mozilla :: #firefox

15 Mar 2017
00:24Caspy7and it might not have happened (as soon) if not for decebalus being so consistently annoying, er, persistent
00:40TimvdeCaspy7: What might not have happened?
00:40Caspy7Timvde: https://bugzilla.mozilla.org/show_bug.cgi?id=1131937#c26
00:40firebotBug 1131937 NEW, nobody@mozilla.org Dramatically high CPU usage on facebook group page
00:41TimvdeCaspy7: Ah, yes :)
00:42* Timvde points at comment 22
00:42TimvdeI hope they find a solution
00:42Caspy7yes, good job
00:42Caspy7sounds like he thinks there can be one short term
00:42TimvdeIsn't the bug fix "Facebook should stop doing that"? :/
00:43Caspy7\_()_/
00:43Caspy7have to talk to a web dev about what they're actually and why
00:43Caspy7Timvde: the range was found like 2 month ago, what prompted you to go digging now?
00:44TimvdeCaspy7: I think decebalus said something about it in this channel
00:45Caspy7ah
00:45Caspy7wonder why no one did that when he first mentioned it
00:45TimvdeAnd I actually did the digging before and suspected bug 539356 already, so it triggered me to talk about it with emilio (who replied at that time he would look into it)
00:45firebothttps://bugzil.la/539356 FIXED, matt.woodrow@gmail.com Replace Invalidate() calls in reflow with display list analysis
00:46TimvdeHe confirmed that I probably was right, so I decided to post it
00:46* Caspy7 nods
00:53Operatorso Timvde why did you lie on your post
00:53Operatorhttps://bugzilla.mozilla.org/show_bug.cgi?id=938464#c12
00:53firebotBug 938464 REOPENED, nobody@mozilla.org google search results page takes 100-500k in sessionstore file
00:54OperatorYou posted that minutes after a conversation with me, where I repeatedly said that "Google Instant is disabled"
00:54Operatorthen you went on the Bugzilla and outright lied
00:54Caspy7Operator: oh my gosh, really? He did not lie. This was covered earlier
00:54Caspy7please stop this
00:54OperatorYes do you want to see the log
00:54Operatorhe fabricated it, he said he talked to someone on IRC and they determined it's caused by Instant search
00:54Caspy7Operator: as was covered, *other people* tested it and found it to be the case
00:55OperatorNo actually nobody said that
00:55OperatorI checked the log
00:55OperatorHe made that part up
00:55TimvdeOperator: I did not lie. *Two* people tested this.
00:56TimvdeThere might be *other causes*, which you are hitting, but Google Instant *definitely* causes bigger session restore files
00:58Operatorand why didn't you note that in your post
00:58Operatoryou said it was determined that Instant Search causes it
00:58OperatorYou only posted what you wanted to confirm
00:58Caspy7Operator: because he didn't know you'd treat this like a federal case and go crazy about it
00:58Operatoryou didn't say "other users reported that it occurs with Instant Search disabled"
00:58TimvdeOperator: I said that disabling Google Instant improves the situation
00:58dolskeOperator: it doesn't feel like you're being very constructive here.
00:58TimvdeWhich is truee.
00:59TimvdeOf course, if you already have Google Instant disabled, it can't help you any further
00:59TimvdeBut it would probably be a lot worse if you enabled it
00:59Operatorsome of my "Google Search" tabs are already 5 MB with Instant Search disabled
01:00Operatorand I don't see how we're ever going to make any progress with this bug since it's been happening for 5 years now
01:00Caspy7Operator: just because you have a problem, doesn't mean that every problem is caused by the same thing.
01:00Operatorand half the people think it's caused by 'Google Instant'
01:00Operatorso they aren't even looking into the root cause
01:00Caspy7Operator: just because you're frustrated does not give you the right to lash out and attack people
01:00TimvdeOperator: If you truly want to be helpful, you can always try to find out the cause
01:00TimvdeIt's kinda hard for us, when we can't reproduce your issue
01:00dolskeOperator: what is the specific problem you're having?
01:01Caspy7Operator: Google Instant does cause an issue. It's just that you're experiencing something else. How hard is that?
01:01dolskeis it just the session store usage? 5MB per tab?
01:04Operatordolske, I use a large session with many tabs. The sessionstore file is ~250 MB. It causes the browser to hang during exit with SessionStore.jsm saying "A script on this page may be busy" and I have to press Continue. So I looked into the problem using the about:sessionstore plugin. I can see that everything is normal, except for "Google Search" tabs taking between 100K and 6MB of memory each.
01:04Operatorbug 938464
01:05firebothttps://bugzil.la/938464 REOPENED, nobody@mozilla.org google search results page takes 100-500k in sessionstore file
01:05Operatorbug 669603
01:05firebothttps://bugzil.la/669603 NEW, nobody@mozilla.org large sessionStorage data causes session restore to block the UI
01:05dolskeOperator: ok, and have you worked out how to reproduce this in a minimal environment?
01:05Caspy7dolske: I'm not highly savvy to session store, but is it possible this is from multiple searches in the history for those particular tabs? Would that get saved in sessionstore?
01:06Caspy7given that it's a range
01:06Caspy7wide range
01:06dolskewell, anything is possible. I know Google has been a problem in this area before.
01:06Operatordolske, well, it's fairly easy to reproduce, considering I've been having this problem for years now, on every Firefox browser, on all 3 of my computers, etc
01:06Operatorand others are reporting it
01:06Caspy7one thought is to do a refresh in order to get sanitized sessionstore/tabs
01:07Caspy7bit of jiggering
01:07Caspy7that way it dumps the history of the tabs, but keeps the open tabs
01:07dolskeThe relevant bit here is to see if it's reproducable (ideally) in a new profile, so that someone can diagnose the cause.
01:08Caspy7Operator: do you tend to do repeated google searches in the same tabs? like you keep that tab open and do another search and another?
01:08Operatorno
01:08OperatorNew tab
01:08TimvdeCaspy7: Yes, sessionstore should store up to `browser.sessionstore.max_serialize_back` history items
01:08Operatordolske, as far as I'm aware, the steps to reproduce are: Make a fresh profile, Google Search for some stuff, open a bunch of tabs.. Come back later. Look at about:sessionstore
01:09TimvdeOperator: It might be an add-on that you're using
01:09TimvdeThat's why it's important to try in a clean profile, without any add-ons installed
01:09OperatoruBlock Origin
01:10Operatorit's also causing my browser to write 100GB to the disk per day
01:11OperatorTimvde, how shall I keep a fresh profile open without disrupting my work?
01:12SoItBeginsabout:profiles
01:12SoItBeginsand then...
01:12TimvdeOperator: firefox -P -no-remote
01:12SoItBeginsyou can create a new profile...
01:12SoItBeginsand have it open in a separate browser...
01:12TimvdeNo clue about GUI :P
01:12SoItBeginsat the same time!
01:12SoItBegins(Launch profile in new browser)
01:13Caspy7yeah, use the -no-remote flag to open profiles at the same time
01:13dolskeyeah, I can reproduce in a new profile. Starts off around 400K, and a couple more searches bumped it up a few hundred K per.
01:13OperatorIt's not really worth me reproducing, cause it's already so easy to reproduce
01:13Operatordolske: *that*
01:13Timvdedolske: with Google Instant disabled?
01:13Caspy7Operator: not worth fixing? Ok.
01:13Timvde(It is enabled by default)
01:14OperatorCaspy7, I use Nightly
01:14OperatorI was going to download Plain firefox just for this
01:14Caspy7ok?
01:14Caspy7Nightly will work too, whatever
01:14Operatorand I'm saying, it's not worth me doing it again, since dolske can see the problem easily, and others can see it
01:14OperatorI already tested it in the past with a fresh profile
01:14Operatorand I got the same problem
01:15dolskeTimvde: oh, I forgot about that. sounds like from your comment it still grows but much more slowly?
01:15Timvdedolske: It is *an* issue (but apparently not the only one)
01:15dolskeand ideally we'll need a fix that doesn't require users to twiddle Google settings.
01:15Operatordolske, I get the problem with Google Instant disabled
01:16dolskeI don't know enough about the issues to say what a fix would be here, but it's blocking bug 669034 (and in turn bug 1330635), so it's on the list of sessionstore issues being tracked.
01:16firebothttps://bugzil.la/669034 NEW [meta] Re-architect session restore to avoid periodic freezes
01:16firebothttps://bugzil.la/1330635 NEW Sessionstore performance tracking
01:17dolskejust a lot of work to do to make session store better. help wanted!
01:17OperatorThis problem has been here for 5 years
01:18Caspy7and? Looks optimistic as session store is getting actual priority right now
01:18Operatornot to mention, that when my sessionstore file hits 256 MB, that's the end
01:18Operatorit starts losing data
01:18Operatorit does not grow past 256MB, everything written to it is corrupted
01:19weaksauceinteresting... are there any plans to do what chrome is doing just recently to limit the background tabs to 1%/core
01:19Operatorso I'm always paranoid about how many "Google Search" tabs are open
01:19Operatorsince it can corrupt my session
01:19dolskemy best suggestion for now would be to close Google tabs you don't need.
01:19Caspy7hm, how about using a different site as proxy?
01:19Caspy7there's at least one...start page is it?
01:20Caspy7does google searches
01:20Timvdeweaksauce: Afaik, no, but I can be very wrong
01:21Caspy7weaksauce: right now content process is one process and in the future (55 I think) it will be 4 processes. But it will likely never be per-tab
01:21Caspy7Chrome doesn't even do per-tab actually
01:21dolsketelemetry data says only 0.14% of users have a session store file over 50MB, so it's not a common problem...
01:21OperatorCaspy7, I don't see any work being done in https://bugzil.la/938464
01:21TimvdeCaspy7: That's not what he asked? :P
01:21firebotBug 938464 REOPENED, nobody@mozilla.org google search results page takes 100-500k in sessionstore file
01:21dolskehttps://telemetry.mozilla.org/new-pipeline/dist.html#!cumulative=1&end_date=2017-03-12&keys=__none__!__none__!__none__&max_channel_version=nightly%252F55&measure=FX_SESSION_RESTORE_FILE_SIZE_BYTES&min_channel_version=null&processType=*&product=Firefox&sanitize=1&sort_keys=submissions&start_date=2017-03-08&table=0&trim=1&use_submission_date=0
01:21Caspy7though everyone seems to think that
01:22TimvdeCaspy7: http://blog.strml.net/2017/01/chrome-56-now-aggressively-throttles.html
01:22Caspy7Timvde: I'll have to read up on it, but I assume they're doing this through the OS, because they can limit the separate processes
01:22Caspy7and that's not feasible with Firefox's setup
01:23Operatordolske, what should this indicate?
01:23TimvdeCaspy7: Looks like they're just manipulating the timers accordingly
01:24Caspy7weaksauce: there is currently a project called Quantum DOM to have greater control over background tab usage
01:24Caspy7weaksauce: you can find it listed here: https://wiki.mozilla.org/Quantum
01:25Caspy7which lists further links on description
01:26OperatorI figured something out
01:26Caspy7"When Quantum DOM is finished, JS code for different tabs (and possibly different iframes) will run in separate cooperatively scheduled threads; the code for some background tabs will never run at all. "
01:26dolskeWe also do some various throttling of background tabs. (I don't know how it compares to Chrome, though)
01:27Operatorfor all Google Image Search results, the size of all the image previews is going into the sessionstore
01:27Operatorsimply open a Google Image Search, scroll down to view some previews, then check your tab
01:27Operatorit will have multiple megabytes in the sessionstore
01:27dolskehttps://blog.wanderview.com/blog/2017/03/13/firefox-52-settimeout-changes/ for example
01:28dolskealso media stuff, and some other things
01:29Caspy7weaksauce: I skimmed that page...where does it say 1% CPU?
01:29Caspy7I see "throttle background tabs' timers."
01:35OperatorI feel bad for Christopher D. Coleman who commented 12 days ago: "I am also continuing to see this issue and have for a years. I do a log of google searching simultaneously. It has become such a problem I have had to invest in more system memory."
01:36OperatorNobody should have to buy hardware cause of a bug
01:38weaksauceCaspy7 https://hardware.slashdot.org/story/17/03/14/2055249/chrome-57-limits-background-tabs-usage-to-1-per-cpu-core
01:38dolskeOperator: you can also try setting dom.storage.default_quota to a lower value. Setting it to 100 (~100kb) seems to reduce the storage usage I See.
01:42dolskeIt seems like this is effectively a per-tab limit, so I'd suspect this only becomes a major problem for people who keep lots and lots of google search tabs open.
01:42Operator<--
01:43dolskeyep, so again: the workaround for now is to aggressively close tabs with google seach results.
01:43OperatorI&#39;m wondering what&#39;s actually even being stored, when Google Instant is off, and it&#39;s just a page with some text
01:43Operatorhow does that come to 3 Megabytes
01:44dolske&quot;ask google&quot; :)
01:44Operatoryou could store a massive amount of information and results in 3 MB
01:44dolskeor use the devtools storage inspector.
01:44Caspy7weaksauce: I find this a bit confusing... the title says limit to 1% per core, but the actual text says &quot;Chrome 57 will temporarily delay a background tab&#39;s JavaScript timers if that tab is using more than 1% of a CPU core&quot;
01:44OperatorI inspected the JSON
01:44dolskehow do I turn off google instant?
01:44Operatorit&#39;s a blob of incomprehensible encoded data
01:45Operatorit needs further investigation to figure out what that data actually is
01:45dolskeyou can use a prettyprinter like http://jsonprettyprint.com/
01:45dolskewell, you probably shouldn&#39;t paste real data in there
01:45dolskesince you&#39;d be exposing sensitive cookies to some random 3rd party site
01:45Operatordolske: on Google search
01:46OperatorSettings -> Search settings -> When should we show you results as you type? -> (x) Never show Instant results
01:46dolskeah
01:46OperatorBUT! I just noticed something... There&#39;s a &quot;Private&quot; results setting as well
01:47Operatormaybe that&#39;s causing it
01:49FirefoxCommandLineIs there a firefox command line that opens firefox on a web-page, with the caveat that if the web-page is already open then it switches to it rather than opening a new copy?
01:51weaksauceCaspy7 not sure what the ramifications will be on it yeah and i have no idea how they are actually implementing it
01:51FirefoxCommandLineWhen I tab complete firefox on zsh (a zsh feature, not an &quot;official&quot; firefox list of params&quot;), I see a 00new-tab command param. I used --old-tab param in the hope that it does what I need but it didn&#39;t work
01:52FirefoxCommandLineShould have looked for this: http://kb.mozillazine.org/Command_line_arguments But yeah, this doesn&#39;t have what I need.
01:52Caspy7weaksauce: well, I think you can add what I said about Quantum DOM (hope you read up on it) and what dolske said. So, with quantum dom at least seems it should be possible, maybe not in the exact same implementation, but in a way to significantly save performance by limiting background tabs. So same idea
01:53TimvdeFirefoxCommandLine: Afaik, not possible without an add-on, sorry
01:53TimvdeI&#39;m also not sure whether an add-on exists
01:53weaksauceCaspy7 good... that&#39;s needed
01:54Operatordolske, can you try that same search you did earlier, with &quot;Private&quot; results disabled?
01:54dolskealas I&#39;ve got to run
01:54dolskesorry, hope the workaround help somewhat.
01:55OperatorI can&#39;t confirm yet, but it seems that it may have been caused also by the &quot;Private&quot; results
01:56FirefoxCommandLineTimvde: Ok. Noted. Thank you.
02:19Modsgotthewronganonhey bois
07:16engblomIf there are any firefox devs here: I have been a firefox user since before even Phoenix became stable. I have not jumped on Chromium and other buzz browsers. However recent changes is forcing me to switch browser unless you begin reverting.
07:16engblomI am talking about the stupid, idiotic pulseaudio requirement.
07:17engblomAnd no, I am not going to install pulseaudio, just for one single application. I have a minimalistic system and my audio mixer is alsamixer, which gives me fine control and ease of use.
07:17engblomThere is not even one console pulseaudio mixer giving the same control and ease of use.
07:28kentuckyfriedtakaheengblom: why don&#39;t you want to install pulseaudio?
07:29kentuckyfriedtakaheis it the quality of console mixers that is the issue?
07:30engblomkentuckyfriedtakahe: Yes, the biggest reason. I want the fine control of alsamixer and with the same ease of use.
07:31kentuckyfriedtakaheso you would install it if not for the lack of a console mixer?
07:31engblomkentuckyfriedtakahe: If really forced, but not happily. I hate extra layers being potential bug nests and resource wasters.
07:34kentuckyfriedtakaheengblom: I&#39;ll make a mental note of that and see any common issues with pulse emerge
07:34kentuckyfriedtakahefyi we moved away from supporting ALSA because pulseaudio is significantly less buggy
07:35engblomI never had any problem with alsa and firefox. And currently Firefox is the only application in my computer requiring pulseaudio
07:36kentuckyfriedtakaheengblom: a lot of people did
07:36engblomCould it not be a about:config option?
07:36engbloman*
07:37kentuckyfriedtakaheengblom: no
07:37engblomI do not mind it using pulseaudio as default even, if pulseaudio is available, but it would be nice if it did fall back on alsa
07:37kentuckyfriedtakahethe problem is that it falls back to broken for a lot of people
07:37kentuckyfriedtakahewhich is in stark contrast to pulseaudio which seems to fix everyone&#39;s audio issues
07:38engblomRight now it tells me to install pulseaudio. Would it not be enough for those people actually having trouble to see that message? I mean, then the ball is back at the user again and the user get a choice to continuse with the &quot;buggy&quot; alsa or to switch to pulseaudio?
07:39kentuckyfriedtakaheExcept people complain when audio doesn&#39;t work and don&#39;t know why
07:39kentuckyfriedtakahepulse-audio not being installed is more easily fixable
07:41kentuckyfriedtakaheengblom: fyi there are a number of people forming an angry mob in bug 1345661
07:41firebothttps://bugzil.la/1345661 NEW, nobody@mozilla.org PulseAudio requirement breaks Firefox on ALSA-only systems
07:43kentuckyfriedtakaheengblom: the thing about my job is that I have to make opportunity cost decisions
07:44engblomWhy not fixing the bugs in the alsa code instead?
07:45kentuckyfriedtakaheengblom: because it has the highest number of bugs of any audio backend. more than windows, more than mac, more than oss
07:46kentuckyfriedtakaheWe&#39;re also rewriting all the backends
07:46kentuckyfriedtakaheto make them work in a sandbox
07:47engblomIf you are anyway rewriting the backends, why not also include alsa in the rewrite?
07:47kentuckyfriedtakaheengblom: because pulseaudio already works
07:47kentuckyfriedtakaheand the effort could be spent somewhere productive
07:48engblomWill you rewrite the oss support?
07:48kentuckyfriedtakaheno
07:48kentuckyfriedtakahethat is tier-3 and maintained by bsd folks
07:50kentuckyfriedtakahealsa is tier-3
07:50kentuckyfriedtakahewe&#39;re still accepting patches but not compiling in by default
08:20TewSoni - APNG support in Chrome, https://chromium.googlesource.com/chromium/src/+/7d2b8c45afc9c0230410011293cc2e1dbb8943a7 (https://news.ycombinator.com/item?id=13871809)
08:30Mardegwow! Finally.
08:33Mardegso that leaves just IE/Edge not supporting it
08:35Corkguess hoping for that ugly hack to die is slowly going away...
08:36Mardegdon&#39;t worry the butthurt over MNG going away will never end
08:39Corkit isn&#39;t so much about mng as it is &quot;here is an image, but ya, it isn&#39;t actually an image&quot;
09:25__marcoHello, does the google talk plugin work with FF?
09:25__marcoFF52
09:27Tokoyami__marco: on 52 no, on 52esr maybe (not 100% sure)
09:28__marcoHow do I know if I am using the ESR version? The about page does not mention it
09:28Tokoyamithen you probably aren&#39;t
09:29Tewif google talk = hangouts, last I heard it was still not-fixed (goog side)
09:29Tokoyamihttps://www.mozilla.org/en-US/firefox/organizations/all/
09:29Tokoyamithis is the download page for ESR
09:30__marcoI am using the official Debian package
09:30Tewabout:support has &quot;Update Channel&quot; item that indicates &#39;esr&#39;
09:31Tewmost linux do use esr.. but I wonder how fast they are to switch to the new one? (there is a transitional period when e.g. ff45 ESR still releases while ff52 ESR first .minor versions are already available)
09:31TokoyamiTew: distribution builds usually have the Update Channel disabled because the package manager handles the updates
09:32TewTokoyami - good point, I did not know that.
09:32__marcohttps://packages.debian.org/search?keywords=firefox&searchon=names&suite=all&section=all&sourceid=mozilla-search
09:33__marcofirefox-esr 52 is only in experimental
09:33__marcook, thanks
09:34Tokoyami45.8 is the last 45 ESR that released with 52.0 ESR so they&#39;ll probably update at some point (if debian does such updates in a single major debian version)
09:34Tokoyamipossibly at 52.1
09:35TewTokoyami - no, there should still come 45.9.0
09:35Tokoyamioh
09:35TokoyamiI see
09:35Tewit overlaps 2 .minor versions
09:35Tew(https://www.mozilla.org/en-US/firefox/organizations/faq/)
09:45sam113101is it possible to write proprietary addons for firefox? or are they open source by default?
09:56Tew&quot;Yes; many are (you can check Version if the addon has licence stated and View the source)&quot;, but maybe rephrase your goal-question? there are multiple definitions of open source - as long as you do have the addon package, you can view the code inside it (which may not be really readable as-is).
10:00TewI remember https://addons.mozilla.org/en-US/firefox/addon/flash-debugger/ as an example of paid extensions (but the licence is apparently GPL 3.0); _many_ companies have extensions written for internal use / to accompany their software (oh the fun we had watching them scramble over addon signing... /s)
10:15sam113101Tew: I didn&#39;t know there were paid extensions
10:18decebalusanybody know an addon wich can show thr time loading for webpages?
10:19decebalus*the loading time
10:26DuClareI&#39;d imagine the devtools do that already?
10:26DuClareSurely there&#39;s a timeline somewhere
10:26Tewhe&#39;s gone
10:26DuClareOh I&#39;m so dumb again
10:27Tewnah, your client just doesn&#39;t highlight exits glaringly much :)
10:32decebalustried an addon but don&#39;t work
10:33HavvyDuClare: He&#39;s back.
10:33Tewat the bottom right of devtools - Network tab there is a clock and request and time, you can even click it to compare fresh vs primed cache.. I have no idea which page load event it is; it&#39;s one of those things where chrome is better, http://i.imgur.com/PcGnd9t.png Firebug should show it better too, but why would someone keep firebug just for this.. (no idea how it currently is, but as you
10:33Tewcomplain about your perf, you probably should be careful)
10:34decebalusDuClare: devtools do that but it continue to record after complete loading pages and I don&#39;t want that
10:35Tewyour search keywords here are (most likely) &#39;DOMContentLoaded&#39; and/or &#39;(page) load event&#39;
10:40decebalusTew: I don&#39;t want to do thing like coding and keep constantly open devtools.I just want something simple .for long term, like this:https://chrome.google.com/webstore/detail/page-load-time/fploionmjgeclbkemipmkogoaohcdbig
10:41DuClareYea keeping devtools open is problematic
10:41DuClareIt&#39;ll eat all your ram
10:44Tew:)
10:48decebalusfound the solution :D
10:50Tewdon&#39;t tell us, don&#39;t tell us!
10:51decebalusInstalled &quot;chrome store foxified&quot; then the above addon for chrome ;)
10:51decebalus*from chrome
10:53Tewcool
10:56decebalusbut I don&#39;t know if it works properly
11:00decebalusand the results are different from devtools :/
11:01decebalusand not working for every page..
11:02decebalusso..goodby addon
11:04DuClarewelcome to the android store
11:06decebalus?
11:53cbauerhow do I find out if a particular add-on is e10s compatible without having to simply test for &quot;does it work&quot;?
12:13Corkcbauer: you could check on http://arewee10syet.com/
12:13Corkcbauer: also if they are &quot;need restart&quot; on amo they most likely are NOT e10s compatible (they might still work though)
12:14Corki don&#39;t think there is any list anywhere for extentions that are explicitly compatible
12:14cbaueramo
12:14Corkthough with the move to webextensions one could look if the extensions is comming from chrome, as other extensions will go away when xul extensions are dropped
12:23cbaueramo?
13:46guybrushhey guys! do i need to use a cloud-service to use the web push-api?
13:55Corkguybrush: you need to use the service the browser uses to get push notifications yes
13:56Corkguybrush: you can push notifications without a need for a cloud service, but then you have to have a page open for the page
13:56guybrushthanks :/
13:56guybrushyes that is what i thought
13:57guybrushi guess there is no way arround this, for security reasons
14:01Corkguybrush: not security reasons, but there has to be a way to get the data from random servers to the specific browser
14:03guybrushah, i see
14:03guybrushthanks for pointing that out
14:04Cork(and having a browser keeping X connections to X different servers just in case any of them want to push a notification won&#39;t scale)
14:04guybrushyeah makes totally sense now
14:07DaggerCork: although they combined the permissions for the two :/
14:07Daggerso you can&#39;t get notifications from an open page without also letting it show you notifications all the time
14:07CorkDagger: i think that makes sense
14:09DaggerI don&#39;t. I want open pages to be able to show notifications, I don&#39;t want them to be able to do it when I&#39;m not even on the site
14:09Daggeror well, depends on the notification
14:09Daggerweb skype uses them to tell you about new messages. I definitely don&#39;t want them doing that when I don&#39;t even have it open
14:10DaggerI can imagine a lot of cases where you do want notifications from not-open sites, sure, but that&#39;s not always what I want
15:07whoisxyDagger: do you know if this holds on exit for example when restarting firefox will it attempt to repopen a connection with any push services ?
15:08DaggerI belive so
15:11whoisxyCork: clarify what you mena by &quot;cloud service&quot; when you say &quot;you can push notifications without a need for a cloud
15:12whoisxy service, but then you have to have a page open for the page&quot;
15:12Corkwhoisxy: there are several apis you can use to push a message to a page you have open from a site
15:12whoisxybecasue from the looks of the documentation that any push notifications don&#39;t require the page to be open, only to initiate
15:12Corksimple push, or websocket to name two
15:13Corkexactly
15:13Corkpush notifications requires that you register a transfer key with the service the browser uses for the push notification
15:13Corkbut simple push and websocket goes directly from the server to the page so no such key is needed
15:14Cork(at the cost of having to have the page open)
15:16whoisxyCork: presumably this is your first example https://developer.mozilla.org/en-US/docs/Web/API/Push_API, I assume the &quot;simple push and websocket&quot; method is reliant on the sites JS to initiate a websocket and display messages which are pushed as apposed to an API provided by FF ?
15:17whoisxyother than ofcourse the websocket API
15:20DOOMGuySorry for my bad English
15:21DOOMGuyI have a problem. For the latest Firefox does not work for me unity web player, and wants to play in the game based on this engine? Can someone help me.
15:22DOOMGuyI have a problem. For the latest Firefox does not work unity web player, and i want to play in the game based on this engine? Can someone help me.
15:25Caspy7DOOMGuy: assuming you are talking about the old NPAPI unity plugin, NPAPI plugin have all been deprecated/removed. They can be used in Firefox 52 ESR though.
15:25Caspy7DOOMGuy: Unity should have announced and made clear that they are moving forward and not recommending the plugin. Developers can now export to web technologies rather than using the plugin
15:26DOOMGuySo what do I do to work these games?
15:26whoisxyCork: ? is ther a separate API provided by FF for a simple push or is that completely down to the users implmentation
15:27whoisxythe user being the website in this example
15:27whoisxyprobably should have said service instead
15:28Caspy7DOOMGuy: have you checked to see that these games are now available as native web apps (no plugin needed)?
15:29DOOMGuyCaspy7 Which for example?
15:29Caspy7DOOMGuy: as I said, you can use them in Firefox 52 ESR, which is the Extended Support Release of Firefox, but it will only be supported for a year
15:29Caspy7DOOMGuy: I don&#39;t know. I don&#39;t know what games you&#39;re talking about.
15:29DOOMGuyCaspy7 a minute
15:31DOOMGuyCaspy7 http://www.cda.pl/gry-online/23192d4/Combat-3-Combat3-Multiplayer
15:33Caspy7DOOMGuy: if they are not offering the non-plugin version, I recommend contacting them about this issue because plugin support is now gone in Firefox. You will be able to use it in 52 ESR for a year, but that is it. They need to re-publish their game to web technologies
15:34Caspy7brb one second
15:34Corkwhoisxy: sorry was ages since i used it so i mixed up the name
15:34Corkwhoisxy: this is what i was referring to https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events
15:35Cork(that and websocket, where i would quite strongly recommend websocket)
15:36Caspy7DOOMGuy: if using 52 ESR is your only option, you may consider creating a separate Firefox profile just for the game. Then you can create a shortcut just for running it using the options: -no-remote -P &quot;profile name&quot;
15:37whoisxyCork: ahh Okay, wasn&#39;t aware they had so many things to are similar I wonder why they don&#39;t depricate it in favour of websockets
15:37Corkwhoisxy: they are very different
15:37CorkserverEvent is a basic open a connection wait for a data package close and reopen
15:37DOOMGuyCaspy7 I hope. Because I recently browsed the chrome experiments and one of the applications 3d was very advanced, so I hope that the html5 version of the game will be as the advanced
15:38Corkso it is easy to implement in stateless languages like php
15:38Corkwhoisxy: websocket is a constant multiplexed connection you can send data back and forth through, but you need a connection
15:38DuClareStateless language??
15:38CorkDuClare: you don&#39;t have a state between requests
15:38DuClareOh. That has nothing to do with the language though.
15:39Caspy7DOOMGuy: what technologies did it use?
15:39Corkwhoisxy: and then there is the push api where you need to register a key and then you can push without any other needed data
15:39DOOMGuyCaspy7 WebGL
15:39Corkwhoisxy: so they don&#39;t really overlap as they are very different
15:40DOOMGuyCaspy7 JS
15:40Caspy7DOOMGuy: Firefox support WebGL, WebGL 2, and now WebAssembly
15:40whoisxyCork: Okay I get that still however If you need to have a persistant connection thenyou should just do it right the overhead of initiating new TCP can be significant
15:40Caspy7they released WebAssembly with 52 just before Chrome did
15:40Corkwhoisxy: i agree, but ex in a shared hosting you might not have that option
15:41DOOMGuyCaspy7 thanks
15:51DOOMGuyhelp!
15:51firebotDOOMGuy: Welcome to #firefox! Just ask your question and someone should be with you shortly. If you don&#39;t get an answer after a while, you can try taking a look through the support site: https://support.mozilla.org/
15:51DOOMGuyhelp!
15:51Caspy7DOOMGuy: is youre question not answered?
15:52Caspy7*your
15:52DOOMGuyCaspy7 no i only tested firebot
15:52Caspy7firebot: good boy
15:52* firebot smiles
15:53Caspy7firebot: Open the pod bay doors
15:53firebotCaspy7: I&#39;m afraid I cant do that
15:55DOOMGuygoodbye everyone
15:57Caspy7o/
17:13debianuserkentuckyfriedtakahe: &quot;We&#39;re also rewriting all the backends&quot;, &quot;pulseaudio already works&quot; - Ah, you mean you already have pulseaudio backend rewritten? Can you give a link to it? I wonder how much rewriting was needed, maybe I can do that, or learn something new at least...
17:34devslashis there a way that i can paste my own javascript in the browser console and add a breakpoint
17:37MossopYou might be able to use the debugger statement in the scratchpad
17:55kentuckyfriedtakahedebianuser: bug 1346665
17:55firebothttps://bugzil.la/1346665 NEW, dglastonbury@mozilla.com Implement PulseAudio backend for Cubeb in Rust
17:56kentuckyfriedtakahemaybe you can do what?
17:57kentuckyfriedtakahethe ipc isn&#39;t done yet. so far we have a rust port of the C code
18:00firebotJust appeared in Planet Mozilla - http://planet.mozilla.org/ :
18:00firebot Air Mozilla: Building Habit-Forming Products with Nir Eyal
18:00firebot https://air.mozilla.org/habit-forming-products-nir-eyal-2017-03-15/
18:00firebot Air Mozilla: The Joy of Coding - Episode 95
18:00firebot https://air.mozilla.org/the-joy-of-coding-episode-95/
18:03debianuserkentuckyfriedtakahe: &quot;we have a rust port of the C code&quot; -- is the code public? Bug 1346665 looks empty, no code, no patches...
18:07kentuckyfriedtakahedebianuser: the patch should be there soon. you can ask kamidphish on #media in about 5 hours time
18:19debianuserkentuckyfriedtakahe: Ah, it&#39;s not implemented yet. Ok, thank you! I&#39;ll ask kamidphish later.
18:19kentuckyfriedtakahedebianuser: it is implemented. the patch isn&#39;t on the bug
18:20firebotJust appeared in Planet Mozilla - http://planet.mozilla.org/ :
18:20firebot The Mozilla Blog: Five issues that will determine the future of Internet Health
18:20firebot https://blog.mozilla.org/blog/2017/03/15/five-issues-that-will-determine-the-future-of-internet-health/
18:35firebotJust appeared in Planet Mozilla - http://planet.mozilla.org/ :
18:35firebot Myk Melez: Introducing qbrt
18:35firebot https://mykzilla.org/2017/03/15/introducing-qbrt/
18:49Timvdeqbrt looks cool :)
18:53tonymecNo Nightly 55.0a1 nightly (yet?) for Linux today. Anything wrong?
19:01Caspy7tonymec: they&#39;re called *nightlies* ;) (honestly, I don&#39;t know what time they typically ship, but it&#39;s still today here.
19:06WaltS48Typically ship in the morning
19:06tonymecCaspy7: they used to be available no later that 4pm my time, it is now 8pm. (Since the US has gone into DST and I haven&#39;t, I&#39;m temporarily on California time + 8 hours.) It ought to be a few minutes after 12 noon in California IIUC.
19:07tonymecWaltS48: --^
19:07Caspy7I presume it sometimes gets delayed
19:07Caspy7tonymec: you&#39;re welcome to ask in #nightly
19:08tonymecCaspy7: thx
19:10* WaltS48 kinda hinted he would like to know why in #nightly already. Seems there isn&#39;t an update for Android either.
19:11wagleis there only #xul (which is mostly dead) or is there a better channel to ask xul questions?
19:25Mossopwagle: #extdev for extension development or #fx-team for front-end development
19:26wagleMossop: porting xul overlays to firefox directly seems like fx-team then?
19:27wagleMossop: thanks!
19:32decebalusI think Maxthon is the only browser wich use one process for each tab :) http://imgur.com/a/csDDw
20:21CypherousIs there a way to limit how much upstream firefox is allowed to use for uploading to websites?
20:26WaltS48What are you uploading?
20:27WaltS48Cypherous: ^
20:27CypherousVideos to youtube, issue is it uses all of my upstream, which means i can&#39;t do anything else :P
20:28CypherousSo i wanted to throttle it back so i can still use stuff reliant on low latency
20:31WaltS48I see, but have no idea how one would go about that.
20:31Daggera non-shit router ought to be able to handle high upload without increasing latency for everything else
20:31Daggerunfortunately, most routers are shit :(
20:31WaltS48Anything in Developer Tools that may help?
20:32MekliCZHello guys, I need somebody who is responsible for Fennec publishing. I tried few channels, Mozillians.org and wiki, but didn&#39;t find anything. Anybody knows where I can find somebody? :)
20:32CypherousDagger: well its ISP kit, its rubbish but its what i have access to :P
20:32CypherousWaltS48: nothing that i can see :(
20:32CypherousThere used to be an extension apparently, but that broke back with FF4 :P
20:33philippMekliCZ: what would you need in this regard?
20:34CypherousBut yeah its just seeing the upload as normal web traffic and just letting it go full ham on my upstream
20:34WaltS48Cypherous: In all my years using Firefox and following support questions, I don&#39;t recall your question ever being asked.
20:35CypherousWaltS48: well google says there are lots of people asking it, but all being told to install 3rd party software to do the job :P
20:35Cypherousits probably not something that people need much
20:35WaltS48There be a lot of YouTube videos
20:36CypherousWell i meant specifically from the browser itself
20:36WaltS48Understood
20:36MekliCZphilipp: I sent you direct message
20:36CypherousI&#39;m j