On 5/28/2013 01:05, RW wrote:
On Mon, 27 May 2013 22:33:53 +0200
John Marino wrote:
On 5/27/2013 22:09, RW wrote:
On Mon, 27 May 2013 20:38:11 +0200
John Marino wrote:
No, that's something you just made up. It is however vague and
anecdotal. We have only one data point that we know is from this
year and not self-inflicted, even if the others are, for all we
know it could still be fast most of the time.
Some monitoring would be useful.
However you slice it, a distinfo file with 1000+ entries is
completely absurd. 95% of the blame goes to Vim developers.
However, it is within the realm of feasibility to pre-package patches
in batches of 100 (or conversely 1 tarball of patches rolled for
every time patch count hits multiple of 100).
In other words downloading every patch twice.
No. That's not what those words mean.
Please stop assuming that somebody builds Vim repeatedly and start
assuming it's built for the very first time. Also, given these patches
are a couple of kilobytes at most, a compressed tarball of 100 patches
(or even 700 patches) is negligible. Even if somebody with a cache
downloaded it twice, so what? It's not even noticeable.
At the very, very least maybe only HTTP hosts are listed for VIM (I
just checked bsd.sites.mk, the ftp sites are all at the end of the
list now)
All 13 http links would have to fail before the ftp links are
tried.
So what's the point of having them on the list? Isn't 13 mirrors enough?
I may have still been on the old bsd.sites.mk with a site> 10
seconds per file. (this is yet another data point)
We already knew that it was slow before January, so that's irrelevant.
It validated my story as more than anecdotal.
_______________________________________________
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
