[ https://issues.jboss.org/browse/RF-12537?page=com.atlassian.jira.plugin.s... ] Lukáš Fryč updated RF-12537: ---------------------------- Description: The order of resources rendered in a head is currently un-managed and can cause resource ordering issues. (Which may be partially fixed by using different placement of h:outputScript tags). Furthermore a performance of application perceived by user might be significantly enhanced when ordering head resources correctly (see Yahoo recommendations). The proposal of {{HeadRenderer}} by Adrian Mitev: https://gist.github.com/3899465 ---- This proposal has several drawbacks: * doesn't count with mime-types, depends on resource names (Resource provided by ResourceHandler may be used here to obtain mime-type) * resource names stored in attribute "name" are not part of JSF spec (afaik) and thus implementation-dependent * the implementation does not count with resource library notion (.reslib) * ECSS is unsupported, but it can be simply changed with using ResourceHandler as described above Other than that the proposed solution seems like enough generic. was: The order of resources rendered in a head is currently un-managed and can cause resource ordering issues. (Which may be partially fixed by using different placement of h:outputScript tags). Furthermore a performance of application perceived by user might be significantly enhanced when ordering head resources correctly (see Yahoo recommendations). The proposal of {{HeadRenderer}} by Adrian Mitev: https://gist.github.com/3899465 ---- This proposal has several drawbacks: * doesn't count with mime-types, depends on resource names (Resource provided by ResourceHandler may be used here to obtain mime-type) * resource names stored in attribute "name" are not part of JSF spec (afaik) and thus implementation-dependent * the implementation does not count with resource library notion * ECSS is unsupported, but it can be simply changed with using ResourceHandler as described above Other than that the proposed solution seems like enough generic. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://issues.jboss.org/browse/RF-12537?page=com.atlassian.jira.plugin.s... ] Lukáš Fryč commented on RF-12537: --------------------------------- Another idea is the mechanism of sorting resources plugable (using ServiceTracker). -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://issues.jboss.org/browse/RF-12537?page=com.atlassian.jira.plugin.s... ] Ken Finnigan commented on RF-12537: ----------------------------------- After discussions with JSF guys at Oracle today I've found out that facets within h:head are ignored if they contain h:outputStylesheet and h:outputScript elements. The problem is that the JSF resource relocation bit finds all the outputStylesheet and outputScript elements within the head, even if they're in facets, and relocates them to be children of the head itself. This is expected behavior as of current JSF 2 spec. http://java.net/jira/browse/JAVASERVERFACES_SPEC_PUBLIC-1139 has been raised to add the ability for facets to be added into h:head without them being ignored. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://issues.jboss.org/browse/RF-12537?page=com.atlassian.jira.plugin.s... ] Ken Finnigan commented on RF-12537: ----------------------------------- Lukas, unfortunately it goes beyond that. I had my own head renderer and would see no facets inside encodeEnd() of that renderer if the facets just had outputScript or outputStylesheet elements present. The JSF resource relocation, which supports targeting resources at the head of page from inside the body, is the bit that "relocates" the outputScript and outputStylesheet elements from inside the facet to directly on the view root. And as there is nothing else in the facet the facet is no longer there. If you have other tags in the facet other than outputScript and outputStylesheet, the facet is present with only those non JS/CSS tags. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://issues.jboss.org/browse/RF-12537?page=com.atlassian.jira.plugin.s... ] Lukáš Fryč edited comment on RF-12537 at 11/5/12 4:23 AM: ---------------------------------------------------------- Ken, could we use custom tag (e.g. {{<r:resources target="head-begin" />}}) instead of {{f:facet}}? was (Author: lfryc): Ken, could we use custom tag (e.g. {{<r:resources target="begin" />}}) instead of {{f:facet}}? -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://issues.jboss.org/browse/RF-12537?page=com.atlassian.jira.plugin.s... ] Ken Finnigan commented on RF-12537: ----------------------------------- Lukas, I'm not sure. You'd need to try to find out, but I have a nasty feeling the Resource relocation bits would still find stylesheet and script JSF tags and pull them out. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira