The Official Win32 20160629 builds are out

[Josa]

Previous Nightly Builds thread 20160628

The Official Win32 20160629 NIGHTLY build is out
The Official Win32 20160629 AURORA build is out
.The Official Firefox 48 Beta 5 builds are not yet out

Previous NIGHTLY: 20160628 (Fx 50.0a1)
Previous AURORA: 20160628 (Fx 49.0a2)
Previous BETA: 20160628 (Fx 48.0b4)

Hourly NIGHTLY builds: mozilla-central-linux | mozilla-central-linux64 | mozilla-central-macosx | mozilla-central-win32 | mozilla-central-win64 | NIGHTLY Tinderbox | NIGHTLY Changelog | AURORA Tinderbox | AURORA changelog

Fixed:

1. NIGHTLY +45
2. #1270537
3. #1280043
4. #1282113
5. #1281632 [Core:Audio/Video: Playback]-Refactor create MediaDataDecoder arguments into CreateDecoderParams struct [Uns][]
6. #1281991 [Core:Canvas: WebGL]-Clean up WebGLSync objects before WebGL context lost [All][]
7. #1281687 [Core:Canvas: WebGL]-Update ANGLE to chromium/2776 [Win][]
8. #1244590 [Core:DOM: Animation]-implement KeyframeEffectReadOnly spacing modes [Uns][]
9. #1221730 [Core:DOM: Device Interfaces]-Move gamepad API to PBackground [Uns][]
10. #1280106 [Core:DOM: Push Notifications]-Crash in mozilla::dom::PushSubscriptionChangeDispatcher::NotifyWorkers [Win][btpp-active]
11. #1207744 [Core:DOM: Push Notifications]-unregisters lost on reconnect [Uns][btpp-active]
12. #1188642 [Core:DOM: Security]-Use channel->ascynOpen2 in dom/base/nsObjectLoadingContent.cpp [Uns][[domsecurity-active]]
13. #1269162 [Core:DOM: Workers]-Remove separate worker binding for XMLHttpRequest and friends [Uns][btpp-backlog]
14. #1282253 [Core:DOM]-Get rid of XHR::ResponseTypeEnum [Uns][]
15. #1282341 [Core:Graphics: Layers]-Add ImageContainer::ClearCachedResources() [Uns][]
16. #1261166 [Core:Graphics: Layers]-Consider using IOSurface for sharing content-painted textures to the compositor on OS X [Mac][gfx-noted]
17. #1282347 [Core:Graphics: Layers]-Use gfx::IntSize in CompositorBridgeParent ctor [Uns][]
18. #1282248 [Core:Graphics: Text]-constify a bunch of gfxTextRun methods [Uns][]
19. #1281105 [Core:Graphics]-6.9 - 25.13% sessionrestore / sessionrestore_no_auto_restore / tp5o Main_RSS / tresize / ts_paint (linux64) regression on push b3930a21b6ed (Wed Jun 8 2016) [Uns][[talos_regression] gfx-noted]
20. #1282397 [Core:Graphics]-[Static Analysis][Explicit null dereferenced] In function ssse3_fetch_bilinear_cover [All][CID 1362893]
21. #1272908 [Core:JavaScript Engine]-Assertion failure: !cx->isExceptionPending(), at js/src/vm/Debugger.cpp:1866 [Mac][[jsbugmon:update]]
22. #1268526 [Core:JavaScript Engine]-Crash [@ js::Debugger::removeDebuggeeGlobal] [Mac][[jsbugmon:update]]
23. #1280042 [Core:JavaScript Engine]-If class B extends class A and constructor is omitted for B, then var b = new B(); b.constructor.name returns 'A' [All][]
24. #1202665 [Core:JavaScript Engine]-In an ES6 class, `"constructor"(){}` with double quotes is not recognized as a constructor [Uns][]
25. #1270331 [Core:JavaScript Engine]-Pass JS_IS_CONSTRUCTING as |this| to constructing functions from bytecode [Uns][]
26. #1281450 [Core:JavaScript Engine]-Speed up receiver wrapping in CrossCompartmentWrapper::{get, set} [All][]
27. #1281988 [Core:Layout: Text]-The bidi engine resolves levels incorrectly when there are continuous formatting characters [Uns][]
28. #1281424 [Core:Layout: Text]-text-orientation:upright should not affect bidi directionality in sideways-* writing modes [Uns][]
29. #1279340 [Core:Plug-ins]-PPluginSurfaceChild tries to send a delete message from PluginInstanceChild::ActorDestroy() [Win][]
30. #1281457 [Core:Selection]-Word with "text-align: justify" moves as text selection changes [Uns][]
31. #1280796 [Core:Widget: Gtk]-Undoing committed strings by Input Method is strange. [Uns][]
32. #1249352 [Core:XPCOM]-Make NS_EscapeURL more fallible [Uns][]
33. #1145655 [Firefox:Developer Tools: Debugger]-Remove remnants of Browser Debugger [All][[good first bug]]
34. #1281489 [Firefox:Developer Tools: Shared Components]-Reps: Grips rep renders only 3 properties in long mode [Uns][[devtools-html]]
35. #1272531 [Firefox:General]-Advanced panel does not occupy full content width in about:net/certerror. [Uns][[fxprivacy]]
36. #1282223 [Firefox:General]-Remove gPermissionObserver [Uns][]
37. #1241746 [Firefox:General]-layout of the "secure connection failed" results in poor wrapping of the word "sites" [Uns][[good first bug]]
38. #1279120 [Firefox:Search]-ContentSearchUI.js leaks a blob URL when you change default search engine [Uns][[MemShrink]]
39. #1243034 [Firefox:Sync]-"browser.download.useDownloadDir" preference not synced [Uns][]
40. #1229927 [Firefox:Tours]-Allow `showHeartbeat` to change the icon [Uns][[lang=js][good first bug]]
41. #1277295 [Toolkit:Add-ons Manager]-Extension installation failing with NS_ERROR_XPC_GS_RETURNED_FAILURE for Services.storage [Uns][triaged]
42. #1279034 [Toolkit:Add-ons Manager]-[Disco Pane] Expose the abilty to enable/disable an add-on from content. [Uns][]
43. #1272585 [Toolkit:Application Update]-Handle NS_ERROR_DOCUMENT_NOT_CACHED when downloading mar files. [Uns][]
44. #1233086 [Toolkit:Notifications and Alerts]-http icons do not work on linux native notifications [Uns][]
45. #1025965 [Toolkit:Safe Browsing]-Rename browser.safebrowsing.enabled to browser.safebrowsing.phishing.enabled [All][]
46. #1233986 [Toolkit:Telemetry]-Move clientId loading, generation and serialization earlier during startup [All][[measurement:client]]

Partial Landings/Diagnostic Patches:

1. #1247977 [Core:Canvas: WebGL]-crash in mozilla::WebGLContext::FakeBlackTexture::FakeBlackTexture regressing in 45 [Win][gfx-noted]
2. #1239188 [Core:Graphics: Layers]-Timeout in D3D lock mutex timeout, followed by a crash [Win][]
3. #1271649 [Core:JavaScript Engine]-Implement a C++ interface for Debugger.Environment instances. [Uns][[devtools-html]]
4. #1281529 [Core:JavaScript Engine]-Make JSContext inherit from JSRuntime [All][]
5. #1229813 [Core:JavaScript Engine: JIT]-Enable branch pruning by default. [Uns][]

1. AURORA +8
2. #1279898 [Core:JavaScript Engine: JIT]-Wrong alias analysis for buffer shared with two typed arrays [Uns][]
3. #1271523 [Core:Widget: Gtk]-tooltips are not themed when running against gtk+ 3.20 [Uns][[uplift to aurora after bug 1277818]]
4. #1275700 [Core:Widget: Gtk]-[Ubuntu] AwesomeBar and Search Pop-up is too Dark [Lin][[fxsearch]]
5. #1277818 [Core:Widget: Gtk]-[GTK+ 3.20] Export CreateCSSNode() and potentially divide pre/post 3.20 code. [Lin][]
6. #1277199 [Firefox:Developer Tools]-The command-button-frames in toolbox-buttons does not have the open attribute [Uns][[devtools-html]]
7. #1279649 [Firefox:Developer Tools: Inspector]-The inspector-pane-toggle should be rotated 90 deg, when docking to side of browser and vertical mode [All][[devtools-html]]
8. #1272531 [Firefox:General]-Advanced panel does not occupy full content width in about:net/certerror. [Uns][[fxprivacy]]
9. #1263774 [Toolkit:Breakpad Integration]-No memory reports in content process crash reports [Uns][[fce-active]]

1. BETA +2
2. #1275700 [Core:Widget: Gtk]-[Ubuntu] AwesomeBar and Search Pop-up is too Dark [Lin][[fxsearch]]
3. #1272531 [Firefox:General]-Advanced panel does not occupy full content width in about:net/certerror. [Uns][[fxprivacy]]

Nightly 50.0 fixes since 20160606 (Gecko 49) ~727
Aurora 49.0 fixes since 20160425 (Gecko 48) ~1972
Beta 48.0 fixes since 20160307 (Gecko 47) ~2585

Release tracking Firefox 48-50 see HERE

[mdew]

Is it, or has nightly performance has degraded recently?

[dbcooper.dk]

For those with AMD GPUs who are concerned by black screens etc:

https://bugzilla.mozilla.org/show_bug.cgi?id=1280081

--- Comment #13 from Milan Sreckovic [:milan] <milan@mozilla.com> 2016-06-28 10:55:46 PDT ---
A number of these should be resolved in the (not yet released) 16.300.*
drivers.

[Omega X]

Is it, or has nightly performance has degraded recently?

Its been that way for a while now. Some sort of memory leak is keeping ghost windows around.

[Josa]

Is it, or has nightly performance has degraded recently?

Its been that way for a while now. Some sort of memory leak is keeping ghost windows around.

You guys did set network.predictor.enable-prefetch to false already, right?

[itisomegakai]

Is it, or has nightly performance has degraded recently?

Its been that way for a while now. Some sort of memory leak is keeping ghost windows around.

if you guys use ABP disable 2.7.3 beta,
it causing hangs & high cpu/gpu usage.

also TMP
http://tabmixplus.org/forum/viewtopic.p ... 18&p=70855

can anyone notify the developer of ABP

[GHM113]

can anyone notify the developer of ABP

Hello! That's my sheet over there, ABP developer has already been notified:
https://bugzilla.mozilla.org/show_bug.c ... 192585#c18
https://bugzilla.mozilla.org/show_bug.cgi?id=1280368#c1
Also, shortly afterwards this bug was created as an attempt to improve ABP&e10s performance: https://bugzilla.mozilla.org/show_bug.cgi?id=1280368

Raymond Hill, the creator of uBlock Origin, likely saw that sheet too but it looks like I couldn't convince him uBlock Origin causes performance problems. I guess I'll also post my sheet here so that everyone could see what I am talking about:

[bjherbison]

The biggest problem I see appears to be an interaction between Tab Mix Plus and uBlock/AdBlock Plus. Have you talked to the TMP developer about that?

[skootb]

why am I now getting this message?

"a recommended security and stability update is available, but you do not have the system permissions required to install it"

[WildcatRay]

The biggest problem I see appears to be an interaction between Tab Mix Plus and uBlock/AdBlock Plus. Have you talked to the TMP developer about that?

http://tabmixplus.org/forum/viewtopic.php?f=3&t=19618

[Grantius]

Anyone found todays build to be much lighter on RAM than the previous? I'm quite surprised, nothing else has changed E.G. addons

[itisomegakai]

can anyone notify the developer of ABP

Hello! That's my sheet over there, ABP developer has already been notified:
https://bugzilla.mozilla.org/show_bug.c ... 192585#c18
https://bugzilla.mozilla.org/show_bug.cgi?id=1280368#c1
Also, shortly afterwards this bug was created as an attempt to improve ABP&e10s performance: https://bugzilla.mozilla.org/show_bug.cgi?id=1280368

Raymond Hill, the creator of uBlock Origin, likely saw that sheet too but it looks like I couldn't convince him uBlock Origin causes performance problems. I guess I'll also post my sheet here so that everyone could see what I am talking about:

Thanks for the info,
but i can see he issues with e10s = off too.

ABP seems to be causing it & TMP usage makes it more often,
It cant only be that API problem if it happens with e10s=off if i understand correctly.

Also it seems recent ABP updates (last two or three seems to bring this problem)
cause high/cpu/memory usage

[itisomegakai]

Anyone found todays build to be much lighter on RAM than the previous? I'm quite surprised, nothing else has changed E.G. addons

Not much but memory usage increases slowly than yesterdays build but still not perfect like 120616 build,
can you or anyone say which bug caused this problem?
Or what bug related was fixed today that brought some improvement?


EDIT 1:

can some one verify that once the usage in memory increases it take quite long to come down?
Just open a few tabs so the memory goes around 1500 mb, then close all tabs and leave it for 5mins/or browse
memory usage still remains quite high

EDIT 2: Disabling network.http.spdy.enabled=false solves most of the page keeps loading problem for anyone else?

Edit 3: ABP is causing huge spikes in CPU/Memory usage in today's build even higher than previous build.



What the hell is going on with nightly for the last 2weeks, performance is taking a serious quality drop.

Edit 4: ABP 2.7.3.4175-beta is the one which works perfectly & builds just after that causing issues so switch to that if you want & please notify the developer about it.

[Wladimir Palant]

Yes, the latest Adblock Plus development build (Adblock Plus 2.7.3.4184-beta) indeed causes a memory leak. This is a regression from https://issues.adblockplus.org/ticket/4139 - I reopened that issue, the memory leak should be resolved shortly (I'm waiting for review there).

[itisomegakai]

Yes, the latest Adblock Plus development build (Adblock Plus 2.7.3.4184-beta) indeed causes a memory leak. This is a regression from https://issues.adblockplus.org/ticket/4139 - I reopened that issue, the memory leak should be resolved shortly (I'm waiting for review there).

are you sure?
ABP 2.7.3.4175-beta is the one which works perfectly & builds just after that causing issues for me