29-May-2003 11:12 Ian Collier wrote: > On Thu, May 29, 2003 at 10:50:46AM +0900, Henry Nelson wrote:
>> In many cases it means that each page from >> the server is unique, i.e., each time you access the URL, the page >> that the server returns is different from any page you may have >> received on previous accesses to the same URL. A cached document in >> that situation perhaps needs to have it's base URL time stamped so >> that its content is not confused with the constantly changing content >> of the original URL. > If I view the source then I generally want to see the source of the page > I'm looking at right now, not the source of the page that would have > been generated if I'd asked for it right now (and without the Referer > header, which makes a big difference on certain kinds of site). You are right. Henry mix two completely different concepts: source cache is used for browser' presentation purposes only, the document already received - it is not a server business whether you will include links to images or exclude them or change from one to another... While the proxy is HTTP agent, it should follow HTTP/1.x expiration model - check timestamps, interpret several HTTP headers (for example, http server may ask proxy to not store a copy of the document). The waste of resources is minimal: by default, lynx caches last 10 documents (HText) in memory, source cache will cache .html either (if stored in memory, it doubles the memory usage in a worst case). Compare with MSIE which stores hundreds and hundreds of images and pages as temp files. >> I don't have any idea how cacheing of https would work. Does the cache >> somehow keep state on the handshake between the server and client? > It keeps the source of the document which you are viewing - I don't see > what handshaking has to do with it. >> Along >> lines that Doug mentioned, I'd be rather concerned about any confidential >> information that the server may have put into the cached source, and about >> any information I may have entered into a form. > But I've already got the document - if anything was confidential then > it's too late! No one else is going to see it if it's only cached in > memory. >> While most of you probably look upon me as reactionary, I still think that >> the client should not be doing cacheing at all. Lynx has so many ways to >> easily plug into a proxy. In addition, unlike the past, it is pretty easy >> for anyone to install a fully compliant and configurable cacheing proxy. >> It's very much like the recent questions about Lynx's news capabilities. >> Lynx's news support is totally out-of-date and non-compliant. Look back >> in the archives to see the status on that. Like news, is it worth >> implementing cacheing if it's not going to be at least as good as what's >> already available, e.g., squid? > That's ridiculous... most Linux installations don't come with squid > unless you ask for a full or server install, and I'm not about to > bloat my system with that just so I can use Lynx properly. > imc > ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to [EMAIL PROTECTED] ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to [EMAIL PROTECTED]
