For me, it's OK to do another RC.

Peter

Am 21.02.2018 um 16:48 schrieb Marshall Schor:
Hi Richard,

Once again, thanks for your testing :-)

Well,  it looks like I fixed the setters and getters for the various non-FS
types, but forgot to include a fix for the feature structure value type.
This will need a new fix.

I'm happy to do another RC, but if others think this is not worth delaying the
release, then we can add this for the next official release.
Personally, I'm kind of a perfectionist, so would like to fix this and do
another RC, but I'm aware that others spend some time doing testing of each 
RC...
Please weigh in (if you think it's OK to have another RC, or if you think this
is too minor, and ought to be shifted to the next release).

-Marshall

On 2/19/2018 5:05 PM, Richard Eckart de Castilho wrote:
Compared SVN tag to source release: OK
Build from source release with empty repo: OK
Spot checked release notes: OK
Spot checked issues list: OK
Spot checked signature: OK
Built uimaFIT 3.0.0-SNAPSHOT against the RC: OK
Built DKPro Core 2.0.0-SNAPSHOT against the RC: OK
- Actually I had a strange effect that enabling V2_PRETTY_PRINT did not work
   in two specific cases, but I assume the cause does not lie in the RC

Testing (again) with WebAnno...

Looks like methods like 
org.apache.uima.cas.impl.FeatureStructureImplC.getFeatureValue(Feature)
(still?) do not check if the returned value actually is valid for the method. 
E.g.
I can call this method for "Annotation:end" and no exception is thrown despite
this being an integer feature and not a FS feature.

   @Override
   public TOP getFeatureValue(Feature feat) {
     if (IS_ENABLE_RUNTIME_FEATURE_VALIDATION) featureValidation(feat);
     return _getFeatureValueNc((FeatureImpl) feat);
   }

According to the documentation, a CASRuntimeException should be thrown:
  * If there is a typing violation, i.e., if <code>feat</code> is not defined 
for
  * the type of this FS, or the range type of <code>feat</code> is Float, 
Integer or
  * String.

I think we came across this earlier... so is it intentional or forgotten?

Cheers,

-- Richard

On 05.02.2018, at 22:28, Marshall Schor <m...@schor.com> wrote:

Hi,

Here's rc4 of uimaj 3.0.0 sdk.

The change since rc3: some fixes to the logging APIs to avoid "double scanning"
of substitution when using logrb style calls, some formatting of the messages
fixes for logging, and some caching added for internationalization of messages
using message bundles.

-----------------

(Copied from previous vote msg):

The change since rc2
   - fix for the class path in the uimaj examples eclipse project, to correspond
to the upgrades done for slf4j and jackson-core.

Change since rc1:
   - bug when JCas range is unknown
   - exception thrown unnecessarily, making Component Descriptor Editor (CDE)
fail on bad JCas classes
   - changed CDE to bypass loading user-defined JCas classes.

Changes since the beta are mainly 2 (plus some bug fixes):
   1) support for serialization / deserialization and subsequent access using
LowLeveCAS.getFSFromRef(int) API.
   2) support for applications which have JCas class classes, and load multiple
type systems (under the same class loader) with some of these having subsets of
the features.

Also, added 2 more semi-built-in types: one is a variant of the FSHashSet, using
linked Hash Set, so iterating gives a reasonable order. The other is a general
map from ints to Feature Structures, intended to let users eventually replace
use of the LowLevelCAS API method referenced above.
-----------------------

The Jiras for this release are here:

https://issues.apache.org/jira/issues/?jql=project%20%3D%2012310570%20AND%20fixVersion%20%3D%2012341578%20ORDER%20BY%20priority%20DESC%2C%20key%20ASC

Maven artifacts:
https://repository.apache.org/content/repositories/orgapacheuima-1183/

Source and binary zip/tar staged to:
https://dist.apache.org/repos/dist/dev/uima/uimaj/3.0.0-rc4/artifacts/

Eclipse update subsite:
https://dist.apache.org/repos/dist/dev/uima/uimaj/3.0.0-rc4/uimaj-uv3

SVN tag: https://svn.apache.org/repos/asf/uima/uv3/uimaj-v3/tags/uimaj-3.0.0/

Please vote on release:

[ ] +1 OK to release
[ ] 0 Don't care
[ ] -1 Not OK to release, because ...

Thanks.

-Marshall


--
Peter Klügl
R&D Text Mining/Machine Learning

Averbis GmbH
Tennenbacher Str. 11
79106 Freiburg
Germany

Fon: +49 761 708 394 0
Fax: +49 761 708 394 10
Email: peter.klu...@averbis.com
Web: https://averbis.com

Headquarters: Freiburg im Breisgau
Register Court: Amtsgericht Freiburg im Breisgau, HRB 701080
Managing Directors: Dr. med. Philipp Daumke, Dr. Kornél Markó

Reply via email to