First I run the c version of the Sample_mpio.c and I suppose it worked:
Proc 2: all tests passed
Proc 3: all tests passed
Proc 1: all tests passed
Proc 0: all tests passed
"SIG_SETMASK undeclared ... #define HDsigsetjmp(J,N) sigsetjmp(J,N)
... in expansion of macro ..."
error.
Since there are no separate instructions for cygwin, like for the
serial case, I thought it might be an issue of cygwin. Does anyone
know how to fix this?
One last question about the installation is szip. I built from source
into the home folder but I have not changed any path variables. Will
hdf be able to find it or is it not important?
First I run the c version of the Sample_mpio.c and I suppose it worked:
Proc 2: all tests passed
Proc 3: all tests passed
Proc 1: all tests passed
Proc 0: all tests passed
"SIG_SETMASK undeclared ... #define HDsigsetjmp(J,N) sigsetjmp(J,N) ... in expansion of macro ..."
error.
the problem is not due to the --enable-parallel
Trying to rebuild the serial 1.8.18 I see the same issue,
so it is likely a consequence of changes in the cygwing headers.
Since there are no separate instructions for cygwin, like for the serial case, I thought it might be an issue of cygwin. Does anyone know how to fix this?
I will look on that, but it could be in September
One last question about the installation is szip. I built from source into the home folder but I have not changed any path variables. Will hdf be able to find it or is it not important?
Thank you for the helpful comments. I also contacted the hdf support and
apparently parallel hdf is not supported or tested for cygwin, therefore I
expect that there might be more problems on the way.
Anyhow, I used make CFLAGS=-D_POSIX_C_SOURCE and this error does not appear
anymore. However, "make check" gives a segmentation fault error for
testhdf5.exe.
I have also opened a thread in stackoverflow
<https://stackoverflow.com/questions/45494332/building-parallel-hdf5-with-cygwin>,
if anyone would like to contribute.
