What's done?
A cross compiling nqp-js (running on top of parrot) passes all tests in nqp/t/nqp with the exception of test 49 (which does an equivalent of eval, which would be really tricky do in a cross compiler).
It also passes the serialization tests in nqp/t/serialization.
It also passes the same tests as nqp-parrot in nqp/t/qregex.
We use a custom harness for those to avoid compiling regex at runtime.
A standalone nqp-js (compiled with a cross compiling nqp-js and using a setting compiled by a cross compiling nqp-js) passes most of the tests in nqp/t/nqp (it fails 5 of them).
During the program I also extended the nqp test suite, implemented nqp::shell on nqp-parrot.
What's missing?
Currently for building the nqp-js I'm keeping a nqp-compiler.nqp, which is basically the most of NQP concatenated into one file. The real nqp concatenates it files into 2 gigantic files.
My nqp-compiler.nqp doesn't currently register itself using the nqp::bindcomp ops correctly so it's failing the roles tests (https://github.com/perl6/nqp/blob/master/t/nqp/56-role.t and https://github.com/perl6/nqp/blob/master/t/nqp/66-pararole.t) as that's needed for roles.
Some bugs are making test 32 fail on the standalone compiler.
Test 49 needs loading of the compiler and runtime.
And test 55 has some problems with cloning multi methods.
As the standalone compiler had trouble with roles and multi methods I didn't get to create a bootstraped nqp-js (nqp-js compiled with standalone nqp-js).
Running in the browser is mostly a matter of figuring out how to load our pure JavaScript dependencies with RequireJS (or something similiar as I haven't got around to figuring that), and figuring out how to replace our bigint dependency.
The other none pure JavaScript dependencies are for stuff browsers don't support anyway.
What's next?
The short term things I plan to do now once the GSoC program has ended are:
* Instead of keeping a nqp-compiler.nqp in the repo build it (in the Makefile) the same way as other backends do from the files in the nqp repo.
* Finish bootstraping.
* Code cleanup:
I need to seperate stuff into more files, adopt some sort of automatically verified coding standard on the js part (I have experimented with adapting the Google Style Guide before).
I also need to make things stick to a naming convention. That is decide where do use CamelCase and where to use under_scores. Translating stuff from the real nqp which is also sometime confused about that didn't help with that.
Middle term things:
* Write a nqptidy ;)
* Source maps. We need them to use fancy node debuggers I recently found exist.
* Optimalisations. Now that I figured out how most stuff works in nqp I can try to emit decent JavaScript code