David Nalesnik [Tue, 23 Jul 2013 01:20:33 +0000 (20:20 -0500)]
Issue 3472: several fixes to input/regression/scheme-text-spanner.ly
* The file `input/regression/scheme-text-spanner.ly' assigns a constant
pair to the variable `event-drul' and subsequently attempts to modify
the pair. This can lead to scheme-text-spanners not appearing when
requested in simultaneous contexts. This patch corrects the faulty usage,
replacing `event-drul' with two variables `event-start' and `event-stop'.
* The variable `current-event' is not necessary, and has been removed.
* Tabs in the engraver definition have been replaced with spaces.
David Kastrup [Wed, 24 Jul 2013 10:47:40 +0000 (12:47 +0200)]
Issue 3388: -dpreview retains non-working page links in table of contents
This just disables mark_page_link in the EPS header. When including
EPS files in LilyPond itself, BeginEPSF and EndEPSF should limit the
effect of this redefinition to the EPS file itself, restoring the
userdict afterwards.
Since integers are accepted in this kind of usage, not accepting
floating point numbers appears inconsistent.
While in INITIAL mode like inside of layout blocks real numbers can
be written like 4. or -.3 for historic reasons, permitting this inside
of music could easily lead to confusion with durations or
articulations, so those spellings remain invalid in music modes.
Decimal fractions are also not accepted in \chordmode since this would
preclude chord entry like c:13.11^3.7 and similar.
Thomas Morley [Wed, 24 Jul 2013 17:14:16 +0000 (19:14 +0200)]
Issue 3469: adding markup-commands \oval and \ellipse
\oval and \ellipse are drawing an oval respectively an ellipse around their
argument.
Also adding a reg-test for markup-commands framing their argument.
David Kastrup [Tue, 9 Jul 2013 18:22:24 +0000 (20:22 +0200)]
Issue 3449: Shift accordion register symbols down by half a line width
This makes the documented proposals for combining the accordion
register symbols together with the accordion dot symbol match in
positioning. It would have been feasible instead to shift the
(centered) accordion dot, but some register symbols work with
different line widths.
The net result will now exhibit a slight descender (namely half a
linewidth) of the accordion register symbols when used in a running
line.
David Kastrup [Fri, 5 Jul 2013 16:23:28 +0000 (18:23 +0200)]
Issue 3439: Create a two-argument form of define-event-class
This definition of define-event-class just specifies the event class
symbol itself as well as its immediate parent class. Redefining
existing event classes is not (yet?) supported.
Federico Bruni [Sun, 2 Jun 2013 09:45:59 +0000 (11:45 +0200)]
Doc-it: fix credits for translation checkers
I want to keep track of translated files that still need a review
from a translation checker. And I don't want someone to be credited
for something he never did.
David Kastrup [Wed, 19 Jun 2013 09:42:27 +0000 (11:42 +0200)]
Issue 3414: Mark c0-position as an internal grob property
Since its value is explicitly initialized from either
middleCClefPosition or middleCPosition for the relevant grobs
KeySignature and KeyCancellation, overriding or tweaking it has no
visible effect.
David Kastrup [Sun, 16 Jun 2013 20:13:39 +0000 (22:13 +0200)]
Issue 3413: Remove patches from elisp
The files emacsclient.patch and server.el.patch applied to ancient
versions of Emacs. They have been unnecessary for more than 8 years
now. There is no conceivable use for them any more.
where only the reverts get conflated (and thus Flag color stays in
place). That's less than perfect, but probably better than leaving
reverts alone (for Flag.transparent, they might have been created
automatically by an earlier convert-ly rule).
David Kastrup [Thu, 6 Jun 2013 13:36:48 +0000 (15:36 +0200)]
Issue 3401: Parser error with void function in music list
Problem code fixed:
test =
(symbol-list?))
{
\test Symbol
}
This patch is sort of a so-so fix for the problem: for one thing,
embedded_scm_closed should not have been able to end in a symbol list
in the first place since parsing those requires lookahead for '.'.
The grammar has now been changed to use embedded_scm, and just backs
up in the case a scheme function/identifier returns a markup in lyrics
mode (which can then be assembled with duration and post-events into a
more complex music expression). If parsing the source for the markup
required lookahead, the same error will be thrown.
But that use case of Scheme functions in Lyrics mode for generating
markups to be turned into lyric events is a lot more exotic than the
one triggering this issue. The current fix at least results in
predictable behavior outside of using Lyrics mode, and in particular
for uses of define-void-function.