git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff
@ 2021-03-07  6:57 ZheNing Hu via GitGitGadget
  2021-03-08 20:18 ` Taylor Blau
  2021-03-09  8:33 ` [PATCH v2] [GSOC][RFC] format-patch: pass --left-only " ZheNing Hu via GitGitGadget
  0 siblings, 2 replies; 16+ messages in thread
From: ZheNing Hu via GitGitGadget @ 2021-03-07  6:57 UTC (permalink / raw)
  To: git
  Cc: Johannes Schindelin, Denton Liu, Junio C Hamano, Taylor Blau,
	ZheNing Hu, ZheNing Hu

From: ZheNing Hu <adlternative@gmail.com>

In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
Taylor Blau proposing 'git format-patch --cover-letter
--range-diff' may mistakenly place upstream commit in the
range-diff output. Teach `format-patch` pass `--right-only`
to range-diff, maybe can avoid this kind of mistake.

Because `git rebase --apply` will internally call `git
format-patch -k --stdout --full-index --cherry-pick --right-only
--src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter
--pretty=mboxrd --topo-order --no-base`, In order to avoid the
ambiguity of `--right-only`, users can call `format-patch
--range-diff --only-right` instead of `--right-only` to only emit
output related to the second range.

Signed-off-by: ZheNing Hu <adlternative@gmail.com>
---
    [GSOC][RFC] format-patch: pass --right-only to range-diff
    
    Transfer parameters are relatively simple. But because I can' t
    reproduce the situation, There's no big difference between adding
    ---only-right and not adding ---only-right. So I haven't written the
    test files for the time being.
    
    I may need to ask reviewers opinion first.
    
    this want to fix #876 Thanks.

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-898%2Fadlternative%2Fformat-patch-range-diff-right-only-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-898/adlternative/format-patch-range-diff-right-only-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/898

 Documentation/git-format-patch.txt |  5 ++++-
 builtin/log.c                      | 15 ++++++++++-----
 2 files changed, 14 insertions(+), 6 deletions(-)

diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt
index 3e49bf221087..5dae34c3090b 100644
--- a/Documentation/git-format-patch.txt
+++ b/Documentation/git-format-patch.txt
@@ -27,7 +27,7 @@ SYNOPSIS
 		   [--[no-]encode-email-headers]
 		   [--no-notes | --notes[=<ref>]]
 		   [--interdiff=<previous>]
-		   [--range-diff=<previous> [--creation-factor=<percent>]]
+		   [--range-diff=<previous> [--creation-factor=<percent>] [--right-only]]
 		   [--filename-max-length=<n>]
 		   [--progress]
 		   [<common diff options>]
@@ -301,6 +301,9 @@ material (this may change in the future).
 	creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
 	for details.
 
+--only-right:
+	Used with `--range-diff`, only emit output related to the second range.
+
 --notes[=<ref>]::
 --no-notes::
 	Append the notes (see linkgit:git-notes[1]) for the commit
diff --git a/builtin/log.c b/builtin/log.c
index f67b67d80ed1..5d2f39fd19a7 100644
--- a/builtin/log.c
+++ b/builtin/log.c
@@ -1153,7 +1153,7 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			      struct commit *origin,
 			      int nr, struct commit **list,
 			      const char *branch_name,
-			      int quiet)
+			      int quiet, int right_only)
 {
 	const char *committer;
 	struct shortlog log;
@@ -1228,7 +1228,8 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			.creation_factor = rev->creation_factor,
 			.dual_color = 1,
 			.diffopt = &opts,
-			.other_arg = &other_arg
+			.other_arg = &other_arg,
+			.right_only = right_only
 		};
 
 		diff_setup(&opts);
@@ -1732,7 +1733,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 	struct strbuf rdiff2 = STRBUF_INIT;
 	struct strbuf rdiff_title = STRBUF_INIT;
 	int creation_factor = -1;
-
+	int right_only = 0;
 	const struct option builtin_format_patch_options[] = {
 		OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
 			    N_("use [PATCH n/m] even with a single patch"),
@@ -1814,6 +1815,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 			     parse_opt_object_name),
 		OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
 			   N_("show changes against <refspec> in cover letter or single patch")),
+		OPT_BOOL(0, "only-right", &right_only,
+			 N_("only emit output related to the second range")),
 		OPT_INTEGER(0, "creation-factor", &creation_factor,
 			    N_("percentage by which creation is weighted")),
 		OPT_END()
@@ -2087,7 +2090,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
 	else if (!rdiff_prev)
 		die(_("--creation-factor requires --range-diff"));
-
+	if (right_only && !rdiff_prev)
+		die(_("--right_only requires --range-diff"));
 	if (rdiff_prev) {
 		if (!cover_letter && total != 1)
 			die(_("--range-diff requires --cover-letter or single patch"));
@@ -2134,7 +2138,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 		if (thread)
 			gen_message_id(&rev, "cover");
 		make_cover_letter(&rev, !!output_directory,
-				  origin, nr, list, branch_name, quiet);
+				  origin, nr, list, branch_name, quiet,
+					right_only);
 		print_bases(&bases, rev.diffopt.file);
 		print_signature(rev.diffopt.file);
 		total++;

base-commit: be7935ed8bff19f481b033d0d242c5d5f239ed50
-- 
gitgitgadget

^ permalink raw reply related	[flat|nested] 16+ messages in thread

* Re: [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff
  2021-03-07  6:57 [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff ZheNing Hu via GitGitGadget
@ 2021-03-08 20:18 ` Taylor Blau
  2021-03-09  7:28   ` ZheNing Hu
  2021-03-12 21:55   ` Junio C Hamano
  2021-03-09  8:33 ` [PATCH v2] [GSOC][RFC] format-patch: pass --left-only " ZheNing Hu via GitGitGadget
  1 sibling, 2 replies; 16+ messages in thread
From: Taylor Blau @ 2021-03-08 20:18 UTC (permalink / raw)
  To: ZheNing Hu via GitGitGadget
  Cc: git, Johannes Schindelin, Denton Liu, Junio C Hamano,
	Taylor Blau, ZheNing Hu

On Sun, Mar 07, 2021 at 06:57:19AM +0000, ZheNing Hu via GitGitGadget wrote:
> From: ZheNing Hu <adlternative@gmail.com>
>
> In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
> Taylor Blau proposing 'git format-patch --cover-letter
> --range-diff' may mistakenly place upstream commit in the
> range-diff output. Teach `format-patch` pass `--right-only`
> to range-diff, maybe can avoid this kind of mistake.
>
> Because `git rebase --apply` will internally call `git
> format-patch -k --stdout --full-index --cherry-pick --right-only
> --src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter
> --pretty=mboxrd --topo-order --no-base`, In order to avoid the
> ambiguity of `--right-only`, users can call `format-patch
> --range-diff --only-right` instead of `--right-only` to only emit
> output related to the second range.

A couple of times this new option is called "--right-only" and other
times it is "--only-right". This comment makes me think that you meant
to call it "--only-right", but...

> Signed-off-by: ZheNing Hu <adlternative@gmail.com>
> ---
>     [GSOC][RFC] format-patch: pass --right-only to range-diff
>
>     Transfer parameters are relatively simple. But because I can' t
>     reproduce the situation, There's no big difference between adding
>     ---only-right and not adding ---only-right. So I haven't written the
>     test files for the time being.

Note that I think (and Johannes--cc'd--could confirm) that what you want
is '--left-only' to discard rebased changes from the upstream branch:

In either case, this should help you reproduce it (and would be a good
starting point for adding a test case). Here, we're using '--left-only'
to discard changes from upstream (here, the 'main' branch) after
rebasing:

    #!/bin/sh

    set -e

    rm -fr repo
    git init repo
    cd repo

    git branch -M main

    echo "base" >base
    git add base
    git commit -m "base"

    git checkout -b my-feature
    echo "feature" >feature
    git add feature
    git commit -m "feature"

    base="$(git rev-parse main)"
    old="$(git rev-parse my-feature)"

    git checkout main
    echo "other" >>base
    git add base
    git commit -m "new"

    git checkout my-feature
    git rebase $base --onto main

    tip="$(git rev-parse my-feature)"

    git range-diff $base $old $tip
    git range-diff --left-only $base $old $tip

>     I may need to ask reviewers opinion first.
>
>     this want to fix #876 Thanks.
>
> Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-898%2Fadlternative%2Fformat-patch-range-diff-right-only-v1
> Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-898/adlternative/format-patch-range-diff-right-only-v1
> Pull-Request: https://github.com/gitgitgadget/git/pull/898
>
>  Documentation/git-format-patch.txt |  5 ++++-
>  builtin/log.c                      | 15 ++++++++++-----
>  2 files changed, 14 insertions(+), 6 deletions(-)
>
> diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt
> index 3e49bf221087..5dae34c3090b 100644
> --- a/Documentation/git-format-patch.txt
> +++ b/Documentation/git-format-patch.txt
> @@ -27,7 +27,7 @@ SYNOPSIS
>  		   [--[no-]encode-email-headers]
>  		   [--no-notes | --notes[=<ref>]]
>  		   [--interdiff=<previous>]
> -		   [--range-diff=<previous> [--creation-factor=<percent>]]
> +		   [--range-diff=<previous> [--creation-factor=<percent>] [--right-only]]
>  		   [--filename-max-length=<n>]
>  		   [--progress]
>  		   [<common diff options>]
> @@ -301,6 +301,9 @@ material (this may change in the future).
>  	creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
>  	for details.
>
> +--only-right:
> +	Used with `--range-diff`, only emit output related to the second range.
> +

s/--only-right/--right-only?

>  --notes[=<ref>]::
>  --no-notes::
>  	Append the notes (see linkgit:git-notes[1]) for the commit
> diff --git a/builtin/log.c b/builtin/log.c
> index f67b67d80ed1..5d2f39fd19a7 100644
> --- a/builtin/log.c
> +++ b/builtin/log.c
> @@ -1153,7 +1153,7 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
>  			      struct commit *origin,
>  			      int nr, struct commit **list,
>  			      const char *branch_name,
> -			      int quiet)
> +			      int quiet, int right_only)
>  {
>  	const char *committer;
>  	struct shortlog log;
> @@ -1228,7 +1228,8 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
>  			.creation_factor = rev->creation_factor,
>  			.dual_color = 1,
>  			.diffopt = &opts,
> -			.other_arg = &other_arg
> +			.other_arg = &other_arg,
> +			.right_only = right_only
>  		};
>
>  		diff_setup(&opts);
> @@ -1732,7 +1733,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
>  	struct strbuf rdiff2 = STRBUF_INIT;
>  	struct strbuf rdiff_title = STRBUF_INIT;
>  	int creation_factor = -1;
> -
> +	int right_only = 0;

Nit: the newline between creation_factor's declaration and
builtin_format_patch_options's is good to keep.

>  	const struct option builtin_format_patch_options[] = {
>  		OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
>  			    N_("use [PATCH n/m] even with a single patch"),
> @@ -1814,6 +1815,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
>  			     parse_opt_object_name),
>  		OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
>  			   N_("show changes against <refspec> in cover letter or single patch")),
> +		OPT_BOOL(0, "only-right", &right_only,
> +			 N_("only emit output related to the second range")),
>  		OPT_INTEGER(0, "creation-factor", &creation_factor,
>  			    N_("percentage by which creation is weighted")),
>  		OPT_END()
> @@ -2087,7 +2090,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
>  		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
>  	else if (!rdiff_prev)
>  		die(_("--creation-factor requires --range-diff"));
> -
> +	if (right_only && !rdiff_prev)
> +		die(_("--right_only requires --range-diff"));

These may be good to combine into one big conditional that dies whenever
an option requiring --range-diff is set, but --range-diff is not:

  if (!rdiff_prev) {
    if (creation_factor)
      die(_("--creation-factor requires --range-diff"));
    if (right_only)
      die(_("--right-only requires --range-diff"));
  }

Thanks,
Taylor

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff
  2021-03-08 20:18 ` Taylor Blau
@ 2021-03-09  7:28   ` ZheNing Hu
  2021-03-12 21:55   ` Junio C Hamano
  1 sibling, 0 replies; 16+ messages in thread
From: ZheNing Hu @ 2021-03-09  7:28 UTC (permalink / raw)
  To: Taylor Blau
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Junio C Hamano, Taylor Blau

Taylor Blau <ttaylorr@github.com> 于2021年3月9日周二 上午4:18写道:
>
> On Sun, Mar 07, 2021 at 06:57:19AM +0000, ZheNing Hu via GitGitGadget wrote:
> > From: ZheNing Hu <adlternative@gmail.com>
> >
> > In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
> > Taylor Blau proposing 'git format-patch --cover-letter
> > --range-diff' may mistakenly place upstream commit in the
> > range-diff output. Teach `format-patch` pass `--right-only`
> > to range-diff, maybe can avoid this kind of mistake.
> >
> > Because `git rebase --apply` will internally call `git
> > format-patch -k --stdout --full-index --cherry-pick --right-only
> > --src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter
> > --pretty=mboxrd --topo-order --no-base`, In order to avoid the
> > ambiguity of `--right-only`, users can call `format-patch
> > --range-diff --only-right` instead of `--right-only` to only emit
> > output related to the second range.
>
> A couple of times this new option is called "--right-only" and other
> times it is "--only-right". This comment makes me think that you meant
> to call it "--only-right", but...
>
> > Signed-off-by: ZheNing Hu <adlternative@gmail.com>
> > ---
> >     [GSOC][RFC] format-patch: pass --right-only to range-diff
> >
> >     Transfer parameters are relatively simple. But because I can' t
> >     reproduce the situation, There's no big difference between adding
> >     ---only-right and not adding ---only-right. So I haven't written the
> >     test files for the time being.
>
> Note that I think (and Johannes--cc'd--could confirm) that what you want
> is '--left-only' to discard rebased changes from the upstream branch:
>
> In either case, this should help you reproduce it (and would be a good
> starting point for adding a test case). Here, we're using '--left-only'
> to discard changes from upstream (here, the 'main' branch) after
> rebasing:
>
>     #!/bin/sh
>
>     set -e
>
>     rm -fr repo
>     git init repo
>     cd repo
>
>     git branch -M main
>
>     echo "base" >base
>     git add base
>     git commit -m "base"
>
>     git checkout -b my-feature
>     echo "feature" >feature
>     git add feature
>     git commit -m "feature"
>
>     base="$(git rev-parse main)"
>     old="$(git rev-parse my-feature)"
>
>     git checkout main
>     echo "other" >>base
>     git add base
>     git commit -m "new"
>
>     git checkout my-feature
>     git rebase $base --onto main
>
>     tip="$(git rev-parse my-feature)"
>
>     git range-diff $base $old $tip
>     git range-diff --left-only $base $old $tip
>
Thanks, I have seen an obvious phenomenon from
your shell script, and I wrote test code based on it.

> >     I may need to ask reviewers opinion first.
> >
> >     this want to fix #876 Thanks.
> >
> > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-898%2Fadlternative%2Fformat-patch-range-diff-right-only-v1
> > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-898/adlternative/format-patch-range-diff-right-only-v1
> > Pull-Request: https://github.com/gitgitgadget/git/pull/898
> >
> >  Documentation/git-format-patch.txt |  5 ++++-
> >  builtin/log.c                      | 15 ++++++++++-----
> >  2 files changed, 14 insertions(+), 6 deletions(-)
> >
> > diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt
> > index 3e49bf221087..5dae34c3090b 100644
> > --- a/Documentation/git-format-patch.txt
> > +++ b/Documentation/git-format-patch.txt
> > @@ -27,7 +27,7 @@ SYNOPSIS
> >                  [--[no-]encode-email-headers]
> >                  [--no-notes | --notes[=<ref>]]
> >                  [--interdiff=<previous>]
> > -                [--range-diff=<previous> [--creation-factor=<percent>]]
> > +                [--range-diff=<previous> [--creation-factor=<percent>] [--right-only]]
> >                  [--filename-max-length=<n>]
> >                  [--progress]
> >                  [<common diff options>]
> > @@ -301,6 +301,9 @@ material (this may change in the future).
> >       creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
> >       for details.
> >
> > +--only-right:
> > +     Used with `--range-diff`, only emit output related to the second range.
> > +
>
> s/--only-right/--right-only?
Now it's all "--left-only".
>
> >  --notes[=<ref>]::
> >  --no-notes::
> >       Append the notes (see linkgit:git-notes[1]) for the commit
> > diff --git a/builtin/log.c b/builtin/log.c
> > index f67b67d80ed1..5d2f39fd19a7 100644
> > --- a/builtin/log.c
> > +++ b/builtin/log.c
> > @@ -1153,7 +1153,7 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
> >                             struct commit *origin,
> >                             int nr, struct commit **list,
> >                             const char *branch_name,
> > -                           int quiet)
> > +                           int quiet, int right_only)
> >  {
> >       const char *committer;
> >       struct shortlog log;
> > @@ -1228,7 +1228,8 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
> >                       .creation_factor = rev->creation_factor,
> >                       .dual_color = 1,
> >                       .diffopt = &opts,
> > -                     .other_arg = &other_arg
> > +                     .other_arg = &other_arg,
> > +                     .right_only = right_only
> >               };
> >
> >               diff_setup(&opts);
> > @@ -1732,7 +1733,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
> >       struct strbuf rdiff2 = STRBUF_INIT;
> >       struct strbuf rdiff_title = STRBUF_INIT;
> >       int creation_factor = -1;
> > -
> > +     int right_only = 0;
>
> Nit: the newline between creation_factor's declaration and
> builtin_format_patch_options's is good to keep.
>
> >       const struct option builtin_format_patch_options[] = {
> >               OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
> >                           N_("use [PATCH n/m] even with a single patch"),
> > @@ -1814,6 +1815,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
> >                            parse_opt_object_name),
> >               OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
> >                          N_("show changes against <refspec> in cover letter or single patch")),
> > +             OPT_BOOL(0, "only-right", &right_only,
> > +                      N_("only emit output related to the second range")),
> >               OPT_INTEGER(0, "creation-factor", &creation_factor,
> >                           N_("percentage by which creation is weighted")),
> >               OPT_END()
> > @@ -2087,7 +2090,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
> >               creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
> >       else if (!rdiff_prev)
> >               die(_("--creation-factor requires --range-diff"));
> > -
> > +     if (right_only && !rdiff_prev)
> > +             die(_("--right_only requires --range-diff"));
>
> These may be good to combine into one big conditional that dies whenever
> an option requiring --range-diff is set, but --range-diff is not:
>
Good tip.
>   if (!rdiff_prev) {
>     if (creation_factor)
>       die(_("--creation-factor requires --range-diff"));
>     if (right_only)
>       die(_("--right-only requires --range-diff"));
>   }
>
> Thanks,
> Taylor

Thanks,
ZheNing Hu

^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v2] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-07  6:57 [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff ZheNing Hu via GitGitGadget
  2021-03-08 20:18 ` Taylor Blau
@ 2021-03-09  8:33 ` ZheNing Hu via GitGitGadget
  2021-03-09  9:00   ` Eric Sunshine
  2021-03-09 10:28   ` [PATCH v3] " ZheNing Hu via GitGitGadget
  1 sibling, 2 replies; 16+ messages in thread
From: ZheNing Hu via GitGitGadget @ 2021-03-09  8:33 UTC (permalink / raw)
  To: git
  Cc: Johannes Schindelin, Denton Liu, Junio C Hamano, Taylor Blau,
	Taylor Blau, ZheNing Hu, ZheNing Hu

From: ZheNing Hu <adlternative@gmail.com>

In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
Taylor Blau proposing `git format-patch --cover-letter
--range-diff` may mistakenly place upstream commit in the
range-diff output. Teach `format-patch` pass `--left-only`
to range-diff,can avoid this kind of mistake.

Signed-off-by: ZheNing Hu <adlternative@gmail.com>
---
    [GSOC][RFC] format-patch: pass --left-only to range-diff
    
    With the help of Taylor Blau, I understood why the upstream commit
    appeared in the range-diff, and completed the writing of the test.
    
    this want to fix #876 Thanks.

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-898%2Fadlternative%2Fformat-patch-range-diff-right-only-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-898/adlternative/format-patch-range-diff-right-only-v2
Pull-Request: https://github.com/gitgitgadget/git/pull/898

Range-diff vs v1:

 1:  590bd06ee5ee ! 1:  8daffd4f7546 [GSOC][RFC] format-patch: pass --right-only to range-diff
     @@ Metadata
      Author: ZheNing Hu <adlternative@gmail.com>
      
       ## Commit message ##
     -    [GSOC][RFC] format-patch: pass --right-only to range-diff
     +    [GSOC][RFC] format-patch: pass --left-only to range-diff
      
          In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
     -    Taylor Blau proposing 'git format-patch --cover-letter
     -    --range-diff' may mistakenly place upstream commit in the
     -    range-diff output. Teach `format-patch` pass `--right-only`
     -    to range-diff, maybe can avoid this kind of mistake.
     -
     -    Because `git rebase --apply` will internally call `git
     -    format-patch -k --stdout --full-index --cherry-pick --right-only
     -    --src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter
     -    --pretty=mboxrd --topo-order --no-base`, In order to avoid the
     -    ambiguity of `--right-only`, users can call `format-patch
     -    --range-diff --only-right` instead of `--right-only` to only emit
     -    output related to the second range.
     +    Taylor Blau proposing `git format-patch --cover-letter
     +    --range-diff` may mistakenly place upstream commit in the
     +    range-diff output. Teach `format-patch` pass `--left-only`
     +    to range-diff,can avoid this kind of mistake.
      
          Signed-off-by: ZheNing Hu <adlternative@gmail.com>
      
     @@ Documentation/git-format-patch.txt: SYNOPSIS
       		   [--no-notes | --notes[=<ref>]]
       		   [--interdiff=<previous>]
      -		   [--range-diff=<previous> [--creation-factor=<percent>]]
     -+		   [--range-diff=<previous> [--creation-factor=<percent>] [--right-only]]
     ++		   [--range-diff=<previous> [--creation-factor=<percent>] [--left-only]]
       		   [--filename-max-length=<n>]
       		   [--progress]
       		   [<common diff options>]
     @@ Documentation/git-format-patch.txt: material (this may change in the future).
       	creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
       	for details.
       
     -+--only-right:
     -+	Used with `--range-diff`, only emit output related to the second range.
     ++--left-only:
     ++	Used with `--range-diff`, only emit output related to the first range.
      +
       --notes[=<ref>]::
       --no-notes::
     @@ builtin/log.c: static void make_cover_letter(struct rev_info *rev, int use_separ
       			      int nr, struct commit **list,
       			      const char *branch_name,
      -			      int quiet)
     -+			      int quiet, int right_only)
     ++			      int quiet, int left_only)
       {
       	const char *committer;
       	struct shortlog log;
     @@ builtin/log.c: static void make_cover_letter(struct rev_info *rev, int use_separ
       			.diffopt = &opts,
      -			.other_arg = &other_arg
      +			.other_arg = &other_arg,
     -+			.right_only = right_only
     ++			.left_only = left_only
       		};
       
       		diff_setup(&opts);
     @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *pre
       	struct strbuf rdiff2 = STRBUF_INIT;
       	struct strbuf rdiff_title = STRBUF_INIT;
       	int creation_factor = -1;
     --
     -+	int right_only = 0;
     ++	int left_only = 0;
     + 
       	const struct option builtin_format_patch_options[] = {
       		OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
     - 			    N_("use [PATCH n/m] even with a single patch"),
      @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *prefix)
       			     parse_opt_object_name),
       		OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
       			   N_("show changes against <refspec> in cover letter or single patch")),
     -+		OPT_BOOL(0, "only-right", &right_only,
     -+			 N_("only emit output related to the second range")),
     ++		OPT_BOOL(0, "left-only", &left_only,
     ++			 N_("only emit output related to the first range")),
       		OPT_INTEGER(0, "creation-factor", &creation_factor,
       			    N_("percentage by which creation is weighted")),
       		OPT_END()
      @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *prefix)
     + 
     + 	if (creation_factor < 0)
       		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
     - 	else if (!rdiff_prev)
     - 		die(_("--creation-factor requires --range-diff"));
     +-	else if (!rdiff_prev)
     +-		die(_("--creation-factor requires --range-diff"));
      -
     -+	if (right_only && !rdiff_prev)
     -+		die(_("--right_only requires --range-diff"));
     ++	else if (!rdiff_prev) {
     ++		if (creation_factor >= 0)
     ++			die(_("--creation-factor requires --range-diff"));
     ++		if (left_only)
     ++			die(_("--left-only requires --range-diff"));
     ++	}
       	if (rdiff_prev) {
       		if (!cover_letter && total != 1)
       			die(_("--range-diff requires --cover-letter or single patch"));
     @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *pre
       		make_cover_letter(&rev, !!output_directory,
      -				  origin, nr, list, branch_name, quiet);
      +				  origin, nr, list, branch_name, quiet,
     -+					right_only);
     ++					left_only);
       		print_bases(&bases, rev.diffopt.file);
       		print_signature(rev.diffopt.file);
       		total++;
     +
     + ## t/t3206-range-diff.sh ##
     +@@ t/t3206-range-diff.sh: test_expect_success '--left-only/--right-only' '
     + 	test_cmp expect actual
     + '
     + 
     ++test_expect_success 'format-patch --range-diff --left-only' '
     ++	rm -fr repo &&
     ++	git init repo &&
     ++	cd repo &&
     ++	git branch -M main &&
     ++	echo "base" >base &&
     ++	git add base &&
     ++	git commit -m "base" &&
     ++	git checkout -b my-feature &&
     ++	echo "feature" >feature &&
     ++	git add feature &&
     ++	git commit -m "feature" &&
     ++	base="$(git rev-parse main)" &&
     ++	old="$(git rev-parse my-feature)" &&
     ++	git checkout main &&
     ++	echo "other" >>base &&
     ++	git add base &&
     ++	git commit -m "new" &&
     ++	git checkout my-feature &&
     ++	git rebase $base --onto main &&
     ++	tip="$(git rev-parse my-feature)" &&
     ++	git format-patch --range-diff $base $old $tip --cover-letter  &&
     ++	grep  "> 1: .* feature$"  0000-cover-letter.patch &&
     ++	git format-patch --range-diff $base $old $tip --left-only --cover-letter &&
     ++	! grep  "> 1: .* feature$"  0000-cover-letter.patch
     ++'
     ++
     ++
     + test_done


 Documentation/git-format-patch.txt |  5 ++++-
 builtin/log.c                      | 20 ++++++++++++++------
 t/t3206-range-diff.sh              | 28 ++++++++++++++++++++++++++++
 3 files changed, 46 insertions(+), 7 deletions(-)

diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt
index 3e49bf221087..8c5eca0ba1e3 100644
--- a/Documentation/git-format-patch.txt
+++ b/Documentation/git-format-patch.txt
@@ -27,7 +27,7 @@ SYNOPSIS
 		   [--[no-]encode-email-headers]
 		   [--no-notes | --notes[=<ref>]]
 		   [--interdiff=<previous>]
-		   [--range-diff=<previous> [--creation-factor=<percent>]]
+		   [--range-diff=<previous> [--creation-factor=<percent>] [--left-only]]
 		   [--filename-max-length=<n>]
 		   [--progress]
 		   [<common diff options>]
@@ -301,6 +301,9 @@ material (this may change in the future).
 	creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
 	for details.
 
+--left-only:
+	Used with `--range-diff`, only emit output related to the first range.
+
 --notes[=<ref>]::
 --no-notes::
 	Append the notes (see linkgit:git-notes[1]) for the commit
diff --git a/builtin/log.c b/builtin/log.c
index f67b67d80ed1..5620dfafb37e 100644
--- a/builtin/log.c
+++ b/builtin/log.c
@@ -1153,7 +1153,7 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			      struct commit *origin,
 			      int nr, struct commit **list,
 			      const char *branch_name,
-			      int quiet)
+			      int quiet, int left_only)
 {
 	const char *committer;
 	struct shortlog log;
@@ -1228,7 +1228,8 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			.creation_factor = rev->creation_factor,
 			.dual_color = 1,
 			.diffopt = &opts,
-			.other_arg = &other_arg
+			.other_arg = &other_arg,
+			.left_only = left_only
 		};
 
 		diff_setup(&opts);
@@ -1732,6 +1733,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 	struct strbuf rdiff2 = STRBUF_INIT;
 	struct strbuf rdiff_title = STRBUF_INIT;
 	int creation_factor = -1;
+	int left_only = 0;
 
 	const struct option builtin_format_patch_options[] = {
 		OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
@@ -1814,6 +1816,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 			     parse_opt_object_name),
 		OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
 			   N_("show changes against <refspec> in cover letter or single patch")),
+		OPT_BOOL(0, "left-only", &left_only,
+			 N_("only emit output related to the first range")),
 		OPT_INTEGER(0, "creation-factor", &creation_factor,
 			    N_("percentage by which creation is weighted")),
 		OPT_END()
@@ -2085,9 +2089,12 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 
 	if (creation_factor < 0)
 		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
-	else if (!rdiff_prev)
-		die(_("--creation-factor requires --range-diff"));
-
+	else if (!rdiff_prev) {
+		if (creation_factor >= 0)
+			die(_("--creation-factor requires --range-diff"));
+		if (left_only)
+			die(_("--left-only requires --range-diff"));
+	}
 	if (rdiff_prev) {
 		if (!cover_letter && total != 1)
 			die(_("--range-diff requires --cover-letter or single patch"));
@@ -2134,7 +2141,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 		if (thread)
 			gen_message_id(&rev, "cover");
 		make_cover_letter(&rev, !!output_directory,
-				  origin, nr, list, branch_name, quiet);
+				  origin, nr, list, branch_name, quiet,
+					left_only);
 		print_bases(&bases, rev.diffopt.file);
 		print_signature(rev.diffopt.file);
 		total++;
diff --git a/t/t3206-range-diff.sh b/t/t3206-range-diff.sh
index 1b26c4c2ef91..abb0cf7fb465 100755
--- a/t/t3206-range-diff.sh
+++ b/t/t3206-range-diff.sh
@@ -748,4 +748,32 @@ test_expect_success '--left-only/--right-only' '
 	test_cmp expect actual
 '
 
+test_expect_success 'format-patch --range-diff --left-only' '
+	rm -fr repo &&
+	git init repo &&
+	cd repo &&
+	git branch -M main &&
+	echo "base" >base &&
+	git add base &&
+	git commit -m "base" &&
+	git checkout -b my-feature &&
+	echo "feature" >feature &&
+	git add feature &&
+	git commit -m "feature" &&
+	base="$(git rev-parse main)" &&
+	old="$(git rev-parse my-feature)" &&
+	git checkout main &&
+	echo "other" >>base &&
+	git add base &&
+	git commit -m "new" &&
+	git checkout my-feature &&
+	git rebase $base --onto main &&
+	tip="$(git rev-parse my-feature)" &&
+	git format-patch --range-diff $base $old $tip --cover-letter  &&
+	grep  "> 1: .* feature$"  0000-cover-letter.patch &&
+	git format-patch --range-diff $base $old $tip --left-only --cover-letter &&
+	! grep  "> 1: .* feature$"  0000-cover-letter.patch
+'
+
+
 test_done

base-commit: be7935ed8bff19f481b033d0d242c5d5f239ed50
-- 
gitgitgadget

^ permalink raw reply related	[flat|nested] 16+ messages in thread

* Re: [PATCH v2] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-09  8:33 ` [PATCH v2] [GSOC][RFC] format-patch: pass --left-only " ZheNing Hu via GitGitGadget
@ 2021-03-09  9:00   ` Eric Sunshine
  2021-03-09  9:35     ` ZheNing Hu
  2021-03-09 10:28   ` [PATCH v3] " ZheNing Hu via GitGitGadget
  1 sibling, 1 reply; 16+ messages in thread
From: Eric Sunshine @ 2021-03-09  9:00 UTC (permalink / raw)
  To: ZheNing Hu via GitGitGadget
  Cc: Git List, Johannes Schindelin, Denton Liu, Junio C Hamano,
	Taylor Blau, Taylor Blau, ZheNing Hu

On Tue, Mar 9, 2021 at 3:35 AM ZheNing Hu via GitGitGadget
<gitgitgadget@gmail.com> wrote:
> In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
> Taylor Blau proposing `git format-patch --cover-letter
> --range-diff` may mistakenly place upstream commit in the
> range-diff output. Teach `format-patch` pass `--left-only`
> to range-diff,can avoid this kind of mistake.
>
> Signed-off-by: ZheNing Hu <adlternative@gmail.com>
> ---
> diff --git a/builtin/log.c b/builtin/log.c
> @@ -2085,9 +2089,12 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
>         if (creation_factor < 0)
>                 creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
> -       else if (!rdiff_prev)
> -               die(_("--creation-factor requires --range-diff"));
> -
> +       else if (!rdiff_prev) {
> +               if (creation_factor >= 0)
> +                       die(_("--creation-factor requires --range-diff"));
> +               if (left_only)
> +                       die(_("--left-only requires --range-diff"));
> +       }

This logic starts getting difficult to reason about. It would be
easier to understand if written like this:

    if (!rdiff_prev) {
        if (creation_factor >= 0)
            die(_("--creation-factor requires --range-diff"));
        if (left_only)
            die(_("--left-only requires --range-diff"));
    }

    if (creation_factor < 0)
        creation_factor = RANGE_DIFF_..._DEFAULT;

or this (which reduces the indentation a bit):

    if (creation_factor >= 0 && !rdiff_prev)
        die(_("--creation-factor requires --range-diff"));
    if (left_only && !rdiff_prev)
        die(_("--left-only requires --range-diff"));

    if (creation_factor < 0)
        creation_factor = RANGE_DIFF_..._DEFAULT;

Subjective; not necessarily worth a re-roll.

> @@ -2134,7 +2141,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
>                 make_cover_letter(&rev, !!output_directory,
> -                                 origin, nr, list, branch_name, quiet);
> +                                 origin, nr, list, branch_name, quiet,
> +                                       left_only);

Nit: This indentation looks odd. One would expect `origin` and
`left_only` to have the same indentation.

> diff --git a/t/t3206-range-diff.sh b/t/t3206-range-diff.sh
> @@ -748,4 +748,32 @@ test_expect_success '--left-only/--right-only' '
> +test_expect_success 'format-patch --range-diff --left-only' '
> +       rm -fr repo &&
> +       git init repo &&
> +       ...
> +       ! grep  "> 1: .* feature$"  0000-cover-letter.patch
> +'
> +
> +
>  test_done

Nit: One blank line before test_done() is typical, not two.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-09  9:00   ` Eric Sunshine
@ 2021-03-09  9:35     ` ZheNing Hu
  0 siblings, 0 replies; 16+ messages in thread
From: ZheNing Hu @ 2021-03-09  9:35 UTC (permalink / raw)
  To: Eric Sunshine
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Junio C Hamano, Taylor Blau, Taylor Blau

Eric Sunshine <sunshine@sunshineco.com> 于2021年3月9日周二 下午5:00写道:
>
> On Tue, Mar 9, 2021 at 3:35 AM ZheNing Hu via GitGitGadget
> <gitgitgadget@gmail.com> wrote:
> > In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
> > Taylor Blau proposing `git format-patch --cover-letter
> > --range-diff` may mistakenly place upstream commit in the
> > range-diff output. Teach `format-patch` pass `--left-only`
> > to range-diff,can avoid this kind of mistake.
> >
> > Signed-off-by: ZheNing Hu <adlternative@gmail.com>
> > ---
> > diff --git a/builtin/log.c b/builtin/log.c
> > @@ -2085,9 +2089,12 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
> >         if (creation_factor < 0)
> >                 creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
> > -       else if (!rdiff_prev)
> > -               die(_("--creation-factor requires --range-diff"));
> > -
> > +       else if (!rdiff_prev) {
> > +               if (creation_factor >= 0)
> > +                       die(_("--creation-factor requires --range-diff"));
> > +               if (left_only)
> > +                       die(_("--left-only requires --range-diff"));
> > +       }
>
> This logic starts getting difficult to reason about. It would be
> easier to understand if written like this:
>
>     if (!rdiff_prev) {
>         if (creation_factor >= 0)
>             die(_("--creation-factor requires --range-diff"));
>         if (left_only)
>             die(_("--left-only requires --range-diff"));
>     }
>
>     if (creation_factor < 0)
>         creation_factor = RANGE_DIFF_..._DEFAULT;
>
> or this (which reduces the indentation a bit):
>
>     if (creation_factor >= 0 && !rdiff_prev)
>         die(_("--creation-factor requires --range-diff"));
>     if (left_only && !rdiff_prev)
>         die(_("--left-only requires --range-diff"));
>
>     if (creation_factor < 0)
>         creation_factor = RANGE_DIFF_..._DEFAULT;
>
Emm,I would prefer to use the above.
> Subjective; not necessarily worth a re-roll.
>
> > @@ -2134,7 +2141,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
> >                 make_cover_letter(&rev, !!output_directory,
> > -                                 origin, nr, list, branch_name, quiet);
> > +                                 origin, nr, list, branch_name, quiet,
> > +                                       left_only);
>
> Nit: This indentation looks odd. One would expect `origin` and
> `left_only` to have the same indentation.
>
> > diff --git a/t/t3206-range-diff.sh b/t/t3206-range-diff.sh
> > @@ -748,4 +748,32 @@ test_expect_success '--left-only/--right-only' '
> > +test_expect_success 'format-patch --range-diff --left-only' '
> > +       rm -fr repo &&
> > +       git init repo &&
> > +       ...
> > +       ! grep  "> 1: .* feature$"  0000-cover-letter.patch
> > +'
> > +
> > +
> >  test_done
>
> Nit: One blank line before test_done() is typical, not two.

Thanks for the reminder.
I'm going to modify these format errors.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-09  8:33 ` [PATCH v2] [GSOC][RFC] format-patch: pass --left-only " ZheNing Hu via GitGitGadget
  2021-03-09  9:00   ` Eric Sunshine
@ 2021-03-09 10:28   ` ZheNing Hu via GitGitGadget
  2021-03-12 22:50     ` Junio C Hamano
  2021-03-14  8:10     ` [PATCH v4] [GSOC] " ZheNing Hu via GitGitGadget
  1 sibling, 2 replies; 16+ messages in thread
From: ZheNing Hu via GitGitGadget @ 2021-03-09 10:28 UTC (permalink / raw)
  To: git
  Cc: Johannes Schindelin, Denton Liu, Junio C Hamano, Taylor Blau,
	Taylor Blau, Eric Sunshine, ZheNing Hu, ZheNing Hu

From: ZheNing Hu <adlternative@gmail.com>

In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
Taylor Blau proposing `git format-patch --cover-letter
--range-diff` may mistakenly place upstream commit in the
range-diff output. Teach `format-patch` pass `--left-only`
to range-diff,can avoid this kind of mistake.

Signed-off-by: ZheNing Hu <adlternative@gmail.com>
---
    [GSOC][RFC] format-patch: pass --left-only to range-diff
    
    With the help of Taylor Blau, I understood why the upstream commit
    appeared in the range-diff, and completed the writing of the test.
    
    this want to fix #876 Thanks.

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-898%2Fadlternative%2Fformat-patch-range-diff-right-only-v3
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-898/adlternative/format-patch-range-diff-right-only-v3
Pull-Request: https://github.com/gitgitgadget/git/pull/898

Range-diff vs v2:

 1:  8daffd4f7546 ! 1:  5c58eb186d41 [GSOC][RFC] format-patch: pass --left-only to range-diff
     @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *pre
       			    N_("percentage by which creation is weighted")),
       		OPT_END()
      @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *prefix)
     + 					     _("Interdiff against v%d:"));
     + 	}
       
     - 	if (creation_factor < 0)
     - 		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
     --	else if (!rdiff_prev)
     --		die(_("--creation-factor requires --range-diff"));
     --
     -+	else if (!rdiff_prev) {
     ++	if (!rdiff_prev) {
      +		if (creation_factor >= 0)
      +			die(_("--creation-factor requires --range-diff"));
      +		if (left_only)
      +			die(_("--left-only requires --range-diff"));
      +	}
     ++
     + 	if (creation_factor < 0)
     + 		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
     +-	else if (!rdiff_prev)
     +-		die(_("--creation-factor requires --range-diff"));
     + 
       	if (rdiff_prev) {
       		if (!cover_letter && total != 1)
     - 			die(_("--range-diff requires --cover-letter or single patch"));
      @@ builtin/log.c: int cmd_format_patch(int argc, const char **argv, const char *prefix)
       		if (thread)
       			gen_message_id(&rev, "cover");
       		make_cover_letter(&rev, !!output_directory,
      -				  origin, nr, list, branch_name, quiet);
      +				  origin, nr, list, branch_name, quiet,
     -+					left_only);
     ++				  left_only);
       		print_bases(&bases, rev.diffopt.file);
       		print_signature(rev.diffopt.file);
       		total++;
     @@ t/t3206-range-diff.sh: test_expect_success '--left-only/--right-only' '
      +	git rebase $base --onto main &&
      +	tip="$(git rev-parse my-feature)" &&
      +	git format-patch --range-diff $base $old $tip --cover-letter  &&
     -+	grep  "> 1: .* feature$"  0000-cover-letter.patch &&
     ++	grep "> 1: .* feature$" 0000-cover-letter.patch &&
      +	git format-patch --range-diff $base $old $tip --left-only --cover-letter &&
     -+	! grep  "> 1: .* feature$"  0000-cover-letter.patch
     ++	! grep "> 1: .* feature$" 0000-cover-letter.patch
      +'
     -+
      +
       test_done


 Documentation/git-format-patch.txt |  5 ++++-
 builtin/log.c                      | 20 +++++++++++++++-----
 t/t3206-range-diff.sh              | 27 +++++++++++++++++++++++++++
 3 files changed, 46 insertions(+), 6 deletions(-)

diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt
index 3e49bf221087..8c5eca0ba1e3 100644
--- a/Documentation/git-format-patch.txt
+++ b/Documentation/git-format-patch.txt
@@ -27,7 +27,7 @@ SYNOPSIS
 		   [--[no-]encode-email-headers]
 		   [--no-notes | --notes[=<ref>]]
 		   [--interdiff=<previous>]
-		   [--range-diff=<previous> [--creation-factor=<percent>]]
+		   [--range-diff=<previous> [--creation-factor=<percent>] [--left-only]]
 		   [--filename-max-length=<n>]
 		   [--progress]
 		   [<common diff options>]
@@ -301,6 +301,9 @@ material (this may change in the future).
 	creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
 	for details.
 
+--left-only:
+	Used with `--range-diff`, only emit output related to the first range.
+
 --notes[=<ref>]::
 --no-notes::
 	Append the notes (see linkgit:git-notes[1]) for the commit
diff --git a/builtin/log.c b/builtin/log.c
index f67b67d80ed1..21fed9db82d6 100644
--- a/builtin/log.c
+++ b/builtin/log.c
@@ -1153,7 +1153,7 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			      struct commit *origin,
 			      int nr, struct commit **list,
 			      const char *branch_name,
-			      int quiet)
+			      int quiet, int left_only)
 {
 	const char *committer;
 	struct shortlog log;
@@ -1228,7 +1228,8 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			.creation_factor = rev->creation_factor,
 			.dual_color = 1,
 			.diffopt = &opts,
-			.other_arg = &other_arg
+			.other_arg = &other_arg,
+			.left_only = left_only
 		};
 
 		diff_setup(&opts);
@@ -1732,6 +1733,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 	struct strbuf rdiff2 = STRBUF_INIT;
 	struct strbuf rdiff_title = STRBUF_INIT;
 	int creation_factor = -1;
+	int left_only = 0;
 
 	const struct option builtin_format_patch_options[] = {
 		OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
@@ -1814,6 +1816,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 			     parse_opt_object_name),
 		OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
 			   N_("show changes against <refspec> in cover letter or single patch")),
+		OPT_BOOL(0, "left-only", &left_only,
+			 N_("only emit output related to the first range")),
 		OPT_INTEGER(0, "creation-factor", &creation_factor,
 			    N_("percentage by which creation is weighted")),
 		OPT_END()
@@ -2083,10 +2087,15 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 					     _("Interdiff against v%d:"));
 	}
 
+	if (!rdiff_prev) {
+		if (creation_factor >= 0)
+			die(_("--creation-factor requires --range-diff"));
+		if (left_only)
+			die(_("--left-only requires --range-diff"));
+	}
+
 	if (creation_factor < 0)
 		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
-	else if (!rdiff_prev)
-		die(_("--creation-factor requires --range-diff"));
 
 	if (rdiff_prev) {
 		if (!cover_letter && total != 1)
@@ -2134,7 +2143,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 		if (thread)
 			gen_message_id(&rev, "cover");
 		make_cover_letter(&rev, !!output_directory,
-				  origin, nr, list, branch_name, quiet);
+				  origin, nr, list, branch_name, quiet,
+				  left_only);
 		print_bases(&bases, rev.diffopt.file);
 		print_signature(rev.diffopt.file);
 		total++;
diff --git a/t/t3206-range-diff.sh b/t/t3206-range-diff.sh
index 1b26c4c2ef91..6178a12dd4b1 100755
--- a/t/t3206-range-diff.sh
+++ b/t/t3206-range-diff.sh
@@ -748,4 +748,31 @@ test_expect_success '--left-only/--right-only' '
 	test_cmp expect actual
 '
 
+test_expect_success 'format-patch --range-diff --left-only' '
+	rm -fr repo &&
+	git init repo &&
+	cd repo &&
+	git branch -M main &&
+	echo "base" >base &&
+	git add base &&
+	git commit -m "base" &&
+	git checkout -b my-feature &&
+	echo "feature" >feature &&
+	git add feature &&
+	git commit -m "feature" &&
+	base="$(git rev-parse main)" &&
+	old="$(git rev-parse my-feature)" &&
+	git checkout main &&
+	echo "other" >>base &&
+	git add base &&
+	git commit -m "new" &&
+	git checkout my-feature &&
+	git rebase $base --onto main &&
+	tip="$(git rev-parse my-feature)" &&
+	git format-patch --range-diff $base $old $tip --cover-letter  &&
+	grep "> 1: .* feature$" 0000-cover-letter.patch &&
+	git format-patch --range-diff $base $old $tip --left-only --cover-letter &&
+	! grep "> 1: .* feature$" 0000-cover-letter.patch
+'
+
 test_done

base-commit: be7935ed8bff19f481b033d0d242c5d5f239ed50
-- 
gitgitgadget

^ permalink raw reply related	[flat|nested] 16+ messages in thread

* Re: [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff
  2021-03-08 20:18 ` Taylor Blau
  2021-03-09  7:28   ` ZheNing Hu
@ 2021-03-12 21:55   ` Junio C Hamano
  2021-03-12 22:09     ` Junio C Hamano
  1 sibling, 1 reply; 16+ messages in thread
From: Junio C Hamano @ 2021-03-12 21:55 UTC (permalink / raw)
  To: Taylor Blau
  Cc: ZheNing Hu via GitGitGadget, git, Johannes Schindelin,
	Denton Liu, Taylor Blau, ZheNing Hu

Taylor Blau <ttaylorr@github.com> writes:

> Note that I think (and Johannes--cc'd--could confirm) that what you want
> is '--left-only' to discard rebased changes from the upstream branch:

Does that mean range-diff got --left-only and --right-only
backwards?  IOW, doesn't

$ git log --left-only A...B

show what is shown in "git log B..A", and if so, shouldn't

$ git range-diff --left-only A...B

which is a synonym for

$ git range-diff B..A A..B

also give commits in the B..A range?

Puzzled....


^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff
  2021-03-12 21:55   ` Junio C Hamano
@ 2021-03-12 22:09     ` Junio C Hamano
  0 siblings, 0 replies; 16+ messages in thread
From: Junio C Hamano @ 2021-03-12 22:09 UTC (permalink / raw)
  To: Taylor Blau
  Cc: ZheNing Hu via GitGitGadget, git, Johannes Schindelin,
	Denton Liu, Taylor Blau, ZheNing Hu

Junio C Hamano <gitster@pobox.com> writes:

> Taylor Blau <ttaylorr@github.com> writes:
>
>> Note that I think (and Johannes--cc'd--could confirm) that what you want
>> is '--left-only' to discard rebased changes from the upstream branch:
>
> Does that mean range-diff got --left-only and --right-only
> backwards?  IOW, doesn't
>
> $ git log --left-only A...B
>
> show what is shown in "git log B..A", and if so, shouldn't
>
> $ git range-diff --left-only A...B
>
> which is a synonym for
>
> $ git range-diff B..A A..B
>
> also give commits in the B..A range?
>
> Puzzled....

Ah, it was only that "range-diff --help" uses confusing
double-negative phrasing.  I.e.

    --left-only::
            Suppress commits that are missing from the first specified range
            (or the "left range" when using the `<rev1>...<rev2>` format).

If you "suppress those that are missing from X", you are effectively
including only those that appear in X, so two conclusions I draw are

 - "range-diff" does not get "--left/right-only" backwards;

 - its documentation should be improved.

Perhaps

    --left-only::
	Include only commits that are in the first range (or the
	"left range" when using the `<rev1>...<rev2>` format).

wouldn't have made me puzzled.

Thanks.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-09 10:28   ` [PATCH v3] " ZheNing Hu via GitGitGadget
@ 2021-03-12 22:50     ` Junio C Hamano
  2021-03-13  4:01       ` ZheNing Hu
  2021-03-14  8:10     ` [PATCH v4] [GSOC] " ZheNing Hu via GitGitGadget
  1 sibling, 1 reply; 16+ messages in thread
From: Junio C Hamano @ 2021-03-12 22:50 UTC (permalink / raw)
  To: ZheNing Hu via GitGitGadget
  Cc: git, Johannes Schindelin, Denton Liu, Taylor Blau, Taylor Blau,
	Eric Sunshine, ZheNing Hu

"ZheNing Hu via GitGitGadget" <gitgitgadget@gmail.com> writes:

> From: ZheNing Hu <adlternative@gmail.com>
>
> In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
> Taylor Blau proposing `git format-patch --cover-letter
> --range-diff` may mistakenly place upstream commit in the
> range-diff output. Teach `format-patch` pass `--left-only`
> to range-diff,can avoid this kind of mistake.

The above is a bit too dense for average readers to grok.  Even if
the readers refer to the external reference, it is unclear where the
"may mistakenly" can come from and why "--left-only" would be
useful (and our log message should not depend on external material
so heavily to begin with).

So let's think aloud to see what use case this may be helpful, and
how the proposed solution makes the world a better place.

If I understand correctly, the use case this tries to help is this:

 * You had sent the v1 iteration of topic.  It was in the range
   B1..T1 where B1 is the tip of the integration branch (like
   'master') from the upstream.

 * To prepare for the v2 iteration, not only you updated individual
   commits, you rebased the series on a new upstream.  Now the topic
   is in the range B2..T2, where B2 is the tip of the integration
   branch from the upstream, and it is very likely that B2 is a
   descendant of B1.

And you want to find out how your commits in T2 (new iteration)
compares with those in T1 (old iteration).  Normally,

    $ git range-diff T1...T2

would be the shortest-to-type and correct version but that is
invalidated because you rebased.

    ---o---B1--b---b---b---B2
            \               \
	     t---t---T1      s---s---s---T2

You'd have commits B1..T1 on the left hand side of the range-diff,
while the right hand side has not just B2..T2 but also commits in
the range B1..B2, too.

By using --left-only (i.e. show only those pair that maps from
commits in the left range), you can exclude the commits in the
B1..B2.

    $ git range-diff --left-only T1...T2

I however wonder what --left-only (Suppress commits that are missing
from the first range) would do to commits in range B2..T2 (they are
all yours) that are (1) added since the v1 iteration, or (2)
modified so drastically that no matching commit is found.  With the
right invocation, of course,

    $ git range-diff B1..T1 B2..T2

you would not have such a problem.  If 2 't's in B1..T1 correspond
to 2 of the 3 's's in B2..T2, at least the presense of the third 's'
that did not match would show up in the output, making it clear that
you have one more commit relative to the earlier iteration.  If use
of --left-only filters it out, the output may be misleading to the
readers, no?

I started writing (or "thinking aloud") hoping that I can help
coming up with a better log message to describe the problem being
solved, but I ended up with "does this make the system better?"

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-12 22:50     ` Junio C Hamano
@ 2021-03-13  4:01       ` ZheNing Hu
  2021-03-13 23:23         ` Junio C Hamano
  0 siblings, 1 reply; 16+ messages in thread
From: ZheNing Hu @ 2021-03-13  4:01 UTC (permalink / raw)
  To: Junio C Hamano
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Taylor Blau, Taylor Blau, Eric Sunshine

Junio C Hamano <gitster@pobox.com> 于2021年3月13日周六 上午6:51写道:
>
> "ZheNing Hu via GitGitGadget" <gitgitgadget@gmail.com> writes:
>
> > From: ZheNing Hu <adlternative@gmail.com>
> >
> > In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
> > Taylor Blau proposing `git format-patch --cover-letter
> > --range-diff` may mistakenly place upstream commit in the
> > range-diff output. Teach `format-patch` pass `--left-only`
> > to range-diff,can avoid this kind of mistake.
>
> The above is a bit too dense for average readers to grok.  Even if
> the readers refer to the external reference, it is unclear where the
> "may mistakenly" can come from and why "--left-only" would be
> useful (and our log message should not depend on external material
> so heavily to begin with).
>

You are right, commit information with the original thread link may make
it difficult for readers to read. I will pay attention.

> So let's think aloud to see what use case this may be helpful, and
> how the proposed solution makes the world a better place.
>
> If I understand correctly, the use case this tries to help is this:
>
>  * You had sent the v1 iteration of topic.  It was in the range
>    B1..T1 where B1 is the tip of the integration branch (like
>    'master') from the upstream.
>
>  * To prepare for the v2 iteration, not only you updated individual
>    commits, you rebased the series on a new upstream.  Now the topic
>    is in the range B2..T2, where B2 is the tip of the integration
>    branch from the upstream, and it is very likely that B2 is a
>    descendant of B1.
>
> And you want to find out how your commits in T2 (new iteration)
> compares with those in T1 (old iteration).  Normally,
>
>     $ git range-diff T1...T2
>
> would be the shortest-to-type and correct version but that is
> invalidated because you rebased.
>
>     ---o---B1--b---b---b---B2
>             \               \
>              t---t---T1      s---s---s---T2
>
> You'd have commits B1..T1 on the left hand side of the range-diff,
> while the right hand side has not just B2..T2 but also commits in
> the range B1..B2, too.
>
> By using --left-only (i.e. show only those pair that maps from
> commits in the left range), you can exclude the commits in the
> B1..B2.
>
>     $ git range-diff --left-only T1...T2
>
> I however wonder what --left-only (Suppress commits that are missing
> from the first range) would do to commits in range B2..T2 (they are
> all yours) that are (1) added since the v1 iteration, or (2)
> modified so drastically that no matching commit is found.  With the
> right invocation, of course,
>
>     $ git range-diff B1..T1 B2..T2
>
> you would not have such a problem.  If 2 't's in B1..T1 correspond
> to 2 of the 3 's's in B2..T2, at least the presense of the third 's'
> that did not match would show up in the output, making it clear that
> you have one more commit relative to the earlier iteration.  If use
> of --left-only filters it out, the output may be misleading to the
> readers, no?
>
> I started writing (or "thinking aloud") hoping that I can help
> coming up with a better log message to describe the problem being
> solved, but I ended up with "does this make the system better?"

Junio, thank you for elaborating this issue in detail and clearly.
I probably understand what you mean by "git range-diff B1..T1 B2..T2"
 to correctly output the commits on my two version topic branch, without
including the upstream commits of B1..B2.So we don’t even need to specify
the `--left-only` to avoid the output of B1...B2, right?

The only thing I can think of now is that if users tend to use T1...T2
to compare
 the differences between the two topics, will the upstream commit in
B1...B2 appear
more abrupt?

Thanks.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-13  4:01       ` ZheNing Hu
@ 2021-03-13 23:23         ` Junio C Hamano
  2021-03-14  2:16           ` ZheNing Hu
  0 siblings, 1 reply; 16+ messages in thread
From: Junio C Hamano @ 2021-03-13 23:23 UTC (permalink / raw)
  To: ZheNing Hu
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Taylor Blau, Taylor Blau, Eric Sunshine

ZheNing Hu <adlternative@gmail.com> writes:

> Junio C Hamano <gitster@pobox.com> 于2021年3月13日周六 上午6:51写道:
>>
>> "ZheNing Hu via GitGitGadget" <gitgitgadget@gmail.com> writes:
>>
>> > From: ZheNing Hu <adlternative@gmail.com>
>> >
>> > In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
>> > Taylor Blau proposing `git format-patch --cover-letter
>> > --range-diff` may mistakenly place upstream commit in the
>> > range-diff output. Teach `format-patch` pass `--left-only`
>> > to range-diff,can avoid this kind of mistake.
>>
>> The above is a bit too dense for average readers to grok.  Even if
>> the readers refer to the external reference, it is unclear where the
>> "may mistakenly" can come from and why "--left-only" would be
>> useful (and our log message should not depend on external material
>> so heavily to begin with).
>>
>
> You are right, commit information with the original thread link may make
> it difficult for readers to read. I will pay attention.
>
>> So let's think aloud to see what use case this may be helpful, and
>> how the proposed solution makes the world a better place.
>>
>> If I understand correctly, the use case this tries to help is this:
>>
>>  * You had sent the v1 iteration of topic.  It was in the range
>>    B1..T1 where B1 is the tip of the integration branch (like
>>    'master') from the upstream.
>>
>>  * To prepare for the v2 iteration, not only you updated individual
>>    commits, you rebased the series on a new upstream.  Now the topic
>>    is in the range B2..T2, where B2 is the tip of the integration
>>    branch from the upstream, and it is very likely that B2 is a
>>    descendant of B1.
>>
>> And you want to find out how your commits in T2 (new iteration)
>> compares with those in T1 (old iteration).  Normally,
>>
>>     $ git range-diff T1...T2
>>
>> would be the shortest-to-type and correct version but that is
>> invalidated because you rebased.
>>
>>     ---o---B1--b---b---b---B2
>>             \               \
>>              t---t---T1      s---s---s---T2
>>
>> You'd have commits B1..T1 on the left hand side of the range-diff,
>> while the right hand side has not just B2..T2 but also commits in
>> the range B1..B2, too.
>>
>> By using --left-only (i.e. show only those pair that maps from
>> commits in the left range), you can exclude the commits in the
>> B1..B2.
>>
>>     $ git range-diff --left-only T1...T2
>>
>> I however wonder what --left-only (Suppress commits that are missing
>> from the first range) would do to commits in range B2..T2 (they are
>> all yours) that are (1) added since the v1 iteration, or (2)
>> modified so drastically that no matching commit is found.  With the
>> right invocation, of course,
>>
>>     $ git range-diff B1..T1 B2..T2
>>
>> you would not have such a problem.  If 2 't's in B1..T1 correspond
>> to 2 of the 3 's's in B2..T2, at least the presense of the third 's'
>> that did not match would show up in the output, making it clear that
>> you have one more commit relative to the earlier iteration.  If use
>> of --left-only filters it out, the output may be misleading to the
>> readers, no?
>>
>> I started writing (or "thinking aloud") hoping that I can help
>> coming up with a better log message to describe the problem being
>> solved, but I ended up with "does this make the system better?"
>
> Junio, thank you for elaborating this issue in detail and clearly.
> I probably understand what you mean by "git range-diff B1..T1 B2..T2"
>  to correctly output the commits on my two version topic branch, without
> including the upstream commits of B1..B2.So we don’t even need to specify
> the `--left-only` to avoid the output of B1...B2, right?
>
> The only thing I can think of now is that if users tend to use T1...T2
> to compare
>  the differences between the two topics, will the upstream commit in
> B1...B2 appear
> more abrupt?

Yes, it would be, but that is why you need to educate users what
causes it, and what the right way to avoid unrelated commits from
appearing, and how this --left-only fits in the solution.

If some of the time, "--left-only T1...T2" would give you the same
result as the more strict "B1..T1 B2..T2", that may be why users may
want to use the "--left-only" instead as an easy/lazy alternative.

But I suspect that it would give an incorrect result some of the
time---for example, in the above example, wouldn't one of the
commits labeled as 's' be completely hidden?  And if that is the
case, the end-user documentation would need to warn about it, and
explain that it is a easy/lazy alternative that can produce
incorrect result in the log message.

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-13 23:23         ` Junio C Hamano
@ 2021-03-14  2:16           ` ZheNing Hu
  2021-03-14  2:37             ` ZheNing Hu
  0 siblings, 1 reply; 16+ messages in thread
From: ZheNing Hu @ 2021-03-14  2:16 UTC (permalink / raw)
  To: Junio C Hamano
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Taylor Blau, Taylor Blau, Eric Sunshine

Junio C Hamano <gitster@pobox.com> 于2021年3月14日周日 上午7:23写道:

>
> Yes, it would be, but that is why you need to educate users what
> causes it, and what the right way to avoid unrelated commits from
> appearing, and how this --left-only fits in the solution.
>
> If some of the time, "--left-only T1...T2" would give you the same
> result as the more strict "B1..T1 B2..T2", that may be why users may
> want to use the "--left-only" instead as an easy/lazy alternative.
>
> But I suspect that it would give an incorrect result some of the
> time---for example, in the above example, wouldn't one of the
> commits labeled as 's' be completely hidden?  And if that is the
> case, the end-user documentation would need to warn about it, and
> explain that it is a easy/lazy alternative that can produce
> incorrect result in the log message.

Thanks, I will try to illustrate these issues in the document.

My another thinking is:
Since `--left-only` inhibits "B1..B2" and "B2..T2" ( let the user
choose the left B1..T1 ), To some extent, `--right-only` can also
add ( let the user choose the right B2..T2 ). A separate `--left-only`
will be strange to the user ( If user call T2...T1 ). Since the
`git rebase --apply` will internal call `git format-patch -k --stdout
--full-index --cherry-pick --right-only ...`, I don't know what to deal
with this `--right-only` yet, because I don't how to teach git to judge
if the `--right-only` is pass from user or `git rebase --apply`, Is there
any good way?  If can solve this problem,  the user can choose the left
or right side of the free choice. (Or users don't need `--right-only?`)

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-14  2:16           ` ZheNing Hu
@ 2021-03-14  2:37             ` ZheNing Hu
  2021-03-14  2:41               ` ZheNing Hu
  0 siblings, 1 reply; 16+ messages in thread
From: ZheNing Hu @ 2021-03-14  2:37 UTC (permalink / raw)
  To: Junio C Hamano
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Taylor Blau, Taylor Blau, Eric Sunshine

ZheNing Hu <adlternative@gmail.com> 于2021年3月14日周日 上午10:16写道:
>
> Junio C Hamano <gitster@pobox.com> 于2021年3月14日周日 上午7:23写道:
>
> >
> > Yes, it would be, but that is why you need to educate users what
> > causes it, and what the right way to avoid unrelated commits from
> > appearing, and how this --left-only fits in the solution.
> >
> > If some of the time, "--left-only T1...T2" would give you the same
> > result as the more strict "B1..T1 B2..T2", that may be why users may
> > want to use the "--left-only" instead as an easy/lazy alternative.
> >
> > But I suspect that it would give an incorrect result some of the
> > time---for example, in the above example, wouldn't one of the
> > commits labeled as 's' be completely hidden?  And if that is the
> > case, the end-user documentation would need to warn about it, and
> > explain that it is a easy/lazy alternative that can produce
> > incorrect result in the log message.
>
> Thanks, I will try to illustrate these issues in the document.
>
> My another thinking is:
> Since `--left-only` inhibits "B1..B2" and "B2..T2" ( let the user
> choose the left B1..T1 ), To some extent, `--right-only` can also
> add ( let the user choose the right B2..T2 ). A separate `--left-only`
> will be strange to the user ( If user call T2...T1 ). Since the
> `git rebase --apply` will internal call `git format-patch -k --stdout
> --full-index --cherry-pick --right-only ...`, I don't know what to deal
> with this `--right-only` yet, because I don't how to teach git to judge
> if the `--right-only` is pass from user or `git rebase --apply`, Is there
> any good way?  If can solve this problem,  the user can choose the left
> or right side of the free choice. (Or users don't need `--right-only?`)

Let me refute my own point just now :`--right-only` can not show "B2..T2",
but "B1..T1", and it may be useful only when user want an inverted "T2...T1".

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v3] [GSOC][RFC] format-patch: pass --left-only to range-diff
  2021-03-14  2:37             ` ZheNing Hu
@ 2021-03-14  2:41               ` ZheNing Hu
  0 siblings, 0 replies; 16+ messages in thread
From: ZheNing Hu @ 2021-03-14  2:41 UTC (permalink / raw)
  To: Junio C Hamano
  Cc: ZheNing Hu via GitGitGadget, Git List, Johannes Schindelin,
	Denton Liu, Taylor Blau, Taylor Blau, Eric Sunshine

> Let me refute my own point just now :`--right-only` can not show "B2..T2",
> but "B1..T1", and it may be useful only when user want an inverted "T2...T1".

sorry,
s/B1..T1/B1..T2/

^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v4] [GSOC] format-patch: pass --left-only to range-diff
  2021-03-09 10:28   ` [PATCH v3] " ZheNing Hu via GitGitGadget
  2021-03-12 22:50     ` Junio C Hamano
@ 2021-03-14  8:10     ` ZheNing Hu via GitGitGadget
  1 sibling, 0 replies; 16+ messages in thread
From: ZheNing Hu via GitGitGadget @ 2021-03-14  8:10 UTC (permalink / raw)
  To: git
  Cc: Johannes Schindelin, Denton Liu, Junio C Hamano, Taylor Blau,
	Taylor Blau, Eric Sunshine, ZheNing Hu, ZheNing Hu

From: ZheNing Hu <adlternative@gmail.com>

When two different iterative versions t1 and t2 are base upstream
commit b1 and b2, b2 inherits from b1. If the user base on b1 to
generate cover-letter, but `git format-patch --range-diff=b1..t1
b1..t2 --cover-letter` may mistakenly place upstream commit in the
range-diff output in cover-letter.

Teaching `format-patch` pass `--left-only` to `range-diff`, it will
suppress the output on the upstream commits.At the same time, it has
a disadvantage: the iterative version on the right side will be ignored
too. So using `--left-only` is just a lazy way to avoid upstream output.

Signed-off-by: ZheNing Hu <adlternative@gmail.com>
---
    [GSOC][RFC] format-patch: pass --left-only to range-diff
    
    With the help of Taylor Blau and Junio, I understand why the upstream
    commit appeared in the range-diff, and completed the writing of the
    test.
    
    But I notice that in https://github.com/gitgitgadget/git/issues/876 and
    https://lore.kernel.org/git/xmqqpn0456lr.fsf@gitster.g/ Both Junio and
    Johannes Schindelin have questions about the practicality of this
    --left-only, and I am also starting to have confusion now, and
    format-patch --range-diff is more suitable for two commit ranges to
    compare the differences. It is not suitable for T1...T2, which further
    proves that the practicability of this patch may not be as good as
    previously imagined. Should I close it?

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-898%2Fadlternative%2Fformat-patch-range-diff-right-only-v4
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-898/adlternative/format-patch-range-diff-right-only-v4
Pull-Request: https://github.com/gitgitgadget/git/pull/898

Range-diff vs v3:

 1:  5c58eb186d41 ! 1:  3738a00129b5 [GSOC][RFC] format-patch: pass --left-only to range-diff
     @@ Metadata
      Author: ZheNing Hu <adlternative@gmail.com>
      
       ## Commit message ##
     -    [GSOC][RFC] format-patch: pass --left-only to range-diff
     +    [GSOC] format-patch: pass --left-only to range-diff
      
     -    In https://lore.kernel.org/git/YBx5rmVsg1LJhSKN@nand.local/,
     -    Taylor Blau proposing `git format-patch --cover-letter
     -    --range-diff` may mistakenly place upstream commit in the
     -    range-diff output. Teach `format-patch` pass `--left-only`
     -    to range-diff,can avoid this kind of mistake.
     +    When two different iterative versions t1 and t2 are base upstream
     +    commit b1 and b2, b2 inherits from b1. If the user base on b1 to
     +    generate cover-letter, but `git format-patch --range-diff=b1..t1
     +    b1..t2 --cover-letter` may mistakenly place upstream commit in the
     +    range-diff output in cover-letter.
     +
     +    Teaching `format-patch` pass `--left-only` to `range-diff`, it will
     +    suppress the output on the upstream commits.At the same time, it has
     +    a disadvantage: the iterative version on the right side will be ignored
     +    too. So using `--left-only` is just a lazy way to avoid upstream output.
      
          Signed-off-by: ZheNing Hu <adlternative@gmail.com>
      
     @@ Documentation/git-format-patch.txt: material (this may change in the future).
       
      +--left-only:
      +	Used with `--range-diff`, only emit output related to the first range.
     ++	This option only can be used in this situation: The first iteration `t1`
     ++	based on the upstream commit `b1`, the second iteration `t2` based on the
     ++	upstream commit `b2`, `b2` inherited from the `b1`, then If the user want
     ++	output the range-diff between this two iterations (t1 and t2) in cover-letter,
     ++	and just use `b1` as the same base for `--range-diff`, but they don't want the
     ++	upstream range `b1..b2` included on the right side of the range-diff output.
     ++	By using `git format-patch --range-diff=b1..t1 b1..t2 --cover-letter --left-only`,
     ++	all content on the right side will be removed, leaving only the range
     ++	`b1..t1` on the left side.
     +++
     ++Note that this `--left-only` is just a lazy way to let user use same base
     ++and avoid outputting upstream commits in cover-letter, and the side effect
     ++is that `b2..t2` on the right side will not be outputted.
     ++
      +
       --notes[=<ref>]::
       --no-notes::
     @@ t/t3206-range-diff.sh: test_expect_success '--left-only/--right-only' '
      +	git checkout my-feature &&
      +	git rebase $base --onto main &&
      +	tip="$(git rev-parse my-feature)" &&
     -+	git format-patch --range-diff $base $old $tip --cover-letter  &&
     -+	grep "> 1: .* feature$" 0000-cover-letter.patch &&
     -+	git format-patch --range-diff $base $old $tip --left-only --cover-letter &&
     ++	git format-patch --range-diff $base..$old $base..$tip --cover-letter  &&
     ++	grep "> 1: .* new$" 0000-cover-letter.patch &&
     ++	git format-patch --range-diff $base..$old $base..$tip --left-only --cover-letter &&
      +	! grep "> 1: .* feature$" 0000-cover-letter.patch
      +'
      +


 Documentation/git-format-patch.txt | 19 ++++++++++++++++++-
 builtin/log.c                      | 20 +++++++++++++++-----
 t/t3206-range-diff.sh              | 27 +++++++++++++++++++++++++++
 3 files changed, 60 insertions(+), 6 deletions(-)

diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt
index 3e49bf221087..99a2d87a859f 100644
--- a/Documentation/git-format-patch.txt
+++ b/Documentation/git-format-patch.txt
@@ -27,7 +27,7 @@ SYNOPSIS
 		   [--[no-]encode-email-headers]
 		   [--no-notes | --notes[=<ref>]]
 		   [--interdiff=<previous>]
-		   [--range-diff=<previous> [--creation-factor=<percent>]]
+		   [--range-diff=<previous> [--creation-factor=<percent>] [--left-only]]
 		   [--filename-max-length=<n>]
 		   [--progress]
 		   [<common diff options>]
@@ -301,6 +301,23 @@ material (this may change in the future).
 	creation/deletion cost fudge factor. See linkgit:git-range-diff[1])
 	for details.
 
+--left-only:
+	Used with `--range-diff`, only emit output related to the first range.
+	This option only can be used in this situation: The first iteration `t1`
+	based on the upstream commit `b1`, the second iteration `t2` based on the
+	upstream commit `b2`, `b2` inherited from the `b1`, then If the user want
+	output the range-diff between this two iterations (t1 and t2) in cover-letter,
+	and just use `b1` as the same base for `--range-diff`, but they don't want the
+	upstream range `b1..b2` included on the right side of the range-diff output.
+	By using `git format-patch --range-diff=b1..t1 b1..t2 --cover-letter --left-only`,
+	all content on the right side will be removed, leaving only the range
+	`b1..t1` on the left side.
++
+Note that this `--left-only` is just a lazy way to let user use same base
+and avoid outputting upstream commits in cover-letter, and the side effect
+is that `b2..t2` on the right side will not be outputted.
+
+
 --notes[=<ref>]::
 --no-notes::
 	Append the notes (see linkgit:git-notes[1]) for the commit
diff --git a/builtin/log.c b/builtin/log.c
index f67b67d80ed1..21fed9db82d6 100644
--- a/builtin/log.c
+++ b/builtin/log.c
@@ -1153,7 +1153,7 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			      struct commit *origin,
 			      int nr, struct commit **list,
 			      const char *branch_name,
-			      int quiet)
+			      int quiet, int left_only)
 {
 	const char *committer;
 	struct shortlog log;
@@ -1228,7 +1228,8 @@ static void make_cover_letter(struct rev_info *rev, int use_separate_file,
 			.creation_factor = rev->creation_factor,
 			.dual_color = 1,
 			.diffopt = &opts,
-			.other_arg = &other_arg
+			.other_arg = &other_arg,
+			.left_only = left_only
 		};
 
 		diff_setup(&opts);
@@ -1732,6 +1733,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 	struct strbuf rdiff2 = STRBUF_INIT;
 	struct strbuf rdiff_title = STRBUF_INIT;
 	int creation_factor = -1;
+	int left_only = 0;
 
 	const struct option builtin_format_patch_options[] = {
 		OPT_CALLBACK_F('n', "numbered", &numbered, NULL,
@@ -1814,6 +1816,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 			     parse_opt_object_name),
 		OPT_STRING(0, "range-diff", &rdiff_prev, N_("refspec"),
 			   N_("show changes against <refspec> in cover letter or single patch")),
+		OPT_BOOL(0, "left-only", &left_only,
+			 N_("only emit output related to the first range")),
 		OPT_INTEGER(0, "creation-factor", &creation_factor,
 			    N_("percentage by which creation is weighted")),
 		OPT_END()
@@ -2083,10 +2087,15 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 					     _("Interdiff against v%d:"));
 	}
 
+	if (!rdiff_prev) {
+		if (creation_factor >= 0)
+			die(_("--creation-factor requires --range-diff"));
+		if (left_only)
+			die(_("--left-only requires --range-diff"));
+	}
+
 	if (creation_factor < 0)
 		creation_factor = RANGE_DIFF_CREATION_FACTOR_DEFAULT;
-	else if (!rdiff_prev)
-		die(_("--creation-factor requires --range-diff"));
 
 	if (rdiff_prev) {
 		if (!cover_letter && total != 1)
@@ -2134,7 +2143,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
 		if (thread)
 			gen_message_id(&rev, "cover");
 		make_cover_letter(&rev, !!output_directory,
-				  origin, nr, list, branch_name, quiet);
+				  origin, nr, list, branch_name, quiet,
+				  left_only);
 		print_bases(&bases, rev.diffopt.file);
 		print_signature(rev.diffopt.file);
 		total++;
diff --git a/t/t3206-range-diff.sh b/t/t3206-range-diff.sh
index 1b26c4c2ef91..8e537793947b 100755
--- a/t/t3206-range-diff.sh
+++ b/t/t3206-range-diff.sh
@@ -748,4 +748,31 @@ test_expect_success '--left-only/--right-only' '
 	test_cmp expect actual
 '
 
+test_expect_success 'format-patch --range-diff --left-only' '
+	rm -fr repo &&
+	git init repo &&
+	cd repo &&
+	git branch -M main &&
+	echo "base" >base &&
+	git add base &&
+	git commit -m "base" &&
+	git checkout -b my-feature &&
+	echo "feature" >feature &&
+	git add feature &&
+	git commit -m "feature" &&
+	base="$(git rev-parse main)" &&
+	old="$(git rev-parse my-feature)" &&
+	git checkout main &&
+	echo "other" >>base &&
+	git add base &&
+	git commit -m "new" &&
+	git checkout my-feature &&
+	git rebase $base --onto main &&
+	tip="$(git rev-parse my-feature)" &&
+	git format-patch --range-diff $base..$old $base..$tip --cover-letter  &&
+	grep "> 1: .* new$" 0000-cover-letter.patch &&
+	git format-patch --range-diff $base..$old $base..$tip --left-only --cover-letter &&
+	! grep "> 1: .* feature$" 0000-cover-letter.patch
+'
+
 test_done

base-commit: be7935ed8bff19f481b033d0d242c5d5f239ed50
-- 
gitgitgadget

^ permalink raw reply related	[flat|nested] 16+ messages in thread

end of thread, other threads:[~2021-03-14  8:11 UTC | newest]

Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-03-07  6:57 [PATCH] [GSOC][RFC] format-patch: pass --right-only to range-diff ZheNing Hu via GitGitGadget
2021-03-08 20:18 ` Taylor Blau
2021-03-09  7:28   ` ZheNing Hu
2021-03-12 21:55   ` Junio C Hamano
2021-03-12 22:09     ` Junio C Hamano
2021-03-09  8:33 ` [PATCH v2] [GSOC][RFC] format-patch: pass --left-only " ZheNing Hu via GitGitGadget
2021-03-09  9:00   ` Eric Sunshine
2021-03-09  9:35     ` ZheNing Hu
2021-03-09 10:28   ` [PATCH v3] " ZheNing Hu via GitGitGadget
2021-03-12 22:50     ` Junio C Hamano
2021-03-13  4:01       ` ZheNing Hu
2021-03-13 23:23         ` Junio C Hamano
2021-03-14  2:16           ` ZheNing Hu
2021-03-14  2:37             ` ZheNing Hu
2021-03-14  2:41               ` ZheNing Hu
2021-03-14  8:10     ` [PATCH v4] [GSOC] " ZheNing Hu via GitGitGadget

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 NNTP newsgroup(s).