summaryrefslogtreecommitdiff
path: root/README
diff options
context:
space:
mode:
Diffstat (limited to 'README')
-rw-r--r--README49
1 files changed, 49 insertions, 0 deletions
diff --git a/README b/README
new file mode 100644
index 00000000..68c970f6
--- /dev/null
+++ b/README
@@ -0,0 +1,49 @@
+This is the configuration we use to extract changesets and run gitdm on them.
+
+* Tweaking your affiliation
+
+This mostly happens in domain_map; there should be no need to touch
+the 'oracle' and 'go-oo' files, which tag legacy account affiliation.
+
+* Running mkstats.sh
+
+To run this, you will also need a patched gitdm next to this repo;
+something like this:
+
+cd ..
+git clone git://git.lwn.net/gitdm.git
+cd gitdm
+for p in ../gitdm-config/gitdm-patches/*; do
+ patch -p1 < $p
+done
+
+Then - simply run:
+
+./mkstats.sh
+
+(which will clone a lot of repos under ./checkouts, make sure the
+filesystem has some GB of space)
+
+This no longer handles the old split-repo source setup, which was ~impossible
+to extract accurate statistics out of, anyway. As an upshot, it now
+handles almost every git repo that is under the TDF umbrella.
+
+Having got that right - you should end up with a stats-$year-$week
+directory containing the required data.
+
+ Best of luck interpreting it.
+
+Authors:
+ Cedric Bosdonnat
+ Michael Meeks
+ Thorsten Behrens
+
+
+
+File information:
+
+domain-name, contains all companies who have a blanket license
+knownn-licensee.csv, contains all known licensee (from email, wiki, legal@ or blanket license)
+audit, contains the audit history
+developers-dump.json, is extended information, needed by the esc-reporting tool
+