| Summary: | [nrwt] Get rid of DriverProxy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Balazs Kelemen <kbalazs> | ||||||||
| Component: | Tools / Tests | Assignee: | Nobody <webkit-unassigned> | ||||||||
| Status: | REOPENED --- | ||||||||||
| Severity: | Normal | CC: | abarth, dglazkov, dpranke, tony, webkit.review.bot | ||||||||
| Priority: | P2 | Keywords: | NRWT | ||||||||
| Version: | 528+ (Nightly build) | ||||||||||
| Hardware: | Unspecified | ||||||||||
| OS: | Unspecified | ||||||||||
| Attachments: |
|
||||||||||
|
Description
Balazs Kelemen
2012-08-06 07:31:35 PDT
Created attachment 156690 [details]
WIP patch
WIP, because I have some problems with webkitpy tests, virtual tests are causing a key error exception with this.
Created attachment 156691 [details]
test-webkitpy output
This is my output of test-webkitpy with the patch. I don't really understand what's happening with these virtual test thingies.
Heh, yeah, Chromium uses the DriverProxy for virtual tests as well. A virtual test suite is basically a way to say "re-run the tets in directory X with command line flags Y and Z". We use them, for example, to test the compositing tests using both the regular and accelerated code paths.. You can get rid of the DriverProxy on other ports but not Chromium yet. since you've done the legwork, I'll try to get the chromium ports updated this week. Comment on attachment 156690 [details] WIP patch Attachment 156690 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/13446842 New failing tests: platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect.globalcomposite.html fast/loader/loadInProgress.html fast/loader/unload-form-post-about-blank.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect+fillRect.alpha0.html http/tests/xmlhttprequest/zero-length-response.html platform/chromium/virtual/gpu/fast/canvas/arc360.html fast/frames/cached-frame-counter.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect.negative.html platform/chromium/virtual/gpu/fast/canvas/2d.fillText.gradient.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect.nonfinite.html http/tests/security/script-crossorigin-loads-correctly.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.canvas.reference.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect.basic.html platform/chromium/virtual/gpu/fast/canvas/arc-crash.html platform/chromium/virtual/gpu/fast/canvas/DrawImageSinglePixelStretch.html fast/canvas/webgl/shader-precision-format.html platform/chromium/virtual/gpu/fast/canvas/2d.text.draw.fill.maxWidth.gradient.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect+fillRect.alpha0.5.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect.globalalpha.html platform/chromium/virtual/gpu/fast/canvas/2d.text.draw.fill.maxWidth.veryLarge.html platform/chromium/virtual/gpu/fast/canvas/2d.composite.globalAlpha.fillPath.html platform/chromium/virtual/gpu/fast/canvas/2d.text.draw.fill.maxWidth.negative.html platform/chromium/virtual/gpu/fast/canvas/access-zero-sized-canvas.html platform/chromium/virtual/gpu/fast/canvas/2d.text.draw.fill.maxWidth.verySmall.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.canvas.readonly.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect.clip.html platform/chromium/virtual/gpu/canvas/philip/tests/2d.clearRect+fillRect.basic.html Created attachment 157007 [details]
Archive of layout-test-results from gce-cr-linux-07
The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: gce-cr-linux-07 Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'> Platform: Linux-2.6.39-gcg-201203291735-x86_64-with-Ubuntu-10.04-lucid
Comment on attachment 156690 [details]
WIP patch
Looks like this is not ready for review.
indeed, I thought we had cleared that flag already. thanks! I'm not going to finish this, as Chromium use the feature and for other ports there should only be one (real) driver anyway. Also probably other ports will use virtual suites as well in the future. I assume we can close this. I still want to get rid of DriverProxy; we should be able to pass arbitrary command line arguments with each test, and it's a hack that we have to run multiple DRTs per worker and switch between them. |