Community
Participate
Working Groups
Build Identifier: According to the almighty BuckyBook, the <matcher> for the URL reader type currently works as follows: - crawl the given base URL for <a href=".."> elements and collect them in a list of links - apply the <matcher> element on the list of links I am currently trying to use the SpringSource Enterprise Bundle Repository in a Buckminster build (as probably many have done before). They do not offer a P2 repository, so I have to use a URL reader. I would like to be able to use a matcher. However the only "overview page" (in this case an RSS feed) listing all their bundles doesn't contain <a href="..">, but <link href="..">: http://www.springsource.com/repository/app/search/atom?query=&type=bundlesByName It would therefore be very helpful, if the search pattern used to create the link list on which the <matcher> is applied, would be configurable (e.g. look for <link> elements instead of <a>). Reproducible: Always
Did you try using the maven reader and the maven repository?
(In reply to comment #1) > Did you try using the maven reader and the maven repository? Indeed I did. However, as SpringSource has a very "interesting" maven repo setup, it seems to be pretty much unusable with Buckminster. Let me give you an example: http://www.springsource.com/repository/app/bundle/version/detail?name=com.springsource.net.sf.cglib&version=2.2.0&searchType=bundlesByName&searchQuery=cglib In order to use this bundle from the SpringSource maven repository, I would have to manually configure groupID and artifactID. At least the groupID cannot be easily computed automatically. This means I would have to modify the searchPath in my RMAP for every single bundle I want to use. This makes the whole repository idea pretty much useless. Using the URL matcher would work, if the matcher was more configurable.
