[Geotools-devel] Re: Maven 2.0.4 and 2.2.x

[Martin D. <mar...@no...>]

Cory Horner a =E9crit :
> Maven 2.0.4 is due out in a few days, and it seems to solve our issues=20
> with surefire (silent failures).  Snapshots are available:
> http://maven.zones.apache.org/~maven/builds/branches/maven-2.0.x/

I tend to believe that our current trouble with Maven 2 may be a Maven bu=
g (or if it is a mistake of=20
mine, at least it would be nice if Maven error message were improved). My=
 hope was that Maven 2.0.3=20
would fix the bug, and it didn't happen. If Maven 2.0.4 fix it, it would =
be very nice... I tend to=20
believe that a Maven issue is more likely than a dependency issue, becaus=
e in such case I would=20
expect Maven to fails at compile time or at test execution time (not a si=
lent failure before tests=20
are executed).



> I've added 2 dependencies to module/main/pom.xml, but still have 2=20
> exceptions hanging things up in main:
>=20
> java.lang.ClassCastException
>        at org.geotools.data.DataTestCase.dataSetUp(DataTestCase.java:11=
9)
>=20
> In this case, roadType.create(...) is returning null, and the cast to=20
> SimpleFeature doesn't like it.  The odd part, is that running this as a=
=20
> singular test "mvn -Dtest=3DFeatureTypeTest test" from /module/main pas=
ses=20
> without a hitch.

I noticed this exception a while ago and didn't had the time to fix it. M=
y current bet is that this=20
test case try to load (directly or indirectly) a file from the root packa=
ge using Class.getResource=20
or something equivalent. It sound like that loading resources at the root=
 package is not allowed in=20
Maven 2, so we may need to finish to move all test data to test-data subd=
irectory. I already moved=20
some of them (mostly the shapefile), but didn't finished.

Thanks for all yours Maven 2 effort!

     Martin.