Re: [RFC PATCH 2/3] grep: prepare grep for submodules

On 10/1/10, Chris Packham [off-list ref] wrote:

 > You can make setup_explicit_git_dir() realize that situation (current
 > working directory outside $GIT_WORK_TREE), then calculate and save the
 > submodule prefix in startup_info struct. Then "git grep" or any
 > commands can just read startup_info to find out the submodule prefix.


Here's my first naive attempt at implementing what you describe. Needs
 tests, more comments, sign-off etc.

Thanks.

 One situation that could be handled better is when the cwd is a
 subdirectory of the specified worktree. At the moment this ends up
 giving the full path to the worktree, the output would look much nicer
 if it gave the relative path (e.g. ../../).

Hmm.. if cwd is inside a worktree, prefix (the 3rd parameter in
cmd_grep) should be correctly set and "git grep" should also show
relative path. Or are you talking about another command?

 ---8<---
 From: Chris Packham [off-list ref]
 Date: Thu, 30 Sep 2010 11:19:29 -0700
 Subject: [RFC PATCH] save the work tree prefix in startup_info

 This is the relative path between the cwd and the worktree or the
 absolute path of the worktree if the worktree is not a subdirectory
 of the worktree.
 ---
  cache.h |    1 +
  dir.c   |   26 ++++++++++++++++++++++++++
  dir.h   |    1 +
  setup.c |    4 ++++
  4 files changed, 32 insertions(+), 0 deletions(-)

 diff --git a/cache.h b/cache.h
 index e1d3ffd..f320e78 100644
 --- a/cache.h
 +++ b/cache.h
 @@ -1111,6 +1111,7 @@ const char *split_cmdline_strerror(int cmdline_errno);
  /* git.c */
  struct startup_info {
        int have_repository;
 +       const char *prefix;

You should use another name here to avoid confusion with the current
prefix, relative to worktree toplevel directory. I'm thinking of
outer_prefix or cwd_prefix, but I'm usually bad at naming.

  };
  extern struct startup_info *startup_info;

 diff --git a/dir.c b/dir.c
 index 58ec1a1..2148730 100644
 --- a/dir.c
 +++ b/dir.c
 @@ -1036,6 +1036,32 @@ char *get_relative_cwd(char *buffer, int size,
 const char *dir)
        }
  }

 +char *get_relative_wt(char *buffer, int size, const char *dir)
 +{
 +       char *cwd = buffer;
 +
 +       if (!dir)
 +               return NULL;
 +       if (!getcwd(buffer, size))
 +               die_errno("can't find the current directory");
 +       if (!is_absolute_path(dir))
 +               dir = make_absolute_path(dir);
 +       if (strstr(dir, cwd)) {

Why not strncmp?

 +               dir += strlen(cwd);
 +               switch(*dir){
 +               case '\0':
 +                       return NULL;
 +               case '/':
 +                       dir++;
 +                       break;

Yeah.

 +               default:
 +                       break;

Should we properly handle relative path that includes ".." here too?

 +               }
 +       }
 +       strncpy(buffer, dir, size);

So if "cwd" is inside "dir", an absolute "dir" is returned? That does
not look like a prefix to me.

-- 
Duy