$include_dir="/home/hyper-archives/geometry/include"; include("$include_dir/msg-header.inc") ?>
Subject: Re: [geometry] Extensions maintenance (was: Setting up Travis CI for Boost.Geometry)
From: Adam Wulkiewicz (adam.wulkiewicz_at_[hidden])
Date: 2014-11-10 21:10:43
Hi,
Barend Gehrels wrote:
>
> Some are "soon" to be released (geographic strategies, projections) 
> and should stay there.
>
> Same for some formats (WKB)
>
> Some are quite often used by others (such as dissolve) and should stay 
> there, and indeed we can think about releasing them too.
>
> Some others are not actively worked on (such as sphere).
>
Actually I'm keeping the nsphere green, though indeed many algorithms 
aren't implemented for this Geometry. However some time ago I 
implemented a subset allowing to store nspheres in the rtree.
FYI, I'm planning to propose to use similar interface for 
ellipsoid/sphere models we were talking about, to use core/radius as a 
part of the concept. So this could be released as well.
>
> They are in extensions because they are not released. They are not in 
> the master, so they are not in the main repository.
>
> With this location, we have automatic access to the regression matrix 
> too.
>
> Especially for the last reason, I don't think they should be moved. 
> But I agree that they should be made "green" and kept "green". 
I agree, I took care of some of them already (nsphere). Most of them are 
failing becasue recent changes in distance and robustness upgrades. But 
dissolve() is failing since I can remember.
Regards,
Adam