[ 
https://issues.apache.org/jira/browse/SLING-9662?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17180571#comment-17180571
 ] 

Georg Henzler edited comment on SLING-9662 at 8/19/20, 2:18 PM:
----------------------------------------------------------------

One open question is how we deal with the lifecycle of ResourceResolvers (and 
ResourceResolverFactory resp.) compared to the ResourceToUriMapper services - 
ATM the lifecycle is independent [1], references to 
ResourceResolver/ResourceResolverFactory remain intact if a ResourceToUriMapper 
vanishes (e.g. during a deployment) and the respective mapping will just be 
"missing" during this time (which could potentially lead to problems, but a 
ServiceUnavailableFilter can be configured to mitigate). I think this behaviour 
is in line with changes to /etc/maps (here the bundle is also not restarted 
upon changes IIRC). We could introduce a config property 
"requiredResourceToUriMapperServices" that the activator for the 
ResourceResolverFactory waits for, but if not needed I would like to avoid the 
complexity.

[~cziegeler] [~rombert] Do you see a problem with the current referencing 
approach as it is on branch [1]? 

[1] 
https://github.com/apache/sling-org-apache-sling-resourceresolver/blob/9a703ac38bee64c6ebe73b5c4df3b56307fdaed0/src/main/java/org/apache/sling/resourceresolver/impl/mappingchain/ResourceUriMappingChain.java#L53


was (Author: henzlerg):
One open question is how we deal the the lifecycle of the ResourceResolver (and 
ResourceResolverFactory resp.) compared to the ResourceToUriMapper services - 
ATM the lifecycle is independent [1], references to 
ResourceResolver/ResourceResolverFactory remain intact if a ResourceToUriMapper 
vanishes (e.g. during a deployment) and the respective mapping will be just 
"missing" during this time (which could potentially lead to problems, but a 
ServiceUnavailableFilter can be configured to mitigate). I think this behaviour 
is in line with changes to /etc/maps (here the bundle is also not restarted 
upon changes IIRC). We could introduce a config property 
"requiredResourceToUriMapperServices" that the activator for the 
ResourceResolverFactory waits for, but if not needed I would like to avoid the 
complexity.

[~cziegeler] [~rombert] Do you see a problem with the current approach as it is 
on branch [1]? 

[1] 
https://github.com/apache/sling-org-apache-sling-resourceresolver/blob/9a703ac38bee64c6ebe73b5c4df3b56307fdaed0/src/main/java/org/apache/sling/resourceresolver/impl/mappingchain/ResourceUriMappingChain.java#L53

> Introduce a Resource Mapping SPI
> --------------------------------
>
>                 Key: SLING-9662
>                 URL: https://issues.apache.org/jira/browse/SLING-9662
>             Project: Sling
>          Issue Type: New Feature
>          Components: API
>            Reporter: Georg Henzler
>            Assignee: Georg Henzler
>            Priority: Major
>
> Introduce a simple SPI that allows to contribute to the resource resolver's 
> map() and resolve() methods:
> *ResourceUri:*
> General purpose class to represent a ResourceUri (like e.g. a link in an html 
> <a> tag). Immutable itself but adjustable using the builder pattern. Part of 
> the Sling API and can be used anywhere to simplify handling/modification of 
> Sling ResourceUri. Implements RequestPathInfo. ResourceUri can be created 
> easily from a String, a Request, a Resource, a URI or a RequestPathInfo.
> *ResourceToUriMapper:*
> SPI interface to be implemented as OSGi Service. All registered services 
> build a conceptual chain sorted by service ranking. The resource link is 
> passed through the chain while any ResourceToUriMapper chain member may or 
> may not make adjustments to the resource link.
>  rr.resolve() passes through the chain starting at the ResourceToUriMapper 
> with the <strong>highest</strong> service ranking and rr.map() passes through 
> the chain starting at the ResourceToUriMapper with the 
> <strong>lowest</strong> service ranking
> _Mailing List References:_
>  [https://www.mail-archive.com/dev@sling.apache.org/msg93537.html] 
>  [https://www.mail-archive.com/dev@sling.apache.org/msg87736.html]



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to