On Sun, Jan 27, 2008 at 11:45:25AM +0200, Marko Kreen wrote:
The minus:
- some memory i/o overhead which may or may not matter
If a string is short, it will probably reside in the processor cache, so
there is no real memory i/o overhead here. For more longer strings, it
may be better to do that in short chunks, so each chunk can reside in
the cache. But I don't think filenames are long, so it is not an issue
here.
- the parts would not be fully generic, but special to hashing
I think the second part can be rather generic to be reused for hashing
something else that does not require filename specific case-folding.
Dmitry