> -----Original Message-----
> From: Alexey Borzov [mailto:borz_off@cs.msu.su]
> Sent: 22 December 2004 14:48
> To: Dave Page
> Cc: Oleg Bartunov; Steve Simms; pgsql-www@postgresql.org
> Subject: Re: [pgsql-www] Ready to launch?
>
> Hi,
>
> Dave Page wrote:
> >>How much pages rebuildes and does everything rebuilded each run,
> >>regardless on changes ?
> >>20+ minutes looks too much for medium sized server.
> >
> > It builds around 10,000 pages, and does so by spidering the
> site as an
> > indexer would, and saving the output. It's clearly not the most
> > efficient way, but given the number of different ways that pages get
> > built on the site, it is certainly the most maintainable.
>
> Do you spider all the static docs every time? They only need to be
> regenerated on design change or docs update after a new release.
At the moment it spiders the whole site, exactly as you wrote it (with
the exception that it ignores /redir).
> > I have a suspicion that it might not be avoiding saving
> duplicate links
> > but I haven't checked that yet.
>
> If it was *that* stupid, the mirroring would take *much* more than 20
> minutes. ;]
Would it? I don't know how many links there are to the doc sets. Mind
you, I suppose each doc links to the next and previous, so it would hit
each at least 3 times which it clearly isn't.
Regards, Dave.