On Sun, Nov 18, 2001 at 08:12:59PM -0700, Chris Tillman wrote: > > So, you're thinking along the lines of a cgi that would initially > respond to a user's 'littledoc' request by extracting just the pieces > we want from the .tgz, and present the info to the user. And then it
Any request from package foo would cause all the relevant files from that file to be put in the cache. Also, the files would be extracted from the .deb since they are in a know location there. > would store that info in a disk cache file so the next user to request the > same info wouldn't cause it to re-extract? Each server's cache file Yes. > would have different contents based on what requests it's > seen. Wouldn't that make servers' resource needs hard to predict? > It's not clear what you mean by "each server's cache". The cgi script would be handled from one machine so only that machine would need the cache. My guess is you believe there is a connection between the debian web stuff and the debian archive. There isn't. > BTW, I mentioned the README because it sometimes would give a better > idea of what the package does. Some package descriptions are obtuse or > just plain too brief to convey understanding. > The README is missing from enough packages that creating a link for it would confuse many people. Also, it is frequently used for the same thing as an INSTALL file making the results even less useful. In fact, a file containing a README that only contains installation instructions should have a bug filed against it. If a package has a long description that isn't useful, file a bug. Even better, include a possible replacement text. -- James (Jay) Treacy [EMAIL PROTECTED]
