Re: what is in modules vs what is in the core

2009-03-30 Thread M. Brian Akins


On Mar 30, 2009, at 7:37 PM, Paul Querna wrote:


mod_watchdog is the latest offender in a series of modules that expose
additional functions to the API. (mod_proxy and mod_cache do too!)

What happened to all functions that are not inside server/* must be
either dynamic optional functions or hooks?



Some modules (mostly 3rd party??) allow it either way - optional  
function or just linkage.  I'm personally a fan of hooks and  
providers.  (With providers, I usually just do the lookup once in,  
say, post-config, and "cache" the results in the "subscribing" module  
- this saves some hash lookups on potentially every single request.)


As I hack on some lua stuff, it's useful to have the symbols for  
functions.  That may just be because I'm lazy, because I could do  
optional function lookups in library opens, I suppose.  OT, but I like  
my Lua glue in a lua module and just use "require  
'apache2.memcache'" (or whatever) to do the linking.  This works  
really well with per thread lua states that are all loaded at  
startup... (hint, hint)


--Brian


Re: mod_serf: now somewhat working

2009-03-28 Thread M. Brian Akins


On Mar 28, 2009, at 11:09 AM, Paul Querna wrote:



- Much simpler configuration than mod_proxy, using location blocks
(or locationMatch), rather than ProxyPass' hacking of URI stuff way
earlier.



It'd be nice to be able to configure and do stuff at run time using Lua.

Someone had to ask...

--Brian


Re: mod_dbd analogue for memcached

2009-03-06 Thread M. Brian Akins


On Mar 6, 2009, at 1:25 AM, Kevac Marko wrote:


And after that you have to restart apache? It's not enough dynamic  
for us.




graceful restart.



Re: [PATCH] mod_dbd with more than one pool

2009-03-05 Thread M. Brian Akins


On Mar 5, 2009, at 4:52 PM, Kevac Marko wrote:


80 columns is a little bit ancient requirement in the world of 22"
LCDs, but ok, i'll fix that too.


Not when you have 4 code branches side by side...

Also, netbooks are pretty popular as well.  Most of my coding is done  
on a 15" laptop screen.


--Brian


Re: mod_dbd analogue for memcached

2009-03-05 Thread M. Brian Akins


On Mar 5, 2009, at 4:35 PM, Kevac Marko wrote:

ise.

mod_memcache, if we are talking about
http://code.google.com/p/modmemcache/, is too simple.



Works fine for me :)


I need multiple name pools to multiple servers. Something like




Will you implement a "default" pool like mod_dbd does?


This feature enables us to create dynamic, not static servers list.
For high availability clusters for example.



Sounds interesting, I suppose. I just generate my configs from  
templates and they generate the correct server list based on whatever







Re: [PATCH] mod_dbd with more than one pool

2009-02-11 Thread M. Brian Akins


On Feb 11, 2009, at 6:22 PM, Graham Leggett wrote:


Would it make sense for mod_memcache to become a provider beneath  
mod_socache, or am I missing something?


mod_memcache really just provides the config "glue" for apr_memcache  
so that every module that wants to use apr_memcache doesn't have to  
write all the config "glue" themselves.  Yeah, it could probably add  
some socache hooks as well. I'm adding some Lua glue right now.   
mod_memcache is already Apache licensed.




Re: [mod_lua] vm management

2009-02-06 Thread M. Brian Akins


On Feb 6, 2009, at 12:17 AM, Brian McCallister wrote:



* VMs are attached to apr pools.


+1



* Concurrency is up to the client


What are you defining as "the client"?


 We should not expose configuration options which will create
 concurrency issues, such as attaching a VM to the server_rec
 pool. It is very possible for someone to programmatically use the
 module to do things like that, but if they do any locking or
 resource pooling is up to them.


Sure, bcs attaching to server_rec pool is silly.  However, we do, IMO,  
need a way to have some other way besides r->pool in stock mod_lua.   
This is just too slow for a lot of things.




* One entry point to obtain VMs


+1.  I personally do not like exposing tons of struct info, but I'm -0  
on that.





We want to be able to get back to "stock" mod_lua, but it's just too  
darned slow right now :(  Having some type of per-allocation and long  
lived LuaStates helps this allot.  All the C stuff is done that way  
(in memory once, ran many times) and I'd like to be able to do the Lua  
stuff the same way without having to bolt on yet another bakins- 
specific module.


OT: toying with lua server pages :)