[<<][staapl][>>][..]
Tue Dec 18 16:21:16 CET 2007

benchmarking

the current reader is problematic.. it's slow, and i don't understand
the reason. i don't think it's usage of streams, since it wat slow
before, and it's not read-char, since i tried that..

so...

  1. make a test for the current reader
  2. replace it with a new reader
  3. build 'read-syntax'


first text: the problem seems to be somewhere else..

  (define f (forth-load-in-path "monitor.f" '("prj/CATkit" "pic18" )))

is virtually instantaneous like it should be..

so where did i get the idea that this is slow?

indeed:

  '(file monitor) prjfile prj-path forth-load-in-path

is instantaneous also.

otoh, 'forth->code/macro' isn't instantaneous at all..

compiling the code 'code/macro!' is instantaneous also. i think i got
it. why is the code/macro splitter so slow?

tracking down to forth.ss : forth->macro.code which uses
@forth->macro/code which uses @moses

it can't be @moses since that's just a filter.. so it's probably down
the stream in the macro processor. need to test that separately.

running into some inconsistencies.. probably best to switch everything
to syntax objects, including a syntax-reader.



[Reply][About]
[<<][staapl][>>][..]