Issue 53 in pencil-code: Reading start.in, print.in and run.in does not work on all systems

[penci...@googlecode.com]

Status: Accepted
Owner: Bourdin.KIS
CC: ajo...@strw.leidenuniv.nl
Labels: Type-Defect Priority-High

New issue 53 by ajo...@strw.leidenuniv.nl: Reading start.in, print.in and
run.in does not work on all systems
http://code.google.com/p/pencil-code/issues/detail?id=53

The parallel_open and parallel_close way to read start.in etc. does not
work on all systems. On the Milleotto system in Lund the code insists that
there is a problem in an arbitrary name list (in our case &hydro_run_pars).
The problem could be fixed by downdating register.f90 and param_io.f90 to
v. 13670, before the new scheme was introduced.

Has anyone else encountered a similar problem?

Should we have an option for choosing between the new and the old scheme
for reading namelists?

[penci...@googlecode.com]

Comment #1 on issue 53 by ajo...@strw.leidenuniv.nl: Reading start.in,

print.in and run.in does not work on all systems
http://code.google.com/p/pencil-code/issues/detail?id=53

We managed to solve the problem today by compiling the syscalls_ansi.c in
32 bit. Somehow Fortran 90 was compiled in 32 bit and C in 64 bit. The
troubling thing is that the linker would complai, but not stop when linking
64 bit files to the mainly 32 bit created by Fortran.

[penci...@googlecode.com]

Updates:
Status: WontFix
Labels: -Type-Defect Type-Other

Comment #2 on issue 53 by Bourdin.KIS: Reading start.in, print.in and

run.in does not work on all systems
http://code.google.com/p/pencil-code/issues/detail?id=53

Since this was resolved without changes in the source, "WontFix"
and "Type-Other" is the right choice, here.