Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

git-mergetool: improve error code paths and messages #1827

Open
wants to merge 5 commits into
base: master
Choose a base branch
from

Conversation

phil-blain
Copy link

@phil-blain phil-blain commented Nov 11, 2024

Changes in v2:
As suggested by Junio:

  • 3/5: moved the error message to setup_tool itself, and adjusted the commit message
  • 3/5: made the test more robust
  • 4/5: adjusted the error message

v1:
These are a few improvements to improve existing error messages in 'git mergetool', and make sure that errors are not quiet, along with a small completion update in 1/1.

CC: Seth House [email protected]
CC: David Aguilar [email protected]
CC: Johannes Sixt [email protected]

In git-mergetool--lib.sh::get_merge_tool_path, we check if the chosen
tool is valid via valid_tool and exit with an error message if not. This
error message mentions "Unknown merge tool", even if the command the
user tried was 'git difftool --tool=unknown'. Use the global 'TOOL_MODE'
variable for a more correct error message.

Signed-off-by: Philippe Blain <[email protected]>
@phil-blain
Copy link
Author

/submit

2 similar comments
@phil-blain
Copy link
Author

/submit

@phil-blain
Copy link
Author

/submit

Copy link

gitgitgadget bot commented Nov 13, 2024

Submitted as [email protected]

To fetch this version into FETCH_HEAD:

git fetch https://github.com/gitgitgadget/git/ pr-1827/phil-blain/absent-mergetool-v1

To fetch this version to local tag pr-1827/phil-blain/absent-mergetool-v1:

git fetch --no-tags https://github.com/gitgitgadget/git/ tag pr-1827/phil-blain/absent-mergetool-v1

@phil-blain
Copy link
Author

Hi @dscho !

My first two tries (one, two) at /submit failed because (if I understand correctly) the node.js v18.20.5 released today was not yet up at https://nodejs.org/dist/v18.20.5/, although the folder existed already. I guess this is just bad luck and there is nothing much Gitgitgadget can do ? And changes would need to go into the NodeTool Azure task to make it more robust ?

@@ -474,7 +474,7 @@ get_merge_tool_path () {
merge_tool="$1"
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Junio C Hamano wrote (reply to this):

"Philippe Blain via GitGitGadget" <[email protected]> writes:

> From: Philippe Blain <[email protected]>
>
> In git-mergetool--lib.sh::get_merge_tool_path, we check if the chosen
> tool is valid via valid_tool and exit with an error message if not. This
> error message mentions "Unknown merge tool", even if the command the
> user tried was 'git difftool --tool=unknown'. Use the global 'TOOL_MODE'
> variable for a more correct error message.

Makes sense.  Is this something we can easily test to catch future
regression, or is it too trivial to matter?

I wouldn't mind if the answer were "the latter" ;-)

Thanks.

> Signed-off-by: Philippe Blain <[email protected]>
> ---
>  git-mergetool--lib.sh | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/git-mergetool--lib.sh b/git-mergetool--lib.sh
> index 1ff26170ffc..269a60ea44c 100644
> --- a/git-mergetool--lib.sh
> +++ b/git-mergetool--lib.sh
> @@ -474,7 +474,7 @@ get_merge_tool_path () {
>  	merge_tool="$1"
>  	if ! valid_tool "$merge_tool"
>  	then
> -		echo >&2 "Unknown merge tool $merge_tool"
> +		echo >&2 "Unknown $TOOL_MODE tool $merge_tool"
>  		exit 1
>  	fi
>  	if diff_mode

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Philippe Blain wrote (reply to this):

Hi Junio (sorry for a late response),

Le 2024-11-12 à 20:27, Junio C Hamano a écrit :
> "Philippe Blain via GitGitGadget" <[email protected]> writes:
> 
>> From: Philippe Blain <[email protected]>
>>
>> In git-mergetool--lib.sh::get_merge_tool_path, we check if the chosen
>> tool is valid via valid_tool and exit with an error message if not. This
>> error message mentions "Unknown merge tool", even if the command the
>> user tried was 'git difftool --tool=unknown'. Use the global 'TOOL_MODE'
>> variable for a more correct error message.
> 
> Makes sense.  Is this something we can easily test to catch future
> regression, or is it too trivial to matter?
> 
> I wouldn't mind if the answer were "the latter" ;-)

With the changes in the next commit of the series, this particular error
becomes hard to trigger, as setup_user_tool will return with an error
before the error message change in this patch is reached. So I would way
it is not worth to add a test for this particular code path since it seems like
it becomes unreachable in the next commit (but I could be wrong). So mostly
"the latter" is my answer.

Thanks,
Philippe.

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Philippe Blain wrote (reply to this):

Le 2024-11-22 à 13:57, Philippe Blain a écrit :
> Hi Junio (sorry for a late response),
> 
> Le 2024-11-12 à 20:27, Junio C Hamano a écrit :
>> "Philippe Blain via GitGitGadget" <[email protected]> writes:
>>
>>> From: Philippe Blain <[email protected]>
>>>
>>> In git-mergetool--lib.sh::get_merge_tool_path, we check if the chosen
>>> tool is valid via valid_tool and exit with an error message if not. This
>>> error message mentions "Unknown merge tool", even if the command the
>>> user tried was 'git difftool --tool=unknown'. Use the global 'TOOL_MODE'
>>> variable for a more correct error message.
>>
>> Makes sense.  Is this something we can easily test to catch future
>> regression, or is it too trivial to matter?
>>
>> I wouldn't mind if the answer were "the latter" ;-)
> 
> With the changes in the next commit of the series,

correction: with the changes in 3/5 and 5/5,

> this particular error
> becomes hard to trigger, as setup_user_tool will return with an error
> before the error message change in this patch is reached. So I would way
> it is not worth to add a test for this particular code path since it seems like
> it becomes unreachable in the next commit (but I could be wrong). So mostly
> "the latter" is my answer.

@@ -159,14 +159,18 @@ check_unchanged () {
}
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Junio C Hamano wrote (reply to this):

"Philippe Blain via GitGitGadget" <[email protected]> writes:

>  valid_tool () {
> -	setup_tool "$1" && return 0
> +	setup_tool "$1" 2>/dev/null && return 0
>  	cmd=$(get_merge_tool_cmd "$1")
>  	test -n "$cmd"
>  }

As we are checking if a tool is valid, it is normal for setup_tool
to fail when we are checking is not valid (aka "fails to get set
up").  There is no need to show an error message for such a failure,
as the callers of valid_tool would do so if they wish.  OK.

>  setup_user_tool () {
>  	merge_tool_cmd=$(get_merge_tool_cmd "$tool")
> -	test -n "$merge_tool_cmd" || return 1
> +	if test -z "$merge_tool_cmd"
> +	then
> +		echo >&2 "error: ${TOOL_MODE}tool.$tool.cmd not set for tool '$tool'"
> +		return 1
> +	fi

There are only two callers of setup_user_tool, and one of them
squelches this message by sending it to /dev/null.  The error
message composed here does not use anything that is unique to the
function (in other words, $tool and ${TOOL_MODE} are available to
its callers).

I wonder if it is a better design to leave this one as-is, and
instead show the error message from the other caller of
setup_user_tool that does not squelch the message?  Are we planning
to add more callers of this function that want to show the same
message?

>  	diff_cmd () {
>  		( eval $merge_tool_cmd )
> @@ -255,7 +259,7 @@ setup_tool () {
>  
>  	# Now let the user override the default command for the tool.  If
>  	# they have not done so then this will return 1 which we ignore.
> -	setup_user_tool
> +	setup_user_tool 2>/dev/null

If we did that, then this change can be dropped.  Instead, a few
lines above this hunk, we can give the error message ourselves from
this setup_tool function.

>  	if ! list_tool_variants | grep -q "^$tool$"
>  	then
> diff --git a/t/t7610-mergetool.sh b/t/t7610-mergetool.sh
> index 22b3a85b3e9..82a88107850 100755
> --- a/t/t7610-mergetool.sh
> +++ b/t/t7610-mergetool.sh
> @@ -898,4 +898,12 @@ test_expect_success 'mergetool with guiDefault' '
>  	git commit -m "branch1 resolved with mergetool"
>  '
>  
> +test_expect_success 'mergetool with non-existent tool' '
> +	test_when_finished "git reset --hard" &&
> +	git checkout -b test$test_count branch1 &&
> +	test_must_fail git merge main &&
> +	yes "" | test_must_fail git mergetool --tool=absent >out 2>&1 &&
> +	test_grep -i "not set for tool" out
> +'

Why "-i"?  I do not offhand see the reason why we want to be loose
here.

The "${TOOL_MODE}tool" part may also want to be verified, perhaps,
which was related to the topic of the fix in [2/5]?

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Philippe Blain wrote (reply to this):

Hi Junio,

Le 2024-11-12 à 20:48, Junio C Hamano a écrit :
> "Philippe Blain via GitGitGadget" <[email protected]> writes:
> 
>>  setup_user_tool () {
>>  	merge_tool_cmd=$(get_merge_tool_cmd "$tool")
>> -	test -n "$merge_tool_cmd" || return 1
>> +	if test -z "$merge_tool_cmd"
>> +	then
>> +		echo >&2 "error: ${TOOL_MODE}tool.$tool.cmd not set for tool '$tool'"
>> +		return 1
>> +	fi
> 
> There are only two callers of setup_user_tool, and one of them
> squelches this message by sending it to /dev/null.  The error
> message composed here does not use anything that is unique to the
> function (in other words, $tool and ${TOOL_MODE} are available to
> its callers).
> 
> I wonder if it is a better design to leave this one as-is, and
> instead show the error message from the other caller of
> setup_user_tool that does not squelch the message?  Are we planning
> to add more callers of this function that want to show the same
> message?

I don't think we are planning to add more callers, no.

> 
>>  	diff_cmd () {
>>  		( eval $merge_tool_cmd )
>> @@ -255,7 +259,7 @@ setup_tool () {
>>  
>>  	# Now let the user override the default command for the tool.  If
>>  	# they have not done so then this will return 1 which we ignore.
>> -	setup_user_tool
>> +	setup_user_tool 2>/dev/null
> 
> If we did that, then this change can be dropped.  Instead, a few
> lines above this hunk, we can give the error message ourselves from
> this setup_tool function.

I agree it could be done this way, I can change if it we wish.


>>  	if ! list_tool_variants | grep -q "^$tool$"
>>  	then
>> diff --git a/t/t7610-mergetool.sh b/t/t7610-mergetool.sh
>> index 22b3a85b3e9..82a88107850 100755
>> --- a/t/t7610-mergetool.sh
>> +++ b/t/t7610-mergetool.sh
>> @@ -898,4 +898,12 @@ test_expect_success 'mergetool with guiDefault' '
>>  	git commit -m "branch1 resolved with mergetool"
>>  '
>>  
>> +test_expect_success 'mergetool with non-existent tool' '
>> +	test_when_finished "git reset --hard" &&
>> +	git checkout -b test$test_count branch1 &&
>> +	test_must_fail git merge main &&
>> +	yes "" | test_must_fail git mergetool --tool=absent >out 2>&1 &&
>> +	test_grep -i "not set for tool" out
>> +'
> 
> Why "-i"?  I do not offhand see the reason why we want to be loose
> here.

Indeed this is a leftover from my bisection test in which I had to 
be a bit more loose. I'll remove that flag.
 
> The "${TOOL_MODE}tool" part may also want to be verified, perhaps,
> which was related to the topic of the fix in [2/5]?

Yes, I guess I could make the pattern stricter. I'll update that.

@@ -255,10 +259,11 @@ setup_tool () {

# Now let the user override the default command for the tool. If
# they have not done so then this will return 1 which we ignore.
setup_user_tool
setup_user_tool 2>/dev/null

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Junio C Hamano wrote (reply to this):

"Philippe Blain via GitGitGadget" <[email protected]> writes:

> From: Philippe Blain <[email protected]>
>
> In setup_tool, we check if the given tool is a known variant of a tool,
> and quietly return with an error if not. This leads to the following
> invocation quietly failing:
>
> 	git mergetool --tool=vimdiff4
>
> Add an error message before returning in this case.

Makes sense, but ...

> Signed-off-by: Philippe Blain <[email protected]>
> ---
>  git-mergetool--lib.sh | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/git-mergetool--lib.sh b/git-mergetool--lib.sh
> index f4786afc63f..9a00fabba27 100644
> --- a/git-mergetool--lib.sh
> +++ b/git-mergetool--lib.sh
> @@ -263,6 +263,7 @@ setup_tool () {
>  
>  	if ! list_tool_variants | grep -q "^$tool$"
>  	then
> +		echo "error: unknown ${tool%[0-9]} variant '$tool'" >&2

... I do not understand why you strip a single digit from the end.

    git mergetool --tool=nvimdiff4

says 'nvimdiff4' is not known as a variant of 'nvimdiff', but
wouldn't it still be a variant of 'vimdiff'?  Of course,

    git mergetool --tool=nvimdiff48

gets a vastly different error message ;-)

Saying

	echo >&2 "error: unknown variant '$tool'"

may be sufficient, perhaps?  I dunno.


>  		return 1
>  	fi

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On the Git mailing list, Philippe Blain wrote (reply to this):

Le 2024-11-12 à 21:01, Junio C Hamano a écrit :
> "Philippe Blain via GitGitGadget" <[email protected]> writes:
> 
>> From: Philippe Blain <[email protected]>
>>
>> In setup_tool, we check if the given tool is a known variant of a tool,
>> and quietly return with an error if not. This leads to the following
>> invocation quietly failing:
>>
>> 	git mergetool --tool=vimdiff4
>>
>> Add an error message before returning in this case.
> 
> Makes sense, but ...
> 
>> Signed-off-by: Philippe Blain <[email protected]>
>> ---
>>  git-mergetool--lib.sh | 1 +
>>  1 file changed, 1 insertion(+)
>>
>> diff --git a/git-mergetool--lib.sh b/git-mergetool--lib.sh
>> index f4786afc63f..9a00fabba27 100644
>> --- a/git-mergetool--lib.sh
>> +++ b/git-mergetool--lib.sh
>> @@ -263,6 +263,7 @@ setup_tool () {
>>  
>>  	if ! list_tool_variants | grep -q "^$tool$"
>>  	then
>> +		echo "error: unknown ${tool%[0-9]} variant '$tool'" >&2
> 
> ... I do not understand why you strip a single digit from the end.
> 
>     git mergetool --tool=nvimdiff4
> 
> says 'nvimdiff4' is not known as a variant of 'nvimdiff', but
> wouldn't it still be a variant of 'vimdiff'?  Of course,
> 
>     git mergetool --tool=nvimdiff48
> 
> gets a vastly different error message ;-)
> 
> Saying
> 
> 	echo >&2 "error: unknown variant '$tool'"
> 
> may be sufficient, perhaps?  I dunno.

the stripping of the last digit is just because I copied from 
the 'if' a few lines above, where we source "$MERGE_TOOLS_DIR/${tool%[0-9]}".
In MERGE_TOOLS_DIR we have 'nvimdiff' and 'gvimdiff' that simply source vimdiff,
so this works. But I agree that we can simplify the error message, I'll do that.

Copy link

gitgitgadget bot commented Nov 13, 2024

This patch series was integrated into seen via git@a35aba8.

@gitgitgadget gitgitgadget bot added the seen label Nov 13, 2024
@dscho
Copy link
Member

dscho commented Nov 13, 2024

My first two tries (one, two) at /submit failed because (if I understand correctly) the node.js v18.20.5 released today was not yet up at https://nodejs.org/dist/v18.20.5/, although the folder existed already. I guess this is just bad luck and there is nothing much Gitgitgadget can do ? And changes would need to go into the NodeTool Azure task to make it more robust ?

@phil-blain sorry for the woes! Your analysis is indeed correct.

I now changed the Pipeline definition to use v1 of that task (instead of v0), and I checked off the "Check for Latest Version" checkbox. That should be good enough.

Copy link

gitgitgadget bot commented Nov 15, 2024

This patch series was integrated into seen via git@37181c6.

Copy link

gitgitgadget bot commented Nov 16, 2024

This patch series was integrated into seen via git@5006b1c.

Copy link

gitgitgadget bot commented Nov 18, 2024

This patch series was integrated into seen via git@6d9c2d4.

Copy link

gitgitgadget bot commented Nov 19, 2024

This patch series was integrated into seen via git@011d5cd.

Copy link

gitgitgadget bot commented Nov 20, 2024

This patch series was integrated into seen via git@ad8e306.

Copy link

gitgitgadget bot commented Nov 21, 2024

This patch series was integrated into seen via git@208a0e2.

Copy link

gitgitgadget bot commented Nov 22, 2024

This patch series was integrated into seen via git@67867a8.

Copy link

gitgitgadget bot commented Nov 22, 2024

This patch series was integrated into seen via git@4ff66cb.

Copy link

gitgitgadget bot commented Nov 22, 2024

User Philippe Blain <[email protected]> has been added to the cc: list.

In git-mergetool--lib.sh::setup_tool, we check if the given tool is a
known builtin tool, a known variant, or a user-defined tool by calling
setup_user_tool, and we return with the exit code from setup_user_tool
if it was called. setup_user_tool checks if {diff,merge}tool.$tool.cmd
is set and quietly returns with an error if not.

This leads to the following invocation quietly failing:

	git mergetool --tool=unknown

which is not very user-friendly. Adjust setup_tool to output an error
message before returning if setup_user_tool returned with an error.

Note that we do not check the result of the second call to
setup_user_tool in setup_tool, as this call is only meant to allow users
to redefine 'cmd' for a builtin tool; it is not an error if they have
not done so.

Note that this behaviour of quietly failing is a regression dating back
to de8dafb (mergetool: break setup_tool out into separate
initialization function, 2021-02-09), as before this commit an unknown
mergetool would be diagnosed in get_merge_tool_path when called from
run_merge_tool.

Signed-off-by: Philippe Blain <[email protected]>
In setup_tool, we check if the given tool is a known variant of a tool,
and quietly return with an error if not. This leads to the following
invocation quietly failing:

	git mergetool --tool=vimdiff4

Add an error message before returning in this case.

Signed-off-by: Philippe Blain <[email protected]>
Since the introduction of 'initialize_merge_tool' in de8dafb
(mergetool: break setup_tool out into separate initialization function,
2021-02-09), any errors from this function are ignored in
git-difftool--helper.sh::launch_merge_tool, which is not the case for
its call in git-mergetool.sh::merge_file.

Despite the in-code comment, initialize_merge_tool (via its call to
setup_tool) does different checks than run_merge_tool, so it makes sense
to abort early if it encounters errors. Add exit calls if
initialize_merge_tool fails.

Signed-off-by: Philippe Blain <[email protected]>
@phil-blain
Copy link
Author

/submit

Copy link

gitgitgadget bot commented Nov 22, 2024

Submitted as [email protected]

To fetch this version into FETCH_HEAD:

git fetch https://github.com/gitgitgadget/git/ pr-1827/phil-blain/absent-mergetool-v2

To fetch this version to local tag pr-1827/phil-blain/absent-mergetool-v2:

git fetch --no-tags https://github.com/gitgitgadget/git/ tag pr-1827/phil-blain/absent-mergetool-v2

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants