Geometry Boost List: Re: [geometry] Running Geometry libs test suite.

Subject: Re: [geometry] Running Geometry libs test suite.
From: Nicolas Troncoso (ntroncos_at_[hidden])
Date: 2018-10-25 14:47:53

On Thu, Oct 25, 2018, 01:40 Mateusz Loskot via Geometry <
geometry_at_[hidden]> wrote:

> On Wed, 24 Oct 2018 at 23:14, Nicolas Troncoso <ntroncos_at_[hidden]> wrote:
> > On Wed, Oct 24, 2018 at 12:44 PM Mateusz Loskot via Geometry <
> geometry_at_[hidden]> wrote:
> >>
> >> It seems that DB-based approach may be easier for you than C++.
> >>
> >> I'd rather use PostGIS to validate geometries, especially regarding
> >> OGC compliance
> >>
> >
> > As you suggested I went on the DB route. These are my findings for
> future reference:
> >  I ran the same (or equivalent where applicable) query in:
> > MySql 5.6: Success
> > MySql 5.7.23: Failure
> > MySql 8.0.12: Success
> > Postgres server 10.5: Success
> > MSSQL 2016: Success
> >
> > I attach the query for the curious.
>
> Thanks.
>
> I think the results are enough to assume validity checks in MySQL 5.7
> are not reliable.
>
> I've checked your geometry with PostGIS 2.1. It's valid there too.
>

Yesterday I opened a bug report in MySQL

https://bugs.mysql.com/?id=92937

Thanks for the help.

Cheers.

>