mozilla :: #e10s

8 Aug 2017
20:35fffsfs35how can i figure out why multiprocessmode is not running?
20:35fffsfs35about:e10s or sth else?
20:36fffsfs350/2 (deaktiviert wegen Add-on
20:36RyanVMlooks like it's disabled due to an incompatible addon
20:37fffsfs35but which?
20:38RyanVMthe addons manager should say
20:42fffsfs35says all alter oddon type
20:43RyanVMwhich one(s)?
20:47fffsfs35classic theme restorer
20:47fffsfs35firestorage plus
20:47fffsfs35privacy badger
20:48fffsfs35proxy selector
20:48fffsfs35mute tab
20:48fffsfs35google search link fix
20:48fffsfs35and some deactivated
20:56* RyanVM doesn't see a Proxy Selector extension on AMO
20:57RyanVMlooks like nearly all of those are marked as legacy extensions
21:00RyanVMFireStorage Plus, Muter, RefControl, storeTab are all disabling e10s
21:00RyanVM(marked as not compatible)
21:00fffsfs35so there is no successor?
21:00RyanVMI have no clue about any of those
21:01RyanVMCTR, QuickJava, Firebug are marked as legacy and won't be supported in 57+, but aren't disabling e10s at the moment
21:01fffsfs35thats bad :/
21:01RyanVMPrivacy Badger, Mute Tab, and Google search link fix are Web Extensions are will work fine with e10s/57
21:01RyanVM has more info
21:01fffsfs35the most worst thing is, that java and flash is blocked in x64
21:03fffsfs35java is needed for so many remote server view tools, security cams, etc
21:04RyanVMthe ESR52 release is supported until mid-2018
21:04fffsfs35thats soon
21:04RyanVMsorry for the bad news :(
21:04RyanVMthis has all been the public plan for a long time now
21:04fffsfs35and i hope classic theme restorer will function in future
21:05RyanVMtheir maintainers have said they don't intend to try making it work with 57+
21:05fffsfs35of theme restorer?
21:05fffsfs35then i have to use the cripple ff interface? omg
21:06RyanVMfwiw, the Firefox theme is changing significantly in 57
21:06RyanVMyou can see what it looks like in current nightly builds
21:08fffsfs35i prefer the oldest guis
21:08fffsfs35windowns, ff, etc
21:09RyanVMI can't speak to what themes will or won't exist for Fx57, but I have an overall feeling that if there's demand for something more like CTR, people will make it
21:09RyanVMmaybe I'm just an optimist by nature though :)
21:09RyanVMtheming certainly isn't going away
21:09fffsfs35too optimistic
21:10fffsfs35ff is getting more worse and worse since they decided to use major version stepping for commercial reasons :P
21:10RyanVMsorry you feel that way
21:10RyanVMmany of us obviously feel differently in that respect :)
21:14fffsfs35maybe the latest 2 majors are getting a bit better
21:14fffsfs35i hope it will get better
21:16The_8472performance is getting better at the expense of many addons
21:18mrbkapbillm: returning nsStrings is the way to do things now? We're able to rely on RVO to avoid copying data?
21:19billmmrbkap: even without RVO, won't the move constructor make it reasonably fast?
21:19billmwell, if there is one. now I'm not sure.
21:20mrbkapbillm: I can't find one.
21:20mrbkapbillm: this probably changed at some point, I was just curious.
21:20billmmrbkap: ok, well, I guess I can use an outparam. I honestly wasn't thinking much about it.
21:20mrbkapbillm: also, do you have plans to write any tests for this nsINamed JS-wrapper stuff?
21:20mrbkapbillm: lemme verify, I'm pretty sure I've seen other patches returning strings.
21:21billmmrbkap: sort of. I think I could test it from JS somehow.
21:21mrbkapbillm: oh, you might need a C++ test :(
21:21mrbkapbillm: IIRC, it's hard or impossible to get your hands on a double-wrapped object these days.
21:21billmmrbkap: if I double-wrap some JS thing, I think it might work
21:22billmmrbkap: can I pass it through an observer?
21:22mrbkapbillm: oh, that might work.
21:28fffsfs35at least it would be good to add a classic gui and java support for x64 ;)
21:29fffsfs35cya l8r and thx
22:15mrbkapbillm: JS names can be UCS-2 right?
22:15mrbkapbillm: what happens if we get an emoji in a function name in a CString?
22:42mrbkapfelipe: ping?
22:54felipemrbkap: pong
22:55mrbkapfelipe: Hey, did you get a chance to run your e10s-disabled test script again?
22:55mrbkapfelipe: there's some discussion about turning off non-e10s tests on the newsgroups and dbaron was wondering what the plan was.
22:55mrbkapnewsgroups = dev.platform
22:56felipemrbkap: not yet, but I saw the thread and that reminded me that I have to run that again
22:56mrbkapfelipe: ok, cool.
22:56mrbkapfelipe: I was going to file a bug to remove e10srollout on 57 (Nightly), btw. That sound good to you?
22:57felipemrbkap: yep, it does. And I'm gonna file one to remove the front-end buttons and stuff related to it
22:58mrbkapfelipe: great.
22:58RyanVMfelipe: i still have no clue what the plan is for mochitest-chrome :(
22:58RyanVMdavidb at least has his team working on getting mochitest-a11y working with e10s enabled
22:59mrbkapWhat code do we get to rip out by killing mochitest-chrome?
22:59mrbkapWhat's the benefit of killing it?
23:00mrbkapother than de-XULification?
23:00RyanVMi think the question is more "what are we willing to do to keep it alive if/when non-e10s breaks?"
23:00RyanVMif it just keeps on working, great
23:02mrbkapWell, we're going to have to be able to load XUL pages in the parent process for the foreseeable future anyway, right? Because that's how preferences works, if for no other reason.
23:04* mrbkap still doesn't know what people are talking about when they say that non-e10s will "break".
23:06mccr8Isn't mochitest-chrome just running... chrome? It doesn't seem like you need to support running content in the main process for it to work.
23:08mrbkapmccr8: yes.
23:09felipeI don't think mochitest-chrome will break by itself so I believe we can just keep running it.. any major change that would break it will come up to be discussed and then we can figure that then
23:10felipebut in any case, as I understand it, the majority of the tests there are testing XUL implementation stuff, so I believe the platform team should be responsible for it
23:11mrbkapfelipe: Except that the "platform team" doesn't exist anymore.
23:11mrbkapfelipe: maybe you mean the "runtime team" :x
23:11felipeyeah, maybe that :P
23:11felipemochitest-chrome should probably be tested with Stylo
23:29jimmmrbkap: you're on the runtime team aren't you?
23:29mrbkapjimm: Yes.
23:30jimmI'd love to see a list of all the tests that remain.. I wonder how much is left
23:31* jimm looks at th test run totals
9 Aug 2017
No messages
Last message: 12 days and 20 hours ago