MozillaZine

The first official 20190524 builds are out

Discussion about official Mozilla Firefox builds
Josa
 
Posts: 3905
Joined: July 28th, 2009, 4:52 pm

Post Posted May 23rd, 2019, 2:02 pm

Previous Nightly Builds thread 20190523

The first official 201905024 NIGHTLY builds are out
The Official Firefox 68 Beta 4 builds are out

Previous NIGHTLY: 20190523 (Fx 69.0a1)
Previous BETA: 20190522 (Fx 68.0b3)

Mozilla Inbound Builds
Treeherder
Release Notes: NIGHTLY | BETA
Changelogs: NIGHTLY | BETA


Fixed:
    NIGHTLY +71
  1. #1543092
  2. #1550567 [Core:Audio/Video: Playback]-Flip media.rdd-vorbis.enabled to true on Linux [Uns][]
  3. #1534081 [Core:Audio/Video: Playback]-Flip media.rdd-vorbis.enabled to true on macOS [Uns][]
  4. #1548629 [Core:Audio/Video: Playback]-Intermittent dom/media/test/test_streams_element_capture.html | Got 0.2699319727891156, expected at least 0.27; gizmo-short.mp4-56 current time at end [Uns][]
  5. #1553555 [Core:Audio/Video: Playback]-Release RemoteDecoderChild on manager thread in the case of an InitIPDL error [Uns][]
  6. #1552354 [Core:Audio/Video: Playback]-Video without audio ended too early before it really reaches to the end [Uns][]
  7. #1429301 [Core:CSS Parsing and Computation]-Implement offset-rotate for motion path [Uns][]
  8. #1541233 [Core:CSS Parsing and Computation]-Stop loading userContent.css or userChrome.css by default unless a preference is set [Uns][[layout:triage-discuss][fxperf:p1][fxperfsize:XS]]
  9. #1553252 [Core:CSS Parsing and Computation]-Use cbindgen for text-overflow. [Uns][]
  10. #1536672 [Core:Canvas: WebGL]-Implement webgl2 draft ext OVR_multiview2 [Uns][gfx-noted]
  11. #1553021 [Core:DOM: Animation]-Update getAnimations parameter to match latest spec [All][]
  12. #1552268 [Core:DOM: Content Processes]-Allow omitting `moduleURI` for JSWindowActor declarations [Uns][]
  13. #1552751 [Core:DOM: Core & HTML]-MessagePort should not work, silently failing, when created from removed windows. [Uns][]
  14. #1370630 [Core:DOM: Events]-preventDefault() on form.dispatchEvent(new Event('submit'))? [Uns][]
  15. #1553400 [Core:DOM: Networking]-XMLHttpRequestDoneNotifier should own XHR, not vice-versa [Uns][[necko-triaged]]
  16. #1380905 [Core:Disability Access APIs]-Assertion failure: mRawPtr != nullptr (You can't dereference a NULL RefPtr with operator->().) [@mozilla::a11y::HTMLComboboxAccessible::Shutdown()] [Uns][a11y:crash-shutdown]
  17. #1551825 [Core:Disability Access APIs]-Intermittent TEST-UNEXPECTED-FAIL | LeakSanitizer | leak at js_pod_arena_malloc, maybe_pod_malloc, js::Nursery::allocateBuffer, js::Nursery::reallocateBuffer in accessible/tests/browser/events/ [Uns][[retriggered][stockwell disable-recommended]]
  18. #1534394 [Core:Editor]-Crash [@get] near [@mozilla::WSRunObject::GetWSNodes()] [Uns][]
  19. #1552549 [Core:General]-Update Rand crate dependency and accept the license that comes with it [Uns][]
  20. #1552687 [Core:Graphics: Text]-Crash in [@ ComPtr<T>::~ComPtr<T>] [Win][]
  21. #1551520 [Core:Graphics: WebRender]-1.53 - 3.93% displaylist_mutate (linux64-shippable-qr, windows10-64-shippable-qr) regression on push 73254a69497b209b372ed209ce9d004080ea8052 (Fri May 10 2019) [All][]
  22. #1531142 [Core:Graphics: WebRender]-Long pinch actions seem to continuously increase CPU (compositor) time [Uns][[gfx-noted][wr-amvp][wr-q2][wr-april]]
  23. #1552984 [Core:Graphics: WebRender]-Refactor parts of the WR batching and flattening code, to support future picture caching improvements. [Uns][]
  24. #1553701 [Core:Graphics: WebRender]-Remove the clip chain stack during the prepare_prims pass in WR [Uns][]
  25. #1525720 [Core:Graphics]-Fission: Enable/disable rendering of OOP iframes when switching tabs [Uns][]
  26. #1546432 [Core:Internationalization]-Migrate Fluent's MutationObserver to C++ [Uns][]
  27. #1368266 [Core:JavaScript Engine: JIT]-Crash in js::jit::BaselineScript::nativeCodeForPC [Lin][]
  28. #1543458 [Core:JavaScript Engine]-Convert over various wholly-internal, non-web-exposed users of UTF-8 compilation to non-inflating compilation [Uns][]
  29. #1552875 [Core:JavaScript Engine]-Field initialized to unnamed function should name the function [Uns][]
  30. #1552528 [Core:JavaScript Engine]-Hook up structured spewer to Gecko profiler start/stop [Uns][]
  31. #1552470 [Core:JavaScript Engine]-TypeScript refactorings to prepare for bug 1551796 [All][]
  32. #1553491 [Core:JavaScript: GC]-Renumber MarkColor to match intuition that 'black' is more marked than 'gray' [Uns][]
  33. #1553710 [Core:JavaScript: GC]-Sort the available chunks list [Uns][[MemShrink?]]
  34. #1552401 [Core:JavaScript: GC]-When decommit runs on the main thread it is accounted in the compact phase. [Uns][]
  35. #1553022 [Core:Layout: Scrolling and Overflow]-Make scroll-snap-type-on-root-element.html in wpt pass [Uns][]
  36. #1546219 [Core:Layout: Scrolling and Overflow]-Rename mScrollSnapType{X,Y} to mScrollSnapStrictness{X,Y} to match what it is [Uns][]
  37. #1553548 [Core:Layout: Tables]-CSS `contain` needs to inherit from table to table-wrapper-box [Uns][]
  38. #1544242 [Core:Layout]-Assertion failure: !IsLazilyCascadedPseudoElement() (Lazy pseudos can't inherit lazy pseudos) with ::marker { display: list-item } [Uns][]
  39. #1550105 [Core:Layout]-Negative values in meta viewport result in different initial viewports on Firefox and Chrome [Uns][[webcompat]]
  40. #1552287 [Core:Layout]-[css-contain] contain:layout on various frames doesn't fully suppress baseline [Uns][]
  41. #1553680 [Core:Networking: Cookies]-Read the Cookie response headers from a `white-space: pre' element so that any white-space in them would be preserved in the expectation value [Uns][]
  42. #1492305 [Core:Security: PSM]-NSS-implemented OSKeyStore promises does not reject after hitting cancel on master passsword prompt [Uns][[psm-backlog] [psm-wouldtake]]
  43. #1552310 [Core:Security: PSM]-fix deleting preloaded intermediates by using the right field [Uns][[psm-blocked][psm-assigned]]
  44. #1553766 [Core:Serializers]-Encapsulate range boundary specific members in `nsDocumentEncoder` [Uns][]
  45. #1553480 [Core:Serializers]-Make resetting `nsIDocumentEncoder`'s members consistent [Uns][]
  46. #1428913 [Core:User events and focus handling]-Don't allow entering fullscreen on right / middle click [Uns][]
  47. #1553213 [Core:WebRTC]-AddressSanitizer: SEGV /builds/worker/workspace/build/src/obj-firefox/dist/include/mozilla/RefPtr.h:268:27 in get near [mozilla::dom::MediaDevices::GetDisplayMedia] [Uns][[fuzzblocker]]
  48. #1553567 [Core:XPCOM]-Using -1 to initialize nsThread::mCurrentEventLoopDepth is cruel [All][]
  49. #1551320 [Core:XUL]-In XUL documents replace all callers of createElement with createXULElement [Uns][]
  50. #1553473 [DevTools:Console]-Warning: Failed prop type: The prop * is marked as required [Uns][]
  51. #1552453 [DevTools:Debugger]-Breakpoint doesn't pause in file during page load [Uns][[debugger-mvp]]
  52. #1550495 [DevTools:Debugger]-Set directory root does not work for debugger.html [Uns][[debugger-mvp]]
  53. #1553556 [DevTools:General]-Blurry warning icon (themes/images/alert.svg) on 1x displays [All][]
  54. #1341155 [DevTools:Netmonitor]-Show full path in File column in developer tools [Uns][]
  55. #1357774 [DevTools:Responsive Design Mode]-Update orientation state (window.screen.orientation) in Responsive Design Mode [Uns][[rdm-mvp] [dt-q]]
  56. #1553387 [Firefox:General]-'Window' menu mostly empty (missing the list of windows) since switching to browser.xhtml [Uns][]
  57. #1551334 [Firefox:General]-Update hiddenWindow.xul on OSX to xhtml [Uns][]
  58. #1517475 [Firefox:Normandy Client]-Execute recipe runner on Remote Settings "sync" event [Uns][]
  59. #1553198 [Firefox:Normandy Client]-Preference Experiments are expired in the same session they are fetched with “app.normandy.dev_mode” set to true [All][]
  60. #1553062 [Firefox:Normandy Client]-Send events when the recipe runner starts and ends [Uns][]
  61. #1553142 [Firefox:Remote Settings Client]-Leverage Console.jsm to add more logging to Remote Settings [Uns][]
  62. #1553498 [Firefox:Security]-Create about:certificates [Uns][]
  63. #1552537 [Toolkit:Add-ons Manager]-[about:addons] Plugins are missing preferences link [Uns][]
  64. #1552853 [Toolkit:Application Update]-Update downloads get stuck at 'Applying update' state in the UI (updates continue to work) when Mozilla Maintenance Service is disabled [Win][]
  65. #1543829 [Toolkit:Storage]-add LMDB's OpenLDAP Public License (Version 2.8) to about:license [Uns][]
  66. #1550692 [Toolkit:Telemetry]-Process subsection headers in about:telemetry should be clear that they're about processes [Uns][[good second bug][lang=js]]
  67. #1544228 [Toolkit:Themes]-about:buildconfig styling has bad contrast, links with Adwaita-dark GTK theme [Uns][]
  68. #1547795 [Toolkit:Video/Audio Controls]-Hide the Picture-in-Picture toggles on videos where it'll likely be more annoying than not [Uns][]
  69. #1528459 [Toolkit:Video/Audio Controls]-Tab should indicate that a <video> inside of it is being viewed in Picture-in-Picture, and perhaps not display the audio indicator [Uns][]
  70. #1519577 [Toolkit:XUL Widgets]-Convert toolbarbutton bindings to custom elements [Uns][[xbl-available]]
  71. #1543384 [WebExtensions:General]-AsyncShutdownTimeout timeout in "Startup: Run manifest" [Uns][]

    Partial Landings/Diagnostic Patches:
  1. #1532174 [Core:Graphics: WebRender]-Fix inconsistent transform flattening logic [Uns][[wr-q2][wr-may]]
  2. #1551796 [Core:JavaScript Engine: JIT]-Merge TypeScript and ICScript into JitScript [All][]
  3. #1383650 [Core:SVG]-Implement SVG geometry properties in CSS [Uns][[webcompat][wptsync upstream]]

    Beta 3 -> Beta 4 Changelog +10
  1. #1547757
  2. #1552998
  3. #1552719 [Core:Layout]-nsBulletFrame::Ordinal is slow [Uns][[qf]]
  4. #1551729 [Core:Networking: Cookies]-Cannot login to Verizon Fios Quantum Gateway router [All][[necko-triaged]]
  5. #1553028 [DevTools:about:debugging]-[remote-dbg-next] new about:debugging should be enabled only on Nightly and Dev Edition [Uns][[qa-triaged]]
  6. #1553029 [DevTools:about:debugging]-[remote-dbg-next] WebIDE deprecation message should only be displayed if new about:debugging is enabled [Uns][[qa-triaged]]
  7. #1553031 [DevTools:about:debugging]-[remote-dbg-next] remove "Remote Debugging" menu entry if new about:debugging is disabled [Uns][[qa-triaged]]
  8. #416771 [Firefox:Tabbed Browser]-Allow window.focus() to switch tabs at times when popups are allowed [All][[platform-rel-Outlook][webcompat:p1]]
  9. #1550697 [Firefox:Toolbars and Customization]-The "Restart to update Firefox" menu item gives no visual indication when selected through the keyboard [Uns][]
  10. #1552248 [WebExtensions:Request Handling]-proxy.onRequest is limited to http/https [Uns][]

Nightly 69 fixes since 20190520 (Gecko 68) ~226
Beta 68 fixes since 20190318 (Gecko 67) ~2764

Nightly Blog
Release Calendar
Release Tracking
MozRegression Guide
Last edited by Josa on May 24th, 2019, 6:02 pm, edited 1 time in total.

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 23rd, 2019, 2:05 pm

DinGo wrote:The memory leak issue is still there in 68.0 beta4.


Yep, has the bug been reported?

A temp quick fix, it is just one exe instance, kill it and you are good to go for good, and nothing will crash, not even a single tab.

Edit: Beta 04 has been releases https://archive.mozilla.org/pub/firefox ... es/68.0b4/

Are You A Wiiizard?
 
Posts: 380
Joined: August 30th, 2012, 4:27 am

Post Posted May 23rd, 2019, 10:46 pm

It is just one process, if you killed it, Firefox will make another process doesn't increased.
Is this bug reported?

A temp quick fix, it is just one exe instance, kill it and you are good to go for good, and nothing will crash, not even a single tab.


I'm not sure if it is reported but it's been in Nightly now for a long time, surely someone has but I'm not good at finding reported bugs.
Killing the process isn't an option for me as it's always the main process that experiences the leak. Trying to close the browser doesn't work either (it visibly closes but the process stays open with the main process using max CPU). That means I need to kill the main process just to have the browser in a working state (if I let the browser open for much longer than an hour everything gets extremely laggy).

Update: Firefox 68 beta 04 doesn't fix the memory bug problem!


Considering it's not fixed in Nightly, it will probably be a long time before it's fixed in beta.

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 24th, 2019, 5:52 am

Are You A Wiiizard? wrote:
I'm not sure if it is reported but it's been in Nightly now for a long time, surely someone has but I'm not good at finding reported bugs.
Killing the process isn't an option for me as it's always the main process that experiences the leak. Trying to close the browser doesn't work either (it visibly closes but the process stays open with the main process using max CPU). That means I need to kill the main process just to have the browser in a working state (if I let the browser open for much longer than an hour everything gets extremely laggy).


I am not sure if has something to do with nVidia, Tested in virtualbox couldn't address the issue,

Did you try disabling hardware acceleration.

About killing the process, and this why I am suspecting nvidia, cause it is the first process which uses GPU, the leak in VRAM will start long before the leak in RAM. Notice the VRAM in the picture

look at the image, and try using process explorer instead of windows task manager:

https://i.imgur.com/FBlHi0O.png
Last edited by smsmith on May 24th, 2019, 8:08 pm, edited 1 time in total.
Reason: removed image tags as the size of the image was breaking the forum layout - smsmith/moderator

TheVisitor
 
Posts: 4899
Joined: May 13th, 2012, 10:43 am

Post Posted May 24th, 2019, 6:04 am

The highlighted line in your graphic is the 'Total' of all the Firefox processes listed below that line. That is NOT a leak IMO. The top line of 442meg is most likely the VRAM used by the GPU/SYSTEM, and I would also not consider that a leak.

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 24th, 2019, 6:07 am

TheVisitor wrote:The highlighted line in your graphic is the 'Total' of all the Firefox processes listed below that line. That is NOT a leak IMO. The top line of 442meg is most likely the VRAM used by the GPU/SYSTEM, and I would also not consider that a leak.

Nope, it is not the total,

Here is a new pic after few minutes, the second picture is after killing the process, https://imgur.com/a/azjipEv

First image

https://i.imgur.com/kAsKPVN.png

Second image after killing the process


https://i.imgur.com/1bhJS0k.png
Last edited by smsmith on May 24th, 2019, 8:09 pm, edited 3 times in total.
Reason: removed image tags as the size of the image was breaking the forum layout - smsmith/moderator

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 24th, 2019, 6:08 am

jeny wrote:
TheVisitor wrote:The highlighted line in your graphic is the 'Total' of all the Firefox processes listed below that line. That is NOT a leak IMO. The top line of 442meg is most likely the VRAM used by the GPU/SYSTEM, and I would also not consider that a leak.

Nope, it is not the total,

Here is a new pic after few minutes, the second picture is after killing the process, https://imgur.com/a/azjipEv


BTW, killing the process will not crash any of the opened tabs

knowguy
 
Posts: 5
Joined: February 13th, 2018, 8:01 am

Post Posted May 24th, 2019, 6:19 am

Has anyone reported this bug or is it already being tracked?

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 24th, 2019, 6:24 am

knowguy wrote:Has anyone reported this bug or is it already being tracked?


I send a feedback using the submit feedback in the help and not happy,

Not sure if being reported, but I googled a lot and found nothing, I am thinking it is not being reported

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 24th, 2019, 6:35 am

TheVisitor wrote:The highlighted line in your graphic is the 'Total' of all the Firefox processes listed below that line. That is NOT a leak IMO. The top line of 442meg is most likely the VRAM used by the GPU/SYSTEM, and I would also not consider that a leak.



Just noticed, the VRAM is to the right, just scroll to the right

The PID 1440 only, which has the leak, has VRAM of over 2GB Almost 3GB in the second picture

You have read the whole thing so wrong.


Edit: Leaving the process will use over 10gb RAM just within few minutes

Are You A Wiiizard?
 
Posts: 380
Joined: August 30th, 2012, 4:27 am

Post Posted May 24th, 2019, 7:01 am

jeny wrote:
Are You A Wiiizard? wrote:I'm not sure if it is reported but it's been in Nightly now for a long time, surely someone has but I'm not good at finding reported bugs.
Killing the process isn't an option for me as it's always the main process that experiences the leak. Trying to close the browser doesn't work either (it visibly closes but the process stays open with the main process using max CPU). That means I need to kill the main process just to have the browser in a working state (if I let the browser open for much longer than an hour everything gets extremely laggy).

I am not sure if has something to do with nVidia, Tested in virtualbox couldn't address the issue,
Did you try disabling hardware acceleration.


I'm using AMD (RX570 4GB) though but it may be the graphics part of Firefox since FF is using 2.4GB of dedicated VRAM for having essentially nothing open which seems wrong. I haven't had any time to test anything yet.

jeny
 
Posts: 25
Joined: May 22nd, 2019, 11:55 am

Post Posted May 24th, 2019, 7:03 am

Are You A Wiiizard? wrote:
jeny wrote:
Are You A Wiiizard? wrote:I'm not sure if it is reported but it's been in Nightly now for a long time, surely someone has but I'm not good at finding reported bugs.
Killing the process isn't an option for me as it's always the main process that experiences the leak. Trying to close the browser doesn't work either (it visibly closes but the process stays open with the main process using max CPU). That means I need to kill the main process just to have the browser in a working state (if I let the browser open for much longer than an hour everything gets extremely laggy).

I am not sure if has something to do with nVidia, Tested in virtualbox couldn't address the issue,
Did you try disabling hardware acceleration.


I'm using AMD (RX570 4GB) though but it may be the graphics part of Firefox since FF is using 2.4GB of dedicated VRAM for having essentially nothing open which seems wrong. I haven't had any time to test anything yet.


Couldn't address this issue in virtualbox, but then in virtual box FF didn't use VRAM at all although 2d and 3d are enabled,

So it has something to do with the graphic card

knowguy
 
Posts: 5
Joined: February 13th, 2018, 8:01 am

Post Posted May 24th, 2019, 8:32 am

68 Beta 3 was great for me but I am noticing a slow creep up in memory in Beta 4 for some reason.

Fanolian
 
Posts: 859
Joined: February 21st, 2011, 10:14 am

Post Posted May 24th, 2019, 9:41 am

jeny wrote:Couldn't address this issue in virtualbox, but then in virtual box FF didn't use VRAM at all although 2d and 3d are enabled,
So it has something to do with the graphic card


Go to about:memory and take a measure. It may tell you what is using your RAM. Also on the right side it shows what the PID is responsible for.

bjherbison
 
Posts: 836
Joined: October 6th, 2003, 5:40 am
Location: Bolton, MA, US

Post Posted May 24th, 2019, 10:10 am

And when you file the bug it would also be useful to know how much GC , CC, and Minimize memory use reduce the size.

Return to Firefox Builds


Who is online

Users browsing this forum: No registered users and 2 guests