> 
> I thought that caching-point pipelines work similarly to the 
> situation that for every new component in the processed
> pipeline something like new subpipeline is being created. 
> That means if we have generator G, transformer T and
> serializer S, the cache entries would be:
> G, GT, GTS (every processing stage is cached)
> 
> Am I correct?

I must confess I have never tried anything with the caching-point pipeline, so 
perhaps somebody else can explain.

My story holds for the normal caching pipelines. I have never seen a reason to 
use caching-point pipelines, 
because I always could solve it by seperating a single pipeline into multiple 
ones. 

As I understand from the caching-point pipeline, you can set the 
autoCachingPoint to on/off (default on) and
when off, you can indicate by hand caching points 
[pipeline-hints="caching-point"].

IMHO, I would never use autoCachingPoint algorithm, because IMO, it is an easy 
way to pollute your cache. 
Setting the caching points by hand....well, then I just use a seperate pipeline 
instead.

I never saw any real benefit in the cachin point pipelines.

Ard

> 
> -- 
> Grzegorz Kossakowski
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
> 

Reply via email to