measures/apps/CMakeLists.txt | 2 1 1 - 0 !
1 file changed, 1 insertion( ), 1 deletion(-)

 do not install test and demonstration executables

casacore_{assay,floatcheck} are removed and only used with scons.
measuresdata.csh and ms2uvfits should be included in casacore-doc as examples.
change measuresdata.csh shebang to /usr/bin/tcsh as in Debian.

doxygen.cfg | 6 3 3 - 0 !
1 file changed, 3 insertions( ), 3 deletions(-)

 disable class and collaboration graph generation

The graphviz-generated class and collaboration graph
occupies 1GB disk space.  It is too big to deal with.  We turn them
off only to save space.  Output image format is specified as svg to
further save space (bug #751947).

coordinates/Coordinates/test/CMakeLists.txt | 12 6 6 - 0 !
images/Images/test/CMakeLists.txt | 40 20 20 - 0 !
images/Regions/test/CMakeLists.txt | 8 4 4 - 0 !
lattices/LatticeMath/test/CMakeLists.txt | 2 1 1 - 0 !
meas/MeasUDF/test/CMakeLists.txt | 14 7 7 - 0 !
measures/Measures/test/CMakeLists.txt | 37 16 21 - 0 !
measures/TableMeasures/test/CMakeLists.txt | 2 1 1 - 0 !
ms/MSOper/test/CMakeLists.txt | 8 4 4 - 0 !
ms/MeasurementSets/test/CMakeLists.txt | 2 1 1 - 0 !
9 files changed, 60 insertions( ), 65 deletions(-)

 disable tests that require data tables

The data tables are not available during the build of casacore (yet).

casa/IO/test/CMakeLists.txt | 4 2 2 - 0 !
casa/OS/test/CMakeLists.txt | 9 4 5 - 0 !
images/Images/test/CMakeLists.txt | 2 1 1 - 0 !
lattices/LatticeMath/test/CMakeLists.txt | 2 1 1 - 0 !
scimath/Fitting/test/CMakeLists.txt | 4 2 2 - 0 !
tables/TaQL/test/CMakeLists.txt | 4 2 2 - 0 !
6 files changed, 12 insertions( ), 13 deletions(-)

 disable known test failures

https: //bugs.debian.org/cgi-bin/bugreport.cgi?bug=838424#15


These tests fail, probably due to the test itself. From
I'm not sure what to do about the other two tests that fail on arm64,
"tLSQaips" and "tLSQFit", because those two generate a load of output
text containing floating-point numbers and compare that text with an
expected output. It might not be so easy to make that work on all
architectures. Perhaps one could simply disable those two tests.
Preferably one would still run them, so that we can see what happens
in the logs, but not have the build fail if only those tests fail.

scimath/Mathematics/test/tFFTServer.cc | 2 1 1 - 0 !
scimath/Mathematics/test/tFFTServer2.cc | 2 1 1 - 0 !
2 files changed, 2 insertions( ), 2 deletions(-)

 loose some tests tfftserver tests

This shall help to compile for mips64el and hppa.

casa/OS/Directory.cc | 19 14 5 - 0 !
1 file changed, 14 insertions( ), 5 deletions(-)

 make the check for nfs a bit more portable (bsd)

And provide a fallback (f.e. for HURD).However, there is probably no
real use case for that, since also other file systems may be slow or
lack certain features.

casa/aipsenv.h | 2 1 1 - 0 !
1 file changed, 1 insertion( ), 1 deletion(-)

 enable hostinfo for kfreebsd

casa/OS/HostInfo.cc | 7 7 0 - 0 !
casa/OS/HostInfoHurd.h | 157 157 0 - 0 !
casa/OS/Path.cc | 6 4 2 - 0 !
casa/OS/malloc.cc | 2 1 1 - 0 !
casa/aipsenv.h | 7 7 0 - 0 !
mirlib/bug.c | 2 1 1 - 0 !
6 files changed, 177 insertions( ), 4 deletions(-)

 fix compilation for gnu hurd

However, the tests still fail.

python3/CMakeLists.txt | 2 1 1 - 0 !
1 file changed, 1 insertion( ), 1 deletion(-)

 fall back to older cmake functions to find python 3

This is a workaround that cmake will fail when the default Python
version is not the newest one. Reference: #945825

build-tools/casacore_floatcheck | 2 1 1 - 0 !
casa/apps/countcode | 2 1 1 - 0 !
casa/apps/plotmemory | 2 1 1 - 0 !
casa/apps/watchmemory | 2 1 1 - 0 !
python/Converters/test/tConvert.py | 2 1 1 - 0 !
5 files changed, 5 insertions( ), 5 deletions(-)

 use python3 in bang path of test utilities

msfits/MSFits/FitsIDItoMS.cc | 48 18 30 - 0 !
scimath/CMakeLists.txt | 2 0 2 - 0 !
scimath/Mathematics/FFTW.cc | 232 123 109 - 0 !
scimath/Mathematics/FFTW.h | 101 70 31 - 0 !
scimath_f/CMakeLists.txt | 2 0 2 - 0 !
5 files changed, 211 insertions( ), 174 deletions(-)

 use fftw instead of fftpack in fits-idi to ms table converter

The FitsIDItoMS class is the only class that was still dependent on the FFTPack
Fortran code. By rewriting it to use FFTW, the FFTPack code can be removed.
This required implementing a wrapper for the real even FFT in the FFTW class,
as that was not available.

The FFTW class should be rewritten to have a more sensible interface. I've added
some TODO comments for that. Despite the use of a mutex, the FFTW wrapper is
also not thread safe.

Closes: #957065

scimath_f/lawson.f | 4 2 2 - 0 !
1 file changed, 2 insertions( ), 2 deletions(-)

 fix type of dummy variable