GWM is now in. Couple of notes:
- GWM_ERR actually had an error in it. "PARAMTER" instead of PARAMETER.
This implies that noone is writing fortran code that includes this
file!
- Since gwm_err.h is made up of both messgen output and internal
C error codes, how do I deal with that? Currently messgen will
overwrite the file. There is some comment in the header file
that ems is not used for those constants because it should be
portable, but as far as I can tell it doesn't really matter what
values those error codes have so why is there a problem in
generating a load of new EMS error codes for those error conditions?
Everything should still work (since that's the point). It's not
like using different constants will suddenly result in new dependencies
(and will it matter to fortran code that new internal codes are now
available).
- make check fails if you are building from scratch and don't have
"." in your path (since GWM needs gwmXrefresh in your path)
- Should we be using starX_link in the gwm_link script or should
I simply use gwm_link.in and insert the equivalent directly?
PDA is also now in but there are some subtle changes in the test output.
Can someone take a look please? [I did update pda_d1mach and i1mach to
newer automatic versions but they return the same values as the originals]
--
Tim Jenness
JAC software
http://www.jach.hawaii.edu/~timj
|