These two platforms are the easiest to get started with - as well as accounting for the great majority
of desktop Firefox users.
This patch is based on the OS vendors' lists of fonts shipped with the current version of each OS.
Differential Revision: https://phabricator.services.mozilla.com/D66125
--HG--
extra : moz-landing-system : lando
This replaces and extends the "hidden" flag we currently use on macOS to mark internal system fonts
like .LastResort and .Keyboard that should not be exposed; rather than just a boolean "hidden" flag
we'll have several levels of visibility, some of which the user may opt in to exposing (at the cost
of potentially becoming more fingerprintable).
The current patch assumes three levels besides always-hidden:
Base - fonts that are part of the base OS install and always available
LangPack - fonts that are provided by the OS subject to user's chosen language options
User - user-installed fonts that were not provided by the OS
(This categorization may be subject to revision as we learn more about real-world needs and
configurations.)
Differential Revision: https://phabricator.services.mozilla.com/D66124
--HG--
extra : moz-landing-system : lando
Replace MOZ_CRASH with mozalloc_handle_oom for DOMArena OOM
because it captures the requested size.
Differential Revision: https://phabricator.services.mozilla.com/D69255
--HG--
extra : moz-landing-system : lando
The test case is based on the existing heap-analysis/byteSize-of-string.js test file.
Differential Revision: https://phabricator.services.mozilla.com/D69158
--HG--
extra : moz-landing-system : lando
This prevents a race condition where the watchman hook can potentially
invalidate the cache in-between the call to 'isfile' and 'getmtime'.
Depends on D69189
Differential Revision: https://phabricator.services.mozilla.com/D69190
--HG--
extra : moz-landing-system : lando
* Updated the toolkit/components/passwordmgr/test/mochitest/test_autocomplete_new_password.html test to include a third form that has enough signal for the new-password heuristics to detect its password field as a new password field.
* Factored out most of the test_autofillAutocompletePassword_withGeneration task to be rechecked for this new form.
* Created some new helper functions and promoted them to the global scope in the test, so they can be used by other tasks as well.
Differential Revision: https://phabricator.services.mozilla.com/D68499
--HG--
extra : moz-landing-system : lando
Make the names of the data structures used inside HttpObserverManager
match the webRequest event names.
Differential Revision: https://phabricator.services.mozilla.com/D68736
--HG--
extra : moz-landing-system : lando
Currently, we build android emulator packages manually and upload to
tooltool.
This patch switches it to be pulled from the toolchain built artifacts.
This also allows android tests to run in the staging environment.
Differential Revision: https://phabricator.services.mozilla.com/D68617
--HG--
extra : moz-landing-system : lando
The stub implementation seems like it conflicts on platforms where we have an
implementation, is registered in the child process where we don't need it,
and the implementation itself uses getters where it should use fields.
The code that tries to find a sharepicker implementation handles one not
being registered just fine as far as I can see, so just remove the stub
implementation now that we have actual implementations for people to
look at.
Differential Revision: https://phabricator.services.mozilla.com/D69147
--HG--
extra : moz-landing-system : lando
This file has an entry in kWidgetContracts for kNS_SHAREPICKER_CID, but not in
kWidgetCIDs, so there's no implementation. I think this is causing the "Could
not map contract" warning message. We never use the share picker in the content
process, so just remove this.
Differential Revision: https://phabricator.services.mozilla.com/D69146
--HG--
extra : moz-landing-system : lando
The stack overflow handler is based very heavily on the old implementation. There may be a nicer way to write it.
After this patch, we have a complete RegExpMacroAssembler implementation.
Differential Revision: https://phabricator.services.mozilla.com/D69002
--HG--
extra : moz-landing-system : lando
V8 (and our old import) supports a global cache optimization where we find multiple global matches at once and then iterate over those when finding the next match, instead of jumping back into the generated code each time. We never actually hooked that optimization up to anything, though, and I'm not sure whether it would add enough performance to justify the complexity.
I've pulled all of that logic out, so the success handler here is significantly simpler than it was in the old implementation.
(I had already written most of the code by the time I realized what was going on, so I have an optional patch in my stack to re-add the support if we decide we want to implement the global cache.)
Differential Revision: https://phabricator.services.mozilla.com/D69001
--HG--
extra : moz-landing-system : lando
Miko Mynttinen
Matt Woodrow
Bianca Danforth
Kagami Sascha Rosylight
Kartikaya Gupta
Jonathan Kew
Jared Wein
Yannan Tuo ytuo
aarushivij
Bert Peers
Soumya Lahiri
Sean Feng
André Bargull
Erik Nordin
Michael Froman
James Willcox
Ciure Andrei
Bogdan Tara
Andrew Halberstadt
Glenn Watson
Dragana Damjanovic
Emilio Cobos Álvarez
Zibi Braniecki
Kris Maglione
Thom Chiovoloni
Tom Schuster
Andrew Swan
Gijs Kruitbosch
Wander Lairson Costa
Andrew McCreight
David Walsh
Iain Ireland