Bug 199844 - Make ANGLE work inside WebKit2's sandbox
Summary: Make ANGLE work inside WebKit2's sandbox
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebGL (show other bugs)
Version: WebKit Nightly Build
Hardware: Mac All
: P2 Normal
Assignee: Kenneth Russell
URL:
Keywords: InRadar
Depends on:
Blocks: webglangle
  Show dependency tree / graph
 
Reported: 2019-07-16 16:51 PDT by Kenneth Russell
Modified: 2019-07-17 10:50 PDT (History)
9 users (show)

See Also:

Attachments
Patch (24.22 KB, patch)
2019-07-16 17:01 PDT, Kenneth Russell 		no flags Details | Formatted Diff | Diff
Archive of layout-test-results from ews212 for win-future (13.75 MB, application/zip)
2019-07-16 18:45 PDT, EWS Watchlist 		no flags Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Kenneth Russell 2019-07-16 16:51:10 PDT 
WebKit's current ANGLE snapshot doesn't work inside WebKit2's sandbox because it makes a deprecated API call, CGDisplayIOServicePort, which is only used to determine the "active" GPU on dual-GPU MacBook Pros. It doesn't otherwise impact ANGLE's internal logic, nor consumers like WebCore. Top-of-tree ANGLE has removed this API call in http://crbug.com/angleproject/3383 . Cherry-picking that commit makes WebKit's ANGLE backend for WebGL work in WebKit2.
Comment 1 Kenneth Russell 2019-07-16 17:01:08 PDT 
Created attachment 374264 [details]
Patch
Comment 2 EWS Watchlist 2019-07-16 17:03:42 PDT 
Note that there are important steps to take when updating ANGLE. See http://trac.webkit.org/wiki/UpdatingANGLE
Comment 3 Kenneth Russell 2019-07-16 17:04:22 PDT 
The attached patch is just an application of https://chromium-review.googlesource.com/1584360 with one merge conflict resolved, and also a tiny fix to WebGLLayer.mm fixing a build failure with USE_ANGLE=1 that I introduced at the last minute in my most recent patch (sorry!). It makes WebKit's ANGLE / WebGL backend run in WebKit2 mode in the minibrowser.
Comment 4 EWS Watchlist 2019-07-16 18:45:23 PDT 
Comment on attachment 374264 [details]
Patch

Attachment 374264 [details] did not pass win-ews (win):
Output: https://webkit-queues.webkit.org/results/12754839

New failing tests:
http/tests/css/filters-on-iframes-transform.html
Comment 5 EWS Watchlist 2019-07-16 18:45:25 PDT 
Created attachment 374269 [details]
Archive of layout-test-results from ews212 for win-future

The attached test failures were seen while running run-webkit-tests on the win-ews.
Bot: ews212  Port: win-future  Platform: CYGWIN_NT-10.0-17763-3.0.5-338.x86_64-x86_64-64bit
Comment 6 Alexey Proskuryakov 2019-07-16 20:49:26 PDT 
Could you please elaborate on what the user visible issue is? If we didn’t notice a full “doesn’t work”, that would be sad and surprising.
Comment 7 Radar WebKit Bug Importer 2019-07-16 20:49:48 PDT 
<rdar://problem/53186410>
Comment 8 Alex Christensen 2019-07-17 10:22:10 PDT 
Comment on attachment 374264 [details]
Patch

Kenneth is working on a WebGL implementation that not only uses ANGLE for shader translation, but also for rendering instead of using OpenGL/OpenGLES directly like we do now.  This affects that WebGL implementation.  ANGLE as a text-in-text-out translator does not need this change, so shipping WebGL is not broken and is unaffected by this patch.  r=me
Comment 9 WebKit Commit Bot 2019-07-17 10:50:51 PDT 
Comment on attachment 374264 [details]
Patch

Clearing flags on attachment: 374264

Committed r247521: <https://trac.webkit.org/changeset/247521>
Comment 10 WebKit Commit Bot 2019-07-17 10:50:53 PDT 
All reviewed patches have been landed.  Closing bug.