(In reply to Nick Alexander :nalexander from bug 1260874 comment #4) > Re: questions in IRC about mozharness cfg, you found it. But the error > message suggests we could use a path directly, rather than a "shortname". > Try it, and if that's true, I'd much rather just refer to the file than > route through this mozharness nonsense. The message Nick refers to in comment 4 is: Whoops! '--custom-build-variant' was passed but an appropriate config file could not be determined. Tried using: 'builds/releng_sub_android_configs/64_api_15_frontend.py' but it was either not: -- a valid shortname: ['cross-debug', 'b2g-debug', 'graphene', 'code-coverage', 'x86', 'api-15-gradle-dependencies', 'api-15-partner-sample1', 'api-9', 'source', 'stat-and-debug', 'api-11-debug', 'asan-and-debug', 'api-15', 'api-11', 'mulet', 'api-9-debug', 'asan', 'api-11-partner-sample1', 'api-15-debug', 'tsan', 'horizon', 'debug', 'api-15-frontend', 'cross-opt'] -- a valid path in ['.', '/home/worker/workspace/build/src/testing/mozharness/scripts/../configs', '/home/worker/workspace/build/src/testing/mozharness/scripts/../../configs'] -- a valid variant for the given platform and bits. --- Depends on bug 1260874 because the code will conflict, not because there is a dependency in theory.
It appears we fail on this line : variant_cfg = BuildOptionParser.build_variants[build_variant] Because build_variant is not a key. And build_variant is: 'builds/releng_sub_android_configs/64_lint.py' Which is the value I directly put in for the MH_CUSTOM_BUILD_VARIANT_CFG . The command line arg was: --custom-build-variant-cfg=builds/releng_sub_android_configs/64_lint.py We could try: * using `%s` in the string, as it is when we use short-names  * changing the path (e.g. using the full path) : https://dxr.mozilla.org/mozilla-central/rev/e5a10bc7dac4ee2453d8319165c1f6578203eac7/testing/mozharness/mozharness/mozilla/building/buildbase.py#634 : https://hg.mozilla.org/try/rev/cf44e608ab0b62d0d97548629f87fc09128e8b4b : https://dxr.mozilla.org/mozilla-central/rev/e5a10bc7dac4ee2453d8319165c1f6578203eac7/testing/mozharness/mozharness/mozilla/building/buildbase.py#369
I also missed this push that should be after comment 4 but before comment 6 (but is not comment 5! ;): https://treeherder.mozilla.org/#/jobs?repo=try&revision=9743d0c153430c622ca71c560c0ef4f549ac4fe2 I prefixed the path with "/", just to see.
fwiw, this is too low priority to keep moving it forward – I'll wait for results from bug 1271342.
(In reply to Michael Comella (:mcomella) from comment #8) > fwiw, this is too low priority to keep moving it forward – I'll wait for > results from bug 1271342. see https://bugzilla.mozilla.org/show_bug.cgi?id=1271342#c1
(In reply to Jordan Lund (:jlund) from comment #9) > (In reply to Michael Comella (:mcomella) from comment #8) > > fwiw, this is too low priority to keep moving it forward – I'll wait for > > results from bug 1271342. > > see https://bugzilla.mozilla.org/show_bug.cgi?id=1271342#c1 Sounds like this is unlikely to happen for a while - unassigning.