|
From: Vadim Z. <vz...@ze...> - 2011-12-11 16:35:04
|
On Fri, 09 Dec 2011 20:40:22 +0000 William S Fulton <ws...@fu...> wrote: WSF> Unfortunately I can see that something is broken, but am not quite sure WSF> how you've managed it. By misunderstanding the original code purpose. I hadn't realized that the value we're inheriting could already exist for the derived class (this is not normally the case but can happen if it's set explicitly, as minherit2 test does). And because of this I didn't check for the value existence in the correct hash. This is fixed now in r12874. WSF> What I normally do with changes in the core like WSF> this is to run the partialcheck make targets, eg 'make WSF> partialcheck-java-test-suite' Unfortunately I still can't run the test suite for Java because of the problem with doing it from a build directory that is different from source directory, see this (very old) thread: http://thread.gmane.org/gmane.comp.programming.swig.devel/20043/ As I wrote there, top_srcdir value in Makefile is wrong in this case. But I did test minherit2, contract and director_frob manually and they all generate correct (i.e. the same as before) code now. Sorry again for the breakage and thanks a lot for noticing it! VZ |