Instead of trying to figure out which pointer is nullptr, add a more
powerful assertion. Since the nullptr already causes a crash, this won't
cause any regression.
Differential Revision: https://phabricator.services.mozilla.com/D22085
--HG--
extra : moz-landing-system : lando
By allowing the creation of StrongWorkerRefs in the Canceling state we
ensure that IPC will not fail and lead to crashes.
Differential Revision: https://phabricator.services.mozilla.com/D21920
--HG--
extra : moz-landing-system : lando
If main thread is busy handling runnables in the normal priority queue, control-type of messages from
workers are possibly postponed to run after those. That can lead to bad performance, if the page
expects workers to be able to proceed simultanously with the main thread.
This patch makes the control messages to use medium high priority queue in order to try to
ensure they are handled in timely manner.
Pref dom.worker.use_medium_high_event_queue can be set to false to disable this new behavior.
Differential Revision: https://phabricator.services.mozilla.com/D22128
--HG--
extra : rebase_source : 447dec6dbcccaa0206a1815c21ccf713c523fc91
This will mean that in places like the tight loop in GetTypeIndex()
we would no longer require calling strlen() on the input type argument
once per loop iteration.
Depends on D20236
Differential Revision: https://phabricator.services.mozilla.com/D20237
--HG--
extra : moz-landing-system : lando
Replacing js and text occurences of asyncOpen2
Replacing open2 with open
Differential Revision: https://phabricator.services.mozilla.com/D16885
--HG--
rename : layout/style/test/test_asyncopen2.html => layout/style/test/test_asyncopen.html
extra : moz-landing-system : lando
For cases where the class has direct calls (that is, we cast `this` to the
subclass before making the call) no longer declare Recv/Answer methods on the
base class at all. This should ensure that slots for them are not generated in
vtables, and also allow the derived class to choose the method signature (e.g.
whether it wants to take something by reference or by value).
Differential Revision: https://phabricator.services.mozilla.com/D18132
--HG--
extra : moz-landing-system : lando
When calling a Recv/Alloc/Dealloc method on most types, cast `this` to the
derived class.
There is a heuristic to figure out what the correct derived type is. There is a
blacklist of types which we can't do direct calls on for the moment, as well as
an override for types that do work with direct calls but which don't match the
heuristic.
Differential Revision: https://phabricator.services.mozilla.com/D16492
--HG--
extra : moz-landing-system : lando
***
Bug 1514594: Part 3a - Change ChromeUtils.import to return an exports object; not pollute global. r=mccr8
This changes the behavior of ChromeUtils.import() to return an exports object,
rather than a module global, in all cases except when `null` is passed as a
second argument, and changes the default behavior not to pollute the global
scope with the module's exports. Thus, the following code written for the old
model:
ChromeUtils.import("resource://gre/modules/Services.jsm");
is approximately the same as the following, in the new model:
var {Services} = ChromeUtils.import("resource://gre/modules/Services.jsm");
Since the two behaviors are mutually incompatible, this patch will land with a
scripted rewrite to update all existing callers to use the new model rather
than the old.
***
Bug 1514594: Part 3b - Mass rewrite all JS code to use the new ChromeUtils.import API. rs=Gijs
This was done using the followng script:
https://bitbucket.org/kmaglione/m-c-rewrites/src/tip/processors/cu-import-exports.jsm
***
Bug 1514594: Part 3c - Update ESLint plugin for ChromeUtils.import API changes. r=Standard8
Differential Revision: https://phabricator.services.mozilla.com/D16747
***
Bug 1514594: Part 3d - Remove/fix hundreds of duplicate imports from sync tests. r=Gijs
Differential Revision: https://phabricator.services.mozilla.com/D16748
***
Bug 1514594: Part 3e - Remove no-op ChromeUtils.import() calls. r=Gijs
Differential Revision: https://phabricator.services.mozilla.com/D16749
***
Bug 1514594: Part 3f.1 - Cleanup various test corner cases after mass rewrite. r=Gijs
***
Bug 1514594: Part 3f.2 - Cleanup various non-test corner cases after mass rewrite. r=Gijs
Differential Revision: https://phabricator.services.mozilla.com/D16750
--HG--
extra : rebase_source : 359574ee3064c90f33bf36c2ebe3159a24cc8895
extra : histedit_source : b93c8f42808b1599f9122d7842d2c0b3e656a594%2C64a3a4e3359dc889e2ab2b49461bab9e27fc10a7
This does the following:
- It introduces a controlling ifdef ENABLE_WASM_REFTYPES that enables
exactly those features that are in the reftypes proposal, excluding
those in the gc proposal. Any remaining features (namely, ref.eq,
(ref T) types, struct types) are still under ENABLE_WASM_GC control.
ENABLE_WASM_GC requires ENABLE_WASM_REFTYPES and this is checked.
- It introduces a new TestingFunctions predicate, wasmReftypesEnabled,
that distinguishes reftype-proposal support from gc-proposal
support. We keep wasmGcEnabled to test for gc-proposal support.
- It segregates test cases so that gc-proposal relevant tests are in
their own files, and tests relevant to the reftypes-proposal are now
guarded by wasmReftypesEnabled.
- It renames the predicate HasGcSupport() as HasReftypesSupport(),
since that is what the predicate tests for.
- It has a drive-by fix for the DEBUG-only function wasm::Classify()
to properly put ref.null and ref.is_null under ifdef control.
Reftypes will soon be enabled unconditionally in Nightly (once we can
trace pointers from Ion frames) while gc-types will remain conditional
until Ion supports all the new instructions for struct types. Therefore:
- The command line switch and about:config option are still called
--wasm-gc and j.o.wasm_gc, respectively, which is fine since they will
fairly soon control only gc-proposal features.
- Internal names still use "Gc" rather than "Reftypes", eg,
HasGcTypes, wasmGc_, and so on. This is most appropriate since it
reduces the scope of the patch and these names will pertain mainly
to the gc feature in the future.
--HG--
extra : rebase_source : 51cf3bfe67da594e89195472e4ce1ccfa36c146d
These tests have always been skipped silently. Now that https tests are
enabled to actually run, we find that some (all?) actually fail on Android:
now explicitly skipped to allow for green runs of tier 1 suites.
Verify that the script URI is not null before using it.
If it's the case, we can't really continue because that url
identifies the worker in the metrics.
Differential Revision: https://phabricator.services.mozilla.com/D15187
--HG--
extra : moz-landing-system : lando
Summary: Really sorry for the size of the patch. It's mostly automatic
s/nsIDocument/Document/ but I had to fix up in a bunch of places manually to
add the right namespacing and such.
Overall it's not a very interesting patch I think.
nsDocument.cpp turns into Document.cpp, nsIDocument.h into Document.h and
nsIDocumentInlines.h into DocumentInlines.h.
I also changed a bunch of nsCOMPtr usage to RefPtr, but not all of it.
While fixing up some of the bits I also removed some unneeded OwnerDoc() null
checks and such, but I didn't do anything riskier than that.
This code races with the WeakWorkerRef shutdown code that sets both `mWorkerState` and `mWorkerPrivate` (though on different threads). This patch is based on the observation that except for failure cases, we can't get to `RuntimeServiceWorker::UnregisterWorker` without having already notified the `WorkerRef`s.
Differential Revision: https://phabricator.services.mozilla.com/D14176
--HG--
extra : moz-landing-system : lando
By taking an rvalue reference and using std::move more aggressively, we can
avoid at least two copies of these closures, which avoids a bit of atomic
refcounting.
Differential Revision: https://phabricator.services.mozilla.com/D14807
--HG--
extra : moz-landing-system : lando
This member was never changed after it was set, but it couldn't be const
because WorkerLoadInfo used a custom StealFrom() function instead of an
idiomatic move constructor and move assignment operator. This commit
replaces StealFrom with its idiomatic counterparts, adds a function that
determines a new worker's secure context status, and uses that function
to initialize the now const mIsSecureContext.
Making constant members const is part of a greater effort to ensure that
data is either modified from a single thread only, or access to it is
synchronized properly.
Differential Revision: https://phabricator.services.mozilla.com/D12030
--HG--
extra : moz-landing-system : lando
We only need to expose an intercept controller in SharedWorkers if we're on
the non-parent-intercept version of ServiceWorkers or if e10s is off.
nsDocShell already does this dance and we have to mirror it.
Differential Revision: https://phabricator.services.mozilla.com/D12490
--HG--
extra : moz-landing-system : lando
Add some memory usage information to the Performance counters and make everything asynchronous.
Differential Revision: https://phabricator.services.mozilla.com/D7984
--HG--
extra : moz-landing-system : lando
This member was never changed after it was set, but it couldn't be const
because WorkerLoadInfo used a custom StealFrom() function instead of an
idiomatic move constructor and move assignment operator. This commit
replaces StealFrom with its idiomatic counterparts, adds a function that
determines a new worker's secure context status, and uses that function
to initialize the now const mIsSecureContext.
Making constant members const is part of a greater effort to ensure that
data is either modified from a single thread only, or access to it is
synchronized properly.
Differential Revision: https://phabricator.services.mozilla.com/D12030
--HG--
extra : moz-landing-system : lando
In the next patch I get rid of the duplicated arguments.
Differential Revision: https://phabricator.services.mozilla.com/D11826
--HG--
extra : rebase_source : 2c21364406b9a757a78ae523e212a180f848ff9b
In order to fix the problem mentioned in comment 91 & co, we need to hold onto
the URI object that we resolve in the child process when we construct the
SharedWorker. Otherwise, we risk the Blob getting deallocated from under us.
This patch isn't sufficient to fix that problem, however, because the worker
code itself ends up going back through strings. I fix that in the next couple
of patches.
Differential Revision: https://phabricator.services.mozilla.com/D11825
--HG--
extra : rebase_source : c77854f00c0d7a102e73e0c81f59cc217f43fd69
This separates out the parent process from the list of child processes and
makes our handling of non-e10s more explicit.
Differential Revision: https://phabricator.services.mozilla.com/D11824
--HG--
extra : rebase_source : 3d4a5c5b427a5ad709468b4063793cb7bce27117
This implements the behavior that as long as there's one non-frozen or
non-suspended actor, we resume or thaw the manager.
Differential Revision: https://phabricator.services.mozilla.com/D11822
--HG--
extra : rebase_source : d5dd0e87581c94b383f1e155b03f63398b2a14a7
It's possible for RuntimeService to be created after 'xpcom-shutdown' has fired. In this case, it
will receive 'xpcom-shutdown-threads' and perform Cleanup() but not Shutdown(). This means that
mShuttingDown will not be set to 'true', but mIdleThreadTimer will be destroyed. This can cause
crashes if a NoteIdleThread callback runs after Cleanup(). This has been observed to happen in
xpcshell tests.
I think the easiest way to handle this is to manually call Shutdown() in Cleanup() when we
see that mShuttingDown == false. This means that Shutdown() might be called in GetOrCreateService()
if we fail to create the service, but it looks like the code can handle this.
Differential Revision: https://phabricator.services.mozilla.com/D10288
--HG--
extra : source : e28fa79bc2f94ca3b72456b47353f2e2dda8da1a
extra : amend_source : f09508b5c62fb1fe4b5997f822b5d4e5f7ef8076
It's possible for RuntimeService to be created after 'xpcom-shutdown' has fired. In this case, it
will receive 'xpcom-shutdown-threads' and perform Cleanup() but not Shutdown(). This means that
mShuttingDown will not be set to 'true', but mIdleThreadTimer will be destroyed. This can cause
crashes if a NoteIdleThread callback runs after Cleanup(). This has been observed to happen in
xpcshell tests.
I think the easiest way to handle this is to manually call Shutdown() in Cleanup() when we
see that mShuttingDown == false. This means that Shutdown() might be called in GetOrCreateService()
if we fail to create the service, but it looks like the code can handle this.
Differential Revision: https://phabricator.services.mozilla.com/D10288
--HG--
extra : rebase_source : 3c4a9cb76b81c4aef87b6373548e9da8ca64075e
extra : amend_source : d17d7a0e35e8bd9fcfbbd567e387d9af857bfd8a
In this patch, I went through any place in DOM fetch code, where there are
ReadableStreams and update the locked, disturbed, readable checks.
Because we expose streams more often, we need an extra care in the use of
ErrorResult objects. JS streams can now throw exceptions and we need to handle
them.
This patch also fixes a bug in FileStreamReader::CloseAndRelease() which could
be called in case mReader creation fails.
- Make ServiceWorkerGlobalScope.importScripts() throw a NetworkError when receiving a
bad (i.e. non-JavaScript) MIME type
- Correct registration-tests-mime-types.js to expect TypeError when registering
a service worker that calls importScripts() with a bad MIME type, per spec
- Add WPT import-scripts-mime-types.https.html to test importScripts success/failure,
depending on MIME type
Depends on D6416
Differential Revision: https://phabricator.services.mozilla.com/D9886
--HG--
extra : moz-landing-system : lando
Remove WorkerPrivate::mQueuedRunnables and its associated functions. The
approach they implement can never be correct, as the parent window gets
'resumed' whenever the debugger resumes execution after a breakpoint. The
interrupted JavaScript invocation has not yet completed, so it is not yet time
to run mQueuedRunnables. Simply re-enqueing them at that point can cause
messages from the worker to arrive out of order.
Instead, we create a separate ThrottledEventQueue,
WorkerPrivate::mMainThreadDebuggeeEventTarget especially for
WorkerDebuggeeRunnables, runnables sent from the worker to the main thread that
should not be delivered to a paused or frozen content window. This queue is
paused and resumed by WorkerPrivate::Freeze, WorkerPrivate::Thaw,
WorkerPrivate::ParentWindowPaused, and WorkerPrivate::ParentWindowResumed.
Since this affects when WorkerDebuggeeRunnables are delivered relative to other
administrative worker runnables, WorkerDebuggeeRunnable must use a
ThreadSafeWorkerRef to ensure that the WorkerPrivate sticks around long enough
for them to run properly.
Depends on D9219
Differential Revision: https://phabricator.services.mozilla.com/D9220
--HG--
extra : moz-landing-system : lando
Separating these runnables out under a separate subclass will let us delay their
delivery while the content window is paused in the debugger.
CancelingOnParentRunnable, used when the worker calls self.close(), to close the
worker from the parent thread, must also be a WorkerDebuggeeRunnable, since it
must be processed only after all prior messages/errors from the worker.
Depends on D9218
Differential Revision: https://phabricator.services.mozilla.com/D9219
--HG--
extra : moz-landing-system : lando
If class A is derived from class B, then an instance of class A can be
converted to B via a static cast, so a slower QI is not needed.
Differential Revision: https://phabricator.services.mozilla.com/D6861
--HG--
extra : moz-landing-system : lando
These maps hold strong references which complicate nsThread lifetime handling
considerably, and only have a couple of fringe uses. We have a linked list of
active threads that the thread manager can use for its internal enumeration
purposes, and the external uses are easily done away with, so there doesn't
seem to be much reason to keep the map around.
MozReview-Commit-ID: x7dsj6C4x8
--HG--
extra : source : 5f870621361012ba459943212d8c68a9ff81cb16
extra : intermediate-source : 89a0c0874d400dd324df6fc3627c0c47d130df19
extra : histedit_source : bbd7900e3d754bde925a411c10aa30a1d6e22edd
These maps hold strong references which complicate nsThread lifetime handling
considerably, and only have a couple of fringe uses. We have a linked list of
active threads that the thread manager can use for its internal enumeration
purposes, and the external uses are easily done away with, so there doesn't
seem to be much reason to keep the map around.
MozReview-Commit-ID: x7dsj6C4x8
--HG--
extra : rebase_source : 897e2d32d1dfee24d51459065925fb9b41fa543a
extra : source : 5f870621361012ba459943212d8c68a9ff81cb16
These maps hold strong references which complicate nsThread lifetime handling
considerably, and only have a couple of fringe uses. We have a linked list of
active threads that the thread manager can use for its internal enumeration
purposes, and the external uses are easily done away with, so there doesn't
seem to be much reason to keep the map around.
MozReview-Commit-ID: x7dsj6C4x8
--HG--
extra : rebase_source : 88c56fa4f5da97f33ade08d892c3d8c42666307e
The patch in the next part will need a handle to the object in
TryPreserveWrapper.
Differential Revision: https://phabricator.services.mozilla.com/D6197
--HG--
extra : moz-landing-system : lando
The logic here to move our check was right, but our check was wrong.
Also, we landed a test that checked for our wrong implementation.
We need to correct our implementation and re-think the test. The
right test might just be a mochitest, possibly with some Firefox-only
hooks involved.
--HG--
extra : rebase_source : 4d6b9a120adcee835f626098e8547c440a39f595
This allows JS callers to automatically get the correct types during
interation, without having to explicitly specify them.
Differential Revision: https://phabricator.services.mozilla.com/D3728
--HG--
extra : rebase_source : b708f382d8ea571d199c669bfed5b5a7ca9ffac4
extra : histedit_source : 7df6feb82088c8a5ca45dc28fe4d2b852c177fee
In order to allow JS callers to use nsISimpleEnumerator instances with the JS
iteration protocol, we'll need to additional methods to every instance. Since
we currently have a large number of unrelated implementations, it would be
best if they could share the same implementation for the JS portion of the
protocol.
This patch adds a stub nsSimpleEnumerator base class, and updates all existing
implementations to inherit from it. A follow-up will add a new base interface
to this class, and implement the additional functionality required for JS
iteration.
Differential Revision: https://phabricator.services.mozilla.com/D3725
--HG--
extra : rebase_source : ad66d7b266856d5a750c772e4710679fab9434b1
extra : histedit_source : a83ebffbf2f0b191ba7de9007f73def6b9a955b8
This patch introduces a new cookie behavior policy called
BEHAVIOR_REJECT_TRACKER. It also makes it possible to override that
behavior with cookie permissions similar to other cookie behaviors.
AutoSafeJSContext enters the unprivileged junk scope, so CreateSandbox returned a wrapper and then we used the wrapper's global instead of the sandbox global. We now use AutoJSAPI with a null realm.
This should hopefully prevent races where we run promise handlers during
worker shutdown, when we're canceling regular events anyway.
Differential Revision: https://phabricator.services.mozilla.com/D2376
--HG--
extra : moz-landing-system : lando
This new id is added in the PerformanceInfo data and helps consumers distinguish
counters.
MozReview-Commit-ID: 7kEmqJcVggM
--HG--
extra : rebase_source : 40cca4c937f846db93ec1315036ad1bac04bc762
DocShells are associated with outer DOM Windows, rather than Documents, so
having the getter on the document is a bit odd to begin with. But it's also
considerably less convenient, since most of the times when we want a docShell
from JS, we're dealing most directly with a window, and have to detour through
the document to get it.
MozReview-Commit-ID: LUj1H9nG3QL
--HG--
extra : source : fcfb99baa0f0fb60a7c420a712c6ae7c72576871
extra : histedit_source : 5be9b7b29a52a4b8376ee0bdfc5c08b12e3c775a
DocShells are associated with outer DOM Windows, rather than Documents, so
having the getter on the document is a bit odd to begin with. But it's also
considerably less convenient, since most of the times when we want a docShell
from JS, we're dealing most directly with a window, and have to detour through
the document to get it.
MozReview-Commit-ID: LUj1H9nG3QL
--HG--
extra : rebase_source : a13c59d1a5ed000187c7fd8e7339408ad6e2dee6
Brindusan Cristian
Yaron Tausky
Ehsan Akhgari
Brian Hackett
Daniel Varga
Olli Pettay
Ryan Hunt
Cosmin Sabou
Jonathan Kingston
Ciure Andrei
Andrea Marchesini
Andreea Pavel
Gurzau Raul
Christoph Kerschbaumer
Bogdan Tara
Benjamin Bouvier
Luke Wagner
Chris Peterson
Alex Gaynor
Dorel Luca
Boris Zbarsky
Tom Schuster
Kris Maglione
Jeff Walden
Jon Coppeard
Sylvestre Ledru
Lars T Hansen
Dave Townsend
Geoff Brown
Tarek Ziadé
Emilio Cobos Álvarez
Margareta Eliza Balazs
Blake Kaplan
Mark Banner
James Willcox
Jason Orendorff
Alexandre Poirot
Gabriele Svelto
Razvan Maries
Sebastian Hengst
Gerald Squelart
Perry Jiang
Jim Blandy
arthur.iakab
Andrew McCreight
Noemi Erli
Narcis Beleuzu
Nicholas Nethercote
Tiberius Oros
Mike Hommey
Jan de Mooij
Andrew Sutherland
Jean-Yves Avenard
Zibi Braniecki
Tooru Fujisawa