The first official 20180512 builds are out

[Josa]

Previous Nightly Builds thread 20180511

The first official 20180512 NIGHTLY builds are out
.The Official Firefox 61 Beta 5 builds are not yet out

Previous NIGHTLY: 20180511 (Fx 62.0a1)
Previous BETA: 20180511 (Fx 61.0b4)

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


Fixed:

1. NIGHTLY +40
2. #1460655 [Core:CSS Parsing and Computation]-Support 'x' as a resolution unit. [Uns][]
3. #1459655 [Core:DOM]-Crash in mozilla::ipc::PrincipalInfoToPrincipal [All][]
4. #1460807 [Core:DOM]-Drop nsIDOMWindowUtils.requestCompositorProperty [Uns][]
5. #1001469 [Core:DOM]-Re-enable cycle collector optimization for nodes in the shadow tree [All][]
6. #1438193 [Core:Disability Access APIs]-State and NativeState should be const [Uns][]
7. #1451853 [Core:Gecko Profiler]-Include the Firefox build id in the profiler's meta information [All][]
8. #1459935 [Core:Graphics: WebRender]-Update webrender to 4811a6e7c06f9dd1b4056e5f5e66983842983ba0 [Uns][[gfx-noted]]
9. #1457466 [Core:Graphics: WebRender]-Use async scene building better (i.e. make scrolling smooth on pages with slow scene builds) [Uns][[gfx-noted]]
10. #1452390 [Core:Graphics: WebRender]-We don't implement paint skipping for WebRender (less paints during scrolling) [Uns][]
11. #1460988 [Core:Graphics: WebRender]-Yet another case of holding the layer trees lock while blocking on WR threads [Uns][[gfx-noted]]
12. #1461023 [Core:Graphics]-Disable the assertion that checks that we have surely the previous opacity value when we skip computing opacity animation value on the compositor [Uns][]
13. #1460127 [Core:IPC]-Failed to connect GPU process [Lin][]
14. #1348792 [Core:JavaScript Engine: JIT]-Try to share more CacheIR compiler code [All][[qf-]]
15. #1459552 [Core:JavaScript Engine: JIT]-Wasm: memory.fill and memory.copy: fix opcode assignments per recent committee decision [Uns][]
16. #1459220 [Core:JavaScript Engine]-Add LazyScript::isEnclosingScriptLazy and replace existing LazyScript::sourceObject callsite which checks if the enclosing script is lazy [Uns][]
17. #1415202 [Core:JavaScript Engine]-Always use the equivalent year when determining the time zone offset and name in PRMJ_FormatTime [Mac][]
18. #1459127 [Core:JavaScript Engine]-Store ScriptSourceObject reference into LazyScript inside LazyScript [Uns][]
19. #1460636 [Core:JavaScript: GC]-Long CC pauses from tracing group_property in js::ObjectGroup::traceChildren(JSTracer* trc) [Uns][[qf]]
20. #1427660 [Core:Layout: Text]-Hangul jamo in the range of U+A960–U+A97C do not form a syllable [Uns][]
21. #1460368 [Core:Layout]-nsContentUtils::GetClosestNonNativeAnonymousAncestor seems to be dead code [Uns][]
22. #1460313 [Core:Networking: HTTP]-Is nsHttpTransaction::mForTakeResponseHead used? [Uns][[necko-triaged]]
23. #1460559 [Core:WebRTC: Audio/Video]-Enable tracing of new real time logger in dom/media/webrtc/* files [Uns][]
24. #1460810 [Core:Widget: Gtk]-[Wayland] - Segfault while pasting text [Lin][]
25. #1450314 [Core:XPCOM]-Unused memory-pressure handling bits [Uns][]
26. #1460119 [Firefox:Activity Streams: Newtab]-gSnippetsMap.getTotalBookmarksCount() never resolves [Uns][]
27. #1455462 [Firefox:Developer Tools]-When DevTools are zoomed in meatball menu and overflow menu are in the wrong place [Uns][]
28. #1458205 [Firefox:Developer Tools]-telemetry.js: Remove actionOccurred [All][]
29. #1458203 [Firefox:Developer Tools]-telemetry.js: Rename all log* methods [All][]
30. #1458204 [Firefox:Developer Tools]-telemetry.js: Rename startTimer/stopTimer [All][]
31. #1457620 [Firefox:Normandy Client]-Expose study opt-out string to all locales in Firefox Nightly for desktop only [Uns][]
32. #1435875 [Firefox:Normandy Client]-Localize about:studies [Uns][]
33. #1460654 [Firefox:Preferences]-The "OK", "Cancel" and "Help" buttons from the "Languages" pop-up are cut off [Win][]
34. #1453308 [Mozilla Localizations:uk / Ukrainian]-Change e-commerce search engine in Firefox for uk [Uns][]
35. #1460629 [Toolkit:Add-ons Manager]-All dictionaries are seen as incompatible in Firefox 62 even if the install button is active [Win][]
36. #1460717 [Toolkit:Add-ons Manager]-Cleanup XPIDatabaseReconcile.processStartupChanges [Uns][]
37. #1385057 [Toolkit:Add-ons Manager]-Remove support for unpacked side-loaded extensions [Uns][]
38. #1460748 [Toolkit:Add-ons Manager]-WebExtension dictionaries load .dic file as affix file [Mac][]
39. #1458910 [Toolkit:Places]-Merge nsPIPlacesDatabase into nsINavHistoryService [Uns][[fxsearch]]
40. #1451795 [Toolkit:WebExtensions: Developer Tools]-Store the devtools pages definitions into the instance of the ExtensionAPI class instead of a global Map. [Uns][]
41. #1419132 [Toolkit:WebExtensions: Frontend]-deprecate windowTypes for windows.get, windows.getCurrent and windows.getLastFocused [All][[good first bug][windows]]

1. Partial Landings/Diagnostic Patches:
2. #1446900 [Core:IPC]-Crash in mozilla::ipc::IToplevelProtocol::OtherPid [Win][]
3. #1460341 [Core:JavaScript: GC]-Remove GCPolicy<T>::initial [Uns][]

1. Beta 4 -> Beta 5 Changelog +0
   None

Nightly 62 fixes since 20180507 (Gecko 61) ~255
Beta 61 fixes since 20180312 (Gecko 60) ~2560

Nightly Blog
Release Calendar
Release Tracking
MozRegression Guide

[Orangutan100]

Anybody seeing an issue where background tabs seem to retain their state until you switch to them?

STR:
1. Go to https://blog.mozilla.org/addons/. Let's call this tab1
2. Scroll down to the bottom, click the `Older posts` link
3. Immediately after, switch to another tab, tab2. It might be better reproducible if you manage to switch to tab2 while tab1 is still connecting (grey tab throbber)
4. Once the blue throbber animation on tab1 finishes, switch back to it.

ER: content area should show https://blog.mozilla.org/addons/page/2/.

AR: I get a glimpse of https://blog.mozilla.org/addons/, lasting a couple of seconds on some occasion, before it is replaced by https://blog.mozilla.org/addons/page/2/. The urlbar displays https://blog.mozilla.org/addons/page/2/ during the glimpse of stale content though.

I don't have Webrender enabled. I'm on 62.0a1 (2018-05-11) (64-bit), Windows 10. Graphics section from about:support: https://pastebin.com/k168Gwi5

[mayankleoboy1]

Anybody seeing an issue where background tabs seem to retain their state until you switch to them?

STR:
1. Go to https://blog.mozilla.org/addons/. Let's call this tab1
2. Scroll down to the bottom, click the `Older posts` link
3. Immediately after, switch to another tab, tab2. It might be better reproducible if you manage to switch to tab2 while tab1 is still connecting (grey tab throbber)
4. Once the blue throbber animation on tab1 finishes, switch back to it.

ER: content area should show https://blog.mozilla.org/addons/page/2/.

AR: I get a glimpse of https://blog.mozilla.org/addons/, lasting a couple of seconds on some occasion, before it is replaced by https://blog.mozilla.org/addons/page/2/. The urlbar displays https://blog.mozilla.org/addons/page/2/ during the glimpse of stale content though.

I don't have Webrender enabled. I'm on 62.0a1 (2018-05-11) (64-bit), Windows 10. Graphics section from about:support: https://pastebin.com/k168Gwi5

https://bugzilla.mozilla.org/show_bug.cgi?id=1460905 is probably it.

[ashleylai87]

Recently, I tried to play around with DOH (dns over https).
So, I manually change some firefox pref:

network.trr.mode 3 (As I want it to be 100% DOH only)
network.trr.uri https://cloudflare-dns.com/dns-query (Cloudflare doh resolver)

But after I set these two preferences, firefox is unable to resolve all the websites.
Wireshark log doesn't show anything at all.

Then someone informed me that one must set network.trr.bootstrapAddress to match the DOH server. So, I insert 1.1.1.1 (cloudflare) at network.trr.bootstrapAddress. After that, all websites resolve properly.
Wireshark also indicates DOH is being used.

So, my question, do I really have to setup network.trr.bootstrapAddress even though after i setup network.trr.uri https://cloudflare-dns.com/dns-query ?

[johnp_]

Normally, the system resolver should be used if `network.trr.bootstrapAddress` is empty: https://searchfox.org/mozilla-central/r ... ll.js#5305
But from a quick glance at the code it doesn't seem to be implemented: https://searchfox.org/mozilla-central/r ... ce.cpp#352
And I also can't find an open bug regarding this...

[ashleylai87]

Normally, the system resolver should be used if `network.trr.bootstrapAddress` is empty: https://searchfox.org/mozilla-central/r ... ll.js#5305
But from a quick glance at the code it doesn't seem to be implemented: https://searchfox.org/mozilla-central/r ... ce.cpp#352
And I also can't find an open bug regarding this...

Then there is a weird bug too.
Firefox cannot make use of IPv6 if DOH is used.
I changed the network.trr.bootstrapAddress to 2606:4700:4700::1111, dns request is successful over cloudflare ipv6 address, however firefox will only connect to website via ipv4 instead of ipv6.

[TheVisitor]

I'm curious, what is the benefit of using DOH, looks like a lot of trouble to me....

[johnp_]

Maybe try enabling `network.trr.early-AAAA`? This should "force" IPv6 if it's returned by the resolver before the A-Record/IPv4. Otoh I don't know enough about the IPv4/IPv6 choice semantics in Firefox.

[ashleylai87]

I'm curious, what is the benefit of using DOH, looks like a lot of trouble to me....

To be honest, I am not sure myself.

Seem to me, firefox DOH isn't working well/stable yet.

[ashleylai87]

Maybe try enabling `network.trr.early-AAAA`? This should "force" IPv6 if it's returned by the resolver before the A-Record/IPv4. Otoh I don't know enough about the IPv4/IPv6 choice semantics in Firefox.

Flipping this seems to work.

[Ver Greeneyes]

The weird tab ordering bug (new tabs opening in between pinned tabs) should be fixed in today's Nightly.

[Josa]

This thread is closed, please use the next one.