Gentoo Archives: gentoo-dev

From: "Michał Górny" <mgorny@g.o>
To: gentoo-dev@l.g.o
Subject: [gentoo-dev] Stats: Gentoo developer commit timeline
Date: Mon, 16 Jan 2017 09:35:47
1 Hi, everyone.
3 Just a quick side project we've done a while ago. It's a timeline of
4 developer commit activity [1]. Code for data processing in [2]. I did
5 the data, Amynka prepared a nice JS to graph it.
7 This version is running purely on git + historical cvs->git conversion.
8 Dates are done using committer timestamp (so they may shift a little
9 when people fail to set times), authors are mapped by author +
10 committer emails. This means that two people can get one commit
11 attributed -- since the goal is only to check activity, not attribute
12 work.
14 The emails are further mapped to developer accounts either via LDAP
15 (mail + gentooAlias), or via stripping ''. People outside
16 Gentoo are skipped since it's hard to track their email changes.
18 The graphs are interrupted if a developer had no commits for 90 days.
19 This is meant to visualize returning developers better -- though I
20 guess it more often points out developers who rarely commit at all ;-).
21 Possibly it also emphasizes the cases when developers slowly lost
22 interest in Gentoo.
24 Of course, it's all approximate. Non-commit activity is not included.
25 Some of more recent developers are listed 'earlier' due to their
26 proxy-maint activity. Some of the data may be inconsistent, some of
27 proxy-maint work may be missing (if we don't have mapping from old
28 e-mail address), some developer may be duplicated (if we don't have
29 mapping for their username changes).
31 To avoid any disputable proxy-maint data, I could go for committer
32 addresses only + --first-parent to skip over merge commits. However, I
33 know that a few developers failed at merge commits in the past, so that
34 would also skip some legitimate data.
36 [1]:
37 [2]:
39 --
40 Best regards,
41 Michał Górny
42 <>