GMT

Beware of dubious hyphens.

I'm using GMT 6.0.0_r20280 provided by conda-forge (https://github.com/conda-forge/gmt-feedstock).

The following two commands are expected to be equivalent for GMT6. However, the classic one succeeds and the modern one fails.

$ gmt --version
6.0.0_r20280

$ gmt psbasemap -R69:30/71:45/-17/-15:15 -Jt70/1:1000000 -Bafg -B+t"Survey area" > test.ps

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/1:1000000 -Bafg -B+t"Survey area" -pdf test  
psbasemap [ERROR]: Projection standard latitude is not known; cannot convert -R<unit>... to geographic corners
psbasemap [ERROR]: Syntax error -R option. Correct syntax:
    -R<xmin>/<xmax>/<ymin>/<ymax>[/<zmin>/<zmax>]
      Append r if giving lower left and upper right coordinates
    -Rg or -Rd for global domain
    -R<grdfile> to take the domain from a grid file
psbasemap [ERROR]: Offending option -R69:30/71:45/-17/-15:15

Try cleaning your .gmt/sessions dir. It works fine for me.

Joaquim wrote:

Try cleaning your .gmt/sessions dir. It works fine for me.

Still failing after removing .gmt/sessions directory:

$ rm -rv ~/.gmt/sessions
removed directory: ‘/home/seisman/.gmt/sessions’
$ gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/1:1000000 -Bafg -B+t"Survey area" -pdf test
psbasemap [ERROR]: Projection standard latitude is not known; cannot convert -R<unit>... to geographic corners
psbasemap [ERROR]: Syntax error -R option. Correct syntax:
    -R<xmin>/<xmax>/<ymin>/<ymax>[/<zmin>/<zmax>]
      Append r if giving lower left and upper right coordinates
    -Rg or -Rd for global domain
    -R<grdfile> to take the domain from a grid file
psbasemap [ERROR]: Offending option -R69:30/71:45/-17/-15:15

All other plot commands (for different projections) in the psbasemap documentation work, except this one.

Hope no Ubuntu devil in that GMT build. It works fine in Wibuntu too (... which is based on Ubuntu).
It's a very weird error because standard lat is not mandatory in transverse mercator. You can try variations like

gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/0/1:1000000 -Bafg -B+t"Survey area" -pdf test

or

gmt basemap -R69:30/71:45/-17/-15:15 -Jtmerc70/1:1000000 -Bafg -B+t"Survey area" -pdf test

but that's getting into the witchcraft world.

I'm feeling lost and confused with the results. Maybe there is something wrong with the conda-forge package?

$ gmt --version
6.0.0_r20280

$ which gmt
~/.anaconda/bin/gmt

$ rm -rv ~/.gmt/sessions
rm: cannot remove ‘/home/seisman/.gmt/sessions’: No such file or directory

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/1:1000000 -Bafg -B+t"Survey area" -pdf test                                                                                                 
psbasemap [ERROR]: Projection standard latitude is not known; cannot convert -R<unit>... to geographic corners
psbasemap [ERROR]: Syntax error -R option. Correct syntax:
    -R<xmin>/<xmax>/<ymin>/<ymax>[/<zmin>/<zmax>]
      Append r if giving lower left and upper right coordinates
    -Rg or -Rd for global domain
    -R<grdfile> to take the domain from a grid file
psbasemap [ERROR]: Offending option -R69:30/71:45/-17/-15:15

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/0/1:1000000 -Bafg -B+t"Survey area" -pdf test  
ERROR: Caught signal number 11 (Segmentation fault) at
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(gmt_geo_to_xy_line+0x6c)[0x7fc7517edb5c]
[0x0]
Stack backtrace:
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(sig_handler+0x257)[0x7fc7516b4cd7]
/lib64/libpthread.so.0(+0xf5e0)[0x7fc75025b5e0]
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(gmt_geo_to_xy_line+0x6c)[0x7fc7517edb5c]
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(gmt_linearx_grid+0x336)[0x7fc751806f66]
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(gmt_map_basemap+0x2596)[0x7fc751822c66]
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(GMT_psbasemap+0xb1c)[0x7fc75196dd2c]
/home/seisman/.anaconda/bin/../lib/libgmt.so.6(GMT_Call_Module+0x115)[0x7fc7516cb595]
gmt(main+0x1fd)[0x40117d]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7fc74b066c05]
gmt[0x4019f5]

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jtmerc70/1:1000000 -Bafg -B+t"Survey area" -pdf test                                                                                             
psbasemap [ERROR]: Could not decode merc70, return NaN.
psbasemap [ERROR]: Syntax error -J option. Correct syntax:
    -Jt<lon0>/[<lat0>/]<scale> OR -JT<lon0>/[<lat0>/]<width>
      <scale> is <1:xxxx> or cm/degree, or use <width> in cm
psbasemap [ERROR]: Offending option -Jtmerc70/1:1000000
psbasemap [ERROR]: Central meridian is not known; cannot convert -R<unit>... to geographic corners
psbasemap [ERROR]: Syntax error -R option. Correct syntax:
    -R<xmin>/<xmax>/<ymin>/<ymax>[/<zmin>/<zmax>]
      Append r if giving lower left and upper right coordinates
    -Rg or -Rd for global domain
    -R<grdfile> to take the domain from a grid file
psbasemap [ERROR]: Offending option -R69:30/71:45/-17/-15:15

Sorry, my -Jtmerc example was wrong. Should have been -Jtmerc/70/1:1000000

Regarding the other, I have no explanation. But why don't you build GMT yourself? I recently build it under Cygwin (what a pain) but compiling instruction I used was (just adapt for your case)

cmake --config Release /cygdrive/c/progs_cygw/GMTdev/gmt5/trunk -DDO_EXAMPLES=ON -DDO_TESTS=ON -DCOPY_GSHHG=TRUE -DCOPY_DCW=TRUE -DSUPPORT_EXEC_IN_BINARY_DIR=TRUE -DGSHHG_ROOT=/cygdrive/c/progs_cygw/GMTdev/gmt5/trunk/share/coast -DDCW_ROOT=/cygdrive/c/progs_cygw/GMTdev/gmt5/trunk/share/dcw -DGMT_OPENMP=ON

make -j

I built GMT r20292 myself. Below are the results:

$ gmt --version
6.0.0

$ which gmt
/opt/GMT6-dev/bin/gmt

$ rm -rv ~/.gmt/sessions
removed directory: ‘/home/seisman/.gmt/sessions’

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/0/1:1000000 -Bafg -B+t"Survey area" -pdf test
ERROR: Caught signal number 11 (Segmentation fault) at
/opt/GMT6-dev/lib64/libgmt.so.6(gmt_geo_to_xy_line+0x6c)[0x7f00f54de5ec]
[0x0]
Stack backtrace:
/opt/GMT6-dev/lib64/libgmt.so.6(sig_handler+0x257)[0x7f00f53a5f37]
/lib64/libc.so.6(+0x35270)[0x7f00eec6b270]
/opt/GMT6-dev/lib64/libgmt.so.6(gmt_geo_to_xy_line+0x6c)[0x7f00f54de5ec]
/opt/GMT6-dev/lib64/libgmt.so.6(gmt_linearx_grid+0x336)[0x7f00f54f77f6]
/opt/GMT6-dev/lib64/libgmt.so.6(gmt_map_basemap+0x2597)[0x7f00f5512637]
/opt/GMT6-dev/lib64/libgmt.so.6(GMT_psbasemap+0xb1c)[0x7f00f565c62c]
/opt/GMT6-dev/lib64/libgmt.so.6(GMT_Call_Module+0x115)[0x7f00f53bc805]
gmt(main+0x1fd)[0x400ecd]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f00eec57c05]
gmt[0x401743]

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jtmerc/70/1:1000000 -Bafg -B+t"Survey area" -pdf test
psbasemap [ERROR]: Projection standard latitude is not known; cannot convert -R<unit>... to geographic corners
psbasemap [ERROR]: Syntax error -R option. Correct syntax:
    -R<xmin>/<xmax>/<ymin>/<ymax>[/<zmin>/<zmax>]
      Append r if giving lower left and upper right coordinates
    -Rg or -Rd for global domain
    -R<grdfile> to take the domain from a grid file
psbasemap [ERROR]: Offending option -R69:30/71:45/-17/-15:15

$ gmt basemap -R69:30/71:45/-17/-15:15 -Jt70/1:1000000 -Bafg -B+t"Survey area" -pdf test  
psbasemap [ERROR]: Projection standard latitude is not known; cannot convert -R<unit>... to geographic corners
psbasemap [ERROR]: Syntax error -R option. Correct syntax:
    -R<xmin>/<xmax>/<ymin>/<ymax>[/<zmin>/<zmax>]
      Append r if giving lower left and upper right coordinates
    -Rg or -Rd for global domain
    -R<grdfile> to take the domain from a grid file
psbasemap [ERROR]: Offending option -R69:30/71:45/-17/-15:15

OK, I'm as lost as you. My last attempt, but even if it works it won't explain previous errors. Try with proj syntax

gmt basemap -R69:30/71:45/-17/-15:15 -J+proj=tmerc+lon_0=70/1:1000000 -Bafg -B+t"Survey area" -pdf test

Unfortunately, it's still failing for me.
Joaquim wrote:

OK, I'm as lost as you. My last attempt, but even if it works it won't explain previous errors. Try with proj syntax

gmt basemap -R69:30/71:45/-17/-15:15 -J+proj=tmerc+lon_0=70/1:1000000 -Bafg -B+t"Survey area" -pdf test

But what was the error message?

Joaquim wrote:

But was is the error message?

$ gmt basemap -R69:30/71:45/-17/-15:15 -J+proj=tmerc+lon_0=70/1:1000000 -Bafg -B+t"Survey area" -pdf test
ERROR: Caught signal number 11 (Segmentation fault) at
[0x18a1840]
[0x18a1840]
Stack backtrace:
/opt/GMT6-dev/lib64/libgmt.so.6(sig_handler+0x257)[0x7fe7085d0f37]
/lib64/libc.so.6(+0x35270)[0x7fe701e96270]
[0x18a1840]

So, only debugging can tell what's going on in your system.

Hi Dongdong, remind us again of your OS and if build with debug or for relelase. Are you able to run your examine in a debugger to see where/why it crashes?

Also, please run command with -Vd.

Paul wrote:

The message about standard latitude is only given by gmtinit_rectR_to_geoR which is only called when -R is given in projected units. That lead me to check how it might get in there and I found that an array double rect of length 4 is not initialized and could have random junk exceeding 360. Then, if the read for the meters fails then we may be comparing junk to see if values are degrees or meters. I have fixed this by initializing the array to zero and also check if the read is successfully converting 4 numbers. I suspect this was the problem. Let me know. In r20295.

r20295 fixed the issue. The commands all succeed.