On Thu, Jun 16, 2005 at 08:36:53AM -0700, Brian Dessent wrote: >Igor Pechtchanski wrote: >>If it's ok to make cygcheck depend on wget, see >><http://cygwin.com/ml/cygwin/2005-06/msg00726.html>. >> >>FWIW, I'd suggest using the -p (--package-find) option for this in >>cygcheck. > >I've noticed that the package-grep.cgi script can somwtimes take a >while to complete, which is not surprising given that I've read that >sourceware.org tends to be overworked.
sourceware.org is somewhat less overworked these days after some recent performance tweaks. I'm doing more spam blocking at the SMTP level, we've remounted some of the disks using EXT2 rather than EXT3, and I'm throttling some problematic http connections -- all seem to account for some improvement. There is also a new super machine on order so this will not be an issue by the end of the year. >Adding a programmatic query to this would seem to just add more work >for the poor server. Would it make sense to dump the "tar tjf" output >of all the packages to a single static file on cygwin.com in a daily >cronjob, and then just grab that file and search locally? Or is >something similar already in place to facilitate the >"cygwin.com/packages/foo/foo-1.2.3"-type pages? That would be static, out-of-date data, just ripe for "I just did a 'cygcheck --whatprovides cygintl-2.dll' and the package don't exist!!!!" type of messages. OTOH, Red Hat gets away with this by providing a massive rpm static database for --redhatprovides queries. Of course Red Hat also provides monolithic releases, unlike Cygwin. cgf -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
