26.08.2013, 00:41, "John Regehr" <[email protected]>:
>>  Btw, I had cases where pass_lines::0 and pass_lines::1 where running for
>>  a long time with almost no progress, while pass_lines::2 removed more than
>>  50% quickly after that.
>
> We've seen this too!  There might be some clever way to figure out which
> line pass to run, I'm not sure...

Maybe just run all pass_lines with granularity limit, e.g. stop at 5 lines 
chunk?

-- 
Regards,
Konstantin

Reply via email to