freenode/#sicl - IRC Chatlog
Search
17:49:30
beach
Bike: Is there any reason why we wait and introduce the INITIALIZE-CLOSURE-INSTRUCTION (that, as I have said, I would rather call SET-STATIC-ENVIRONMENT-INSTRUCTION) only when we segregate lexicals, as opposed to immediately in AST-to-HIR?
17:50:37
Colleen
Bike: kpoeck said 7 hours, 36 minutes ago: Do you see any problems with extending ttps://github.com/clasp-developers/clasp/blob/dev/src/lisp/kernel/cleavir/translate.lisp#L850? This is to compile https://github.com/3b/3bgl-shader/blob/master/example.lisp#L187 with cst
17:50:37
Colleen
Bike: kpoeck said 7 hours, 35 minutes ago: : Meant extending *additional-clasp-character-names* to accept "ESC" in addition to "ESCAPE"
17:50:50
Bike
without segregate-lexicals we don't know the set of variables closed over, and we have no cells.
17:52:26
beach
In segregate-lexicals, we would just add cells and shared variables as inputs to the new instruction.
17:56:36
Bike
i think we might need to have the initialize instruction placed not immediately after the enclose, like for mutually recursive functions. that might be difficult if we make it earlier.
17:57:50
beach
I thought we added instructions, like CREATE-CELL and such, and we just made sure they were placed between the two.
18:05:18
beach
For things like LABELS, all instructions for setting the static environment must come after the last ENCLOSE.
18:11:34
beach
... unless we place them right initially. But that may require doing it at the AST level.
18:19:05
beach
Anyway, I am off to spend time with my (admittedly small) family. I'll be back tomorrow morning (UTC+2).