One use case we are looking at is using Ivy to implement dependency checking on the set of tools and their versions that our toolkit requires. For this I would imagine some kind of custom ToolsResolver would be needed, but based on my previous questions about custom resolvers, it seems there is quite a strong dependency on file-like objects, downloading, caches and so on. Does a tools resolver make sense and how might it best be implemented? There would be nothing to download but resolution would involve running each tool in a certain way to check its version (at least that is how we implemented our current checker which is less flexible). In this case would it be more appropriate to derive from AbstractResolver? Is it important to still have some kind of custom cache class also? thanks
paul
