Used to separate the behaviour without the variance experiment flag
turned off and with the flag turned on. Allows for more specific error
descriptions and allows the backend to avoid emitting variance
annotations for legacy covariance and only when the experiment is
enabled, otherwise there is no way of knowing whether the variance is
user specified.
Change-Id: Ia3070a1358e9b0fef10a3ce5057df881b9659832
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/121140
Commit-Queue: Kallen Tu <kallentu@google.com>
Reviewed-by: Dmitry Stefantsov <dmitryas@google.com>
Reviewed-by: Leaf Petersen <leafp@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Add an incremental serializer that can be used to avoid re-serializing
the same thing again and again.
It does this by grouping libraries into Components / bundles,
serializing them individually and using the concatenated dill feature to
output the wanted data --- just potentially faster if we had the data
in cache.
Note that the serialized output might contain *more* than the input given
if we cached the wanted data into a bigger bundle.
The output will always be "closed" though, i.e. if the stuff that is
included that is too much added a new dependency, that dependency will be
included as well.
This should generally make it safe, although one can imagine situations
where it could pull in lots of dependencies that it wouldn't otherwise
have.
It is being driven by the incremental compiler which makes sure to
invalidate the cache when the data changes.
Except for situations where the libraries are changed externally after
being serialized, but before being serialized again, where one then wants
the updated library serialized, the feature should be safe.
Change-Id: I2a504abe6dbb68434c3b04abff13480ef72a6a6e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120786
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
Prior to this change, loading the sdk dill twice for instance would
result in additional exports as well as parameter counts doubling,
essentially destroying the result.
Change-Id: I4d694bec1926d6df6572e9bb36e8a88ec0822e27
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/120785
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
The interpretation of the subtype check result depends on the mode. In partial
NNBD mode all type errors that wouldn't be errors in non-NNBD programs should
become warnings. In full NNBD mode all such errors are errors. This CL adds
the mode explicitly to the interface of the subtype check as a parameter.
Bug: http://dartbug.com/38673
Change-Id: I14bcdd260618530cfdd8189c01abba7469a16679
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119545
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Dmitry Stefantsov <dmitryas@google.com>
In my measurements this changes the time it takes to write the sdk by
-11.1274% +/- 0.624366% (via slightly modified
pkg/kernel/test/binary_bench.dart AstToBinary out/ReleaseX64/vm_platform_strong.dill)
Change-Id: I4c5cbe17f42b70ef361b2edda0b33c8fc1727b63
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/119540
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
Returns types on generic methods in generic extension that contained
type variables declared on the generic method were not substituted,
leaving these as unindexed type parameters during serialization.
Change-Id: I4f53005318ba2be014bfe6e92f4f6c93288f2a4f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/118995
Commit-Queue: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Dmitry Stefantsov <dmitryas@google.com>
Pass variance data to field in TypeParameter through the listeners.
ast_to_text will print variances for classes and mixins if specified.
Avoids serialization/deserialization (impl in future CL).
Change-Id: I298537604823710f0d30001f4cb5f1e81530959f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/118464
Commit-Queue: Kallen Tu <kallentu@google.com>
Reviewed-by: Dmitry Stefantsov <dmitryas@google.com>
This refactoring achieves the following:
- Collect the integer operation semantics into one place instead of
having it sprinkled all over the constant evaluator with many
"if (targetingJavaScript)" tests.
- Avoid emitting the internal JavaScriptIntConstant node, which is
serialized as normal DoubleConstant nodes, thus achieving parity
between direct consumers and consumers seeing output that has been
serialized.
- Implement the JavaScript semantics for unsigned right shift.
- Fix truncating divide with a result outside int64 range incorrectly
clamping the result for JS targets.
- Fix positive hex constants >= 2^63 through int.fromEnvironment
producing negative values in dart2js.
- Clarify in Kernel documentation how numeric constants are represented
on VM vs JS targets.
Change-Id: If30bb2c2c77c54eff120b611b059c2ec726c99a0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/116525
Reviewed-by: Mayank Patke <fishythefish@google.com>
We have two layers of laziness.
Lazy loading functions (the old one) disables itself when loading a
concatenated dill file.
Lazy loading class content (the new one) forgot to do that while not
being geared towards it either.
This means that it - when loading lazily - can look into the wrong
string table, the wrong link table etc.
This CL brings the laziness for classes on par with the laziness for
functions in that it is now disabled when loading a concatenated dill
file.
Change-Id: If3449d5aff34287221a14fb7e940779d93b1e7b3
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/117145
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
Removes StatementJudgment, ExpressionJudgment, and InitializerJudgment
classes that are identical to existing nodes.
Change-Id: If67c1972387bb158bfe775cdf091cae4eadd47c9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/116061
Reviewed-by: Dmitry Stefantsov <dmitryas@google.com>
This CL gives the option to load classes lazily (on by default),
meaning that fields, procedures, constructors and redirecting
factory constructors of classes are not initially read,
but instead read when (or if) needed.
The idea being that many classes aren’t needed, and that
spending time on reading them is thus just a waste of time.
This is especially true in modular compilation where a lot
of modules needs to be given as input as the compilation at
hand might technically rely on it, but in practice only very
few of the classes are actually needed (or very few classes
actually need their members).
The below numbers are comparing a VM without this change and without
the lazy class hierarchy member lists change with
b31566b297 (the base change for those
two changes).
For running a simple hello-world script, on my machine 500 times
and doing statistics on it I can measure the following differences:
With verbose, from kernel_service startup to compile complete:
Difference at 95.0% confidence
-17.8 +/- 0.555817
-24.3455% +/- 0.760206%
(Student's t, pooled s = 4.48379)
With verbose, total runtime measure by `time`
Difference at 95.0% confidence
-19.346 +/- 0.905217
-12.0333% +/- 0.563051%
(Student's t, pooled s = 7.30242)
Without verbose, total runtime measure by `time`
Difference at 95.0% confidence
-17.862 +/- 0.842905
-12.0635% +/- 0.569276%
(Student's t, pooled s = 6.79974)
For running a number of modular compilations I’ve recorded an
improvement in actual runtime in the order of 20-37% depending
on the circumstance. In a real build system, though, the
difference seems to be more in the order of around 5%.
Change-Id: Id329bcf2b01d12c12d7a49f2b8abacd9c2447f05
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/115703
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Kallen Tu