simon-git: gonville (master): Simon Tatham

[Commits to Tartarus hosted VCS]

TL;DR:
  a58dd2b Add updated coverage tests.
  fbac966 Add another exclusion to --lilycheck.
  170e336 Shorten the variant segno by one loop.
  ad2e488 Add scripts.varsegno to the Lilypond glyph table.
  a162825 Fix LILC bounding boxes for lyric ties.
  9be29ca Fix small version of tenorised G clef.

Repository:     https://git.tartarus.org/simon/gonville.git
On the web:     https://git.tartarus.org/?p=simon/gonville.git
Branch updated: master
Committer:      Simon Tatham <anakin at pobox.com>
Date:           2020-03-17 18:10:02

commit a58dd2bf4c5958d6153d35f99ab36b59b2453484
web diff https://git.tartarus.org/?p=simon/gonville.git;a=commitdiff;h=a58dd2bf4c5958d6153d35f99ab36b59b2453484;hp=dbbc53bf726df78b04ff5dae4a3164e5970c3e45
Author: Simon Tatham <anakin at pobox.com>
Date:   Tue Mar 17 17:57:44 2020 +0000

    Add updated coverage tests.
    
    I've written a lilytest-2.21.ly file which exercises all the glyphs I
    could reach in current Lilypond git master. Then I cut it down by a
    very small amount (just deleting the extra note flags, the dynamic 'n'
    letter, and a couple of types of fermata sign) to be a coverage test
    for Lilypond 2.20.

 lilytest-2.20.ly | 363 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 lilytest-2.21.ly | 367 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 730 insertions(+)

commit fbac966b3d0a87eae23deadbdd296b15f00cc364
web diff https://git.tartarus.org/?p=simon/gonville.git;a=commitdiff;h=fbac966b3d0a87eae23deadbdd296b15f00cc364;hp=a58dd2bf4c5958d6153d35f99ab36b59b2453484
Author: Simon Tatham <anakin at pobox.com>
Date:   Tue Mar 17 17:46:44 2020 +0000

    Add another exclusion to --lilycheck.
    
    I haven't found any situation in which 'scripts.trillelement' is used
    by Lilypond. 'scripts.trill_element' seems to be the thing used for a
    wiggly line indicating a long trill.

 glyphs.py | 3 +++
 1 file changed, 3 insertions(+)

commit 170e33671c6f1310c98e9d117312c73b117f58fb
web diff https://git.tartarus.org/?p=simon/gonville.git;a=commitdiff;h=170e33671c6f1310c98e9d117312c73b117f58fb;hp=fbac966b3d0a87eae23deadbdd296b15f00cc364
Author: Simon Tatham <anakin at pobox.com>
Date:   Tue Mar 17 17:53:36 2020 +0000

    Shorten the variant segno by one loop.
    
    When I came to test it in a real score, I immediately noticed that it
    looked a bit too tall. But it took me longer to spot the reason why,
    even though it should have been obvious: the components of the glyph
    were fine, but I had just repeated the looping motif in the middle one
    time too many compared to existing examples.
    
    Reduced by one loop, it's now the same height as Lilypond's version.

 font.py | 1 -
 1 file changed, 1 deletion(-)

commit ad2e488866e0dddb812f3a78ad4d96fe1e2278ed
web diff https://git.tartarus.org/?p=simon/gonville.git;a=commitdiff;h=ad2e488866e0dddb812f3a78ad4d96fe1e2278ed;hp=170e33671c6f1310c98e9d117312c73b117f58fb
Author: Simon Tatham <anakin at pobox.com>
Date:   Tue Mar 17 17:44:52 2020 +0000

    Add scripts.varsegno to the Lilypond glyph table.
    
    I drew the actual _symbol_ in commit 732ddb773a, but forgot to give it
    its Lilypond name, so you couldn't actually select it. Ahem.

 glyphs.py | 1 +
 1 file changed, 1 insertion(+)

commit a162825bdea061520c63d377fbd64aa26f105108
web diff https://git.tartarus.org/?p=simon/gonville.git;a=commitdiff;h=a162825bdea061520c63d377fbd64aa26f105108;hp=ad2e488866e0dddb812f3a78ad4d96fe1e2278ed
Author: Simon Tatham <anakin at pobox.com>
Date:   Tue Mar 17 17:45:43 2020 +0000

    Fix LILC bounding boxes for lyric ties.
    
    The logical bounding boxes for those glyphs are zero-width and centred
    on the tie's midline. Using the physical bounding box (which is my
    build script's default if you don't override it) means Lilypond will
    add extra space when it draws a tie.

 glyphs.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

commit 9be29ca4b8aed9df598d371366d32bb5fc530eb7
web diff https://git.tartarus.org/?p=simon/gonville.git;a=commitdiff;h=9be29ca4b8aed9df598d371366d32bb5fc530eb7;hp=a162825bdea061520c63d377fbd64aa26f105108
Author: Simon Tatham <anakin at pobox.com>
Date:   Tue Mar 17 17:47:36 2020 +0000

    Fix small version of tenorised G clef.
    
    All the other 'small' versions of clefs are simply the original clef
    shrunk. (They're used when you change clefs in the middle of a line.)
    
    But the tenorised G clef is an exception, because its spiral is
    supposed to be centred on one stave line (representing a G), and its
    upper protrusions are supposed to bracket another stave line
    (representing the D four notes higher). So if you just shrink it in
    the obvious way, then the distance between those two vertical
    positions becomes wrong by a factor of 0.8.
    
    Instead, the small tenorised G clef must be built in the same way as
    the large one, but starting with scaled-down versions of the
    components. Then the vertical distance is preserved.
    
    (Incidentally, I claimed in the log message of commit 521eb894cb that
    the protrusions borrowed from the variant C clef were supposed to
    bracket middle C. That made so much logical sense to me that I didn't
    actually check it. But in fact it's nonsense, because two stave lines
    above a G is not a C, it's a D! In this clef - as you'd expect from a
    treble clef transposed down by an octave - middle C is in a gap
    between stave lines, not on a stave line at all.)

 font.py | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)