Strange warning message with subplot and -Ba+ap

[Esteban82]

I run this a command withsubplot -B+apand I got this warning:

gmt subplot begin 2x2 $REGION -Fs20c -Bxaf -Byaf+ap -Srl -Scb -A1+gwhite+jTR -M0.05c/0.75c
subplot [WARNING]: Modifier +a detected but not a valid modifier! - ignored

The figure is well created.

The issue is only about the warning message. I don't know why I got it. I think it should NOT appear.

[joa-quim]

Don't know. But +a should be followed by a number. Not a p|s directly.

[Esteban82]

But +a should be followed by a number.

+a accepts two shorthands

+an can be used as a shorthand for normal (i.e., +a90) [Default for y-axis] and +ap for parallel (i.e., +a0) annotations [Default for x-axis].

https://docs.generic-mapping-tools.org/dev/reference/options.html#map-frame-and-axes-annotations-the-b-option

[seisman]

gmt/src/subplot.c

Lines 730 to 752 in fe9a841

	if (By) { /* Did get y-axis annotation settings */
	if ((c = gmt_first_modifier (GMT, By->arg, "lpsu"))) { /* Gave valid axes modifiers for custom labels */
	pos = 0;
	while (gmt_getmodopt (GMT, 'B', c, "lpsu", &pos, p, &error) && error == 0) {
	switch (p[0]) {
	case 'l': /* Regular y-axis label */
	Ctrl->S[GMT_Y].label[GMT_PRIMARY] = strdup (&p[1]); break;
	case 'p': /* Annotation prefix */
	Ctrl->S[GMT_Y].prefix[GMT_PRIMARY] = strdup (&p[1]); break;
	case 's': /* Secondary y-axis label */
	Ctrl->S[GMT_Y].label[GMT_SECONDARY] = strdup (&p[1]); break;
	case 'u': /* Annotation units */
	Ctrl->S[GMT_Y].unit[GMT_PRIMARY] = strdup (&p[1]); break;
	default: /* These are caught in gmt_getmodopt so break is just for Coverity */
	break;
	}
	}
	if (Ctrl->S[GMT_Y].label[GMT_PRIMARY] || Ctrl->S[GMT_Y].label[GMT_SECONDARY]) Ctrl->S[GMT_Y].has_label = true;
	c[0] = '\0'; /* Chop off for now */
	}
	Ctrl->S[GMT_Y].b = subplot_prep_annot_args (Ctrl, By, GMT_Y);
	if (c) c[0] = '+';
	}

subplot has its own -B parser. From the source code, it seems currently it only allows modifiers +l/+p/+s/+u. I don't know why it was written like this, perhaps we should try to support +a and see if there are any side-effects.