Move common libraries of wasm and wasm_js_compatibility to wasm_common,
remove wasm_base.
Change-Id: I0293cfd11b58be3d921413fdb2fff7171f0b3d4e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/364320
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
- Fix typos in "Get finalizer callback" -> "Call finalizer callback"
in documentation of an `attach` call.
- Make it clear that the `detach` parameter of `attach` is not used by
the finalizer.
- Make it clear what "the same value" means in the `detach` parameter
documentation.
CoreLibraryReviewExempt: Documentation only
Change-Id: I8c1dc9e5fd6926e5c594c26a50228a044e892974
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/364100
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Previously, Internal_randomInstructionsOffsetInsideAllocateObjectStub
assumed the entry point is at the start of the instructions payload when
choosing a random offset into the instructions payload. However, most
of our architectures have non-zero entry point offsets.
Instead, replace it with two methods, one that returns the start
(inclusive) and one that returns the end (exclusive) of the
AllocateObject stub instructions payload and have the test choose
points inside and outside of that range to test.
TEST=vm/dart/use_dwarf_stack_traces_flag_test
Fixes: https://github.com/dart-lang/sdk/issues/50286
Cq-Include-Trybots: luci.dart.try:vm-aot-dwarf-linux-product-x64-try,vm-aot-linux-release-x64-try,vm-aot-linux-debug-x64-try,vm-aot-mac-release-arm64-try,vm-aot-mac-product-arm64-try
Change-Id: Ia8965c72fbc6f6d2c178778a32f8083923a8b243
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/363080
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Auto-Submit: Tess Strickland <sstrickl@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
The recent rewrite of [0] kept the existing semantics of the finalizer
implementation but migrated to the new interop semantics.
Turns out in that existing semantics there's a bug, namely that it
casted the peer object of to a non-nullable [Object] which is incorrect.
This CL
* fixes the cast to non-nullable [Object]
* moves the (currently) web-specific tests for
[Finalizer]/[WeakReference] into `corelib` so it runs on all backends
* expands the finalizer tests to check finalizers get invoked
* expands the weak reference test to ensure weak reference is cleared
[0] https://dart-review.googlesource.com/c/sdk/+/363082/
Issue https://github.com/dart-lang/sdk/issues/55474
Change-Id: Ibd8c186b39100cff9e2f437f1a737034a5364830
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/363581
Reviewed-by: Ömer Ağacan <omersa@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
HTML spec says the timeout argument in `setInterval` and `setTimeout`
will be 0 when negative [1], update our mocked versions in the d8 script
to implement this.
Fixes tests on dart2wasm-linux-d8:
- co19/LibTest/async/Timer/Timer.periodic_A02_t01
- co19/LibTest/async/Timer/Timer_A02_t01
These tests already pass on Firefox and Chrome.
Fixes#54598.
[1]: https://html.spec.whatwg.org/multipage/timers-and-user-prompts.html#dom-setinterval-dev
in "Timer initialization steps" step 4: "If timeout is less than 0, then
set timeout to 0."
Change-Id: I716f5942b5dff0fbd5a45c2bb6e0ef2732a5ea67
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362260
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Nate Biggs <natebiggs@google.com>
With the new type system type arguments appearing in extends are
no longer represented by a class definition and do not need to
trigger any lazy logic.
Change-Id: I3f80f9f972bec69c678909c12d25a604a8fe2c59
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/353205
Commit-Queue: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Mark Zhou <markzipan@google.com>
Also removes a redundant variable.
Change-Id: I6ff36f2a007f4248b9ec4b3c1fc3fe78a2bb6e85
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/363560
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This enables the currently dart2js/ddc-only tests for [WeakReference] /
[Finalizer] on dart2wasm.
As part of that we add a polyfill for the two implementations (just like
dart2js/ddc) do, to make the polyfill test pass.
We also migrate the implementation to use static interop.
Issue https://github.com/dart-lang/sdk/issues/55474
TEST=tests/web/wasm_js_shared/weak_reference_*test.dart
Change-Id: Ic08ae41b6fc78fd9857172ca24695a7f91e06a78
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/363082
Reviewed-by: Ömer Ağacan <omersa@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
This is a reland of commit e6c9eaaf6b
Original change's description:
> [SDK] - Ensure we only build an AOT snapshot for the frontend-server
>
> Currently we seem to build both the JIT and AOT snapshots of the
> frontend-server, ensure we only build the AOT snapshot on all
> architectures except IA32
>
> Change-Id: Ie79dc6d88fd9fa680267571e5d65081ca57609c6
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359100
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Siva Annamalai <asiva@google.com>
Change-Id: Id9a695f33e4e808463b4f65bdb3cd0e546a695d1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/363480
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Siva Annamalai <asiva@google.com>
Fix trim methods returning a new string (instead of the argument) when
there's nothing to trim.
A second bug fixed in NEL handling in `trim`: the last line should be
`result.substring(...)` instead of `this.substring(...)`. The bug wasn't
caught by any of the tests: NEL needs to be handled specially on the
browser, but mixing NEL with other whitespace wasn't tested in the
existing tests `string_trim_test` and `string_trimlr_test`, so a new
test file added with this case.
Change-Id: Ibf84e185de1b26c9811e3ea58c2ad3f223da8515
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/363081
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Lasse Nielsen <lrn@google.com>
This is a reland of commit a7e2dce0cc
Fixes the crash during code generation when typed data view
LoadIndexed was used on an internal typed data array
(in the unreachable code path) without loading its payload.
TEST=runtime/tests/vm/dart/regress_b_334128316_test.dart
TEST=manually verified repro from b/334128316
Original change's description:
> [vm] Implement typed data view GetIndexed in flow graph builder
>
> Typed data view 'operator []' is now implemented in the flow
> graph builder. This unifies all typed data GetIndexed operations
> and removes unnecessary address calculations.
>
> TEST=ci
>
> Change-Id: I8d2ca6e7c7bf18b2590536a643f92cad5beb6d95
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/361283
> Reviewed-by: Tess Strickland <sstrickl@google.com>
> Commit-Queue: Alexander Markov <alexmarkov@google.com>
> Reviewed-by: Slava Egorov <vegorov@google.com>
Change-Id: I3786fa17e971f8ba6cd01a57f1a68504bf24bc47
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362821
Reviewed-by: Tess Strickland <sstrickl@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
This reverts commit 102f6c7f33.
Reason for revert: Instead of using VM's implementation we will
refactor and fix the old implementation that calls JS `trim`.
Since the original CL fixed some tests, revert will break them again.
Original change's description:
> [dart2wasm] Fix trim methods of JS strings
>
> We can't use JS `String.p.trim` (and the left/right trims) as we need to
> return the original argument when itd oesn't have any whitespace to
> remove.
>
> They also don't handle the NEL character as expected by Dart.
>
> Instead use the VM's trim implementations.
>
> Fix extracted from https://dart-review.googlesource.com/c/sdk/+/361362
> to keep the CL small.
>
> Change-Id: I5cd6cc2d4349f7b55544257666faff3bbeb00b8c
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362388
> Reviewed-by: Slava Egorov <vegorov@google.com>
> Commit-Queue: Ömer Ağacan <omersa@google.com>
Change-Id: I9c4a3402213f5ace0da55584e737bf62bb51924b
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362802
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
We can't use JS `String.p.trim` (and the left/right trims) as we need to
return the original argument when itd oesn't have any whitespace to
remove.
They also don't handle the NEL character as expected by Dart.
Instead use the VM's trim implementations.
Fix extracted from https://dart-review.googlesource.com/c/sdk/+/361362
to keep the CL small.
Change-Id: I5cd6cc2d4349f7b55544257666faff3bbeb00b8c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362388
Reviewed-by: Slava Egorov <vegorov@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
When calling `String.p.replace` we need to escape `$` characters in
replacement argument as `$$` as Dart doesn't have the `$` replacement
syntax.
Fix extracted from https://dart-review.googlesource.com/c/sdk/+/361362
to keep the CL small.
Change-Id: I23a26dda48ed0d5ccb7fea5353f16350a2b046f3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362387
Commit-Queue: Ömer Ağacan <omersa@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
Converting the int to JS `double` fails with large integers. Pass the
`int` as Wasm `i64`, which will implicitly convert it to `BigInt`, and
then use `BigInt.p.toString`.
Fix extracted from https://dart-review.googlesource.com/c/sdk/+/361362
to keep the CL small.
Change-Id: Ia5723632d11fd8544789cf453bb5c0e36ac8f6bb
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362461
Reviewed-by: Slava Egorov <vegorov@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
Typed data view 'operator []' is now implemented in the flow
graph builder. This unifies all typed data GetIndexed operations
and removes unnecessary address calculations.
TEST=ci
Change-Id: I8d2ca6e7c7bf18b2590536a643f92cad5beb6d95
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/361283
Reviewed-by: Tess Strickland <sstrickl@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
Vm flag --null-assertions only makes sense without sound null safety.
Now that VM no longer supports unsound mode, the flag is obsolete and
can be removed.
TEST=ci
Change-Id: I69eb2012ac483f082fd43ecc9743076e0e1928c8
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362181
Reviewed-by: Siva Annamalai <asiva@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
This reverts commit ff5f391c0a.
Reason for revert: The expect library is used by Flutter engine, and some of its tests use assertStatementsEnabled. There should be a migration path that doesn't require an atomic change, like adding the replacement api before removing the old one.
Original change's description:
> Tweak `expect.dart` library.
>
> Make API more consistent for a few methods.
> Reduce the number of language features used in tests:
> * Never iterating an iterable, always converting it
> using `.toList()` first and iterating using indices
> (fx `setEquals`).
> Also require a `List` in places where an `Iterable`
> wasn't necessary.
> * Avoid doing complicated computations that are also
> used for the error message. Do simple check first,
> then recompute to get better error messages
> (fx `allDistinct`).
>
> Renamed some rarely used members for consistency
> (`stringContainsInOrder`->`containsInOrder`,
> where other string-contains functions just start
> with `contains`, and `containsOneOf` -> `containsAny`
> to match `Iterable.any` phrasing, and also it accepts
> if containing at least one, not precisely one.)
>
> Removed a function that wasn't used anywhere.
>
> Moved `assertStatementsEnabled` to `variations.dart` as `asserts`.
> Removed `typeAssertionsEnabled` and `checkedModeEnabled`. The former used in one place, where it was replaced with `checkedImplicitDowncasts` from `variations.dart`, the latter wasn't used anywhere.
>
> Deprecates `package:expect/minitest.dart`. It was never intended
> to be used for new tests, only as a help to convert existing tests
> written against `package:unit_test`.
> All existing imports marked as `// ignore: deprecated_member_use`.
>
> Change-Id: I07e21d4c0f3ccf11b82ee34af2668fdbb22264d2
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/352360
> Reviewed-by: Slava Egorov <vegorov@google.com>
> Reviewed-by: Ömer Ağacan <omersa@google.com>
> Reviewed-by: Nate Bosch <nbosch@google.com>
> Reviewed-by: Stephen Adams <sra@google.com>
> Commit-Queue: Lasse Nielsen <lrn@google.com>
Change-Id: I360b4347470a0bb2b63c3108e2b83ee2a771bf3f
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/362020
Reviewed-by: Nate Bosch <nbosch@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
Reviewed-by: Stephen Adams <sra@google.com>
Reviewed-by: Leaf Petersen <leafp@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: William Hesse <whesse@google.com>
Make API more consistent for a few methods.
Reduce the number of language features used in tests:
* Never iterating an iterable, always converting it
using `.toList()` first and iterating using indices
(fx `setEquals`).
Also require a `List` in places where an `Iterable`
wasn't necessary.
* Avoid doing complicated computations that are also
used for the error message. Do simple check first,
then recompute to get better error messages
(fx `allDistinct`).
Renamed some rarely used members for consistency
(`stringContainsInOrder`->`containsInOrder`,
where other string-contains functions just start
with `contains`, and `containsOneOf` -> `containsAny`
to match `Iterable.any` phrasing, and also it accepts
if containing at least one, not precisely one.)
Removed a function that wasn't used anywhere.
Moved `assertStatementsEnabled` to `variations.dart` as `asserts`.
Removed `typeAssertionsEnabled` and `checkedModeEnabled`. The former used in one place, where it was replaced with `checkedImplicitDowncasts` from `variations.dart`, the latter wasn't used anywhere.
Deprecates `package:expect/minitest.dart`. It was never intended
to be used for new tests, only as a help to convert existing tests
written against `package:unit_test`.
All existing imports marked as `// ignore: deprecated_member_use`.
Change-Id: I07e21d4c0f3ccf11b82ee34af2668fdbb22264d2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/352360
Reviewed-by: Slava Egorov <vegorov@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
Reviewed-by: Nate Bosch <nbosch@google.com>
Reviewed-by: Stephen Adams <sra@google.com>
Commit-Queue: Lasse Nielsen <lrn@google.com>
These were added in 1436635d27.
CoreLibraryReviewExempt: No functional change. Only adding since version metadata.
Change-Id: Ifc20265df2694d8ce410025f1d1f6a65f5da81d2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/361500
Commit-Queue: Marya Belanger <mbelanger@google.com>
Reviewed-by: Marya Belanger <mbelanger@google.com>
Reviewed-by: Lasse Nielsen <lrn@google.com>
This reverts commit e6c9eaaf6b.
Reason for revert: We seem to have issues rolling Dart into Flutter frame works because code outside of the Dart SDK seems to rely on the existence of this snapshot.
Original change's description:
> [SDK] - Ensure we only build an AOT snapshot for the frontend-server
>
> Currently we seem to build both the JIT and AOT snapshots of the
> frontend-server, ensure we only build the AOT snapshot on all
> architectures except IA32
>
> Change-Id: Ie79dc6d88fd9fa680267571e5d65081ca57609c6
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359100
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Siva Annamalai <asiva@google.com>
Change-Id: I0e79cc688aaec7dc5428ccbbaff7f1dbea11a962
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/361285
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Siva Annamalai <asiva@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
The `dart development-service` command will be used by tooling to launch
DDS from the SDK instead of shipping DDS via package:dds.
TEST=Existing service test suite
Change-Id: Ib928aa5b8961caf87d7074884c3d226b5c096ccd
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/361180
Reviewed-by: Derek Xu <derekx@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
R=iinozemtsev@google.com
CoreLibraryReviewExempt: Documentation change
Change-Id: I3354f7d4a0c9e5a7281ea8764431139deeac1014
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/361220
Auto-Submit: 枯枯坐 <kukuzuo@gmail.com>
Reviewed-by: Alexander Thomas <athom@google.com>
Reviewed-by: Lasse Nielsen <lrn@google.com>
Commit-Queue: Alexander Thomas <athom@google.com>
- Add a runtime flag to enable checks on non-nullable APIs that
return values from non-static JavaScript interop.
- Call a new helper method at the call site when these APIs are
detected to perform the null check.
- Add test file for the cases we can detect and enforce.
NOTE: This does not make non-static JavaScript interop sound.
This only adds more checks to enforce soundness with respect
to nullability in some cases. There are still holes that will
never be closed due to the permissive nature of this form of
JavaScript interop.
Change-Id: I2f88d1543a683fdc84d764e2b0eaafeb0ca73107
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/358581
Reviewed-by: Samuel Rawlins <srawlins@google.com>
Reviewed-by: Mayank Patke <fishythefish@google.com>
Commit-Queue: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Srujan Gaddam <srujzs@google.com>
Reviewed-by: Sigmund Cherem <sigmund@google.com>
Currently we seem to build both the JIT and AOT snapshots of the
frontend-server, ensure we only build the AOT snapshot on all
architectures except IA32
Change-Id: Ie79dc6d88fd9fa680267571e5d65081ca57609c6
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359100
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Siva Annamalai <asiva@google.com>
Change-Id: Idab8e7823512898b614de0f944af4de720dced7e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/360860
Reviewed-by: Kevin Moore <kevmoo@google.com>
Commit-Queue: Ömer Ağacan <omersa@google.com>
The previous logic for building the path to dds.dart.snapshot would result
in the Windows drive letter being dropped from the path:
\path\to\dart-sdk\bin\dds.dart.snapshot
This works most of the time since a leading slash is treated as a reference
to the current drive, which often contains the Dart SDK. However, if the SDK
is on a different drive than the current (e.g., in a container with two drives),
the VM will fail to find the snapshot.
This change uses the File(...) APIs from dart:io to build the path rather than
trying to use the Uri class to manually hack together a path.
TEST=N/A, not reproducible without a second Windows drive
Change-Id: I78e9f2980ac583f3784d0c7a5333ec2d235cc113
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/360920
Reviewed-by: Derek Xu <derekx@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
This makes it explicit on the call site what the integer is,
simplifying follow up CLs.
TEST=tests/ffi
CoreLibraryReviewExempt: implementation change only.
Bug: https://github.com/dart-lang/sdk/issues/54739
Change-Id: Ib5bc09f4194e9103cf42eb3851b2308553fd8990
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/360640
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
This is a reland of commit d24b5d1f5e
On top of the original change, the following is fixed and improved:
1) Flow graph builder (FGB) body is now also applied to
dynamic invocation forwarders (similarly to graph
intrinsics).
2) Frame can be omitted for functions which have a
call on a shared slow path. This is needed to make
FGB implementation of GetIndexed frameless, as it has
GenericCheckBound which calls on shared slow path.
(Graph intrinsics are frameless).
3) Range analysis is enabled for force-optimized functions,
so more efficient code can be generated for boxing
instructions. Range analysis is fixed to avoid crashes
and correctly intersect ranges with constant boundaries
(needed for some force-optimized FFI functions).
4) EliminateStackOverflowChecks pass is enabled for
force-optimized functions so CheckStackOverflow can be
eliminated.
Original change's description:
> [vm] Replace array GetIndexed graph intrinsics with flow graph builder implementation
>
> _Array, _GrowableList, internal and external typed data 'operator []'
> are now implemented in the flow graph builder.
>
> Unlike graph intrinsics, flow graph created in the flow graph builder
> can be used by the inliner. Corresponding graph intrinsics and native
> methods are removed.
>
> Also, this change adds missing external typed data indexing operations.
>
> TEST=ci
>
> Change-Id: Ic19784481feadf54c096a587413e67b4e18353dc
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359940
> Reviewed-by: Ryan Macnak <rmacnak@google.com>
> Commit-Queue: Alexander Markov <alexmarkov@google.com>
TEST=ci
Change-Id: I04ef008a04238d432683d7543cd047e35bad17c3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/360560
Reviewed-by: Tess Strickland <sstrickl@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
In some cases the code downloaded through this hook is corrupted and one or more part files don't get registered correctly. The implementer of this hook just gets a 'download success' event and calls the success callback. It's not until dart2js itself checks if the part file as registered that we know it failed. In this case we should retry the load request with only the failed part files.
This matches the retry logic from the non-multi deferred load hook.
Change-Id: I655acd9f519cba1837d5cb367365a7be7254df03
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/360380
Reviewed-by: Mayank Patke <fishythefish@google.com>
Commit-Queue: Nate Biggs <natebiggs@google.com>
This reverts commit d24b5d1f5e.
Reason for revert: performance regression
(Graph intrinsics are applied to dynamic invocation forwarders,
but flow graph builder implementation isn't.)
Original change's description:
> [vm] Replace array GetIndexed graph intrinsics with flow graph builder implementation
>
> _Array, _GrowableList, internal and external typed data 'operator []'
> are now implemented in the flow graph builder.
>
> Unlike graph intrinsics, flow graph created in the flow graph builder
> can be used by the inliner. Corresponding graph intrinsics and native
> methods are removed.
>
> Also, this change adds missing external typed data indexing operations.
>
> TEST=ci
>
> Change-Id: Ic19784481feadf54c096a587413e67b4e18353dc
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359940
> Reviewed-by: Ryan Macnak <rmacnak@google.com>
> Commit-Queue: Alexander Markov <alexmarkov@google.com>
Change-Id: I2b31b06edaa4b8a09d256f25f923d4489e28518b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/360305
Auto-Submit: Alexander Markov <alexmarkov@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
_Array, _GrowableList, internal and external typed data 'operator []'
are now implemented in the flow graph builder.
Unlike graph intrinsics, flow graph created in the flow graph builder
can be used by the inliner. Corresponding graph intrinsics and native
methods are removed.
Also, this change adds missing external typed data indexing operations.
TEST=ci
Change-Id: Ic19784481feadf54c096a587413e67b4e18353dc
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359940
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
This CL
* adds a `kind` to `CheckWritableInstr`,
* starts using `CheckWritableInstr` in JIT mode, and
* adds support for `CheckWritableInstr` ia32.
TEST=tests/ffi/vmspecific_native_finalizer_deeply_immutable_test.dart
Closes: https://github.com/dart-lang/sdk/issues/55067
Change-Id: I0b397daba12cfc8b885401169889f7cd7c040166
Cq-Include-Trybots: dart-internal/g3.dart-internal.try:g3-cbuild-try
Cq-Include-Trybots: dart/try:vm-aot-android-release-arm64c-try,vm-aot-android-release-arm_x64-try,vm-aot-linux-debug-x64-try,vm-aot-linux-debug-x64c-try,vm-aot-mac-release-arm64-try,vm-aot-mac-release-x64-try,vm-aot-obfuscate-linux-release-x64-try,vm-aot-optimization-level-linux-release-x64-try,vm-aot-win-debug-arm64-try,vm-aot-win-debug-x64c-try,vm-aot-win-release-x64-try,vm-appjit-linux-debug-x64-try,vm-asan-linux-release-x64-try,vm-checked-mac-release-arm64-try,vm-eager-optimization-linux-release-ia32-try,vm-eager-optimization-linux-release-x64-try,vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64c-try,vm-ffi-qemu-linux-release-arm-try,vm-ffi-qemu-linux-release-riscv64-try,vm-linux-debug-ia32-try,vm-linux-debug-x64-try,vm-linux-debug-x64c-try,vm-mac-debug-arm64-try,vm-mac-debug-x64-try,vm-msan-linux-release-x64-try,vm-reload-linux-debug-x64-try,vm-reload-rollback-linux-debug-x64-try,vm-ubsan-linux-release-x64-try,vm-win-debug-arm64-try,vm-win-debug-x64-try,vm-win-debug-x64c-try,vm-win-release-ia32-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359223
Auto-Submit: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Daco Harkes <dacoharkes@google.com>
This CL changes compounds (structs, unions, and arrays) to be backed
by a TypedDataBase and an int offset.
Before this CL, the compounds where only backed by a TypedDataBase.
This leads to the following issues:
1. Access to nested structs required code for allocating new typed data
views or pointers, which the optimizer then had to prevent from
being allocated after inlining.
2. Runtime branching on whether the TypedDataBase was a Pointer or
TypedData increased code size and prevented inlining.
https://github.com/dart-lang/sdk/issues/54892
This could not be properly optimized if in AOT both typed
data and pointer were flowing into the same compound.
3. Constructing TypedData views required calculating the length of the
view.
After this CL, accessing nested compounds will lead to accesses on the
original TypedDataBase with an extra offset.
This removes the polymorphism on TypedData vs Pointer, because the
final int/float/Pointer accesses in nested compounds operate on
TypedDataBase.
Also, it simplifies creating an `offsetBy` accessor, because it will
no longer have to be polymorphic in typed data vs pointer, nor will it
have to calculate the length of the field.
Implementation details:
* The changes in the CFE and patch files are straightforward.
* VM: Struct-by-value returns (or callback params) are initialized
with an offsetInBytes of 0.
* VM: Struct-by-value arguments (and callback return) need to read out
the offsetInBytes. Before this CL we were passing in the TypedData
as tagged value. With this CL we are passing the TypedData as tagged
value and the offset as unboxed int, from 1 IL input to 2 IL
inputs. (The alternative would have been to take the compound as
a tagged value, but that would have prevented optimizations from not
allocating the compound object in the optimizer.
The FfiCallInstr is updated to have two definitions for the case
where we were passing in the TypedData previously.
The NativeReturnInstr is refactored to be able to take two inputs
instead of 1. (Note that we don't have VariadicInstr only
VariadicDefinition in the code base. So the instruction is _not_
implemented as variadic, rather as having a fixed length of 2.)
* dart2wasm does no longer support nested compounds due to the
compound implementation only storing a pointer address.
https://github.com/dart-lang/sdk/issues/55083
Intending to land this after
https://dart-review.googlesource.com/c/sdk/+/353101.
TEST=test/ffi
CoreLibraryReviewExempt: VM and WASM-only implementation change.
Closes: https://github.com/dart-lang/sdk/issues/54892
Bug: https://github.com/dart-lang/sdk/issues/44589
Change-Id: I8749e21094bf8fa2d5ff1e48b6b002c375232eb5
Cq-Include-Trybots: dart-internal/g3.dart-internal.try:g3-cbuild-try
Cq-Include-Trybots: dart/try:vm-aot-android-release-arm64c-try,vm-aot-android-release-arm_x64-try,vm-aot-linux-debug-x64-try,vm-aot-linux-debug-x64c-try,vm-aot-mac-release-arm64-try,vm-aot-mac-release-x64-try,vm-aot-obfuscate-linux-release-x64-try,vm-aot-optimization-level-linux-release-x64-try,vm-aot-win-debug-arm64-try,vm-aot-win-debug-x64-try,vm-aot-win-debug-x64c-try,vm-appjit-linux-debug-x64-try,vm-asan-linux-release-x64-try,vm-checked-mac-release-arm64-try,vm-eager-optimization-linux-release-ia32-try,vm-eager-optimization-linux-release-x64-try,vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64c-try,vm-ffi-qemu-linux-release-arm-try,vm-ffi-qemu-linux-release-riscv64-try,vm-fuchsia-release-x64-try,vm-linux-debug-ia32-try,vm-linux-debug-x64-try,vm-linux-debug-x64c-try,vm-mac-debug-arm64-try,vm-mac-debug-x64-try,vm-msan-linux-release-x64-try,vm-reload-linux-debug-x64-try,vm-reload-rollback-linux-debug-x64-try,vm-ubsan-linux-release-x64-try,vm-win-debug-arm64-try,vm-win-debug-x64-try,vm-win-debug-x64c-try,vm-win-release-ia32-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/354226
Reviewed-by: Tess Strickland <sstrickl@google.com>
Closes https://github.com/dart-lang/sdk/issues/55187
Adds a faster way for users to pass opaque Dart values to
JS without the need for boxing like in JSBoxedDartObject.
This does mean, however, that this new type can't be a JS type,
and therefore cannot have interop members declared on it.
Refactors existing code to handle that distinction.
CoreLibraryReviewExempt: Backend-specific library that's been reviewed by both dart2wasm and JS compiler teams.
Change-Id: Ia86f1fe3476512fc0e5f382e05739713b687f092
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/358224
Reviewed-by: Sigmund Cherem <sigmund@google.com>
This reverts commit 9967075787.
Reason for revert: Ignoring `done` is asymmetrical with other IOSinks.
Original change's description:
> [io] Fix a bug where Process.stdin.add exceptions could not be caught
>
> Change-Id: I2383a74bfa6950ab8f8934087fb68218f06dd681
> Bug:https://github.com/dart-lang/sdk/issues/48501
> Tested: Unit test
> CoreLibraryReviewExempt: dart:io only
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/351380
> Reviewed-by: Alexander Aprelev <aam@google.com>
> Commit-Queue: Brian Quinlan <bquinlan@google.com>
Bug: https://github.com/dart-lang/sdk/issues/48501
Change-Id: Ib5356c640d4dddc30e561b9d60d93f794a935103
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359680
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Brian Quinlan <bquinlan@google.com>
This change aligns all typed data implementations with Int8 typed data
and allows compiler to recognized and replace typed data access
operations earlier, before method inlining. Consequently, this increases
chances for [] and []= to be inlined. In JIT, this also prevents
non-inlined calls to _getIndexed/_setIndexed as [] and []= are graph
intrinsics which bypass calls to _getIndexed/_setIndexed.
TEST=ci
Change-Id: I95a85963da07fd1c04c06e0af09f76eb15f93353
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359520
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Related issues:
https://github.com/dart-lang/sdk/issues/55024https://github.com/dart-lang/sdk/issues/55267
These operators were initially broken in 3.3 and were exposed
as returning JSBoolean but implemented as returning bool. They
were fixed to return bool in the public API, but we should
prefer to have them return JS types as they're likely to be used
in cases where implicit conversions are not useful.
CoreLibraryReviewExempt: Fixing type mismatch in backend-specific library.
Change-Id: I3b0e60550dcac78918f8399d11238dcfa34982cd
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/359180
Commit-Queue: Srujan Gaddam <srujzs@google.com>
Reviewed-by: Sigmund Cherem <sigmund@google.com>