Inset within subplot panel fails

[PaulWessel]

As reported on the forum, inset used in subplot panels fails in terms of placement. Clearly the -Dj setting is computing strange placement coordinates. In the original post, the BR box was plotted way to the right and later covered up by other plots.

A simple example contrasts plain inset versus inset inside a panel:

#!/bin/bash
# Shows problem with inset within subplot panel
gmt begin
    gmt figure inset_in_panel png
    gmt subplot begin 1x1 -Fs5c/4c -A
        gmt subplot set 0
        gmt coast -R70/138/13/56 -Glightgray -B
        gmt inset begin -DjTL+w0.5c/0.8c+o0.1c -F+p1p,red
        gmt inset end
    gmt subplot end

    gmt figure plain_inset png
    gmt coast -R70/138/13/56 -Glightgray -B -JQ5c
    gmt inset begin -DjTL+w0.5c/0.8c+o0.1c -F+p1p,red
    gmt inset end

gmt end show

Only showing the failing inset_in_panel plot here (the other is fine):

[gd-a]

Related to #6288 no?

[seisman]

I had the same problem today. Ping @PaulWessel just in case you forgot it.

[PaulWessel]

Not truly forgotten, but on the road...

[seisman]

Here is another example that doens't work as expected:

gmt begin map
	gmt subplot begin 1x2 -Fs15c/15c
	gmt subplot set 0
		gmt coast -R-74/-69.5/41/43 -JM? -Baf -Wthin -Clightblue -Glightyellow
		gmt inset begin -DjBL+o0.5c/0.2c -F+pblack+gwhite+c0.2c/0.2c -R-80/-65/35/50 -JM3c
			gmt coast -Ggray -Cwhite -EUS.MA+gred -W1/thin -N1 -N2
	    gmt inset end
	gmt subplot set 1
		gmt coast -R-74/-69.5/41/43 -JM? -Baf -Wthin -Clightblue -Glightyellow
	    gmt inset begin -DjBL+o0.5c/0.2c -F+pblack+gwhite+c0.2c/0.2c -R-80/-65/35/50 -JM3c
		    gmt coast -Ggray -Cwhite -EUS.MA+gred -W1/thin -N1 -N2
	    gmt inset end
    gmt subplot end
gmt end show