I think a provider is providing data, and if you call a URL then the provider handling that data should provide the data, always. As you point out, its more consistent, and in the case of the file provider the 2 things are the same anyway.
Besides, if you want the file itself with no processing, why would you use ANY provider at all? In that case it seems like the best thing would be to have a seperate URL where you can get "raw" data. One that AxKit doesn't even know about... On Monday 20 May 2002 09:22, Matt Sergeant wrote: > This is a follow up on a chat on IRC with Kip... > > I was wondering if the current passthru system is doing the wrong thing. > At the moment it merely "punts", and passes the request through, allowing > Apache to handle it. But what if you have a provider installed for (say) > POD files or for OpenOffice files, or for emails. The provider is > responsible for extracting XML from those files in whatever way it can. > > Should a passthru on a request like http://server/test.sxw?passthru=1 > return you: > > a) The OpenOffice document stored on the filesystem. > b) The XML extracted from the above OpenOffice document (i.e. an AxKit > handled request, just without any stylesheets applied). > > When you've come up with an answer, now reconsider for the case when the > provider gets the XML from somewhere other than the filesystem. When > that's the case, b seems like the more sane option (which isn't what we do > right now). > > So overall the question becomes do we change passthru, or do we add some > new passthru-like operator? I'd prefer to fix passthru, as I hate to add > extra cruft, but I'm open to other thoughts on this (including ones that > say "Do nothing - it ain't broke don't fix it"). --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
