Re: [Gnu-arch-users] Dirname caching for leroy's tla on cygwin

[John Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Right, but you have to compute the hash by reading the whole string. It
really just depends on the changes of collisions and how the hashtree
handles it.

But actually, I found a different problem. Dependencies.... Basically, I
can't use libawk/associative.h because libawk depends on libhackerlab,
and you can have libhackerlab depend on libawk, and libawk depend on
libhackerlab.

I would like to move associative.h out of libawk and into hackerlab.
Probably create an hackerlab/assoc directory and put it there. For now,
I'm just copying the code and adding some prefixes so that hackerlab
will have a duplicate.

Now associative is really only a thin layer over hackerlab/hashtree, but
it is a _useful_ layer.

Anyway, I don't really know why associative is where it is, but to see
if it works, I'm doing the copy. I would depend on someone more "in the
know" to tell me if moving from libawk to libhackerlab is okay.

John
=:->

Aaron Bentley wrote:

| John Meinel wrote:
|
|> The reason I wonder is that hash always has to traverse the entire
|> string, while a comparison would only have to traverse until a mismatch.
|
|
| I haven't examined at this implementation, but hashtables usually use
| the hash value as an index.  As long as you don't have too many
| duplicate hash values (or near-duplicates), it's extremely fast.
|
| Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Cygwin)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFA6cG6JdeBCYSNAAMRAgmsAKC6CotvhjtMYf8wR3SmvRr1DXoljgCfctak
1kwna65HxQgLQHZ3ow61lJw=
=ath5
-----END PGP SIGNATURE-----