Stefan Bodewig wrote: > On Wed, 09 Oct 2002, Christopher Lenz <[EMAIL PROTECTED]> wrote: > > >>This looks like an error with a test case not in sync with API >>changes in the client lib, right? > > > You are keeping some oldish copy of the commons-httpclient code in > slide's CVS module and compile against it, so you don't get this error > when you work with your local installation of Slide.
... because we needed a stable httpclient-lib, which commons-httpclient hasn't been providing for quite some time... but that's another story ;-) > Gump forces Ant to use the httpclient code that is on its CLASSPATH, > which is the latest CVS version. Wow, I wasn't aware of Gump trying to use commons-httpclient, even though we don't declare it as a dependancy (or do we? we shouldn't). In that case keeping the code in our tree is complete non-sense. > This problem has now been present for several months. The days > without a Gump nagging mail have just been days where slide hasn't > been built at all because of failed dependencies (usually Tomcat-4). > > On a side not, commons-httpclient is rather close to release version > 2.0. If its API had changed in a way that makes it no longer suitable > for Slide, I bet the commons-httpclient committers would want to know > before they do the final release. Detecting such problems is one of > Gump's main points. We've been through that long ago, with the effect that Remy re-imported the code (which originally came from Slide) into our CVS. Anyway, moving back to commons-httpclient has just become pretty urgent, I guess ;-) Thanks for the clarifications, Stefan. -- Christopher Lenz /=/ cmlenz at gmx.de -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
