Paul Ramsey

Having just spent a little time in the code, this is a bit harder than I thought, as I seem to have an assumption of unique source names, so that...

Looks like you have a winner, @dbaston

Sorry, why change a perfectly lovely pure C library to a C++ library?

Sounds OK in principle I suppose. We have our copy in-tree and probably won't change that. We managed to maintain a pure-C stance for a long time, though that eroded...

Reproduces for me under GEOS test. ``` diff --git a/tests/unit/geom/GeometryCollectionTest.cpp b/tests/unit/geom/GeometryCollectionTest.cpp index 41477bb2f..e42e7f358 100644 --- a/tests/unit/geom/GeometryCollectionTest.cpp +++ b/tests/unit/geom/GeometryCollectionTest.cpp @@ -181,4 +181,27 @@ void object::test ensure(gc->hasDimension(geos::geom::Dimension::A)); } + +// https://github.com/libgeos/geos/issues/982 +template...

The control runs all the way down into RelateOp, and the DE9IM matrix returned looks bad to me. ``` select st_relate('POINT(0 0)', 'GEOMETRYCOLLECTION(POINT(0 0), LINESTRING(0 0, 1 0))'); F0F FFF...

I'm going to close this as "finished".

I don't think we managed to convince ourselves this juice was worth the squeeze.

Going to close this absent any feedback. I cannot replicate on Apple Silicon and standard Apple build tools.

Is anything going to happen here? Requires agreement and a good PR.