Question

I moved a project from svn to git. I used git-filter-branch with –index-filter to remove all the unnecessary directories, and pruned all empty commits with the –prune-empty option.

After I recloned the repository, I am still seeing these strange commits with paths pointing to a ancestor directory. This is best explained by the output of git log –stat

commit 222222b80e791e4ef5b9a027c8b10f64be5e2222
Author: someuser <someuser@somedomain>
Date:   Mon Sep 14 17:52:08 2009

    some commit msg

 hello/.project                                |   11 ++++
 hello/bin/hello.fla                           |  Bin 0 -> 43008 bytes
 .../other/Plugin.as                           |   10 ++++
 .../other/Constants.as                        |    2 +
 hello/Base.as                                 |   52 ++++++++++++++++++++
 .../other/HelloWorld.as                       |   21 ++++++++
 .../target/Hello.swf                          |  Bin 0 -> 1039 bytes
 7 files changed, 96 insertions(+), 0 deletions(-)

How does this happen, and how do I fix this?

Btw, the command was run at the root of the git repository.

Answer 1

Those aren’t parent directories. git is abbreviating longer pathnames to ... because it doesn’t have room to display them (presumably).

From the git log --help manual page:

   --stat[=<width>[,<name-width>[,<count>]]]
       Generate a diffstat. You can override the default output width
       for 80-column terminal by --stat=<width>. The width of the
       filename part can be controlled by giving another width to it
       separated by a comma. By giving a third parameter <count>, you
       can limit the output to the first <count> lines, followed by
       ...  if there are more.

       These parameters can also be set individually with
       --stat-width=<width>, --stat-name-width=<name-width> and
       --stat-count=<count>.

Try resizing your terminal to 120 characters wide, and then do this:

git log --stat=120,80

You should have plenty of room to see the full pathnames then.