At the moment, everything we're putting in a wasm sandbox is not
performance-critical, so we don't need PGO. It's also not clear that
PGO would actually work properly with code that's been run through
wasm. Let's leave figuring that out until, at the very least, the wasm
toolchain is a little more mature.
Differential Revision: https://phabricator.services.mozilla.com/D58513
--HG--
extra : moz-landing-system : lando
There was a race between the add-on adding its pref listener and us flipping the enabled pref. By waiting for
the balrog migration, we give the add-on time to start listening for the pref before setting it.
Differential Revision: https://phabricator.services.mozilla.com/D58422
--HG--
extra : moz-landing-system : lando
Avoid intermittent task timeouts by increasing the number of test chunks on the
affected platform.
Differential Revision: https://phabricator.services.mozilla.com/D58391
--HG--
extra : moz-landing-system : lando
Errors that arise from use of the protocol, for example by the client
passing in the wrong argument or an operationg being unsupported, are
returned to the client. Since these are not inherent implementation
problems there is no point to double-logging them in the server.
This is a left-over from when the remote agent was in a prototype
stage and we couldn't always rely on the Puppeteer client to
interact with Firefox correctly.
Differential Revision: https://phabricator.services.mozilla.com/D58508
--HG--
extra : moz-landing-system : lando
When remote.log.level is Log.Level.Info or above, verbose logging
is enabled and we pretty-print JSON payloads in requests to the
HTTPD in JSONHandler.
This patch matches the behaviour with JSON payloads being transmitted
across WebSocket connections.
Differential Revision: https://phabricator.services.mozilla.com/D58506
--HG--
extra : moz-landing-system : lando
Moving the JSON payload sanitisation function to Protocol.jsm
means we can share it across modules.
The patch also adds new tests.
Differential Revision: https://phabricator.services.mozilla.com/D58505
--HG--
extra : moz-landing-system : lando
Over in bug 1604761 I realized that GTK has a high-contrast accessibility
setting, that GTK does not honor the same way as windows does.
There's no good reason for that IMO, and high contrast mode on windows is
getting some love.
This hooks the relevant parts of the GTK widget code so that we also avoid
websites overriding high contrast colors (via PreferenceStyleSheet and co).
Differential Revision: https://phabricator.services.mozilla.com/D58278
--HG--
extra : moz-landing-system : lando
Calculate and set correct opaque regions for all wl_surfaces, which are:
- mShell - toplevel window with CSD decorations, opaque mask needs to be shifted by titlebar/CSD size and subtracted transparent corners.
- mContainer - child window with subtracted transparent corners.
Differential Revision: https://phabricator.services.mozilla.com/D57797
--HG--
extra : moz-landing-system : lando
The configuration may specify multiple locales for a search engine. The search engine selector already splits these out into different engine objects, but keeps the locale field as an array. This patch changes the 'locales' array into a 'locale' string and applies appropriate simplifications to the search service.
Differential Revision: https://phabricator.services.mozilla.com/D58510
--HG--
extra : moz-landing-system : lando
I also took the opportunity to fix the SSL proxy label - we shouldn't label
things "SSL" anymore in 2019, unless they really are. All the comments in
bug 969282, bug 1577862 and bug 1601871 use 'HTTPS'.
Differential Revision: https://phabricator.services.mozilla.com/D58120
--HG--
extra : moz-landing-system : lando
This is the right thing to do, and prevents a race where we try to log
after the stdout capture queue is closed
Differential Revision: https://phabricator.services.mozilla.com/D56567
--HG--
extra : moz-landing-system : lando
The problem is the compartment checking tracer currently gets to see the target of a WeakRef, which may be in another compartment. I updated the trace method to check the traceWeakEdges() flag and set this to false for the GCMarking tracer. I was surprised that this wasn't the case before but I did a try run and everything looks fine.
Differential Revision: https://phabricator.services.mozilla.com/D57962
--HG--
extra : moz-landing-system : lando
JIT compilation and bytecode analysis now no longer depend on any source notes.
The deleted code to print the offset-to-catch in Disassemble1 goes back to
bug 680228, before the newer code coverage implementation. It's unlikely that
anyone is still depending on this and it's also confusing to print an offset
in the bytecode disassembly that's not actually an operand.
Differential Revision: https://phabricator.services.mozilla.com/D58430
--HG--
extra : moz-landing-system : lando
This fixes the following case to just execute the script instead of using
the interactive shell:
```
dist/bin/js < file.js
```
The -i flag can be used to force interactive mode in this case.
Differential Revision: https://phabricator.services.mozilla.com/D56334
--HG--
extra : moz-landing-system : lando
This was removed in Bug 1549396 for causing stability issues. Since then, a
number of GC and TI bugs have been fixed and this should now be stable.
Differential Revision: https://phabricator.services.mozilla.com/D58451
--HG--
extra : moz-landing-system : lando
Changes:
Remove `crashtest`, `mochitest-gpu` from running with ubuntu1604.
Add the above tests to run with ubuntu1804 image under `linux1804-64` label.
Remove `awsy` from linux64-qr/opt and linux64-shippable-qr/opt platforms.
Add `awsy` to linux1804-qr/opt and linux1804-shippable-qr/opt platforms.
Add new test set `linux180r-qr-tests` to serve as placeholder as `linux-qr-tests` are slowly migrated to the new test set.
Replace the usage of `linux1804-tests` for quantumrenderer platforms with the `linux-qr-tests` set.
Differential Revision: https://phabricator.services.mozilla.com/D58464
--HG--
extra : moz-landing-system : lando
This patch adds the following functionality to the CSSStyleSheet WebIDL API
under the feature flag layout.css.constructable-stylesheets.enabled:
- constructor()
- replace()
- replaceSync()
constructor(), replace() and replaceSync() are currently stubs that lack full
functionality.
Differential Revision: https://phabricator.services.mozilla.com/D58326
--HG--
extra : source : 6e675d6399cee28f204b3501de3475ea3febd12c
Changes:
As part of python3 compatibility work, remove files that are no longer being used as far as I can tell (looking in searchfox).
Differential Revision: https://phabricator.services.mozilla.com/D58472
--HG--
extra : moz-landing-system : lando
This commit changes IsBidiLeaf to mark ruby boxes not a bidi leaf, since
conceptually they are not. I auditted all uses of it, and I believe it
makes more sense this way.
Differential Revision: https://phabricator.services.mozilla.com/D58290
--HG--
extra : moz-landing-system : lando
Brindusan Cristian
Henrik Skupin
Nathan Froyd
Nihanth Subramanya
Jonathan Kingston
Sylvestre Ledru
Geoff Brown
Ted Campbell
Andreas Tolfsen
Emilio Cobos Álvarez
Cosmin Sabou
Martin Stransky
Mark Banner
Gijs Kruitbosch
ffxbld
Oriol Brufau
James Graham
Jon Coppeard
Jan de Mooij
Alastor Wu
Bogdan Tara
Ciure Andrei
championshuttler
Edwin Takahashi
nordzilla
Xidorn Quan
Mihai Alexandru Michis