I am not XWiki developer and hardly ever would be. Even if i have my favorite mysfeatures. Also here are script-making and script-using guys. So i don't think this question would be complete off-topic. === It seems that XWiki was never tested on Windows as none of developers had it. It seems that XWiki GitHub just cannot be cloned to Windows now. And i don't know if can be compiled there (hopefully JVM can overcome it, but who knows) But what i noticed again, is that hugely redundant file naming xwiki-platform-core/xwiki-platform-classloader/xwiki-platfor m-classloader-protocols/xwiki-platform-classloader-protocol-attachmentjar/src/te st/java/org/xwiki/classloader/internal/protocol/attachmentjar/AttachmentURLStrea mHandlerTest.java Why that repeating time and gain, like if you use flat filesystem rather than tree ? It looks like calling an object org.org-xwiki.org-xwiki-classloader.org-xwiki-classloader-internal.org-xwiki-classloader-internal-protocol.org-xwiki-classloader-internal-protocol-attachmentjar.org-xwiki-classloader-internal-protocol-attachmentjar-AttachmentURLStreamHandlerTest Nonsense ? Truly so. Yet on file system level Java developers usually do it, not only XWiki but many teams. Why ? Aren't directories given to suppress such redundancy ? Not only that give overly long unobservable paths, it also disables some programs on Windows. I heard that git has internal 4KB file path length limitations. I wonder what UNIX guys would say about sanity if one day they would try to download from Windows file with name about 16KB long... -- View this message in context: http://xwiki.475771.n2.nabble.com/sources-filename-why-that-redundancy-tp75… Sent from the XWiki- Users mailing list archive at Nabble.com.