Conflict between 'legend specfile' and 'plot -l'

Q&A
1

In order to plot an own symbol via legend, I would like to combine gmt plot -l and gmt legend specfile commands in modern mode. But in a given begin–end environment legend ignores the specfile if, in a previous plot command, the -l flag was given:

gmt begin a pdf
        gmt basemap -JX5/5 -R0/5/0/4 -BEWSN
        echo "-1 -1" | gmt plot -Sx0.2 -Wthick -l'blub'
        gmt legend -DjTL+o0.25c/0c <<- END
        S - x 0.0787402i - thick,black - nominal height 
        END
gmt end show

This results in “blub” is written instead of “nominal height” into the legend. The flag -M also does not help here, as in this case both information is considered: “blub” and “nominal height” are written.
Is their a possibility to force legend to ignore the hidden auto-legend information?

I am using gmt 6.3 on linux.

2

btw. from the manual it should work but also with -Me both, hidden and explicit information is read. So, is it a bug?

3

Well, you are actually not giving any spec file - you are expecting legend to somehow know it should read stdin. If you instead save that info to a file and pass it to legend and use -M then both things show.

4

See -M to ignore one or the other.

5

Thanks for the quick answer, but the problem remains.

gmt begin a pdf
        gmt basemap -JX5/5 -R0/5/0/4 -BEWSN
        echo "-1 -1" | gmt plot -Sx0.2 -Wthick -l'blub'
        gmt legend -DjTL+o0.25c/0c
<<- END
        S - x 0.0787402i - thick,black - nominal height 
        END
gmt end show
6

Well, did you make the changes? What you just pasted is still the wrong way of doing it - read my email above.

7

sorry tiping error
The problem remains

gmt begin a pdf
        gmt basemap -JX5/5 -R0/5/0/4 -BEWSN
        echo "-1 -1" | gmt plot -Sx0.2 -Wthick -l'blub'
        cat <<- END > legend.txt
        S - x 0.0787402i - thick,black - nominal height 
        END
        gmt legend -DjTL+o0.25c/0c -Me legend.txt
gmt end show

results in the same output both entries appear in the legend.

8

Works for me - only nominal height shows. What GMT version are you using?

9

6.3.0

10

The golden rule of GMT is to always run the very latest. This is the output of your script (changed to PNG for showing here) when I run it on master branch (6.5). That is what you expect, no?:

a
a598×598 2.27 KB

11

This is what I expected. So I have a motivation to update. But I depend a bit on my sysadmins regarding updates of the system. So I wanted to check first, where the error came from. So thanks a lot for the quick assist.

12

Addendum: Careful reading of the version related man pages, showed it also, -M[eh] was implemented in 6.4.0.