo foo >file2 &&
> + git stash &&
> + echo bar >file2 &&
> + git stash &&
> + test-chmtime =123456789 file2 &&
> + for type in apply pop "branch stash-branch"
> + do
> + test_must_fail git stash $type stash@{0} stash@{1} 2>err &&
> + test_i18ngrep "Too many" err &&
> + test 123456789 = $(test-chmtime -v +0 file2 | sed 's/[^0-9].*$//') || return 1
Not your problem, but if there is future work on this (read: if I get to
mentor a GSoC student, and if I get them to work on it: this idiom
`test-chmtime -v +0 ... | sed ...` is too common, there really *should* be
a `test-chmtime --get ...`).
Any prospective GSoC student: you know what I have in stock for you ;-)
> + done
> +'
> +
> +test_expect_success 'giving too many ref agruments to drop does not drop anything' '
s/agruments/arguments/
Also, we try to keep the lines to <= 80 columns. So maybe
+test_expect_success 'drop: too many arguments errors out (does nothing)' '
> + git stash list > stashlist1 &&
Again, before Eric can beat me to it: we prefer the syntax `>file` (i.e.
without a space between the `>` and the file name).
BTW is there a public branch with your patches? I am not so much of a fan
of reviewing and having all the work being punted back to the contributor,
especially when we failed you for so long in reviewing these patches. I
would love, for example, to open a PR on GitHub (even if that would only
apply to code changes, not so much to commit message changes, not until we
have rebase support reword! and drop! keywords).
> + test_must_fail git stash drop stash@{0} stash@{1} 2>err &&
> + test_i18ngrep "Too many" err &&
> + git stash list > stashlist2 &&
> + test_cmp stashlist1 stashlist2
> +'
It might make sense to rename stashlist1 to `expect` an `stashlist2` to
`actual`, to clarify the roles.
> +
> +test_expect_success 'giving too many ref agruments to show does not show anything' '
Maybe
test_expect_success 'show: error on too may arguments (show nothing)'
to keep within the 80 columns/line limit?
The rest looks obviously correct to me.
Thanks!
Dscho
