|
+1 to that as well. Also, re: build times - We could make some of the tests not start up two instances of Spring/CXF. What exactly are we trying to test for when we do that? Client/Server should be completely isolated from each other even without that. With the exception of WS-RM/Addr I don't really see this as adding a lot of benefit. Maybe if we could establish what types of errors we're testing for here and then create a smaller set of test cases around that as opposed to making ever system test start two instances. Benson Margulies wrote: I'm all in favor of #1. In my perhaps unlucky experience, I keep running into 'standard' impls that are to quirky to reuse.On Mon, 2007-12-17 at 09:49 -0500, Daniel Kulp wrote: -- Dan Diephouse MuleSource http://mulesource.com | http://netzooid.com/blog |
- [IMPASSIONED PLEA] Naming and structuring test materials Benson Margulies
- Re: [IMPASSIONED PLEA] Naming and structuring test m... Fred Dushin
- Re: [IMPASSIONED PLEA] Naming and structuring te... Daniel Kulp
- Re: [IMPASSIONED PLEA] Naming and structurin... Benson Margulies
- Build times [was Re: [IMPASSIONED PLEA] ... Dan Diephouse
