Adam Wulkiewicz

Btw, my test program is more or less: ``` #include #include #include template inline void read_wkt(std::string const& wkt, Geometry & geometry) { namespace bg = boost::geometry; bg::read_wkt(wkt, geometry); bg::correct(geometry); }...

I've checked the case reported in this issue with your branch (`barendgehrels:bugfix/disable_start_turns`) but the result didn't change, still the same turns are generated. I'm going to find the exact commit...

@barendgehrels I found the cause. It seems my commit introduced this regression and your initial thought about rescaling was correct. Here is the commit: https://github.com/boostorg/geometry/commit/5ea1abcedc301bcacf6fcff157b4cd88184eab2f Reverting it fixes the problem...

Reverting https://github.com/boostorg/geometry/commit/5ea1abcedc301bcacf6fcff157b4cd88184eab2f causes the following `difference` cases to fail as it was before https://github.com/boostorg/geometry/pull/550 was merged: ``` algorithms_difference difference: mysql_23023665_6_b not valid Multi-polygon has intersecting interiors type: d difference: mysql_23023665_6_s...

The above means this issue is still on the table.

@barendgehrels I'm ok with leaving it as it is for now, with this case failing, this issue open (as a reminder) and waiting for the disabling of rescaling in the...

Just an update. This issue is in 1.72.

Hi and thanks for this proposal! I think that a function removing self-intersections should be a part of the library. AFAIK @barendgehrels was working on this and the result was...

One way would be to provide a "complete" solution, several ways of calculating this resulting valid polygon. Then the user would be able to choose the one that is the...

> And also sometimes multipolygons are drawn like one polygon: This means that for my second example the small triangle should be preserved because this might be a user drawing...