]> git.donarmstrong.com Git - lilypond.git/blobdiff - scripts/lilypond-book.py
(Paper size): \paper is in
[lilypond.git] / scripts / lilypond-book.py
index dbc6965780191bbb1d82e3e8629b209330a5073b..e3e62de2dabee0d492cf4094bcc15b025a778457 100644 (file)
 #!@PYTHON@
 
 '''
-TODO:
-      ly-options: intertext, quote ?
-      --linewidth?
-      eps in latex?
-      check latex parameters, twocolumn
-      multicolumn?
-      papersizes?
-      ly2dvi/notexidoc?
-      
 Example usage:
 
 test:
-     filter-lilypond-book --filter="tr '[a-z]' '[A-Z]'" BOOK
-         
+     lilypond-book --filter="tr '[a-z]' '[A-Z]'" BOOK
+
 convert-ly on book:
-     filter-lilypond-book --filter="convert-ly --no-version --from=1.6.11 -" BOOK
+     lilypond-book --filter="convert-ly --no-version --from=1.6.11 -" BOOK
 
 classic lilypond-book:
-     filter-lilypond-book --process="lilypond-bin" BOOK.tely
+     lilypond-book --process="lilypond" BOOK.tely
+
+TODO:
+
+    *  this script is too complex. Modularize.
+    
+    *  ly-options: intertext?
+    *  --linewidth?
+    *  eps in latex / eps by lilypond -b ps?
+    *  check latex parameters, twocolumn, multicolumn?
+    *  use --png --ps --pdf for making images?
+
+    *  Converting from lilypond-book source, substitute:
+       @mbinclude foo.itely -> @include foo.itely
+       \mbinput -> \input
 
-   must substitute:
-     @mbinclude foo.itely -> @include foo.itely
-     \mbinput -> \input
-     
 '''
 
-import string
 import __main__
+import glob
+import stat
+import string
 
-################################################################
 # Users of python modules should include this snippet
 # and customize variables below.
 
-# We'll suffer this path init stuff as long as we don't install our
-# python packages in <prefix>/lib/pythonx.y (and don't kludge around
-# it as we do with teTeX on Red Hat Linux: set some environment var
-# (PYTHONPATH) in profile)
+# We'll suffer this path initialization stuff as long as we don't install
+# our python packages in <prefix>/lib/pythonX.Y
 
-# If set, LILYPONDPREFIX must take prevalence
-# if datadir is not set, we're doing a build and LILYPONDPREFIX
+# If set, LILYPONDPREFIX must take prevalence.
+# if datadir is not set, we're doing a build and LILYPONDPREFIX.
 import getopt, os, sys
 datadir = '@local_lilypond_datadir@'
 if not os.path.isdir (datadir):
        datadir = '@lilypond_datadir@'
-if os.environ.has_key ('LILYPONDPREFIX') :
+if os.environ.has_key ('LILYPONDPREFIX'):
        datadir = os.environ['LILYPONDPREFIX']
        while datadir[-1] == os.sep:
                datadir= datadir[:-1]
 
 sys.path.insert (0, os.path.join (datadir, 'python'))
 
-# Customize these
+# Customize these.
 #if __name__ == '__main__':
 
 import lilylib as ly
+import fontextract
 global _;_=ly._
 global re;re = ly.re
 
-
-# lilylib globals
+# Lilylib globals.
 program_version = '@TOPLEVEL_VERSION@'
-program_name = 'lilypond-book'
+program_name = os.path.basename (sys.argv[0])
 verbose_p = 0
 pseudo_filter_p = 0
 original_dir = os.getcwd ()
+backend = 'ps'
 
+help_summary = _ (
+'''Process LilyPond snippets in hybrid HTML, LaTeX, or texinfo document.
+Example usage:
 
-help_summary = _ ("""Process LilyPond snippets in hybrid HTML, LaTeX or texinfo document.  Example usage:
-
-   filter-lilypond-book --filter="tr '[a-z]' '[A-Z]'" BOOK
-   filter-lilypond-book --filter="convert-ly --no-version --from=2.0.0 -" BOOK
-   filter-lilypond-book --process='lilypond-bin -I include' BOOK
+   lilypond-book --filter="tr '[a-z]' '[A-Z]'" BOOK
+   lilypond-book --filter="convert-ly --no-version --from=2.0.0 -" BOOK
+   lilypond-book --process='lilypond -I include' BOOK
+''')
 
-""")
-
-copyright = ('Jan Nieuwenhuizen <janneke@gnu.org>>',
+copyright = ('Jan Nieuwenhuizen <janneke@gnu.org>',
             'Han-Wen Nienhuys <hanwen@cs.uu.nl>')
 
 option_definitions = [
-       (_ ("EXT"), 'f', 'format', _ ("use output format EXT (texi [default], texi-html, latex, html)")),
-       (_ ("FILTER"), 'F', 'filter', _ ("pipe snippets through FILTER [convert-ly -n -]")),
-       ('', 'h', 'help', _ ("print this help")),
-       (_ ("DIR"), 'I', 'include', _ ("add DIR to include path")),
-       (_ ("COMMAND"), 'P', 'process', _ ("process ly_files using COMMAND FILE...")),
-       (_ ("DIR"), 'o', 'output', _ ("write output to DIR")),
-       ('', 'V', 'verbose', _ ("be verbose")),
-       ('', 'v', 'version', _ ("print version information")),
-       ('', 'w', 'warranty', _ ("show warranty and copyright")),
-       ]
+       (_ ("FMT"), 'f', 'format',
+         _ ('''use output format FMT (texi [default], texi-html,
+               latex, html)''')),
+       (_ ("FILTER"), 'F', 'filter',
+         _ ("pipe snippets through FILTER [convert-ly -n -]")),
+       ('', 'h', 'help',
+         _ ("print this help")),
+       (_ ("DIR"), 'I', 'include',
+         _ ("add DIR to include path")),
+       (_ ("DIR"), 'o', 'output',
+         _ ("write output to DIR")),
+       (_ ("COMMAND"), 'P', 'process',
+         _ ("process ly_files using COMMAND FILE...")),
+       (_(''), '', 'psfonts',
+        _ ('''extract all PostScript fonts into INPUT.psfonts for LaTeX
+        must use this with dvips -h INPUT.psfonts''')),
+       ('', 'V', 'verbose',
+         _ ("be verbose")),
+       ('', 'v', 'version',
+         _ ("print version information")),
+       ('', 'w', 'warranty',
+         _ ("show warranty and copyright")),
+]
 
 include_path = [ly.abspath (os.getcwd ())]
-lilypond_binary = os.path.join ('@bindir@', 'lilypond-bin')
+lilypond_binary = os.path.join ('@bindir@', 'lilypond')
 
-# only use installed binary  when we're installed too.
+# Only use installed binary when we are installed too.
 if '@bindir@' == ('@' + 'bindir@') or not os.path.exists (lilypond_binary):
-       lilypond_binary = 'lilypond-bin'
-
+       lilypond_binary = 'lilypond'
 
+psfonts_p = 0
 use_hash_p = 1
 format = 0
-output_name = 0
+output_name = ''
 latex_filter_cmd = 'latex "\\nonstopmode \input /dev/stdin"'
 filter_cmd = 0
-process_cmd = lilypond_binary
-default_ly_options = { }
+process_cmd = ''
+default_ly_options = { 'alt': "[image of music]" }
 
+#
+# Is this pythonic?  Personally, I find this rather #define-nesque. --hwn
+#
 AFTER = 'after'
 BEFORE = 'before'
+EXAMPLEINDENT = 'exampleindent'
+FILTER = 'filter'
+FRAGMENT = 'fragment'
 HTML = 'html'
+INDENT = 'indent'
 LATEX = 'latex'
+LAYOUT = 'layout'
 LINEWIDTH = 'linewidth'
+NOFRAGMENT = 'nofragment'
+NOINDENT = 'noindent'
+NOQUOTE = 'noquote'
 NOTES = 'body'
+NOTIME = 'notime'
 OUTPUT = 'output'
+OUTPUTIMAGE = 'outputimage'
+PACKED = 'packed'
 PAPER = 'paper'
 PREAMBLE = 'preamble'
+PRINTFILENAME = 'printfilename'
+QUOTE = 'quote'
+RAGGEDRIGHT = 'raggedright'
+RELATIVE = 'relative'
+STAFFSIZE = 'staffsize'
+TEXIDOC = 'texidoc'
 TEXINFO = 'texinfo'
 VERBATIM = 'verbatim'
+FONTLOAD = 'fontload'
+FILENAME = 'filename'
+ALT = 'alt'
+
+
+# NOTIME has no opposite so it isn't part of this dictionary.
+# NOQUOTE is used internally only.
+no_options = {
+       NOFRAGMENT: FRAGMENT,
+       NOINDENT: INDENT,
+}
 
 
-# Recognize special sequences in the input 
+# Recognize special sequences in the input.
 #
-# (?P<name>regex) -- assign result of REGEX to NAME
-# *? -- match non-greedily.
-# (?m) -- multiline regex: make ^ and $ match at each line
-# (?s) -- make the dot match all characters including newline
+#   (?P<name>regex) -- Assign result of REGEX to NAME.
+#   *? -- Match non-greedily.
+#   (?m) -- Multiline regex: Make ^ and $ match at each line.
+#   (?s) -- Make the dot match all characters including newline.
+#   (?x) -- Ignore whitespace in patterns.
 no_match = 'a\ba'
 snippet_res = {
+       ##
        HTML: {
-       'include':  no_match,
-       'lilypond' : '(?m)(?P<match><lilypond((?P<options>[^:]*):)(?P<code>.*?)/>)',
-       'lilypond-block': r'''(?ms)(?P<match><lilypond(?P<options>[^>]+)?>(?P<code>.*?)</lilypond>)''',
-       'lilypond-file': r'(?m)(?P<match><lilypondfile(?P<options>[^>]+)?>\s*(?P<filename>[^<]+)\s*</lilypondfile>)',
-       'multiline-comment': r"(?sm)\s*(?!@c\s+)(?P<code><!--\s.*?!-->)\s",
-       'singleline-comment': no_match,
-       'verb': r'''(?P<code><pre>.*?</pre>)''',
-       'verbatim': r'''(?s)(?P<code><pre>\s.*?</pre>\s)''',
+               'include':
+                 no_match,
+
+               'lilypond':
+                 r'''(?mx)
+                   (?P<match>
+                   <lilypond
+                     (\s*(?P<options>.*?)\s*:)?\s*
+                     (?P<code>.*?)
+                   />)''',
+
+               'lilypond_block':
+                 r'''(?msx)
+                   (?P<match>
+                   <lilypond
+                     \s*(?P<options>.*?)\s*
+                   >
+                   (?P<code>.*?)
+                   </lilypond>)''',
+
+               'lilypond_file':
+                 r'''(?mx)
+                   (?P<match>
+                   <lilypondfile
+                     \s*(?P<options>.*?)\s*
+                   >
+                   \s*(?P<filename>.*?)\s*
+                   </lilypondfile>)''',
+
+               'multiline_comment':
+                 r'''(?smx)
+                   (?P<match>
+                   \s*(?!@c\s+)
+                   (?P<code><!--\s.*?!-->)
+                   \s)''',
+
+               'singleline_comment':
+                 no_match,
+
+               'verb':
+                 r'''(?x)
+                   (?P<match>
+                     (?P<code><pre>.*?</pre>))''',
+
+               'verbatim':
+                 r'''(?x)
+                   (?s)
+                   (?P<match>
+                     (?P<code><pre>\s.*?</pre>\s))''',
        },
 
+       ##
        LATEX: {
-       'include': r'(?m)^[^%\n]*?(?P<match>\\input{(?P<filename>[^}]+)})',
-       'lilypond' : r'(?m)^[^%\n]*?(?P<match>\\lilypond\s*(\[(?P<options>.*?)\])?\s*{(?P<code>.*?)})',
-       'lilypond-block': r"(?sm)^[^%\n]*?(?P<match>\\begin\s*(\[(?P<options>.*?)\])?\s*{lilypond}(?P<code>.*?)\\end{lilypond})",
-       'lilypond-file': r'(?m)^[^%\n]*?(?P<match>\\lilypondfile\s*(\[(?P<options>.*?)\])?\s*\{(?P<filename>.+)})',
-       'multiline-comment': no_match,
-       'singleline-comment': r"(?m)^.*?(?P<match>(?P<code>^%.*$\n+))",
-       'verb': r"(?P<code>\\verb(?P<del>.).*?(?P=del))",
-       'verbatim': r"(?s)(?P<code>\\begin\s*{verbatim}.*?\\end{verbatim})",
+               'include':
+                 r'''(?smx)
+                   ^[^%\n]*?
+                   (?P<match>
+                   \\input\s*{
+                     (?P<filename>\S+?)
+                   })''',
+
+               'lilypond':
+                 r'''(?smx)
+                   ^[^%\n]*?
+                   (?P<match>
+                   \\lilypond\s*(
+                   \[
+                     \s*(?P<options>.*?)\s*
+                   \])?\s*{
+                     (?P<code>.*?)
+                   })''',
+
+               'lilypond_block':
+                 r'''(?smx)
+                   ^[^%\n]*?
+                   (?P<match>
+                   \\begin\s*(
+                   \[
+                     \s*(?P<options>.*?)\s*
+                   \])?\s*{lilypond}
+                     (?P<code>.*?)
+                   ^[^%\n]*?
+                   \\end\s*{lilypond})''',
+
+               'lilypond_file':
+                 r'''(?smx)
+                   ^[^%\n]*?
+                   (?P<match>
+                   \\lilypondfile\s*(
+                   \[
+                     \s*(?P<options>.*?)\s*
+                   \])?\s*\{
+                     (?P<filename>\S+?)
+                   })''',
+
+               'multiline_comment':
+                 no_match,
+
+               'singleline_comment':
+                 r'''(?mx)
+                   ^.*?
+                   (?P<match>
+                     (?P<code>
+                     %.*$\n+))''',
+
+               'verb':
+                 r'''(?mx)
+                   ^[^%\n]*?
+                   (?P<match>
+                     (?P<code>
+                     \\verb(?P<del>.)
+                       .*?
+                     (?P=del)))''',
+
+               'verbatim':
+                 r'''(?msx)
+                   ^[^%\n]*?
+                   (?P<match>
+                     (?P<code>
+                     \\begin\s*{verbatim}
+                       .*?
+                     \\end\s*{verbatim}))''',
        },
 
+       ##
        TEXINFO: {
-       'include':  '(?m)^[^%\n]*?(?P<match>@include\s+(?P<filename>\S*))',
-       'lilypond' : '(?m)^(?P<match>@lilypond(\[(?P<options>[^]]*)\])?{(?P<code>.*?)})',
-       'lilypond-block': r'''(?ms)^(?P<match>@lilypond(\[(?P<options>[^]]*)\])?\s(?P<code>.*?)@end lilypond)\s''',
-       'lilypond-file': '(?m)^(?P<match>@lilypondfile(\[(?P<options>[^]]*)\])?{(?P<filename>[^}]+)})',
-       'multiline-comment': r"(?sm)^\s*(?!@c\s+)(?P<code>@ignore\s.*?@end ignore)\s",
-       'singleline-comment': r"(?m)^.*?(?P<match>(?P<code>@c([ \t][^\n]*|)\n))",
-       'verb': r'''(?P<code>@code{.*?})''',
-       'verbatim': r'''(?s)(?P<code>@example\s.*?@end example\s)''',
+               'include':
+                 r'''(?mx)
+                   ^(?P<match>
+                   @include\s+
+                     (?P<filename>\S+))''',
+
+               'lilypond':
+                 r'''(?smx)
+                   ^[^\n]*?(?!@c\s+)[^\n]*?
+                   (?P<match>
+                   @lilypond\s*(
+                   \[
+                     \s*(?P<options>.*?)\s*
+                   \])?\s*{
+                     (?P<code>.*?)
+                   })''',
+
+               'lilypond_block':
+                 r'''(?msx)
+                   ^(?P<match>
+                   @lilypond\s*(
+                   \[
+                     \s*(?P<options>.*?)\s*
+                   \])?\s+?
+                   ^(?P<code>.*?)
+                   ^@end\s+lilypond)\s''',
+
+               'lilypond_file':
+                 r'''(?mx)
+                   ^(?P<match>
+                   @lilypondfile\s*(
+                   \[
+                     \s*(?P<options>.*?)\s*
+                   \])?\s*{
+                     (?P<filename>\S+)
+                   })''',
+
+               'multiline_comment':
+                 r'''(?smx)
+                   ^(?P<match>
+                     (?P<code>
+                     @ignore\s
+                       .*?
+                     @end\s+ignore))\s''',
+
+               'singleline_comment':
+                 r'''(?mx)
+                   ^.*
+                   (?P<match>
+                     (?P<code>
+                     @c([ \t][^\n]*|)\n))''',
+
+       # Don't do this: It interferes with @code{@{}.
+       #       'verb': r'''(?P<code>@code{.*?})''',
+
+               'verbatim':
+                 r'''(?sx)
+                   (?P<match>
+                     (?P<code>
+                     @example
+                       \s.*?
+                     @end\s+example\s))''',
        },
-       }
+}
 
 format_res = {
        HTML: {
-       'option-sep' : '\s*',
-       'intertext': r',?\s*intertext=\".*?\"',
+               'intertext': r',?\s*intertext=\".*?\"',
+               'option_sep': '\s*',
        },
+
        LATEX: {
-       'intertext': r',?\s*intertext=\".*?\"',
-       'option-sep' : ',\s*',
+               'intertext': r',?\s*intertext=\".*?\"',
+               'option_sep': '\s*,\s*',
        },
+
        TEXINFO: {
-       'intertext': r',?\s*intertext=\".*?\"',
-       'option-sep' : ',\s*',
+               'intertext': r',?\s*intertext=\".*?\"',
+               'option_sep': '\s*,\s*',
        },
-       }
+}
+
+# Options without a pattern in ly_options.
+simple_options = [
+       EXAMPLEINDENT,
+       FRAGMENT,
+       NOFRAGMENT,
+       NOINDENT,
+       PRINTFILENAME,
+       TEXIDOC,
+       VERBATIM,
+       FONTLOAD,
+       FILENAME,
+       ALT
+]
 
 ly_options = {
+       ##
        NOTES: {
-       'relative': r'''\relative c%(relative_quotes)s''',
+               RELATIVE: r'''\relative c%(relative_quotes)s''',
        },
+
+       ##
        PAPER: {
-       'indent' : r'''
-    indent = %(indent)s''',
-       'linewidth' : r'''
-    linewidth = %(linewidth)s''',
-       'noindent' : r'''
-    indent = 0.0\mm''',
-       'notime' : r'''
-    \translator {
-        \StaffContext
-        \remove Time_signature_engraver
-    }''',
-       'raggedright' : r'''
-    indent = 0.0\mm
-    raggedright = ##t''',
+               INDENT: r'''indent = %(indent)s''',
+
+               LINEWIDTH: r'''linewidth = %(linewidth)s''',
+
+               QUOTE: r'''linewidth = %(linewidth)s - 2.0 * %(exampleindent)s''',
+
+               RAGGEDRIGHT: r'''raggedright = ##t''',
+
+               PACKED: r'''packed = ##t''',
        },
+
+       ##
+       LAYOUT: {
+               NOTIME: r'''
+  \context {
+    \Score
+    timing = ##f
+  }
+  \context {
+    \Staff
+    \remove Time_signature_engraver
+  }''',
+       },
+
+       ##
        PREAMBLE: {
-       'staffsize': r'''
-#(set-global-staff-size %(staffsize)s)''',
+               STAFFSIZE: r'''#(set-global-staff-size %(staffsize)s)''',
        },
-       }
+}
 
 output = {
-       HTML : {
-       AFTER: '',
-       BEFORE: '',
-       OUTPUT: r'''<img align="center" valign="center"
-border="0" src="%(base)s.png" alt="[picture of music]">''',
-       VERBATIM: r'''<pre>
+       ##
+       HTML: {
+               FILTER: r'''<lilypond %(options)s>
+%(code)s
+</lilypond>
+''',
+
+               AFTER: r'''
+  </a>
+</p>''',
+
+               BEFORE: r'''<p>
+  <a href="%(base)s.ly">''',
+
+               OUTPUT: r'''
+    <img align="center" valign="center"
+        border="0" src="%(image)s" alt="%(alt)s">''',
+
+               PRINTFILENAME: '<p><tt><a href="%(base)s.ly">%(filename)s</a></tt></p>',
+
+               QUOTE: r'''<blockquote>
+%(str)s
+</blockquote>
+''',
+
+               VERBATIM: r'''<pre>
 %(verb)s</pre>''',
        },
-       
-       LATEX : {
-       AFTER: '',
-       BEFORE: '',
-       OUTPUT: r'''{\parindent 0pt
+
+       ##
+       LATEX: {
+               OUTPUT: r'''{%%
+\parindent 0pt
 \catcode`\@=12
-\ifx\preLilyPondExample\preLilyPondExample\fi
-\def\lilypondbook{}
-\input %(base)s.tex
-\ifx\preLilyPondExample\postLilyPondExample\fi
-\catcode`\@=0}''',
-       VERBATIM: r'''\begin{verbatim}
+\ifx\preLilyPondExample \undefined
+  \relax
+\else
+  \preLilyPondExample
+\fi
+\def\lilypondbook{}%%
+\input %(base)s-systems.tex
+\ifx\postLilyPondExample \undefined
+  \relax
+\else
+  \postLilyPondExample
+\fi
+}''',
+
+               PRINTFILENAME: '''\\texttt{%(filename)s}
+       ''',
+
+               QUOTE: r'''\begin{quotation}
+%(str)s
+\end{quotation}
+''',
+
+               VERBATIM: r'''\noindent
+\begin{verbatim}
 %(verb)s\end{verbatim}
 ''',
+
+               FILTER: r'''\begin{lilypond}[%(options)s]
+%(code)s
+\end{lilypond}''',
        },
-       
-       TEXINFO :       {
-       BEFORE: '',
-       AFTER: '',
-       VERBATIM: r'''@example
-%(verb)s@end example''',
-       
+
+       ##
+       TEXINFO: {
+               FILTER: r'''@lilypond[%(options)s]
+%(code)s
+@lilypond''',
+
+               OUTPUT: r'''
+@iftex
+@include %(base)s-systems.texi
+@end iftex
+''',
+
+               OUTPUTIMAGE: r'''@noindent
+@ifinfo
+@image{%(base)s,,,%(alt)s,%(ext)s}
+@end ifinfo
+@html
+<p>
+  <a href="%(base)s.ly">
+    <img align="center" valign="center"
+        border="0" src="%(image)s" alt="%(alt)s">
+  </a>
+</p>
+@end html
+''',
+
+               PRINTFILENAME: '''@file{%(filename)s}
+       ''',
+
+               QUOTE: r'''@quotation
+%(str)s@end quotation
+''',
+
+               NOQUOTE: r'''@format
+%(str)s@end format
+''',
+
+               VERBATIM: r'''@exampleindent 0
+@example
+%(verb)s@end example
+''',
        },
-       
-       }
+}
+
+PREAMBLE_LY = r'''%%%% Generated by %(program_name)s
+%%%% Options: [%(option_string)s]
+
+#(set! toplevel-score-handler ly:parser-print-score)
+#(set! toplevel-music-handler (lambda (p m)
+                              (ly:parser-print-score
+                               p (ly:music-scorify m p))))
+
+#(ly:set-option (quote no-point-and-click))
 
-PREAMBLE_LY = r'''%% Generated by %(program_name)s
-%% Options: [%(option_string)s]
+#(define version-seen? #t)
 %(preamble_string)s
-\paper {%(paper_string)s
+
+
+%% ********************************
+%% Start cut-&-pastable-section 
+%% ********************************
+
+\paper {
+  #(define dump-extents #t)
+  %(font_dump_setting)s
+  %(paper_string)s
 }
-''' 
-
-FRAGMENT_LY = r'''\score{
-    \notes%(notes_string)s{
-        %(code)s    }
-}'''
-FULL_LY = '%(code)s'
-
-def classic_lilypond_book_compatibility (o):
-       if o == 'singleline':
-               return 'raggedright'
-       m = re.search ('relative\s*([-0-9])', o)
-       if m:
-               return 'relative=%s' % m.group (1)
-       m = re.match ('([0-9]+)pt', o)
-       if m:
-               return 'staffsize=%s' % m.group (1)
-       m = re.match ('indent=([-.0-9]+)(cm|in|mm|pt)', o)
+
+\layout {
+  %(layout_string)s
+}
+'''
+
+FRAGMENT_LY = r'''
+%(notes_string)s
+{
+%% ly snippet contents follows:
+%(code)s
+%% end ly snippet
+}
+'''
+
+FULL_LY = '''
+%% ly snippet:
+%(code)s
+%% end ly snippet
+'''
+
+texinfo_linewidths = {
+       '@afourpaper': '160\\mm',
+       '@afourwide': '6.5\\in',
+       '@afourlatex': '150\\mm',
+       '@smallbook': '5\\in',
+       '@letterpaper': '6\\in',
+}
+
+def classic_lilypond_book_compatibility (key, value):
+       if key == 'singleline' and value == None:
+               return (RAGGEDRIGHT, None)
+
+       m = re.search ('relative\s*([-0-9])', key)
        if m:
-               f = float (m.group (1))
-               return 'indent=%f\\%s' % (f, m.group (2))
-       m = re.match ('linewidth=([-.0-9]+)(cm|in|mm|pt)', o)
+               return ('relative', m.group (1))
+
+       m = re.match ('([0-9]+)pt', key)
        if m:
-               f = float (m.group (1))
-               return 'linewidth=%f\\%s' % (f, m.group (2))
-       return None
+               return ('staffsize', m.group (1))
 
-def compose_ly (code, option_string):
-       options = []
-       # urg
-       for i in default_ly_options.keys ():
-               options.append (i)
-               vars ()[i] = default_ly_options[i]
+       if key == 'indent' or key == 'linewidth':
+               m = re.match ('([-.0-9]+)(cm|in|mm|pt|staffspace)', value)
+               if m:
+                       f = float (m.group (1))
+                       return (key, '%f\\%s' % (f, m.group (2)))
 
-       if option_string:
-               options = options + split_options (option_string)
-       
-       m = re.search (r'''\\score''', code)
-       if not m and (not options \
-                     or not 'nofragment' in options \
-                     or 'fragment' in options):
-               options.append ('raggedright')
-               body = FRAGMENT_LY
-       else:
-               body = FULL_LY
-
-       # defaults
-       relative = 0
-       staffsize = '16'
-
-       notes_options = []
-       paper_options = []
-       preamble_options = []
-       for i in options:
-               c = classic_lilypond_book_compatibility (i)
-               if c:
-                       ly.warning (_ ("deprecated ly-option used: %s" % i))
-                       ly.warning (_ ("compatibility mode translation: %s" \
-                                      % c))
-                       i = c
-               
-               if string.find (i, '=') > 0:
-                       key, value = string.split (i, '=')
-                       # hmm
-                       vars ()[key] = value
-               else:
-                       key = i
-
-               if key in ly_options[NOTES].keys ():
-                       notes_options.append (ly_options[NOTES][key])
-               elif key in ly_options[PREAMBLE].keys ():
-                       preamble_options.append (ly_options[PREAMBLE][key])
-               elif key in ly_options[PAPER].keys ():
-                       paper_options.append (ly_options[PAPER][key])
-               elif key not in ('fragment', 'nofragment',
-                                'relative', 'verbatim'):
-                       ly.warning (_("ignoring unknown ly option: %s") % i)
-
-       relative_quotes = (",,,", ",,", ",", "", "'", "''", "'''")[relative-3]
-       program_name = __main__.program_name
-       notes_string = string.join (notes_options, '\n    ') % vars ()
-       paper_string = string.join (paper_options, '\n    ') % vars ()
-       preamble_string = string.join (preamble_options, '\n    ') % vars ()
-       return (PREAMBLE_LY + body) % vars ()
-
-
-# BARF
-# use lilypond-bin for latex (.lytex) books,
-# and lilypond --preview for html, texinfo books?
-def to_eps (file):
-       cmd = r'latex "\nonstopmode \input %s"' % file
-       # Ugh.  (La)TeX writes progress and error messages on stdout
-       # Redirect to stderr
-       cmd = '(( %s  >&2 ) >&- )' % cmd
-       ly.system (cmd)
-       ly.system ('dvips -Ppdf -u+lilypond.map -E -o %s.eps %s' \
-                  % (file, file))
+       return (None, None)
 
 def find_file (name):
        for i in include_path:
                full = os.path.join (i, name)
                if os.path.exists (full):
                        return full
-       ly.error (_ ('file not found: %s\n' % name))
+       ly.error (_ ("file not found: %s") % name + '\n')
        ly.exit (1)
        return ''
-       
+
 def verbatim_html (s):
        return re.sub ('>', '&gt;',
                       re.sub ('<', '&lt;',
@@ -361,141 +639,479 @@ def verbatim_texinfo (s):
                               re.sub ('@', '@@', s)))
 
 def split_options (option_string):
-       return re.split (format_res[format]['option-sep'], option_string)
+       if option_string:
+               if format == HTML:
+                       options = re.findall('[\w\.-:]+(?:\s*=\s*(?:"[^"]*"|\'[^\']*\'|\S+))?',option_string)
+                       for i in range(len(options)):
+                               options[i] = re.sub('^([^=]+=\s*)(?P<q>["\'])(.*)(?P=q)','\g<1>\g<3>',options[i])
+                       return options
+               else:
+                       return re.split (format_res[format]['option_sep'],
+                                        option_string)
+       return []
 
+def invokes_lilypond ():
+       return re.search ('^[\'\"0-9A-Za-z/]*lilypond', process_cmd)
 
-## make index static of Snippet?
-index = 0
+def set_default_options (source):
+       global default_ly_options
+       if not default_ly_options.has_key (LINEWIDTH):
+               if format == LATEX:
+                       textwidth = get_latex_textwidth (source)
+                       default_ly_options[LINEWIDTH] = \
+                         '''%.0f\\pt''' % textwidth
+               elif format == TEXINFO:
+                       for (k, v) in texinfo_linewidths.items ():
+                               # FIXME: @layout is usually not in
+                               # chunk #0:
+                               #
+                               #  \input texinfo @c -*-texinfo-*-
+                               #
+                               # Bluntly search first K items of
+                               # source.
+                               # s = chunks[0].replacement_text ()
+                               if re.search (k, source[:1024]):
+                                       default_ly_options[LINEWIDTH] = v
+                                       break
+
+class Chunk:
+       def replacement_text (self):
+               return ''
 
-class Snippet:
-       def __init__ (self, type, source, index, match):
-               self.type = type
+       def filter_text (self):
+               return self.replacement_text ()
+
+       def ly_is_outdated (self):
+               return 0
+
+       def png_is_outdated (self):
+               return 0
+
+class Substring (Chunk):
+       def __init__ (self, source, start, end, line_number):
                self.source = source
-               self.index = index
+               self.start = start
+               self.end = end
+               self.line_number = line_number
+
+       def replacement_text (self):
+               return self.source[self.start:self.end]
+
+class Snippet (Chunk):
+       def __init__ (self, type, match, format, line_number):
+               self.type = type
                self.match = match
                self.hash = 0
+               self.option_dict = {}
+               self.format = format
+               self.line_number = line_number
 
-       def start (self, s):
-               return self.index + self.match.start (s)
-
-       def end (self, s):
-               return self.index + self.match.end (s)
+       def replacement_text (self):
+               return self.match.group ('match')
 
        def substring (self, s):
-               return self.source[self.start (s):self.end (s)]
+               return self.match.group (s)
+
+       def __repr__ (self):
+               return `self.__class__` + ' type = ' + self.type
+
+class Include_snippet (Snippet):
+       def processed_filename (self):
+               f = self.substring ('filename')
+               return os.path.splitext (f)[0] + format2ext[format]
+
+       def replacement_text (self):
+               s = self.match.group ('match')
+               f = self.substring ('filename')
+
+               return re.sub (f, self.processed_filename (), s)
+
+class Lilypond_snippet (Snippet):
+       def __init__ (self, type, match, format, line_number):
+               Snippet.__init__ (self, type, match, format, line_number)
+               os = match.group ('options')
+               self.do_options (os, self.type)
 
        def ly (self):
-               s = ''
-               if self.type == 'lilypond-block' or self.type == 'lilypond':
-                       s = self.substring ('code')
-               elif self.type == 'lilypond-file':
-                       name = self.substring ('filename')
-                       s = open (find_file (name)).read ()
-               return s
-               
+               return self.substring ('code')
+
        def full_ly (self):
                s = self.ly ()
                if s:
-                       return compose_ly (s, self.match.group ('options'))
+                       return self.compose_ly (s)
                return ''
-       
+
+       def do_options (self, option_string, type):
+               self.option_dict = {}
+
+               options = split_options (option_string)
+
+               for i in options:
+                       if string.find (i, '=') > 0:
+                               (key, value) = re.split ('\s*=\s*', i)
+                               self.option_dict[key] = value
+                       else:
+                               if i in no_options.keys ():
+                                       if no_options[i] in self.option_dict.keys ():
+                                               del self.option_dict[no_options[i]]
+                               else:
+                                       self.option_dict[i] = None
+
+               has_linewidth = self.option_dict.has_key (LINEWIDTH)
+               no_linewidth_value = 0
+
+               # If LINEWIDTH is used without parameter, set it to default.
+               if has_linewidth and self.option_dict[LINEWIDTH] == None:
+                       no_linewidth_value = 1
+                       del self.option_dict[LINEWIDTH]
+
+               for i in default_ly_options.keys ():
+                       if i not in self.option_dict.keys ():
+                               self.option_dict[i] = default_ly_options[i]
+
+               if not has_linewidth:
+                       if type == 'lilypond' or FRAGMENT in self.option_dict.keys ():
+                               self.option_dict[RAGGEDRIGHT] = None
+
+                       if type == 'lilypond':
+                               if LINEWIDTH in self.option_dict.keys ():
+                                       del self.option_dict[LINEWIDTH]
+                       else:
+                               if RAGGEDRIGHT in self.option_dict.keys ():
+                                       if LINEWIDTH in self.option_dict.keys ():
+                                               del self.option_dict[LINEWIDTH]
+
+                       if QUOTE in self.option_dict.keys () or type == 'lilypond':
+                               if LINEWIDTH in self.option_dict.keys ():
+                                       del self.option_dict[LINEWIDTH]
+
+               if not INDENT in self.option_dict.keys ():
+                       self.option_dict[INDENT] = '0\\mm'
+
+               # The QUOTE pattern from ly_options only emits the `linewidth'
+               # keyword.
+               if has_linewidth and QUOTE in self.option_dict.keys ():
+                       if no_linewidth_value:
+                               del self.option_dict[LINEWIDTH]
+                       else:
+                               del self.option_dict[QUOTE]
+
+       def compose_ly (self, code):
+               if FRAGMENT in self.option_dict.keys ():
+                       body = FRAGMENT_LY
+               else:
+                       body = FULL_LY
+
+               # Defaults.
+               relative = 1
+               override = {}
+               # The original concept of the `exampleindent' option is broken.
+               # It is not possible to get a sane value for @exampleindent at all
+               # without processing the document itself.  Saying
+               #
+               #   @exampleindent 0
+               #   @example
+               #   ...
+               #   @end example
+               #   @exampleindent 5
+               #
+               # causes ugly results with the DVI backend of texinfo since the
+               # default value for @exampleindent isn't 5em but 0.4in (or a smaller
+               # value).  Executing the above code changes the environment
+               # indentation to an unknown value because we don't know the amount
+               # of 1em in advance since it is font-dependent.  Modifying
+               # @exampleindent in the middle of a document is simply not
+               # supported within texinfo.
+               #
+               # As a consequence, the only function of @exampleindent is now to
+               # specify the amount of indentation for the `quote' option.
+               #
+               # To set @exampleindent locally to zero, we use the @format
+               # environment for non-quoted snippets.
+               override[EXAMPLEINDENT] = r'0.4\in'
+               override[LINEWIDTH] = texinfo_linewidths['@smallbook']
+               override.update (default_ly_options)
+
+               option_list = []
+               for (key, value) in self.option_dict.items ():
+                       if value == None:
+                               option_list.append (key)
+                       else:
+                               option_list.append (key + '=' + value)
+               option_string = string.join (option_list, ',')
+
+               compose_dict = {}
+               compose_types = [NOTES, PREAMBLE, LAYOUT, PAPER]
+               for a in compose_types:
+                       compose_dict[a] = []
+
+               for (key, value) in self.option_dict.items ():
+                       (c_key, c_value) = \
+                         classic_lilypond_book_compatibility (key, value)
+                       if c_key:
+                               if c_value:
+                                       ly.warning \
+                                         (_ ("deprecated ly-option used: %s=%s" \
+                                           % (key, value)))
+                                       ly.warning \
+                                         (_ ("compatibility mode translation: %s=%s" \
+                                           % (c_key, c_value)))
+                               else:
+                                       ly.warning \
+                                         (_ ("deprecated ly-option used: %s" \
+                                           % key))
+                                       ly.warning \
+                                         (_ ("compatibility mode translation: %s" \
+                                           % c_key))
+
+                               (key, value) = (c_key, c_value)
+
+                       if value:
+                               override[key] = value
+                       else:
+                               if not override.has_key (key):
+                                       override[key] = None
+
+                       found = 0
+                       for type in compose_types:
+                               if ly_options[type].has_key (key):
+                                       compose_dict[type].append (ly_options[type][key])
+                                       found = 1
+                                       break
+
+                       if not found and key not in simple_options:
+                               ly.warning (_ ("ignoring unknown ly option: %s") % key)
+
+               # URGS
+               if RELATIVE in override.keys () and override[RELATIVE]:
+                       relative = string.atoi (override[RELATIVE])
+
+               relative_quotes = ''
+
+               # 1 = central C
+               if relative < 0:
+                       relative_quotes += ',' * (- relative)
+               elif relative > 0:
+                       relative_quotes += "'" * relative
+
+               program_name = __main__.program_name
+
+               paper_string = string.join (compose_dict[PAPER],
+                                           '\n  ') % override
+               layout_string = string.join (compose_dict[LAYOUT],
+                                            '\n  ') % override
+               notes_string = string.join (compose_dict[NOTES],
+                                           '\n  ') % vars ()
+               preamble_string = string.join (compose_dict[PREAMBLE],
+                                              '\n  ') % override
+
+               font_dump_setting = ''
+               if FONTLOAD in self.option_dict:
+                       font_dump_setting = '#(define-public force-eps-font-include #t)\n'
+
+               return (PREAMBLE_LY + body) % vars ()
+
+       # TODO: Use md5?
        def get_hash (self):
                if not self.hash:
-                       self.hash = abs (hash (self.ly ()))
+                       self.hash = abs (hash (self.full_ly ()))
                return self.hash
 
        def basename (self):
+               if FILENAME in self.option_dict:
+                       return self.option_dict[FILENAME]
                if use_hash_p:
                        return 'lily-%d' % self.get_hash ()
                raise 'to be done'
 
        def write_ly (self):
-               if self.type == 'lilypond-block' or self.type == 'lilypond'\
-                      or self.type == 'lilypond-file':
-                       h = open (self.basename () + '.ly', 'w')
-                       h.write (self.full_ly ())
+               outf = open (self.basename () + '.ly', 'w')
+               outf.write (self.full_ly ())
+
+               open (self.basename () + '.txt', 'w').write ('image of music')
+
+       def ly_is_outdated (self):
+               base = self.basename ()
+
+               tex_file = '%s.tex' % base
+               eps_file = '%s.eps' % base
+               system_file = '%s-systems.tex' % base
+               ly_file = '%s.ly' % base
+               ok = os.path.exists (ly_file) \
+                    and os.path.exists (system_file)\
+                    and os.stat (system_file)[stat.ST_SIZE] \
+                    and re.match ('% eof', open (system_file).readlines ()[-1])
+               if ok and (not use_hash_p or FILENAME in self.option_dict):
+                       ok = (self.full_ly () == open (ly_file).read ())
+               if ok:
+                       # TODO: Do something smart with target formats
+                       #       (ps, png) and m/ctimes.
+                       return None
+               return self
+
+       def png_is_outdated (self):
+               base = self.basename ()
+               ok = self.ly_is_outdated ()
+               if format == HTML or format == TEXINFO:
+                       ok = ok and (os.path.exists (base + '.png')
+                                    or glob.glob (base + '-page*.png'))
+               return not ok
+       def texstr_is_outdated (self):
+               if backend == 'ps':
+                       return 0
+
+               base = self.basename ()
+               ok = self.ly_is_outdated ()
+               ok = ok and (os.path.exists (base + '.texstr'))
+               return not ok
+
+       def filter_text (self):
+               code = self.substring ('code')
+               s = run_filter (code)
+               d = {
+                       'code': s,
+                       'options': self.match.group ('options')
+               }
+               # TODO
+               return output[self.format][FILTER] % d
+
+       def replacement_text (self):
+               func = Lilypond_snippet.__dict__['output_' + self.format]
+               return func (self)
+
+       def get_images (self):
+               base = self.basename ()
+               # URGUGHUGHUGUGH
+               single = '%(base)s.png' % vars ()
+               multiple = '%(base)s-page1.png' % vars ()
+               images = (single,)
+               if os.path.exists (multiple) \
+                  and (not os.path.exists (single) \
+                       or (os.stat (multiple)[stat.ST_MTIME] \
+                           > os.stat (single)[stat.ST_MTIME])):
+                       images = glob.glob ('%(base)s-page*.png' % vars ())
+               return images
 
        def output_html (self):
+               str = ''
                base = self.basename ()
-               option_string = self.match.group ('options')
-               if option_string and VERBATIM in split_options (option_string)\
-                  and format == HTML:
-                       verb = verbatim_html (self.substring ('code'))
-                       h.write (output[HTML][VERBATIM] % vars ())
-               h.write (output[HTML][BEFORE])
-               h.write (output[HTML][OUTPUT] % vars ())
-               h.write (output[HTML][AFTER])
-                       
+               if format == HTML:
+                       str += self.output_print_filename (HTML)
+                       if VERBATIM in self.option_dict:
+                               verb = verbatim_html (self.substring ('code'))
+                               str += write (output[HTML][VERBATIM] % vars ())
+                       if QUOTE in self.option_dict:
+                               str = output[HTML][QUOTE] % vars ()
+
+               str += output[HTML][BEFORE] % vars ()
+               for image in self.get_images ():
+                       (base, ext) = os.path.splitext (image)
+                       alt = self.option_dict[ALT]
+                       str += output[HTML][OUTPUT] % vars ()
+               str += output[HTML][AFTER] % vars ()
+               return str
+
+       def output_info (self):
+               str = ''
+               for image in self.get_images ():
+                       (base, ext) = os.path.splitext (image)
+
+                       # URG, makeinfo implicitly prepends dot to extension.
+                       # Specifying no extension is most robust.
+                       ext = ''
+                       alt = self.option_dict[ALT]
+                       str += output[TEXINFO][OUTPUTIMAGE] % vars ()
+
+               base = self.basename ()
+               str += output[format][OUTPUT] % vars ()
+               return str
+
        def output_latex (self):
-               option_string = self.match.group ('options')
-               if option_string and VERBATIM in split_options (option_string)\
-                  and format == LATEX:
-                       verb = self.substring ('code')
-                       h.write (output[LATEX][VERBATIM] % vars ())
-               h.write (output[LATEX][BEFORE])
+               str = ''
                base = self.basename ()
-               h.write (output[LATEX][OUTPUT] % vars ())
-               h.write (output[LATEX][AFTER])
-                       
+               if format == LATEX:
+                       str += self.output_print_filename (LATEX)
+                       if VERBATIM in self.option_dict:
+                               verb = self.substring ('code')
+                               str += (output[LATEX][VERBATIM] % vars ())
+                       if QUOTE in self.option_dict:
+                               str = output[LATEX][QUOTE] % vars ()
+
+               str += (output[LATEX][OUTPUT] % vars ())
+               return str
+
+       def output_print_filename (self, format):
+               str = ''
+               if PRINTFILENAME in self.option_dict:
+                       base = self.basename ()
+                       filename = self.substring ('filename')
+                       str = output[format][PRINTFILENAME] % vars ()
+
+               return str
+
        def output_texinfo (self):
-               option_string = self.match.group ('options')
-               if option_string and VERBATIM in split_options (option_string):
-                       verb = verbatim_texinfo (self.substring ('code'))
-                       h.write (output[TEXINFO][VERBATIM] % vars ())
-               h.write ('\n@tex\n')
-               self.output_latex ()
-               h.write ('\n@end tex\n')
-               
-               h.write ('\n@html\n')
-               self.output_html ()
-               h.write ('\n@end html\n')
-                       
-       def outdated_p (self):
-               if self.type != 'lilypond-block' and self.type != 'lilypond'\
-                      and self.type != 'lilypond-file':
-                       return None
+               str = ''
+               if self.output_print_filename (TEXINFO):
+                       str += ('@html\n'
+                               + self.output_print_filename (HTML)
+                               + '\n@end html\n')
+                       str += ('@tex\n'
+                               + self.output_print_filename (LATEX)
+                               + '\n@end tex\n')
                base = self.basename ()
-               if os.path.exists (base + '.ly') \
-                  and os.path.exists (base + '.tex') \
-                  and (use_hash_p \
-                       or self.ly () == open (base + '.ly').read ()):
-                       # TODO: something smart with target formats
-                       # (ps, png) and m/ctimes
-                       return None
-               return self
+               if TEXIDOC in self.option_dict:
+                       texidoc = base + '.texidoc'
+                       if os.path.exists (texidoc):
+                               str += '@include %(texidoc)s\n\n' % vars ()
 
-       def filter_code (self):
-               global index
-               # Hmm, why is verbatim's group called 'code'; rename to 'verb'?
-               #if snippet.match.group ('code'):
-               # urg
-               if self.type == 'lilypond' or self.type == 'lilypond-block':
-                       h.write (self.source[index:self.start ('code')])
-                       h.write (run_filter (self.substring ('code')))
-                       h.write (self.source[self.end ('code'):self.end (0)])
-               else:
-                       h.write (self.source[index:self.end (0)])
-               index = self.end (0)
-
-       def compile_output (self):
-               global index
-               # Hmm, why is verbatim's group called 'code'; rename to 'verb'?
-               # if snippet.match.group ('code'):
-               # urg
-               if self.type == 'lilypond' \
-                      or self.type == 'lilypond-block'\
-                      or self.type == 'lilypond-file':
-                       h.write (self.source[index:self.start (0)])
-                       snippet_output = eval ("Snippet.output_" + format)
-                       snippet_output (self)
-               elif self.type == 'include':
-                       h.write (self.source[index:self.start ('filename')])
-                       base = os.path.splitext (self.substring ('filename'))[0]
-                       h.write (base + format2ext[format])
-                       h.write (self.source[self.end ('filename'):self.end (0)])
-               else:
-                       h.write (self.source[index:self.end (0)])
-               index = self.end (0)
+               if VERBATIM in self.option_dict:
+                       verb = verbatim_texinfo (self.substring ('code'))
+                       str += (output[TEXINFO][VERBATIM] % vars ())
+                       if not QUOTE in self.option_dict:
+                               str = output[TEXINFO][NOQUOTE] % vars ()
+
+               str += self.output_info ()
+
+#              str += ('@ifinfo\n' + self.output_info () + '\n@end ifinfo\n')
+#              str += ('@tex\n' + self.output_latex () + '\n@end tex\n')
+#              str += ('@html\n' + self.output_html () + '\n@end html\n')
+
+               if QUOTE in self.option_dict:
+                       str = output[TEXINFO][QUOTE] % vars ()
+
+               # need par after image
+               str += '\n'
+
+               return str
+
+class Lilypond_file_snippet (Lilypond_snippet):
+       def ly (self):
+               name = self.substring ('filename')
+               return '\\renameinput \"%s\"\n%s' \
+                        % (name, open (find_file (name)).read ())
+
+snippet_type_to_class = {
+       'lilypond_file': Lilypond_file_snippet,
+       'lilypond_block': Lilypond_snippet,
+       'lilypond': Lilypond_snippet,
+       'include': Include_snippet,
+}
+
+def find_linestarts (s):
+       nls = [0]
+       start = 0
+       end = len (s)
+       while 1:
+               i = s.find ('\n', start)
+               if i < 0:
+                       break
+
+               i = i + 1
+               nls.append (i)
+               start = i
+
+       nls.append (len (s))
+       return nls
 
 def find_toplevel_snippets (s, types):
        res = {}
@@ -504,33 +1120,83 @@ def find_toplevel_snippets (s, types):
 
        snippets = []
        index = 0
-       found = {}.fromkeys (types)
+       ## found = dict (map (lambda x: (x, None),
+       ##                    types))
+       ## urg python2.1
+       found = {}
+       map (lambda x, f = found: f.setdefault (x, None),
+            types)
+
+       line_starts = find_linestarts (s)
+       line_start_idx = 0
+       # We want to search for multiple regexes, without searching
+       # the string multiple times for one regex.
+       # Hence, we use earlier results to limit the string portion
+       # where we search.
+       # Since every part of the string is traversed at most once for
+       # every type of snippet, this is linear.
+
        while 1:
-               first = 0
+               first = None
                endex = 1 << 30
-               for i in types:
-                       if not found[i] or found[i].start (0) < index:
-                               found[i] = 0
-                               m = res[i].search (s[index:endex])
-                               if m:
-                                       found[i] = Snippet (i, s, index, m)
-                       if found[i] \
-                              and (not first \
-                                   or found[i].start (0) < found[first].start (0)):
-                               first = i
-                               endex = found[first].start (0)
+               for type in types:
+                       if not found[type] or found[type][0] < index:
+                               found[type] = None
+                               m = res[type].search (s[index:endex])
+                               if not m:
+                                       continue
+
+                               cl = Snippet
+                               if snippet_type_to_class.has_key (type):
+                                       cl = snippet_type_to_class[type]
+
+
+                               start = index + m.start ('match')
+                               line_number = line_start_idx
+                               while (line_starts[line_number] < start):
+                                       line_number += 1
+
+                               line_number += 1
+                               snip = cl (type, m, format, line_number)
+                               found[type] = (start, snip)
+
+                       if found[type] \
+                          and (not first \
+                               or found[type][0] < found[first][0]):
+                               first = type
+
+                               # FIXME.
+
+                               # Limiting the search space is a cute
+                               # idea, but this *requires* to search
+                               # for possible containing blocks
+                               # first, at least as long as we do not
+                               # search for the start of blocks, but
+                               # always/directly for the entire
+                               # @block ... @end block.
+
+                               endex = found[first][0]
+
                if not first:
+                       snippets.append (Substring (s, index, len (s), line_start_idx))
                        break
-               snippets.append (found[first])
-               index = found[first].end (0)
-               
+
+               while (start > line_starts[line_start_idx+1]):
+                       line_start_idx += 1
+
+               (start, snip) = found[first]
+               snippets.append (Substring (s, index, start, line_start_idx + 1))
+               snippets.append (snip)
+               found[first] = None
+               index = start + len (snip.match.group ('match'))
+
        return snippets
 
 def filter_pipe (input, cmd):
        if verbose_p:
-               ly.progress (_ ("Opening filter `%s\'") % cmd)
-               
-       stdin, stdout, stderr = os.popen3 (cmd)
+               ly.progress (_ ("Opening filter `%s'") % cmd)
+
+       (stdin, stdout, stderr) = os.popen3 (cmd)
        stdin.write (input)
        status = stdin.close ()
 
@@ -539,36 +1205,62 @@ def filter_pipe (input, cmd):
                output = stdout.read ()
                status = stdout.close ()
                error = stderr.read ()
-               
+
        if not status:
                status = 0
        signal = 0x0f & status
        if status or (not output and error):
                exit_status = status >> 8
-               ly.error (_ ("`%s\' failed (%d)") % (cmd, exit_status))
+               ly.error (_ ("`%s' failed (%d)") % (cmd, exit_status))
                ly.error (_ ("The error log is as follows:"))
                sys.stderr.write (error)
                sys.stderr.write (stderr.read ())
                ly.exit (status)
-       
+
        if verbose_p:
                ly.progress ('\n')
 
        return output
-       
+
 def run_filter (s):
        return filter_pipe (s, filter_cmd)
 
-def process_snippets (cmd, snippets):
-       names = filter (lambda x:x, map (Snippet.basename, snippets))
-       if names:
-               ly.system (string.join ([cmd] + names))
-
-       if format == HTML or format == TEXINFO:
-               for i in names:
-                       if os.path.exists (i + '.tex'):
-                               to_eps (i)
-                               ly.make_ps_images (i + '.eps', resolution=110)
+def is_derived_class (cl, baseclass):
+       if cl == baseclass:
+               return 1
+       for b in cl.__bases__:
+               if is_derived_class (b, baseclass):
+                       return 1
+       return 0
+
+def process_snippets (cmd, ly_snippets, texstr_snippets, png_snippets):
+       ly_names = filter (lambda x: x,
+                          map (Lilypond_snippet.basename, ly_snippets))
+       texstr_names = filter (lambda x: x,
+                          map (Lilypond_snippet.basename, texstr_snippets))
+       png_names = filter (lambda x: x,
+                           map (Lilypond_snippet.basename, png_snippets))
+
+       status = 0
+       def my_system (cmd):
+               status = ly.system (cmd,
+                                   ignore_error = 1, progress_p = 1)
+
+               if status:
+                       ly.error ('Process %s exited unsuccessfully.' % cmd)
+                       raise Compile_error
+
+       # UGH
+       # the --process=CMD switch is a bad idea
+       # it is too generic for lilypond-book.
+       if texstr_names and invokes_lilypond:
+               my_system (string.join ([cmd, '--backend texstr',
+                                        'snippet-map.ly'] + texstr_names))
+               for l in texstr_names:
+                       my_system ('latex %s.texstr' % l)
+
+       if ly_names:
+               my_system (string.join ([cmd, 'snippet-map.ly'] + ly_names))
 
 LATEX_DOCUMENT = r'''
 %(preamble)s
@@ -578,12 +1270,13 @@ LATEX_DOCUMENT = r'''
 \makeatletter\if@twocolumn\typeout{columns=2}\fi\makeatother
 \end{document}
 '''
-#need anything else besides textwidth?
+
+# Do we need anything else besides `textwidth'?
 def get_latex_textwidth (source):
-       m = re.search (r'''(?P<preabmle>\\begin\s*{document})''', source)
+       m = re.search (r'''(?P<preamble>\\begin\s*{document})''', source)
        preamble = source[:m.start (0)]
        latex_document = LATEX_DOCUMENT % vars ()
-        parameter_string = filter_pipe (latex_document, latex_filter_cmd)
+       parameter_string = filter_pipe (latex_document, latex_filter_cmd)
 
        columns = 0
        m = re.search ('columns=([0-9.]*)', parameter_string)
@@ -596,7 +1289,7 @@ def get_latex_textwidth (source):
                columnsep = string.atof (m.group (1))
 
        textwidth = 0
-       m = re.search('textwidth=([0-9.]*)pt', parameter_string)
+       m = re.search ('textwidth=([0-9.]*)pt', parameter_string)
        if m:
                textwidth = string.atof (m.group (1))
                if columns:
@@ -604,129 +1297,211 @@ def get_latex_textwidth (source):
 
        return textwidth
 
-
 ext2format = {
-       '.html' : HTML,
-       '.itely' : TEXINFO,
-       '.lytex' : LATEX,
-       '.tely' : TEXINFO,
+       '.html': HTML,
+       '.itely': TEXINFO,
+       '.latex': LATEX,
+       '.lytex': LATEX,
+       '.tely': TEXINFO,
        '.tex': LATEX,
-       '.texi' : TEXINFO,
-       '.texinfo' : TEXINFO,
-       '.xml' : HTML,
-       }
-                              
+       '.texi': TEXINFO,
+       '.texinfo': TEXINFO,
+       '.xml': HTML,
+}
+
 format2ext = {
        HTML: '.html',
-       #TEXINFO: '.texinfo',
+       # TEXINFO: '.texinfo',
        TEXINFO: '.texi',
        LATEX: '.tex',
-       }
+}
 
-def do_file (input_filename):
-       #ugh
-       global format
-       if not format:
-               e = os.path.splitext (input_filename)[1]
-               if e in ext2format.keys ():
-                       #FIXME
-                       format = ext2format[e]
-               else:
-                       ly.error (_ ("cannot determine format for: %s" \
-                                    % input_filename))
+class Compile_error:
+       pass
 
-       ly.progress (_ ("Reading %s...") % input_filename)
-       if not input_filename or input_filename == '-':
-               ih = sys.stdin
+def write_file_map (lys, name):
+       snippet_map = open ('snippet-map.ly', 'w')
+       snippet_map.write ("""
+#(define version-seen? #t)
+#(ly:add-file-name-alist '(
+""")
+       for ly in lys:
+               snippet_map.write ('("%s.ly" . "%s:%d (%s.ly)")\n'
+                                  % (ly.basename (),
+                                     name,
+                                     ly.line_number,
+                                     ly.basename ()))
+
+       snippet_map.write ('))\n')
+
+def do_process_cmd (chunks, input_name):
+       all_lys = filter (lambda x: is_derived_class (x.__class__,
+                                                     Lilypond_snippet),
+                         chunks)
+
+       write_file_map (all_lys, input_name)
+       ly_outdated = \
+         filter (lambda x: is_derived_class (x.__class__,
+                                             Lilypond_snippet)
+                           and x.ly_is_outdated (),
+                 chunks)
+       texstr_outdated = \
+         filter (lambda x: is_derived_class (x.__class__,
+                                             Lilypond_snippet)
+                           and x.texstr_is_outdated (),
+                 chunks)
+       png_outdated = \
+         filter (lambda x: is_derived_class (x.__class__,
+                                             Lilypond_snippet)
+                           and x.png_is_outdated (),
+                 chunks)
+
+       ly.progress (_ ("Writing snippets..."))
+       map (Lilypond_snippet.write_ly, ly_outdated)
+       ly.progress ('\n')
+
+       if ly_outdated:
+               ly.progress (_ ("Processing..."))
+               ly.progress ('\n')
+               process_snippets (process_cmd, ly_outdated, texstr_outdated, png_outdated)
        else:
-               ih = open (input_filename)
-       source = ih.read ()
+               ly.progress (_ ("All snippets are up to date..."))
        ly.progress ('\n')
 
-       ly.progress (_ ("Dissecting..."))
-       #snippets = find_toplevel_snippets (source, snippet_res[format].keys ())
-       snippet_types = (
-               'lilypond-block',
-               'verb',
-               'verbatim',
-               'singleline-comment',
-               'multiline-comment',
-               'lilypond-file',
-               'include',
-               'lilypond', )
-       
-       snippets = find_toplevel_snippets (source, snippet_types)
+def guess_format (input_filename):
+       format = None
+       e = os.path.splitext (input_filename)[1]
+       if e in ext2format.keys ():
+               # FIXME
+               format = ext2format[e]
+       else:
+               ly.error (_ ("can't determine format for: %s" \
+                            % input_filename))
+               ly.exit (1)
+       return format
+
+def write_if_updated (file_name, lines):
+       try:
+               f = open (file_name)
+               oldstr = f.read ()
+               new_str = string.join (lines, '')
+               if oldstr == new_str:
+                       ly.progress (_ ("%s is up to date.") % file_name)
+                       ly.progress ('\n')
+                       return
+       except:
+               pass
+
+       ly.progress (_ ("Writing `%s'...") % file_name)
+       open (file_name, 'w').writelines (lines)
        ly.progress ('\n')
 
-       global h
-       if output_name == '-' or not output_name:
-               h = sys.stdout
-               output_filename = '-'
+def do_file (input_filename):
+       # Ugh.
+       if not input_filename or input_filename == '-':
+               in_handle = sys.stdin
+               input_fullname = '<stdin>'
        else:
-               if not os.path.isdir (output_name):
-                       os.mkdir (output_name, 0777)
-               if input_filename == '-':
-                       input_base = 'stdin'
+               if os.path.exists (input_filename):
+                       input_fullname = input_filename
+               elif format == LATEX:
+                       # urg python interface to libkpathsea?
+                       input_fullname = ly.read_pipe ('kpsewhich '
+                                                      + input_filename)[:-1]
                else:
-                       input_base = os.path.splitext (input_filename)[0]
-               output_filename = output_name + '/' + input_base \
-                                 + format2ext[format]
-               h = open (output_filename, 'w')
-               os.chdir (output_name)
+                       input_fullname = find_file (input_filename)
+               in_handle = open (input_fullname)
 
-       global default_ly_options
-       textwidth = 0
-       if format == LATEX and LINEWIDTH not in default_ly_options.keys ():
-               textwidth = get_latex_textwidth (source)
-               default_ly_options[LINEWIDTH] = '''%.0f\pt''' % textwidth
-
-       global index
-       if filter_cmd:
-               index = 0
-               map (Snippet.filter_code, snippets)
-               h.write (source[index:])
-       elif process_cmd:
-               outdated = filter (lambda x:x,
-                                  map (Snippet.outdated_p, snippets))
-               ly.progress (_ ("Writing snippets..."))
-               map (Snippet.write_ly, snippets)
+       if input_filename == '-':
+               input_base = 'stdin'
+       else:
+               input_base = os.path.basename \
+                            (os.path.splitext (input_filename)[0])
+
+       # Only default to stdout when filtering.
+       if output_name == '-' or (not output_name and filter_cmd):
+               output_filename = '-'
+               output_file = sys.stdout
+       else:
+               if output_name:
+                       if not os.path.isdir (output_name):
+                               os.mkdir (output_name, 0777)
+                       os.chdir (output_name)
+
+               output_filename = input_base + format2ext[format]
+               if os.path.exists (input_filename) \
+                  and os.path.exists (output_filename) \
+                  and os.path.samefile (output_filename, input_fullname):
+                       ly.error (
+                         _ ("Output would overwrite input file; use --output."))
+                       ly.exit (2)
+
+       try:
+               ly.progress (_ ("Reading %s...") % input_fullname)
+               source = in_handle.read ()
                ly.progress ('\n')
-               
-               if outdated:
-                       ly.progress (_ ("Processing..."))
-                       process_snippets (process_cmd, outdated)
-               else:
-                       ly.progress (_ ("All snippets are up to date..."))
+
+               set_default_options (source)
+
+
+               # FIXME: Containing blocks must be first, see
+               #        find_toplevel_snippets.
+               snippet_types = (
+                       'multiline_comment',
+                       'verbatim',
+                       'lilypond_block',
+       #               'verb',
+                       'singleline_comment',
+                       'lilypond_file',
+                       'include',
+                       'lilypond',
+               )
+               ly.progress (_ ("Dissecting..."))
+               chunks = find_toplevel_snippets (source, snippet_types)
                ly.progress ('\n')
+
+               if filter_cmd:
+                       write_if_updated (output_filename,
+                                         [c.filter_text () for c in chunks])
+               elif process_cmd:
+                       do_process_cmd (chunks, input_fullname)
+                       ly.progress (_ ("Compiling %s...") % output_filename)
+                       ly.progress ('\n')
+                       write_if_updated (output_filename,
+                                         [s.replacement_text ()
+                                          for s in chunks])
                
-               ly.progress (_ ("Compiling %s...") % output_filename)
-               index = 0
-               map (Snippet.compile_output, snippets)
-               h.write (source[index:])
-               ly.progress ('\n')
+               def process_include (snippet):
+                       os.chdir (original_dir)
+                       name = snippet.substring ('filename')
+                       ly.progress (_ ("Processing include: %s") % name)
+                       ly.progress ('\n')
+                       return do_file (name)
+
+               include_chunks = map (process_include,
+                                     filter (lambda x: is_derived_class (x.__class__,
+                                                                         Include_snippet),
+                                             chunks))
 
-       if h != sys.stdout:
-               h.close ()
 
-       def process_include (snippet):
+               return chunks + reduce (lambda x,y: x + y, include_chunks, [])
+               
+       except Compile_error:
                os.chdir (original_dir)
-               name = snippet.substring ('filename')
-               ly.progress (_ ('Processing include: %s') % name)
+               ly.progress (_ ("Removing `%s'") % output_filename)
                ly.progress ('\n')
-               do_file (name)
-
-       map (process_include, filter (lambda x: x.type == 'include', snippets))
+               raise Compile_error
 
 def do_options ():
-       global format, output_name
+       global format, output_name, psfonts_p
        global filter_cmd, process_cmd, verbose_p
-       
+
        (sh, long) = ly.getopt_args (option_definitions)
        try:
                (options, files) = getopt.getopt (sys.argv[1:], sh, long)
        except getopt.error, s:
                sys.stderr.write ('\n')
-               ly.error (_ ("getopt says: `%s\'" % s))
+               ly.error (_ ("getopt says: `%s'" % s))
                sys.stderr.write ('\n')
                ly.help ()
                ly.exit (2)
@@ -744,6 +1519,8 @@ def do_options ():
                        format = a
                        if a == 'texi-html' or a == 'texi':
                                format = TEXINFO
+               elif o == '--tex-backend ':
+                       backend = 'tex'
                elif o == '--help' or o == '-h':
                        ly.help ()
                        sys.exit (0)
@@ -752,8 +1529,6 @@ def do_options ():
                                                           ly.abspath (a)))
                elif o == '--output' or o == '-o':
                        output_name = a
-               elif o == '--outdir':
-                       output_name = a
                elif o == '--process' or o == '-P':
                        process_cmd = a
                        filter_cmd = 0
@@ -762,6 +1537,8 @@ def do_options ():
                        sys.exit (0)
                elif o == '--verbose' or o == '-V':
                        verbose_p = 1
+               elif o == '--psfonts':
+                       psfonts_p = 1 
                elif o == '--warranty' or o == '-w':
                        if 1 or status:
                                ly.warranty ()
@@ -770,10 +1547,66 @@ def do_options ():
 
 def main ():
        files = do_options ()
+       if not files or len (files) > 1:
+               ly.help ()
+               ly.exit (2)
+
+       file = files[0]
+
+       basename = os.path.splitext (file)[0]
+       basename = os.path.split (basename)[1]
+       
+       global process_cmd, format
+       if not format:
+               format = guess_format (files[0])
+
+       formats = 'ps'
+       if format == TEXINFO or format == HTML:
+               formats += ',png'
+       if process_cmd == '':
+               process_cmd = lilypond_binary \
+                             + ' --formats=%s --backend eps ' % formats
+
+       if process_cmd:
+               process_cmd += string.join ([(' -I %s' % p)
+                                            for p in include_path])
+
        ly.identify (sys.stderr)
        ly.setup_environment ()
-       if files:
-               do_file (files[0])
+
+       try:
+               chunks = do_file (file)
+               if psfonts_p and invokes_lilypond ():
+                       fontextract.verbose = verbose_p
+                       snippet_chunks = filter (lambda x: is_derived_class (x.__class__,
+                                                                             Lilypond_snippet),
+                                                chunks)
+
+                       psfonts_file = basename + '.psfonts' 
+                       if not verbose_p:
+                               ly.progress (_ ("Writing fonts to %s...") % psfonts_file)
+                       fontextract.extract_fonts (psfonts_file,
+                                                  [x.basename() + '.eps'
+                                                   for x in snippet_chunks])
+                       if not verbose_p:
+                               ly.progress ('\n')
+                       
+       except Compile_error:
+               ly.exit (1)
+
+       if format == TEXINFO or format == LATEX:
+               if not psfonts_p:
+                       ly.warning (_ ("option --psfonts not used"))
+                       ly.warning (_ ("processing with dvips will have no fonts"))
+
+               psfonts_file = os.path.join (output_name, basename + '.psfonts')
+               output = os.path.join (output_name, basename +  '.dvi' )
+               
+               ly.progress ('\n')
+               ly.progress (_ ("DVIPS usage:"))
+               ly.progress ('\n')
+               ly.progress ("    dvips -h %(psfonts_file)s %(output)s" % vars ())
+               ly.progress ('\n')
 
 if __name__ == '__main__':
        main ()