649360 - Convert some of the EventUtils.js over to SpecialPowers

Status: RESOLVED DUPLICATE of bug 674323

(Testing :: Mochitest, defect)

Description

[Doug Turner (:dougt)]

I converted some of the EventUtils over to use SpecialPowers so that I can start gettings some of the mochitests passing.

One question -- how do I tell from the SpecialPowers component if we are doing remote e10s layers or not? maybe the pref: browser.remote.tabs?

Comment 1

[Doug Turner (:dougt)]

Attachment: patch v.1

Comment 2

[(not currently active) Ted Mielczarek]

The only thing I could tell you to check would be nsIXULRuntime.processType. However, we've been writing everything to simply assume we're running with a content process, so you might as well continue on that path. Firefox is going to get there at some point, so it doesn't seem worthwhile adding extra complexity to do things differently for IPC vs. non-IPC.

Comment 3

[Joel Maher ( :jmaher ) (UTC -8)]

I have a simple check in my ipc.js file:
if (xulRuntime.processType == 2) {

http://mxr.mozilla.org/mozilla-central/source/testing/mochitest/ipc.js#82


this has been working as it runs on all desktop and mobile platforms for the last 9 months or so.

Comment 4

[(not currently active) Ted Mielczarek]

Technically you want to be testing that against Components.interfaces.nsIXULRuntime.PROCESS_TYPE_CONTENT:
http://mxr.mozilla.org/mozilla-central/source/xpcom/system/nsIXULRuntime.idl#95

but regardless, I'm not sure why we need to test this. Shouldn't we just fix all the tests to assume they're running in a content process, since all the IPC code works fine in current desktop Firefox anyway? Seems silly to write conditional logic for a case that's going to go away within a year (non-IPC).

Comment 5

[Joel Maher ( :jmaher ) (UTC -8)]

ted I think this is different because this code is the harness, not the tests themselves.

Comment 6

[(not currently active) Ted Mielczarek]

Still, I don't think it's worth writing two codepaths, one of which we know is going to be thrown away, unless there's a real compelling reason.

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

Attachment: patch + test fixes (WIP)

this patch actually implements and uses the snapshotwindow, comparewindow, synthesizeKey and synthesizeMouse.

The problem is we can't check this in because the chrome tests depend on these utilities and ATM chrome doesn't have access to SpecialPowers.

good news is with this patch on desktop firefox and fennec we are really really close to green for layout/base in mochitest-plain!

Comment 8

[Joel Maher ( :jmaher ) (UTC -8)]

Attachment: updated patch with latest specialpowers changes (WIP)