This the second part of fixing #32077 - dealing with mints.
Example:
The runtime of a program adding 100,000 different mints to
a list and prints the last one:
Before: ~12.5 seconds
Now: ~1.2 seconds.
Change-Id: Iad6da1749886bab4e4d3941748cdbab51661c2c3
Reviewed-on: https://dart-review.googlesource.com/41161
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
This the first part of fixing #32077 - dealing with doubles.
Example:
The runtime of a program adding 100,000 different doubles to
a list and prints the last one:
Before: ~12 seconds
Now: ~1.3 seconds.
Change-Id: I7f64b8763fdd37dae5d32d3275f8b9c991d4e43a
Reviewed-on: https://dart-review.googlesource.com/40300
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Currently front-end embeds absolute paths to build time locations into
Kernel binaries. Reporting those to user make stack trace unreadable and
also breaks stack trace filters that attempt to filter out implementation
frames from dart:async:
#1 _RootZone.runUnary (file:///Users/vegorov/src/dart/sdk/sdk/lib/async/zone.dart:1381:54)
same frame in legacy mode looks like this:
#1 _RootZone.runUnary (dart:async/zone.dart:1381)
This CL changes stack frame formatting so that even in Dart 2 mode result looks
closer to legacy mode result.
This is a workaround for https://github.com/dart-lang/sdk/issues/32087
Change-Id: Ic69cb039ff7d40d0e8e5340a9c1dc067427479ea
Reviewed-on: https://dart-review.googlesource.com/40522
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
The original change is available in patchset 1.
The changes since then are just status file updates.
Change-Id: I0c45c4b09a187bdd60c6a38bf23f6b7a0cf80c97
Reviewed-on: https://dart-review.googlesource.com/34400
Reviewed-by: Jens Johansen <jensj@google.com>
E.g.
```
main() {
new Foo();
}
class Foo {
Foo() {
print("hello from foo");
}
}
```
debugging with --pause_isolates_on_start, the constructor (or anything else
in class Foo) would not be breakable from the Observatory interface.
This CL fixes that by collecting token positions from unfinalized classes.
Bug:
Change-Id: I93a20f3322b0a343a619625c71236b59c885cb7e
Reviewed-on: https://dart-review.googlesource.com/30441
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Kevin Millikin <kmillikin@google.com>
The optimizer uses the "is more specific than" relation to eliminate type tests.
In strong mode, the "is more specific than" relation is supposed to be identical
to the "is subtype of" relation. However, it is not quite true, and this is the
cause of this bug. In non-strong mode (Dart 1), the "is more specific than"
relation uses covariance for both result type and parameter types of functions.
In strong mode, we need covariance of result type and contravariance of
parameter types.
Added regression test and updated status files.
Change-Id: I6315643890db75213af6720aa99d8da7b15f99d6
Bug:
Reviewed-on: https://dart-review.googlesource.com/29900
Reviewed-by: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Régis Crelier <regis@google.com>
The info_array is needed to visit an ObjectPool's pointers, requiring the compactor to move the info_array's body before forwarding the ObjectPool's pointers. Moving the info_array inline remove this constraint on the compactor.
Also saves 3 words per ObjectPool modulo allocation size rounding.
Bug: https://github.com/dart-lang/sdk/issues/30978
Change-Id: I94de0e4b7356d46fb145efee7ab14abd7473eb4c
Reviewed-on: https://dart-review.googlesource.com/27480
Reviewed-by: Erik Corry <erikcorry@google.com>
is of class Future, when test type is FutureOr.
This was missing only when testing an instance, not when testing a type or a
parameterized class.
Change-Id: I1609f6f7df80f15a8f36d18bf504df444dfbb928
Reviewed-on: https://dart-review.googlesource.com/27083
Reviewed-by: Régis Crelier <regis@google.com>
This reverts commit cd4e5ee866.
Reason for revert: This broke the 32bit Linux SDK build.
Original change's description:
> Keep the line starts delta encoded in the VM heap as well.
>
> The line starts in the kernel binary are delta encoded. By decoding and
> storing them as int32 values, we are unneccesarily bloating the VM heap.
> This change keeps the line starts delta encoded in the VM heap as well.
> This brings down the core snapshot size by 275KB (but will of course
> affect the debug performance.)
>
> Bug:
> Change-Id: I77769efe7339484096d04c37180367adf96e51f4
> Reviewed-on: https://dart-review.googlesource.com/22304
> Commit-Queue: Siva Chandra <sivachandra@google.com>
> Reviewed-by: Siva Annamalai <asiva@google.com>
TBR=sivachandra@google.com,asiva@google.com
Change-Id: I6e78bd510de7033d8583c089c00b5b261fdf69c1
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/24800
Reviewed-by: Alexander Thomas <athom@google.com>
Commit-Queue: Alexander Thomas <athom@google.com>
The line starts in the kernel binary are delta encoded. By decoding and
storing them as int32 values, we are unneccesarily bloating the VM heap.
This change keeps the line starts delta encoded in the VM heap as well.
This brings down the core snapshot size by 275KB (but will of course
affect the debug performance.)
Bug:
Change-Id: I77769efe7339484096d04c37180367adf96e51f4
Reviewed-on: https://dart-review.googlesource.com/22304
Commit-Queue: Siva Chandra <sivachandra@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
- Remove random build-id.
- Replace build time in embedded version string with commit time.
- Remove timestamps from Observatory tarball.
- Zero-initialize skipped bytes in snapshot streams.
- Fix uninitialized fields in PatchClass, Script and Library.
- Disable (under flag) random identity hashes and concurrent GC.
Bug: https://github.com/dart-lang/sdk/issues/31427
Change-Id: I3e95de679c8372841cd27ca60df78d9b00ffbfe1
Reviewed-on: https://dart-review.googlesource.com/22901
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Zach Anderson <zra@google.com>
The introduced "constants" transformation can evaluate constant expressions. The
original use-sites of constant expressions are replaced by a new [ConstantExpression]
node, which points to a subclass of a new [Constant] class hierarchy. Constant
[Field]s and [VariableDeclarations]s will be removed, since all use-sites are
re-written.
The [Constant] class hierarchy is, similarly to the [DartType] class hierarchy, not
part of the AST tree (also has no parent pointer). The constants form a
DAG (directed acyclic graph).
There is no canonicalization requirement of the [Constant] objects referenced by the
AST (via [ConstantExpression]). Although it is beneficial to canonicalize them during
construction, since it reduces time spent in operator==/hashCode.
This CL furthermore adds support for a constant table in the binary format. Similarly
to [String]s, we canonicalize the constants before writing the table to the binary.
The constant table entries in the binary are written in a post-order way, to ensure
easy construction on the backend side.
The text format will be augmented with a "constants { ... }" section at the end,
which lists the constants in the same order as in the binary format.
The transformation can be used by those backends who choose to do so. It is not
enabled by default atm. It should therefore not affect analyzer, fasta or other
components.
Change-Id: I57cd9624fedcf537ab6870db76246149647bed21
Reviewed-on: https://dart-review.googlesource.com/14382
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Kevin Millikin <kmillikin@google.com>
It was previously an array of Smi values. By converting to TypedData
array of int32_t values, we shave off 70+KB from the kernel based core
snapshots. While 70KB might not seem too high, it is 5% of the excess
size we have wrt to the token stream based core snapshots.
Change-Id: I5b58a1d1ac27391eca1efff179301bb04162e97b
Reviewed-on: https://dart-review.googlesource.com/19800
Commit-Queue: Siva Chandra <sivachandra@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Coverage uses the value of an execution counter to detect executed functions. However non-positive execution counter does not necessary mean that a function was never executed because we sometimes reset the counter.
Introduce additional bit on a function that is set whenever we reset positive execution counter and use this bit when generating coverage data.
Bug: https://github.com/dart-lang/sdk/issues/31326
Change-Id: I5357109603defad7e7c9415c433203f16bcf88f4
Reviewed-on: https://dart-review.googlesource.com/19760
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This change brings down core snapshot size by ~750KB, and brings
down app-jit snapshot size of simple "Hello, World" dart script by
~650KB. The bot cycle times will also come down by around ~20%.
Change-Id: I2a01c98bedc7ebfa2a653983995486a71504daf3
Reviewed-on: https://dart-review.googlesource.com/16323
Commit-Queue: Siva Chandra <sivachandra@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Jens Johansen <jensj@google.com>
Remove parsing support for 'super' that was meant to declare lower bounds of
function type parameters, since this feature did not make it into the spec.
Change-Id: I4087d0e108c5984437083a05a065fe785888d2f7
Reviewed-on: https://dart-review.googlesource.com/17360
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
This allows the normal VM metadata lookup code to find metadata attached
to members which come from mixed-in classes.
This allows e.g. putting a @AlwaysInline / @NeverInline annotation on
methods of mixin classes.
Change-Id: I70639dc78fc5bab98ac7cee107c2042f84d07888
Reviewed-on: https://dart-review.googlesource.com/17164
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
Currently, we have some tests that run with reified generic functions ON and
some with it OFF. However, this causes inconsistencies when running the Kernel
FE from snapshot in the VM, since the VM requires that the flag be consistent
with the snapshot. Our solution is to turn the reified generic functions flag
into a per-isolate flag, which is always disabled in the Kernel FE's isolate.
Bug:
Change-Id: Ia91e9f0ff5fc059edf4bdbd79ef14abb288b4a49
Reviewed-on: https://dart-review.googlesource.com/16020
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Includes the following changes:
- Disable type checking in the service isolate when strong mode is
turned on. We don't yet have strong mode clean vmservice Kernel
binary available (Issue #31203);
- Introduce new Dart VM API to allow creating List<T> (T in {String, int}).
Current API only allows to create List<dynamic> and in strong mode
List<dynamic> is not assignable to List<T>. For now we limit this API to
a fixed number of core types for performance considerations (type
arguments can be canonicalized and cache ahead of time). We will
extend the API allowing creation of arbitrary List<T> if we will
discover the need for it later;
- Use this new API to fix CreateRuntimeOptions to create List<String>
and not List<dynamic>;
- Likewise fix OneByteString_splitWithCharCode to return List<String>
and not List<dynamic>;
Additionally this CL refactors how getters and setters are created
in object_store.h removing 500 lines of largely boilerplate code.
Bug: https://github.com/dart-lang/sdk/issues/31052
Change-Id: I3fdcd2d54ff3f307db0913c67a7c2f009ea9d7a7
Reviewed-on: https://dart-review.googlesource.com/15884
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
For 1-byte strings we don't need the full power of UTF-8 conversion with
support for 3 and 4-byte sequences and reassembly of UTF-16 surrogate pairs.
This optimization was originally intended to improve the reverse-complement
benchmark. It also seems to have a positive effect (ca. 1% on average) on
dart2js.
After implementing this it doesn't seem worth having a special case in
ToCString for ASCII strings. That special case would allocate the string twice
for all Latin1 strings, which feels unfortunate.
R=vegorov@google.com
Bug:
Change-Id: If852f3ecb4e06118da8357cfa2c843ad0df0edb9
Reviewed-on: https://dart-review.googlesource.com/8920
Commit-Queue: Erik Corry <erikcorry@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Régis Crelier