This is the mail archive of the
mailing list for the Cygwin project.
Re: Package file lists archive
- From: Warren Young <warren at etr-usa dot com>
- To: cygwin at cygwin dot com
- Date: Mon, 09 Jun 2014 09:17:32 -0600
- Subject: Re: Package file lists archive
- Authentication-results: sourceware.org; auth=none
- References: <CAAXzdLUx+zsvXGDfYyBkounDNLsMy8oW1ZmGZtLoOA-XfeDQHg at mail dot gmail dot com> <20140526201210 dot GB3903 at ednor dot casa dot cgf dot cx> <CAAXzdLUZ-0iR4NR2j30UPJpmr4LfPtqBXfPtc7vh-nY3ZfF9Ag at mail dot gmail dot com> <20140527221151 dot GA6521 at ednor dot casa dot cgf dot cx>
On 5/27/2014 16:11, Christopher Faylor wrote:
This would ease the burden of the cygwin.com site, as it is the only site with
that information currently
It is surviving just fine.
Speaking of, is the source for package-grep.cgi available somewhere? I
could quarter the page weight, and probably improve the render time in
browsers considerably, too. It currently takes about 16 seconds to
completely load in Firefox on my system, most of which is just page
rendering time, due to the huge table used for layout.
(I'd also need either a static copy of its input, or a way to easily
generate it locally.)
The /packages page load time is a considerable fraction of this, at
roughly 2-4 seconds. It's only returning about 120 kB of text
(compressed) so that comes to roughly half a megabit/sec. I assume that
means the script is running every time, re-generating the whole page
text anew instead of caching the output, despite the fact that the input
probably changes less often than once a day on average. Caching should
give a page load time of a few hundred ms, and save some CPU time to boot.
Problem reports: http://cygwin.com/problems.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple