The title of this integration is misleading and Clement Bera clarified
it in the Pull Request comments:
Here is a copy of his comment:
«Just to clarify. This slice makes all tests green which means the bytecode
set is ready for integration, it does not integrate it. Integration of the
new bytecode set is now difficult in Pharo 7 with the Bootstrap followed by
the code loaded with Hermes that should be modified. I wish I had
integrated it in Pharo 6 (at this point it was just a single slice with a
single DoIt) so I could have done it very quickly but it's not the case.
The main advantages of the new bytecode set over the existing one are:
- Increased size of bytecode instructions limitation (jump size, number of
inst var, number of literals, etc.)
- Support for Lowcode/Sista inlined primitives
- Support for FullBlockClosure
- Easier to decode (instructions sorted by number of bytes, simpler closure
instruction, etc.)
- Bytecode instructions to push immediate objects instead of wasting
bytes/literal slots
- No more DoubleExtendedDoAnything bytecode crap»
On Fri, Nov 17, 2017 at 12:35 PM, <
[hidden email]> wrote:
> There is a new Pharo build available!
>
> The status of the build #299 was: SUCCESS.
>
> The Pull Request #484 was integrated: "20603-Integration-of-new-bytecode-set-Sista-V1-by-default-green-tests"
> Pull request url:
https://github.com/pharo-project/pharo/pull/484>
> Issue Url:
https://pharo.fogbugz.com/f/cases/20603> Build Url:
https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%20and%20branch%20Pipeline/job/development/299/--
Cyril Ferlicot
https://ferlicot.frhttp://www.synectique.eu2 rue Jacques Prévert 01,
59650 Villeneuve d'ascq France
Locked
