Re: More on classpath order semantics (and lookup)

This is a little tricky in IntelliJ.

In IntelliJ you can easily configure several web roots, the problem is that
you can't configure their precedence. With an quick test i managed to figure
out how IntelliJ solves conflicts between resources in diferent roots, but
this is internal behaviour. Even knowing this, like you say, there's no way
of knowing how some other application that builds the war will handle the
conflict.

So, i think that context conflicts should be treated as an abnormal
situation, something that shouldn't normally happen. More, if such a
conflict exists it's problably because a developer created it without
noticing that a resource with the same name and path already existed. As
such, i think that the correct thing to do would be to show a warning to the
user informing him of the conflict and advising him to resolve it. The
implementation of lookup in this case should try to figure out in the best
way it can which resource it should hide and which it should return, as long
as the user doesn't solve the conflict.


What do you think ?

Hugo


On 08/02/06, Geoff Longman <[email protected]> wrote:
>
> So, after all the verbal diarrhea in the last message I want to talk
> some more about lookup and classpath order semantics.
>
> I asked myself why is this a big deal with classpath order semantics
> in the context root? There are no hidden files in an exploded war
> layout.
>
> This is true if you follow Spindle's example and 'require' the project
> context to be in exploded war format.
>
> But there is nothing saying that an exploded war format is *required*
> by the core. On having more than one folder represent the context
> root, some people actually prefer it.
>
> It's an IDE implementers choice whether they want to support more than
> one 'root' folder in the context root. I mean the core is ignorant of
> the implementation of IResourceRoot/ICoreResource/ISearch.
>
> But I think in this case classpath order semantics should come into play.
>
> It could be the case that a file in one context root folder is
> 'hidden' by a file by the same name in another. I guess it depends on
> how a real war file is built for deployment from all these folders but
> would guess that one would want *something* like classpath order
> semantics to apply!
>
> Or am I off base here?
>
> Geoff
>
> --
> The Spindle guy.          http://spindle.sf.net
> Get help with Spindle:
> http://lists.sourceforge.net/mailman/listinfo/spindle-user
> Blog:                     http://jroller.com/page/glongman
> Feature Updates:          http://spindle.sf.net/updates
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log
> files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://sel.as-us.falkag.net/sel?cmdlnk&kid3432&bid#0486&dat1642
> _______________________________________________
> Spindle-developer mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/spindle-developer
>