PROJ/src, branch 6.3.1

Update version and libtool number to 6.3.1

2020-02-10T09:34:57+00:00

Fix numerical precision issues in vandg and robin

2020-02-08T09:50:32+00:00

Refs #1906
Fix remaining issues of https://github.com/OSGeo/PROJ/issues/1906#issuecomment-583168348
as found with gcc 8.2.0 -m32 -O2

Fix test issues on i386

2020-02-08T09:50:32+00:00

Fix a few issues of #1906 found when running the test suite
on Ubuntu 16.04 with gcc 5.5 -m32. When applied on top of
the fix of #1912, make check succeeds

cart: Avoid discontinuity at poles in the inverse case

2020-02-08T09:50:31+00:00

This should avoid issues with numerical stability as uncovered in
https://github.com/OSGeo/PROJ/issues/1906.

Practically speaking this change isn't going to affect real life
scenarios since the position of the center of the Earth is rarely
expressed in geodetic coordinates.

Fix identification of ESRI-style datum names starting with D_ but without alias

2020-02-06T13:18:39+00:00

Fixes #1911

Fix performance issue, affecting projinfo EPSG:7842

2020-02-06T12:55:06+00:00

Fixes #1913

AuthorityFactory::createBetweenGeodeticCRSWithDatumBasedIntermediates() issued
a complex SQL query that pushes the SQLite3 query plan optimizer to its limits.
Was working reasonably with sqlite 3.11, but not with later versions.
So put less constraints in the main query and do post-processing checks and
auxiliary requests to avoid such issues.

For some unknown reason, this slightly slows down a bit execution time of the
whole test_cpp_api binary (~ 10%), but couldn't come with something better,
despite trying many variations of the main SQL query. It seems that in the
general case the non-filter LEFT JOIN on the supersession table helped,
except on this EPSG:7842 case.

Fix ingestion of +proj=cea with +k_0

2020-02-06T12:45:17+00:00

Fixes #1881

Digging into the implementation of proj=cea, it appears that
k_0 and lat_ts are intended to be exclusive ways of specifying the
same concept. EPSG only models the variant using lat_s.
So if k_0 is found and lat_ts is absent, compute the equivalent
value of lat_ts from k_0.

Note: k_0 should normally be in the [0,1] range. In case creative users
would use something outside, we raise an exception, even if the cea
implementation could potentially deal with any k_0 value. Hopefully
this is a (reasonable) limitation that will address nominal use cases.

Merge pull request #1875 from rouault/backport_fix_wrong_use_of_derivingConversion

2020-01-23T22:26:26+00:00

[6.3 backport] Fix wrong use of derivingConversionRef() that caused GDAL bug

Fix wrong use of derivingConversionRef() that caused GDAL bug

2020-01-23T18:11:05+00:00

Hopefully final cut at solving the same class of bug that the one
that affected QGIS in December.

This time, this hit GDAL in the situation of
https://lists.osgeo.org/pipermail/gdal-dev/2020-January/051500.html
The bug fix for that particular issue is in
PROJStringParser::createFromPROJString()

But grepping more in the code base, I could find other potential smelly
situations (might not be issues, but better be safe than sorry),
so let's fix them too.

Bottom line is:
derivingConversionRef() should *only* be used for consultation, and
never to create a new ProjectedCRS()

Database: add a geoid_like value for proj_method column of grid_alternatives, fix related entries and simplify/robustify logic to deal with EPSG 'Geographic3D to GravityRelatedHeight' methods

2020-01-23T18:05:50+00:00