* [PATCH] submodule add: show 'add --dry-run' stderr when aborting
@ 2020-01-08 0:31 Kyle Meyer
2020-01-08 21:41 ` Josh Steadmon
2020-01-16 16:17 ` Johannes Schindelin
0 siblings, 2 replies; 5+ messages in thread
From: Kyle Meyer @ 2020-01-08 0:31 UTC (permalink / raw)
To: git; +Cc: Kyle Meyer, Yaroslav O Halchenko
Unless --force is specified, 'submodule add' checks if the destination
path is ignored by calling 'git add --dry-run --ignore-missing', and,
if that call fails, aborts with a custom "path is ignored" message (a
slight variant of what 'git add' shows). Aborting early rather than
letting the downstream 'git add' call fail is done so that the command
exits before cloning into the destination path. However, in rare
cases where the dry-run call fails for a reason other than the path
being ignored---for example, due to a preexisting index.lock
file---displaying the "ignored path" error message hides the real
source of the failure.
Instead of displaying the tailored "ignored path" message, let's
report the standard error from the dry run to give the caller more
accurate information about failures that are not due to an ignored
path.
For the ignored path case, this leads to the following change in the
error message:
The following [-path is-]{+paths are+} ignored by one of your .gitignore files:
<destination path>
Use -f if you really want to add [-it.-]{+them.+}
The new phrasing is a bit awkward, because 'submodule add' is only
dealing with one destination path. Alternatively, we could continue
to use the tailored message when the exit code is 1 (the expected
status for a failure due to an ignored path) and relay the standard
error for all other non-zero exits. That, however, risks hiding the
message of unrelated failures that share an exit code of 1, so it
doesn't seem worth doing just to avoid a clunkier, but still clear,
error message.
Signed-off-by: Kyle Meyer <kyle@kyleam.com>
---
git-submodule.sh | 14 ++++++++------
t/t7400-submodule-basic.sh | 15 +++++++++++++--
2 files changed, 21 insertions(+), 8 deletions(-)
diff --git a/git-submodule.sh b/git-submodule.sh
index aaa1809d24..afcb4c0948 100755
--- a/git-submodule.sh
+++ b/git-submodule.sh
@@ -241,13 +241,15 @@ cmd_add()
die "$(eval_gettext "'\$sm_path' does not have a commit checked out")"
fi
- if test -z "$force" &&
- ! git add --dry-run --ignore-missing --no-warn-embedded-repo "$sm_path" > /dev/null 2>&1
+ if test -z "$force"
then
- eval_gettextln "The following path is ignored by one of your .gitignore files:
-\$sm_path
-Use -f if you really want to add it." >&2
- exit 1
+ dryerr=$(git add --dry-run --ignore-missing --no-warn-embedded-repo "$sm_path" 2>&1 >/dev/null)
+ res=$?
+ if test $res -ne 0
+ then
+ echo >&2 "$dryerr"
+ exit $res
+ fi
fi
if test -n "$custom_name"
diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
index 7f75bb1be6..42a00f95b9 100755
--- a/t/t7400-submodule-basic.sh
+++ b/t/t7400-submodule-basic.sh
@@ -156,9 +156,9 @@ test_expect_success 'submodule add to .gitignored path fails' '
(
cd addtest-ignore &&
cat <<-\EOF >expect &&
- The following path is ignored by one of your .gitignore files:
+ The following paths are ignored by one of your .gitignore files:
submod
- Use -f if you really want to add it.
+ Use -f if you really want to add them.
EOF
# Does not use test_commit due to the ignore
echo "*" > .gitignore &&
@@ -191,6 +191,17 @@ test_expect_success 'submodule add to reconfigure existing submodule with --forc
)
'
+test_expect_success 'submodule add relays add --dry-run stderr' '
+ test_when_finished "rm -rf addtest/.git/index.lock" &&
+ (
+ cd addtest &&
+ : >.git/index.lock &&
+ ! git submodule add "$submodurl" sub-while-locked 2>output.err &&
+ test_i18ngrep "^fatal: .*index\.lock" output.err &&
+ test_path_is_missing sub-while-locked
+ )
+'
+
test_expect_success 'submodule add --branch' '
echo "refs/heads/initial" >expect-head &&
cat <<-\EOF >expect-heads &&
base-commit: 042ed3e048af08014487d19196984347e3be7d1c
--
2.24.1
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [PATCH] submodule add: show 'add --dry-run' stderr when aborting
2020-01-08 0:31 [PATCH] submodule add: show 'add --dry-run' stderr when aborting Kyle Meyer
@ 2020-01-08 21:41 ` Josh Steadmon
2020-01-09 2:39 ` Kyle Meyer
2020-01-16 16:17 ` Johannes Schindelin
1 sibling, 1 reply; 5+ messages in thread
From: Josh Steadmon @ 2020-01-08 21:41 UTC (permalink / raw)
To: Kyle Meyer; +Cc: git, Yaroslav O Halchenko
On 2020.01.07 19:31, Kyle Meyer wrote:
> Unless --force is specified, 'submodule add' checks if the destination
> path is ignored by calling 'git add --dry-run --ignore-missing', and,
> if that call fails, aborts with a custom "path is ignored" message (a
> slight variant of what 'git add' shows). Aborting early rather than
> letting the downstream 'git add' call fail is done so that the command
> exits before cloning into the destination path. However, in rare
> cases where the dry-run call fails for a reason other than the path
> being ignored---for example, due to a preexisting index.lock
> file---displaying the "ignored path" error message hides the real
> source of the failure.
>
> Instead of displaying the tailored "ignored path" message, let's
> report the standard error from the dry run to give the caller more
> accurate information about failures that are not due to an ignored
> path.
>
> For the ignored path case, this leads to the following change in the
> error message:
>
> The following [-path is-]{+paths are+} ignored by one of your .gitignore files:
> <destination path>
> Use -f if you really want to add [-it.-]{+them.+}
>
> The new phrasing is a bit awkward, because 'submodule add' is only
> dealing with one destination path. Alternatively, we could continue
> to use the tailored message when the exit code is 1 (the expected
> status for a failure due to an ignored path) and relay the standard
> error for all other non-zero exits. That, however, risks hiding the
> message of unrelated failures that share an exit code of 1, so it
> doesn't seem worth doing just to avoid a clunkier, but still clear,
> error message.
>
> Signed-off-by: Kyle Meyer <kyle@kyleam.com>
> ---
> git-submodule.sh | 14 ++++++++------
> t/t7400-submodule-basic.sh | 15 +++++++++++++--
> 2 files changed, 21 insertions(+), 8 deletions(-)
>
> diff --git a/git-submodule.sh b/git-submodule.sh
> index aaa1809d24..afcb4c0948 100755
> --- a/git-submodule.sh
> +++ b/git-submodule.sh
> @@ -241,13 +241,15 @@ cmd_add()
> die "$(eval_gettext "'\$sm_path' does not have a commit checked out")"
> fi
>
> - if test -z "$force" &&
> - ! git add --dry-run --ignore-missing --no-warn-embedded-repo "$sm_path" > /dev/null 2>&1
> + if test -z "$force"
> then
> - eval_gettextln "The following path is ignored by one of your .gitignore files:
> -\$sm_path
> -Use -f if you really want to add it." >&2
> - exit 1
> + dryerr=$(git add --dry-run --ignore-missing --no-warn-embedded-repo "$sm_path" 2>&1 >/dev/null)
> + res=$?
> + if test $res -ne 0
> + then
> + echo >&2 "$dryerr"
> + exit $res
> + fi
> fi
>
> if test -n "$custom_name"
Seems reasonable: we move the dry-run add inside the if block, and
capture its stderr, then display the message only if the return code is
non-zero.
> diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
> index 7f75bb1be6..42a00f95b9 100755
> --- a/t/t7400-submodule-basic.sh
> +++ b/t/t7400-submodule-basic.sh
> @@ -156,9 +156,9 @@ test_expect_success 'submodule add to .gitignored path fails' '
> (
> cd addtest-ignore &&
> cat <<-\EOF >expect &&
> - The following path is ignored by one of your .gitignore files:
> + The following paths are ignored by one of your .gitignore files:
> submod
> - Use -f if you really want to add it.
> + Use -f if you really want to add them.
> EOF
> # Does not use test_commit due to the ignore
> echo "*" > .gitignore &&
> @@ -191,6 +191,17 @@ test_expect_success 'submodule add to reconfigure existing submodule with --forc
> )
> '
>
> +test_expect_success 'submodule add relays add --dry-run stderr' '
> + test_when_finished "rm -rf addtest/.git/index.lock" &&
> + (
> + cd addtest &&
> + : >.git/index.lock &&
> + ! git submodule add "$submodurl" sub-while-locked 2>output.err &&
> + test_i18ngrep "^fatal: .*index\.lock" output.err &&
> + test_path_is_missing sub-while-locked
> + )
> +'
> +
> test_expect_success 'submodule add --branch' '
> echo "refs/heads/initial" >expect-head &&
> cat <<-\EOF >expect-heads &&
I had to look up what ":" does, but it looks like it's reasonably widely
used in other tests so that seems fine. However, it looks like you don't
even need the : command and can just ">.git/index.lock" by itself (see
the "setup - initial commit" test case in this file for an example).
> base-commit: 042ed3e048af08014487d19196984347e3be7d1c
> --
> 2.24.1
>
Looks good to me. Thanks for the patch!
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] submodule add: show 'add --dry-run' stderr when aborting
2020-01-08 21:41 ` Josh Steadmon
@ 2020-01-09 2:39 ` Kyle Meyer
0 siblings, 0 replies; 5+ messages in thread
From: Kyle Meyer @ 2020-01-09 2:39 UTC (permalink / raw)
To: Josh Steadmon; +Cc: git, Yaroslav O Halchenko
Josh Steadmon <steadmon@google.com> writes:
> On 2020.01.07 19:31, Kyle Meyer wrote:
[...]
>> +test_expect_success 'submodule add relays add --dry-run stderr' '
>> + test_when_finished "rm -rf addtest/.git/index.lock" &&
>> + (
>> + cd addtest &&
>> + : >.git/index.lock &&
[...]
> I had to look up what ":" does, but it looks like it's reasonably widely
> used in other tests so that seems fine. However, it looks like you don't
> even need the : command and can just ">.git/index.lock" by itself (see
> the "setup - initial commit" test case in this file for an example).
Indeed. I spot some recent commits that added new instances of ": >",
so I suspect it's not strongly discouraged in this project, but I'm not
particularly attached to the unneeded colon :>
> Looks good to me. Thanks for the patch!
Thanks for the review!
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] submodule add: show 'add --dry-run' stderr when aborting
2020-01-08 0:31 [PATCH] submodule add: show 'add --dry-run' stderr when aborting Kyle Meyer
2020-01-08 21:41 ` Josh Steadmon
@ 2020-01-16 16:17 ` Johannes Schindelin
2020-01-16 20:23 ` Junio C Hamano
1 sibling, 1 reply; 5+ messages in thread
From: Johannes Schindelin @ 2020-01-16 16:17 UTC (permalink / raw)
To: Kyle Meyer, Junio C Hamano; +Cc: git, Yaroslav O Halchenko
Hi (in particular Junio),
On Tue, 7 Jan 2020, Kyle Meyer wrote:
> diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
> index 7f75bb1be6..42a00f95b9 100755
> --- a/t/t7400-submodule-basic.sh
> +++ b/t/t7400-submodule-basic.sh
> @@ -156,9 +156,9 @@ test_expect_success 'submodule add to .gitignored path fails' '
> (
> cd addtest-ignore &&
> cat <<-\EOF >expect &&
> - The following path is ignored by one of your .gitignore files:
> + The following paths are ignored by one of your .gitignore files:
> submod
> - Use -f if you really want to add it.
> + Use -f if you really want to add them.
I think this got mis-merged when merging down `km/submodule-add-errmsg`:
it needs to be prefixed with `hint:` because of `hw/advice-add-nothing`,
i.e.
-- snipsnap --
diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
index 42a00f95b9d..a6973a3003c 100755
--- a/t/t7400-submodule-basic.sh
+++ b/t/t7400-submodule-basic.sh
@@ -158,7 +158,7 @@ test_expect_success 'submodule add to .gitignored path fails' '
cat <<-\EOF >expect &&
The following paths are ignored by one of your .gitignore files:
submod
- Use -f if you really want to add them.
+ hint: Use -f if you really want to add them.
EOF
# Does not use test_commit due to the ignore
echo "*" > .gitignore &&
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH] submodule add: show 'add --dry-run' stderr when aborting
2020-01-16 16:17 ` Johannes Schindelin
@ 2020-01-16 20:23 ` Junio C Hamano
0 siblings, 0 replies; 5+ messages in thread
From: Junio C Hamano @ 2020-01-16 20:23 UTC (permalink / raw)
To: Johannes Schindelin; +Cc: Kyle Meyer, git, Yaroslav O Halchenko
Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:
> I think this got mis-merged when merging down `km/submodule-add-errmsg`:
> it needs to be prefixed with `hint:` because of `hw/advice-add-nothing`,
> i.e.
>
> -- snipsnap --
> diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
> index 42a00f95b9d..a6973a3003c 100755
> --- a/t/t7400-submodule-basic.sh
> +++ b/t/t7400-submodule-basic.sh
> @@ -158,7 +158,7 @@ test_expect_success 'submodule add to .gitignored path fails' '
> cat <<-\EOF >expect &&
> The following paths are ignored by one of your .gitignore files:
> submod
> - Use -f if you really want to add them.
> + hint: Use -f if you really want to add them.
> EOF
> # Does not use test_commit due to the ignore
> echo "*" > .gitignore &&
Thanks, will prepare merge-fix material for this.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2020-01-16 20:23 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-01-08 0:31 [PATCH] submodule add: show 'add --dry-run' stderr when aborting Kyle Meyer
2020-01-08 21:41 ` Josh Steadmon
2020-01-09 2:39 ` Kyle Meyer
2020-01-16 16:17 ` Johannes Schindelin
2020-01-16 20:23 ` Junio C Hamano
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).