1 @c -*- coding: utf-8; mode: texinfo; -*-
4 Translation of GIT committish: FILL-IN-HEAD-COMMITTISH
6 When revising a translation, copy the HEAD committish of the
7 version that you are working on. For details, see the Contributors'
8 Guide, node Updating translation committishes..
13 @node Interfaces for programmers
14 @chapter Interfaces for programmers
16 Advanced tweaks may be performed by using Scheme. If you are
17 not familiar with Scheme, you may wish to read our
18 @ref{Scheme tutorial}.
21 * LilyPond code blocks::
26 * Contexts for programmers::
27 * Callback functions::
28 * Inline Scheme code::
32 @node LilyPond code blocks
33 @section LilyPond code blocks
35 @cindex LilyPond code blocks
36 @cindex code blocks, LilyPond
37 @funindex #@{ @dots{} #@}
41 Creating music expressions in Scheme can be tedious, as they are
42 heavily nested and the resulting Scheme code is large. For some
43 simple tasks this can be avoided by using LilyPond code blocks,
44 which enable common LilyPond syntax to be used within Scheme.
46 LilyPond code blocks look like
49 #@{ @var{LilyPond code} #@}
52 Here is a trivial example:
54 @lilypond[verbatim,quote]
55 ritpp = #(define-event-function (parser location) ()
62 LilyPond code blocks can be used anywhere where you can write Scheme
63 code. The Scheme reader actually is changed for accommodating
64 LilyPond code blocks and can deal with embedded Scheme expressions
65 starting with @code{$} and@w{ }@code{#}.
67 @cindex parser (function argument)
70 The reader extracts the LilyPond code block and generates a runtime
71 call to the LilyPond @code{parser} to interpret the LilyPond code.
72 Scheme expressions embedded in the LilyPond code are evaluated in the
73 lexical environment of the LilyPond code block, so all local variables
74 and function parameters available at the point the LilyPond code block
75 is written may be accessed. Variables defined in other Scheme modules,
76 like the modules containing @code{\header} and @code{\layout} blocks,
77 are not accessible as Scheme variables, i.e. prefixed
78 with@tie{}@code{#}, but they are accessible as LilyPond variables, i.e.
79 prefixed with@tie{}@code{\}.
81 If @code{location} (see @ref{Scheme functions}) refers to a valid
82 input location (which it usually does inside of music/@/Scheme
83 functions), all music generated inside the code block has its
84 @samp{origin} set to @code{location}.
86 A LilyPond code block may contain anything that you can use on the
87 right side of an assignment. In addition, an empty LilyPond block
88 corresponds to a void music expression, and a LilyPond block
89 containing multiple music events gets turned into a sequential music
92 @node Scheme functions
93 @section Scheme functions
94 @cindex Scheme functions (LilyPond syntax)
96 @emph{Scheme functions} are Scheme procedures that can create Scheme
97 expressions from input written in LilyPond syntax. They can be called
98 in pretty much all places where using @code{#} for specifying a value in
99 Scheme syntax is allowed. While Scheme has functions of its own, this
100 chapter is concerned with @emph{syntactic} functions, functions that
101 receive arguments specified in LilyPond syntax.
104 * Scheme function definitions::
105 * Scheme function usage::
106 * Void scheme functions::
109 @node Scheme function definitions
110 @subsection Scheme function definitions
111 @funindex define-scheme-function
113 The general form for defining scheme functions is:
117 #(define-scheme-function
118 (parser location @var{arg1} @var{arg2} @dots{})
119 (@var{type1?} @var{type2?} @dots{})
126 @multitable @columnfractions .33 .66
128 @tab needs to be literally @code{parser} in order to give LilyPond code
129 blocks (@code{#@{}@dots{}@code{#@}}) access to the parser.
131 @item @code{location}
132 @tab needs to be literally @code{location} in order to provide access
133 to the input location object, which is used to provide error messages
134 with file names and line numbers.
136 @item @code{@var{argN}}
137 @tab @var{n}th argument
139 @item @code{@var{typeN?}}
140 @tab a Scheme @emph{type predicate} for which @code{@var{argN}}
141 must return @code{#t}. Some of these predicates are specially
142 recognized by the parser, see below. There is also a special form
143 @code{(@emph{predicate?} @emph{default})} for specifying optional
144 arguments. If the actual argument is missing when the function is being
145 called, the default value is substituted instead. Default values are
146 evaluated at definition time (including LilyPond code blocks!), so if
147 you need a default calculated at runtime, instead write a special value
148 you can easily recognize. If you write the predicate in parentheses but
149 don't follow it with a default value, @code{#f} is used as the default.
150 Default values are not verified with @emph{predicate?} at either
151 definition or run time: it is your responsibility to deal with the
152 values you specify. Default values that happen to be music expressions
153 are copied while setting @code{origin} to the @code{location} parameter.
155 @item @code{@var{body}}
156 @tab A sequence of Scheme forms evaluated in order, the last one being
157 used as the return value of the scheme function. It may contain
158 LilyPond code blocks enclosed in hashed braces
159 (@tie{}@w{@code{#@{@dots{}#@}}}@tie{}), like described in
160 @ref{LilyPond code blocks}. Within LilyPond code blocks, use @code{#}
161 to reference function arguments (eg., @samp{#arg1}) or to start an
162 inline Scheme expression containing function arguments (eg.,
163 @w{@samp{#(cons arg1 arg2)}}). Where normal Scheme expressions using
164 @code{#} don't do the trick, you might need to revert to immediate
165 Scheme expressions using @code{$}, for example as @samp{$music}.
167 If your function returns a music expression, it is given a useful value
172 Some type predicates are specially handled by the parser since it
173 can't recognize the arguments reliably otherwise. Currently these are
174 @code{ly:pitch?} and @code{ly:duration?}.
176 Suitability of arguments for all other predicates is determined by
177 actually calling the predicate after LilyPond has already converted them
178 into a Scheme expression. As a consequence, the argument can be
179 specified in Scheme syntax if desired (introduced with @code{#} or as
180 the result of calling a scheme function), but LilyPond will also convert
181 a number of LilyPond constructs into Scheme before actually checking the
182 predicate on them. Currently, those include music, postevents, simple
183 strings (with or without quotes), numbers, full markups and markup
184 lists, score, book, bookpart, context definition and output definition
187 For some kinds of expression (like most music not enclosed in braces)
188 LilyPond needs to look further than the expression itself in order to
189 determine its end. If such an expression were considered for an
190 optional argument by evaluating its predicate, LilyPond would not be
191 able to ``backup'' when it decides the expression does not fit the
192 parameter. So some forms of music might need to be enclosed in braces
193 to make them acceptable to LilyPond. There are also some other
194 ambiguities that LilyPond sorts out by checking with predicate
195 functions: is @samp{-3} a fingering postevent or a negative number? Is
196 @code{"a" 4} in lyric mode a string followed by a number, or a lyric
197 event of duration @code{4}? LilyPond decides by asking the predicates.
198 That means that you should avoid permissive predicates like
199 @code{scheme?} if you have a particular use in mind instead of a general
202 For a list of available predefined type predicates, see
203 @ruser{Predefined type predicates}.
208 @ruser{Predefined type predicates}.
211 @file{lily/music-scheme.cc},
215 @node Scheme function usage
216 @subsection Scheme function usage
218 Scheme functions can be called pretty much anywhere where a Scheme
219 expression starting with @code{#} can be written. You call a scheme
220 function by writing its name preceded by @code{\}, followed by its
221 arguments. Once an optional argument predicate does not match an
222 argument, LilyPond skips this and all following optional arguments,
223 replacing them with their specified default, and @q{backs up} the
224 argument that did not match to the place of the next mandatory argument.
225 Since the backed up argument needs to go somewhere, optional arguments
226 are not actually considered optional unless followed by a mandatory
229 There is one exception: if you write @code{\default} in the place of an
230 optional argument, this and all following optional arguments are skipped
231 and replaced by their default. This works even when no mandatory
232 argument follows since @code{\default} does not need to get backed up.
233 The @code{mark} and @code{key} commands make use of that trick to
234 provide their default behavior when just followed by @code{\default}.
236 Apart from places where a Scheme value is required, there are a few
237 places where @code{#} expressions are currently accepted and evaluated
238 for their side effects but otherwise ignored. Mostly those are the
239 places where an assignment would be acceptable as well.
241 Since it is a bad idea to return values that can be misinterpreted in
242 some context, you should use normal scheme functions only for those
243 cases where you always return a useful value, and use void scheme
244 functions (@pxref{Void scheme functions}) otherwise.
246 @node Void scheme functions
247 @subsection Void scheme functions
248 @funindex define-void-function
251 Sometimes a procedure is executed in order to perform an action rather
252 than return a value. Some programming languages (like C and Scheme) use
253 functions for either concept and just discard the returned value
254 (usually by allowing any expression to act as statement, ignoring the
255 result). This is clever but error-prone: most C compilers nowadays
256 offer warnings for various non-``void'' expressions being discarded.
257 For many functions executing an action, the Scheme standards declare the
258 return value to be unspecified. LilyPond's Scheme interpreter Guile has
259 a unique value @code{*unspecified*} that it usually (such when using
260 @code{set!} directly on a variable) but unfortunately not consistently
261 returns in such cases.
263 Defining a LilyPond function with @code{define-void-function} makes
264 sure that this special value (the only value satisfying the predicate
265 @code{void?}) will be returned.
269 #(define-void-function
272 (ly:set-option 'point-and-click #f))
274 \noPointAndClick % disable point and click
277 If you want to evaluate an expression only for its side-effect and
278 don't want any value it may return interpreted, you can do so by
279 prefixing it with @code{\void}:
282 \void #(hashq-set! some-table some-key some-value)
285 That way, you can be sure that LilyPond will not assign meaning to the
286 returned value regardless of where it encounters it. This will also
287 work for music functions such as @code{\displayMusic}.
289 @node Music functions
290 @section Music functions
292 @cindex music functions
294 @emph{Music functions} are Scheme procedures that can create music
295 expressions automatically, and can be used to greatly simplify the
299 * Music function definitions::
300 * Music function usage::
301 * Simple substitution functions::
302 * Intermediate substitution functions::
303 * Mathematics in functions::
304 * Functions without arguments::
305 * Void music functions::
309 @node Music function definitions
310 @subsection Music function definitions
311 @cindex defining music functions
312 @funindex define-music-function
314 The general form for defining music functions is:
318 #(define-music-function
319 (parser location @var{arg1} @var{arg2} @dots{})
320 (@var{type1?} @var{type2?} @dots{})
325 quite in analogy to @ref{Scheme function definitions}. More often than
326 not, @var{body} will be a @ref{LilyPond code blocks, LilyPond code block}.
328 For a list of available type predicates, see
329 @ruser{Predefined type predicates}.
334 @ruser{Predefined type predicates}.
337 @file{lily/music-scheme.cc},
342 @node Music function usage
343 @subsection Music function usage
344 Music functions may currently be used in several places. Depending on
345 where they are used, restrictions apply in order to be able to parse
346 them unambiguously. The result a music function returns must be
347 compatible with the context in which it is called.
351 At top level in a music expression. No restriction apply here.
354 As a post-event, explicitly started with a direction indicator (one of
355 @code{-}, @code{^}, @w{and @code{_}}). Note that returning a post-event
356 will be acceptable for music functions called as normal music, leading
357 to a result roughly equivalent to
362 In this case, you can't use an @emph{open} music expression as the last
363 argument, one that would end with a music expression able to accept
364 additional postevents.
367 As a chord constituent. The returned expression must be of
368 @code{rhythmic-event} type, most likely a @code{NoteEvent}.
372 The special rules for trailing arguments make it possible to write
373 polymorphic functions like @code{\tweak} that can be applied to
374 different constructs.
376 @node Simple substitution functions
377 @subsection Simple substitution functions
379 Simple substitution functions are music functions whose output
380 music expression is written in LilyPond format and contains
381 function arguments in the output expression. They are described
382 in @ruser{Substitution function examples}.
385 @node Intermediate substitution functions
386 @subsection Intermediate substitution functions
388 Intermediate substitution functions involve a mix of Scheme code
389 and LilyPond code in the music expression to be returned.
391 Some @code{\override} commands require an argument consisting of
392 a pair of numbers (called a @emph{cons cell} in Scheme).
394 The pair can be directly passed into the music function,
395 using a @code{pair?} variable:
399 #(define-music-function
400 (parser location beg-end)
403 \once \override Beam.positions = #beg-end
407 \manualBeam #'(3 . 6) c8 d e f
411 Alternatively, the numbers making up the pair can be
412 passed as separate arguments, and the Scheme code
413 used to create the pair can be included in the
416 @lilypond[quote,verbatim,ragged-right]
418 #(define-music-function
419 (parser location beg end)
422 \once \override Beam.positions = #(cons beg end)
426 \manualBeam #3 #6 c8 d e f
431 @cindex temporary overrides
432 @cindex overrides, temporary
433 @cindex properties, popping previous value
435 Properties are maintained conceptually using one stack per property
436 per grob per context. Music functions may need to override one or
437 several properties for the duration of the function, restoring them
438 to their previous value before exiting. However, normal overrides
439 pop and discard the top of the current property stack before
440 pushing to it, so the previous value of the property is lost when it
441 is overridden. When the previous value must be preserved, prefix the
442 @code{\override} command with @code{\temporary}, like this:
445 \temporary \override @dots{}
448 The use of @code{\temporary} causes the (usually set) @code{pop-first}
449 property in the override to be cleared, so the previous value is not
450 popped off the property stack before pushing the new value onto it.
451 When a subsequent @code{\revert} pops off the temporarily overriden
452 value, the previous value will re-emerge.
454 In other words, calling @code{\temporary \override} and @code{\revert}
455 in succession on the same property will have a net effect of zero.
456 Similarly, pairing @code{\temporary} and @code{\undo} on the same
457 music containing overrides will have a net effect of zero.
459 Here is an example of a music function which makes use of this.
460 The use of @code{\temporary} ensures the values of the
461 @code{cross-staff} and @code{style} properties are restored on exit
462 to whatever values they had when the @code{crossStaff} function was
463 called. Without @code{\temporary} the default values would have been
468 #(define-music-function (parser location notes) (ly:music?)
469 (_i "Create cross-staff stems")
471 \temporary \override Stem.cross-staff = #cross-staff-connect
472 \temporary \override Flag.style = #'no-flag
474 \revert Stem.cross-staff
480 @node Mathematics in functions
481 @subsection Mathematics in functions
483 Music functions can involve Scheme programming in
484 addition to simple substitution,
486 @lilypond[quote,verbatim,ragged-right]
488 #(define-music-function
489 (parser location mag)
492 \override Stem.length = #(* 7.0 mag)
493 \override NoteHead.font-size =
494 #(inexact->exact (* (/ 6.0 (log 2.0)) (log mag)))
499 \revert NoteHead.font-size
504 \AltOn #1.5 c c \AltOff c2
509 This example may be rewritten to pass in music expressions,
511 @lilypond[quote,verbatim,ragged-right]
513 #(define-music-function
514 (parser location mag music)
517 \override Stem.length = #(* 7.0 mag)
518 \override NoteHead.font-size =
519 #(inexact->exact (* (/ 6.0 (log 2.0)) (log mag)))
522 \revert NoteHead.font-size
526 c2 \withAlt #0.5 { c4 c }
527 \withAlt #1.5 { c c } c2
532 @node Functions without arguments
533 @subsection Functions without arguments
535 In most cases a function without arguments should be written
539 dolce = \markup@{ \italic \bold dolce @}
542 However, in rare cases it may be useful to create a music function
547 #(define-music-function
550 (if (eq? #t (ly:get-option 'display-bar-numbers))
551 #@{ \once \override Score.BarNumber.break-visibility = ##f #@}
555 To actually display bar numbers where this function is called,
556 invoke @command{lilypond} with
559 lilypond -d display-bar-numbers FILENAME.ly
563 @node Void music functions
564 @subsection Void music functions
566 A music function must return a music expression. If you want to
567 execute a function only for its side effect, you should use
568 @code{define-void-function}. But there may be cases where you
569 sometimes want to produce a music expression, and sometimes not (like
570 in the previous example). Returning a @code{void} music expression
571 via @code{#@{ #@}} will achieve that.
573 @node Event functions
574 @section Event functions
575 @funindex define-event-function
576 @cindex event functions
578 To use a music function in the place of an event, you need to write a
579 direction indicator before it. But sometimes, this does not quite match
580 the syntax of constructs you want to replace. For example, if you want
581 to write dynamics commands, those are usually attached without direction
582 indicator, like @code{c'\pp}. Here is a way to write arbitrary
584 @lilypond[quote,verbatim,ragged-right]
585 dyn=#(define-event-function (parser location arg) (markup?)
586 (make-dynamic-script arg))
587 \relative c' { c\dyn pfsss }
589 You could do the same using a music function, but then you always would
590 have to write a direction indicator before calling it, like
591 @code{@w{c-\dyn pfsss}}.
594 @node Markup functions
595 @section Markup functions
597 Markups are implemented as special Scheme functions which produce a
598 @code{Stencil} object given a number of arguments.
601 * Markup construction in Scheme::
602 * How markups work internally::
603 * New markup command definition::
604 * New markup list command definition::
608 @node Markup construction in Scheme
609 @subsection Markup construction in Scheme
611 @cindex defining markup commands
612 @funindex \displayScheme
614 Markup expressions are internally represented in Scheme using the
621 To see a markup expression in its Scheme form, use the
622 @code{\displayScheme} command:
628 \line @{ \bold \italic "hello" \raise #0.4 "world" @}
629 \larger \line @{ foo bar baz @}
635 Compiling the code above will send the following to the display
643 (#:bold (#:italic "hello") #:raise 0.4 "world")
646 (#:simple "foo" #:simple "bar" #:simple "baz")))))
649 To prevent the markup from printing on the page, use
650 @w{@samp{\void \displayScheme @var{markup}}}. Also, as with the
651 @code{\displayMusic} command, the output of @code{\displayScheme}
652 can be saved to an external file. See
653 @ref{Displaying music expressions}.
656 This example demonstrates the main translation rules between regular
657 LilyPond markup syntax and Scheme markup syntax. Using @code{#@{
658 @dots{} #@}} for entering in LilyPond syntax will often be most
659 convenient, but we explain how to use the @code{markup} macro to get a
660 Scheme-only solution.
663 @multitable @columnfractions .3 .3
664 @item @b{LilyPond} @tab @b{Scheme}
665 @item @code{\markup markup1} @tab @code{(markup markup1)}
666 @item @code{\markup @{ markup1 markup2 @dots{} @}} @tab
667 @code{(markup markup1 markup2 @dots{} )}
668 @item @code{\markup-command} @tab @code{#:markup-command}
669 @item @code{\variable} @tab @code{variable}
670 @item @code{\center-column @{ @dots{} @}} @tab
671 @code{#:center-column ( @dots{} )}
672 @item @code{string} @tab @code{"string"}
673 @item @code{#scheme-arg} @tab @code{scheme-arg}
677 The whole Scheme language is accessible inside the
678 @code{markup} macro. For example, You may use function calls inside
679 @code{markup} in order to manipulate character strings. This is
680 useful when defining new markup commands (see
681 @ref{New markup command definition}).
686 The markup-list argument of commands such as @code{#:line},
687 @code{#:center}, and @code{#:column} cannot be a variable or
688 the result of a function call.
691 (markup #:line (function-that-returns-markups))
695 is invalid. One should use the @code{make-line-markup},
696 @code{make-center-markup}, or @code{make-column-markup} functions
700 (markup (make-line-markup (function-that-returns-markups)))
704 @node How markups work internally
705 @subsection How markups work internally
710 \raise #0.5 "text example"
714 @code{\raise} is actually represented by the @code{raise-markup}
715 function. The markup expression is stored as
718 (list raise-markup 0.5 (list simple-markup "text example"))
721 When the markup is converted to printable objects (Stencils), the
722 @code{raise-markup} function is called as
727 @var{list of property alists}
729 @var{the "text example" markup})
732 The @code{raise-markup} function first creates the stencil for the
733 @code{text example} string, and then it raises that Stencil by 0.5
734 staff space. This is a rather simple example; more complex examples
736 of this section, and in @file{scm/define-markup-commands.scm}.
739 @node New markup command definition
740 @subsection New markup command definition
742 This section discusses the definition of new markup commands.
745 * Markup command definition syntax::
747 * A complete example::
748 * Adapting builtin commands::
751 @node Markup command definition syntax
752 @unnumberedsubsubsec Markup command definition syntax
754 New markup commands can be defined using the
755 @code{define-markup-command} Scheme macro, at top-level.
758 (define-markup-command (@var{command-name} @var{layout} @var{props} @var{arg1} @var{arg2} @dots{})
759 (@var{arg1-type?} @var{arg2-type?} @dots{})
760 [ #:properties ((@var{property1} @var{default-value1})
762 @dots{}command body@dots{})
768 @item @var{command-name}
769 the markup command name
771 the @q{layout} definition.
773 a list of associative lists, containing all active properties.
775 @var{i}th command argument
776 @item @var{argi-type?}
777 a type predicate for the i@var{th} argument
780 If the command uses properties from the @code{props} arguments,
781 the @code{#:properties} keyword can be used to specify which
782 properties are used along with their default values.
784 Arguments are distinguished according to their type:
786 @item a markup, corresponding to type predicate @code{markup?};
787 @item a list of markups, corresponding to type predicate
789 @item any other scheme object, corresponding to type predicates such as
790 @code{list?}, @code{number?}, @code{boolean?}, etc.
793 There is no limitation on the order of arguments (after the
794 standard @code{layout} and @code{props} arguments). However,
795 markup functions taking a markup as their last argument are
796 somewhat special as you can apply them to a markup list, and the
797 result is a markup list where the markup function (with the
798 specified leading arguments) has been applied to every element of
799 the original markup list.
801 Since replicating the leading arguments for applying a markup
802 function to a markup list is cheap mostly for Scheme arguments,
803 you avoid performance pitfalls by just using Scheme arguments for
804 the leading arguments of markup functions that take a markup as
809 @funindex interpret-markup
810 Markup commands have a rather complex life cycle. The body of a
811 markup command definition is responsible for converting the
812 arguments of the markup command into a stencil expression which is
813 returned. Quite often this is accomplished by calling the
814 @code{interpret-markup} function on a markup expression, passing
815 the @var{layout} and @var{props} arguments on to it. Those
816 arguments are usually only known at a very late stage in
817 typesetting. Markup expressions have their components assembled
818 into markup expressions already when @code{\markup} in a LilyPond
819 expression or the @code{markup} macro in Scheme is expanded. The
820 evaluation and typechecking of markup command arguments happens at
821 the time @code{\markup}/@code{markup} are interpreted.
823 But the actual conversion of markup expressions into stencil
824 expressions by executing the markup function bodies only happens
825 when @code{interpret-markup} is called on a markup expression.
828 @unnumberedsubsubsec On properties
830 The @code{layout} and @code{props} arguments of markup commands bring a
831 context for the markup interpretation: font size, line width, etc.
833 The @code{layout} argument allows access to properties defined in
834 @code{paper} blocks, using the @code{ly:output-def-lookup} function.
835 For instance, the line width (the same as the one used in scores) is
839 (ly:output-def-lookup layout 'line-width)
842 The @code{props} argument makes some properties accessible to markup
843 commands. For instance, when a book title markup is interpreted, all
844 the variables defined in the @code{\header} block are automatically
845 added to @code{props}, so that the book title markup can access the book
846 title, composer, etc. It is also a way to configure the behaviour of a
847 markup command: for example, when a command uses font size during
848 processing, the font size is read from @code{props} rather than having a
849 @code{font-size} argument. The caller of a markup command may change
850 the value of the font size property in order to change the behaviour.
851 Use the @code{#:properties} keyword of @code{define-markup-command} to
852 specify which properties shall be read from the @code{props} arguments.
854 The example in next section illustrates how to access and override
855 properties in a markup command.
857 @node A complete example
858 @unnumberedsubsubsec A complete example
860 The following example defines a markup command to draw a double box
861 around a piece of text.
863 Firstly, we need to build an approximative result using markups.
864 Consulting the @ruser{Text markup commands} shows us the @code{\box}
867 @lilypond[quote,verbatim,ragged-right]
868 \markup \box \box HELLO
871 Now, we consider that more padding between the text and the boxes is
872 preferable. According to the @code{\box} documentation, this command
873 uses a @code{box-padding} property, which defaults to 0.2. The
874 documentation also mentions how to override it:
876 @lilypond[quote,verbatim,ragged-right]
877 \markup \box \override #'(box-padding . 0.6) \box A
880 Then, the padding between the two boxes is considered too small, so we
883 @lilypond[quote,verbatim,ragged-right]
884 \markup \override #'(box-padding . 0.4) \box
885 \override #'(box-padding . 0.6) \box A
888 Repeating this lengthy markup would be painful. This is where a markup
889 command is needed. Thus, we write a @code{double-box} markup command,
890 taking one argument (the text). This draws the two boxes, with some
894 #(define-markup-command (double-box layout props text) (markup?)
895 "Draw a double box around text."
896 (interpret-markup layout props
897 #@{\markup \override #'(box-padding . 0.4) \box
898 \override #'(box-padding . 0.6) \box @{ #text @}#@}))
904 #(define-markup-command (double-box layout props text) (markup?)
905 "Draw a double box around text."
906 (interpret-markup layout props
907 (markup #:override '(box-padding . 0.4) #:box
908 #:override '(box-padding . 0.6) #:box text)))
911 @code{text} is the name of the command argument, and @code{markup?} its
912 type: it identifies it as a markup. The @code{interpret-markup}
913 function is used in most of markup commands: it builds a stencil, using
914 @code{layout}, @code{props}, and a markup. In the second case, this
915 markup is built using the @code{markup} scheme macro, see @ref{Markup
916 construction in Scheme}. The transformation from @code{\markup}
917 expression to scheme markup expression is straight-forward.
919 The new command can be used as follow:
922 \markup \double-box A
925 It would be nice to make the @code{double-box} command customizable:
926 here, the @code{box-padding} values are hard coded, and cannot be
927 changed by the user. Also, it would be better to distinguish the
928 padding between the two boxes, from the padding between the inner box
929 and the text. So we will introduce a new property,
930 @code{inter-box-padding}, for the padding between the two boxes. The
931 @code{box-padding} will be used for the inner padding. The new code is
935 #(define-markup-command (double-box layout props text) (markup?)
936 #:properties ((inter-box-padding 0.4)
938 "Draw a double box around text."
939 (interpret-markup layout props
940 #@{\markup \override #`(box-padding . ,inter-box-padding) \box
941 \override #`(box-padding . ,box-padding) \box
945 Again, the equivalent version using the markup macro would be:
948 #(define-markup-command (double-box layout props text) (markup?)
949 #:properties ((inter-box-padding 0.4)
951 "Draw a double box around text."
952 (interpret-markup layout props
953 (markup #:override `(box-padding . ,inter-box-padding) #:box
954 #:override `(box-padding . ,box-padding) #:box text)))
957 Here, the @code{#:properties} keyword is used so that the
958 @code{inter-box-padding} and @code{box-padding} properties are read from
959 the @code{props} argument, and default values are given to them if the
960 properties are not defined.
962 Then, these values are used to override the @code{box-padding}
963 properties used by the two @code{\box} commands. Note the backquote and
964 the comma in the @code{\override} argument: they allow you to introduce
965 a variable value into a literal expression.
967 Now, the command can be used in a markup, and the boxes padding be
970 @lilypond[quote,verbatim,ragged-right]
971 #(define-markup-command (double-box layout props text) (markup?)
972 #:properties ((inter-box-padding 0.4)
974 "Draw a double box around text."
975 (interpret-markup layout props
976 #{\markup \override #`(box-padding . ,inter-box-padding) \box
977 \override #`(box-padding . ,box-padding) \box
980 \markup \double-box A
981 \markup \override #'(inter-box-padding . 0.8) \double-box A
982 \markup \override #'(box-padding . 1.0) \double-box A
985 @node Adapting builtin commands
986 @unnumberedsubsubsec Adapting builtin commands
988 A good way to start writing a new markup command, is to take example on
989 a builtin one. Most of the markup commands provided with LilyPond can be
990 found in file @file{scm/define-markup-commands.scm}.
992 For instance, we would like to adapt the @code{\draw-line} command, to
993 draw a double line instead. The @code{\draw-line} command is defined as
994 follow (documentation stripped):
997 (define-markup-command (draw-line layout props dest)
1000 #:properties ((thickness 1))
1001 "@dots{}documentation@dots{}"
1002 (let ((th (* (ly:output-def-lookup layout 'line-thickness)
1006 (make-line-stencil th 0 0 x y)))
1009 To define a new command based on an existing one, copy the definition,
1010 and change the command name. The @code{#:category} keyword can be
1011 safely removed, as it is only used for generating LilyPond
1012 documentation, and is of no use for user-defined markup commands.
1015 (define-markup-command (draw-double-line layout props dest)
1017 #:properties ((thickness 1))
1018 "@dots{}documentation@dots{}"
1019 (let ((th (* (ly:output-def-lookup layout 'line-thickness)
1023 (make-line-stencil th 0 0 x y)))
1026 Then, a property for setting the gap between two lines is added, called
1027 @code{line-gap}, defaulting e.g. to 0.6:
1030 (define-markup-command (draw-double-line layout props dest)
1032 #:properties ((thickness 1)
1034 "@dots{}documentation@dots{}"
1038 Finally, the code for drawing two lines is added. Two calls to
1039 @code{make-line-stencil} are used to draw the lines, and the resulting
1040 stencils are combined using @code{ly:stencil-add}:
1042 @lilypond[quote,verbatim,ragged-right]
1043 #(define-markup-command (my-draw-line layout props dest)
1045 #:properties ((thickness 1)
1048 (let* ((th (* (ly:output-def-lookup layout 'line-thickness)
1052 (w (/ line-gap 2.0))
1053 (x (cond ((= dx 0) w)
1055 (else (/ w (sqrt (+ 1 (* (/ dx dy) (/ dx dy))))))))
1056 (y (* (if (< (* dx dy) 0) 1 -1)
1059 (else (/ w (sqrt (+ 1 (* (/ dy dx) (/ dy dx))))))))))
1060 (ly:stencil-add (make-line-stencil th x y (+ dx x) (+ dy y))
1061 (make-line-stencil th (- x) (- y) (- dx x) (- dy y)))))
1063 \markup \my-draw-line #'(4 . 3)
1064 \markup \override #'(line-gap . 1.2) \my-draw-line #'(4 . 3)
1068 @node New markup list command definition
1069 @subsection New markup list command definition
1070 @funindex define-markup-list-command
1071 @funindex interpret-markup-list
1072 Markup list commands are defined with the
1073 @code{define-markup-list-command} Scheme macro, which is similar to the
1074 @code{define-markup-command} macro described in
1075 @ref{New markup command definition}, except that where the latter returns
1076 a single stencil, the former returns a list of stencils.
1078 In a similar vein, @code{interpret-markup-list} is used instead of
1079 @code{interpret-markup} for converting a markup list into a list
1082 In the following example, a @code{\paragraph} markup list command is
1083 defined, which returns a list of justified lines, the first one being
1084 indented. The indent width is taken from the @code{props} argument.
1087 #(define-markup-list-command (paragraph layout props args) (markup-list?)
1088 #:properties ((par-indent 2))
1089 (interpret-markup-list layout props
1090 #@{\markuplist \justified-lines @{ \hspace #par-indent #args @} #@}))
1094 The version using just Scheme is more complex:
1096 #(define-markup-list-command (paragraph layout props args) (markup-list?)
1097 #:properties ((par-indent 2))
1098 (interpret-markup-list layout props
1099 (make-justified-lines-markup-list (cons (make-hspace-markup par-indent)
1103 Besides the usual @code{layout} and @code{props} arguments, the
1104 @code{paragraph} markup list command takes a markup list argument, named
1105 @code{args}. The predicate for markup lists is @code{markup-list?}.
1107 First, the function gets the indent width, a property here named
1108 @code{par-indent}, from the property list @code{props}. If the
1109 property is not found, the default value is @code{2}. Then, a
1110 list of justified lines is made using the built-in markup list command
1111 @code{\justified-lines}, which is related to the
1112 @code{make-justified-lines-markup-list} function. A
1113 horizontal space is added at the beginning using @code{\hspace} (or the
1114 @code{make-hspace-markup} function). Finally, the markup list is
1115 interpreted using the @code{interpret-markup-list} function.
1117 This new markup list command can be used as follows:
1121 The art of music typography is called \italic @{(plate) engraving.@}
1122 The term derives from the traditional process of music printing.
1123 Just a few decades ago, sheet music was made by cutting and stamping
1124 the music into a zinc or pewter plate in mirror image.
1126 \override-lines #'(par-indent . 4) \paragraph @{
1127 The plate would be inked, the depressions caused by the cutting
1128 and stamping would hold ink. An image was formed by pressing paper
1129 to the plate. The stamping and cutting was completely done by
1135 @node Contexts for programmers
1136 @section Contexts for programmers
1139 * Context evaluation::
1140 * Running a function on all layout objects::
1143 @node Context evaluation
1144 @subsection Context evaluation
1146 @cindex calling code during interpreting
1147 @funindex \applyContext
1149 Contexts can be modified during interpretation with Scheme code. The
1152 \applyContext @var{function}
1155 @code{@var{function}} should be a Scheme function that takes a
1156 single argument: the context in which the @code{\applyContext}
1157 command is being called. The following code will print the
1158 current bar number on the standard output during the compile:
1163 (format #t "\nWe were called in barnumber ~a.\n"
1164 (ly:context-property x 'currentBarNumber)))
1169 @node Running a function on all layout objects
1170 @subsection Running a function on all layout objects
1173 @cindex calling code on layout objects
1174 @funindex \applyOutput
1176 The most versatile way of tuning an object is @code{\applyOutput} which
1177 works by inserting an event into the specified context
1178 (@rinternals{ApplyOutputEvent}). Its syntax is
1180 \applyOutput @var{Context} @var{proc}
1184 where @code{@var{proc}} is a Scheme function, taking three arguments.
1186 When interpreted, the function @code{@var{proc}} is called for
1187 every layout object found in the context @code{@var{Context}} at
1188 the current time step, with the following arguments:
1190 @item the layout object itself,
1191 @item the context where the layout object was created, and
1192 @item the context where @code{\applyOutput} is processed.
1196 In addition, the cause of the layout object, i.e., the music
1197 expression or object that was responsible for creating it, is in the
1198 object property @code{cause}. For example, for a note head, this is a
1199 @rinternals{NoteHead} event, and for a stem object,
1200 this is a @rinternals{Stem} object.
1202 Here is a function to use for @code{\applyOutput}; it blanks
1203 note-heads on the center-line and next to it:
1205 @lilypond[quote,verbatim,ragged-right]
1206 #(define (blanker grob grob-origin context)
1207 (if (and (memq 'note-head-interface (ly:grob-interfaces grob))
1208 (< (abs (ly:grob-property grob 'staff-position)) 2))
1209 (set! (ly:grob-property grob 'transparent) #t)))
1212 a'4 e8 <<\applyOutput #'Voice #blanker a c d>> b2
1216 To have @var{function} interpreted at the @code{Score} or @code{Staff}
1217 level use these forms
1220 \applyOutput #'Score #@var{function}
1221 \applyOutput #'Staff #@var{function}
1225 @node Callback functions
1226 @section Callback functions
1228 Properties (like @code{thickness}, @code{direction}, etc.) can be
1229 set at fixed values with @code{\override}, e.g.
1232 \override Stem.thickness = #2.0
1235 Properties can also be set to a Scheme procedure,
1237 @lilypond[fragment,verbatim,quote,relative=2]
1238 \override Stem.thickness = #(lambda (grob)
1239 (if (= UP (ly:grob-property grob 'direction))
1246 In this case, the procedure is executed as soon as the value of the
1247 property is requested during the formatting process.
1249 Most of the typesetting engine is driven by such callbacks.
1250 Properties that typically use callbacks include
1254 The printing routine, that constructs a drawing for the symbol
1256 The routine that sets the horizontal position
1258 The routine that computes the width of an object
1261 The procedure always takes a single argument, being the grob.
1263 If routines with multiple arguments must be called, the current grob
1264 can be inserted with a grob closure. Here is a setting from
1265 @code{AccidentalSuggestion},
1269 ,(ly:make-simple-closure
1271 ,(ly:make-simple-closure
1272 (list ly:self-alignment-interface::centered-on-x-parent))
1273 ,(ly:make-simple-closure
1274 (list ly:self-alignment-interface::x-aligned-on-self)))))
1278 In this example, both @code{ly:self-alignment-interface::x-aligned-on-self} and
1279 @code{ly:self-alignment-interface::centered-on-x-parent} are called
1280 with the grob as argument. The results are added with the @code{+}
1281 function. To ensure that this addition is properly executed, the whole
1282 thing is enclosed in @code{ly:make-simple-closure}.
1284 In fact, using a single procedure as property value is equivalent to
1287 (ly:make-simple-closure (ly:make-simple-closure (list @var{proc})))
1291 The inner @code{ly:make-simple-closure} supplies the grob as argument
1292 to @var{proc}, the outer ensures that result of the function is
1293 returned, rather than the @code{simple-closure} object.
1295 From within a callback, the easiest method for evaluating a markup is
1296 to use grob-interpret-markup. For example:
1299 my-callback = #(lambda (grob)
1300 (grob-interpret-markup grob (markup "foo")))
1303 @node Inline Scheme code
1304 @section Inline Scheme code
1306 TODO: after this section had been written, LilyPond has improved
1307 to the point that finding a @emph{simple} example where one would
1308 @emph{have} to revert to Scheme has become rather hard.
1310 Until this section gets a rewrite, let's pretend we don't know.
1312 The main disadvantage of @code{\tweak} is its syntactical
1313 inflexibility. For example, the following produces a syntax error
1314 (or rather, it did so at some point in the past).
1317 F = \tweak font-size #-3 -\flageolet
1325 Using Scheme, this problem can be avoided. The route to the
1326 result is given in @ref{Adding articulation to notes (example)},
1327 especially how to use @code{\displayMusic} as a helping guide.
1330 F = #(let ((m (make-music 'ArticulationEvent
1331 'articulation-type "flageolet")))
1332 (set! (ly:music-property m 'tweaks)
1333 (acons 'font-size -3
1334 (ly:music-property m 'tweaks)))
1343 Here, the @code{tweaks} properties of the flageolet object
1344 @code{m} (created with @code{make-music}) are extracted with
1345 @code{ly:music-property}, a new key-value pair to change the
1346 font size is prepended to the property list with the
1347 @code{acons} Scheme function, and the result is finally
1348 written back with @code{set!}. The last element of the
1349 @code{let} block is the return value, @code{m} itself.
1353 @node Difficult tweaks
1354 @section Difficult tweaks
1356 There are a few classes of difficult adjustments.
1362 One type of difficult adjustment involves the appearance of
1363 spanner objects, such as slurs and ties. Usually, only one
1364 spanner object is created at a time, and it can be adjusted with
1365 the normal mechanism. However, occasionally a spanner crosses a
1366 line break. When this happens, the object is cloned. A separate
1367 object is created for every system in which the spanner appears.
1368 The new objects are clones of the original object and inherit all
1369 properties, including @code{\override}s.
1372 In other words, an @code{\override} always affects all pieces of a
1373 broken spanner. To change only one part of a spanner at a line break,
1374 it is necessary to hook into the formatting process. The
1375 @code{after-line-breaking} callback contains the Scheme procedure that
1376 is called after the line breaks have been determined and layout
1377 objects have been split over different systems.
1379 In the following example, we define a procedure
1380 @code{my-callback}. This procedure
1384 determines if the spanner has been split across line breaks
1386 if yes, retrieves all the split objects
1388 checks if this grob is the last of the split objects
1390 if yes, it sets @code{extra-offset}.
1393 This procedure is installed into @rinternals{Tie}, so the last part
1394 of the broken tie is repositioned.
1396 @lilypond[quote,verbatim,ragged-right]
1397 #(define (my-callback grob)
1399 ;; have we been split?
1400 (orig (ly:grob-original grob))
1402 ;; if yes, get the split pieces (our siblings)
1403 (siblings (if (ly:grob? orig)
1404 (ly:spanner-broken-into orig)
1407 (if (and (>= (length siblings) 2)
1408 (eq? (car (last-pair siblings)) grob))
1409 (ly:grob-set-property! grob 'extra-offset '(-2 . 5)))))
1412 \override Tie.after-line-breaking =
1420 When applying this trick, the new @code{after-line-breaking} callback
1421 should also call the old one, if such a default exists. For example,
1422 if using this with @code{Hairpin}, @code{ly:spanner::kill-zero-spanned-time}
1423 should also be called.
1427 Some objects cannot be changed with @code{\override} for
1428 technical reasons. Examples of those are @code{NonMusicalPaperColumn}
1429 and @code{PaperColumn}. They can be changed with the
1430 @code{\overrideProperty} function, which works similar to @code{\once
1431 \override}, but uses a different syntax.
1435 Score.NonMusicalPaperColumn % Grob name
1436 . line-break-system-details % Property name
1437 . next-padding % Optional subproperty name
1441 Note, however, that @code{\override}, applied to
1442 @code{NonMusicalPaperColumn} and @code{PaperColumn}, still works as
1443 expected within @code{\context} blocks.
1447 @node LilyPond Scheme interfaces
1448 @chapter LilyPond Scheme interfaces
1450 This chapter covers the various tools provided by LilyPond to help
1451 Scheme programmers get information into and out of the music streams.
1453 TODO -- figure out what goes in here and how to organize it