Dissonant documentation of ExtractPolygonalPrismData

[CMun]

Hello,

the "bug fix for the corrected output of 2D Convex Hulls" from @aichim leads to a change in behaviour from pcl 1.6 to pcl 1.7.

More specific, one could specify in PCL 1.6 a list of n points on a plane, defining a convex hull (without using the pcl::ConvexHull class). This list described an implicitly closed polygon, so the first point was not equal to the last point.

After @aichim 's bug fix, however, that doesn't work anymore. See the notes for the commit for further details. However, if you create the convex hull by specifying n+1 points (the first beeing identical to the last), the current code works again.

However, it is then not clear, what good the "// And a last check for the polygon line formed by the last and the first points..." code change (duplication, possibly superfluous) brought.

If it is really neccessary (e.g. for cooperation with the convex hull class), than the documentation should be updated to specifically mention the "first = last" precondition.

(my use case is millions of calls for dynamically changing 12 points forming a circle -> creating a cylinder for point extraction, so the overhead for calling pcl::ConvexHull is not negligible).

Thanks for consideration,
Chris

[jspricke]

@aichim can you please comment on this? If it's a regression, I would propose to simply revert the commit.

[taketwo]

I was told recently that the code in one of my past projects started to fail after updating to PCL 1.7, and it was quite hard to trace down the problem, which turned out to be exactly this introduced first = last assumption.

I think we have to revert back to implicitly closing polygons.

[jspricke]

done in moster.