[postgis-devel] [postgis-tickets] [SCM] PostGIS branch master updated. 3.3.0rc2-128-g1bd0a8410

[Sandro Santilli]

On Mon, Sep 26, 2022 at 01:47:46PM -0700, git at osgeo.org wrote:
> commit 1bd0a84100ed511a0466d24b45ccb62d5e2b4b4c
> Author: Regina Obe <lr at pcorp.us>
> Date:   Mon Sep 26 16:47:05 2022 -0400
> 
>     Remove a relate_bnr that results in different answer in patched GEOS.  References #5245 for PostGIS 3.4.0dev

Everytime a test is removed from the testsuite I feel pain.

Rather than hiding the bug under a carpet, what about letting
the testsuite do it's job and show RED (failure) if someone
is using a GEOS version which gives a BOGUS result ?

Bugs should be fixed in patch-level releases of GEOS, maybe
let's file a GEOS ticket to request such backporting ?

I do understand the rationale of wanting a green CI but I'm afraid
unless we have a blocker ticket requesting to re-enable the test
the result will just be that a test is removed and never gets back,
as I think happened for other tests.

--strk;