These are chat archives for dropbox/pyston

5th
Jan 2015
Bob Fang
@dorafmon
Jan 05 2015 12:20
Hi, this might be a dumb question
but for the inline cache the assembly is generated by hand written assembler right, but for the jitted code we use llvm’s assembler right?
Why not use the same assembler?
Kevin Modzelewski
@kmod
Jan 05 2015 21:10
Well, they have slightly different priorities
I think it would be quite a bit of work to get LLVM's assembler working in this context
since it's not really the kind of use case that it's designed for (jitting lots of small sections each with potentially a different ABI)
It probably could be done though and we should probably look into it more seriously at some point
Bob Fang
@dorafmon
Jan 05 2015 21:11
cool thanks that make sense