Date: Sun, 22 Aug 2004 12:50:37 -0700 From: Doug Barton <DougB@DougBarton.net> To: Colin Percival <colin.percival@wadham.ox.ac.uk> Cc: Colin Percival <cperciva@FreeBSD.org> Subject: Re: cvs commit: src/etc services Message-ID: <4128F90D.8080506@DougBarton.net> In-Reply-To: <6.1.0.6.1.20040816045454.02a34168@popserver.sfu.ca> References: <200408161152.i7GBqMlq084661@repoman.freebsd.org> <6.1.0.6.1.20040816045454.02a34168@popserver.sfu.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
Colin Percival wrote: > At 04:52 16/08/2004, Colin Percival wrote: > >>cperciva 2004-08-16 11:52:22 UTC >> Modified files: >> etc services >> Log: >> Add 3653/tcp and 3653/udp, since they were specifically requested. > > > As discussed earlier, I have another 6400 lines ready to > merge from IANA's list. Since some benchmarks have been > produced showing that adding that many new entries into > /etc/services could have a significant performance impact, > I'm going to wait until after 5-STABLE is branched, and > only commit these to 6-CURRENT. If/when the performance > impact becomes insignificant (due to the addition of a > lookup daemon, for example), I'll merge them into STABLE. Please don't commit these changes to 6-current without discussion on -arch, which I would like to be involved in. This topic has come up many times before (including a long time ago instigated by me) and the consensus has always been to add lines from the IANA table when they are specifically requested, IF they add value to a significant percentage of our userbase. I don't see any reason to change that philosophy with 6-current, and I'd like to see significant discussion as to why it's a value add to change the philosophy before any changes are made to the file. Also, I'd like to point out that many of the port numbers in the current IANA list are not related to active products, so adding them all would be a waste of time, and degrade performance for no good reason. Finally, the solution that has most often been discussed for improving the performance of getservbyname() and friends is a Berkeley db implementation similar to the way login.conf is done now. I'm not sure a lookup daemon is justified, and I'd like to see some discussion about this topic too. Thanks, Doug -- If you're never wrong, you're not trying hard enough
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4128F90D.8080506>