This updates the Initializer classes such that InvalidInitializer has a message and will throw at runtime, similar to InvalidExpression. All initializers have positions and only FieldInitializer and SuperInitializer have an isSynthetic field.
TEST=pkg/front_end/testcases/general/invalid_initializer.dart
Change-Id: Ic444bb1dc273b433ac70249240527d115deaa332
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/463661
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
Now moves mixin applications into their own libraries that can be
imported separately from the rest of the code in the originating
library.
Then adds an import to the new library containing the mixin application
to any libraries that use it.
This ensures the import graph is accurate while only making sure it
remains minimal.
TEST=Added case to pkg/vm testcases.
Bug: https://github.com/dart-lang/sdk/issues/43089
Change-Id: I56b6ea1db8261e0d0508c2d7e62d0c13fbe9678b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/457260
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Nate Biggs <natebiggs@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
As a part of this CL, the following was changed:
* Instead of `VariableInitialization` implementing `VariableDeclaration`, `VariableDeclaration` now implements `VariableInitialization`.
* `InternalLocalVariable` now implements `LocalVariable`. It delegates most of its members to the variable it wraps around, and the implementations of the `TreeNode` members are throwing.
* Some tests in the bytecode generator under the experimental type started to fail and are skipped.
Part of https://github.com/dart-lang/sdk/issues/61572
Change-Id: I06b391cc762d5935f528d208ff5e27eccf311a47
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/462981
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Chloe Stefantsova <cstefantsova@google.com>
This adds an internal wrapper expression for redirecting factory invocations. This holds a reference to the original redirecting factory constructor along with the resulting invocation of the generative or factory constructor. The node is removed during constant evaluation and only available to backends through the pre-modular transformations.
TEST=existing
Change-Id: I35c19766117ab2942aefddb4ccebcd2df66babf3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/454820
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Chloe Stefantsova <cstefantsova@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This CL reduces the noise of future CFE experiments in the
backends. Whenever new nodes will be added as an experiment, the
related exception-throwing implementations will be added to the
mixins, eliminating the need to add the visit methods to the backend
code directly. The experimental type addresses a similar issue in
exhaustive switch statements and expressions.
TEST=existing
Change-Id: I1a021e11a159a992500c42287cbe6ef027bf54a2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/455080
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Nate Biggs <natebiggs@google.com>
Reviewed-by: Ömer Ağacan <omersa@google.com>
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Mayank Patke <fishythefish@google.com>
This CL adds two tools
* A dill extractor tool which extracts the source files inside the dill
to a new directory, trying to recrease the package config too so the
sources can be compiled again. Thought as being useful for doing
benchmarks between the CFE and the analyzer to make sure they look at
the exact same source files. No more. No less.
* A CFE compile entry point that compiles all files in a directory.
Again thought as being useful to compare the CFE with the Analyzer.
Additionally it introduces "--gcs=<int>" to `benchmarker.dart` so it can
do multiple runs and do statistics on combined gc times.
Usage example:
```
$ out/ReleaseX64/dart pkg/front_end/tool/compile.dart pkg/front_end/tool/compile.dart
$ out/ReleaseX64/dart-sdk/bin/dart pkg/kernel/bin/dill_extractor.dart pkg/front_end/tool/compile.dart.dill /tmp/extracted_compile_dart_compile
Done. Wrote 687 source files.
$ out/ReleaseX64/dart-sdk/bin/dart compile aot-snapshot pkg/front_end/tool/compile_files_in_folders.dart
Generated: [...]/pkg/front_end/tool/compile_files_in_folders.aot
$ out/ReleaseX64/dart-sdk/bin/dart compile aot-snapshot pkg/analyzer_cli/bin/analyzer.dart
Generated: [...]/pkg/analyzer_cli/bin/analyzer.aot
$ time out/ReleaseX64/dart-sdk/bin/dartaotruntime pkg/front_end/tool/compile_files_in_folders.aot /tmp/extracted_compile_dart_compile/
Got 626 libraries.
Finished in 0:00:03.496817
real 0m3.530s
user 0m4.985s
sys 0m0.257s
$ time out/ReleaseX64/dart-sdk/bin/dartaotruntime pkg/analyzer_cli/bin/analyzer.aot /tmp/extracted_compile_dart_compile/
Analyzing /tmp/extracted_compile_dart_compile...
warning • Target of URI doesn't exist: 'package:compiler/src/io/source_file.dart'. • package:_fe_analyzer_shared/src/scanner/utf8_bytes_scanner.dart:5:16 • uri_does_not_exist_in_doc_import
1 warning found.
real 0m8.479s
user 0m9.997s
sys 0m1.048s
$ out/ReleaseX64/dart pkg/front_end/tool/benchmarker.dart --silent --iterations=10 --gcs=5 --snapshot=pkg/front_end/tool/compile_files_in_folders.aot --snapshot=pkg/analyzer_cli/bin/analyzer.aot --arguments="/tmp/extracted_compile_dart_compile/"
Will now run 10 iterations with 2 snapshots.
..............................
Comparing snapshot #1 (compile_files_in_folders.aot) with snapshot #2 (analyzer.aot)
msec task-clock:u: 141.1012% +/- 1.5550% (6281.55 +/- 69.22) (4451.81 -> 10733.36)
page-faults:u: 47.5414% +/- 0.7786% (45602.90 +/- 746.84) (95922.50 -> 141525.40)
cycles:u: 119.3182% +/- 1.5751% (22131655652.90 +/- 292155654.08) (18548436164.50 -> 40680091817.40)
instructions:u: 140.2463% +/- 0.0340% (30410953910.60 +/- 7375789.67) (21683959961.20 -> 52094913871.80)
branch-misses:u: 97.9816% +/- 7.3049% (73071988.00 +/- 5447773.30) (74577266.40 -> 147649254.40)
seconds time elapsed: 141.0789% +/- 1.5479% (6.28 +/- 0.07) (4.45 -> 10.74)
seconds user: 124.4165% +/- 1.8322% (5.31 +/- 0.08) (4.26 -> 9.57)
seconds sys: 521.0422% +/- 17.4291% (0.98 +/- 0.03) (0.19 -> 1.16)
Comparing GC:
Combined GC time: 63.8901% +/- 1.2946% (1008.76 +/- 20.44) (1578.90 -> 2587.66)
```
(note that benchmarker passes `--deterministic` and limits the run to one cpu, both of which (potentially) makes it slower --- which is why the stats say 4.45 s instead of ~3.53 s and 10.74 s instead of ~8.48 s)
Change-Id: I785620ce40af11ca1f7b6c88a0ac863b2200f7d3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/449180
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
It's not necessary for `package:kernel` and
`package:_fe_analyzer_shared` to have duplicate implementations of
this method. `package:kernel` depends on
`package:_fe_analyzer_shared`, so they can share one definition.
Change-Id: I6a6a696462a0e22a2ea818c42f58187c45821afc
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/447842
Commit-Queue: Paul Berry <paulberry@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
I've been experimenting with using the Jj source control tool, and
since it's not natively supported by `depot_tools`, this means I have
to manually invoke `git cl presubmit`. `git cl presubmit` accepts a
single argument which is the name of the upstream branch.
The way this is intended to work is that the upstream branch name gets
passed into the `PRESUBMIT.py` scripts, and they use it rather than
making assumptions about what the upstream branch is.
Prior to this change, our presubmit scripts for _fe_analyzer_shared,
front_end, frontend_server, and kernel were ignoring the upstream
branch and instead using git's `@{u}` shorthand (see
https://git-scm.com/docs/git-rev-parse). This caused them to behave
strangely when no upstream branch is set, which sometimes happens when
running `git cl presubmit` manually.
This change avoids the strange behavior by getting the upstream branch
from the input to `PRESUBMIT.py`, as intended.
Change-Id: I6a6a696423221d7b945b083fd72585f1f5a7e312
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/447626
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Commit-Queue: Paul Berry <paulberry@google.com>
For expression evaluation we want to be "more than dart" in that if for
instance we can see (in the debugger) that a List contains `B`s (even
if it's typed as containing `A`s) we'd like to be able to access things
on `B` (without manually having to cast to either `B` or `dynamic`).
Furthermore - when we in the debugger can see that it's a `B`, and that
`B` has, say, a field `_privateField` or a method `_privateMethod` we'd
like to be able to access that even if `B` is in another library.
This CL - for expression evaluation - makes dynamic accesses and calls
where we would normally issue a "missing getter" (etc) error, and tries
to create a `Name` so private access is possible.
Change-Id: I887318a50413e9a5f11ec685b27719edd312dca0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/446260
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Note: Const locals are still off for VM aot and dart2js for the entry
points I've found in an attempt to retain the old behaviour there.
It might be better if those targets could remove such locals in a
whole-world analysis instead.
* Keep const locals by default (except as noted above). Update the
verifier to accept that. For the platforms this has increased the
size by at most 6584 bytes. With this the VM will pass in any const
locals as it does normal locals, but as the variable is never
captured it will never pass a const local defined in a method when
inside a local function in that method.
* Change the dart scope calculation(s) to return the found variables
instead of just the types of the found variables.
* When the incremental compilers expression compilation - via the dart
scope calculation - finds a const local that it wasn't told about, it
will pass it on as an extra variable that it knows about, allowing
for evaluating const locals in the case not covered by the first
bullet.
With luck this can in future CLs be extended to know about other
variables that we're not told about, allowing to give a message saying
something like
"yes, we know what 'foo' is, but you can't currently use it" as wanted
in for instance https://github.com/dart-lang/sdk/issues/60316 and
https://github.com/dart-lang/sdk/issues/53996.
Tested: Existing tests for existing functionality; new tests for the new
Change-Id: I1ec24350273e6f81574bb2888f6bf46e3b8b1b47
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/445461
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Nicholas Shahan <nshahan@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
For the generated files (e.g. codes) we had gotten into a scenario
where the auto generated files was forced to be formatted as short
style, but if asking the dart formatter to format the
_fe_analyzer_shared file it would be formatted in long style because the
minimum version for _fe_analyzer_shared is now 3.7.
This CL updates the generators for the generated files to pass the
language version from the package config to the formatter.
Change-Id: I986ce1b5ff65244499ab5a277e78124f372ee46d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/444300
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
These are from `_Array.iterator` whereas the previous ones was from,
iirc, `List.iterator`.
* 7.41mb Import.finalizeImports
* 1.13mb FindTypeVisitor.visitInterfaceType
* 856.39kb new NamedTypeBuilderImpl
* 815.92kb BodyBuilder.createSuperParametersAsArguments
(values from uprobe)
When compiling the CFE with the CFE I get this:
```
msec task-clock:u: -5.6347% +/- 0.5257% (-321.96 +/- 30.04) (5713.89 -> 5391.92)
page-faults:u: 5.9842% +/- 0.0970% (6310.26 +/- 102.33) (105449.48 -> 111759.74)
cycles:u: -6.0829% +/- 0.5338% (-1447602857.06 +/- 127036484.81) (23797839542.02 -> 22350236684.96)
instructions:u: -4.9550% +/- 0.0008% (-1364113249.40 +/- 214444.14) (27529935457.16 -> 26165822207.76)
branch-misses:u: -5.7909% +/- 2.2880% (-5395643.76 +/- 2131857.44) (93173948.12 -> 87778304.36)
seconds time elapsed: -5.6275% +/- 0.5257% (-0.32 +/- 0.03) (5.72 -> 5.40)
seconds user: -6.0582% +/- 0.5861% (-0.33 +/- 0.03) (5.47 -> 5.13)
Scavenge( new space) goes from 63 to 61
Evacuate(store buffer) goes from 2 to 3
MarkSweep( old space) goes from 1 to 0
```
As before the changes here are mostly "bogus" and likely caused by the
GC being weird, see also
https://dart-review.googlesource.com/c/sdk/+/438722 and
https://dart-review.googlesource.com/c/sdk/+/438681.
It does seem to consistently change the number of new space GCs though
(65 -> 63 -> (here) 61).
Change-Id: I18ef712394c1d98b8ed15c7fe6c4ac32fa9bc2e3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/438682
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Chloe Stefantsova <cstefantsova@google.com>
This adds `LookupResult.isInvalidLookup` to the handle invalid lookup
results and uses this to avoid a lot of cascading error messages.
This is a step towards removing ProblemBuilder, AmbiguousBuilder and
reliance on `NamedBuilder.isDuplicate` in lookups.
Change-Id: Ia9d558ce55b45567607282295dd1e54e6187f9c5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/441880
Reviewed-by: Chloe Stefantsova <cstefantsova@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Erik Ernst <eernst@google.com>
* Remove ~2.8mb of iterators in SourceMethodBuilder
* Remove ~0.7mb of iterators in calculateBounds
* Remove ~1.7 mb of iterators in ClassMembersNodeBuilder.build
* Remove ~1.9mb of iterators in BuilderFactory.computeBuildersByName
* Remove ~3mb of iterables and iterators in findRecordUseAnnotation
* Remove ~3mb of allocations from List.iterator in InferenceVisitorBase._inferInvocation
* Another ~6.3mb of iterables in various places in chunks of ~100-800kb.
Running a benchmark of the CFE compiling the CFE I get this:
```
msec task-clock:u: -5.9719% +/- 0.3845% (-342.77 +/- 22.07) (5739.61 -> 5396.84)
page-faults:u: -2.6307% +/- 0.0630% (-2951.12 +/- 70.63) (112178.18 -> 109227.06)
cycles:u: -6.1751% +/- 0.3921% (-1471343183.50 +/- 93434420.74) (23827045867.84 -> 22355702684.34)
instructions:u: -5.0116% +/- 0.0006% (-1385443477.90 +/- 176957.44) (27644543467.86 -> 26259099989.96)
branch-misses:u: -6.0357% +/- 1.7661% (-5562845.14 +/- 1627775.60) (92165665.40 -> 86602820.26)
seconds time elapsed: -5.9868% +/- 0.3830% (-0.34 +/- 0.02) (5.75 -> 5.40)
seconds user: -6.0365% +/- 0.4515% (-0.33 +/- 0.02) (5.47 -> 5.14)
seconds sys: -4.6542% +/- 3.6415% (-0.01 +/- 0.01) (0.27 -> 0.26)
Scavenge( new space) goes from 64 to 63
MarkSweep( promotion) goes from 9 to 8
Evacuate(store buffer) goes from 2 to 1
```
Worth noticing, though, is that the before numbers (e.g. 5739.61 ms
task-clock:u) is very close to the before numbers from
https://dart-review.googlesource.com/c/sdk/+/438681 (5767.31 ms) and not
the after numbers on that cl (5478.56 ms) --- so something weird is
going on here.
Change-Id: I67248bbde47900435ababcb1a2d5ecff2dde5c13
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/438722
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
This document was created during the early stages of the NNBD
development. The goal was to aid the migration to the new model. The
document has served its purpose and is not maintained for some time
now, so it can be deleted.
Change-Id: I398331cc41a0070e28cf3d1f94da958dd2d3ecfa
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/436602
Commit-Queue: Chloe Stefantsova <cstefantsova@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
This adds the library reference of private names of other libraries in
the ast-to-text of member declarations. This prepares for a fix of the
library used for private constructor tear offs.
TEST=existing
Change-Id: I65a30f6446b6e2cd15291327babb7fcacb36c4ff
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/434640
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
Under the hood, there is no difference between
`isSubtypeWhenUsingNullabilities` and
`isSubtypeWhenIgnoringNullabilities` by now. This CL makes the
corresponding renames and removals in the CFE and its clients.
TEST=existing
Change-Id: I22060c29834179c30ba62562aa254146b1d7530d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/433480
Reviewed-by: Ömer Ağacan <omersa@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Mayank Patke <fishythefish@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Johnni Winther