613269 - Some identifiers which are recognized on mxr are not recognized on mxr-test

Status: RESOLVED FIXED

(Webtools Graveyard :: MXR, defect)

Description

[(no longer active)]

For example, see:

<http://mxr.mozilla.org/mozilla-central/ident?i=RegisterCommand>
<http://mxr-test.konigsberg.mozilla.org/mozilla-central/ident?i=RegisterCommand>

Comment 1

[timeless]

so, mxr-test went crazy after nov 13. I killed all the jobs last night.

mxr-test also has an experimental indexer available, but I don't think it's supposed to be active.

I'm running an update-xref (using cron commands) now,

the last successful run start and end times:
Sat Nov 13 03:13:33 CET 2010
...
Sat Nov 13 04:24:54 CET 2010

so it should take ~70mins for the run to complete.

Comment 2

[(no longer active)]

Hmm, the problem still remains the same, as far as I can tell...

Comment 3

[timeless]

yeah, it didn't finish, so it just started getting stuck again.

i turned off the "silence logging" flag and ran again.

genxref on mozilla-central got stuck on this file:
js/src/tests/js1_5/Regress/regress-229006.js
... which is rather mean, but mxr has dealt w/ it for years, so I'm not sure if that's really what's going on, i've filed that as bug 614207. And I'll investigate after lunch.

Comment 4

[timeless]

ok, this is a bit strange. the fix in bug 614207 will fix the underlying problem.
since this bug was more of a "something's wrong" and I split out the underlying problem, I'm going to resolve this bug as fixed first.

A quick check shows that both mozilla-central and mozilla-1.9.1 managed to finish their indexing tasks.

I'll resolve the other bug probably Sunday when I push the fix upstream. -- Possibly including a couple of other fixes.

Note that I'm restoring the patch that I disabled initially, so there should be another xref run in roughly 8 hours, if things go bad, i'll know tomorrow when I come into work.