[iliev-blnk]

Fixes #144, fixes #147, fixes #151

Cause

AATStateMachine.process looks up the entry for the end-of-text class and moves to its new state, but never passes the entry to the subtable. So the actions a font puts on its end-of-text transitions are dropped:

• MORX-10 and MORX-11 ([morx] Should take end-of-text transition #144): the font marks the last glyph and runs a rearrangement verb at the end of the text.
• MORX-17 ([morx] Failure for MORX-17: Rearrangement with DontAdvance at End of Text #147): the same, with DontAdvance set on the end-of-text entry.
• MORX-20 ([morx] Contextual Glyph Substitution at End-of-Text #151): the font substitutes the marked glyph and the last glyph at the end of the text.

Fix

The state machine now passes the end-of-text entry to the subtable with a null glyph, and still never stays at the end of the text, whatever DontAdvance says. Each subtable then follows HarfBuzz, which matches CoreText here:

• Rearrangement: MarkLast at the end of the text marks the last glyph (this is the clamp from Rearrange morx glyphs only within a long enough marked range (#142, #143) #433), and the verb runs as usual.
• Contextual substitution: nothing happens unless a glyph was marked with SetMark. Otherwise the mark substitution applies to the marked glyph and the current substitution to the last glyph.
• Ligature and glyph insertion: unchanged, they still skip the end of the text. HarfBuzz forms no ligatures there either. It does insert glyphs at the end of the text, which would be a separate change on top of Fix AAT glyph insertion (#171, #172, #173, #174) #429.

	#433	this PR	HarfBuzz 15.0.0
MORX-10 ABABAB	A B A B A B	A B A B B A	A B A B B A
MORX-11 BABBAABX	B A B B A A B X	B A B B A X A B	B A B B A X A B
MORX-17 AB	A B	B A	B A
MORX-20 ABCDE	A B C D E	A B C.alt D E.alt1	A B C.alt D E.alt1
MORX-20 AE	A E	A.alt E.alt1	A.alt E.alt1

Across all 174 MORT/MORX cases of text-rendering-tests, this goes from 131 to 139 passing on top of #433: MORX-10, MORX-11, MORX-17 and the five failing MORX-20 cases. No other results change.

This builds on #433, because the rearrangement cases need its MarkLast clamp. Only the last commit is new here.

Tests

• test/issues.js: MORX-10, MORX-11, MORX-17 and all seven MORX-20 inputs. All three tests fail without the fix.
• The test fonts are TestMORXTen.ttf, TestMORXEleven.ttf, TestMORXSeventeen.ttf and TestMORXTwenty.ttf from text-rendering-tests, added to test/data/fonttest/ under the Apache License 2.0 in test/data/fonttest/LICENSE.txt.
• npm test: 715 passing. npm run format:check passes.