[v7,0/3] builtin/ls-files.c:add git ls-file --dedup option

Message

Philippe Blain via GitGitGadget Jan. 24, 2021, 10:54 a.m. UTC

I am reading the source code of git ls-files and learned that git ls-files
may have duplicate files name when there are unmerged path in a branch merge
or when different options are used at the same time. Users may fell confuse
when they see these duplicate file names.

As Junio C Hamano said ,it have odd behaviour.

Therefore, we can provide an additional option to git ls-files to delete
those repeated information.

This fixes https://github.com/gitgitgadget/git/issues/198

Thanks!

ZheNing Hu (3):
  ls_files.c: bugfix for --deleted and --modified
  ls_files.c: consolidate two for loops into one
  ls-files.c: add --deduplicate option

 Documentation/git-ls-files.txt |  8 ++++
 builtin/ls-files.c             | 85 ++++++++++++++++++++--------------
 t/t3012-ls-files-dedup.sh      | 66 ++++++++++++++++++++++++++
 3 files changed, 124 insertions(+), 35 deletions(-)
 create mode 100755 t/t3012-ls-files-dedup.sh


base-commit: 6d3ef5b467eccd2769f1aa1c555d317d3c8dc707
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-832%2Fadlternative%2Fls-files-dedup-v7
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-832/adlternative/ls-files-dedup-v7
Pull-Request: https://github.com/gitgitgadget/git/pull/832

Range-diff vs v6:

 1:  fbc38ce9075 ! 1:  8b02367a359 ls_files.c: bugfix for --deleted and --modified
     @@ Commit message
          ie_modified() when lstat() has failed.
      
          Signed-off-by: ZheNing Hu <[e-mail]>
     +    [jc: fixed misindented code]
     +    Signed-off-by: Junio C Hamano <[e-mail]>
      
       ## builtin/ls-files.c ##
      @@ builtin/ls-files.c: static void show_files(struct repository *repo, struct dir_struct *dir)
 2:  3997d390883 ! 2:  f9d5e44d2c0 ls_files.c: consolidate two for loops into one
     @@ Metadata
       ## Commit message ##
          ls_files.c: consolidate two for loops into one
      
     -    Refactor the two for loops into one,skip showing the ce if it
     -    has the same name as the previously shown one, only when doing so
     -    won't lose information.
     +    This will make it easier to show only one entry per filename in the
     +    next step.
      
          Signed-off-by: ZheNing Hu <[e-mail]>
     +    [jc: corrected the log message]
     +    Signed-off-by: Junio C Hamano <[e-mail]>
      
       ## builtin/ls-files.c ##
      @@ builtin/ls-files.c: static void show_files(struct repository *repo, struct dir_struct *dir)
 3:  07b603fd97c ! 3:  384f77a4c18 ls-files.c: add --deduplicate option
     @@ Metadata
       ## Commit message ##
          ls-files.c: add --deduplicate option
      
     -    In order to provide users a better experience
     -    when viewing information about files in the index
     -    and the working tree, the `--deduplicate` option will suppress
     -    some duplicate name under some conditions.
     +    During a merge conflict, the name of a file may appear multiple
     +    times in "git ls-files" output, once for each stage.  If you use
     +    both `--delete` and `--modify` at the same time, the output may
     +    mention a deleted file twice.
      
     -    In a merge conflict, one file name of "git ls-files" output may
     -    appear multiple times. For example,now there is an unmerged path
     -    `a.c`,`a.c` will appear three times in the output of
     -    "git ls-files".We can use "git ls-files --deduplicate" to output
     -    `a.c` only one time.(unless `--stage` or `--unmerged` is
     -    used to view all the detailed information in the index)
     +    When none of the '-t', '-u', or '-s' options is in use, these
     +    duplicate entries do not add much value to the output.
      
     -    In addition, if you use both `--delete` and `--modify` at
     -    the same time, The `--deduplicate` option
     -    can also suppress file name output.
     -
     -    Additional instructions:
     -    In order to display entries information,`deduplicate` suppresses
     -    the output of duplicate file names, not the output of duplicate
     -    entries information, so under the option of `-t`, `--stage`, `--unmerge`,
     -    `--deduplicate` will have no effect.
     +    Introduce a new '--deduplicate' option to suppress them.
      
          Signed-off-by: ZheNing Hu <[e-mail]>
     +    [jc: extended doc and rewritten commit log]
     +    Signed-off-by: Junio C Hamano <[e-mail]>
      
       ## Documentation/git-ls-files.txt ##
      @@ Documentation/git-ls-files.txt: SYNOPSIS
     @@ Documentation/git-ls-files.txt: OPTIONS
       	See OUTPUT below for more information.
       
      +--deduplicate::
     -+	Suppress duplicate entries when there are unmerged paths in index
     -+	or `--deleted` and `--modified` are combined.
     ++	When only filenames are shown, suppress duplicates that may
     ++	come from having multiple stages during a merge, or giving
     ++	`--deleted` and `--modified` option at the same time.
     ++	When any of the `-t`, `--unmerged`, or `--stage` option is
     ++	in use, this option has no effect.
      +
       -x <pattern>::
       --exclude=<pattern>::