GSoC weekly raport - my rakudo breaking adventure

I tried to mimick the way subroutines are wrapped on rakudo and make test passed on parrot I commited the code.
And it broke nqp on the jvm. And I realised that rakudo on the jvm uses the HEAD revision on nqp and not a version specified in tools/build/NQP_REVISION. And after my commit make test broke on nqp.
I reverted the commit, and Jonathan Worthington figured out that I misunderstood what p6capture_lex did.
NQP opcodes propably need some kind of docs. I encoutered a crazy one called locallifetime which is some optimalisation thing I can ignore ( Which one of the primary rakudo/nqp authors wasn't sure what I does, just explaining it's some kind of ignorable optimalisation opcode.
(As a side note p6capture_lex sets the outer of code ref stored in $!do of it's arguement) to the current frame of lexicals.)
I implemented a version of that creatively called capture_lex for nqp (which was slightly more generic way as it didn't assume it's argument was a Perl6 Code).
I implemented the capture_lex opcode on parrot and nqp, and once make test passed, I happlily commited it (
And I broke rakudo. My commits where reverted and put on a branch. I realised that the sensible things was to just work on the more complex role.t test rather then spend a ton of time on hairy nqp refactors.
I'm currently doing that.

One quick solution I'm using to make the generated JavaScript easier to read (which is a large part of debugging) is to emit vim folds (


/* deserialization code {{{*/
A couple of dense screens mostly consisting of a large sha1 encoded string literal
/* }}}*/

/* main {{{*/
/* {{{ */
Things can be arbitrarly nested to provide increasing levels of increasingly ugly details.
/* }}} */
/* }}} */

As I no longer have an distracting exams at university work on my project should be now picking up a faster pace.