Week-end work notes

[Louis Opter]

David gave me a hand this week-end to solve a blocker in the compiler;
but it's causing some regressions I'm writing down here so we don't get
lost when we resume our work.

The blocker was that the type of a chunk parameter wasn't correctly
resolved during a sequence call. Fixing it [1] caused two regressions:

1. As a side-effect it triggered a weird reference reuse which would
yield to a corrupted resval;
2. Somehow, during a Socket::SKBuff::init method call in e1000.blt, it
messes up how parameters are searched across the chunk arguments and
the chunk local variables. The compiler then complains that local is
missing in front of a chunk argument (which is wrong), and when you
add it, it complains that local is used in front of a chunk argument
(which is correct). Either way, it doesn't work/generate.

Hopefully, we'll address that this week and/or next week-end!

Right now everything is in my patch queue, point 1 is addressed in
"rathaxes_compiler_passes_fix_reference_reuse.patch" and point 2 is
being addressed in "wip_fix_regression_in_rathaxes_compiler_passes_correctly_resolve_chunk_args_subidentifiers.patch".

[1] https://www.kalessin.fr/projs/hg/epitech/mq/rathaxes/file/ecba077e0f73/rathaxes_compiler_passes_correctly_resolve_chunk_args_subidentifiers.patch

--
Louis Opter