This commit adds a `dispatch_with_options` method to Rust task
runnables, which takes a builder-style `DispatchOptions`
argument controlling how the runnable should be dispatched.
Currently, the only supported option is `may_block`, which
corresponds to `nsIEventTarget::DISPATCH_EVENT_MAY_BLOCK`.
Differential Revision: https://phabricator.services.mozilla.com/D68688
--HG--
extra : moz-landing-system : lando
Turns out that align-self is enough for sizing to end up like we want (not
getting squashed by the flex container's padding). This prevents double-applying
the padding in cases where the top and bottom paddings don't match.
We also fix another change from my previous patch, which is that we wouldn't
clip stuff if the text editor is much bigger than the flex container. Most
trivial example of this is:
data:text/html,<input type=number style="font-size: 300px; height: 40px;">
I renamed the test from overflow-clip-box-* to padding-*, as we're not
technically exercising overflow-clip-box anymore, and added a test for the
clipping too.
Differential Revision: https://phabricator.services.mozilla.com/D68535
--HG--
rename : layout/reftests/forms/input/number/overflow-clip-box.html => layout/reftests/forms/input/number/padding-notref.html
rename : layout/reftests/forms/input/number/overflow-clip-box-ref.html => layout/reftests/forms/input/number/padding-ref.html
rename : layout/reftests/forms/input/number/overflow-clip-box-notref.html => layout/reftests/forms/input/number/padding.html
extra : moz-landing-system : lando
We've had some bugs where the sandboxed child process and/or the chroot
helper process deadlocks during launch, often reported by end users,
and it's confusing to have the mysterious hanging task inherit the name
of the launching thread; this patch fixes that by giving them more
informative names.
`prctl(PR_SET_NAME, ...)` is used directly, instead of via one of our
wrappers for it, to avoid the possibility of async signal unsafe
operations.
This doesn't name the pre-exec child process in the cases where regular
`fork()` is used, but as far as I know we haven't had any bugs (yet?)
where that would matter.
Differential Revision: https://phabricator.services.mozilla.com/D68134
--HG--
extra : moz-landing-system : lando
ContentBEnd() is equivalent to mBEndEdge except when ContentBSize() is
unconstrained because ContentBEnd() can overflow. However, according to
ContentBEnd()'s documentation, the user shouldn't use ContentBEnd() when
ContentBSize() is constrained, so I add an assertion in ContentBEnd() as
a reminder.
Differential Revision: https://phabricator.services.mozilla.com/D68624
--HG--
extra : moz-landing-system : lando
The word "must" is used to mark rules that valid bytecode must follow or risk
MOZ_CRASH, assertion failures, or undefined behavior. Documenting all of them
is an aspirational goal; this revision notes a bunch that were accidentally
skipped before, and adds the word "must" in some places.
The rule that `hops` and `slot` must be reasonable is documented at the top of
the file, under "Other operands", and does not need to be repeated in the
documentation for any instruction.
Differential Revision: https://phabricator.services.mozilla.com/D62519
--HG--
extra : moz-landing-system : lando
This commit changes the Push client code to check for existing push
subscriptions when connecting to the server, and omits the UAID from
the `hello` handshake if there are none. This, in turn, causes the
server to issue a new UAID, which we keep until either the next
reconnect, or when the user subscribes to push. It's a way to rotate
the UAID and reduce the privacy risk of tying a persistent identifier
to the connection.
Differential Revision: https://phabricator.services.mozilla.com/D66485
--HG--
extra : moz-landing-system : lando
Correcting linter errors
Added new logic for toggling control frames by adding a new control frame check variable and toggle function
Preserve controlFrames option when clearing frames. Changed Control Frames to Control in locale
Correct linter errors
Added a separator between received and control. Changed the dropdown title to reflect control frame state
Changed controlFramesEnabled to showControlFrames
Added a comment about showControlFrames and toggleControlFrames
Differential Revision: https://phabricator.services.mozilla.com/D67849
--HG--
extra : moz-landing-system : lando
I looked at the code generated for IsHugeMemoryEnabled(). It unconditionally locks, using multiple levels of un-inlined function calls in my optimized build. With the previous patch, that all gets called (and the lock taken) during every finalization, which is probably fine but seems worrisome.
Given that c++11 gives us easy access to a correct double-checked locking implementation, I suggest dropping it in here so that subsequent accesses need to do no locking, but only a lightweight memory fence.
Differential Revision: https://phabricator.services.mozilla.com/D68719
--HG--
extra : moz-landing-system : lando
Botond Ballo
Tim Nguyen
emcminn
Kris Maglione
Kartikaya Gupta
Lina Cambridge
Emilio Cobos Álvarez
Jed Davis
Daniel Varga
Ryan VanderMeulen
aarushivij
Jeff Gilbert
Ting-Yu Lin
Jason Orendorff
David Walsh
Henri Sivonen
Dana Keeler
Ian Neal
Michael Ratcliffe
Lars T Hansen
Dorel Luca
Bogdan Tara
Karl Tomlinson
Ken VanDine
Kishlaya Jaiswal
Razvan Maries
Steve Fink
sumagnadas