pstext -D[j|J] is broken

[Chuan1937]

Add additional check for MC position in gmt_support.

Fixes: #8712
Refixing: #8916

The local test was successful:

The following tests passed:
        doc/scripts/GMT_pstext_clearance.sh
        doc/scripts/GMT_pstext_justify.sh
        test/pstext/ISO-8859-2.sh
        test/pstext/book.sh
        test/pstext/boxtext.sh
        test/pstext/caption.sh
        test/pstext/cliptxtbox.sh
        test/pstext/composite.sh
        test/pstext/corner_justify.sh
        test/pstext/delay.sh
        test/pstext/escape.sh
        test/pstext/filled_font.sh
        test/pstext/labeler.sh
        test/pstext/para_offset.sh
        test/pstext/pstext-l2s.sh
        test/pstext/shade_text.sh
        test/pstext/supersub.sh
        test/pstext/textclip.sh
        test/pstext/utf8.sh

100% tests passed, 0 tests failed out of 19

Total Test time (real) =  16.27 sec

[joa-quim]

I still have local failures that seem to be related to this. For example,

        799 - test/psbasemap/radial_labels1.sh (Failed)
        800 - test/psbasemap/radial_labels2.sh (Failed)
        801 - test/psbasemap/radial_labels3.sh (Failed)

They fail because text strings are a bit out of place when compared with references.

[Chuan1937]

I still have local failures that seem to be related to this. For example,

        799 - test/psbasemap/radial_labels1.sh (Failed)
        800 - test/psbasemap/radial_labels2.sh (Failed)
        801 - test/psbasemap/radial_labels3.sh (Failed)

They fail because text strings are a bit out of place when compared with references.

Need to run all the tests, right. I've only run examples related to pstext before. So there are omissions.

[Chuan1937]

In gmt_support.c, the mode property of gmt_smart_justify is extended by adding mode=3/4 as a dedicated branch for pstext -Dj/-DJ. The MC will only trigger on this dedicated branch; the original behavior remains unchanged.

Finally, in three calls in pstext.c, mode is changed from Ctrl->D.justify to Ctrl->D.justify + 2, explicitly routing -Dj/-DJ to the dedicated branch.

[joa-quim]

Now the previously failing test pass.

[seisman]

But there are still a few new failures in the CI

[Chuan1937]

But there are still a few new failures in the CI

I can see that CI works. It makes the follow-up easy

[Chuan1937]

But there are still a few new failures in the CI

@seisman

I don't think these failures are due to my revisions:

The following tests FAILED:
	 80 - doc/scripts/GMT_color_hsv.sh (Failed)
	192 - doc/scripts/GMT_vector.sh (Failed)
	193 - doc/scripts/GMT_vector4.sh (Failed)
	702 - test/mapproject/oblmerc_down.sh (Failed)
	703 - test/mapproject/oblmerc_up.sh (Failed)
Errors while running CTest

[joa-quim]

I'm afraid they are. If you look on why those tests fail, there is always a text element that changed position.

[Chuan1937]

I tested these 5 tests on the master branch and they all fail there too:

• GMT_color_hsv.sh (RMS 0.039)
• GMT_vector.sh (RMS 0.018)
• GMT_vector4.sh (RMS 0.016)
• oblmerc_down.sh (RMS 0.011)
• oblmerc_up.sh (RMS 0.014)

[joa-quim]

Nope, they pass for me with master.