[jboss-dev-forums] [Design of POJO Server] - Re: Regression on VirtualFile.toURL()

"jason.greene(a)jboss.com" wrote : "scott.stark(a)jboss.org" wrote : "jason.greene(a)jboss.com" wrote : "alesj" wrote : | | | | But wouldn't that make Carlo's impl fail on exploded jar? | | | | | | Yes it would most definitely break on exploded jars. | | | | The issue is that you need to know what the original form of the URL the reference was against. You can't start with a compressed jar with a manifest classpath relative to the jar as a file, and then use the same relative path against the jar as a directory. The jar manifest classpath spec does not deal with this distinction. | | | | If we ever stop unpacking archives such that a jar classpath resolves against an ear which is still packed, we will have the opposite problem of the default relative reference resolution pushing the jar to be outside of the ear. | | So to truly solve this issue, vfs urls would need to be completely logical, and have no distinguishing difference between an exploded directory and a compressed jar. Those really are physical aspects. Which seems to indicate that Ales' consistent use of / on non-leaf urls is the correct approach. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4211245#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...