Change-Id: Ia1a803bd52e94e15696c60c77938fd2f5c08bdaf
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105981
Commit-Queue: Brian Wilkerson <brianwilkerson@google.com>
Auto-Submit: Brian Wilkerson <brianwilkerson@google.com>
Reviewed-by: Paul Berry <paulberry@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Copied from a change by keertip@ from internal repo.
Original change ID: 253076358
Change-Id: I997ffe7c4fa0588689a2562b7b1dcd5c01837919
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/106002
Reviewed-by: Vijay Menon <vsm@google.com>
Commit-Queue: Nicholas Shahan <nshahan@google.com>
... and update comment and test output to indicate how files should be
regenerated from tools/experimental_features.yaml
Change-Id: I369b93c35b91f7a93c915f0abf9cebdd00c13f8e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/106003
Commit-Queue: Paul Berry <paulberry@google.com>
Auto-Submit: Dan Rubel <danrubel@google.com>
Reviewed-by: Paul Berry <paulberry@google.com>
There is no codegen yet so almost all code compiled with
--experiment-new-rti will crash the compiler.
Change-Id: Idf542646ac3629cb02a50d44c2a98abd156047ed
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/106001
Commit-Queue: Stephen Adams <sra@google.com>
Reviewed-by: Sigmund Cherem <sigmund@google.com>
Reviewed-by: Mayank Patke <fishythefish@google.com>
Rationale:
Avoids rejecting extra parameter when running
on cluster for nightly testing.
Change-Id: I34a5e02638ec93f602609e32f5e460e774656f35
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105968
Reviewed-by: Alexander Thomas <athom@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
This will allow me to put stronger assertions in the DecoratedType
constructor in a follow-up CL.
Change-Id: I375ae2b33c2f14948d9fedbdfbc000b8a1057b00
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105963
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Commit-Queue: Paul Berry <paulberry@google.com>
This sets the "?.[" endGroup to "]" for proper token stream structure
and parser recovery.
Change-Id: Ie4147ac5dedcc273ea6cfda6dafc5d91e892b1a8
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105980
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Commit-Queue: Dan Rubel <danrubel@google.com>
The return value of the GraphBuilder visit methods should always be
the type of the AST node assuming it's used as an expression. When a
type name is used as an expression, its type is always a non-null
reference to the built-in type `Type`.
(Note: this wasn't causing any failures, but fixing it will allow me
to add stronger assertions in a follow-up CL.)
Change-Id: I62c72dcfc6c0f12f8d8a1a7de34a1d3f94c5e4f4
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105962
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Commit-Queue: Paul Berry <paulberry@google.com>
Change-Id: I4d75bc286fe8d5dfe063b30525a99dacc96b4ec6
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105961
Reviewed-by: Dan Rubel <danrubel@google.com>
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Commit-Queue: Paul Berry <paulberry@google.com>
This removes one location where ErrorToken processing is no longer needed
in the parser, and updates the analyzer parser tests to use scanString
rather than instantiating the scanner directly.
Change-Id: I399d771004af7f9fe25d19c1065fbfd425416edc
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105901
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Commit-Queue: Dan Rubel <danrubel@google.com>
This CL updates the parser to parse extension declarations where
the extension name is omitted. For more details, see
https://github.com/dart-lang/language/pull/303/files#diff-9b84d2e3c88265bde2eb43544ddb757dR66
This change explicitly does not handle the case `extension on on on { … }`.
A subsequent CL should either make `on` a builtin
or add more lookahead in the parser to allow this edge case.
Change-Id: Ided821192770f88a51a3300fed8cb0fdef9eea3f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105900
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Commit-Queue: Dan Rubel <danrubel@google.com>
The recipe switched from --output_directory to --output-directory in
https://crrev.com/c/1643427 but the core dump archiver was still
attempting to detect the old property. This caused it to attempt to
upload dumps with gsutil rather than moving the dumps to the output
directory.
Change-Id: Ia73b294eae280ba45b5acf40ec495aba51732694
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105942
Auto-Submit: Alexander Thomas <athom@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Alexander Thomas <athom@google.com>
* move names to Identifiers
* move resolution parts to front end strategy
* move codegen parts to backend strategy
* remove now unneeded methods from FrontendStrategy and BackendStrategy
Change-Id: I0675d56045dd212ad195177ecd23b27b0d849a80
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105742
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Sigmund Cherem <sigmund@google.com>
During constant evaluation, unused arguments to a const constructor are
thrown away after evaluation, since their values do not affect the
resulting instance constant. If such an unused argument ends up
unevaluated, any errors that would arise in the final evaluation are
not reported.
This CL adds space in the Kernel AST for saving these unevaluated
expressions so they can be checked during final constant evaluation.
Even though this is an incompatible change, no update is needed to the
VM code (except for the version bump), since the VM does not support
InstanceCreation nodes in the first place.
Change-Id: I4752562c1164efbba79eb018c15b07ed8354ce5f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105761
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Aske Simon Christensen <askesc@google.com>
Includes changes to implicit covariance test.
To redescribe this, for some tear-off C<T>.f, the member type may be
contravariant if T is referenced by a parameter of C<T>.f (such as if it
is of type void Function(T)). This must be a runtime failure if C<T> is
a reference which has been covariantly upcast at runtime. For example,
List<Object>.add, where the list is at runtime an instance of List<int>,
results in a static type of void Function(Object), but a runtime type of
void Function(int), which is unsound.
We omit these checks, however, when it is trivially sound for all cases
Previously, this was the case for tearing off `List<Null>.add`. This is
no longer a valid test case for NNBD where the runtime type of that list
could actually be `List<Never>`.
Therefore the test has been changed to expect `List<Null>.add` to result
in a cast, and a new NNBD test case has been added to test `List<Never>`
cases don't get a cast.
Change-Id: I61d024a249e2c2a249ad5a8037cbe3d103e3ea8b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/104361
Commit-Queue: Mike Fairhurst <mfairhurst@google.com>
Reviewed-by: Paul Berry <paulberry@google.com>
I'll build on this in future CLs to address more complex cases, where
the type is instantiated to a bound.
Fixes#37213.
Change-Id: I1e23c5599557ef17177483222f15b1a7985b9ca8
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105726
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
dart-bytecode, arm64: +4.742% geomean
dart-bytecode-jit-unopt, arm64: +12.73% geomean
dart2js-compile, x64: +3.635% geomean
In the polymorphic and unlinked cases, call to a stub the does a linear scan against an ICData.
In the monomorphic case, call to a prologue of the expected target function that checks the expected receiver class. There is additional indirection in the JIT version compared to the AOT version to also tick a usage counter so the inliner can make good decisions.
In the megamorphic case, call to a stub that does a hash table lookup against a MegamorphicCache.
Megamorphic call sites face a loss of precision in usage counts. The call site count is not recorded and the usage counter of the target function is used as an approximation.
Monomorphic and megamorphic calls sites are reset to the polymorphic/unlinked state on hot reload.
Monomorphic and megamorphic calls sites do not check the stepping state, so they are reset to the polymorphic/unlinked state when stepping begins and disabled.
Back-edges now increment the usage counter in addition to checking it. This ensures function with loops containing monomorphic calls will eventually cross the optimization threshold.
Fixed backwards use of kMonomorphicEntryOffset and kPolymorphicEntryOffset.
Fixed C stack overflow when bouncing between the KBC interpreter and a simulator.
Bug: https://github.com/dart-lang/sdk/issues/26780
Bug: https://github.com/dart-lang/sdk/issues/36409
Bug: https://github.com/dart-lang/sdk/issues/36731
Change-Id: I78a49cccd962703a459288e71ce246ed845df474
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/102820
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This will simplify some situations where nullability migration would
otherwise have to visit the source code in a very careful order. For
example, when analyzing top level fields undergoing type inference, we
can simply create a node for the nullability of each top level field,
and then later use the "union" operation to hook up those nodes to the
nodes resulting from analyzing their initializers. Without the union
operation, we would have to be careful to visit the top level fields
in dependency order.
Change-Id: I3b07e1abccc5b1c9f1c7c9fa7b13fb6af60c07d6
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105800
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Commit-Queue: Paul Berry <paulberry@google.com>
Previously, they were static, which made them easier to access, but
made it difficult to track their edges (because we couldn't safely
mutate them, so we had to store their edges in special fields of
NullabilityGraph). Now that we're passing NullabilityGraph all over
the place anyhow, there's no benefit to making them static anymore.
Change-Id: Ia3e7d32ae479f40f505621e5d6df04060e39b1c7
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105723
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Rationale:
Previous method cached graph information (instruction and call site
counts) on a per-function level, not accounting for potential
specializations. The improved method runs an extra constant folding
pass, and only caches per-function information for non-specialized
cases. As a result, we inling much better, see for example, the
added test as illustration.
Since we no longer cache for constants, compile-time may be increased
a bit due to the extra scan. In the long run we should consider
for common constant "situations" as the call site.
https://github.com/dart-lang/sdk/issues/36880
Change-Id: I19f007c7f1860ad0ea88fafb38695dc154189ad5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/105460
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
