list of FLA signees published - need scripting help for next step
Adriaan de Groot
groot at kde.org
Wed Sep 7 11:49:12 UTC 2016
On Wednesday, September 07, 2016 09:00:36 AM Lydia Pintscher wrote:
> Thanks to Ingo we now have a list of contributors who signed the FLA
> with KDE e.V. and who agreed to have their name published. You can
> find it on the FLA page at https://ev.kde.org/rules/fla.php. A few are
> still missing who signed it over the last days. We will add them soon.
> As a next step we need to figure out which part of our codebase is
> actually well-covered and which is not. Is there anyone who can hack
> up a script to give us a list of contributors who would significantly
> increase our coverage of important repositories by signing the FLA?
This is a job for Paul Adams. Or rather, it would be right up his alley.
Recently, token-level blaming has been developed by Daniel Berlin and others,
to give an even better view of copyright provenance of source (e.g. avoiding
pollution of provenance by re-indenting). That might be an intersting too to
Anyway, you are asking basically for a mapping filename -> (mapping
contributor-id -> number of lines) for data collection. Assuming we can
normalize on the contributor-ids, then filter the inner mapping on ids-who-are-
not-in-the-fla-signers-list and coalesce that at some level of granularity.
It's probably an extant script, or an hour with python.
Perhaps the biggest challenge is to get a decent list of which repo's need to
be covered (or is that in the meta-buildinfo?). Or to find the aliases of
signatories, e.g. in the akonadi repository (yeah, I had an hour + python on
the train), counting number of lines per committer-email-address (mangled
here), we get:
There's two of Dan there, and there are other duplicates too (hey, I have 1
line in the akonadi/ repo attributed to me!)
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 819 bytes
Desc: This is a digitally signed message part.
More information about the kde-community