position003.ttml uses height-relative units for horizontal lengths, and vice-versa

[palemieux]

imsc-tests/imsc1_1/ttml/position/position003.ttml

Line 32 in e0213a2

<region style="s1" xml:id="r19" tts:extent="60rw 20rh" tts:position="left 25rw" />

imsc-tests/imsc1_1/ttml/position/position003.ttml

Line 16 in e0213a2

<region style="s1" xml:id="r6" tts:extent="60rw 20rh" tts:position="25rh" />

imsc-tests/imsc1_1/ttml/position/position003.ttml

Line 36 in e0213a2

<region style="s1" xml:id="r23" tts:extent="60rw 20rh" tts:position="right 25rw" />

[nigelmegitt]

@palemieux I think those tests are valid as is; it is permitted to do this. Is the problem here that there are no similar tests where height relative units are used for vertical positions and width relative units are used for horizontal positions? Perhaps it makes sense to add a test for that case rather than change this one.

[palemieux]

@nigelmegitt IMSC 1.1, 7.2.9 prohibits rh in horizontal length components, and rw in vertical length components.

[nigelmegitt]

Ah, thanks for that @palemieux , distant bells are ringing in my memory about discussing this. I agree this test is broken and that the h/v length units are the wrong way around. Sorry for the mistake before.