-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 27 Jan 2003, Jörg Walter wrote:

>      if (!$recreate) {
>          AxKit::Debug(1, "delivering cached copy - all conditions met");
> -        return $cache->deliver();
> +        return OK;
>      }
> 
> Here was that nice DECLINED hack which made Apache deliver the cached file 
> with all the Etag, If-Modified-Since, ... bells and whistles. The manual 
> delivery implemented further down the patch imposes extra reading and 
> copying.
 
> Moreover, AxOutputTransformer was designed to be _behind_ the cache. This 
> patch puts it before the cache. If it is used for anything cacheable, it's 
> purpose wasn't understood.

That was one thing in question - Should it really be behind it? Why do 
we NEED to transform the cache with every cache hit? Since Baud wasn't 
sure about an answer @axkit-dahut this morning this diff changed it that 
way.
This patch has ONE main reason - to get the current CVS version running 
again (for me). I don't know whether it's my config (which I did not 
change before upgrading to current CVS) or sth different - after upgrade I got 
a 
"'Can\'t call method "no_cache" on an undefined value at /usr/loca...') called a
t /usr/local/lib/perl5/site_perl/5.6.1/i686-linux/AxKit.pm line 162". at 
every cache hit. After -briefly- looking at the code I decided to try to 
get rid of the duplicated logic in AxKit::deliver_to_browser and 
AxKit::Cache::deliver while getting a running AxKit again.
I agree with the extra copying part introduced in my patch. Sanity checks 
are also missing. That patch  was supposed to be an answer to baud's 
question, for the line there I get the croaks, which is why I sent it to 
him and not to axkit-dev.
Anyway, in current CVS cache delivering is broken. Question is, how to 
fix it, and with it, whether we want to keep the duplicated delivering 
logic in AxKit.pm and Cache.pm for example.

michael

- -- 
IT Services
University of Innsbruck

-----BEGIN PGP SIGNATURE-----
Version: PGP 6.5.8

iQA/AwUBPjV913jN0CNeKmaIEQJc0QCgpqv7zQTZh0ftjes2GanDk9N4NEAAn1fN
N+J7MmT2BkcFC5QpWX7ES8+K
=SF6G
-----END PGP SIGNATURE-----

Reply via email to