Re: How many GNU Emacs committers?

[Tassilo Horn <tsdh@gnu.org>]

Eli Zaretskii <eliz@gnu.org> writes:

>> Back when emacs used CVS, only the actual committers were logged,
>> though they probably committed tons of changes others sent them per
>> mail but couldn't commit themselves.
>
> The names of the actual authors of the changes are always stated in
> the ChangeLog entries, and always have been.

Yes, sure.  But those are not analyzed by ohloh.net.

> So I think the ChangeLog files are a much more accurate source of this
> kind of information that the VCS logs, including for this reason:
>
>> And there are many emacs packages that have their own repositories
>> (Org, Gnus, ERC, etc.), and not all of them are synched on a
>> per-commit basis with emacs' repository, so parts of their
>> contributors are probably not included in the statistics, too.

Indeed, so here's a try:

  $ cat **/ChangeLog* | grep '^[0-9]' | cut -d" " -f1 --complement \
        | sed -e 's/(.*)//' |sed -e 's/<.*>//' | sed -e 's/^[ ]\+//' \
        | sed -e 's/[ ]\+$//' | sort | uniq | wc -l
  1736

There are some duplicates when people sometimes use, e.g., "Richard
Stallman" and sometimes "Richard M. Stallman", and people with the same
name count only once, but matching only for names seemed to be more
correct than email addresses where sometimes people have used a good
dozen over the years.

Bye,
Tassilo