In order to fix bug 873204 we need to either use the mozcrash library (which depends at least on mozfile and mozlog) or replicated its functionality. The latter seems like a dumb choice so we are stuck with the former. Options: Land this in github, this is the easy course but leads to a duplicate copy in m-c and we have to update it when necessary. Manually copy mozcrash and its bits from the source in m-c into the test package that gets used for m-c tests, this would mean we don't fix 873204 for the main Jetpack tree though. This is pretty easy too probably. Teach the two releng bits that our tests are run by where to find mozcrash, I have no idea if that is easy or hard.

We cheated for the other test harnesses: http://mxr.mozilla.org/mozilla-central/source/build/automation.py.in#27 Longer-term the plan is to get mozharness to install mozbase from the test package into the virtualenv it uses to run the test harnesses.

(In reply to Ted Mielczarek [:ted.mielczarek] from comment #1) > We cheated for the other test harnesses: > http://mxr.mozilla.org/mozilla-central/source/build/automation.py.in#27 > > Longer-term the plan is to get mozharness to install mozbase from the test > package into the virtualenv it uses to run the test harnesses. That's another option but probably wouldn't work for the Jetpack tree again. But maybe that is easier to solve as run_jetpack.py doesn't involve changing buildbot.

We probably need the upstream mozrunner to do this.

No longer needed singe mochitest-jetpack is working