Bug 1793453 - Prevent automatic suppression of popups on 'blank' pages when URL bar is focused r=dao,adw, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D158850

Bug 1802706 - Convert site data settings confirmation dialog to modern flex layout. r=mconley,settings-reviewers, a=dsmith TIL we had preferences dialogs that weren't subdialogs! `contain: size` obviously doesn't work on those, and since they're not subdialogs they don't have the "make them scroll if too tall" code, causing this. Converting this one is trivial tho, so do that. Differential Revision: https://phabricator.services.mozilla.com/D163250

Bug 1802701 - Fix XUL vs. flex layout difference in site data settings. r=dao, a=dsmith This restores the pre-regression behavior for this dialog. Depends on D163162 Differential Revision: https://phabricator.services.mozilla.com/D163163

Bug 1802701 - Fix SubDialog with explicit height < min height > max height. r=dao, a=dsmith If we have frameHeight = 500px, minHeight = 800px and maxHeight = 600px, we still need to make the dialog scrollable, otherwise the frame will be too tall. This fixes the root issue, see following patch to fix the issue specific to this page. Differential Revision: https://phabricator.services.mozilla.com/D163162

Bug 1802139 - Simple patch for beta. r=dao,itiel, a=dsmith This can be tested in Win10+ too, as long as you add browser.display.windows.non_native_menus=0. Differential Revision: https://phabricator.services.mozilla.com/D162856

No bug - Tagging 6973e4e63b5c972a1dd4ab1e04358cc91a53e9d7 with FIREFOX_108_0b8_RELEASE a=release CLOSED TREE DONTBUILD

Automatic version bump CLOSED TREE NO BUG a=release DONTBUILD

No bug - Tagging 6973e4e63b5c972a1dd4ab1e04358cc91a53e9d7 with DEVEDITION_108_0b8_RELEASE a=release CLOSED TREE DONTBUILD

No bug - Tagging 6973e4e63b5c972a1dd4ab1e04358cc91a53e9d7 with DEVEDITION_108_0b8_BUILD1 a=release CLOSED TREE DONTBUILD

No bug - Tagging 6973e4e63b5c972a1dd4ab1e04358cc91a53e9d7 with FIREFOX_108_0b8_BUILD1 a=release CLOSED TREE DONTBUILD

no bug - Bumping Firefox l10n changesets r=release a=l10n-bump de -> ad09c843ee6d1ef925aa436aed1551002dc68e11 es-MX -> e922aa4af6047de024687c20f87e9aa9516ba41a gn -> 1423cf4ad277ec106ab9e74bc1ed12b5bbc616b9 it -> 5b4306597b5a2cac0c28e318c6d231779a6fcd2a ja -> ba87f571838dbd8530a3f586b0df6dcbaeaa185d ja-JP-mac -> 92b2d6580b5c80794cf12af8b6f35bf3138d1146 nb-NO -> 6c0c9d7f15d13005cd847e8cddc4f94bba0a9961

Backed out changeset 917a129b9b92 (bug 1801244) back out for potential risk in beta, a=backout

Bug 1802176 - Fix intermittent on browser_html_sitepermission_addons. r=rpl, a=testonly Set `midi.testing` and `dom.webmidi.enabled` prefs in browser.ini for the test. Differential Revision: https://phabricator.services.mozilla.com/D162991

Bug 1799195 - Disable feature callout resize test. r=mviar, a=testonly We can re-enable this test as part of bug 1795973 when the patch for that is rebased and fully passing. Differential Revision: https://phabricator.services.mozilla.com/D162446

Bug 1801021 - Use BigBuffer for DispatchCommands. r=gfx-reviewers,lsalzman, a=dsmith Using ipc::Shmem causes unbounded shmem use growth until e.g. a Worker yields to the event loop. If a Worker never yields, Shmems sent to WebGLParent are never released. Specifically the manager (PCanvasManager) for WebGLParent calls DestroySharedMemory, which sends/enqueues for WebGLChild's manager a matching call to ShmemDestroyed. However, while WebGLChild refuses to spin its event loop (such as a no-return WASM Worker), the ShmemDestroyed events will just pile up. Closing e.g. the tab frees the shmems, but they accumulate unbounded until the Worker yields to the event loop. This is true for other users of ipc::Shmem (or RaiiShmem) as well, but entrypoints other than DispatchCommands are rarer and can be handled later similarly. Differential Revision: https://phabricator.services.mozilla.com/D162946

Bug 1781167 - Allow stacking calls to Add/RemoveVsyncDispatcher so that we survive the sequence Add,Add,Remove. r=jrmuizel, a=dsmith This fixes a bug which caused Firefox windows to become frozen after some time. Full credit goes to Susan and RandyS for bisecting the regressor of this bug, and to Jeff DeFouw for debugging the issue and finding the cause. The bug here is a "state race" between the VsyncDispatcher state and the VsyncSource state. Both are protected by locks, and the code that runs in those locks respectively can see a different orders of invocations. VsyncDispatcher::UpdateVsyncStatus does this thing where it updates its state inside a lock, gathers some information, and then calls methods on VsyncSource *outside* the lock. Since it calls those methods outside the lock, these calls can end up being executed in a different order than the state changes were observed inside the lock. Here's the bad scenario in detail, with the same VsyncDispatcher being used from two different threads, turning a Remove,Add into an Add,Remove: ``` Thread A Thread B VsyncDispatcher::UpdateVsync | |----> Enter VsyncDispatcher lock | | VsyncDispatcher::UpdateVsync | | state->mIsObservingVsync = false | | | (We want to stop listening) | | | | |<---- Exit VsyncDispatcher lock | | |----> Enter VsyncDispatcher lock | | | | | | state->mIsObservingVsync = true | | | (We want to start listening) | | | | |<---- Exit VsyncDispatcher lock | | | |----> Enter VsyncSource::AddVsyncDispatcher | | | | | |----> Enter VsyncSource lock | | | | | | | | state->mDispatchers.Contains(aVsyncDispatcher) |----> VsyncSource::RemoveVsyncDispatcher | | | VsyncDispatcher already present in list, not doing anything | | | | | | | | |<---- Exit VsyncSource lock | | | | | | |<---- Exit VsyncSource::AddVsyncDispatcher | |----> Enter VsyncSource lock | | | | | | Removing aVsyncDispatcher from state->mDispatchers | | | | |<---- Exit VsyncSource lock | | |<---- Exit VsyncSource::AddVsyncDispatcher ``` Now the VsyncDispatcher thinks it is still observing vsync, but it is no longer registered with the VsyncSource. This patch makes it so that two calls to AddVsyncDispatcher followed by one call to RemoveVsyncDispatcher result in the VsyncDispatcher still being registered. AddVsyncDispatcher is no longer idempotent. Differential Revision: https://phabricator.services.mozilla.com/D162760

Bug 1801102 - Handle corrupted JAR files r=jesup, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D162392

Bug 1801360 - Use selectedBrowser URI for gScreenshots. r=cmkm, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D162747

Bug 1801690 - Fix searching for 2 wide chars in a latin1 string r=iain, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D162749

Bug 1801956 - Move titlebar button handling out of nsBoxFrame. r=tnikkel, a=dsmith This fixes the snap layouts feature on Windows 11. Instead of using a content attribute (which is somewhat expensive to look up) use the default appearance and do this where we deal with other appearance hacks (before building themed backgrounds). Consolidate this inside a DealWithWindowsAppearanceHacks function along with the glass stuff. Differential Revision: https://phabricator.services.mozilla.com/D162757

Bug 1799156 - Test for writing unknown clipboard type. r?nika, a=dsmith Based on https://phabricator.services.mozilla.com/D160944 Differential Revision: https://phabricator.services.mozilla.com/D161900

Bug 1799156 - Filter unknown DataTransferItem flavors. r=nika, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D161487

no bug - Bumping Firefox l10n changesets r=release a=l10n-bump de -> b31e37a2b6a3463594ce009f65e269eadb943187 es-ES -> 86fefa9fee8e19db62f74148df8a9d6bcd679495 fi -> d587416197fd811365e2eeba5f50d7419e62f769 fr -> 676efd86df72b537f05316d787ad1f8d27c3c563 ia -> 070ed989ca6fffa3cc73c1b54c6435e09c83ddd5 oc -> 86b5c7672fe8b13c6148353e3679045210f6dc26 th -> d1046e4e0b23243c7b17a035f7b33e0685fdf4d6 tr -> ab1619c4375401f40dc38ed45483fe35da732ee9

Bug 1800456 - Add preconditions to css-view-transitions tests, r=emilio a=test-only This ensures that the tests don't just timeout when the feature isn't implemented. Differential Revision: https://phabricator.services.mozilla.com/D162554

Bug 1800456 - Add failIfNot function to reftest-wait.js, r=emilio a=test-only This allows failing early for reftests that would otherwise timeout waiting for a screenshot. It assumes that replacing the body with an error message and taking a screenshot will never match the reference. Differential Revision: https://phabricator.services.mozilla.com/D162553

Bug 1801244 - Make Merino opt out on Beta. r=daisuke,nanj,RyanVM,chutten, a=dsmith This makes Merino opt out on Beta. There are two different parts: 1. Decide what "Beta" means. After discussing this with RyanVM, I'm using the same definition [used by telemetry](https://searchfox.org/mozilla-central/rev/2f47e3dacf0d773e9c7f363cecf10cfbea490679/toolkit/components/telemetry/app/TelemetryControllerBase.jsm#103,112-114): (a) if `MOZ_UPDATE_CHANNEL` is "beta" or (b) if it's "release" and `app.update.channel` is "beta". The latter case is true for RC builds delivered to the Beta channel. I moved the relevant telemetry logic to `AppConstants` so I can easily use it too. Bug 1435753 has context on this logic w/r/t telemetry. 2. Change the Merino default so it's enabled on Beta in the offline scenario. I kept the online scenario's disabled default since the only point of delivering the online scenario to the user now is to prompt them to opt in to Merino. This change is in UrlbarPrefs.sys.mjs. The rest of the patch updates tests. The effect of this patch will be to use a new default value of true for the `quicksuggest.dataCollection.enabled` pref on Beta in the offline scenario. If the user didn't touch the pref at all, then its previous default value was false, and the new default will now be true. If the user enabled the pref, it will remain enabled. If the user enabled and then disabled it, it will remain disabled. That's because the pref is sticky, so once it has a value on the user branch, it will keep that value. This patch only changes the value of the pref on the default branch, so it won't override the user's previous choice since that is set on the user branch. Differential Revision: https://phabricator.services.mozilla.com/D162353

Bug 1798494. r=jfkthame, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D162616

Bug 1801459 - Use widget sizemode to decide whether persistent should be saved; r=emilio, a=dsmith Fullscreen state on chrome window might not accurate as it might not be updated yet, e.g. user clicks the window controls button to switch fullscreen mode on macOS. Differential Revision: https://phabricator.services.mozilla.com/D163028

No Bug, mozilla-beta repo-update HSTS HPKP remote-settings - a=repo-update, r=dsmith Differential Revision: https://phabricator.services.mozilla.com/D163223

Bug 1742000 - Don't call hunspell rlbox::destroy_sandbox when create_sandbox failed r=bholley, a=dsmith Differential Revision: https://phabricator.services.mozilla.com/D162359