[
https://issues.apache.org/jira/browse/TUSCANY-2882?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12790246#action_12790246
]
Phil Housley commented on TUSCANY-2882:
---------------------------------------
Well, I completely forget about this issue for most of a year, and so obviously
it has come back to haunt me. I've been looking at porting my patch to 1.5.1
for a current project, and found that things have changed a bit in the
implementation-spring packages. What I have pasted together for now is
changing the assumption that all paths are relative:
{code}
parentPath.substring(0, parentPath.lastIndexOf("/") + 1) + location;
{code}
To consider that they may begin with a "classpath:":
{code}
if (location.startsWith("classpath:")) {
location = location.substring(10);
resource =
Thread.currentThread().getContextClassLoader().getResource(location);
}
{code}
Would this make an acceptable patch, or would it be too hacky? There is
already a "FIXME - need to find a right way of generating this path" in the
code, so I don't think I'm really making things worse.
> Relative path imports don't work in Spring implementations.
> -----------------------------------------------------------
>
> Key: TUSCANY-2882
> URL: https://issues.apache.org/jira/browse/TUSCANY-2882
> Project: Tuscany
> Issue Type: Bug
> Components: Java SCA Spring Implementation Extension
> Affects Versions: Java-SCA-1.4
> Reporter: Phil Housley
>
> Using implementation.spring, the Tuscany bean file parser and Spring's parser
> interpret import links in different ways. Spring sees them as relative,
> tuscany as absolute, so one parser will generally fail unless everything goes
> into a root folder.
> I have a patch for this, using Spring's relative resource system, but I
> cannot guarantee that the two parsers now use the same mechanisms. I'll
> attach when I've cleaned it up.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
