This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH] Use Quicksearch in strstr
- From: Zack Weinberg <zackw at panix dot com>
- To: Ondřej Bílka <neleai at seznam dot cz>
- Cc: Wilco Dijkstra <Wilco dot Dijkstra at arm dot com>, GNU C Library <libc-alpha at sourceware dot org>, nd <nd at arm dot com>
- Date: Mon, 29 Oct 2018 19:24:05 -0400
- Subject: Re: [PATCH] Use Quicksearch in strstr
- References: <DB5PR08MB1030E559B3818F52CA48A63583F30@DB5PR08MB1030.eurprd08.prod.outlook.com> <20181029225421.GA21482@domone>
On Mon, Oct 29, 2018 at 6:54 PM Ondřej Bílka <email@example.com> wrote:
> Real benefit is fixing idiocy
> That "benchmark" is complete garbage
When you dismiss other people's testing this aggressively you make
them not want to cooperate with you. If I had written this patch, my
reaction to this feedback would be: Fine, fix it yourself if you think
my approach is idiotic.
It is perfectly possible to express concerns in a less hostile manner.
For instance, you could have said "I think this benchmark doesn't
account for the cost of branch misprediction, because it calls strstr
many times with the same inputs. Can you please test it on a wider
variety of inputs in a randomized order?"