Re: JIT stealth font-lock tuning parameters

[Stefan Monnier]

> Is there a parameter which controls how much work
> JIT-lock does, once it starts working?

Yes: jit-lock-chunk-size.

> Maybe that parameter should be reduced on a slow machine.

I'm not sure it would help.
I think the long pauses are not due to chunks being too large
but rather to the fact that font-lock sometimes needs to do
some preparatory work that not proportional to chunk.  For example,
it might need to do (parse-partial-sexp (point-min) (point)).

The maximum time spent in `parse-partial-sexp' should be reduced
by `syntax-ppss' so at least that part of the problem should be improved.
Or will be improved since the current code still uses both the old
"slow" code and the new "fast" code.
I haven't gotten much (if any) feedback about the elp-profiling of
those two options, so it seems that the few people who looked at it
found syntax-ppss to be faster (I asked to be told when syntax-ppss
is slower).  I'll hence soon remove the "old slow" code.


        Stefan