httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From CASTELLE Thomas <tcaste...@generali.fr>
Subject RE: mod_cache behaviour in Apache 2.0.44
Date Fri, 24 Jan 2003 16:55:06 GMT
One more question about mod_cache :
 
How does it handle conditionnal requests ? Because, in mod_cache.c, in the
cache_url_handler function, it checks if info->etag and info->lastmods are
set and, if not, make a nonconditionnal request to the end server.
 
BUT, it seems that info->lastmods and info->etag are never set, because they
are not saved on the disk through this structure (see mod_disk_cache's
write_headers function), but they are saved as other headers. So, my
questions are :
 
* Is it a bug or just not yet implemented ?
* If it's a bug, should we take the ETag and the Last-Modified info directly
from the cached headers with : 
   lastmods = apr_table_get(r->headers_out, "Last-Modified");
 
Thanks for helping me with that ! And by the way, if you could look at my
preceding mail... ;-)
 
Thomas.
 
-----Message d'origine-----
De : CASTELLE Thomas [mailto:tcastelle@generali.fr]
Envoyé : vendredi 24 janvier 2003 12:46
À : dev@httpd.apache.org
Objet : mod_cache behaviour in Apache 2.0.44



Hi there ! 

I'm currently looking at mod_cache to see if it is enough RFC compliant to
be used in production environment, and I pointed to something that seems a
bug to me, maybe can you help :

in cache_util.c, in the ap_cache_check_freshness function, 

shouldn't the test : 

if ((-1 < smaxage && age < (smaxage - minfresh)) || 
        (-1 < maxage && age < (maxage + maxstale - minfresh)) || 
        (info->expire != APR_DATE_BAD && age < (apr_time_sec(info->expire
-
info->date) + maxstale - minfresh))) { 

be instead : 

if ((-1 < smaxage && age < (smaxage - minfresh)) || 
        (-1 < maxage && age < (maxage + maxstale - minfresh)) { 
   if (info->expire != APR_DATE_BAD && age < (apr_time_sec(info->expire -
info->date) + maxstale - minfresh)) { 

so that maxage and smaxage could be taken in consideration ? Because when
there is a Last-Modified header, info->expire has a value and then takes
priority over max-age. And I think that Last-Modified shouldn't take
priority over Cache-Control: max-age...

Thanks for looking at this ! 

Thomas. 

-----Message d'origine----- 
De : Graham Leggett [ mailto:minfrin@sharp.fm <mailto:minfrin@sharp.fm> ] 
Envoyé : vendredi 24 janvier 2003 08:04 
À : dev@httpd.apache.org 
Objet : Re: "Group" not working properly 


Jeff Trawick wrote: 

> As others mentioned, that big number has the same representation as 32-bit

> -1.  I checked a couple of systems and gid_t is unsigned 32-bit there, 
> and Apache treats the number as gid_t internally.  The user can code that 
> big unsigned number for group if they want to.  Operationally, the right 
> thing happens either way; there is no bug. 

Is it possible to put a warning in there, because 4294967295 "looks wrong". 

Regards, 
Graham 
-- 
----------------------------------------- 
minfrin@sharp.fm                "There's a moon 
                                        over Bourbon Street 
                                                tonight..." 


Mime
View raw message