So, it seems like DEBUG/RELEASE flags in CMakeLists.txt do enable QT_NO_DEBUG, but QT_NO_DEBUG no longer has any effect on the printout of error messages.  I therefore turned OFF the m_deepDebug flag in ReadGeoModel.  I also check this flag before printing out anything.  In other cases I removed output to cout, or to qDebug (which will go, anyway, when we are no longer using Qt5 in this package. This cured a situation in which pages and pages of output were generated, and actually went a little further by suppressing some of the normal output which was already too verbose. Now the level of output should be almost acceptable to users.

Add the SpecialShapes package to GeoModelIO

See merge request !2

Re-introduce CxxUtils namespace to fix the failure of the CI pipeline related to slc6 and cc7 builds

This reverts commit adb491f4.

This reverts commit 590f1377.

This reverts commit 050d85a4.

fixes #7, the issue with files where a table is missing.  Tables were tested with AND statements, so if a single table was missing, the DBManager started creating all tables. Now, we use OR clauses and the even if only one table is present (which should not occur, however) the file is considered non-empty and the creation of the tables is skipped

CMake Cleanup, master branch (2019.10.30.)

See merge request !1

GeoModelWrite was not using this module anyway. (It does not depend on
Eigen directly.)

In exactly the same way as for GeoModelCore. It should be fine-tuned
later on...

Now that the base project uses the "GeoModelCore::" namespace on its
exported targets...

Fixed the name of the exported version file, so that CMake would
actually take it into account.

Also updated the main project description file to ask for an explicit
(minimum) version of GeoModelCore, and fixed the name of a variable
in the file.

Made the same sort of updates as in GeoModelCore. Though things had to be
done in an even more complicated way in this project.