On Thu, 16 Jan 2003, Sam Ruby <[EMAIL PROTECTED]> wrote: > +1! Both to the sentiment, and to the code. ;-)
Cool. Something that I find a bit ugly is the mangled name, but I couldn't come up with something easier that was unique and human identifiable at the same time. And then I'm not sure whether it could cause any filename or command line length issues as these names tend to be rather long. But the whole mangling code is in a single place, one wouldn't have to change anything but the getCachedNodeName method to use a different algorithm. > I ran a test on windows... appeared to cache the files it could > access, and not die because there were ones that it could not. Great. > Note: this should work on the build machines... If you place the two files from my home dir at icarus (see other mail) into the cache directory, cocoon and a couple of other projects will no longer get dropped. > but one thing I'd like to note: the following step is always done > from a clean checkout, so it will not benefit from the cache: > > http://cvs.apache.org/builds/gump/latest/gump.html > > As this step is for verification purposes, this actually is a good > thing. I agree. Stefan -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
