Re: [PATCH] [GSOC] ref-filter: add %(notes) format atom
- From
ZheNing Hu <adlternative@gmail.com>
- Date
- May 2, 2021, 06:12 UTC
- Message-ID
- <CAOLTT8TGaVxxHJRF_-Dq0Gnd50_H-80C6zPhjQAQruiO59qw2A@mail.gmail.com>
- In-Reply-To
- <38083c1b-f92b-7624-67b7-3a5dcf213d79@web.de>
René Scharfe <l.s.r@web.de> 于2021年5月2日周日 上午3:21写道:
Show 30 quoted lines
> > Am 29.04.21 um 12:24 schrieb ZheNing Hu via GitGitGadget: > > From: ZheNing Hu <adlternative@gmail.com> > > > > Note that `--pretty="%N"` can view notes related > > to the commit. So add `%(notes)` to ref-filter > > seem is a good choice. This atom `%(notes)` view > > the notes associated with the ref, and support > > dereference. > > I can't judge the usefulness of this feature because I don't use notes. > > > Signed-off-by: ZheNing Hu <adlternative@gmail.com> > > --- > > [GSOC] ref-filter: add %(notes) format atom > > > > An important step in the GSOC project Use ref-filter formats in git > > cat-file is to integrate different format atoms into the ref-filter. > > Olga and Hariom have also made a lot of efforts in this area. > > This is done to replace the custom format parser in git cat-file with > ref-filter in order to reduce code duplication, right? > > > Currently, I noticed that there may be some format atoms in "pretty.c" > > that have not been migrated to ref-filter, such as --pretty="%N", > > --pretty="%(describe)". > > git cat-file doesn't support pretty format atoms, so I'm not sure I see > the connection here. >
At present, `git cat-file --batch`, `git log --pretty`, and `git for-each-ref --format` lack consistency.
The ambition of this project is to unify these different formats "%aN","%(authorname)" through an interface, maybe it's `ref-filter` or maybe a new one.
It is estimated that many places in need to be replaced and rebuilt, and it not just to reduce code duplication.
Note that `cat-file --batch` should be a superset of `ref-filter` and `--pretty`, because it supports all objects, while `ref-filter` only supports "commit" and "tag" two kinds of objects, and `--pretty` only supports "commit" one kind of object.
So it may be reasonable to provide `%(notes)` to "commit","tag" objects in `cat-file --batch` in the future, now integrate them into ref-filter firstly.
Show 77 quoted lines
> > So in this patch, I tried to migrate --pretty=%N to --format=%(notes).
> >
> > Hope this will be hopeful !!!
> >
> > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-944%2Fadlternative%2Fformat-notes-atom-v1
> > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-944/adlternative/format-notes-atom-v1
> > Pull-Request: https://github.com/gitgitgadget/git/pull/944
> >
> > Documentation/git-for-each-ref.txt | 4 ++++
> > ref-filter.c | 31 ++++++++++++++++++++++++++++--
> > t/t6300-for-each-ref.sh | 10 ++++++++++
> > 3 files changed, 43 insertions(+), 2 deletions(-)
> >
> > diff --git a/Documentation/git-for-each-ref.txt b/Documentation/git-for-each-ref.txt
> > index 2ae2478de706..07f037a16e13 100644
> > --- a/Documentation/git-for-each-ref.txt
> > +++ b/Documentation/git-for-each-ref.txt
> > @@ -139,6 +139,9 @@ deltabase::
> > given object, if it is stored as a delta. Otherwise it
> > expands to the null object name (all zeroes).
> >
> > +notes::
> > + The notes associated with the ref.
> > +
> > upstream::
> > The name of a local ref which can be considered ``upstream''
> > from the displayed ref. Respects `:short`, `:lstrip` and
> > @@ -302,6 +305,7 @@ git for-each-ref --count=3 --sort='-*authordate' \
> > Subject: %(*subject)
> > Date: %(*authordate)
> > Ref: %(*refname)
> > +Notes: %(*notes)
> >
> > %(*body)
> > ' 'refs/tags'
> > diff --git a/ref-filter.c b/ref-filter.c
> > index a0adb4551d87..42a5608a3056 100644
> > --- a/ref-filter.c
> > +++ b/ref-filter.c
> > @@ -23,6 +23,7 @@
> > #include "worktree.h"
> > #include "hashmap.h"
> > #include "strvec.h"
> > +#include "run-command.h"
> >
> > static struct ref_msg {
> > const char *gone;
> > @@ -506,6 +507,7 @@ static struct {
> > { "objectsize", SOURCE_OTHER, FIELD_ULONG, objectsize_atom_parser },
> > { "objectname", SOURCE_OTHER, FIELD_STR, oid_atom_parser },
> > { "deltabase", SOURCE_OTHER, FIELD_STR, deltabase_atom_parser },
> > + { "notes", SOURCE_OTHER, FIELD_STR },
> > { "tree", SOURCE_OBJ, FIELD_STR, oid_atom_parser },
> > { "parent", SOURCE_OBJ, FIELD_STR, oid_atom_parser },
> > { "numparent", SOURCE_OBJ, FIELD_ULONG },
> > @@ -953,6 +955,24 @@ static int grab_oid(const char *name, const char *field, const struct object_id
> > return 0;
> > }
> >
> > +static int grab_notes(const struct object_id *oid, struct atom_value *v)
> > +{
> > + struct child_process cmd = CHILD_PROCESS_INIT;
> > + struct strbuf out = STRBUF_INIT;
> > + struct strbuf err = STRBUF_INIT;
> > + const char *args[] = { "notes", "show", NULL };
> > + int ret;
> > +
> > + cmd.git_cmd = 1;
> > + strvec_pushv(&cmd.args, args);
> > + strvec_push(&cmd.args, oid_to_hex(oid));
> > + ret = pipe_command(&cmd, NULL, 0, &out, 0, &err, 0);
>
> Nice prototype. Is it possible to avoid the overhead of calling git
> notes as an external process by calling load_display_notes() once at
> parse time and format_display_notes() for each item? Would it cause
> conflicts for code that uses ref-filter and handles notes already?
>This sounds feasible, I will try if it can work.
Thanks! -- ZheNing Hu