WebDriver/RemoteProtocol/Meetings/2019/04/29: Difference between revisions

(→‎Roster: fill in)
Line 23: Line 23:


== Minutes ==
== Minutes ==
[[File:Gutenberg_test.gif|500px|thumb|Gutenberg test running against Firefox]]
* Alex is finally able to run the first Gutenberg test up to completion!
** Had to reverse engineer the precise order of CDP events, which has proven challenging
** ato thinks it makes sense to focus on ensuring it is cross-browser compatible, otherwise end users will find it hard to relate CDP events to web platform primitives and there will be a difference in behaviour between Chrome and Firefox
** Alxo also submitted patches for:
*** Implement <code>Page.frameNavigated</code> ({{bug|1543095}})
*** Implement <code>Runtime.executionContextCreated</code> and <code>Runtime.executionContextDestroyed</code> ({{bug|1543098}})
**** A followup, though not mandatory for WordPress is also to fire <code>Runtime.executionContextDestroyed</code> when the tab closes ({{bug|1547198}})
** Other CDP methods to follow up on:
*** Implement <code>Runtime.evaluate</code> and <code>Runtime.callFunctionOn</code> ({{bug|1543099}})
*** Implement <code>Page.reload</code> ({{bug|1544445}})
* More memory leaks reported for browser-chrome tests
** We disabled browser_cdp.js last week because it was perma-failing the leakcheck, and we thought the memory leak was limied to this one test
** Over the weekend more intermittent leaks cropped up in other tests:
*** {{bug|1547303}}
*** {{bug|1546850}}
** We will proceed with disabling the browser-chrome tests on debug builds
** We suspect there’s a memory leak related to <code>nsIServerSocket</code>, but it’s not a priority to fix it right now
** Memory leaks obviously also affect optimised builds, but since we’re not doing refcounting we don’t notice


== Changelog ==
== Changelog ==
715

edits