From: Claes W. <kl...@hy...> - 2003-05-07 22:58:09
|
On Thu, May 08, 2003 at 12:51:35AM +0200, Claes Wikstrom wrote: > > > > > Hmm, some thoughts... The cache itself should run autonomously, > > monitoring the activity of cached pages, uncaching pages which have > > not been recently accessed, etc. The synchronization of a cached page > > with its filesystem source is really driven by an external event -- the > > updating of the source. Perhaps having an api and matching > > administrative interface which triggers the re-compilation of a single > > file or a directory tree, which content updaters utilize, would be > > fine. Something like a web tool for content managers. (There could be > > other advantages too -- perhaps the content manager wants a set of > > pages to be made live simultaneously, and would like the satisfaction > > of pressing a button to synchronize all files at once, rather than just > > upload and pray...) > > > > This I like, > > 1. upload new content > 2. execute yaws -hup (or something) > which just (atomically) ampties the cache. Clean, fast, nice > and not error prone. > ... hmmm second thoughts. It doesn't jive very well with the situation where some webdeveloper ftp uploads content to a machine where he cannot execute any commands. ?? /klacke -- Claes Wikstrom -- Caps lock is nowhere and http://www.hyber.org -- everything is under control |