GSoC 3: My God, It's Full of Goto

This has been a bit of an eventful week outside of GSoC. I had another housemate move in, so I'm surrounded by boxes and keep getting asked to move this or fix something around the house. But progress has been made:

June 6 PAST::Node in NQP: Conversion of PAST::Node and it's seven subclasses into full NQP code.

The last of the PIR was ejected from the PAST::Node classes on the 2nd. I also discovered make codetest and tweaked both my code and the tests to ensure everything passed. (The tests had to be passed so that generated PIR didn't get tested and similar.)

I've also gotten my commit bit and used it to push my work so far into the main parrot repo.

June 13 PAST::Compiler in NQP: Convert PAST::Compiler to NQP.

This has been far more effort than the previous conversions. There's lots of interesting usage of goto in several of the methods which makes figuring out what's happening very difficult. This has, however, convinced me that this step is required. A lot of the logic of conversion is buried in here and it uses lots of little snippets of PIR in its output. This will all have to be altered to make the PIR-less future of PCT happen.

Next week I'm going to have to investigate all these useful new POST nodes that bacek invented and start figuring out how it all gets put together. But for right I'm wrestling with the minor differences between the way PIR and NQP does things...