Fix meca

[Chuan1937]

Fix test error
#8821

The test failed, but it works correctly when I test it locally.

(base) chuan@DESKTOP-QF9C9FL:~/Code/gmt$ cd /home/chuan/Code/gmt/build && ctest --output-on-failure -R "fc_okb|full_moment_tensor|meca_-Smdz|seis_07" 2>&1 | tail -15
Test project /home/chuan/Code/gmt/build
    Start  692: test/potential/fc_okb.sh
1/4 Test  #692: test/potential/fc_okb.sh ..........   Passed    0.97 sec
    Start 1137: test/seis/full_moment_tensor.sh
2/4 Test #1137: test/seis/full_moment_tensor.sh ...   Passed    3.06 sec
    Start 1138: test/seis/meca_-Smdz.sh
3/4 Test #1138: test/seis/meca_-Smdz.sh ...........   Passed    1.50 sec
    Start 1148: test/seis/seis_07.sh
4/4 Test #1148: test/seis/seis_07.sh ..............   Passed    2.39 sec

100% tests passed, 0 tests failed out of 4

Total Test time (real) =   7.95 sec

@seisman

[seisman]

Are you sure that your changes in #8821 are correct and the baseline images are worng? Need some time to verify it.

[joa-quim]

I get no local failures related to psmeca.

[seisman]

The CI jobs on Linux/macOS/Windows fail and I also get the same failures locally.

[Chuan1937]

I'll check and correct it again

[joa-quim]

I'm seeing the failures now too (probably did not update before). And, for example the seis_07 new result is wrong (bottom row). It is producing a left lateral strike-slip when the script says it's meant to be a right lateral.

[Chuan1937]

Let me think it seriously.

[Chuan1937]

Let me think it seriously.

[seisman]

I think we should revert changes in #8821 until we find a working solution.

[joa-quim]

Yes, agree.