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 * Debugging LilyPond::
12 * Tracing object relationships::
13 * Adding or modifying features::
18 * LilyPond miscellany::
21 @node Overview of LilyPond architecture
22 @section Overview of LilyPond architecture
24 LilyPond processes the input file into graphical and musical output in a
25 number of stages. This process, along with the types of routines that
26 accomplish the various stages of the process, is described in this section. A
27 more complete description of the LilyPond architecture and internal program
28 execution is found in Erik Sandberg's
29 @uref{http://lilypond.org/web/images/thesis-erik-sandberg.pdf, master's
32 The first stage of LilyPond processing is @emph{parsing}. In the parsing
33 process, music expressions in LilyPond input format are converted to music
34 expressions in Scheme format. In Scheme format, a music expression is a list
35 in tree form, with nodes that indicate the relationships between various music
36 events. The LilyPond parser is written in Bison.
38 The second stage of LilyPond processing is @emph{iterating}. Iterating
39 assigns each music event to a context, which is the environment in which the
40 music will be finally engraved. The context is responsible for all further
41 processing of the music. It is during the iteration stage that contexts are
42 created as necessary to ensure that every note has a Voice type context (e.g.
43 Voice, TabVoice, DrumVoice, CueVoice, MensuralVoice, VaticanaVoice,
44 GregorianTranscriptionVoice), that the Voice type contexts exist in
45 appropriate Staff type contexts, and that parallel Staff type contexts exist
46 in StaffGroup type contexts. In addition, during the iteration stage each
47 music event is assigned a moment, or a time in the music when the event
50 Each type of music event has an associated iterator. Iterators are defined in
51 *-iterator.cc. During iteration, an
52 event's iterator is called to deliver that music event to the appropriate
55 The final stage of LilyPond processing is @emph{translation}. During
56 translation, music events are prepared for graphical or midi output. The
57 translation step is accomplished by the polymorphic base class Translator
58 through its two derived classes: Engraver (for graphical output) and
59 Performer (for midi output).
61 Translators are defined in C++ files named *-engraver.cc and *-performer.cc.
62 Much of the work of translating is handled by Scheme functions,
63 which is one of the keys to LilyPond's exceptional flexibility.
65 @sourceimage{architecture-diagram,,,png}
68 @node LilyPond programming languages
69 @section LilyPond programming languages
71 Programming in LilyPond is done in a variety of programming languages. Each
72 language is used for a specific purpose or purposes. This section describes
73 the languages used and provides links to reference manuals and tutorials for
74 the relevant language.
78 The core functionality of LilyPond is implemented in C++.
80 C++ is so ubiquitous that it is difficult to identify either a reference
81 manual or a tutorial. Programmers unfamiliar with C++ will need to spend some
82 time to learn the language before attempting to modify the C++ code.
84 The C++ code calls Scheme/GUILE through the GUILE interface, which is
86 @uref{http://www.gnu.org/software/guile/manual/html_node/index.html, GUILE
91 The LilyPond lexer is implemented in Flex, an implementation of the Unix lex
92 lexical analyser generator. Resources for Flex can be found
93 @uref{http://flex.sourceforge.net/, here}.
97 The LilyPond parser is implemented in Bison, a GNU parser generator. The
98 Bison homepage is found at @uref{http://www.gnu.org/software/bison/,
99 gnu.org}. The manual (which includes both a reference and tutorial) is
100 @uref{http://www.gnu.org/software/bison/manual/index.html, available} in a
105 GNU Make is used to control the compiling process and to build the
106 documentation and the website. GNU Make documentation is available at
107 @uref{http://www.gnu.org/software/make/manual/, the GNU website}.
109 @subsection GUILE or Scheme
111 GUILE is the dialect of Scheme that is used as LilyPond's extension language.
112 Many extensions to LilyPond are written entirely in GUILE. The
113 @uref{http://www.gnu.org/software/guile/manual/html_node/index.html,
114 GUILE Reference Manual} is available online.
116 @uref{http://mitpress.mit.edu/sicp/full-text/book/book.html, Structure and
117 Interpretation of Computer Programs}, a popular textbook used to teach
118 programming in Scheme is available in its entirety online.
120 An introduction to Guile/Scheme as used in LilyPond can be found in the
121 @rextend{Scheme tutorial}.
125 MetaFont is used to create the music fonts used by LilyPond. A MetaFont
126 tutorial is available at @uref{http://metafont.tutorial.free.fr/, the
127 METAFONT tutorial page}.
129 @subsection PostScript
131 PostScript is used to generate graphical output. A brief PostScript tutorial
132 is @uref{http://local.wasp.uwa.edu.au/~pbourke/dataformats/postscript/,
133 available online}. The
134 @uref{http://www.adobe.com/devnet/postscript/pdfs/PLRM.pdf, PostScript Lanugage
135 Reference} is available online in PDF format.
139 Python is used for XML2ly and is used for buillding the documentation and the
142 Python documentation is available at @uref{http://www.python.org/doc/,
145 @node Programming without compiling
146 @section Programming without compiling
148 Much of the development work in LilyPond takes place by changing *.ly or
149 *.scm files. These changes can be made without compiling LilyPond. Such
150 changes are described in this section.
153 @subsection Modifying distribution files
155 Much of LilyPond is written in Scheme or LilyPond input files. These
156 files are interpreted when the program is run, rather than being compiled
157 when the program is built, and are present in all LilyPond distributions.
158 You will find .ly files in the ly/ directory and the Scheme files in the
159 scm/ directory. Both Scheme files and .ly files can be modified and
160 saved with any text editor. It's probably wise to make a backup copy of
161 your files before you modify them, although you can reinstall if the
162 files become corrupted.
164 Once you've modified the files, you can test the changes just by running
165 LilyPond on some input file. It's a good idea to create a file that
166 demonstrates the feature you're trying to add. This file will eventually
167 become a regression test and will be part of the LilyPond distribution.
169 @subsection Desired file formatting
171 Files that are part of the LilyPond distribution have Unix-style line
172 endings (LF), rather than DOS (CR+LF) or MacOS 9 and earlier (CR). Make
173 sure you use the necessary tools to ensure that Unix-style line endings are
174 preserved in the patches you create.
176 Tab characters should not be included in files for distribution. All
177 indentation should be done with spaces. Most editors have settings to
178 allow the setting of tab stops and ensuring that no tab characters are
179 included in the file.
181 Scheme files and LilyPond files should be written according to standard
182 style guidelines. Scheme file guidelines can be found at
183 @uref{http://community.schemewiki.org/?scheme-style}. Following these
184 guidelines will make your code easier to read. Both you and others that
185 work on your code will be glad you followed these guidelines.
187 For LilyPond files, you should follow the guidelines for LilyPond snippets
188 in the documentation. You can find these guidelines at
189 @ref{Texinfo introduction and usage policy}.
191 @node Finding functions
192 @section Finding functions
194 When making changes or fixing bugs in LilyPond, one of the initial
195 challenges is finding out where in the code tree the functions to
196 be modified live. With nearly 3000 files in the source tree,
197 trial-and-error searching is generally ineffective. This section
198 describes a process for finding interesting code.
200 @subsection Using the ROADMAP
202 The file ROADMAP is located in the main directory of the lilypond source.
203 ROADMAP lists all of the directories in the LilPond source tree, along
204 with a brief description of the kind of files found in each directory.
205 This can be a very helpful tool for deciding which directories to search
206 when looking for a function.
209 @subsection Using grep to search
211 Having identified a likely subdirectory to search, the grep utility can
212 be used to search for a function name. The format of the grep command is
215 grep -i functionName subdirectory/*
218 This command will search all the contents of the directory subdirectory/
219 and display every line in any of the files that contains
220 functionName. The @code{-i} option makes @command{grep} ignore
221 case -- this can be very useful if you are not yet familiar with
222 our capitalization conventions.
224 The most likely directories to grep for function names are scm/ for
225 scheme files, ly/ for lilypond input (*.ly) files, and lily/ for C++
229 @subsection Using git grep to search
231 If you have used git to obtain the source, you have access to a
232 powerful tool to search for functions. The command:
235 git grep functionName
238 will search through all of the files that are present in the git
239 repository looking for functionName. It also presents the results
240 of the search using @code{less}, so the results are displayed one page
243 @subsection Searching on the git repository at Savannah
245 You can also use the equivalent of git grep on the Savannah server.
250 Go to http://git.sv.gnu.org/gitweb/?p=lilypond.git
253 In the pulldown box that says commit, select grep.
256 Type functionName in the search box, and hit enter/return
260 This will initiate a search of the remote git repository.
266 This section describes style guidelines for LilyPond
273 * Naming conventions::
282 @subsection Languages
284 C++ and Python are preferred. Python code should use PEP 8.
288 @subsection Filenames
290 Definitions of classes that are only accessed via pointers (*) or
291 references (&) shall not be included as include files.
297 ".cc" Implementation files
298 ".icc" Inline definition files
299 ".tcc" non inline Template defs
303 (setq auto-mode-alist
304 (append '(("\\.make$" . makefile-mode)
305 ("\\.cc$" . c++-mode)
306 ("\\.icc$" . c++-mode)
307 ("\\.tcc$" . c++-mode)
308 ("\\.hh$" . c++-mode)
309 ("\\.pod$" . text-mode)
314 The class Class_name is coded in @q{class-name.*}
318 @subsection Indentation
320 Standard GNU coding style is used. In emacs:
323 (add-hook 'c++-mode-hook
324 '(lambda() (c-set-style "gnu")
328 If you like using font-lock, you can also add this to your
332 (setq font-lock-maximum-decoration t)
333 (setq c++-font-lock-keywords-3
335 c++-font-lock-keywords-3
336 '(("\\b\\(a-zA-Z_?+_\\)\\b" 1 font-lock-variable-name-face) ("\\b\\(A-Z?+a-z_?+\\)\\b" 1 font-lock-type-face))
340 Some source files may not currently have proper indenting. If this
341 is the case, it is desirable to fix the improper indenting when the
342 file is modified, with the hope of continually improving the code.
345 @subheading Indenting files with fixcc.py
347 LilyPond provides a python script that will correct the indentation
351 scripts/auxiliar/fixcc.py lily/my-test-file.cc
354 Be sure you replace @code{my-test-file.cc} with the name of the file
357 If you are editing a file that contains an ADD_TRANSLATOR or ADD_INTERFACE
358 macro, the fixcc.py script will move the final parenthesis up one line
359 from where it should be. Please check the end of the file before
360 you run fixcc.py, and then put the final parenthesis and semicolon
361 back on a line by themselves.
364 @subheading Indenting files with emacs in script mode
366 @c email to wl@gnu.org when I get here.
368 @warning{this is pending some confirmation on -devel. July 2009 -gp}
370 Command-line script to format stuff with emacs:
374 emacs $1 -batch --eval '(indent-region (point-min) (point-max) nil)' -f save-buffer
377 (that's all on one line)
379 Save it as a shell script, then run on the file(s) you modified.
382 @subheading Indenting with vim
384 Although emacs indentation is the LilyPond standard, acceptable
385 indentation can usually be accomplished with vim. Some hints for
397 filetype plugin indent on
399 set ignorecase smartcase
402 set statusline=%F%m%r%h%w\ %{&ff}\ %Y\ [ASCII=\%03.3b]\ [HEX=\%02.2B]\ %04l,%04v\ %p%%\ [LEN=%L]
405 " Remove trailing whitespace on write
406 autocmd BufWritePre * :%s/\s\+$//e
409 With this .vimrc, files can be reindented automatically by highlihting
410 the lines to be indented in visual mode (use V to enter visual mode)
413 A scheme.vim file will help improve the indentation. This one
414 was suggested by Patrick McCarty. It should be saved in
415 ~/.vim/after/syntax/scheme.vim.
418 " Additional Guile-specific 'forms'
419 syn keyword schemeSyntax define-public define* define-safe-public
420 syn keyword schemeSyntax use-modules define-module
421 syn keyword schemeSyntax defmacro-public define-macro
422 syn keyword schemeSyntax define-markup-command
423 syn keyword schemeSyntax define-markup-list-command
424 syn keyword schemeSyntax let-keywords* lambda* define*-public
425 syn keyword schemeSyntax defmacro* defmacro*-public
427 " All of the above should influence indenting too
428 set lw+=define-public,define*,define-safe-public,use-modules,define-module
429 set lw+=defmacro-public,define-macro
430 set lw+=define-markup-command,define-markup-list-command
431 set lw+=let-keywords*,lambda*,define*-public,defmacro*,defmacro*-public
433 " These forms should not influence indenting
437 " Try to highlight all ly: procedures
438 syn match schemeFunc "ly:[^) ]\+"
442 @node Naming conventions
443 @subsection Naming Conventions
445 Naming conventions have been established for LilyPond
448 @subheading Classes and Types
450 Classes begin with an uppercase letter, and words
451 in class names are separated with @code{_}:
459 Member variable names end with an underscore:
467 Macro names should be written in uppercase completely,
468 with words separated by @code{_}:
474 @subheading Variables
476 Variable names should be complete words, rather than abbreviations.
477 For example, it is preferred to use @code{thickness} rather than
478 @code{th} or @code{t}.
480 Multi-word variable names in C++ should have the words separated
481 by the underscore character (@q{_}):
484 cxx_multiword_variable
487 Multi-word variable names in Scheme should have the words separated
491 scheme-multiword-variable
495 @subsection Broken code
497 Do not write broken code. This includes hardwired dependencies,
498 hardwired constants, slow algorithms and obvious limitations. If
499 you can not avoid it, mark the place clearly, and add a comment
500 explaining shortcomings of the code.
502 Ideally, the comment marking the shortcoming would include
503 TODO, so that it is marked for future fixing.
505 We reject broken-in-advance on principle.
509 @subsection Code comments
511 Comments may not be needed if descriptive variable names are used
512 in the code and the logic is straightforward. However, if the
513 logic is difficult to follow, and particularly if non-obvious
514 code has been included to resolve a bug, a comment describing
515 the logic and/or the need for the non-obvious code should be included.
517 There are instances where the current code could be commented better.
518 If significant time is required to understand the code as part of
519 preparing a patch, it would be wise to add comments reflecting your
520 understanding to make future work easier.
523 @node Handling errors
524 @subsection Handling errors
526 As a general rule, you should always try to continue computations,
527 even if there is some kind of error. When the program stops, it
528 is often very hard for a user to pinpoint what part of the input
529 causes an error. Finding the culprit is much easier if there is
530 some viewable output.
532 So functions and methods do not return errorcodes, they never
533 crash, but report a programming_error and try to carry on.
535 Error and warning messages need to be localized.
539 @subsection Localization
541 This document provides some guidelines to help programmers write
543 messages. To help translations, user messages must follow
544 uniform conventions. Follow these rules when coding for LilyPond.
545 Hopefully, this can be replaced by general GNU guidelines in the
546 future. Even better would be to have an English (en_BR, en_AM)
547 guide helping programmers writing consistent messages for all GNU
550 Non-preferred messages are marked with `+'. By convention,
551 ungrammatical examples are marked with `*'. However, such ungrammatical
552 examples may still be preferred.
557 Every message to the user should be localized (and thus be marked
558 for localization). This includes warning and error messages.
561 Do not localize/gettextify:
565 `programming_error ()'s
568 `programming_warning ()'s
574 output strings (PostScript, TeX, etc.)
579 Messages to be localized must be encapsulated in `_ (STRING)' or
580 `_f (FORMAT, ...)'. E.g.:
583 warning (_ ("need music in a score"));
584 error (_f ("cannot open file: `%s'", file_name));
587 In some rare cases you may need to call `gettext ()' by hand. This
588 happens when you pre-define (a list of) string constants for later
589 use. In that case, you'll probably also need to mark these string
590 constants for translation, using `_i (STRING)'. The `_i' macro is
591 a no-op, it only serves as a marker for `xgettext'.
594 char const* messages[] = @{
595 _i ("enable debugging output"),
596 _i ("ignore lilypond version"),
603 puts (gettext (messages i));
607 See also `flower/getopt-long.cc' and `lily/main.cc'.
610 Do not use leading or trailing whitespace in messages. If you need
611 whitespace to be printed, prepend or append it to the translated
615 message ("Calculating line breaks..." + " ");
619 Error or warning messages displayed with a file name and line
620 number never start with a capital, eg,
623 foo.ly: 12: not a duration: 3
626 Messages containing a final verb, or a gerund (`-ing'-form) always
627 start with a capital. Other (simpler) messages start with a
633 Not declaring: `foo'.
637 Avoid abbreviations or short forms, use `cannot' and `do not'
638 rather than `can't' or `don't'
639 To avoid having a number of different messages for the same
640 situation, well will use quoting like this `"message: `%s'"' for all
641 strings. Numbers are not quoted:
644 _f ("cannot open file: `%s'", name_str)
645 _f ("cannot find character number: %d", i)
649 Think about translation issues. In a lot of cases, it is better to
650 translate a whole message. The english grammar must not be imposed
651 on the translator. So, instead of
654 stem at + moment.str () + does not fit in beam
660 _f ("stem at %s does not fit in beam", moment.str ())
664 Split up multi-sentence messages, whenever possible. Instead of
667 warning (_f ("out of tune! Can't find: `%s'", "Key_engraver"));
668 warning (_f ("cannot find font `%s', loading default", font_name));
674 warning (_ ("out of tune:"));
675 warning (_f ("cannot find: `%s', "Key_engraver"));
676 warning (_f ("cannot find font: `%s', font_name));
677 warning (_f ("Loading default font"));
681 If you must have multiple-sentence messages, use full punctuation.
682 Use two spaces after end of sentence punctuation. No punctuation
683 (esp. period) is used at the end of simple messages.
686 _f ("Non-matching braces in text `%s', adding braces", text)
687 _ ("Debug output disabled. Compiled with NPRINT.")
688 _f ("Huh? Not a Request: `%s'. Ignoring.", request)
692 Do not modularize too much; words frequently cannot be translated
693 without context. It is probably safe to treat most occurences of
694 words like stem, beam, crescendo as separately translatable words.
697 When translating, it is preferable to put interesting information
698 at the end of the message, rather than embedded in the middle.
699 This especially applies to frequently used messages, even if this
700 would mean sacrificing a bit of eloquency. This holds for original
701 messages too, of course.
704 en: cannot open: `foo.ly'
705 + nl: kan `foo.ly' niet openen (1)
706 kan niet openen: `foo.ly'* (2)
707 niet te openen: `foo.ly'* (3)
711 The first nl message, although grammatically and stylistically
712 correct, is not friendly for parsing by humans (even if they speak
713 dutch). I guess we would prefer something like (2) or (3).
716 Do not run make po/po-update with GNU gettext < 0.10.35
722 @node Debugging LilyPond
723 @section Debugging LilyPond
725 The most commonly used tool for debugging LilyPond is the GNU
726 debugger gdb. The gdb tool is used for investigating and debugging
727 core Lilypond code written in C++. Another tool is available for
728 debugging Scheme code using the Guile debugger. This section
729 describes how to use both gdb and the Guile Debugger.
732 * Debugging overview::
733 * Debugging C++ code::
734 * Debugging Scheme code::
737 @node Debugging overview
738 @subsection Debugging overview
740 Using a debugger simplifies troubleshooting in at least two ways.
742 First, breakpoints can be set to pause execution at any desired point.
743 Then, when execution has paused, debugger commands can be issued to
744 explore the values of various variables or to execute functions.
746 Second, the debugger can display a stack trace, which shows the
747 sequence in which functions have been called and the arguments
748 passed to the called functions.
750 @node Debugging C++ code
751 @subsection Debugging C++ code
753 The GNU debugger, gdb, is the principal tool for debugging C++ code.
755 @subheading Compiling LilyPond for use with gdb
757 In order to use gdb with LilyPond, it is necessary to compile
758 LilyPond with debugging information. This is accomplished by running
759 the following commands in the main LilyPond source directory.
762 ./configure --disable-optimising
766 This will create a version of LilyPond containing debugging
767 information that will allow the debugger to tie the source code
768 to the compiled code.
770 You should not do @var{make install} if you want to use a debugger
771 with LilyPond. The @var{make install} command will strip debugging
772 information from the LilyPond binary.
774 @subheading Typical gdb usage
776 Once you have compiled the Lilypond image with the necessary
777 debugging information it will have been written to a location in a
778 subfolder of your current working directory:
784 This is important as you will need to let gdb know where to find the
785 image containing the symbol tables. You can invoke gdb from the
786 command line usinga the following:
792 This loads the LilyPond symbol tables into gdb. Then, to run
793 LilyPond on @code{test.ly} under the debugger, enter the following:
802 As an alternative to running gdb at the command line you may try
803 a graphical interface to gdb such as ddd:
809 You can also use sets of standard gdb commands stored in a .gdbinit
810 file (see next section).
812 @subheading Typical .gdbinit files
814 The behavior of gdb can be readily customized through the use of a
815 @var{.gdbinit} file. A @var{.gdbinit} file is a file named
816 @var{.gdbinit} (notice the @qq{.} at the beginning of the file name)
817 that is placed in a user's home directory.
819 The @var{.gdbinit} file below is from Han-Wen. It sets breakpoints
820 for all errors and defines functions for displaying scheme objects
821 (ps), grobs (pgrob), and parsed music expressions (pmusic).
824 file lily/out/lilypond
826 b Grob::programming_error
829 print ly_display_scm($arg0)
832 print ly_display_scm($arg0->self_scm_)
833 print ly_display_scm($arg0->mutable_property_alist_)
834 print ly_display_scm($arg0->immutable_property_alist_)
835 print ly_display_scm($arg0->object_alist_)
838 print ly_display_scm($arg0->self_scm_)
839 print ly_display_scm($arg0->mutable_property_alist_)
840 print ly_display_scm($arg0->immutable_property_alist_)
844 @node Debugging Scheme code
845 @subsection Debugging Scheme code
847 Scheme code can be developed using the Guile command line
848 interpreter @code{top-repl}. You can either investigate
849 interactively using just Guile or you can use the debugging
850 tools available within Guile.
852 @subheading Using Guile interactively with LilyPond
854 In order to experiment with Scheme programming in the LilyPond
855 environment, it is necessary to have a Guile interpreter that
856 has all the LilyPond modules loaded. This requires the following
859 First, define a Scheme symbol for the active module in the .ly file:
862 #(module-define! (resolve-module '(guile-user))
863 'lilypond-module (current-module))
866 Now place a Scheme function in the .ly file that gives an
867 interactive Guile prompt:
873 When the .ly file is compiled, this causes the compilation to be
874 interrupted and an interactive guile prompt to appear. Once the
875 guile prompt appears, the LilyPond active module must be set as the
876 current guile module:
879 guile> (set-current-module lilypond-module)
882 You can demonstrate these commands are operating properly by typing the name
883 of a LilyPond public scheme function to check it has been defined:
886 guile> fret-diagram-verbose-markup
887 #<procedure fret-diagram-verbose-markup (layout props marking-list)>
890 If the LilyPond module has not been correctly loaded, an error
891 message will be generated:
894 guile> fret-diagram-verbose-markup
895 ERROR: Unbound variable: fret-diagram-verbose-markup
896 ABORT: (unbound-variable)
899 Once the module is properly loaded, any valid LilyPond Scheme
900 expression can be entered at the interactive prompt.
902 After the investigation is complete, the interactive guile
903 interpreter can be exited:
909 The compilation of the .ly file will then continue.
911 @subheading Using the Guile debugger
913 To set breakpoints and/or enable tracing in Scheme functions, put
916 \include "guile-debugger.ly"
919 in your input file after any scheme procedures you have defined in
920 that file. This will invoke the Guile command-line after having set
921 up the environment for the debug command-line. When your input file
922 is processed, a guile prompt will be displayed. You may now enter
923 commands to set up breakpoints and enable tracing by the Guile debugger.
925 @subheading Using breakpoints
927 At the guile prompt, you can set breakpoints with
928 the @code{set-break!} procedure:
931 guile> (set-break! my-scheme-procedure)
934 Once you have set the desired breakpoints, you exit the guile repl frame
941 Then, when one of the scheme routines for which you have set
942 breakpoints is entered, guile will interrupt execution in a debug
943 frame. At this point you will have access to Guile debugging
944 commands. For a listing of these commands, type:
950 Alternatively you may code the breakpoints in your Lilypond source
951 file using a command such as:
954 #(set-break! my-scheme-procedure)
957 immediately after the @code{\include} statement. In this case the
958 breakpoint will be set straight after you enter the @code{(quit)}
959 command at the guile prompt.
961 Embedding breakpoint commands like this is particularly useful if
962 you want to look at how the Scheme procedures in the @var{.scm}
963 files supplied with LilyPond work. To do this, edit the file in
964 the relevant directory to add this line near the top:
967 (use-modules (scm guile-debugger))
970 Now you can set a breakpoint after the procedure you are interested
971 in has been declared. For example, if you are working on routines
972 called by @var{print-book-with} in @var{lily-library.scm}:
975 (define (print-book-with parser book process-procedure)
976 (let* ((paper (ly:parser-lookup parser '$defaultpaper))
977 (layout (ly:parser-lookup parser '$defaultlayout))
978 (outfile-name (get-outfile-name parser)))
979 (process-procedure book paper layout outfile-name)))
981 (define-public (print-book-with-defaults parser book)
982 (print-book-with parser book ly:book-process))
984 (define-public (print-book-with-defaults-as-systems parser book)
985 (print-book-with parser book ly:book-process-to-systems))
989 At this point in the code you could add this to set a breakpoint at
993 (set-break! print-book-with)
996 @subheading Tracing procedure calls and evaluator steps
998 Two forms of trace are available:
1001 (set-trace-call! my-scheme-procedure)
1007 (set-trace-subtree! my-scheme-procedure)
1010 @code{set-trace-call!} causes Scheme to log a line to the standard
1011 output to show when the procedure is called and when it exits.
1013 @code{set-trace-subtree!} traces every step the Scheme evaluator
1014 performs in evaluating the procedure.
1016 @node Tracing object relationships
1017 @section Tracing object relationships
1019 Understanding the LilyPond source often boils down to figuring out what
1020 is happening to the Grobs. Where (and why) are they being created,
1021 modified and destroyed? Tracing Lily through a debugger in order to
1022 identify these relationships can be time-consuming and tedious.
1024 In order to simplify this process, a facility has been added to
1025 display the grobs that are created and the properties that are set
1026 and modified. Although it can be complex to get set up, once set up
1027 it easily provides detailed information about the life of grobs
1028 in the form of a network graph.
1030 Each of the steps necessary to use the graphviz utility
1035 @item Installing graphviz
1037 In order to create the graph of the object relationships, it is
1038 first necessary to install Graphviz. graphviz is available for a
1039 number of different platforms:
1042 @uref{http://www.graphviz.org/Download..php}
1045 @item Modifying config.make
1047 In order for the Graphviz tool to work, config.make must be modified.
1048 It is probably a good idea to first save a copy of config.make under
1049 a different name. Then, edit config.make by removing every occurence
1052 @item Rebuilding LilyPond
1054 The executable code of LilyPond must be rebuilt from scratch:
1057 make -C lily clean && make -C lily
1060 @item Create a graphviz-compatible .ly file
1062 In order to use the graphviz utility, the .ly file must include
1063 @file{ly/graphviz-init.ly}, and should then specify the
1064 grobs and symbols that should be tracked. An example of this
1065 is found in @file{input/regression/graphviz.ly}.
1067 @item Run lilypond with output sent to a log file
1069 The Graphviz data is sent to stderr by lilypond, so it is
1070 necessary to redirect stderr to a logfile:
1073 lilypond graphviz.ly 2> graphviz.log
1076 @item Edit the logfile
1078 The logfile has standard lilypond output, as well as the Graphviz
1079 output data. Delete everything from the beginning of the file
1080 up to but not including the first occurence of @code{digraph}.
1082 @item Process the logfile with @code{dot}
1084 The directed graph is created from the log file with the program
1088 dot -Tpdf graphviz.log > graphviz.pdf
1093 The pdf file can then be viewed with any pdf viewer.
1095 When compiled without @code{-DNDEBUG}, lilypond may run slower
1096 than normal. The original configuration can be restored by either
1097 renaming the saved copy of @code{config.make} or rerunning
1098 @code{configure}. Then rebuild lilypond with
1101 make -C lily clean && make -C lily
1105 @node Adding or modifying features
1106 @section Adding or modifying features
1108 When a new feature is to be added to LilyPond, it is necessary to
1109 ensure that the feature is properly integrated to maintain
1110 its long-term support. This section describes the steps necessary
1111 for feature addition and modification.
1116 * Write regression tests::
1117 * Write convert-ly rule::
1118 * Automaticaly update auxiliary information::
1119 * Manually update auxiliary information::
1120 * Edit changes.tely::
1121 * Verify successful build::
1122 * Verify regression tests::
1123 * Post patch for comments::
1125 * Closing the issues::
1128 @node Write the code
1129 @subsection Write the code
1131 You should probably create a new git branch for writing the code, as that
1132 will separate it from the master branch and allow you to continue
1133 to work on small projects related to master.
1135 Please be sure to follow the rules for programming style discussed
1136 earlier in this chapter.
1139 @node Write regression tests
1140 @subsection Write regression tests
1142 In order to demonstrate that the code works properly, you will
1143 need to write one or more regression tests. These tests are
1144 typically .ly files that are found in input/regression.
1146 Regression tests should be as brief as possible to demonstrate the
1147 functionality of the code.
1149 Regression tests should generally cover one issue per test. Several
1150 short, single-issue regression tests are preferred to a single, long,
1151 multiple-issue regression test.
1153 Use existing regression tests as templates to demonstrate the type of
1154 header information that should be included in a regression test.
1157 @node Write convert-ly rule
1158 @subsection Write convert-ly rule
1160 If the modification changes the input syntax, a convert-ly rule
1161 should be written to automatically update input files from older
1164 convert-ly rules are found in python/convertrules.py
1166 If possible, the convert-ly rule should allow automatic updating
1167 of the file. In some cases, this will not be possible, so the
1168 rule will simply point out to the user that the feature needs
1172 @node Automaticaly update auxiliary information
1173 @subsection Automatically update auxiliary information
1175 convert-ly should be used to update the documentation, the snippets,
1176 and the regression tests. This not only makes the necessary syntax
1177 changes, it also tests the convert-ly rules.
1179 The automatic updating is a three step process. First, be sure you
1180 are in the top-level source directory. Then, for the
1184 find Documentation/ -name '*.itely' | xargs convert-ly -e --from @qq{@var{X.Y.Z}}
1188 where @var{X.Y.Z} is the version number of the last released development
1191 Next, for the snippets, do:
1194 find Documentation/snippets/ -name '*.ly' | xargs convert-ly -e --from @qq{@var{X.Y.Z}}
1197 Finally, for the regression tests, do:
1200 find input/regression/ -name '*.ly' | xargs convert-ly -e --from @qq{@var{X.Y.Z}}
1205 @node Manually update auxiliary information
1206 @subsection Manually update auxiliary information
1208 Where the convert-ly rule is not able to automatically update the inline
1209 lilypond code in the documentation (i.e. if a NOT_SMART rule is used), the
1210 documentation must be manually updated. The inline snippets that require
1211 changing must be changed in the English version of the docs and all
1212 translated versions. If the inline code is not changed in the
1213 translated documentation, the old snippets will show up in the
1214 English version of the documentation.
1216 Where the convert-ly rule is not able to automatically update snippets
1217 in Documentation/snippets/, those snippets must be manually updated.
1218 Those snippets should be copied to Documentation/snippets/new. The
1219 comments at the top of the snippet describing its automatic generation
1220 should be removed. All translated texidoc strings should be removed.
1221 The comment @qq{% begin verbatim} should be removed. The syntax of
1222 the snippet should then be manually edited.
1224 Where snippets in Documentation/snippets are made obsolete, the snippet
1225 should be copied to Documentation/snippets/new. The comments and
1226 texidoc strings should be removed as described above. Then the body
1227 of the snippet should be changed to:
1231 This snippet is deprecated as of version X.Y.Z and
1232 will be removed from the documentation.
1237 where X.Y.Z is the version number for which the convert-ly rule was
1240 Update the snippet files by running:
1243 scripts/auxiliar/makelsr.py
1246 Where the convert-ly rule is not able to automatically update regression
1247 tests, the regression tests in input/regression should be manually
1250 Although it is not required, it is helpful if the developer
1251 can write relevant material for inclusion in the Notation
1252 Reference. If the developer does not feel qualified to write
1253 the documentation, a documentation editor will be able to
1254 write it from the regression tests. The text that is added to
1255 or removed from the documentation should be changed only in
1256 the English version.
1259 @node Edit changes.tely
1260 @subsection Edit changes.tely
1262 An entry should be added to Documentation/changes.tely to describe
1263 the feature changes to be implemented. This is especially important
1264 for changes that change input file syntax.
1266 Hints for changes.tely entries are given at the top of the file.
1268 New entries in changes.tely go at the top of the file.
1270 The changes.tely entry should be written to show how the new change
1271 improves LilyPond, if possible.
1274 @node Verify successful build
1275 @subsection Verify successful build
1277 When the changes have been made, successful completion must be
1285 When these commands complete without error, the patch is
1286 considered to function successfully.
1288 Developers on Windows who are unable to build LilyPond should
1289 get help from a Linux or OSX developer to do the make tests.
1292 @node Verify regression tests
1293 @subsection Verify regression tests
1295 In order to avoid breaking LilyPond, it is important to verify that
1296 the regression tests all succeed. This process is described in
1297 @ref{Regression tests}.
1300 @node Post patch for comments
1301 @subsection Post patch for comments
1303 For any change other than a minor change, a patch set should be
1304 posted on @uref{http://codereview.appspot.com/, Rietveld} for comment.
1305 This requires the use of an external package, git-cl, and an email
1308 git-cl is installed by:
1311 git clone git://neugierig.org/git-cl.git
1314 Then, add the git-cl directory to your PATH, or create a
1315 symbolic link to the git-cl and upload.py in one of your
1316 PATH directories (like usr/bin). git-cl is then
1317 configured by entering the command
1324 in the LilyPond git directory and answering the questions that
1325 are asked. If you do not understand the question answer with just
1328 The patch set is posted to Rietveld as follows. Ensure your changes
1329 are committed in a separate branch, which should differ from the
1330 reference branch to be used by just the changes to be uploaded.
1331 If the reference branch is to be origin/master, ensure this is
1332 up-to-date. If necessary, use git rebase to rebase the branch
1333 containing the changes to the head of origin/master. Finally,
1334 check out branch with the changes and enter the command:
1337 git cl upload <reference SHA1 ID>
1341 where <reference SHA1 ID> is the SHA1 ID of the commit to be used
1342 as a reference source for the patch. Generally, this will be the
1343 SHA1 ID of origin/master, and in that case the command
1346 git cl upload origin/master
1352 After prompting for your Google email address and password, the
1353 patch set will be posted to Rietveld.
1355 You should then announce the patch by sending
1356 an email to lilypond-devel, with a subject line
1357 starting with PATCH:, asking for comments on the patch.
1359 As revisions are made in response to comments, successive patch sets
1360 for the same issue can be uploaded by reissuing the git-cl command
1361 with the modified branch checked out.
1363 Sometimes in response to comments on revisions, the best way to
1364 work may require creation of a new branch in git. In order to
1365 associate the new branch with an existing Reitveld issue,
1366 the following command can be used:
1369 git cl issue issue-number
1373 where @code{issue-number} is the number of the existing Rietveld
1379 @subsection Push patch
1381 Once all the comments have been addressed, the patch can be pushed.
1383 If the author has push privileges, the author will push the patch.
1384 Otherwise, a developer with push privileges will push the patch.
1387 @node Closing the issues
1388 @subsection Closing the issues
1390 Once the patch has been pushed, all the relevant issues should be
1393 On Rietveld, the author should log in an close the issue either by
1394 using the @q{Edit Issue} link, or by clicking the circled x icon
1395 to the left of the issue name.
1397 If the changes were in response to a feature request on the Google
1398 issue tracker for LilyPond, the author should change the status to
1399 Fixed and a tag @q{fixed_x_y_z} should be added, where the patch was
1400 fixed in version x.y.z. If
1401 the author does not have privileges to change the status, an email
1402 should be sent to bug-lilypond requesting the BugMeister to change
1406 @node Iterator tutorial
1407 @section Iterator tutorial
1409 TODO -- this is a placeholder for a tutorial on iterators
1411 Iterators are routines written in C++ that process music expressions
1412 and sent the music events to the appropriate engravers and/or
1416 @node Engraver tutorial
1417 @section Engraver tutorial
1419 Engravers are C++ classes that catch music events and
1420 create the appropriate grobs for display on the page. Though the
1421 majority of engravers are responsible for the creation of a single grob,
1422 in some cases (e.g. @code{New_fingering_engraver}), several different grobs
1425 Engravers listen for events and acknowledge grobs. Events are passed to
1426 the engraver in time-step order during the iteration phase. Grobs are
1427 made available to the engraver when they are created by other engravers
1428 during the iteration phase.
1432 * Useful methods for information processing::
1433 * Translation process::
1434 * Preventing garbage collection for SCM member variables::
1435 * Listening to music events::
1436 * Acknowledging grobs::
1437 * Engraver declaration/documentation::
1440 @node Useful methods for information processing
1441 @subsection Useful methods for information processing
1443 An engraver inherits the following public methods from the Translator
1444 base class, which can be used to process listened events and acknowledged
1448 @item @code{virtual void initialize ()}
1449 @item @code{void start_translation_timestep ()}
1450 @item @code{void process_music ()}
1451 @item @code{void process_acknowledged ()}
1452 @item @code{void stop_translation_timestep ()}
1453 @item @code{virtual void finalize ()}
1456 These methods are listed in order of translation time, with
1457 @code{initialize ()} and @code{finalize ()} bookending the whole
1458 process. @code{initialize ()} can be used for one-time initialization
1459 of context properties before translation starts, whereas
1460 @code{finalize ()} is often used to tie up loose ends at the end of
1461 translation: for example, an unterminated spanner might be completed
1462 automatically or reported with a warning message.
1465 @node Translation process
1466 @subsection Translation process
1468 At each timestep in the music, translation proceeds by calling the
1469 following methods in turn:
1471 @code{start_translation_timestep ()} is called before any user
1472 information enters the translators, i.e., no property operations
1473 (\set, \override, etc.) or events have been processed yet.
1475 @code{process_music ()} and @code{process_acknowledged ()} are called
1476 after all events in the current time step have been heard, or all
1477 grobs in the current time step have been acknowledged. The latter
1478 tends to be used exclusively with engravers which only acknowledge
1479 grobs, whereas the former is the default method for main processing
1482 @code{stop_translation_timestep ()} is called after all user
1483 information has been processed prior to beginning the translation for
1487 @node Preventing garbage collection for SCM member variables
1488 @subsection Preventing garbage collection for SCM member variables
1490 In certain cases, an engraver might need to ensure private Scheme
1491 variables (with type SCM) do not get swept away by Guile's garbage
1492 collector: for example, a cache of the previous key signature which
1493 must persist between timesteps. The method
1494 @code{virtual derived_mark () const} can be used in such cases:
1497 Engraver_name::derived_mark ()
1499 scm_gc_mark (private_scm_member_)
1504 @node Listening to music events
1505 @subsection Listening to music events
1507 External interfaces to the engraver are implemented by protected
1508 macros including one or more of the following:
1511 @item @code{DECLARE_TRANSLATOR_LISTENER (event_name)}
1512 @item @code{IMPLEMENT_TRANSLATOR_LISTENER (Engraver_name, event_name)}
1516 where @var{event_name} is the type of event required to provide the
1517 input the engraver needs and @var{Engraver_name} is the name of the
1520 Following declaration of a listener, the method is implemented as follows:
1523 IMPLEMENT_TRANSLATOR_LISTENER (Engraver_name, event_name)
1525 Engraver_name::listen_event_name (Stream event *event)
1527 ...body of listener method...
1532 @node Acknowledging grobs
1533 @subsection Acknowledging grobs
1535 Some engravers also need information from grobs as they are created
1536 and as they terminate. The mechanism and methods to obtain this
1537 information are set up by the macros:
1540 @item @code{DECLARE_ACKNOWLEDGER (grob_interface)}
1541 @item @code{DECLARE_END_ACKNOWLEDGER (grob_interface)}
1544 where @var{grob_interface} is an interface supported by the
1545 grob(s) which should be acknowledged. For example, the following
1546 code would declare acknowledgers for a @code{NoteHead} grob (via the
1547 @code{note-head-interface}) and any grobs which support the
1548 @code{side-position-interface}:
1551 @code{DECLARE_ACKNOWLEDGER (note_head)}
1552 @code{DECLARE_ACKNOWLEDGER (side_position)}
1555 The @code{DECLARE_END_ACKNOWLEDGER ()} macro sets up a spanner-specific
1556 acknowledger which will be called whenever a spanner ends.
1558 Following declaration of an acknowledger, the method is coded as follows:
1562 Engraver_name::acknowledge_interface_name (Grob_info info)
1564 ...body of acknowledger method...
1569 @node Engraver declaration/documentation
1570 @subsection Engraver declaration/documentation
1572 An engraver must have a public macro
1575 @item @code{TRANSLATOR_DECLARATIONS (Engraver_name)}
1579 where @code{Engraver_name} is the name of the engraver. This
1580 defines the common variables and methods used by every engraver.
1582 At the end of the engraver file, one or both of the following
1583 macros are generally called to document the engraver in the
1584 Internals Reference:
1587 @item @code{ADD_ACKNOWLEDGER (Engraver_name, grob_interface)}
1588 @item @code{ADD_TRANSLATOR (Engraver_name, Engraver_doc,
1589 Engraver_creates, Engraver_reads, Engraver_writes)}
1593 where @code{Engraver_name} is the name of the engraver, @code{grob_interface}
1594 is the name of the interface that will be acknowledged,
1595 @code{Engraver_doc} is a docstring for the engraver,
1596 @code{Engraver_creates} is the set of grobs created by the engraver,
1597 @code{Engraver_reads} is the set of properties read by the engraver,
1598 and @code{Engraver_writes} is the set of properties written by
1601 The @code{ADD_ACKNOWLEDGER} and @code{ADD_TRANSLATOR} macros use a
1602 non-standard indentation system. Each interface, grob, read property,
1603 and write property is on its own line, and the closing parenthesis
1604 and semicolon for the macro all occupy a separate line beneath the final
1605 interface or write property. See existing engraver files for more
1609 @node Callback tutorial
1610 @section Callback tutorial
1612 TODO -- This is a placeholder for a tutorial on callback functions.
1614 @node LilyPond scoping
1615 @section LilyPond scoping
1617 The Lilypond language has a concept of scoping, ie you can do
1623 (display (+ foo 2)))
1626 @noindent with @code{\paper}, @code{\midi} and @code{\header} being
1627 nested scope inside the @file{.ly} file-level scope. @w{@code{foo = 1}}
1628 is translated in to a scheme variable definition.
1630 This implemented using modules, with each scope being an anonymous
1631 module that imports its enclosing scope's module.
1633 Lilypond's core, loaded from @file{.scm} files, is usually placed in the
1634 @code{lily} module, outside the @file{.ly} level. In the case of
1641 we want to reuse the built-in definitions, without changes effected in
1642 user-level @file{a.ly} leaking into the processing of @file{b.ly}.
1644 The user-accessible definition commands have to take care to avoid
1645 memory leaks that could occur when running multiple files. All
1646 information belonging to user-defined commands and markups is stored in
1647 a manner that allows it to be garbage-collected when the module is
1648 dispersed, either by being stored module-locally, or in weak hash
1651 @node LilyPond miscellany
1652 @section LilyPond miscellany
1654 This is a place to dump information that may be of use to developers
1655 but doesn't yet have a proper home. Ideally, the length of this section
1656 would become zero as items are moved to other homes.
1660 * Spacing algorithms::
1661 * Info from Han-Wen email::
1662 * Music functions and GUILE debugging::
1665 @node Spacing algorithms
1666 @subsection Spacing algorithms
1668 Here is information from an email exchange about spacing algorithms.
1670 On Thu, 2010-02-04 at 15:33 -0500, Boris Shingarov wrote:
1671 I am experimenting with some modifications to the line breaking code,
1672 and I am stuck trying to understand how some of it works. So far my
1673 understanding is that Simple_spacer operates on a vector of Grobs, and
1674 it is a well-known Constrained-QP problem (rods = constraints, springs
1675 = quadratic function to minimize). What I don't understand is, if the
1676 spacer operates at the level of Grobs, which are built at an earlier
1677 stage in the pipeline, how are the changes necessitated by differences
1678 in line breaking, taken into account? in other words, if I take the
1679 last measure of a line and place it on the next line, it is not just a
1680 matter of literally moving that graphic to where the start of the next
1681 line is, but I also need to draw a clef, key signature, and possibly
1682 other fundamental things -- but at that stage in the rendering
1683 pipeline, is it not too late??
1685 Joe Neeman answered:
1687 We create lots of extra grobs (eg. a BarNumber at every bar line) but
1688 most of them are not drawn. See the break-visibility property in
1692 @node Info from Han-Wen email
1693 @subsection Info from Han-Wen email
1695 In 2004, Douglas Linhardt decided to try starting a document that would
1696 explain LilyPond architecture and design principles. The material below
1697 is extracted from that email, which can be found at
1698 @uref{http://thread.gmane.org/gmane.comp.gnu.lilypond.devel/2992}.
1699 The headings reflect questions from Doug or comments from Han-Wen;
1700 the body text are Han-Wen's answers.
1702 @subheading Figuring out how things work.
1704 I must admit that when I want to know how a program works, I use grep
1705 and emacs and dive into the source code. The comments and the code
1706 itself are usually more revealing than technical documents.
1708 @subheading What's a grob, and how is one used?
1710 Graphical object - they are created from within engravers, either as
1711 Spanners (derived class) -slurs, beams- or Items (also a derived
1712 class) -notes, clefs, etc.
1714 There are two other derived classes System (derived from Spanner,
1715 contaning a "line of music") and Paper_column (derived from Item, it
1716 contains all items that happen at the same moment). They are separate
1717 classes because they play a special role in the linebreaking process.
1719 @subheading What's a smob, and how is one used?
1721 A C(++) object that is encapsulated so it can be used as a Scheme
1722 object. See GUILE info, "19.3 Defining New Types (Smobs)"
1724 @@subheading When is each C++ class constructed and used
1731 In the parser.yy see the macro calls MAKE_MUSIC_BY_NAME().
1736 Constructed during "interpreting" phase.
1741 Executive branch of Contexts, plugins that create grobs, usually one
1742 engraver per grob type. Created together with context.
1752 These are not C++ classes per se. The idea of a Grob interface hasn't
1753 crystallized well. ATM, an interface is a symbol, with a bunch of grob
1754 properties. They are not objects that are created or destroyed.
1759 Objects that walk through different music classes, and deliver events
1760 in a synchronized way, so that notes that play together are processed
1761 at the same moment and (as a result) end up on the same horizontal position.
1763 Created during interpreting phase.
1765 BTW, the entry point for interpreting is ly:run-translator
1766 (ly_run_translator on the C++ side)
1770 @subheading Can you get to Context properties from a Music object?
1772 You can create music object with a Scheme function that reads context
1773 properties (the \applycontext syntax). However, that function is
1774 executed during Interpreting, so you can not really get Context
1775 properties from Music objects, since music objects are not directly
1776 connected to Contexts. That connection is made by the Music_iterators
1778 @subheading Can you get to Music properties from a Context object?
1780 Yes, if you are given the music object within a Context
1781 object. Normally, the music objects enter Contexts in synchronized
1782 fashion, and the synchronization is done by Music_iterators.
1784 @subheading What is the relationship between C++ classes and Scheme objects?
1786 Smobs are C++ objects in Scheme. Scheme objects (lists, functions) are
1787 manipulated from C++ as well using the GUILE C function interface
1790 @subheading How do Scheme procedures get called from C++ functions?
1792 scm_call_*, where * is an integer from 0 to 4.
1793 Also scm_c_eval_string (), scm_eval ()
1795 @subheading How do C++ functions get called from Scheme procedures?
1797 Export a C++ function to Scheme with LY_DEFINE.
1799 @subheading What is the flow of control in the program?
1801 Good question. Things used to be clear-cut, but we have Scheme
1802 and SMOBs now, which means that interactions do not follow a very
1803 rigid format anymore. See below for an overview, though.
1805 @subheading Does the parser make Scheme procedure calls or C++ function calls?
1807 Both. And the Scheme calls can call C++ and vice versa. It's nested,
1808 with the SCM datatype as lubrication between the interactions
1810 (I think the word "lubrication" describes the process better than the
1811 traditional word "glue")
1813 @subheading How do the front-end and back-end get started?
1815 Front-end: a file is parsed, the rest follows from that. Specifically,
1817 Parsing leads to a Music + Music_output_def object (see parser.yy,
1818 definition of toplevel_expression )
1820 A Music + Music_output_def object leads to a Global_context object (see
1821 ly_run_translator ())
1823 During interpreting, Global_context + Music leads to a bunch of
1824 Contexts. (see Global_translator::run_iterator_on_me () )
1826 After interpreting, Global_context contains a Score_context (which
1827 contains staves, lyrics etc.) as a child. Score_context::get_output ()
1828 spews a Music_output object (either a Paper_score object for notation
1829 or Performance object for MIDI).
1831 The Music_output object is the entry point for the backend. (see
1832 ly_render_output () )
1834 The main steps of the backend itself are in
1839 paper-score.cc , Paper_score::process_
1842 system.cc , System::get_lines()
1845 The step, where things go from grobs to output, is in
1846 System::get_line(): each grob delivers a Stencil (a Device
1847 independent output description), which is interpreted by our
1848 outputting backends (scm/output-tex.scm and scm/output-ps.scm)
1849 to produce TeX and PS.
1853 Interactions between grobs and putting things into .tex and .ps files
1854 have gotten a little more complex lately. Jan has implemented
1855 page-breaking, so now the backend also involves Paper_book,
1856 Paper_lines and other things. This area is still heavily in flux, and
1857 perhaps not something you should want to look at.
1859 @subheading How do the front-end and back-end communicate?
1861 There is no communication from backend to front-end. From front-end to
1862 backend is simply the program flow: music + definitions gives
1863 contexts, contexts yield output, after processing, output is written
1866 @subheading Where is the functionality associated with KEYWORDs?
1868 See my-lily-lexer.cc (keywords, there aren't that many) and ly/*.ly
1869 (most of the other backslashed \words are identifiers)
1871 @subheading What Contexts/Properties/Music/etc. are available when they are processed?
1873 What do you mean exactly with this question?
1875 See ly/engraver-init.ly for contexts, see scm/define-*.scm for other
1878 @subheading How do you decide if something is a Music, Context, or Grob property?
1879 Why is part-combine-status a Music property when it seems (IMO)
1880 to be related to the Staff context?
1882 The Music_iterators and Context communicate through two channels
1884 Music_iterators can set and read context properties, idem for
1885 Engravers and Contexts
1887 Music_iterators can send "synthetic" music events (which aren't in
1888 the input) to a context. These are caught by Engravers. This is
1889 mostly a one way communication channel.
1891 part-combine-status is part of such a synthetic event, used by
1892 Part_combine_iterator to communicate with Part_combine_engraver.
1895 @subheading Deciding between context and music properties
1897 I'm adding a property to affect how \autochange works. It seems to
1898 me that it should be a context property, but the Scheme autochange
1899 procedure has a Music argument. Does this mean I should use
1902 \autochange is one of these extra strange beasts: it requires
1903 look-ahead to decide when to change staves. This is achieved by
1904 running the interpreting step twice (see scm/part-combiner.scm , at
1905 the bottom), and storing the result of the first step (where to switch
1906 staves) in a Music property. Since you want to influence that
1907 where-to-switch list, your must affect the code in
1908 make-autochange-music (scm/part-combiner.scm). That code is called
1909 directly from the parser and there are no official "parsing
1910 properties" yet, so there is no generic way to tune \autochange. We
1911 would have to invent something new for this, or add a separate
1915 \autochange #around-central-C ..music..
1919 where around-central-C is some function that is called from
1920 make-autochange-music.
1922 @subheading How do I tell about the execution environment?
1924 I get lost figuring out what environment the code I'm looking at is in when it
1925 executes. I found both the C++ and Scheme autochange code. Then I was trying
1926 to figure out where the code got called from. I finally figured out that the
1927 Scheme procedure was called before the C++ iterator code, but it took me a
1928 while to figure that out, and I still didn't know who did the calling in the
1929 first place. I only know a little bit about Flex and Bison, so reading those
1930 files helped only a little bit.
1932 @emph{Han-Wen:} GDB can be of help here. Set a breakpoint in C++, and run. When you
1933 hit the breakpoint, do a backtrace. You can inspect Scheme objects
1934 along the way by doing
1937 p ly_display_scm(obj)
1940 this will display OBJ through GUILE.
1942 @node Music functions and GUILE debugging
1943 @subsection Music functions and GUILE debugging
1945 Ian Hulin was trying to do some debugging in music functions, and
1946 came up with the following question
1949 I'm working on the Guile Debugger Stuff, and would like to try
1950 debugging a music function definition such as:
1953 conditionalMark = #(define-music-function (parser location) ()
1954 #@{ \tag #'instrumental-part @{\mark \default@} #@} )
1957 It appears conditionalMark does not get set up as an
1958 equivalent of a Scheme
1961 (define conditionalMark = define-music-function(parser location () ...
1965 although something gets defined because Scheme apparently recognizes
1968 #(set-break! conditionalMark)
1972 later on in the file without signalling any Guile errors.
1974 However the breakpoint trap is never encountered as
1975 define-music-function passed things on to ly:make-music-function,
1976 which is really C++ code ly_make_music_function, so Guile never
1977 finds out about the breakpoint.
1979 Han-Wen answered as follows:
1981 You can see the defintion by doing
1984 #(display conditionalMark)
1988 inside the .ly file.
1990 The breakpoint failing may have to do with the call sequence. See
1991 parser.yy, run_music_function(). The function is called directly from
1992 C++, without going through the GUILE evaluator, so I think that is why
1993 there is no debugger trap.