12.07.2013, 19:59, "John Regehr" <[email protected]>:
>>   From my subjective observations, rm-toks-1 works fast, probably faster 
>> than pass_peep::b
>>  Shouldn't it be moved to the list of "normal" passes, and pass_peep::b to 
>> "slow"?
>
> Yes, it also occurred to me that we are close to being able to delete
> peep::b, or move it out of the default set of passes.
>
> Here are a few questions:
>
> - does rm-toks-1 suffice?  doesn't peep::b sometimes remove pairs or
> groups of three tokens?

But do we really need to remove three-token groups in main loop? As for me,
I always re-run creduce with --slow after it finishes first time.

>
> - isn't rm-toks-1 too stupid to be fast?  it tries to remove every
> single token including many that obviously cannot be removed without
> creating a useless test case

Maybe it's stupid, but it does not take so much CPU time to make decisions.

Btw, it might be possible to fix pass_peep::b by using possessive regexes,
but since I don't completely understand what ids it supposed it do, I haven't
tried to fix it.

-- 
Regards,
Konstantin

Reply via email to