mozilla :: #firefox

16 May 2017
01:04concernedDoes this IRC network use an untrusted certificate?
01:18concernedInteresting, this is the third network I am forced to override certificate warnings on. It sounds like a MITM attack on my end.
02:47Caspy7Someone made notes about the upcoming Test Pilots (from the weekly meeting on airmo)
02:48Caspy7Hover Pad, Speak to Me and Peer to Peer file sharing
04:56nikolamFirefox sometimes on right click-open in new tab, opens previous like open previously in new tab and not a new one. Was happening in previous FF releases and now again in FF 52 ESR. (Platform Openindiana/illumos)
05:05Caspy7nikolam: where do you right click to see this new tab option? On the tab bar? On the page background?
05:05nikolami right click on the link on the web page I previously opened, so that link can be opened in a new tab
05:06Caspy7nikolam: how easily are you able to replicate this?
05:07nikolamnot that easy, happens from time to time during the session, like it is random
05:07nikolamI think closer toward the begining of the FF session
05:16nikolamI saw it cross-platform in previous ESRs, like also on MS Win too, also reproduces periodically
05:19Caspy7nikolam: I've never heard of that or seen it before. Addons are the most common cause of uncommon behavior. Do you have any addons that are the same on these installs?
05:20nikolamyes, large number of addons installed... google search link fix comes to the mind..
05:21nikolamOk, so should be able to reproduce without addons, I get it.
06:36firebotJust appeared in Planet Mozilla - :
06:36firebot J.C. Jones: OCSP Telemetry in Firefox
07:16firebotJust appeared in Planet Mozilla - :
07:16firebot J.C. Jones: Analyzing Let's Encrypt statistics via Map/Reduce
07:41alwuhi, does anyone know how to enable Flash by default in the test? I have a test including an embed tag with SWF source, and I want it can be started immediately, instead of checking the "click-to-play" permission for Flash.
07:42Caspy7alwu: go to the addons manager, Plugins tab, find Flash and set to to "Always" play
07:47alwuCaspy7 : I mean in the mochitest
07:49Caspy7alwu: ah, sorry, used to supporting end users :)
08:08konradosIn console, when I do function X(){}; and then console.log(new X()); I can only see that this is an object, it doesn't tell me that this is an instance of X - am I missing something? How can I see in console what is the type/class of the dumped object?
08:19Caspy7konrados: perhaps also try #webdev
08:30konradosok, thanks Caspy7!
08:31Caspy7alwu: perhaps #build would have some insight?
08:32alwuCapsy7 : ok, thx!
08:32Caspy7alwu: that or maybe #introduction
08:33Caspy7they field a lot of build questions
08:46firebotJust appeared in Planet Mozilla - :
08:46firebot Nick Desaulniers: Submitting Your First Patch to the Linux kernel and Responding to Feedback
08:48nikolamDo you think Firefox would crash, if I run it inside Solaris Zone with resouce limits? Like limited amount of RAM Firefox is allowed to use.
08:49nikolamI think Firefox uses all RAM it can get and wonder if limiting that will just make FF work inside that limit or it would crash when there is no more RAM to use?
09:01Corknikolam: if firefox needs more memory and can't get it the allocation with crash the browser
09:01Corknikolam: so it depends on if you limit the physical or virtual memory (the later will crash firefox if it hits it)
09:02Corknikolam: hmm not fully correct, when firefox hits it, it will try to free things it can free, but if it can't it will crash
09:05nikolamSo why firefox doesn't free things. Like I close many tabs and memory use seems not too less.
09:06nikolamSo if I limit it's maximum RAM use outside FF and doesn't give it any swap, then it will crash sooner or later
09:07nikolamhmmwhatsthisdo, unloading unused tabs..
09:07Corknikolam: this is a fairly complex topic
09:08Corkbut it is about caches, memory fragmentation and gc
09:08Corknormally firefox uses as much memory as it see useful as long as the system doesn't say it has to release it
09:09nikolamLike I use Firefox for Facebook, and that pages loads, and loads and loads.. fill all ram and swaps a lot and then when I am closing tab, it swaps long time, then when closing it closes for a long time, I mostly pkill it
09:09Corkwhen the system tells firefox the system is running low on memory (memory pressure) firefox triggers a loop through its allocations and tries to trim caches and throw away precompiled code and so on
09:09Corkthis means firefox will run quite a bit slower afterwards
09:10Corknikolam: facebook, twitter, gmail and a few other sites like that are known to run a huge amount of javascript
09:10Corkfacebook runs a ton of iframes in the page
09:10Corkthese take up a lot of ram
09:10nikolambut when I close the tab..
09:10Corkand there isn't that much the browser can do about it
09:11Corkthere is a recently closed tab cache
09:12Corkso the tab isn't thrown away in case you for example navigate away from the page and then back to it again
09:12Cork(i don't have the full details about how that works though so take it with a bit of salt)
09:12nikolamso if I want to tell FF to clear RAM for it, I think there are some addons for that..
09:13Corkyou could try about:memory and the gc button
09:13Corki think that throws away as much as it can, but donno about the caches though >_>
09:13Corkbut ya, my point is: it is a lot more complex then it first seams
09:14nikolam My issue is more local, that important parts of UI got swapped out, because FF uses so much RAM, was hoping limiting it's use maximums in-Firefo settigns or something
09:14nikolamok Cork , thanks for comments, let me reboot locally.
09:15nikolamfinal deal would be to run it inside the resource-capped zone so it can't affect other parts of the system.
09:15Corknikolam: i do run firefox in a limited memory situation from time to time (cgroups, sandboxie, jails...)
09:15Corknikolam: and you can definitely do it, you just have to make sure you leave enough ram for it to not become horrible to use, or for it to hit oom
09:16nikolamEnd up pkill it and restarting FF multiple times a day, because it's RAm use just grows, no matter how much tabs is closed
09:16nikolamok, rebooting, will be back, thanks for talk, Cork
09:41firebotJust appeared in Planet Mozilla - :
09:42firebot Christian Heilmann: You dont owe the world perfection! keynote at Beyond Tellerand
09:52Springany update on this issue?
09:52firebotBug 1099204 REOPENED, Textarea text scrolls under padding instead of over
09:53SpringBasically if padding-top/bottom is added to a textarea element any scrolled text will clip to the padding height, rather than push it up/down when scrolled (as if the padding were 'inline' so to speak)
09:59Springalternatively is there's any workaround as I can't think of anything that would do the equivalent
10:15CorkSpring: there is a mozilla only css property if i remember correctly (not sure if it is exposed to webpages though)
10:15Corkbut also worth mentioning, the current rendering is how all browsers do it
10:15Corkand as far as i know no other browser has an option for this
10:17SpringCork, this isn't how Chromium handles it, neither do Safari or IE from reading elsewhere
10:17CorkSpring: the current implementation came after firefox having had a bug about it for many many years
10:17firebotBug 157846 FIXED, Incorrect implementation of padding on textarea elements (scrollbars/resizer wrongly positioned)
10:19Specularcan you recall what the moz-only property would be?
10:19Corkif it actually existed in the finnal patches it would be in that bug ^
10:19Corknot sure they actually went with it at the end
10:19Specularsorry, Hexchat froze and the message log is gone
10:21Specularproblem is only Firefox at this point renders the padding like this. Just tried Opera and it follows Chromium (expectedly, given it now shares the same engine)
10:22Specularsome option would be good to have, as a comment in the ticket pointed out
10:33TimvdeSpecular: This channel is publicly logged:
10:44mbeckHi! Since a couple of weeks (beginning with one of the more recent versions) we experience a problem with FF: After a couple of hours of usage no page can be opened anymore (until restarting FF), neither reloads in existing tabs, nor in new tabs or windows. The tab title shows "Connecting..." with the moving circle, but FF does not even attempt to open a TCP connection (as seen by tcpdump(1)). I'm
10:44mbeckcurrently using 53.0.2 on FreeBSD, but we have the same symptoms on Linux and Windows desktops with recent versions of Firefox. I already tried safe mode with the same result. Any ideas we can debug/resolve this?
12:36Tewmbeck - #necko is the networking code channel, although it's not really a user support you may get help/directions there; aiui resetting Connection settings in about:preferences#advanced should reset the network stack (something to try vs full restart ff, even if it's not really useful). describes how to log net stuff
12:36Tewfrom firefox internals.
12:37philippmbeck: sounds like bug 1360574
12:38firebot FIXED, Firefox stops working after 900 connections when using NTLM proxy
12:39Tewyes, do mention that you use proxy if you do :)
13:23* JacopoTore si chiede se gmail non stia funzionando solo a lui
13:24mbecktew, philipp: yes, I can confirm this is the problem I have (NTLM & 900 conns), thanks!
14:44* WaltS48-away updates to 54.0b8 on Linux.
15:32firebotJust appeared in Planet Mozilla - :
15:32firebot Hacks.Mozilla.Org: Having fun with physics and A-Frame
15:52firebotJust appeared in Planet Mozilla - :
15:52firebot Mozilla Addons Blog: Add-on Compatibility for Firefox 55
15:52firebot QMO: Firefox 54 Beta 7 Testday Results
16:31Springhow can I tell which element is receiving the :active CSS target?
16:32SpringI have a table-cell element that should be styled a certain way with :active (tested using the web inspector :active style checkbox enabled) yet when clicked manually it fails to style
16:33Springit's in combination with the parent elements having pointer-events:none while the child elements have pointer-events:auto. Hover works, just not :active.
16:34Springed: tested without pointer-events and the same result.
16:53MardegSpring: not sure if it'll work, but try adding the attribute tabIndex="0" to the table-cell
16:54SpringMardeg, thanks, works. Problem is I'm styling a page I can't modify so I'll have to figure out something else. Might be JS interfering with the click.
16:58Springthe other issue I discovered is transitions aren't supported for background gradients if it uses color stops :/
17:42est31the title is... not really friendly towards anyone in AOM but google
17:44Caspy7yeah, it occurred to me early on that AV1 would frequently be considered a Google project as clearly they're the center of the universe. Also VP9 was chosen as the base codec to start with
17:45Caspy7not to belittle their contributions
17:46est31mhh maybe its because its xda developers, for them everything starts with google
17:47est31probably even linux was created by google lol
17:47Caspy7thanks for the link...may need to read it later when I'm a bit more awake
17:48Caspy7est31: today's press release
17:48est31wow cool
17:49Caspy7I'm unsure what they'll be contributing to the codec. At least, unsure if they'll contribute much beyond coding the implementation
17:50Caspy7as compared to putting the theory into practice, improving bitrates, etc
17:50est31there has to be some reason
17:51est31maybe they just want to support the effort
17:51Caspy7the more the merrier :)
17:51est31in the best case, they apply their knowledge about writing software to the project
17:51Caspy7well, VLC's a somewhat big fish
17:51Caspy7in some circles I suppose
17:52* est31 is VLC user
17:52Caspy7yeah, which is a bit of what the article gets at
17:52est31I will switch to mpv once I'm dead!
17:53Caspy7that press post also says "The organization also supports and funds other open source projects, such as; FFmpeg, Libav, and Handbrake."
17:53Caspy7which may be relevant
17:54Caspy7I think there's been at least one dev who works on VLC and/or ffmpeg that's been coding AV1 stuff. Notably he made a decoder that worked in VLC for Daala
17:54est31not that the members of aom would be in need of money
17:55est31they have plenty to spend, videolan foundation is small compared even to mozilla, and mozilla is small compared to all other companies
17:55est31like, factor of 1000
17:55est31or 100
18:32iMacuBlock Origin released 3 days ago, already on AMO. uMatrix released on April 7th, still not on AMO. Seriously what is going on here
19:02firebotJust appeared in Planet Mozilla - :
19:02firebot Daniel Pocock: Building an antenna and receiving ham and shortwave stations with SDR
19:39TewiMac - fwiw, ~2 weeks ago ppl were still mentioning 50+ days addon review queues
19:39iMacI understand if there's a queue but why did uBlock Origin get to cut in line?
19:40DanielXIs there any way to have a .wmv file save by default?
19:40DanielXI searched and the question is old, but I see no good answer
19:40Caspy7I don't know if the queue is strict, so that you only have one reviewer and the biggest addons/most changes get to hold the line up for a month
19:41Caspy7smaller addons and updates will get reviewed quicker
19:44kazumaDanielX maybe you can set it in options > applications, to always ask
19:44kazumaor you could disable autoplay of media, and install a plugin like download helper to download the videos
19:45DanielXkazuma : it always asks, that's the trouble. I can't check a "always do this" box.
19:45DanielXIt's not tragic, but it's a common enough file type, you'd think it would be in the mime list or such
19:46Caspy7DanielX: I believe that's usually because the server is sending an improper MIME type
19:46Caspy7octet stream or nothing at all or something, then Firefox doesn't know what type of file it is
19:47DanielXCaspy7 : and just the file extension doesn't help?
19:48Caspy7Firefox is strict about MIME type. I believe it's a security consideration so that the browser doesn't assume xls should open to MS Office, then it's actually some other file that exploits it
19:48Caspy7something to that effect
19:48Caspy7but no, it does not fallback to the extension, purposefully
19:48Caspy7if you want a more confident answer you could ask that in #security :)
19:48DanielXGoogle shows this question is really old, though.
19:49DanielXSaving a file with a wmv extension should be fairly safe.
19:49DanielXI'm near to thinking this is a bug.
19:49kazumathere mp4 exploits no so long ago
19:49kazumastage fright i think it was called
19:50Caspy7this is the server's fault for not setting the proper MIME type
19:51DanielXCaspy7 : perhaps. This is a download server, though, so it's not really meant to be played at all.
19:52DanielXWell. Let's call it one more report of the problem, then.
20:33firebotJust appeared in Planet Mozilla - :
20:33firebot Mozilla Marketing Engineering & Ops Blog: MozMEAO SRE Status Report - 5/16/2017
21:23firebotJust appeared in Planet Mozilla - :
21:23firebot Air Mozilla: Rust Libs Team Meeting 2017-05-16
22:33firebotJust appeared in Planet Mozilla - :
22:33firebot The Servo Blog: This Week In Servo 102
23:00fiueqhi, how come install ublock origin from the site gives you a different addon vs installing it from the "get addons" tab in the firefox addons menu
23:02Caspy7fiueq: can you provide links to these two addons?
23:02MossopI just installed it from get add-ons and see the same version number as is listed here
23:03fiueqversion number is the same from and then click the menu in firefox and selecting "addons" and "get addons" then selecting ublock as a featured addon
23:03fiueqbut the author is different
23:03fiueqone is "raymond hill" the other is "all ublock contributors"
23:04Caspy7fiueq: I wonder somehow you ended up at ublock?
23:04MossopBoth say Raymond Hill for me
23:05fiueqMossop, even on a fresh profile/
23:05Caspy7as long as it's uBlock Origin, I don't think there's anything to worry about
23:06MossopPossibly you're seeing a metadata update problem
23:06fiueqI realise its very minor but it was a tad spooky
23:07MossopYeah, the install.rdf specifies that as the creator then after a metadata update from AMO it will change to Raymond Hill. Something might be going wrong there. It's the same add-on though
23:09Caspy7fiueq: just tested on a fresh profile and it said Raymond Hill
23:10fiueqCaspy7, after installing it from the "get addons" tab?
23:10fiueqstrange, for me it gives the contributors one every single time
23:11fiueqon a fresh profile
23:11MossopDo you have AV software installed?
23:13fiueqI think i found a lead, my original tests when I thought I was using a clean profile were incorrect, it was pulling my user.js with some prefs, a truly clean profile it is normal
23:14fiueqis there an about:config change that would cause that?
23:15MossopThere are a few yes, under the extensions branch
23:16fiueqI found it, I had: user_pref("extensions.getAddons.cache.enabled", false);
23:16MossopThat'll do it
23:17fiueqI believe I found that setting on a "privacy" focused site
23:18fiueqsorry for wasting your time, but thank you Mossop and Caspy7 for the help :)
23:20fiueqI think I need to look into what those settings are actually doing because I just copied/pasted them because they said it makes firefox safer
23:22Caspy7fiueq: I reflexively get angry every time I see a list like that get released
23:22Caspy7they're almost always flawed and problematic
23:22MossopThat setting stops Firefox from downloading additional metadata about add-ons from AMO, including some information about whether add-ons are compatible with your version of Firefox or not. I'd argue it makes you less safe
23:22fiueqhow so?
23:22Caspy7crap like this
23:22Caspy7stuff breaking
23:23Caspy7maybe next release, maybe not noticed
23:23Caspy7fiueq: people making changes they don't understand and recommend that others do the same so they "feel" more secure
23:24fiueqyeah pretty much nailed it for me, I don't know what 80% of them do but it does make me "feel" better
23:24Caspy7and it's usually crap
23:25Caspy7the stuff they think they understand is often misunderstood
23:26fiueqmind taking a look at the one I found and see if there are any glaring faults?
23:28Caspy7you're welcome to link here, but security isn't really my bag. Also I'm kind of busy and not fully engaged.
23:28Caspy7(or at least distracted)
23:35fiueqI can't find the original site where I got it but thats what was there
23:35fiueqthe top ones didn't have comments on them
23:36Caspy7fiueq: how does disabling autocomplete from your own bookmarks and history help give you increased privacy?
23:37Caspy7or the PDF viewer?
23:38Daggersomewhat useful if your mother is looking over your shoulder, say
23:39Daggeralthough that'd be better solved with a leading character or somesuch (like the search restriction characters) rather than globally disabling it all the time
23:40Caspy7Dagger: solved:
23:41Daggernot very subtle
23:41Daggerthe workaround I just came up with was to create a keyword search for URL "%s" and keyword , or something easy to type but unlikely to be in history
23:41Caspy7We should call it the Snowden Sweater
23:42Daggerworks fine so long as you don't mind a fixed protocol
23:42Daggeror something *else* easy to type. is very easy on a UK keyboard, for some unknown reason
23:44Dagger...and so long as you're only typing a hostname and no port, path, query string etc
23:46fiueqyeah lol, I think I'm just going to scrap it completely
23:46fiueqI've set windows to auto get security updates and have kaspersky, hopefully thats good enough
23:47fiueqwas there anything "bad" in there? it'd be ironic if something in there caused me to be less secure
23:47Caspy7fiueq: recommend to keep the antivirus out of the browser
23:48Caspy7that includes addons and https interception or whatever
23:49Caspy7fiueq: yeah actually, it disables Safe Browsing. I've seen this a lot, but people are wrong about what they say
23:51Daggersafe browsing does involve sending quite a few URLs off to Google (or hashes of URLs, which frankly isn't much different when you're Google and have a giant lookup table you can use to figure out what the hash corresponds to)
23:51Caspy7fiueq: it does not send back all your URLs to Google. It checks URLs clientside. When it encounters one, it doublechecks with Google, but it's strongly obfuscated. It's like a hash and even then that send 3 separate ones so Google wouldn't know which one was the real one
23:52Caspy7I do not think "quite a few" is accurate here. It's only when you hit something that's in the malware list
23:53Daggerthe local db contains the first N bytes of a hash of the URL. any URL that matches the first N bytes of any URL in the local db will be sent off
23:53Caspy7with the extra two hashes, the data becomes too noisy/inaccurate to be of value
23:55fiueqI thought safebrowsing was enabled user_pref("browser.safebrowsing.enabled", true);
23:57Caspy7fiueq: told you I was tired. I just noticed the "safe browsing" bit. Don't know much about the downloaded files bit
23:57fiueqoh ok sorry
23:59Caspy7oops, didn't explicitly say I was tired. Well, there you go.
17 May 2017
No messages
Last message: 126 days and 18 hours ago