Once SeaMonkey (and TB/Calendar) is using Hg, we need to get mxr up to date with a copy of our entire source tree (including mozilla-central too). As long as our primary Hg sources have links to the log/annotate for each file, we can probably live with (temporary) brokenness on the pulled CVS-files, and mozilla-central files links from that section of links. Steps to update source is (after initial clone): python client.py checkout

MXR will deal w/ cvs mixes automatically. http://mxr-test.landfill.bugzilla.org/mozilla-central/source/js/src/jscntxt.h#984 if you want to actually have both CVS and hg links for a file (as shown above), that requires a bit more work. note that your steps are missing the actual hg clone line (which is unfortunate). please file more complete bug reports :). For the most part, as long as the hg repo is on hg.mozilla.org, the hg links are probably also supposed to work automatically (not certain about this part).

comm-central support in MXR added in changeset 93:ca6033951ee7.

Dumb question: for searches, could it be possible to have 2 "roots", one for comm-central only, one for comm-central + mozilla-central ?

(In reply to comment #3) > Dumb question: > for searches, could it be possible to have 2 "roots", > one for comm-central only, one for comm-central + mozilla-central ? timeless? Anyway, this has been done. http://mxr.mozilla.org/comm-central/

supporting merged trees is on my todo list. it's decidedly nontrivial. in the interim, just setup two trees and make us pay to index twice.

reed, assuming cron is also setup as above, can we please get a one-time |hg pull -u| of the repo, as the initial files have now been pushed.

(In reply to comment #6) > reed, assuming cron is also setup as above, can we please get a one-time |hg > pull -u| of the repo, as the initial files have now been pushed. Done.