Bug 849103 will be enabling IonMonkey for B2G. That plus bug 851450 will mean we can turn off the Android NoIon builds on all trees on which those bugs have landed, since the Android NoIon builds are only being used as a substitute for B2G testing. See bug bug 849103 and bug 851450 comment 1 for more info.

As I understand it, we can proceed with turning off these builds+tests.

My understanding is that we need these until b2g18 is no longer supported, which isn't until March of 2014.

We can definitely turn off any NoIon builds that are currently being done on the m-c branch. Only the b2g18 branch needs NoIon testing.

When I brought this up before, I was told that we need m-c NoIon test coverage for any potential uplifts to b2g18. Not that I'd be heartbroken about it personally, but that's the reason I'd been given for why they needed to stay on in the past.

I think we can live without that. NoIon specific breakage seems pretty unlikely at this point. Unlikely enough that testing on b2g18 branch seems enough.

(In reply to Jonas Sicking (:sicking) from comment #5) > I think we can live without that. NoIon specific breakage seems pretty > unlikely at this point. Unlikely enough that testing on b2g18 branch seems > enough. Given that, do we need it on try (note try has only been expected to follow m-c, and is known to break on occasion if pushing a different branch to it, though usually try to prevent that)?

b2g18 is already severely broken on Try, what's one more?

on m-c NoIon means no JIT since all the remaining JIT compilers are guarded by the condition that Ion is enabled. This only case which is served only by the NoIon builds, is to help people which such as PowerPC (not tier1) people who are looking to disable it until they are able to provide a new assembler backend to IonMonkey. So, I think we can safely remove all the tests, and if we want to support ports of the JS engine to other architectures (which is not a huge concern), we can still have a NoIon build, which does not need to be an Android build. Concerning b2g18, we always compile b2g with Ion being disabled, which only leaves JM. It does not make any sense to keep anything else than NoIon builds (Androidn NoIon & B2G) from the perspective of the JIT, as we are no shipping any b2g18 based on IonMonkey.

Given comment 5, hidden everywhere where we run it (which is only actually mozilla-central, inbound, b2g-inbound and try), other than b2g18. Now that it's hidden, will someone believe us when we say it's not needed? :-)

The android-noion is in the locked platform list for b2g18, so believe taking it out of the default platforms list should just work.

(I hope :-))

Comment on attachment 803789 [details] [diff] [review] Turn off Android NoIon builds everywhere apart from mozilla-b2g18 Switching per IRC :-)

Comment on attachment 803789 [details] [diff] [review] Turn off Android NoIon builds everywhere apart from mozilla-b2g18 in mozilla-tests/mobile_config.py PLATFORMS = { 'android': {}, 'android-armv6': {}, - 'android-noion': {}, 'android-x86': {}, } needs to stay and so does BRANCH_UNITTEST_VARS = { 'hghost': 'hg.mozilla.org', # turn on platforms as we get them running 'platforms': { 'android': {}, 'android-armv6': {}, - 'android-noion': {}, 'android-debug': {}, 'android-x86': {}, }, } for this loop for branch in BRANCHES: - if branch not in ('mozilla-central', 'mozilla-inbound', 'mozilla-b2g18', - 'mozilla-b2g18_v1_0_1', 'mozilla-b2g18_v1_1_0_hd', 'try', - 'b2g-inbound', 'date', - ): - if 'android-noion' in BRANCHES[branch]['platforms']: - del BRANCHES[branch]['platforms']['android-noion'] - it needs to stay but only include the 'mozilla-b2g18' branch. I think similar issues apply to the mozilla/config.py file, I didn't test them in staging, just the mobile ones.

Turn off Android NoIon builds everywhere apart from mozilla-b2g18* Hopefully this is correct now :-)

I'll test this tomorrow, I didn't have time today with other bugs + summit on the weekend, sorry for the delay

No problem at all, I figured it would take a few days for people to get up to speed after the summit (I know it will for me!) :-)

Comment on attachment 813575 [details] [diff] [review] Patch v2 Thank you for the review :-) https://hg.mozilla.org/build/buildbot-configs/rev/146c3d5a1096

in production