threads / patch / 58418

patch, 2 partsblame: Document that --show-name is negatable

Subject: [PATCH 2/2] blame: Document that --show-name is negatable

## tl;dr

7 messages between Sep 12, 2022 and Sep 13, 2022. Diffs are folded; open one to read it.

replies: 6people: 3as markdown or json

Rene Kita· Sep 12, 2022, 05:45 UTC · lore

[PATCH 0/2] blame: Make --show-name negatable

With -s git blame does not show author and date, but will show the file name if the file was renamed some time in the past. With -c git blame will not show the name, even if the file was renamed, but will show author and date.

Working with a code base that already uses long lines and has a deep nested directory structure the output of git blame gets hard to read due to line wrapping.

While looking for a solution on IRC a user was so nice to gave me a patch (and permission to submit it) that allows to disable showing of the file name even if the file was renamed. This allows to show only the commit id with `git blame -s --no-show-name`.

The second patch only adds this informations to the man page.
Rene Kita (1):
  blame: Document that --show-name is negatable
Øystein Walle (1):
  blame: Make --show-name negatable
 Documentation/git-blame.txt | 2 +-
 builtin/blame.c             | 9 +++++++--
 2 files changed, 8 insertions(+), 3 deletions(-)
-- 
2.37.3.544.g5c9b9c0a4e
Rene Kita· Sep 12, 2022, 05:45 UTC · re: Rene Kita · lore
Signed-off-by: Rene Kita <mail@rkta.de>
---
 Documentation/git-blame.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Show changes to Documentation/git-blame.txt +1 −0
diff --git a/Documentation/git-blame.txt b/Documentation/git-blame.txt
index d7a46cc674..75e115f97e 100644
--- a/Documentation/git-blame.txt
+++ b/Documentation/git-blame.txt
@@ -63,7 +63,7 @@ include::blame-options.txt[]
 	of code to have been moved.
 
 -f::
---show-name::
+--[no-]show-name::
 	Show the filename in the original commit.  By default
 	the filename is shown if there is any line that came from a
 	file with a different name, due to rename detection.
-- 
2.37.3.544.g5c9b9c0a4e
Rene Kita· Sep 12, 2022, 05:45 UTC · re: Rene Kita · lore

[PATCH 1/2] blame: Make --show-name negatable

From: Øystein Walle <oystwa@gmail.com>
Signed-off-by: Øystein Walle <oystwa@gmail.com>
---
 builtin/blame.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
Show changes to builtin/blame.c +7 −2
diff --git a/builtin/blame.c b/builtin/blame.c
index a9fe8cf7a6..cbaae91a8f 100644
--- a/builtin/blame.c
+++ b/builtin/blame.c
@@ -49,6 +49,7 @@ static int blank_boundary;
 static int incremental;
 static int xdl_opts;
 static int abbrev = -1;
+static int show_name = -1;
 static int no_whole_file_rename;
 static int show_progress;
 static char repeated_meta_color[COLOR_MAXLEN];
@@ -621,7 +622,8 @@ static void find_alignment(struct blame_scoreboard *sb, int *option)
 		if (compute_auto_abbrev)
 			auto_abbrev = update_auto_abbrev(auto_abbrev, suspect);
 		if (strcmp(suspect->path, sb->path))
-			*option |= OUTPUT_SHOW_NAME;
+			if (show_name == -1)
+				*option |= OUTPUT_SHOW_NAME;
 		num = strlen(suspect->path);
 		if (longest_file < num)
 			longest_file = num;
@@ -867,7 +869,7 @@ int cmd_blame(int argc, const char **argv, const char *prefix)
 		OPT_BOOL(0, "show-stats", &show_stats, N_("show work cost statistics")),
 		OPT_BOOL(0, "progress", &show_progress, N_("force progress reporting")),
 		OPT_BIT(0, "score-debug", &output_option, N_("show output score for blame entries"), OUTPUT_SHOW_SCORE),
-		OPT_BIT('f', "show-name", &output_option, N_("show original filename (Default: auto)"), OUTPUT_SHOW_NAME),
+		OPT_BOOL('f', "show-name", &show_name, N_("show original filename (Default: auto)")),
 		OPT_BIT('n', "show-number", &output_option, N_("show original linenumber (Default: off)"), OUTPUT_SHOW_NUMBER),
 		OPT_BIT('p', "porcelain", &output_option, N_("show in a format designed for machine consumption"), OUTPUT_PORCELAIN),
 		OPT_BIT(0, "line-porcelain", &output_option, N_("show porcelain format with per-line commit information"), OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN),
@@ -943,6 +945,9 @@ int cmd_blame(int argc, const char **argv, const char *prefix)
 	revs.diffopt.flags.follow_renames = 0;
 	argc = parse_options_end(&ctx);
 
+	if (show_name == 1)
+		output_option |= OUTPUT_SHOW_NAME;
+
 	prepare_repo_settings(the_repository);
 	the_repository->settings.command_requires_full_index = 0;
 
-- 
2.37.3.544.g5c9b9c0a4e
Øystein Walle· Sep 12, 2022, 13:25 UTC · re: Rene Kita · lore

Re: [PATCH 1/2] blame: Make --show-name negatable

That was me! :-)

Not long after I sent you a new link with the patch below, but I suppose you missed it. My first rough attempt was incomplete, changing the behavior from on/auto on to auto/off instead of the intended on/auto/off.

The docs changes look good to me (but should perhaps be squashed into the impl patch?). Not sure how many of the negatable long options that exist are actually documented as such, but imho. they should be.

There are no added tests, but there are no tests for -f/--show-name in the first place and I just didn't get around to writing one...

Øsse
---
 builtin/blame.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
Show changes to builtin/blame.c +7 −2
diff --git a/builtin/blame.c b/builtin/blame.c
index a9fe8cf7a6..cbaae91a8f 100644
--- a/builtin/blame.c
+++ b/builtin/blame.c
@@ -49,6 +49,7 @@ static int blank_boundary;
 static int incremental;
 static int xdl_opts;
 static int abbrev = -1;
+static int show_name = -1;
 static int no_whole_file_rename;
 static int show_progress;
 static char repeated_meta_color[COLOR_MAXLEN];
@@ -621,7 +622,8 @@ static void find_alignment(struct blame_scoreboard
*sb, int *option)
         if (compute_auto_abbrev)
             auto_abbrev = update_auto_abbrev(auto_abbrev, suspect);
         if (strcmp(suspect->path, sb->path))
-            *option |= OUTPUT_SHOW_NAME;
+            if (show_name == -1)
+                *option |= OUTPUT_SHOW_NAME;
         num = strlen(suspect->path);
         if (longest_file < num)
             longest_file = num;
@@ -867,7 +869,7 @@ int cmd_blame(int argc, const char **argv, const
char *prefix)
         OPT_BOOL(0, "show-stats", &show_stats, N_("show work cost
statistics")),
         OPT_BOOL(0, "progress", &show_progress, N_("force progress
reporting")),
         OPT_BIT(0, "score-debug", &output_option, N_("show output
score for blame entries"), OUTPUT_SHOW_SCORE),
-        OPT_BIT('f', "show-name", &output_option, N_("show original
filename (Default: auto)"), OUTPUT_SHOW_NAME),
+        OPT_BOOL('f', "show-name", &show_name, N_("show original
filename (Default: auto)")),
         OPT_BIT('n', "show-number", &output_option, N_("show original
linenumber (Default: off)"), OUTPUT_SHOW_NUMBER),
         OPT_BIT('p', "porcelain", &output_option, N_("show in a
format designed for machine consumption"), OUTPUT_PORCELAIN),
         OPT_BIT(0, "line-porcelain", &output_option, N_("show
porcelain format with per-line commit information"),
OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN),
@@ -943,6 +945,9 @@ int cmd_blame(int argc, const char **argv, const
char *prefix)
     revs.diffopt.flags.follow_renames = 0;
     argc = parse_options_end(&ctx);

+    if (show_name == 1)
+        output_option |= OUTPUT_SHOW_NAME;
+
     prepare_repo_settings(the_repository);
     the_repository->settings.command_requires_full_index = 0;
-- 
2.34.1
Junio C Hamano· Sep 12, 2022, 16:30 UTC · re: Øystein Walle · lore

Re: [PATCH 1/2] blame: Make --show-name negatable

Øystein Walle <oystwa@gmail.com> writes:
Show 6 quoted lines
> That was me! :-)
>
> Not long after I sent you a new link with the patch below, but I suppose
> you missed it. My first rough attempt was incomplete, changing the
> behavior from on/auto on to auto/off instead of the intended
> on/auto/off.

Not clear to this bystander what exactly is going on. Is the patch Rene relayed is your rough attempt that was incomplete?

Show 6 quoted lines
> The docs changes look good to me (but should perhaps be squashed into the impl
> patch?). Not sure how many of the negatable long options that exist are
> actually documented as such, but imho. they should be.
>
> There are no added tests, but there are no tests for -f/--show-name in the
> first place and I just didn't get around to writing one...

Thanks. Let's see if we can have an cleaned-up version with tests, then.

Show 59 quoted lines
>
> Øsse
>
> ---
>  builtin/blame.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/builtin/blame.c b/builtin/blame.c
> index a9fe8cf7a6..cbaae91a8f 100644
> --- a/builtin/blame.c
> +++ b/builtin/blame.c
> @@ -49,6 +49,7 @@ static int blank_boundary;
>  static int incremental;
>  static int xdl_opts;
>  static int abbrev = -1;
> +static int show_name = -1;
>  static int no_whole_file_rename;
>  static int show_progress;
>  static char repeated_meta_color[COLOR_MAXLEN];
> @@ -621,7 +622,8 @@ static void find_alignment(struct blame_scoreboard
> *sb, int *option)
>          if (compute_auto_abbrev)
>              auto_abbrev = update_auto_abbrev(auto_abbrev, suspect);
>          if (strcmp(suspect->path, sb->path))
> -            *option |= OUTPUT_SHOW_NAME;
> +            if (show_name == -1)
> +                *option |= OUTPUT_SHOW_NAME;
>          num = strlen(suspect->path);
>          if (longest_file < num)
>              longest_file = num;
> @@ -867,7 +869,7 @@ int cmd_blame(int argc, const char **argv, const
> char *prefix)
>          OPT_BOOL(0, "show-stats", &show_stats, N_("show work cost
> statistics")),
>          OPT_BOOL(0, "progress", &show_progress, N_("force progress
> reporting")),
>          OPT_BIT(0, "score-debug", &output_option, N_("show output
> score for blame entries"), OUTPUT_SHOW_SCORE),
> -        OPT_BIT('f', "show-name", &output_option, N_("show original
> filename (Default: auto)"), OUTPUT_SHOW_NAME),
> +        OPT_BOOL('f', "show-name", &show_name, N_("show original
> filename (Default: auto)")),
>          OPT_BIT('n', "show-number", &output_option, N_("show original
> linenumber (Default: off)"), OUTPUT_SHOW_NUMBER),
>          OPT_BIT('p', "porcelain", &output_option, N_("show in a
> format designed for machine consumption"), OUTPUT_PORCELAIN),
>          OPT_BIT(0, "line-porcelain", &output_option, N_("show
> porcelain format with per-line commit information"),
> OUTPUT_PORCELAIN|OUTPUT_LINE_PORCELAIN),
> @@ -943,6 +945,9 @@ int cmd_blame(int argc, const char **argv, const
> char *prefix)
>      revs.diffopt.flags.follow_renames = 0;
>      argc = parse_options_end(&ctx);
>
> +    if (show_name == 1)
> +        output_option |= OUTPUT_SHOW_NAME;
> +
>      prepare_repo_settings(the_repository);
>      the_repository->settings.command_requires_full_index = 0;
Øystein Walle· Sep 13, 2022, 08:56 UTC · re: Junio C Hamano · lore

Re: [PATCH 1/2] blame: Make --show-name negatable

Hi, Junio
> Not clear to this bystander what exactly is going on.  Is the patch
> Rene relayed is your rough attempt that was incomplete?

Correct! I am the nice IRC that was mentioned in the cover letter. My improved patch is the one in the e-mail you responded to, apart from GMail's text wrapping that was unaware of until now.

> Thanks.  Let's see if we can have an cleaned-up version with tests,
> then.

I can send a corrected (and correctly formatted patch) for review, optionally with the docs squashed in with Rene's permission. Maybe Rene would like to take a stab at writing tests? I am in the process of moving house and my usual workstation is not exactly online, so it'll be a while.

Øsse
Rene Kita· Sep 13, 2022, 09:08 UTC · re: Øystein Walle · lore

Re: [PATCH 1/2] blame: Make --show-name negatable

On Tue, Sep 13, 2022 at 10:56:31AM +0200, Øystein Walle wrote:
Show 8 quoted lines
> Hi, Junio
> 
> > Not clear to this bystander what exactly is going on.  Is the patch
> > Rene relayed is your rough attempt that was incomplete?
> 
> Correct! I am the nice IRC that was mentioned in the cover letter. My
> improved patch is the one in the e-mail you responded to, apart from
> GMail's text wrapping that was unaware of until now.

That's my fault. I saw the corrected patch, but somehow used the old one. Sorry!

Show 5 quoted lines
> > Thanks.  Let's see if we can have an cleaned-up version with tests,
> > then.
> 
> I can send a corrected (and correctly formatted patch) for review,
> optionally with the docs squashed in with Rene's permission.
Of course.
> Maybe Rene would like to take a stab at writing tests? I am in the
> process of moving house and my usual workstation is not exactly
> online, so it'll be a while.

Yes, I can write a test. Will need to investigate, never worked with the code base.

← back to recent threads