Ah, shoot. Sorry for the trouble Michael.

This is a new harness, so it's possible that the old bug crept back in. Do you have a pointer to the old Telemetry bug?

On macOS, according to the logs from -v, the harness thinks it's passing down all of those arguments when starting the browser. Note however that webgl_conformance_integration_test.py already includes the flag:

--js-flags=--expose-gc

What happens if --js-flags is passed multiple times? Do they accumulate?

Is this a bug just on Windows?

See here:

./content/test/gpu/run_gpu_integration_test.py webgl_conformance --test-filter=conformance_attribs_gl_bindAttribLocation_aliasing -v --extra-browser-args="--js-flags=\"--trace-gc --trace-incremental-marking\" "

result:

(INFO) 2016-12-19 21:04:27,852 desktop_browser_backend.Start:292  Starting Chrome ['/Users/kbr/src/chrome/src/out/Release/Chromium.app/Contents/MacOS/Chromium', '--test-type=gpu', '--disable-domain-blocking-for-3d-apis', '--disable-gesture-requirement-for-media-playback', '--disable-gpu-process-crash-limit', '--enable-experimental-canvas-features', '--js-flags=--expose-gc', '--disable-gpu-watchdog', '--js-flags=--trace-gc --trace-incremental-marking', '--enable-net-benchmarking', '--metrics-recording-only', '--no-default-browser-check', '--no-first-run', '--enable-gpu-benchmarking', '--disable-background-networking', '--proxy-server=socks://localhost:56041', '--ignore-certificate-errors', '--disable-component-extensions-with-background-pages', '--disable-default-apps', '--remote-debugging-port=0', '--enable-crash-reporter-for-testing', '--window-size=1280,1024', u'--ppapi-flash-path=/Users/kbr/src/chrome/src/third_party/adobe/flash/binaries/ppapi/mac_64/PepperFlashPlayer.plugin', '--ppapi-flash-version=99.9.999.999', '--user-data-dir=/var/folders/pk/pkx9qlgj4z52ftmp_s0hng14002z4b/T/tmpghv5Re', 'about:blank']



and here:

./content/test/gpu/run_gpu_integration_test.py webgl_conformance --test-filter=conformance_attribs_gl_bindAttribLocation_aliasing -v --extra-browser-args="--js-flags=\"--trace-gc --trace-incremental-marking --use-allocation-folding\" "

result:

(INFO) 2016-12-19 21:05:17,414 desktop_browser_backend.Start:292  Starting Chrome ['/Users/kbr/src/chrome/src/out/Release/Chromium.app/Contents/MacOS/Chromium', '--test-type=gpu', '--disable-domain-blocking-for-3d-apis', '--disable-gesture-requirement-for-media-playback', '--disable-gpu-process-crash-limit', '--enable-experimental-canvas-features', '--js-flags=--expose-gc', '--js-flags=--trace-gc --trace-incremental-marking --use-allocation-folding', '--disable-gpu-watchdog', '--enable-net-benchmarking', '--metrics-recording-only', '--no-default-browser-check', '--no-first-run', '--enable-gpu-benchmarking', '--disable-background-networking', '--proxy-server=socks://localhost:56105', '--ignore-certificate-errors', '--disable-component-extensions-with-background-pages', '--disable-default-apps', '--remote-debugging-port=0', '--enable-crash-reporter-for-testing', '--window-size=1280,1024', u'--ppapi-flash-path=/Users/kbr/src/chrome/src/third_party/adobe/flash/binaries/ppapi/mac_64/PepperFlashPlayer.plugin', '--ppapi-flash-version=99.9.999.999', '--user-data-dir=/var/folders/pk/pkx9qlgj4z52ftmp_s0hng14002z4b/T/tmpXAaHtw', 'about:blank']

I observed this on Windows, yes.

I already reported this in 521560, which got fast initial responses and then didn't see much love :)

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/25da6356a3246812d79aaedcb236e1545ed88c95

commit 25da6356a3246812d79aaedcb236e1545ed88c95
Author: kbr <kbr@chromium.org>
Date: Wed Jan 04 18:15:03 2017

Support --js-flags via --extra-browser-args in WebGL conformance tests.

It's impractical to merge them, so instead, allow overriding them.

Tested manually on Windows.

BUG= 675582 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.android:android_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2617453002
Cr-Commit-Position: refs/heads/master@{#441410}

[modify] https://crrev.com/25da6356a3246812d79aaedcb236e1545ed88c95/content/test/gpu/gpu_tests/webgl_conformance_integration_test.py

This is fixed for this particular test suite above. I don't know how to fix this more generally for the perf tests (Issue 521560), but if there are any other gpu_integration_tests with this problem, please file a new bug and I'll generalize the fix.