[jboss-dev] Improving scanning performance

[Scott Marlow]

Ales Justin wrote:
>> Is there any way that we could persistently cache the results of 
>> scanning for annotations?  
>
> Why would you want to cache something like this in the first place?
> You actually never explained this. ;-)
>

I wanted to explore caching as a possible way to reduce the cost of 
scanning for annotations.  I'm still planning on using the 
jboss-scanning.xml support for excluding classes that don't have 
annotations. 

Question about jboss-scanning.xml, if I specify the "include" list but 
leave the "exclude" list blank or missing.  Will that exclude all 
classes that aren't on the "include" list?  If yes, that will make it 
easier to build the jboss-scanning.xml.

>> Putting aside the issue of how to invalidate the cache (lets discuss 
>> that after), where would the caching logic need to go (maybe near the 
>> code dealing with jboss-scanning.xml?)
>
> Dunno, as I'm still not convinced this is something we should do.
>
> Instead of any caching, I could see some easy-to-do predetermined 
> annotation info (X).
> That would be far more efficient and easy to do, than "proper" caching.
>
>> Are there currently any plans to do persistent caching in the MC (or 
>> some form of pre-compiling)?
>
> No.
>
> Pre-compiling?
> I guess we could hack some tool that would know how to do (X).
> Tattletale looks perfect for such things. ;-)
>
> _______________________________________________
> jboss-development mailing list
> jboss-development at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jboss-development
>