UnaryUnion of a non-simple linestring doesn't node it

[strk]

See ​http://postgis.refractions.net/pipermail/postgis-devel/2011-October/015392.html

I haven't tested 3.4dev and 3.2 branches, 3.3 branch is affected.

[strk]

NOTE: Martin added a test for this in TestUnaryUnion.xml and should add another in UnaryUnionTest.java. To be ported.

[Mike Taves]

I assume an example of this bug is with LINESTRING (50 50, 150 150, 50 150, 150 50), which GEOS="3.3.1-CAPI-1.7.1" returns the exact same non-simple geometry while JTS returns a simple geometry

MULTILINESTRING ((50 50, 100 100), 
  (100 100, 150 150, 50 150, 100 100), 
  (100 100, 150 50))

[strk]

I'm working on this

[strk]

It looks like this bug was introduced by the fix for #392

[strk]

r3513 fixes this w/out loosing the robustness improvement introduced for #392.

Martin, if you're reading this, I suggest you take a look for JTS as well...

[strk]

for the record, r3514 forward-ports this to trunk