1 @c -*- coding: utf-8; mode: texinfo; -*-
3 @chapter Programming work
6 * Overview of LilyPond architecture::
7 * LilyPond programming languages::
8 * Programming without compiling::
11 * Warnings Errors Progress and Debug Output::
12 * Debugging LilyPond::
13 * Tracing object relationships::
14 * Adding or modifying features::
18 * Understanding pure properties::
20 * Scheme->C interface::
21 * LilyPond miscellany::
24 @node Overview of LilyPond architecture
25 @section Overview of LilyPond architecture
27 LilyPond processes the input file into graphical and musical output in a
28 number of stages. This process, along with the types of routines that
29 accomplish the various stages of the process, is described in this section. A
30 more complete description of the LilyPond architecture and internal program
31 execution is found in Erik Sandberg's
32 @uref{http://lilypond.org/website/pdf/thesis-erik-sandberg.pdf, master's
35 The first stage of LilyPond processing is @emph{parsing}. In the parsing
36 process, music expressions in LilyPond input format are converted to music
37 expressions in Scheme format. In Scheme format, a music expression is a list
38 in tree form, with nodes that indicate the relationships between various music
39 events. The LilyPond parser is written in Bison.
41 The second stage of LilyPond processing is @emph{iterating}. Iterating
42 assigns each music event to a context, which is the environment in which the
43 music will be finally engraved. The context is responsible for all further
44 processing of the music. It is during the iteration stage that contexts are
45 created as necessary to ensure that every note has a Voice type context (e.g.
46 Voice, TabVoice, DrumVoice, CueVoice, MensuralVoice, VaticanaVoice,
47 GregorianTranscriptionVoice), that the Voice type contexts exist in
48 appropriate Staff type contexts, and that parallel Staff type contexts exist
49 in StaffGroup type contexts. In addition, during the iteration stage each
50 music event is assigned a moment, or a time in the music when the event
53 Each type of music event has an associated iterator. Iterators are defined in
54 @file{*-iterator.cc}. During iteration, an
55 event's iterator is called to deliver that music event to the appropriate
58 The final stage of LilyPond processing is @emph{translation}. During
59 translation, music events are prepared for graphical or midi output. The
60 translation step is accomplished by the polymorphic base class Translator
61 through its two derived classes: Engraver (for graphical output) and
62 Performer (for midi output).
64 Translators are defined in C++ files named @file{*-engraver.cc}
65 and @file{*-performer.cc}.
66 Much of the work of translating is handled by Scheme functions,
67 which is one of the keys to LilyPond's exceptional flexibility.
69 @sourceimage{architecture-diagram,,,png}
72 @node LilyPond programming languages
73 @section LilyPond programming languages
75 Programming in LilyPond is done in a variety of programming languages. Each
76 language is used for a specific purpose or purposes. This section describes
77 the languages used and provides links to reference manuals and tutorials for
78 the relevant language.
82 The core functionality of LilyPond is implemented in C++.
84 C++ is so ubiquitous that it is difficult to identify either a reference
85 manual or a tutorial. Programmers unfamiliar with C++ will need to spend some
86 time to learn the language before attempting to modify the C++ code.
88 The C++ code calls Scheme/GUILE through the GUILE interface, which is
90 @uref{http://www.gnu.org/software/guile/manual/html_node/index.html, GUILE
95 The LilyPond lexer is implemented in Flex, an implementation of the Unix lex
96 lexical analyser generator. Resources for Flex can be found
97 @uref{http://flex.sourceforge.net/, here}.
101 The LilyPond parser is implemented in Bison, a GNU parser generator. The
102 Bison homepage is found at @uref{http://www.gnu.org/software/bison/,
103 gnu.org}. The manual (which includes both a reference and tutorial) is
104 @uref{http://www.gnu.org/software/bison/manual/index.html, available} in a
109 GNU Make is used to control the compiling process and to build the
110 documentation and the website. GNU Make documentation is available at
111 @uref{http://www.gnu.org/software/make/manual/, the GNU website}.
113 @subsection GUILE or Scheme
115 GUILE is the dialect of Scheme that is used as LilyPond's extension language.
116 Many extensions to LilyPond are written entirely in GUILE. The
117 @uref{http://www.gnu.org/software/guile/manual/html_node/index.html,
118 GUILE Reference Manual} is available online.
120 @uref{http://mitpress.mit.edu/sicp/full-text/book/book.html, Structure and
121 Interpretation of Computer Programs}, a popular textbook used to teach
122 programming in Scheme is available in its entirety online.
124 An introduction to Guile/Scheme as used in LilyPond can be found in the
125 @rextend{Scheme tutorial}.
129 MetaFont is used to create the music fonts used by LilyPond. A MetaFont
130 tutorial is available at @uref{http://metafont.tutorial.free.fr/, the
131 METAFONT tutorial page}.
133 @subsection PostScript
135 PostScript is used to generate graphical output. A brief PostScript tutorial
136 is @uref{http://local.wasp.uwa.edu.au/~pbourke/dataformats/postscript/,
137 available online}. The
138 @uref{http://www.adobe.com/products/postscript/pdfs/PLRM.pdf, PostScript Language
139 Reference} is available online in PDF format.
143 Python is used for XML2ly and is used for building the documentation and the
146 Python documentation is available at @uref{http://www.python.org/doc/,
149 @node Programming without compiling
150 @section Programming without compiling
152 Much of the development work in LilyPond takes place by changing @file{*.ly} or
153 @file{*.scm} files. These changes can be made without compiling LilyPond. Such
154 changes are described in this section.
157 @subsection Modifying distribution files
159 Much of LilyPond is written in Scheme or LilyPond input files. These
160 files are interpreted when the program is run, rather than being compiled
161 when the program is built, and are present in all LilyPond distributions.
162 You will find @file{.ly} files in the @file{ly/} directory and the Scheme files in the
163 @file{scm/} directory. Both Scheme files and @file{.ly} files can be modified and
164 saved with any text editor. It's probably wise to make a backup copy of
165 your files before you modify them, although you can reinstall if the
166 files become corrupted.
168 Once you've modified the files, you can test the changes just by running
169 LilyPond on some input file. It's a good idea to create a file that
170 demonstrates the feature you're trying to add. This file will eventually
171 become a regression test and will be part of the LilyPond distribution.
173 @subsection Desired file formatting
175 Files that are part of the LilyPond distribution have Unix-style line
176 endings (LF), rather than DOS (CR+LF) or MacOS 9 and earlier (CR). Make
177 sure you use the necessary tools to ensure that Unix-style line endings are
178 preserved in the patches you create.
180 Tab characters should not be included in files for distribution. All
181 indentation should be done with spaces. Most editors have settings to
182 allow the setting of tab stops and ensuring that no tab characters are
183 included in the file.
185 Scheme files and LilyPond files should be written according to standard
186 style guidelines. Scheme file guidelines can be found at
187 @uref{http://community.schemewiki.org/?scheme-style}. Following these
188 guidelines will make your code easier to read. Both you and others that
189 work on your code will be glad you followed these guidelines.
191 For LilyPond files, you should follow the guidelines for LilyPond snippets
192 in the documentation. You can find these guidelines at
193 @ref{Texinfo introduction and usage policy}.
195 @node Finding functions
196 @section Finding functions
198 When making changes or fixing bugs in LilyPond, one of the initial
199 challenges is finding out where in the code tree the functions to
200 be modified live. With nearly 3000 files in the source tree,
201 trial-and-error searching is generally ineffective. This section
202 describes a process for finding interesting code.
204 @subsection Using the ROADMAP
206 The file ROADMAP is located in the main directory of the lilypond source.
207 ROADMAP lists all of the directories in the LilyPond source tree, along
208 with a brief description of the kind of files found in each directory.
209 This can be a very helpful tool for deciding which directories to search
210 when looking for a function.
213 @subsection Using grep to search
215 Having identified a likely subdirectory to search, the grep utility can
216 be used to search for a function name. The format of the grep command is
219 grep -i functionName subdirectory/*
222 This command will search all the contents of the directory subdirectory/
223 and display every line in any of the files that contains
224 functionName. The @option{-i} option makes @command{grep} ignore
225 case -- this can be very useful if you are not yet familiar with
226 our capitalization conventions.
228 The most likely directories to grep for function names are @file{scm/} for
229 scheme files, ly/ for lilypond input (@file{*.ly}) files, and @file{lily/} for C++
233 @subsection Using git grep to search
235 If you have used git to obtain the source, you have access to a
236 powerful tool to search for functions. The command:
239 git grep functionName
242 will search through all of the files that are present in the git
243 repository looking for functionName. It also presents the results
244 of the search using @code{less}, so the results are displayed one page
247 @subsection Searching on the git repository at Savannah
249 You can also use the equivalent of git grep on the Savannah server.
254 Go to http://git.sv.gnu.org/gitweb/?p=lilypond.git
257 In the pulldown box that says commit, select grep.
260 Type functionName in the search box, and hit enter/return
264 This will initiate a search of the remote git repository.
270 This section describes style guidelines for LilyPond
277 * Naming conventions::
286 @subsection Languages
288 C++ and Python are preferred. Python code should use PEP 8.
292 @subsection Filenames
294 Definitions of classes that are only accessed via pointers (*) or
295 references (&) shall not be included as include files.
301 ".cc" Implementation files
302 ".icc" Inline definition files
303 ".tcc" non inline Template defs
307 (setq auto-mode-alist
308 (append '(("\\.make$" . makefile-mode)
309 ("\\.cc$" . c++-mode)
310 ("\\.icc$" . c++-mode)
311 ("\\.tcc$" . c++-mode)
312 ("\\.hh$" . c++-mode)
313 ("\\.pod$" . text-mode)
318 The class Class_name is coded in @q{class-name.*}
322 @subsection Indentation
324 Standard GNU coding style is used.
326 @subsubheading Indenting files with @code{fixcc.py} (recommended)
328 LilyPond provides a python script that will adjust the indentation
329 and spacing on a @code{.cc} or @code{.hh} file to very near the
333 scripts/auxiliar/fixcc.py FILENAME
336 This can be run on all files at once, but this is not recommended
337 for normal contributors or developers.
340 scripts/auxiliar/fixcc.py \
341 $(find flower lily -name '*cc' -o -name '*hh' | grep -v /out)
345 @subsubheading Indenting with emacs
347 The following hooks will produce indentation which is similar to
348 our official indentation as produced with @code{fixcc.py}.
351 (add-hook 'c++-mode-hook
354 (setq indent-tabs-mode nil))
357 If you like using font-lock, you can also add this to your
361 (setq font-lock-maximum-decoration t)
362 (setq c++-font-lock-keywords-3
364 c++-font-lock-keywords-3
365 '(("\\b\\(a-zA-Z_?+_\\)\\b" 1 font-lock-variable-name-face) ("\\b\\(A-Z?+a-z_?+\\)\\b" 1 font-lock-type-face))
370 @subsubheading Indenting with vim
372 Although emacs indentation is the GNU standard, correct
373 indentation for C++ files can be achieved by using the settings
375 @url{https://gcc.gnu.org/wiki/FormattingCodeForGCC, GNU GCC Wiki}.
376 Save the following in @file{~/.vim/after/ftplugin/cpp.vim}:
380 setlocal cinoptions=>4,n-2,@{2,^-2,:2,=2,g0,h2,p5,t0,+2,(0,u0,w1,m1
381 setlocal shiftwidth=2
382 setlocal softtabstop=2
383 setlocal textwidth=79
384 setlocal fo-=ro fo+=cql
385 " use spaces instead of tabs
387 " remove trailing whitespace on write
388 autocmd BufWritePre * :%s/\s\+$//e
391 With these settings, files can be reindented automatically by
392 highlighting the lines to be indented in visual mode (use V to
393 enter visual mode) and pressing @code{=}, or a single line
394 correctly indented in normal mode by pressing @code{==}.
396 A @file{scheme.vim} file will help improve the indentation of
397 Scheme code. This one was suggested by Patrick McCarty. It
398 should be saved in @file{~/.vim/after/syntax/scheme.vim}.
401 " Additional Guile-specific 'forms'
402 syn keyword schemeSyntax define-public define*-public
403 syn keyword schemeSyntax define* lambda* let-keywords*
404 syn keyword schemeSyntax defmacro defmacro* define-macro
405 syn keyword schemeSyntax defmacro-public defmacro*-public
406 syn keyword schemeSyntax use-modules define-module
407 syn keyword schemeSyntax define-method define-class
409 " Additional LilyPond-specific 'forms'
410 syn keyword schemeSyntax define-markup-command define-markup-list-command
411 syn keyword schemeSyntax define-safe-public define-music-function
412 syn keyword schemeSyntax def-grace-function
414 " All of the above should influence indenting too
415 setlocal lw+=define-public,define*-public
416 setlocal lw+=define*,lambda*,let-keywords*
417 setlocal lw+=defmacro,defmacro*,define-macro
418 setlocal lw+=defmacro-public,defmacro*-public
419 setlocal lw+=use-modules,define-module
420 setlocal lw+=define-method,define-class
421 setlocal lw+=define-markup-command,define-markup-list-command
422 setlocal lw+=define-safe-public,define-music-function
423 setlocal lw+=def-grace-function
425 " These forms should not influence indenting
429 " Try to highlight all ly: procedures
430 syn match schemeFunc "ly:[^) ]\+"
433 For documentation work on texinfo files, identify the file
434 extensions used as texinfo files in your @file{.vim/filetype.vim}:
437 if exists("did_load_filetypes")
440 augroup filetypedetect
441 au! BufRead,BufNewFile *.itely setfiletype texinfo
442 au! BufRead,BufNewFile *.itexi setfiletype texinfo
443 au! BufRead,BufNewFile *.tely setfiletype texinfo
447 and add these settings in @file{.vim/after/ftplugin/texinfo.vim}:
451 setlocal shiftwidth=2
452 setlocal textwidth=66
455 @node Naming conventions
456 @subsection Naming Conventions
458 Naming conventions have been established for LilyPond
461 @subheading Classes and Types
463 Classes begin with an uppercase letter, and words
464 in class names are separated with @code{_}:
472 Member variable names end with an underscore:
480 Macro names should be written in uppercase completely,
481 with words separated by @code{_}:
487 @subheading Variables
489 Variable names should be complete words, rather than abbreviations.
490 For example, it is preferred to use @code{thickness} rather than
491 @code{th} or @code{t}.
493 Multi-word variable names in C++ should have the words separated
494 by the underscore character (@q{_}):
497 cxx_multiword_variable
500 Multi-word variable names in Scheme should have the words separated
504 scheme-multiword-variable
508 @subsection Broken code
510 Do not write broken code. This includes hardwired dependencies,
511 hardwired constants, slow algorithms and obvious limitations. If
512 you can not avoid it, mark the place clearly, and add a comment
513 explaining shortcomings of the code.
515 Ideally, the comment marking the shortcoming would include
516 TODO, so that it is marked for future fixing.
518 We reject broken-in-advance on principle.
522 @subsection Code comments
524 Comments may not be needed if descriptive variable names are used
525 in the code and the logic is straightforward. However, if the
526 logic is difficult to follow, and particularly if non-obvious
527 code has been included to resolve a bug, a comment describing
528 the logic and/or the need for the non-obvious code should be included.
530 There are instances where the current code could be commented better.
531 If significant time is required to understand the code as part of
532 preparing a patch, it would be wise to add comments reflecting your
533 understanding to make future work easier.
536 @node Handling errors
537 @subsection Handling errors
539 As a general rule, you should always try to continue computations,
540 even if there is some kind of error. When the program stops, it
541 is often very hard for a user to pinpoint what part of the input
542 causes an error. Finding the culprit is much easier if there is
543 some viewable output.
545 So functions and methods do not return errorcodes, they never
546 crash, but report a programming_error and try to carry on.
548 Error and warning messages need to be localized.
552 @subsection Localization
554 This document provides some guidelines to help programmers write
556 messages. To help translations, user messages must follow
557 uniform conventions. Follow these rules when coding for LilyPond.
558 Hopefully, this can be replaced by general GNU guidelines in the
559 future. Even better would be to have an English (en_BR, en_AM)
560 guide helping programmers writing consistent messages for all GNU
563 Non-preferred messages are marked with `+'. By convention,
564 ungrammatical examples are marked with `*'. However, such ungrammatical
565 examples may still be preferred.
570 Every message to the user should be localized (and thus be marked
571 for localization). This includes warning and error messages.
574 Do not localize/gettextify:
578 `programming_error ()'s
581 `programming_warning ()'s
587 output strings (PostScript, TeX, etc.)
592 Messages to be localized must be encapsulated in `_ (STRING)' or
593 `_f (FORMAT, ...)'. E.g.:
596 warning (_ ("need music in a score"));
597 error (_f ("cannot open file: `%s'", file_name));
600 In some rare cases you may need to call `gettext ()' by hand. This
601 happens when you pre-define (a list of) string constants for later
602 use. In that case, you'll probably also need to mark these string
603 constants for translation, using `_i (STRING)'. The `_i' macro is
604 a no-op, it only serves as a marker for `xgettext'.
607 char const* messages[] = @{
608 _i ("enable debugging output"),
609 _i ("ignore lilypond version"),
616 puts (gettext (messages i));
620 See also @file{flower/getopt-long.cc} and @file{lily/main.cc}.
623 Do not use leading or trailing whitespace in messages. If you need
624 whitespace to be printed, prepend or append it to the translated
628 message ("Calculating line breaks..." + " ");
632 Error or warning messages displayed with a file name and line
633 number never start with a capital, eg,
636 foo.ly: 12: not a duration: 3
639 Messages containing a final verb, or a gerund (`-ing'-form) always
640 start with a capital. Other (simpler) messages start with a
646 Not declaring: `foo'.
650 Avoid abbreviations or short forms, use `cannot' and `do not'
651 rather than `can't' or `don't'
652 To avoid having a number of different messages for the same
653 situation, well will use quoting like this `"message: `%s'"' for all
654 strings. Numbers are not quoted:
657 _f ("cannot open file: `%s'", name_str)
658 _f ("cannot find character number: %d", i)
662 Think about translation issues. In a lot of cases, it is better to
663 translate a whole message. English grammar must not be imposed on the
664 translator. So, instead of
667 stem at + moment.str () + does not fit in beam
673 _f ("stem at %s does not fit in beam", moment.str ())
677 Split up multi-sentence messages, whenever possible. Instead of
680 warning (_f ("out of tune! Can't find: `%s'", "Key_engraver"));
681 warning (_f ("cannot find font `%s', loading default", font_name));
687 warning (_ ("out of tune:"));
688 warning (_f ("cannot find: `%s', "Key_engraver"));
689 warning (_f ("cannot find font: `%s', font_name));
690 warning (_f ("Loading default font"));
694 If you must have multiple-sentence messages, use full punctuation.
695 Use two spaces after end of sentence punctuation. No punctuation
696 (esp. period) is used at the end of simple messages.
699 _f ("Non-matching braces in text `%s', adding braces", text)
700 _ ("Debug output disabled. Compiled with NPRINT.")
701 _f ("Huh? Not a Request: `%s'. Ignoring.", request)
705 Do not modularize too much; words frequently cannot be translated
706 without context. It is probably safe to treat most occurrences of
707 words like stem, beam, crescendo as separately translatable words.
710 When translating, it is preferable to put interesting information
711 at the end of the message, rather than embedded in the middle.
712 This especially applies to frequently used messages, even if this
713 would mean sacrificing a bit of eloquence. This holds for original
714 messages too, of course.
717 en: cannot open: `foo.ly'
718 + nl: kan `foo.ly' niet openen (1)
719 kan niet openen: `foo.ly'* (2)
720 niet te openen: `foo.ly'* (3)
724 The first nl message, although grammatically and stylistically
725 correct, is not friendly for parsing by humans (even if they speak
726 dutch). I guess we would prefer something like (2) or (3).
729 Do not run make po/po-update with GNU gettext < 0.10.35
734 @node Warnings Errors Progress and Debug Output
735 @section Warnings, Errors, Progress and Debug Output
737 @unnumberedsubsec Available log levels
739 LilyPond has several loglevels, which specify how verbose the output on
740 the console should be:
742 @item NONE: No output at all, even on failure
743 @item ERROR: Only error messages
744 @item WARN: Only error messages and warnings
745 @item BASIC_PROGRESS: Warnings, errors and basic progress (success, etc.)
746 @item PROGRESS: Warnings, errors and full progress messages
747 @item INFO: Warnings, errors, progress and more detailed information (default)
748 @item DEBUG: All messages, including full debug messages (very verbose!)
751 The loglevel can either be set with the environment variable
752 @code{LILYPOND_LOGLEVEL} or on the command line with the @option{--loglevel=...}
755 @unnumberedsubsec Functions for debug and log output
757 LilyPond has two different types of error and log functions:
761 If a warning or error is caused by an identified position in the input file,
762 e.g. by a grob or by a music expression, the functions of the @code{Input}
763 class provide logging functionality that prints the position of the message
764 in addition to the message.
767 If a message can not be associated with a particular position in an input file,
768 e.g. the output file cannot be written, then the functions in the
769 @code{flower/include/warn.hh} file will provide logging functionality that
770 only prints out the message, but no location.
774 There are also Scheme functions to access all of these logging functions from
775 scheme. In addition, the Grob class contains some convenience wrappers for
776 even easier access to these functions.
778 The message and debug functions in @code{warn.hh} also have an optional
779 argument @code{newline}, which specifies whether the message should always
780 start on a new line or continue a previous message.
781 By default, @code{progress_indication} does NOT start on a new line, but rather
782 continue the previous output. They also do not have a particular input
783 position associated, so there are no progress functions in the Input class.
784 All other functions by default start their output on a new line.
786 The error functions come in three different flavors: fatal error messages,
787 programming error messages and normal error messages. Errors written
788 by the @code{error ()} function will cause LilyPond to exit immediately,
789 errors by @code{Input::error ()} will continue the compilation, but
790 return a non-zero return value of the LilyPond call (i.e. indicate an
791 unsuccessful program execution). All other errors will be printed on the
792 console, but not exit LilyPond or indicate an unsuccessful return code.
793 Their only differences to a warnings are the displayed text and that
794 they will be shown with loglevel @code{ERROR}.
796 If the Scheme option @code{warning-as-error} is set, any warning will be
797 treated as if @code{Input::error} was called.
800 @unnumberedsubsec All logging functions at a glance
802 @multitable @columnfractions 0.16 0.42 0.42
804 @tab C++, no location
805 @tab C++ from input location
808 @tab @code{error ()}, @code{programming_error (msg)}, @code{non_fatal_error (msg)}
809 @tab @code{Input::error (msg)}, @code{Input::programming_error (msg)}
812 @tab @code{warning (msg)}
813 @tab @code{Input::warning (msg)}
816 @tab @code{basic_progress (msg)}
820 @tab @code{progress_indication (msg)}
824 @tab @code{message (msg)}
825 @tab @code{Input::message (msg)}
828 @tab @code{debug_output (msg)}
829 @tab @code{Input::debug_output (msg)}
835 @tab Scheme, music expression
838 @tab @code{Grob::programming_error (msg)}
842 @tab @code{Grob::warning (msg)}
843 @tab @code{(ly:music-warning music msg)}
855 @tab @code{(ly:music-message music msg)}
864 @tab Scheme, no location
865 @tab Scheme, input location
869 @tab @code{(ly:error msg args)}, @code{(ly:programming-error msg args)}
872 @tab @code{(ly:warning msg args)}
873 @tab @code{(ly:input-warning input msg args)}
876 @tab @code{(ly:basic-progress msg args)}
880 @tab @code{(ly:progress msg args)}
884 @tab @code{(ly:message msg args)}
885 @tab @code{(ly:input-message input msg args)}
888 @tab @code{(ly:debug msg args)}
896 @node Debugging LilyPond
897 @section Debugging LilyPond
899 The most commonly used tool for debugging LilyPond is the GNU
900 debugger gdb. The gdb tool is used for investigating and debugging
901 core LilyPond code written in C++. Another tool is available for
902 debugging Scheme code using the Guile debugger. This section
903 describes how to use both gdb and the Guile Debugger.
906 * Debugging overview::
907 * Debugging C++ code::
908 * Debugging Scheme code::
911 @node Debugging overview
912 @subsection Debugging overview
914 Using a debugger simplifies troubleshooting in at least two ways.
916 First, breakpoints can be set to pause execution at any desired point.
917 Then, when execution has paused, debugger commands can be issued to
918 explore the values of various variables or to execute functions.
920 Second, the debugger can display a stack trace, which shows the
921 sequence in which functions have been called and the arguments
922 passed to the called functions.
924 @node Debugging C++ code
925 @subsection Debugging C++ code
927 The GNU debugger, gdb, is the principal tool for debugging C++ code.
929 @subheading Compiling LilyPond for use with gdb
931 In order to use gdb with LilyPond, it is necessary to compile
932 LilyPond with debugging information. This is the current default
933 mode of compilation. Often debugging becomes more complicated
934 when the compiler has optimised variables and function calls away.
935 In that case it may be helpful to run the following command in the
936 main LilyPond source directory:
939 ./configure --disable-optimising
943 This will create a version of LilyPond with minimal optimization
944 which will allow the debugger to access all variables and step
945 through the source code in-order. It may not accurately reproduce
946 bugs encountered with the optimized version, however.
948 You should not do @var{make install} if you want to use a debugger
949 with LilyPond. The @var{make install} command will strip debugging
950 information from the LilyPond binary.
952 @subheading Typical gdb usage
954 Once you have compiled the LilyPond image with the necessary
955 debugging information it will have been written to a location in a
956 subfolder of your current working directory:
962 This is important as you will need to let gdb know where to find the
963 image containing the symbol tables. You can invoke gdb from the
964 command line using the following:
970 This loads the LilyPond symbol tables into gdb. Then, to run
971 LilyPond on @file{test.ly} under the debugger, enter the following:
980 As an alternative to running gdb at the command line you may try
981 a graphical interface to gdb such as ddd:
987 You can also use sets of standard gdb commands stored in a .gdbinit
988 file (see next section).
990 @subheading Typical .gdbinit files
992 The behavior of gdb can be readily customized through the use of a
993 @var{.gdbinit} file. A @var{.gdbinit} file is a file named
994 @var{.gdbinit} (notice the @qq{.} at the beginning of the file name)
995 that is placed in a user's home directory.
997 The @var{.gdbinit} file below is from Han-Wen. It sets breakpoints
998 for all errors and defines functions for displaying scheme objects
999 (ps), grobs (pgrob), and parsed music expressions (pmusic).
1002 file $LILYPOND_GIT/build/out/bin/lilypond
1004 b Grob::programming_error
1007 print ly_display_scm($arg0)
1010 print ly_display_scm($arg0->self_scm_)
1011 print ly_display_scm($arg0->mutable_property_alist_)
1012 print ly_display_scm($arg0->immutable_property_alist_)
1013 print ly_display_scm($arg0->object_alist_)
1016 print ly_display_scm($arg0->self_scm_)
1017 print ly_display_scm($arg0->mutable_property_alist_)
1018 print ly_display_scm($arg0->immutable_property_alist_)
1022 @node Debugging Scheme code
1023 @subsection Debugging Scheme code
1025 Scheme code can be developed using the Guile command line
1026 interpreter @code{top-repl}. You can either investigate
1027 interactively using just Guile or you can use the debugging
1028 tools available within Guile.
1030 @subheading Using Guile interactively with LilyPond
1032 In order to experiment with Scheme programming in the LilyPond
1033 environment, it is necessary to have a Guile interpreter that
1034 has all the LilyPond modules loaded. This requires the following
1037 First, define a Scheme symbol for the active module in the @file{.ly} file:
1040 #(module-define! (resolve-module '(guile-user))
1041 'lilypond-module (current-module))
1044 Now place a Scheme function in the @file{.ly} file that gives an
1045 interactive Guile prompt:
1051 When the @file{.ly} file is compiled, this causes the compilation to be
1052 interrupted and an interactive guile prompt to appear. Once the
1053 guile prompt appears, the LilyPond active module must be set as the
1054 current guile module:
1057 guile> (set-current-module lilypond-module)
1060 You can demonstrate these commands are operating properly by typing the name
1061 of a LilyPond public scheme function to check it has been defined:
1064 guile> fret-diagram-verbose-markup
1065 #<procedure fret-diagram-verbose-markup (layout props marking-list)>
1068 If the LilyPond module has not been correctly loaded, an error
1069 message will be generated:
1072 guile> fret-diagram-verbose-markup
1073 ERROR: Unbound variable: fret-diagram-verbose-markup
1074 ABORT: (unbound-variable)
1077 Once the module is properly loaded, any valid LilyPond Scheme
1078 expression can be entered at the interactive prompt.
1080 After the investigation is complete, the interactive guile
1081 interpreter can be exited:
1087 The compilation of the @file{.ly} file will then continue.
1089 @subheading Using the Guile debugger
1091 To set breakpoints and/or enable tracing in Scheme functions, put
1094 \include "guile-debugger.ly"
1097 in your input file after any scheme procedures you have defined in
1098 that file. This will invoke the Guile command-line after having set
1099 up the environment for the debug command-line. When your input file
1100 is processed, a guile prompt will be displayed. You may now enter
1101 commands to set up breakpoints and enable tracing by the Guile debugger.
1103 @subheading Using breakpoints
1105 At the guile prompt, you can set breakpoints with
1106 the @code{set-break!} procedure:
1109 guile> (set-break! my-scheme-procedure)
1112 Once you have set the desired breakpoints, you exit the guile repl frame
1119 Then, when one of the scheme routines for which you have set
1120 breakpoints is entered, guile will interrupt execution in a debug
1121 frame. At this point you will have access to Guile debugging
1122 commands. For a listing of these commands, type:
1128 Alternatively you may code the breakpoints in your LilyPond source
1129 file using a command such as:
1132 #(set-break! my-scheme-procedure)
1135 immediately after the @code{\include} statement. In this case the
1136 breakpoint will be set straight after you enter the @code{(quit)}
1137 command at the guile prompt.
1139 Embedding breakpoint commands like this is particularly useful if
1140 you want to look at how the Scheme procedures in the @file{.scm}
1141 files supplied with LilyPond work. To do this, edit the file in
1142 the relevant directory to add this line near the top:
1145 (use-modules (scm guile-debugger))
1148 Now you can set a breakpoint after the procedure you are interested
1149 in has been declared. For example, if you are working on routines
1150 called by @var{print-book-with} in @file{lily-library.scm}:
1153 (define (print-book-with book process-procedure)
1154 (let* ((paper (ly:parser-lookup '$defaultpaper))
1155 (layout (ly:parser-lookup '$defaultlayout))
1156 (outfile-name (get-outfile-name book)))
1157 (process-procedure book paper layout outfile-name)))
1159 (define-public (print-book-with-defaults book)
1160 (print-book-with book ly:book-process))
1162 (define-public (print-book-with-defaults-as-systems book)
1163 (print-book-with book ly:book-process-to-systems))
1166 At this point in the code you could add this to set a breakpoint at
1170 (set-break! print-book-with)
1173 @subheading Tracing procedure calls and evaluator steps
1175 Two forms of trace are available:
1178 (set-trace-call! my-scheme-procedure)
1184 (set-trace-subtree! my-scheme-procedure)
1187 @code{set-trace-call!} causes Scheme to log a line to the standard
1188 output to show when the procedure is called and when it exits.
1190 @code{set-trace-subtree!} traces every step the Scheme evaluator
1191 performs in evaluating the procedure.
1193 @node Tracing object relationships
1194 @section Tracing object relationships
1196 Understanding the LilyPond source often boils down to figuring out what
1197 is happening to the Grobs. Where (and why) are they being created,
1198 modified and destroyed? Tracing Lily through a debugger in order to
1199 identify these relationships can be time-consuming and tedious.
1201 In order to simplify this process, a facility has been added to
1202 display the grobs that are created and the properties that are set
1203 and modified. Although it can be complex to get set up, once set up
1204 it easily provides detailed information about the life of grobs
1205 in the form of a network graph.
1207 Each of the steps necessary to use the graphviz utility
1212 @item Installing graphviz
1214 In order to create the graph of the object relationships, it is
1215 first necessary to install Graphviz. Graphviz is available for a
1216 number of different platforms:
1219 @uref{http://www.graphviz.org/Download..php}
1222 @item Modifying config.make
1224 In order for the Graphviz tool to work, config.make must be modified.
1225 It is probably a good idea to first save a copy of config.make under
1228 In order to have the required functionality available, LilyPond
1229 needs to be compiled with the option @option{-DDEBUG}. You can
1230 achieve this by configuring with
1233 ./configure --enable-checking
1236 @item Rebuilding LilyPond
1238 The executable code of LilyPond must be rebuilt from scratch:
1244 @item Create a graphviz-compatible @file{.ly} file
1246 In order to use the graphviz utility, the @file{.ly} file must include
1247 @file{ly/graphviz-init.ly}, and should then specify the
1248 grobs and symbols that should be tracked. An example of this
1249 is found in @file{input/regression/graphviz.ly}.
1251 @item Run LilyPond with output sent to a log file
1253 The Graphviz data is sent to stderr by LilyPond, so it is
1254 necessary to redirect stderr to a logfile:
1257 lilypond graphviz.ly 2> graphviz.log
1260 @item Edit the logfile
1262 The logfile has standard LilyPond output, as well as the Graphviz
1263 output data. Delete everything from the beginning of the file
1264 up to but not including the first occurrence of @code{digraph}.
1266 Also, delete the final LilyPond message about success from the end
1269 @item Process the logfile with @code{dot}
1271 The directed graph is created from the log file with the program
1275 dot -Tpdf graphviz.log > graphviz.pdf
1280 The pdf file can then be viewed with any pdf viewer.
1282 When compiled with @option{-DDEBUG}, LilyPond may run slower
1283 than normal. The original configuration can be restored by rerunning
1284 @code{./configure} with @option{--disable-checking}. Then
1285 rebuild LilyPond with
1292 @node Adding or modifying features
1293 @section Adding or modifying features
1295 When a new feature is to be added to LilyPond, it is necessary to
1296 ensure that the feature is properly integrated to maintain
1297 its long-term support. This section describes the steps necessary
1298 for feature addition and modification.
1303 * Write regression tests::
1304 * Write convert-ly rule::
1305 * Automatically update documentation::
1306 * Manually update documentation::
1307 * Edit changes.tely::
1308 * Verify successful build::
1309 * Verify regression tests::
1310 * Post patch for comments::
1312 * Closing the issues::
1315 @node Write the code
1316 @subsection Write the code
1318 You should probably create a new git branch for writing the code, as that
1319 will separate it from the master branch and allow you to continue
1320 to work on small projects related to master.
1322 Please be sure to follow the rules for programming style discussed
1323 earlier in this chapter.
1326 @node Write regression tests
1327 @subsection Write regression tests
1329 In order to demonstrate that the code works properly, you will
1330 need to write one or more regression tests. These tests are
1331 typically @file{.ly} files that are found in @file{input/regression}.
1333 Regression tests should be as brief as possible to demonstrate the
1334 functionality of the code.
1336 Regression tests should generally cover one issue per test. Several
1337 short, single-issue regression tests are preferred to a single, long,
1338 multiple-issue regression test.
1340 If the change in the output is small or easy to overlook, use bigger
1341 staff size -- 40 or more (up to 100 in extreme cases). Size 30 means
1342 "pay extra attention to details in general".
1344 Use existing regression tests as templates to demonstrate the type of
1345 header information that should be included in a regression test.
1348 @node Write convert-ly rule
1349 @subsection Write convert-ly rule
1351 If the modification changes the input syntax, a convert-ly rule
1352 should be written to automatically update input files from older
1355 convert-ly rules are found in python/convertrules.py
1357 If possible, the convert-ly rule should allow automatic updating
1358 of the file. In some cases, this will not be possible, so the
1359 rule will simply point out to the user that the feature needs
1362 @subsubheading Updating version numbers
1364 If a development release occurs between you writing your patch and
1365 having it approved+pushed, you will need to update the version
1366 numbers in your tree. This can be done with:
1369 scripts/auxiliar/update-patch-version old.version.number new.version.number
1372 It will change all files in git, so use with caution and examine
1376 @node Automatically update documentation
1377 @subsection Automatically update documentation
1379 @command{convert-ly} should be used to update the documentation,
1380 the snippets, and the regression tests. This not only makes the
1381 necessary syntax changes, it also tests the @command{convert-ly}
1384 The automatic updating is performed by moving to the top-level
1385 source directory, then running:
1388 scripts/auxiliar/update-with-convert-ly.sh
1391 If you did an out-of-tree build, pass in the relative path:
1394 LILYPOND_BUILD_DIR=../build-lilypond/ scripts/auxiliar/update-with-convert-ly.sh
1398 @node Manually update documentation
1399 @subsection Manually update documentation
1401 Where the convert-ly rule is not able to automatically update the inline
1402 LilyPond code in the documentation (i.e. if a NOT_SMART rule is used), the
1403 documentation must be manually updated. The inline snippets that require
1404 changing must be changed in the English version of the docs and all
1405 translated versions. If the inline code is not changed in the
1406 translated documentation, the old snippets will show up in the
1407 English version of the documentation.
1409 Where the convert-ly rule is not able to automatically update snippets
1410 in Documentation/snippets/, those snippets must be manually updated.
1411 Those snippets should be copied to Documentation/snippets/new. The
1412 comments at the top of the snippet describing its automatic generation
1413 should be removed. All translated texidoc strings should be removed.
1414 The comment @qq{% begin verbatim} should be removed. The syntax of
1415 the snippet should then be manually edited.
1417 Where snippets in Documentation/snippets are made obsolete, the snippet
1418 should be copied to Documentation/snippets/new. The comments and
1419 texidoc strings should be removed as described above. Then the body
1420 of the snippet should be changed to:
1424 This snippet is deprecated as of version X.Y.Z and
1425 will be removed from the documentation.
1430 where X.Y.Z is the version number for which the convert-ly rule was
1433 Update the snippet files by running:
1436 scripts/auxiliar/makelsr.py
1439 Where the convert-ly rule is not able to automatically update regression
1440 tests, the regression tests in input/regression should be manually
1443 Although it is not required, it is helpful if the developer
1444 can write relevant material for inclusion in the Notation
1445 Reference. If the developer does not feel qualified to write
1446 the documentation, a documentation editor will be able to
1447 write it from the regression tests. In this case the developer
1448 should raise a new issue with the Type=Documentation tag containing
1449 a reference to the original issue number and/or the committish of
1450 the pushed patch so that the need for new documention is not
1453 Any text that is added to or removed from the documentation should
1454 be changed only in the English version.
1457 @node Edit changes.tely
1458 @subsection Edit changes.tely
1460 An entry should be added to Documentation/changes.tely to describe
1461 the feature changes to be implemented. This is especially important
1462 for changes that change input file syntax.
1464 Hints for changes.tely entries are given at the top of the file.
1466 New entries in changes.tely go at the top of the file.
1468 The changes.tely entry should be written to show how the new change
1469 improves LilyPond, if possible.
1472 @node Verify successful build
1473 @subsection Verify successful build
1475 When the changes have been made, successful completion must be
1483 When these commands complete without error, the patch is
1484 considered to function successfully.
1486 Developers on Windows who are unable to build LilyPond should
1487 get help from a GNU/Linux or OSX developer to do the make tests.
1490 @node Verify regression tests
1491 @subsection Verify regression tests
1493 In order to avoid breaking LilyPond, it is important to verify that
1494 the regression tests succeed, and that no unwanted changes are
1495 introduced into the output. This process is described in
1496 @ref{Regtest comparison}.
1498 @subheading Typical developer's edit/compile/test cycle
1506 make [-j@var{X} CPU_COUNT=@var{X}] test-baseline
1507 make [-j@var{X} CPU_COUNT=@var{X}] check
1511 Edit/compile/test cycle:
1514 @emph{## edit source files, then...}
1516 make clean @emph{## only if needed (see below)}
1517 make [-j@var{X}] @emph{## only if needed (see below)}
1518 make [-j@var{X} CPU_COUNT=@var{X}] test-redo @emph{## redo files differing from baseline}
1519 make [-j@var{X} CPU_COUNT=@var{X}] check
1530 If you modify any source files that have to be compiled (such as
1531 @file{.cc} or @file{.hh} files in @file{flower/} or @file{lily/}),
1532 then you must run @command{make} before @command{make test-redo},
1533 so @command{make} can compile the modified files and relink all
1534 the object files. If you only modify files which are interpreted,
1535 like those in the @file{scm/} and @file{ly/} directories, then
1536 @command{make} is not needed before @command{make test-redo}.
1538 Also, if you modify any font definitions in the @file{mf/}
1539 directory then you must run @command{make clean} and
1540 @command{make} before running @command{make test-redo}. This will
1541 recompile everything, whether modified or not, and takes a lot
1544 Running @command{make@tie{}check} will leave an HTML page
1545 @file{out/test-results/index.html}. This page shows all the
1546 important differences that your change introduced, whether in the
1547 layout, MIDI, performance or error reporting.
1549 You only need to use @command{make test-clean} to start from
1550 scratch, prior to running @command{make@tie{}test-baseline}. To
1551 check new modifications, all that is needed is to repeat
1552 @command{make@tie{}test-redo} and @command{make@tie{}test-check}
1553 (not forgetting @command{make} if needed).
1558 @node Post patch for comments
1559 @subsection Post patch for comments
1561 See @ref{Uploading a patch for review}.
1565 @subsection Push patch
1567 Once all the comments have been addressed, the patch can be pushed.
1569 If the author has push privileges, the author will push the patch.
1570 Otherwise, a developer with push privileges will push the patch.
1573 @node Closing the issues
1574 @subsection Closing the issues
1576 Once the patch has been pushed, all the relevant issues should be
1579 On Rietveld, the author should log in and close the issue either by
1580 using the @q{Edit Issue} link, or by clicking the circled x icon
1581 to the left of the issue name.
1583 If the changes were in response to a feature request on the Google
1584 issue tracker for LilyPond, the author should change the status to
1585 Fixed and a tag @q{fixed_x_y_z} should be added, where the patch was
1586 fixed in version x.y.z. If
1587 the author does not have privileges to change the status, an email
1588 should be sent to bug-lilypond requesting the BugMeister to change
1592 @node Iterator tutorial
1593 @section Iterator tutorial
1595 TODO -- this is a placeholder for a tutorial on iterators
1597 Iterators are routines written in C++ that process music expressions
1598 and sent the music events to the appropriate engravers and/or
1601 See a short example discussing iterators and their duties in
1602 @ref{Articulations on EventChord}.
1605 @node Engraver tutorial
1606 @section Engraver tutorial
1608 Engravers are C++ classes that catch music events and
1609 create the appropriate grobs for display on the page. Though the
1610 majority of engravers are responsible for the creation of a single grob,
1611 in some cases (e.g. @code{New_fingering_engraver}), several different grobs
1614 Engravers listen for events and acknowledge grobs. Events are passed to
1615 the engraver in time-step order during the iteration phase. Grobs are
1616 made available to the engraver when they are created by other engravers
1617 during the iteration phase.
1621 * Useful methods for information processing::
1622 * Translation process::
1623 * Preventing garbage collection for SCM member variables::
1624 * Listening to music events::
1625 * Acknowledging grobs::
1626 * Engraver declaration/documentation::
1629 @node Useful methods for information processing
1630 @subsection Useful methods for information processing
1632 An engraver inherits the following public methods from the Translator
1633 base class, which can be used to process listened events and acknowledged
1637 @item @code{virtual void initialize ()}
1638 @item @code{void start_translation_timestep ()}
1639 @item @code{void process_music ()}
1640 @item @code{void process_acknowledged ()}
1641 @item @code{void stop_translation_timestep ()}
1642 @item @code{virtual void finalize ()}
1645 These methods are listed in order of translation time, with
1646 @code{initialize ()} and @code{finalize ()} bookending the whole
1647 process. @code{initialize ()} can be used for one-time initialization
1648 of context properties before translation starts, whereas
1649 @code{finalize ()} is often used to tie up loose ends at the end of
1650 translation: for example, an unterminated spanner might be completed
1651 automatically or reported with a warning message.
1654 @node Translation process
1655 @subsection Translation process
1657 At each timestep in the music, translation proceeds by calling the
1658 following methods in turn:
1660 @code{start_translation_timestep ()} is called before any user
1661 information enters the translators, i.e., no property operations
1662 (\set, \override, etc.) or events have been processed yet.
1664 @code{process_music ()} and @code{process_acknowledged ()} are called
1665 after all events in the current time step have been heard, or all
1666 grobs in the current time step have been acknowledged. The latter
1667 tends to be used exclusively with engravers which only acknowledge
1668 grobs, whereas the former is the default method for main processing
1671 @code{stop_translation_timestep ()} is called after all user
1672 information has been processed prior to beginning the translation for
1676 @node Preventing garbage collection for SCM member variables
1677 @subsection Preventing garbage collection for SCM member variables
1679 In certain cases, an engraver might need to ensure private Scheme
1680 variables (with type SCM) do not get swept away by Guile's garbage
1681 collector: for example, a cache of the previous key signature which
1682 must persist between timesteps. The method
1683 @code{virtual derived_mark () const} can be used in such cases:
1686 Engraver_name::derived_mark ()
1688 scm_gc_mark (private_scm_member_)
1693 @node Listening to music events
1694 @subsection Listening to music events
1696 External interfaces to the engraver are implemented by protected
1697 macros including one or more of the following:
1700 @item @code{DECLARE_TRANSLATOR_LISTENER (event_name)}
1701 @item @code{IMPLEMENT_TRANSLATOR_LISTENER (Engraver_name, event_name)}
1705 where @var{event_name} is the type of event required to provide the
1706 input the engraver needs and @var{Engraver_name} is the name of the
1709 Following declaration of a listener, the method is implemented as follows:
1712 IMPLEMENT_TRANSLATOR_LISTENER (Engraver_name, event_name)
1714 Engraver_name::listen_event_name (Stream event *event)
1716 ...body of listener method...
1721 @node Acknowledging grobs
1722 @subsection Acknowledging grobs
1724 Some engravers also need information from grobs as they are created
1725 and as they terminate. The mechanism and methods to obtain this
1726 information are set up by the macros:
1729 @item @code{DECLARE_ACKNOWLEDGER (grob_interface)}
1730 @item @code{DECLARE_END_ACKNOWLEDGER (grob_interface)}
1733 where @var{grob_interface} is an interface supported by the
1734 grob(s) which should be acknowledged. For example, the following
1735 code would declare acknowledgers for a @code{NoteHead} grob (via the
1736 @code{note-head-interface}) and any grobs which support the
1737 @code{side-position-interface}:
1740 DECLARE_ACKNOWLEDGER (note_head)
1741 DECLARE_ACKNOWLEDGER (side_position)
1744 The @code{DECLARE_END_ACKNOWLEDGER ()} macro sets up a spanner-specific
1745 acknowledger which will be called whenever a spanner ends.
1747 Following declaration of an acknowledger, the method is coded as follows:
1751 Engraver_name::acknowledge_interface_name (Grob_info info)
1753 ...body of acknowledger method...
1757 Acknowledge functions are called in the order engravers are
1758 @code{\consist}-ed (the only exception is if you set
1759 @code{must-be-last} to @code{#t}).
1761 If useful things are to be done to the acknowledged grobs, this
1762 should be deferred until all the acknowledging has finished, i.e.,
1763 store the acknowledged grobs and process the information in a
1764 @code{process-acknowledged ()} or @code{stop-translation-timestep ()}
1768 @node Engraver declaration/documentation
1769 @subsection Engraver declaration/documentation
1771 An engraver must have a public macro
1774 @item @code{TRANSLATOR_DECLARATIONS (Engraver_name)}
1778 where @code{Engraver_name} is the name of the engraver. This
1779 defines the common variables and methods used by every engraver.
1781 At the end of the engraver file, one or both of the following
1782 macros are generally called to document the engraver in the
1783 Internals Reference:
1786 @item @code{ADD_ACKNOWLEDGER (Engraver_name, grob_interface)}
1787 @item @code{ADD_TRANSLATOR (Engraver_name, Engraver_doc,
1788 Engraver_creates, Engraver_reads, Engraver_writes)}
1792 where @code{Engraver_name} is the name of the engraver, @code{grob_interface}
1793 is the name of the interface that will be acknowledged,
1794 @code{Engraver_doc} is a docstring for the engraver,
1795 @code{Engraver_creates} is the set of grobs created by the engraver,
1796 @code{Engraver_reads} is the set of properties read by the engraver,
1797 and @code{Engraver_writes} is the set of properties written by
1800 The @code{ADD_ACKNOWLEDGER} and @code{ADD_TRANSLATOR} macros use a
1801 non-standard indentation system. Each interface, grob, read property,
1802 and write property is on its own line, and the closing parenthesis
1803 and semicolon for the macro all occupy a separate line beneath the final
1804 interface or write property. See existing engraver files for more
1808 @node Callback tutorial
1809 @section Callback tutorial
1811 TODO -- This is a placeholder for a tutorial on callback functions.
1814 @node Understanding pure properties
1815 @section Understanding pure properties
1818 * Purity in LilyPond::
1819 * Writing a pure function::
1820 * How purity is defined and stored::
1821 * Where purity is used::
1826 Pure properties are some of the most difficult properties to understand
1827 in LilyPond but, once understood, it is much easier to work with
1828 horizontal spacing. This document provides an overview of what it means
1829 for something to be @q{pure} in LilyPond, what this purity guarantees,
1830 and where pure properties are stored and used. It finishes by
1831 discussing a few case studies for the pure programmer to save you some
1832 time and to prevent you some major headaches.
1835 @node Purity in LilyPond
1836 @subsection Purity in LilyPond
1837 Pure properties in LilyPond are properties that do not have any
1839 That is, looking up a pure property should never result in calls to the
1840 following functions:
1842 @item @code{set_property}
1843 @item @code{set_object}
1844 @item @code{suicide}
1846 This means that, if the property is calculated via a callback, this callback
1847 must not only avoid the functions above but make sure that any functions
1848 it calls also avoid the functions above. Also, to date in LilyPond, a pure
1849 function will always return the same value before line breaking (or, more
1850 precisely, before any version of @code{break_into_pieces} is called). This
1851 convention makes it possible to cache pure functions and be more flexible
1852 about the order in which functions are called. For example; Stem #'length has
1853 a pure property that will @emph{never} trigger one of the functions listed
1854 above and will @emph{always} return the same value before line breaking,
1855 independent of where it is called. Sometimes, this will be the actual length
1856 of the Stem. But sometimes it will not. For example; stem that links up
1857 with a beam will need its end set to the Y position of the beam at the stem's
1858 X position. However, the beam's Y positions can only be known after the score
1859 is broken up in to several systems (a beam that has a shallow slope on a
1860 compressed line of music, for example, may have a steeper one on an
1861 uncompressed line). Thus, we only call the impure version of the properties
1862 once we are @emph{absolutely certain} that all of the parameters needed to
1863 calculate their final value have been calculated. The pure version provides a
1864 useful estimate of what this Stem length (or any property) will be, and
1865 the art of creating good pure properties is trying to get the estimation
1866 as close to the actual value as possible.
1868 Of course, like Gregory Peck and Tintin, some Grobs will have properties
1869 that will always be pure. For example, the height of a note-head in
1870 not-crazy music will never depend on line breaking or other parameters
1871 decided late in the typesetting process. Inversely, in rare cases,
1872 certain properties are difficult to estimate with pure values. For
1873 example, the height of a Hairpin at a certain cross-section of its
1874 horizontal span is difficult to know without knowing the horizontal
1875 distance that the hairpin spans, and LilyPond provides an
1876 over-estimation by reporting the pure height as the entire height of the
1879 Purity, like for those living in a convent, is more like a contract than
1880 an @emph{a priori}. If you write a pure-function, you are promising
1881 the user (and the developer who may have to clean up after you) that
1882 your function will not be dependent on factors that change at different
1883 stages of the compilation process (compilation of a score, not of
1886 One last oddity is that purity, in LilyPond, is currently limited
1887 exclusively to things that have to do with Y-extent and positioning.
1888 There is no concept of @q{pure X} as, by design, X is always the
1889 independent variable (i.e. from column X1 to column X2, what will be the
1890 Y height of a given grob). Furthermore, there is no purity for
1891 properties like color, text, and other things for which a meaningful notion
1892 of estimation is either not necessary or has not yet been found. For example,
1893 even if a color were susceptible to change at different points of the
1894 compilation process, it is not clear what a pure estimate of this color
1895 would be or how this pure color could be used. Thus, in this document and
1896 in the source, you will see purity discussed almost interchangeably with
1897 Y-axis positioning issues.
1900 @node Writing a pure function
1901 @subsection Writing a pure function
1902 Pure functions take, at a minimum, three arguments: the @var{grob}, the
1903 starting column at which the function is being evaluated (hereafter
1904 referred to as @var{start}), and the end column at which the grob is
1905 being evaluated (hereafter referred to as @var{end}). For items,
1906 @var{start} and @var{end} must be provided (meaning they are not optional)
1907 but will not have a meaningful impact on the result, as items only occupy
1908 one column and will thus yield a value or not (if they are not in the range
1909 from @var{start} to @var{end}). For spanners however, @var{start} and
1910 @var{end} are important, as we may can get a better pure estimation of a
1911 slice of the spanner than considering it on the whole. This is useful
1912 during line breaking, for example, when we want to estimate the Y-extent
1913 of a spanner broken at given starting and ending columns.
1916 @node How purity is defined and stored
1917 @subsection How purity is defined and stored
1918 Purity is defined in LilyPond with the creation of an unpure-pure container
1919 (unpure is not a word, but hey, neither was LilyPond until the 90s). For example:
1925 #(define (bar grob start end)
1928 \override Stem #'length = #(ly:make-unpure-pure-container foo bar)
1931 Note that items can only ever have two pure heights: their actual pure height
1932 if they are between @q{start} and @q{end}, or an empty interval if they are
1933 not. Thus, their pure property is cached to speed LilyPond up. Pure
1934 heights for spanners are generally not cached as they change depending
1935 on the start and end values. They are only cached in certain particular
1936 cases. Before writing a lot of caching code, make sure that it is a
1937 value that will be reused a lot.
1940 @node Where purity is used
1941 @subsection Where purity is used
1942 Pure Y values must be used in any functions that are called before
1943 line breaking. Examples of this can be seen in
1944 @code{Separation_items::boxes} to construct horizontal skylines and in
1945 @code{Note_spacing::stem_dir_correction} to correct for optical
1946 illusions in spacing. Pure properties are also used in the calculation
1947 of other pure properties. For example, the @code{Axis_group_interface}
1948 has pure functions that look up other pure functions.
1950 Purity is also implicitly used in any functions that should only ever
1951 return pure values. For example, extra-spacing-height is only ever used
1952 before line-breaking and thus should never use values that would only be
1953 available after line breaking. In this case, there is no need to create
1954 callbacks with pure equivalents because these functions, by design, need
1957 To know if a property will be called before and/or after line-breaking
1958 is sometimes tricky and can, like all things in coding, be found by
1959 using a debugger and/or adding @var{printf} statements to see where they
1960 are called in various circumstances.
1964 @subsection Case studies
1965 In each of these case studies, we expose a problem in pure properties, a
1966 solution, and the pros and cons of this solution.
1968 @subheading Time signatures
1969 A time signature needs to prevent accidentals from passing over or under
1970 it, but its extent does not necessarily extend to the Y-position of
1971 accidentals. LilyPond's horizontal spacing sometimes makes a line of
1972 music compact and, when doing so, allows certain columns to pass over
1973 each other if they will not collide. This type of passing over is not
1974 desirable with time signatures in traditional engraving. But how do we
1975 know if this passing over will happen before line breaking, as we are
1976 not sure what the X positions will be? We need a pure estimation of how
1977 much extra spacing height the time signatures would need to prevent this
1978 form of passing over without making this height so large as to
1979 overly-distort the Y-extent of an system, which could result in a very
1980 @q{loose} looking score with lots of horizontal space between columns.
1981 So, to approximate this extra spacing height, we use the Y-extent of a
1982 time signature's next-door-neighbor grobs via the pure-from-neighbor
1986 @item pros: By extending the extra spacing height of a time signature to
1987 that of its next-door-neighbors, we make sure that grobs to the right of
1988 it that could pass above or below it do not.
1990 @item cons: This over-estimation of the vertical height could prevent
1991 snug vertical spacing of systems, as the system will be registered as
1992 being taller at the point of the time signature than it actually is.
1993 This approach can be used for clefs and bar lines as well.
1997 As described above, Stems need pure height approximations when they are
1998 beamed, as we do not know the beam positions before line breaking. To
1999 estimate this pure height, we take all the stems in a beam and find
2000 their pure heights as if they were not beamed. Then, we find the union
2001 of all these pure heights and take the intersection between this
2002 interval (which is large) and an interval going from the note-head of a
2003 stem to infinity in the direction of the stem so that the interval stops
2007 @item pros: This is guaranteed to be at least as long as the beamed
2008 stem, as a beamed stem will never go over the ideal length of the
2009 extremal beam of a stem.
2011 @item cons: Certain stems will be estimated as being too long, which
2012 leads to the same problem of too-much-vertical-height as described
2018 @node Debugging tips
2019 @subsection Debugging tips
2020 A few questions to ask yourself when working with pure properties:
2023 @item Is the property really pure? Are you sure that its value could
2024 not be changed later in the compiling process due to other changes?
2026 @item Can the property be made to correspond even more exactly with the
2027 eventual impure property?
2029 @item For a spanner, is the pure property changing correctly depending
2030 on the starting and ending points of the spanner?
2032 @item For an Item, will the item's pure height need to act in horizontal
2033 spacing but not in vertical spacing? If so, use extra-spacing-height
2034 instead of pure height.
2039 @node LilyPond scoping
2040 @section LilyPond scoping
2042 The LilyPond language has a concept of scoping, i.e. you can do:
2048 (display (+ foo 2)))
2051 @noindent with @code{\paper}, @code{\midi} and @code{\header} being
2052 nested scope inside the @file{.ly} file-level scope. @w{@code{foo = 1}}
2053 is translated in to a scheme variable definition.
2055 This implemented using modules, with each scope being an anonymous
2056 module that imports its enclosing scope's module.
2058 LilyPond's core, loaded from @file{.scm} files, is usually placed in the
2059 @code{lily} module, outside the @file{.ly} level. In the case of
2066 we want to reuse the built-in definitions, without changes effected in
2067 user-level @file{a.ly} leaking into the processing of @file{b.ly}.
2069 The user-accessible definition commands have to take care to avoid
2070 memory leaks that could occur when running multiple files. All
2071 information belonging to user-defined commands and markups is stored in
2072 a manner that allows it to be garbage-collected when the module is
2073 dispersed, either by being stored module-locally, or in weak hash
2077 @node Scheme->C interface
2078 @section Scheme->C interface
2080 Most of the C functions interfacing with Guile/Scheme used in LilyPond
2081 are described in the API Reference of the
2082 @uref{http://www.gnu.org/software/guile/manual/html_node/index.html,
2083 GUILE Reference Manual}.
2085 The remaining functions are defined in @file{lily/lily-guile.cc},
2086 @file{lily/include/lily-guile.hh} and
2087 @file{lily/include/lily-guile-macros.hh}.
2088 Although their names are meaningful there's a few things you should know
2097 @subsection Comparison
2099 This is the trickiest part of the interface.
2101 Mixing Scheme values with C comparison operators won't produce any crash
2102 or warning when compiling but must be avoided:
2105 scm_string_p (scm_value) == SCM_BOOL_T
2108 As we can read in the reference, @code{scm_string_p} returns a Scheme
2109 value: either @code{#t} or @code{#f} which are written @code{SCM_BOOL_T}
2110 and @code{SCM_BOOL_F} in C. This will work, but it is not following
2111 to the API guidelines. For further information, read this discussion:
2114 @uref{http://lists.gnu.org/archive/html/lilypond-devel/2011-08/msg00646.html}
2117 There are functions in the Guile reference that returns C values
2118 instead of Scheme values. In our example, a function called
2119 @code{scm_is_string} (described after @code{string?} and @code{scm_string_p})
2120 returns the C value 0 or 1.
2122 So the best solution was simply:
2125 scm_is_string (scm_value)
2128 There a simple solution for almost every common comparison. Another example:
2129 we want to know if a Scheme value is a non-empty list. Instead of:
2132 (scm_is_true (scm_list_p (scm_value)) && scm_value != SCM_EOL)
2135 one can usually use:
2138 scm_is_pair (scm_value)
2141 since a list of at least one member is a pair. This test is
2142 cheap; @code{scm_list_p} is actually quite more complex since it makes
2143 sure that its argument is neither a `dotted list' where the last pair
2144 has a non-null @code{cdr}, nor a circular list. There are few
2145 situations where the complexity of those tests make sense.
2147 Unfortunately, there is not a @code{scm_is_[something]} function for
2148 everything. That's one of the reasons why LilyPond has its own Scheme
2149 interface. As a rule of thumb, tests that are cheap enough to be
2150 worth inlining tend to have such a C interface. So there is
2151 @code{scm_is_pair} but not @code{scm_is_list}, and @code{scm_is_eq}
2152 but not @code{scm_is_equal}.
2154 @subheading General definitions
2156 @subsubheading bool to_boolean (SCM b)
2158 Return @code{true} if @var{b} is @code{SCM_BOOL_T}, else return @code{false}.
2160 This should be used instead of @code{scm_is_true} and
2161 @code{scm_is_false} for properties since in LilyPond, unset properties
2162 are read as an empty list, and by convention unset Boolean properties
2163 default to false. Since both @code{scm_is_true} and
2164 @code{scm_is_false} only compare with @code{##f} in line with what
2165 Scheme's conditionals do, they are not really useful for checking the
2166 state of a Boolean property.
2168 @subsubheading bool ly_is_[something] (args)
2170 Behave the same as scm_is_[something] would do if it existed.
2172 @subsubheading bool is_[type] (SCM s)
2174 Test whether the type of @var{s} is [type].
2175 [type] is a LilyPond-only set of values (direction, axis...). More
2176 often than not, the code checks LilyPond specific C++-implemented
2179 @subsubheading [Type *] unsmob<Type> (SCM s)
2181 This tries converting a Scheme object to a pointer of the desired
2182 kind. If the Scheme object is of the wrong type, a pointer value
2183 of@w{ }@code{0} is returned, making this suitable for a Boolean test.
2186 @subsection Conversion
2188 @subheading General definitions
2190 @subsubheading bool to_boolean (SCM b)
2192 Return @code{true} if @var{b} is @code{SCM_BOOL_T}, else return @code{false}.
2194 This should be used instead of @code{scm_is_true} and @code{scm_is_false}
2195 for properties since empty lists are sometimes used to unset them.
2197 @subsubheading [C type] ly_scm2[C type] (SCM s)
2199 Behave the same as scm_to_[C type] would do if it existed.
2201 @subsubheading [C type] robust_scm2[C type] (SCM s, [C type] d)
2203 Behave the same as scm_to_[C type] would do if it existed.
2204 Return @var{d} if type verification fails.
2207 @node LilyPond miscellany
2208 @section LilyPond miscellany
2210 This is a place to dump information that may be of use to developers
2211 but doesn't yet have a proper home. Ideally, the length of this section
2212 would become zero as items are moved to other homes.
2216 * Spacing algorithms::
2217 * Info from Han-Wen email::
2218 * Music functions and GUILE debugging::
2219 * Articulations on EventChord::
2222 @node Spacing algorithms
2223 @subsection Spacing algorithms
2225 Here is information from an email exchange about spacing algorithms.
2227 On Thu, 2010-02-04 at 15:33 -0500, Boris Shingarov wrote:
2228 I am experimenting with some modifications to the line breaking code,
2229 and I am stuck trying to understand how some of it works. So far my
2230 understanding is that Simple_spacer operates on a vector of Grobs, and
2231 it is a well-known Constrained-QP problem (rods = constraints, springs
2232 = quadratic function to minimize). What I don't understand is, if the
2233 spacer operates at the level of Grobs, which are built at an earlier
2234 stage in the pipeline, how are the changes necessitated by differences
2235 in line breaking, taken into account? in other words, if I take the
2236 last measure of a line and place it on the next line, it is not just a
2237 matter of literally moving that graphic to where the start of the next
2238 line is, but I also need to draw a clef, key signature, and possibly
2239 other fundamental things -- but at that stage in the rendering
2240 pipeline, is it not too late??
2242 Joe Neeman answered:
2244 We create lots of extra grobs (eg. a BarNumber at every bar line) but
2245 most of them are not drawn. See the break-visibility property in
2248 Here is another e-mail exchange. Janek Warchoł asked for a starting point
2249 to fixing 1301 (change clef colliding with notes). Neil Puttock replied:
2251 The clef is on a loose column (it floats before the head), so the
2252 first place I'd look would be lily/spacing-loose-columns.cc (and
2253 possibly lily/spacing-determine-loose-columns.cc).
2254 I'd guess the problem is the way loose columns are spaced between
2255 other columns: in this snippet, the columns for the quaver and tuplet
2256 minim are so close together that the clef's column gets dumped on top
2257 of the quaver (since it's loose, it doesn't influence the spacing).
2259 @node Info from Han-Wen email
2260 @subsection Info from Han-Wen email
2262 In 2004, Douglas Linhardt decided to try starting a document that would
2263 explain LilyPond architecture and design principles. The material below
2264 is extracted from that email, which can be found at
2265 @uref{http://thread.gmane.org/gmane.comp.gnu.lilypond.devel/2992}.
2266 The headings reflect questions from Doug or comments from Han-Wen;
2267 the body text are Han-Wen's answers.
2269 @subheading Figuring out how things work.
2271 I must admit that when I want to know how a program works, I use grep
2272 and emacs and dive into the source code. The comments and the code
2273 itself are usually more revealing than technical documents.
2275 @subheading What's a grob, and how is one used?
2277 Graphical object - they are created from within engravers, either as
2278 Spanners (derived class) -slurs, beams- or Items (also a derived
2279 class) -notes, clefs, etc.
2281 There are two other derived classes System (derived from Spanner,
2282 containing a "line of music") and Paper_column (derived from Item, it
2283 contains all items that happen at the same moment). They are separate
2284 classes because they play a special role in the linebreaking process.
2286 @subheading What's a smob, and how is one used?
2288 A C(++) object that is encapsulated so it can be used as a Scheme
2289 object. See GUILE info, "19.3 Defining New Types (Smobs)"
2291 @subheading When is each C++ class constructed and used?
2298 In the parser.yy see the macro calls MAKE_MUSIC_BY_NAME().
2303 Constructed during "interpreting" phase.
2308 Executive branch of Contexts, plugins that create grobs, usually one
2309 engraver per grob type. Created together with context.
2319 These are not C++ classes per se. The idea of a Grob interface hasn't
2320 crystallized well. ATM, an interface is a symbol, with a bunch of grob
2321 properties. They are not objects that are created or destroyed.
2326 Objects that walk through different music classes, and deliver events
2327 in a synchronized way, so that notes that play together are processed
2328 at the same moment and (as a result) end up on the same horizontal position.
2330 Created during interpreting phase.
2332 BTW, the entry point for interpreting is ly:run-translator
2333 (ly_run_translator on the C++ side)
2337 @subheading Can you get to Context properties from a Music object?
2339 You can create music object with a Scheme function that reads context
2340 properties (the \applycontext syntax). However, that function is
2341 executed during Interpreting, so you can not really get Context
2342 properties from Music objects, since music objects are not directly
2343 connected to Contexts. That connection is made by the Music_iterators
2345 @subheading Can you get to Music properties from a Context object?
2347 Yes, if you are given the music object within a Context
2348 object. Normally, the music objects enter Contexts in synchronized
2349 fashion, and the synchronization is done by Music_iterators.
2351 @subheading What is the relationship between C++ classes and Scheme objects?
2353 Smobs are C++ objects in Scheme. Scheme objects (lists, functions) are
2354 manipulated from C++ as well using the GUILE C function interface
2357 @subheading How do Scheme procedures get called from C++ functions?
2359 scm_call_*, where * is an integer from 0 to 4.
2360 Also scm_c_eval_string (), scm_eval ()
2362 @subheading How do C++ functions get called from Scheme procedures?
2364 Export a C++ function to Scheme with LY_DEFINE.
2366 @subheading What is the flow of control in the program?
2368 Good question. Things used to be clear-cut, but we have Scheme
2369 and SMOBs now, which means that interactions do not follow a very
2370 rigid format anymore. See below for an overview, though.
2372 @subheading Does the parser make Scheme procedure calls or C++ function calls?
2374 Both. And the Scheme calls can call C++ and vice versa. It's nested,
2375 with the SCM datatype as lubrication between the interactions
2377 (I think the word "lubrication" describes the process better than the
2378 traditional word "glue")
2380 @subheading How do the front-end and back-end get started?
2382 Front-end: a file is parsed, the rest follows from that. Specifically,
2384 Parsing leads to a Music + Music_output_def object (see parser.yy,
2385 definition of toplevel_expression )
2387 A Music + Music_output_def object leads to a Global_context object (see
2388 ly_run_translator ())
2390 During interpreting, Global_context + Music leads to a bunch of
2391 Contexts (see Global_translator::run_iterator_on_me ()).
2393 After interpreting, Global_context contains a Score_context (which
2394 contains staves, lyrics etc.) as a child. Score_context::get_output ()
2395 spews a Music_output object (either a Paper_score object for notation
2396 or Performance object for MIDI).
2398 The Music_output object is the entry point for the backend (see
2399 ly_render_output ()).
2401 The main steps of the backend itself are in
2406 @file{paper-score.cc} , Paper_score::process_
2409 @file{system.cc} , System::get_lines()
2412 The step, where things go from grobs to output, is in
2413 System::get_line(): each grob delivers a Stencil (a Device
2414 independent output description), which is interpreted by our
2415 outputting backends (@file{scm/output-tex.scm} and
2416 @file{scm/output-ps.scm}) to produce TeX and PS.
2420 Interactions between grobs and putting things into .tex and .ps files
2421 have gotten a little more complex lately. Jan has implemented
2422 page-breaking, so now the backend also involves Paper_book,
2423 Paper_lines and other things. This area is still heavily in flux, and
2424 perhaps not something you should want to look at.
2426 @subheading How do the front-end and back-end communicate?
2428 There is no communication from backend to front-end. From front-end to
2429 backend is simply the program flow: music + definitions gives
2430 contexts, contexts yield output, after processing, output is written
2433 @subheading Where is the functionality associated with KEYWORDs?
2435 See @file{my-lily-lexer.cc} (keywords, there aren't that many)
2436 and @file{ly/*.ly} (most of the other backslashed @code{/\words} are identifiers)
2438 @subheading What Contexts/Properties/Music/etc. are available when they are processed?
2440 What do you mean exactly with this question?
2442 See @file{ly/engraver-init.ly} for contexts,
2443 see @file{scm/define-*.scm} for other objects.
2445 @subheading How do you decide if something is a Music, Context, or Grob property?
2446 Why is part-combine-status a Music property when it seems (IMO)
2447 to be related to the Staff context?
2449 The Music_iterators and Context communicate through two channels
2451 Music_iterators can set and read context properties, idem for
2452 Engravers and Contexts
2454 Music_iterators can send "synthetic" music events (which aren't in
2455 the input) to a context. These are caught by Engravers. This is
2456 mostly a one way communication channel.
2458 part-combine-status is part of such a synthetic event, used by
2459 Part_combine_iterator to communicate with Part_combine_engraver.
2462 @subheading Deciding between context and music properties
2464 I'm adding a property to affect how \autochange works. It seems to
2465 me that it should be a context property, but the Scheme autochange
2466 procedure has a Music argument. Does this mean I should use
2469 \autochange is one of these extra strange beasts: it requires
2470 look-ahead to decide when to change staves. This is achieved by
2471 running the interpreting step twice (see
2472 @file{scm/part-combiner.scm} , at the bottom), and
2473 storing the result of the first step (where to switch
2474 staves) in a Music property. Since you want to influence that
2475 where-to-switch list, your must affect the code in
2476 make-autochange-music (@file{scm/part-combiner.scm}).
2477 That code is called directly from the parser and there are no
2478 official "parsing properties" yet, so there is no generic way
2479 to tune \autochange. We would have to invent something new
2480 for this, or add a separate argument,
2483 \autochange #around-central-C ..music..
2487 where around-central-C is some function that is called from
2488 make-autochange-music.
2490 @subheading More on context and music properties
2492 From Neil Puttock, in response to a question about transposition:
2494 Context properties (using \set & \unset) are tied to engravers: they
2495 provide information relevant to the generation of graphical objects.
2497 Since transposition occurs at the music interpretation stage, it has
2498 no direct connection with engravers: the pitch of a note is fixed
2499 before a notehead is created. Consider the following minimal snippet:
2505 This generates (simplified) a NoteEvent, with its pitch and duration
2506 as event properties,
2512 (ly:make-duration 2 0 1 1)
2514 (ly:make-pitch 0 0 0)
2517 which the Note_heads_engraver hears. It passes this information on to
2518 the NoteHead grob it creates from the event, so the head's correct
2519 position and duration-log can be determined once it's ready for
2522 If we transpose the snippet,
2525 \transpose c d @{ c' @}
2528 the pitch is changed before it reaches the engraver (in fact, it
2529 happens just after the parsing stage with the creation of a
2530 TransposedMusic music object):
2536 (ly:make-duration 2 0 1 1)
2538 (ly:make-pitch 0 1 0)
2541 You can see an example of a music property relevant to transposition:
2545 \transpose c d @{ c'2 \withMusicProperty #'untransposable ##t c' @}
2548 -> the second c' remains untransposed.
2550 Take a look at @file{lily/music.cc} to see where the transposition takes place.
2553 @subheading How do I tell about the execution environment?
2555 I get lost figuring out what environment the code I'm looking at is in when it
2556 executes. I found both the C++ and Scheme autochange code. Then I was trying
2557 to figure out where the code got called from. I finally figured out that the
2558 Scheme procedure was called before the C++ iterator code, but it took me a
2559 while to figure that out, and I still didn't know who did the calling in the
2560 first place. I only know a little bit about Flex and Bison, so reading those
2561 files helped only a little bit.
2563 @emph{Han-Wen:} GDB can be of help here. Set a breakpoint in C++, and run. When you
2564 hit the breakpoint, do a backtrace. You can inspect Scheme objects
2565 along the way by doing
2568 p ly_display_scm(obj)
2571 this will display OBJ through GUILE.
2573 @node Music functions and GUILE debugging
2574 @subsection Music functions and GUILE debugging
2576 Ian Hulin was trying to do some debugging in music functions, and
2577 came up with the following question (edited and adapted to current
2581 I'm working on the Guile Debugger Stuff, and would like to try
2582 debugging a music function definition such as:
2586 #(define-music-function () ()
2587 #@{ \tag instrumental-part @{\mark \default@} #@} )
2590 It appears @code{conditionalMark} does not get set up as an
2591 equivalent of a Scheme
2594 (define conditionalMark = define-music-function () () ...
2598 although something gets defined because Scheme apparently recognizes
2601 #(set-break! conditionalMark)
2605 later on in the file without signalling any Guile errors.
2607 However the breakpoint trap is never encountered as
2608 @code{define-music-function} passed things on to
2609 @code{ly:make-music-function}, which is really C++ code
2610 @code{ly_make_music_function}, so Guile never finds out about the
2614 The answer in the mailing list archive at that time was less than
2615 helpful. The question already misidentifies the purpose of
2616 @code{ly:make-music-function} which is only called once at the
2617 time of @emph{defining} @code{conditionalMark} but is not involved
2618 in its later @emph{execution}.
2620 Here is the real deal:
2622 A music function is not the same as a GUILE function. It boxes
2623 both a proper Scheme function (with argument list and body from
2624 the @code{define-music-function} definition) along with a call
2625 signature representing the @emph{types} of both function and
2628 Those components can be reextracted using
2629 @code{ly:music-function-extract} and
2630 @code{ly:music-function-signature}, respectively.
2632 When LilyPond's parser encounters a music function call in its
2633 input, it reads, interprets, and verifies the arguments
2634 individually according to the call signature and @emph{then} calls
2635 the proper Scheme function.
2637 While it is actually possible these days to call a music function
2638 @emph{as if} it were a Scheme function itself, this pseudo-call
2639 uses its own wrapping code matching the argument list @emph{as a
2640 whole} to the call signature, substituting omitted optional
2641 arguments with defaults and verifying the result type.
2643 So putting a breakpoint on the music function itself will still
2644 not help with debugging uses of the function using LilyPond
2647 However, either calling mechanism ultimately calls the proper
2648 Scheme function stored as part of the music function, and that is
2649 where the breakpoint belongs:
2652 #(set-break! (ly:music-function-extract conditionalMark))
2655 will work for either calling mechanism.
2657 @node Articulations on EventChord
2658 @subsection Articulations on EventChord
2660 From David Kastrup's email
2661 @uref{http://lists.gnu.org/archive/html/lilypond-devel/2012-02/msg00189.html}:
2663 LilyPond's typesetting does not act on music expressions and music
2664 events. It acts exclusively on stream events. It is the act of
2665 iterators to convert a music expression into a sequence of stream events
2666 played in time order.
2668 The EventChord iterator is pretty simple: it just takes its "elements"
2669 field when its time comes up, turns every member into a StreamEvent and
2670 plays that through the typesetting process. The parser currently
2671 appends all postevents belonging to a chord at the end of "elements",
2672 and thus they get played at the same point of time as the elements of
2673 the chord. Due to this design, you can add per-chord articulations or
2674 postevents or even assemble chords with a common stem by using parallel
2675 music providing additional notes/events: the typesetter does not see a
2676 chord structure or postevents belonging to a chord, it just sees a
2677 number of events occuring at the same point of time in a Voice context.
2679 So all one needs to do is let the EventChord iterator play articulations
2680 after elements, and then adding to articulations in EventChord is
2681 equivalent to adding them to elements (except in cases where the order