Some requests for new features

[jrabbit]

Forgive me if any of these are already implemented - I'm not actually
using mod_cache yet, I've only read the documentation and made a brief
scan of the source. In order of descending importance:

1. Support the 1.5.0 mod_deflate which is now part of the standard
lighttpd release. All my dynamically generated content is deflated,
and that is the content I'd ideally like cached. This is why I'm not
currently using mod_cache. (If this does work, the lighttpd wiki needs
updating)

2. Support Vary HTTP headers from the upstream modules to cache
multiple versions of a URL rather than ignoring any response that
sends vary.

3. This probably already works but I haven't seen it explicitly
mentioned: As well as using headers that originate from
mod_proxy_core, also support those coming from mod_setenv.

4. Support the sending of 304 status codes for GET requests when a
request for a cached URL has an If-Modified-Since header newer than
the last modified date of the cached file.

5. Have an option to only enable caching where explictly requested,
whereby the criteria for something being cached is the presence of
"public" in a Cache-Control header from the upstream modules, rather
than the lack of a no-cache/must-revalidate/private. In this mode. no
Cache-Control header would mean the content is not cached.

6. Look for the the max-age in the Cache-Control header as well as the
Expires header when determine how long to cache content. If both are
present use the shorter period.

[Hongyu]

To implement these new features modcache needs to store cache informations to database. I used to develop another modcache branch which used Berkeley DB to store cache informations. But BDB don't work fast enough as I expected and sometimes blocks lighty, then I drop that branch.

 

Lighty is single process, event driven server. If one request blocks, lighty blocks. Database is one source of block operation. Can anybody tell me which database is suitable?

 

Regards

[sebest]

You can't store these information in flat text files and use aio ?

[Hongyu]

CDB is quite out of date and isn't good as it advertised.

When you want to update the CDB, you have to regenerate cdb file from plain text, so you have to keep copy of original text file.
With the original small text files, so we have no needs to use cdb.

On Fri, Mar 28, 2008 at 2:41 AM, Ming <min...@gmail.com> wrote:

sorry, the Reply (to list) button is missing, so I can only reply to
author.

check out

http://cr.yp.to/cdb.html

for fast db.

[Istra]

Why can't you use the file system? (like apache mod_cache)