1177563 - add regression test to ensure we don't fall off the shared rule processor optimized path

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect)

Description

[Cameron McCormack (:heycam)]

In bug 77999 we share rule processors across documents that use the same set of agent/user style sheets.  We should add a regression test to ensure that we don't accidentally stop sharing rule processors for the simple case of having a couple of blank tabs open.

Comment 1

[Cameron McCormack (:heycam)]

Attachment: patch

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 8626355 [details] [diff] [review]
patch

Please rev the IID for nsIDOMWindowUtils.

r=dbaron with that

Is it worth testing that winUtils.hasSharedRuleProcessor(Ci.nsIStyleSheetService.AGENT_SHEET) is false in some case as well?  (Maybe on the AUTHOR_SHEET level, with a tab that's not about:blank and has a style element?)

(Having a test that tests that hasSharedRuleProcessor returns false in some case also helps test hasSharedRuleProcessor itself.)

Comment 3

[Cameron McCormack (:heycam)]

Yes, that's a good idea.  I've added two additional checks to the test -- that the AUTHOR_SHEET rule processor is not shared, and that adding a new, unique sheet to the agent-level causes its rule processor to be not shared.

Comment 4

[Cameron McCormack (:heycam)]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=07c5e812a56c

Comment 5

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/42e80a07acd6

Comment 6

[Pulsebot]

Backout:
https://hg.mozilla.org/integration/mozilla-inbound/rev/4ed2ded0daef

Comment 7

[Carsten Book [:Tomcat]]

sorry had to back this out for bustage like https://treeherder.mozilla.org/logviewer.html#?job_id=11470292&repo=mozilla-inbound

Comment 8

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/61d73966bfb1

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/61d73966bfb1