Fix NotoSansOriya-MM.glyphs GSUB,GDEF

[KazunariTsuboi]

The work I did was to fix GSUB and GDEF to use glyphs that already existed in the font but weren't used. No modifications have been made to the glyph design or anchors.

I also added a test that uses fontdiff. The result of the test is as follows. I sought to minimize the impact on strings that already exist.

The issues I have modified and the details of the modifications are as follows.

Fixes googlefonts/noto-fonts#2380

Cause: ■ pres

• When ra-oriya.below is involved, the first consonant is not replaced by * .base

■ blws

• Substitutions related to the defined ra-oriya.below do not trigger.

Measures: ■ pres

• Added processing to replace the consonant of the first character with * .base when ra-oriya.below is involved (addition of backtrack)
• Define ignore sub to use RA + Vowl glyphs of GID408-411 when ra-oriya.below and u, uu, r, rr, l, ll are combined

■ blws

• Split the blws lookup to trigger the intended replacement

Fixes googlefonts/noto-fonts#2381

Cause -Not included in the GDEF Markclass

Measures -Fixed glyph properties in Glyphs App. Script to oriya, category to Mark, subcategory to Nospacing.

Fixes googlefonts/noto-fonts#2382

Cause ■ psts, abvs

• The position of reph-oriya in the substitution condition is different from the position reordering Indic after rphf feature.

Measures

• Fixed reph-oriya position

Fixes googlefonts/noto-fonts#2383

Measures

• Fixed abvs and blws conditions

[simoncozens]

This is brilliant, thank you! @marekjez86, before we merge this I would like to add a few commits (bump version/copyright, add fontbakery shaping tests, check GF specs)

[marekjez86]

@KazunariTsuboi : thank you! great result @simoncozens : waiting for bump version/copyright, add fontbakery shaping tests, check GF specs