rads2nc produces segmentation fault

[GoogleCodeExporter]

rads2nc produces segmentation fault

[GoogleCodeExporter]

This only occurs without specifying any variables to output.
With -Vsla, for example, it works fine.
What would be the best approach?
1. Mimic how rads2asc works, i.e. have time_rel_eq,lat,lon implicitly (which I 
actually don't like)
2. Make sure that rads2nc complains when no variables are specified.
Either way, something needs to be changed.

En passant, can you give a suggestion about what to do in rads2asc?
1. Continue the way rads2asc works now, i.e. have time_rel_eq,lat,lon implicitly
2. Only output variables specified by -V.

Original comment by [email protected] on 27 Sep 2013 at 8:33

• Changed state: Accepted

[GoogleCodeExporter]

For rads2nc I'd recommend option 2. I'm not certain that there's a clear 
default -V that users should expect, so a prompt for -V is probably best.

As for rads2asc, my personal preference would be for -V to imply -f. But, I can 
see that is nonstandard and not backwards compatible. So, I'd lean toward 
recommending 2). I don't use time_rel_eq much, so the default often isn't 
useful to me.

Original comment by [email protected] on 27 Sep 2013 at 2:22

[GoogleCodeExporter]

In r552: rads2nc now required at least one variable after -V (or --var= or 
--sel= or sel=)
In r551: rads2asc now also requires at least one variable after -V (or --var= 
or --sel= or sel=). Also -V (or --var) imply -f. Only sel= without -f will 
issue the extra time,lat,lon columns (for backward compatibility).

Original comment by [email protected] on 30 Sep 2013 at 3:01

• Changed state: Fixed

[GoogleCodeExporter]

Original comment by [email protected] on 19 Dec 2013 at 4:35

• Changed state: Verified