> No, I don't think there are any multicharacter breaks like break2PS. It
> seems to me like a reasonable (albeit somewhat specialized) extension to
> FastPackedString.
I think I'm going to change its interface, by the way, to return
(takePS (n+1) ps, dropPS (n+1) ps)
rather than
(takePS n ps, dropPS (n+2) ps)
The former turns is more general (the latter can be obtained from the
former with initPS and tailPS), and in particular can be used for
parsing the committer and author lines without breaking sharing.
> I imagine your implementation could be optimized,
Any hints welcome. ``darcs changes'' (no -s) spends 20% of the time
it takes in that function.
Juliusz
|
Try Searching:
servers, voip, java, networking, microsoft ...
|
|
|
| 
