Hi, Björn!
> > Correct. I'm fixing it.
>
> You're ever so cryptic pavel...
Sorry, of course I meant that I only fixed the comment.
> Anyways, please have a look at the
> attached (inline) patch:
It's great! I really appreciate it!
> + if (list->list[i].f.marked) {
> +g_hash_
On Thu, Dec 06, 2001 at 09:47:33PM -0500, Pavel Roskin wrote:
> > A hash-table would make it 'less big O', scale and more predictable.
> > <...looking at code...> But perhaps not a binary search (see dir.c:658)
> > since the list isn't sorted?
>
> Correct. I'm fixing it.
You're ever so crypti
Hi, Björn!
> A hash-table would make it 'less big O', scale and more predictable.
> <...looking at code...> But perhaps not a binary search (see dir.c:658)
> since the list isn't sorted?
Correct. I'm fixing it.
> I guess I wasn't totally clear, I meant to say that there's no link from
> http
Greetings Pavel,
On Mon, Dec 03, 2001 at 04:44:34PM -0500, Pavel Roskin wrote:
> On Thu, 29 Nov 2001, [iso-8859-1] Bj?rn Eriksson wrote:
> > This small patch tries to skip the check-if-file-was-marked-in-the-
> > old-list-logic if no files we're marked at all. It considered
> > decrementing 'mar
Hi, Björn!
On Thu, 29 Nov 2001, [iso-8859-1] Björn Eriksson wrote:
> I had some problems with /tmp/ksocket/* when it contained ~20'000
> files, all of the same length thus foiling the strlen() 'optimisation'.
Indeed, the time spent in reload_dir() grows as n^2 where n is the number
of entries
I had some problems with /tmp/ksocket/* when it contained ~20'000
files, all of the same length thus foiling the strlen() 'optimisation'.
This small patch tries to skip the check-if-file-was-marked-in-the-
old-list-logic if no files we're marked at all. It considered
decrementing 'marked_cnt' f
