Fix for #70 and some porrectus episemus bugfixes

[henryso]

These are two commits primarily to address #70.

While testing, I noticed some weirdness with the heights of horizontal episemus decorating porrectus and torculus resupinus figures. See the following tests in gregorio-tests to see the difference:

• gabc-output/glyphs/torculus_resupinus_quilisma.gabc
• gabc-output/glyphs/porrectus.gabc
• gabc-output/glyphs/torculus_resupinus.gabc
• gabc-output/glyphs/torculus_resupinus_flexus.gabc

I made some semi-educated decisions around positioning the episemus on the final two notes of a porrectus or torculus resupinus, which may be wrong, so please review these changes carefully before accepting them, and please let me know if I should change anything.

A few notes:

• The widths of the porrectus and ambitus-one porrectus episemus glyphs are the same as they were, but the figure within the glyph is reduced horizontally by the size of PunctumSmall and left-aligned. This is the essential fix for New horizontal episemus on porrectus except last note #70. I did this so that thee calculations for what modern notation calls ledger lines will be correct.
• In the same way that torculus resupinus flexus falls through into porrectus flexus (in the C switch statement), torculus resupinus now falls through into porrectus. Thus the tail handling (of the "porrectus part") should be identical.
• In the "porrectus part" of high-low-high, if the episemus is on the low note, it will always be below and right-aligned with the vertical stem unless there is also an episemus on the first high note. Auctus figures (with the head to the right of the stem) should now be taken into account correctly.
• If the all three notes of a "porrectus part" have episema and the final high note is a deminutus, it will draw a punctum-sized episemus to ensure that the episema melt together properly.
• There are still some anomlies like (d_h_g_j): the height of the episemus above hg is in a different place than the episemus above d. In the older code, the episema would be at same height, but actually end up below the i note. I will look into this on another day unless someone beats me to it.

[henryso]

Might as well get the ASCII chart in as well.

[henryso]

If you are satisfied with the changes thus far, then please merge this, I will tackle the remaining anomalies as part of #40 / #385 / #386 since the episemus code needs to be rewritten to handle those issues anyway.