Andrew Hawryluk [Fri, 4 Sep 2009 04:22:24 +0000 (22:22 -0600)]
Essay: freshened up Baerenreiter images
Remove annotations from baer-ledger.png, add original scan of
Bach's Sarabande, straighten baer-sarabande.png, and add a
high resolution (1200 dpi) b/w version for the PDF.
Joseph Wakeling [Tue, 1 Sep 2009 19:33:41 +0000 (21:33 +0200)]
Extended documentation on Turkish classical music and Makam.
* New section on 'Non-Western note names and accidentals'
in the Pitches section of the Notation Reference.
Contains table of accidentals for makam.ly.
* New section on Turkish classical music in the World
Music section of the Notation Reference.
Patrick McCarty [Mon, 31 Aug 2009 05:32:03 +0000 (22:32 -0700)]
Vim ftdetect: use `set ft' instead of `setf'.
Since filetype.vim is no longer distributed, the `setf' command will
only set the filetype if Vim's default filetype.vim does not detect a
matching filetype first.
Unfortunately, when certain lines begin with #(, the "generic" filetype
is matched, so the LilyPond-specific highlighting is never loaded.
A workaround is to use `set ft'; this sets the filetype to `lilypond'
unconditionally for all files ending in .ly or .ily, which is what we
want.
Use our own ~s ly:format placeholder, since guile is broken with wide chars
We need our own ~S format placeholder to escape quotes, so use
lilypond's string formating to make wide utf-8 chars work in filenames
Guile seems unable to handle wide characters, so we mustn't use
simple-format or format, because that will break with non-ANSI characters
(char code > 255). Instead, Implement our own ~S placeholder, that
wraps the string in double quotes and escapes backslashes, double
quotes, percent and dollar signs.
Use this format placeholder to generate the proper cmd line args
for our calls to gs.
Implement framework for post-fix text (de)cresc spanners (backend only)
The main problem why \decr or \dim were prefix operators was that the
new dynamic engraver was reading only the properties of the Engraver
itself, so there was no way to create a CrescendoEvent and at the same
time specify the hairpin/text behavior.
This patch makes the New_dynamic_engraver also look at the properties
of the (Dec|C)recendoEvent and prefer a setting from there. If the event
does not have a property (which will be the case most of the time), the
Engraver's settings are used (i.e. same behavior as now).
With this change, e.g. \dim can be defined as
dim = #(make-music 'DecrescendoEvent 'span-direction START
'span-type text 'span-text "dim.")
and used as postfix like a4\dim without the need of any overrides
before the affected note.
This is just the backend commit, implementing processing of the event's
properties. The definition of \cresc is still unchanged, leaving the old
prefix notation in place, until we find good upgrade path.
Also add regtests for text cresc spanners and add the regtest to the
snippets.
Werner Lemberg [Thu, 27 Aug 2009 13:33:19 +0000 (15:33 +0200)]
[config] Use fc-list.
In case there are different Century fonts available on the system (for
example, in both PFB and OTF format), fc-match only lists the first match
which can be the non-PFB one, while fc-list lists them all.
Some files (e.g. the CC licenses have weird bounding boxes, starting
e.g. at 300). So far, lilypond inserted them with a huge offset (exactly
shifted by the coordinates of the bounding box), which is unlike any other
eps-handling application I know.
Graham Percival [Tue, 25 Aug 2009 06:07:28 +0000 (07:07 +0100)]
Doc build: remove some false idiosyncracities.
The essay should be built just like any other documentation, and
we do want general as one-big-page.html. (the question came up on
-devel in July, and at least two people were horrified at the
notion that we weren't planning on making a big-page version.
*shrug* )
Joe Neeman [Mon, 24 Aug 2009 21:03:50 +0000 (14:03 -0700)]
Use the full pure Y-extent for complete systems.
For staves within a system, we split up the pure Y-extent into the Y-extent of
the prefatory matter and the Y-extent of the rest (thereby emulating a crude
skyline). Between systems, we use a single extent box. Before this commit, we
were ignoring the Y-extent of the prefatory matter altogether; now the extent
box is correct.