Re: [GitStats] Bling bling or some statistics on the git.git repository

[Johannes Schindelin <Johannes.Schindelin@gmx.de>]

Hi,

On Sat, 12 Jul 2008, Sverre Rabbelier wrote:

> On Fri, Jul 11, 2008 at 11:55 PM, Johannes Schindelin
> <Johannes.Schindelin@gmx.de> wrote:
>
> > On Fri, 11 Jul 2008, Johannes Schindelin wrote:
> >
> >> something else I just realized: you might want to use .mailmap, e.g. 
> >> to coalesce the changes of Shawn "O." Pearce correctly.
> >
> > Yet another thing: while it is true that git-gui is usually pulled in 
> > (with the subtree strategy), some parts were changed in git.git 
> > directly, so you will need to cope with the wholesale rename with 
> > every merge.
> 
> I think I understand the cause, but I'm not sure I understand the
> consequences? Will some files at some times appear to be located at
> /path/to/file.txt and at other times at /subdir/path/to/file.txt? If
> so, how could I possibly handle that? How do I know that when it says
> /path/to/file.txt it means /subdir/path/to/file.txt instead?
>
> > Besides, it is slightly distracting to see the file names differently 
> > from what they are in HEAD^{tree}.  But that may be just me.
> 
> Different how? In that it shows the contents of subdirs? For at least
> my repo this is a Very Good Thing (tm) as all my source-code is in a
> /src directory. If all changes to subdirs were aggregated I wouldn't
> get any useful metrics at all. Or am I misunderstanding the difference
> between the current output and that of HEAD^{tree}?

What I meant is this: you list "git-gui.sh", while I would have expected 
"git-gui/git-gui.sh".

So you might need to rename the keys of your aggregating dict (I haven't 
looked at the code, but suspect that's how you do it) with renaming 
commits (including the merge commits).

Ciao,
Dscho