Isn't -K a global option?

And where is that #define GMT_N_API_ENUMS 263 coming from that makes the Validator check to fail?

Isn't -K a global option?

No, see movie -K for instance. But, I agree there is no point using -K if there are other letters available, e.g., -F, -M . I will change it to -F instead.

If we add an enum we must run src/gmt_make_enum_dicts.sh to update that include file for the API. Idid I forget?

But what is the validator compares against to know the number of enums has changed? And what is right, the 261c or 263?

It looks fine. I test it with this:

gmt greenspline -R0/7.2/-0.2/7 -I0.05 -GMM1993.grd @Table_5_11.txt -Sr0.99 -Z1 -Vi -Kstatistics.txt -Emisfit.txt

Statistics.txt:

# Misfit & Variance Evaluation: Data	Model	Explained(%)	N	Mean	Std.dev	RMS
3.5767e+07	3.5767e+07	100	52	-5.24708e-14	7.94742e-14	9.45932e-14

Small Issue

The creation of that file depends on -E. So, if I use the command below, the file is not generated.

gmt greenspline -R0/7.2/-0.2/7 -I0.05 -GMM1993.grd @Table_5_11.txt -Sr0.99 -Z1 -Vi -Kstatistics2.txt

So, maybe should be clarify in the docs? Or maybe we should a modifier to -E instead of a new letter?

Sorry, forget. Just pushed it to master now. It is 263.

Will switch from -K to -F and will update docs etc to say -E is needed.

Switched to -F, requires -E, and -F must set a file since -E has an optional file and will write to stdout. Cannot both write to stdout.

@Esteban82, please update build and check again since there were several changes.

I got an error:

 gmt greenspline -R0/7.2/-0.2/7 -I0.05 -GMM1993.grd @Table_5_11.txt -Sr0.99 -Z1  -Vi -Emisfit.txt  -Fstats.txt
greenspline [ERROR]: Option -r: Syntax is -r[g|p]
greenspline [ERROR]: Option -r parsing failure. Correct syntax:

  -r[g|p]
     Set (g)ridline- or (p)ixel-registration (-r with no argument sets pixel registration). If -r is not given then we default to gridline registration.
greenspline [ERROR]: Offending option -rstats.txt

Me too! Will fix...

Fixed.

OOPS: More changes:

I will instead use option letter -M. The reason is two-fold:

1. The ancient -F for -r was parsed in greenspline so I took that out - not the best solution.
2. gpsgridder should get exactly the same option, but its -F is taken. -M is available in both modules

Ok. BTW I will be away from my PC for the following days.

OK, the new feature is now set via -E[misfitfile][+rreportfile]. The benefits are

1. No need to messing with complicated -F form the olden days.
2. No need to add a new option, just a modifier
3. This also means we would not have to set -E in addition to the initial -K or -M to do this since it is now a modifier of -E so -E is obviously set.
4. green spline actually has an unimplemented -M for passing in previous eigenvalues. May happen at some point and then it is nice that -M is not taken already.

Give @Esteban82 is off, perhaps @joa-quim can have a look. Should be harmless since only about a report to a file.

Fixed synopsis message and a few issue related to freeing files that were forgotten in gpsgridder and green spline. gpsgridder now has the same -E modifier etc.