We routinely run rubyspec via mspec on a number of OS combos. We also run against Java 5/Java 6. We do not run any tests on OpenVMS as we don't have an instance we can use.
I suspect something more basic with JRuby is broken on OpenVMS, which you have uncovered. If you clone JRuby source can you try 'ant test'. This loads some of our older unit tests and it also may shed more light on what is wrong... -Tom On Fri, Mar 25, 2011 at 6:59 AM, BG - Ben Armstrong <[email protected]> wrote: > I'm trying to get an idea of how well JRuby 1.6.0 works on OpenVMS/Alpha > (Java 1.5.0) and OpenVMS/IA64 (Java 1.6.0), observing which areas have > deficiencies we would need to address in order to port our application to > run on it. To this end, I have checked out MSpec and RubySpec both from git > and have mirrored these to our test systems to run them there. This poses > some challenges on OpenVMS/IA64 in particular (but oddly, not > OpenVMS/Alpha,) as mspec breaks completely after an initial broken test is > encountered (i.e. after one "E", I get a whole string of subsequent "E" > results). My workaround is to break up my mspec runs into smaller runs so > that no more than one "E" occurs per test run, but this takes a very long > time, as startup time on these platforms is poor. I currently have broken > down my single "mspec core" run into 348 invocations of mspec, stretching > out what would otherwise take only a few minutes into a run that takes the > better part of an hour to complete.*** > > So I have two questions: First, does the JRuby project use RubySpec? Second, > if so, has anything like this been encountered before? I would like to > resolve this issue, but may need some assistance in doing so. > > Thanks, > Ben > *** Incidentally, nailgun compiles fine on this platform (untested yet, > though,) so we might be able to improve on this, but obviously a fix to > MSpec itself would be more desirable. > > > --------------------------------------------------------------------- > To unsubscribe from this list, please visit: > > http://xircles.codehaus.org/manage_email > > > -- blog: http://blog.enebo.com twitter: tom_enebo mail: [email protected] --------------------------------------------------------------------- To unsubscribe from this list, please visit: http://xircles.codehaus.org/manage_email
