Closed
Bug 1456790
Opened 6 years ago
Closed 6 years ago
"AttributeError: 'GeckoDriverProtocol' object has no attribute 'base'" after "Pausing until the browser exits"
Categories
(Testing :: web-platform-tests, defect)
Testing
web-platform-tests
Tracking
(firefox60 unaffected, firefox61 affected)
RESOLVED
WORKSFORME
Tracking | Status | |
---|---|---|
firefox60 | --- | unaffected |
firefox61 | --- | affected |
People
(Reporter: whimboo, Unassigned)
References
Details
(Keywords: regression)
There are 5 failures, but maybe due to the wptrunner failures the final report says that the run was ok: ======================================================= 5 failed, 1 warnings in 13.40 seconds ======================================================= 0:19.82 TEST_END: Test OK. Subtests passed 0/5. Unexpected 0 0:19.82 INFO Pausing until the browser exits 0:19.82 ERROR Error running command wait with arguments (): Traceback (most recent call last): File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 91, in run rv = commands[command](*args) File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 111, in wait self.executor.wait() File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/executors/base.py", line 204, in wait self.protocol.base.wait() AttributeError: 'GeckoDriverProtocol' object has no attribute 'base' 0:19.83 ERROR Traceback (most recent call last): File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 360, in run new_state = self.wait_event() File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 435, in wait_event return f(*data) File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 637, in error self.restart_runner() File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 628, in restart_runner assert isinstance(self.state, RunnerManagerState.restarting) AssertionError 0:19.96 WARNING u'runner_teardown': () Exception in thread Thread-TestrunnerManager-1: Traceback (most recent call last): File "/Volumes/data/Users/henrik/.pyenv/versions/2.7.14/lib/python2.7/threading.py", line 801, in __bootstrap_inner self.run() File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 360, in run new_state = self.wait_event() File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 435, in wait_event return f(*data) File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 637, in error self.restart_runner() File "/Volumes/data/code/gecko/testing/web-platform/tests/tools/wptrunner/wptrunner/testrunner.py", line 628, in restart_runner assert isinstance(self.state, RunnerManagerState.restarting) AssertionError 0:19.96 INFO Running testharness tests 0:19.97 INFO No testharness tests to run 0:19.97 INFO Got 0 unexpected results 0:19.97 SUITE_END web-platform-test ~~~~~~~~~~~~~~~~~ Ran 6 checks (1 tests, 5 subtests) Expected results: 6 OK James I mentioned that already a couple of days ago. But it hasn't been changed for me. So a fix you were referring to may not have fixed that?
Flags: needinfo?(james)
Comment 1•6 years ago
|
||
This is a different bug; it's trying to pause after the test is done, but that doesn't make sense for wdspec tests. I thought I fixed that, but maybe not.
Flags: needinfo?(james)
Reporter | ||
Comment 2•6 years ago
|
||
So the uncaught failures in this case are wrongly marked as fail. I filed bug 1456799 to get this fixed.
Reporter | ||
Updated•6 years ago
|
Summary: Broken final test result reporting: ok, while tests are failing → "AttributeError: 'GeckoDriverProtocol' object has no attribute 'base'" after "Pausing until the browser exits"
Reporter | ||
Comment 3•6 years ago
|
||
Looks similar to the MarionetteProtocol as fixed on bug 1454471.
See Also: → 1454471
Reporter | ||
Comment 4•6 years ago
|
||
In case of failure the following executor and protocol is in use: > wptrunner.executors.executormarionette.GeckoDriverProtocol As I can see the `GeckoDriverProtocol` inherits from `WebDriverProtocol` which itself inherits from `Protocol`. None of those classes seem to set `self.base` and as such you will get an error like: > 'GeckoDriverProtocol' object has no attribute 'base' There are no implementations defined for `WebDriverProtocol`, which might explain why we see this problem? Given that I haven't seen this a while ago I suspect this should be somewhat a recent regression. I will have a look if I can figure that out. James, anything you know of what has been changed in that area?
Flags: needinfo?(james)
Keywords: regression,
regressionwindow-wanted
Reporter | ||
Comment 5•6 years ago
|
||
The regression started to happen in this merge from inbound to central: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=e96685584bf7&tochange=6276ec7ebbf3
Reporter | ||
Comment 6•6 years ago
|
||
So the regression bug is bug 1440412.
Blocks: 1440412
Keywords: regressionwindow-wanted
Reporter | ||
Comment 7•6 years ago
|
||
Also flagging Geoffrey who implemented that patch.
Flags: needinfo?(geoffers+mozilla)
Reporter | ||
Updated•6 years ago
|
status-firefox60:
--- → unaffected
status-firefox61:
--- → affected
Comment 8•6 years ago
|
||
I didn't implement that patch? :jgraham did? I just reviewed it.
Flags: needinfo?(geoffers+mozilla)
Reporter | ||
Comment 9•6 years ago
|
||
Sorry, you are correct. James, would you mind having a look at this bug?
Comment 10•6 years ago
|
||
I thought this was fixed, and now I can't reproduce it, so I think it must be.
Status: NEW → RESOLVED
Closed: 6 years ago
Flags: needinfo?(james)
Resolution: --- → WORKSFORME
Reporter | ||
Comment 11•6 years ago
|
||
(In reply to James Graham [:jgraham] from comment #10) > I thought this was fixed, and now I can't reproduce it, so I think it must > be. Ok, so fixed in terms of a github PR which then got downstreamed in one of the latest syncs into mozilla-central. It also works for me now.
You need to log in
before you can comment on or make changes to this bug.
Description
•