`gmt basemap -R... -J... -A` fails as a modern-mode one-liner

[mkononets]

gmt basemap -R... -J... -A works in a multi-call modern mode session, works as a classic mode call but fails as a modern mode one-liner with weird error messages, see below. Tried with gmt version 6.6.0.

from the docs:

-A[file]

    No plotting is performed. Instead, we determine the geographical coordinates of 
the polygon outline for the (possibly oblique) rectangular map domain. The plot domain 
must be given via [-R](https://docs.generic-mapping-tools.org/dev/basemap.html#r) and
 [-J](https://docs.generic-mapping-tools.org/dev/basemap.html#j), with no other options 
allowed. The sampling interval is controlled via [MAP_LINE_STEP](https://docs.generic-mapping-tools.org/dev/gmt.conf.html#term-MAP_LINE_STEP) 
parameter. The coordinates are written to file or to standard output if no file is specified.

gmt basemap -JS0/-90/20c -R3000+uk -Aout.txt fails with the following error message with and without out.txt requested after -A:

gmt [ERROR]: Shared GMT module not found: basemap

ERROR: No module named basemap was found.  This could mean one of four cases:
  1. There actually is no such module; please check your spelling.
  2. You used a modern mode module name while running in GMT classic mode.
  3. Module exists in the GMT supplemental library, but the library could not be found.
  4. Module exists in a GMT custom library, but none was specified via GMT_CUSTOM_LIBS.
Shared libraries must be in standard system paths or set via environmental parameter LD_LIBRARY_PATH.

gmt basemap -JS0/-90/20c -R3000+uk -Aout.txt -png quick also fails:

psbasemap [ERROR]: Option -p (ng): Syntax is -p[x|y|z]<azim>[/<elev>[/<zlevel>]][+w<lon0>/<lat0>[/<z0>]][+v<x0>/<y0>]
psbasemap [ERROR]: Option -p parsing failure. Correct syntax:

  -p[x|y|z]<azim>[/<elev>[/<zlevel>]][+w<lon0>/<lat0>[/<z0>]][+v<x0>/<y0>]
     Select a 3-D pseudo perspective view. Append the <azimuth>/<elevation> of
     the viewpoint [180/90], and when used with -Jz|Z, optionally add /<zlevel>
     for basemap level [bottom of z-axis]. Prepend x or y to plot against the
     "wall" x = level or y = level [z]. For a plain rotation about the z-
     axis, give rotation angle only and optionally use modifiers +w or +v to
     select location of axis:
       +w Specify a fixed coordinate point [region center].
       +v Set a fixed projected point [panel center].
psbasemap [ERROR]: Offending option -png

gmt basemap -JS0/-90/20c -R3000+uk -Aout.txt -png quick -B terminates w/o errors but produces neither the -A output nor the plot.

A classic mode call gmt psbasemap -JS0/-90/20c -R3000+uk -A produces the desired output.

[Esteban82]

I tested this and can confirm that it fails, but I’m not sure this should be considered a bug.

The one-liner syntax is meant for commands that produce a figure, while basemap -A explicitly performs no plotting. Also, a valid one-liner requires the -format prefix option, which is missing here—and would not make sense to add in this case.

So this seems consistent with the documented behavior rather than a bug.

[joa-quim]

Agree, and for that reason I guess we could find more cases involving other modules showing the same behavior.

[mkononets]

this seems consistent with the documented behavior

In case my long and detailed report was not really clear:

The documented behavior is literally this:

-A[file]

    No plotting is performed. Instead, we determine the geographical coordinates of 
the polygon outline for the (possibly oblique) rectangular map domain. The plot domain 
must be given via -R and -J, with no other options allowed... The coordinates are written 
to file or to standard output if no file is specified.

When I call gmt basemap -JS0/-90/20c -R3000+uk -Aout.txt outside session the output file promised by the docs is not created. Instead the following error message is displayed:

gmt [ERROR]: Shared GMT module not found: basemap

ERROR: No module named basemap was found.  This could mean one of four cases:
  1. There actually is no such module; please check your spelling.
  2. You used a modern mode module name while running in GMT classic mode.
  3. Module exists in the GMT supplemental library, but the library could not be found.
  4. Module exists in a GMT custom library, but none was specified via GMT_CUSTOM_LIBS.
Shared libraries must be in standard system paths or set via environmental parameter LD_LIBRARY_PATH.

The expected result is not the above message, but the out.txt file with the coordinates.

I guess we could find more cases involving other modules showing the same behavior.

Not sure how to interpret this. Is it supposed to be like that?

[joa-quim]

The modern mode is the result of a group epiphany while looking at the beautiful La Jolla Bay from a room in Scripts. The intent was to relieve newbies from learning the -K -O dance. Latter, dropping the need for at least -R, -J, -B was added too. And even latter, mostly from pressure from me, Paul added the one-liners. But all of this add in mind figure producing commands, not the data processing ones. What you found is a side effect of a plot creating command, which has a little back-door that creates data. Is it a bug because it prints ERROR: No module named basemap was found.? Maybe, but using data processing commands with modern mode was never the idea behind it.

[mkononets]

Oh yes I remember being that newbie learning gmt version 4: one -K -O typo somewhere in the middle of a long plotting script and the resulting .ps could not have been viewed.
That -K -O thing is definitely not good for people manually writing scripts and reusing old pieces of code by copy-pasting.