To see advantage of running decoding on separate isolate this benchmark decodes 10 json strings.
Sync implementation has to do it sequentially, async version does in parallel.
Change-Id: Ia4e25f532d0a92aea9b16b99e374b898899a11cd
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123100
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Also multiplex the service rather than the logger, so that we can
multiplex a crash reporting instrumentation service with a file logging
instrumentation service.
Change-Id: I7720c55e26225c27d2d9447a0156b2ca83da6598
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/122100
Commit-Queue: Mike Fairhurst <mfairhurst@google.com>
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
This is the follow-up for https://dart-review.googlesource.com/c/sdk/+/123920.
RAW_CAST macro in the interpreter doesn't properly handle objects of
subclasses of a tested class (e.g. RAW_CAST(String, x) fails in debug mode
if x is OneByteString). So this CL replaces a couple of recently added
RAW_CAST with a more reliable RawCast.
Change-Id: Ifa94ce79985a5ae60c6f4d86ff7552b5124f649c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/124102
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
In moving to a pointer mixture network from a traditional language
model I had initially decreased model lookback to 50 tokens to
facilitate faster model training iteration. Now that we've landed
on a configuration that works well, increasing lookback to 100 tokens
improves accuracy further.
I also went ahead and followed a recommendation from Jacob to
oversample Flutter sources since they're greatly underrrepresented
in our training corpus compared to our users' code at prediction time.
Change-Id: I885ddb606270394198f3aca2554e381f4465bdae
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/124083
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Commit-Queue: Ari Aye <ariaye@google.com>
This CL
* adds a GN flag "include_dart2native" off by-default (switched on explicitly for our GN build)
* ensures every GN build of the "create_sdk" target has a working dart2native (except ia32)
* removes not-working dart2native script from ia32 dart-sdk (no AOT support in ia32)
* removes the "dartaotruntime" GN target, since it is the same as "dart_precompiled_runtime"
This should ensure every downstream consumer of the "create_sdk" GN target will
have a working version of dart2native.
For the particular "dart-sdk" we create on dart-sdk-{linux,windows,mac}
we will override the default AOT compiler&runtime with the product mode,
which produces smaller AOT snapshots and has smaller AOT runtime (making
the native executables significantly smaller)
Change-Id: Ib5042589297cdb055c41dc56abdff86a4612ef5c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123724
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Michael Thomsen <mit@google.com>
This prolongs the make_version.py hack for Flutter for the 2.7.0 Dart
release.
Change-Id: Ic8324461b195be65a7c3afd3bb34c7f186b29657
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/122501
Reviewed-by: Siva Annamalai <asiva@google.com>
Auto-Submit: Alexander Thomas <athom@google.com>
Commit-Queue: Alexander Thomas <athom@google.com>
We're deprecating kernels Library.isExternal as it wasn't used for
what it was originally intended. It will hopefully go away entirely
soon.
Change-Id: If363c50af5607febae68865875af452c106fff85
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123721
Commit-Queue: Jens Johansen <jensj@google.com>
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Bytecode format v20 was introduced Sep 06 2019 in
https://dart-review.googlesource.com/c/sdk/+/116120.
This change drops support for older bytecode format versions from VM.
Change-Id: I11b75ba16e6b7570e0346f9ac024265b08c0c801
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123920
Reviewed-by: Régis Crelier <regis@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
VM creates Script objects with kKernelTag only, so this CL cleans up
all uses of script tags and Script::kind() along with code
which is no longer reachable.
Change-Id: Ia765e7757264aa614e18eddd16d21937f855c129
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123889
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
We no longer generate kernel files containing both AST and bytecode, so we don't need a flag to choose between them.
Change-Id: I1a4f7df507c649019c9fe254fa18a5826e2006aa
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/122402
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
This also makes the annotation tracker an instance field rather than a
static field.
Change-Id: I3f636047b7cbddc98536cd10b209e22546d6019b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123694
Commit-Queue: Brian Wilkerson <brianwilkerson@google.com>
Reviewed-by: Paul Berry <paulberry@google.com>
We're going to move this file into _fe_analyzer_shared, so we can't
have it importing CompilerContext. It turns out that we were only
using it to cache the decision of whether to display colors, and to
decide whether to verbosely explain how we made that decision. Those
can be accomplished in other ways.
Change-Id: If1e3a9fee7c0ca919444d8429936045b8408c246
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/123840
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Commit-Queue: Paul Berry <paulberry@google.com>
