Hi everyone, First of all, thanks to everyone for testing :-)
I have a rather big project where I'm testing uima v2 vs v3, and comparing the results, and saw a very occasional difference, which I finally tracked down: it's a bug in the Subiterator implementation for some edge cases. So, I want to fix that, add some test cases, etc. I'll also take Richard's good advice and change to 3.1.0 and mention the change done to align the code with the Javadocs (they were inconsistent), as described in UIMA-5869. -Marshall On 3/25/2019 7:48 AM, Richard Eckart de Castilho wrote: > So, RC builds good and I have tried it out in the context of several of my > projects > (some of which I first had to update since their v3 branches were seriously > outdated). > Except the issue mentioned below, all seems to be fine. > > I think the RC would better be released as v3.1.0 because of the switch from > CASRuntimeException to CASException. Also, I think this changed should be made > prominent in the release notes under "Specific changes in release 3.0.2 > versus 3.0.1 include". > > But anyway, if you want to release as-is, I'm +1 with that as well. > > Cheers, > > -- Richard > >> On 20. Mar 2019, at 14:24, Richard Eckart de Castilho <r...@apache.org> >> wrote: >> >> I'm having a look. >> >> So far, I have: >> - compiled the UIMA Core 3.0.2 code ok >> - compiled the uimaFIT 3.0.0 branch against the 3.0.2 RC - kind of ok >> There are uimaFIT unit tests failing which expect a CASRuntimeException >> when a view >> is missing because this has been changed to generating a CASException. I >> think this >> change should cause the version to be increased in the mid-position (i.e. >> 3.1.0). >> - now trying to compile the v3 branches of DKPro Core and WebAnno against >> the RC... >> >> Cheers, >> >> -- Richard >
