There seems to have been quite a bit of a confusion on this topic. At
some point you might have heard that Flutter used an "interpreter" on
iOS, it didn't. As you probably also know we do have simulators for
the non-Intel instruction sets supported by the VM (arm, arm64 and
mips). This way we can develop and debug all of our JIT
implementations on our desktop machines.
In the initial bringup of Flutter on iOS we were running with the JIT
enabled. As we all know, on iOS we cannot execute the dynamically
generated instructions, instead we reused the existing simulator in
the VM. Some might have referred to this mode as "interpreted". We
refer to this mode as simulated. In contrast a real interpreter would
run bytecodes specifically designed for Dart execution and not ARM
instructions.
Since we now do have a precompilation step, the simulated execution on
iOS is not necessary any longer.
I hope this makes things a bit clearer.
Thanks,
-Ivan
Post by Filipe MorgadoFlutter doesn't use Fletch, not even on iOS.
I guess it just links parts of the Dart VM within the executable (gc, event-queue, APIs, snapshot loader?).
It's probably very early, but I'm curious regarding the performance characteristics of AoT Dart.
I suspect the gains could be significant if the compiler cares about types.
But I guess it would meen a whole new specification is required.
--
For other discussions, see https://groups.google.com/a/dartlang.org/
For HOWTO questions, visit http://stackoverflow.com/tags/dart
To file a bug report or feature request, go to http://www.dartbug.com/new
--
For other discussions, see https://groups.google.com/a/dartlang.org/
For HOWTO questions, visit http://stackoverflow.com/tags/dart
To file a bug report or feature request, go to http://www.dartbug.com/new
To unsubscribe from this group and stop receiving emails from it, send an email to misc+***@dartlang.org.