Thanks Will! I'll definitely try freezing to edge and testing this out. On Wed, Feb 2, 2011 at 3:19 PM, William Ross <w...@spanner.org> wrote:
> On 2 Feb 2011, at 10:19, potapuff wrote: > > > Have same problem: all pages a responsed with 200 code. > > > > First of all i'd check: > > SiteController.cache_timeout ~1 day > > Radiant::Config['dev.host'] is different from production > > Page.find_by_parent_id(nil).cache? true > > I've been looking into this a bit more and I suspect that for both you and > Carl these tests are conducted while you are logged into the admin part of > the site? Rack::Cache by default is configured to pass through any request > that comes with a Cookie header. That would explain why visitors like Anton > didn't see the performance problems that Carl was describing. > > If so it could be considered a testing artefact, but there are other > reasons why a cookie might be present and we don't want them all to disable > the cache. I've just checked in an update to make the cache indifferent to > cookies and explicitly uncache the admin pages instead. I'd be interested to > see if your results are different now. > > There is probably a debate to be had here but imho the radiant codebase > generally presumes that pages are not personalised and if we want to build > for efficient delivery of per-user content then we need to change a lot more > than just the cache headers. I think this update just makes things more > consistent, but I'm happy to be disagreed with. Caching is a tricky subject > and I am often wrong. > > This should also help with extensions like reader and shop that use their > own cookieing scheme. I'm doing some experiments with cached forum pages to > see how that works out. > > best, > > will > > > -- > Radiant CMS Dev Mailing List > Post: radiantcms-...@googlegroups.com > Unsubscribe: radiantcms-dev-unsubscr...@googlegroups.com > Group Site: http://groups.google.com/group/radiantcms-dev/
