* lily/new-lyric-combine-music-iterator.cc (find_voice): allow

[lilypond.git] / Documentation / topdocs / NEWS.texi

\input texinfo @c -*-texinfo-*-
@setfilename NEWS.info
@settitle NEWS

@node Top, , , 
@top
@unnumbered New features in 2.3 since 2.2

@itemize @bullet
@item The melody  for a @code{\lyricsto} text can be changed during a melody.

@item The @code{\lyricsto} keyword now automatically sets @code{\lyrics}.

@item The LilyPond binary  now supports output options
@code{--dvi} @code{--png} @code{--ps} @code{--pdf} and
@code{--preview}.  The old wrapper script for La@TeX{} has been
renamed to @code{lilypond-latex}.

@item Support ambituses has been rewritten. The Ambitus object is now
a grouping object, composed of @code{AmbitusLine},
@code{AmbitusAccidental} and @code{AmbitusNoteHead}. These objects may
be tuned similar to accidentals and note heads.

@item Ledger lines are now shortened to prevent them from colliding.

@item Slur formatting has been rewritten. The new slur code
works similar to the Beam formatter: scores are assigned for all
esthetic components of a slur. A large number of combinations for
begin and end points is then tried out.  Slurs will now also take into
account collisions with staff lines, scripts (like staccato and
accent) and accidentals.


@item In the LilyPond  emacs mode, the @code{|} will now display the
  current beat within the measure. 

@item Colliding notes are now correctly aligned relative to notes in other staves.

@item An experimental GNOME output backend is available for developers.
It depends on several unreleased packages such as gnome-guile TLA and
Pango CVS.  The output interface coverage is sparse, but it already
does support point-and-click on the GNOME Canvas.  See also
@file{scm/output-gnome.scm}.

@item Context definitions are now stored in identifiers that match the
context name, in other words, to modify a @code{Staff} context, the
following code may be used.

@example
\context @{
  \Staff
  .. 
@}
@end example

@item A new block, @code{\bookpaper} has been introduced to
hold settings for paper size and output scaling. Further options
include @code{raggedbottom} (if set, systems are not vertically filled
to reach the bottom of the page), and @code{raggedlastbottom}.



@item Support for fret diagrams  has been contributed by Carl
D. Sorensen. @file{input/test/fret-diagram.ly} contains an example.
 
@item The @code{--safe-mode} has been revisited: GUILE evaluation
is done in the R5RS safe module, with only the basic @code{ly:}
interface available and malicious @TeX{} code is stopped.  However, to
be reasonably safe, you are advised to use the PostScript backend
rather than the @TeX{} backend anyway; and if possible use an UML
sandbox to run gs or latex.

@item Music syntax can now be extended seamlessly. As an example,
here is the new implementation @code{\applymusic},

@example
  applymusic = #(ly:make-music-function
     (list procedure? ly:music?) ; signature
     (lambda (where func music)  ; the function
       (func music)))
@end example

@item @code{\apply} has been renamed to @code{\applymusic}.

@item Music can be used as a markup.
When inserting a @code{score} block as part of a @code{\markup}, it
produces a rendered markup of the music.  An example is in
@file{input/test/markup-score.ly}.

@item LilyPond expressions can be embedded into Scheme. The syntax for
this is @code{#@{ ... #@}}. Within such a block, Scheme forms can be
introduced using a @code{$} character (@code{$$} results in a single
`$' character).  These forms are then inserted in the pattern.

@example
#(define (textoffset dx dy)
   (ly:export
     #@{ \override Voice.TextScript #'extra-offset = #(cons $dx $dy) #@}))

@{
  c'^"normal text"
  %% The following embedded scheme call is the same as
  %% \override Voice.TextScript #'extra-offset = #(cons 2 -3)
  #(textoffset 2 -3)
  c'^"text with offset" 
@}
@end example

@item A music list at toplevel is interpreted as implicit @code{\score}
a @code{\score} block at toplevel is interpreted as an implicit
@code{\book} and @code{\notes} mode is the default lexer mode.  The
result is that

@example
\header @{ title = "The Title" @}
@{ a b c @}
@end example

is interpreted as

@example
\header @{ title = "The Title" @}
\book @{
  \score @{
    \notes @{ a b c @}
  @}
@}    
@end example

This is handled by the function defined in
@code{toplevel-music-handler}.  Similarly, @code{\score} and
@code{\book} are handled by @code{toplevel-score-handler} and a
@code{toplevel-book-handler}.  By changing these variables, different
results can be obtained.


@item Start pitch for @code{relative} music is optional for music lists.
The default value is middle C.

@c update-me?  formal definition of newaddlyrics, once it works
@item Combining lyrics with music can be done with @code{\newlyrics}

@example
    \relative @{
        \clef bass
        d2 d c4 bes a2 \break
    @}
    \newlyrics @{
        My first Li -- ly song,
    @}
    \newlyrics @{
        Not much can go wrong!
    @}
@end example

Syntactically, @code{\newlyrics} acts like an infix operator.

@item The parser is encapsulated in a Scheme function
@code{ly:parse-file}, so the following fragment processes two files

@verbatim
#(ly:parse-file "another.ly")
\score { \notes { c4 }}
@end verbatim

@item The @code{font-encoding} property can now be used to select
differently encoded fonts. The previous default setting
@code{TeX-text} has been replaced by latin1. LilyPond requires the
@uref{http://www.lilypond.org/download/fonts/,ec-fonts-mftraced}
package installed to function properly.

@item The encoding of a file can be switched with
@code{\encoding}. The strings for markup texts are translated
automatically.

@item The toplevel block @code{\book} introduces page layout.
A @code{\book} groups @code{\score} blocks into one page layout
entity. For example, two scores are combined in one book with   

@example
\book @{
     \score @{ ... @}
     \score @{ ... @}
@}
@end example

Titling and page breaks are handled by LilyPond.  For automatic page
breaking, two algorithms are available, the classic ragged pages (the
default), and optimal page breaking.  
Page  breaks may be inserted manually with
@example
  \pageBreak
@end example

@item
The @code{lilypond} program does not generate La@TeX{} titles or page
layout.  If you need La@TeX{} titles, you can use the
@code{lilypond-book} program.

@item There is now less of a noticeable ``hook'' at the end of a long slur.

@item The meaning of the @code{|} in the input can be redefined, by
 assigning a music expression to @code{pipeSymbol}.
 
@end itemize

@unnumbered New features in 2.2 since 2.0

@itemize @bullet

@item Setting @code{raggedlast = ##t} in the @code{\paper} block
causes the last line to be set flush-left instead of justified. 

@item The @code{Timing_engraver} now sets the @code{Timing} alias on
its containing context automatically.

@item The code for font selection has been rewritten. In addition to
existing font selection properties, the property @code{font-encoding}
has been added, which makes the switch between normal @code{text} and
other encodings like @code{braces}, @code{music} and @code{math}.

@item The pmx2ly script has been removed from the distribution.

@item Pedal brackets will now run to the last bar of a piece if they are not
explicitly ended.

@item Context definitions now use the word @code{\context} instead of @code{\translator}.

@item Property functions may be used as an argument to @code{set!},
for example

@example
  (set! (ly:grob-property grob 'beam) ... )
@end example

@item In anticipation of Emacs 21.4 or 22.1, the info documentation contains
images.

@item Cue notes can be quoted directly from the parts that
contain them. This will take into account transposition of source and target
instrument. For example,

@verbatim
\addquote clarinet \notes\relative c' {
  \transposition bes
  fis4 fis fis fis
}

\score {
    \notes \relative c'' {
        c8 d8 \quote 2 oboe es8 gis  
    }
}
@end verbatim

@item The transposition of an instrument can be specified using the
@code{\transposition} command.  An
E-flat alto saxophone is specified as 

@example
  \transposition es'
@end example 

@item The naming of exported Scheme functions now follows Scheme conventions.
Changes be applied to Scheme files with 

@example
        convert-ly -e -n --from=2.1.24 --to=2.1.26 *.scm 
@end example


@item Notes can  be excluded from auto-beaming, by  marking them with
@code{\noBeam}
@example
  c8 c \noBeam c c 
@end example

@noindent
will print two separate eighth notes, and two beamed notes.

@item  Translators and contexts have been split. The result of this
internal cleanup is that @code{Score} no longer is the top context;
@code{Score} is contained in the @code{Global} context. Consequently,
it is possible to tweak @code{Score} as follows:

@example
  \context Score \with @{
    @dots{}
  @}
@end example

@item The number of staff lines  in Tablature notation is now
automatically deduced from the @code{stringTunings} property.

@item The program reference has been cleaned up and revised.

@item The syntax for setting properties has been simplified:
the following table lists the differences:

@example
      (old)                           (new)

\property A.B = #C                \set A.B = #C
\property A.B \unset              \unset A.B
\property A.B \set #C = #D        \override A.B #C = #D 
\property A.B \override #C = #D   (removed)
\property A.B \revert #C          \revert A.B #C
@end example

Furthermore, if @code{A} is left out, the bottommost context is used
by default.  In other words, it is no longer necessary to explicitly
mention @code{Voice}, @code{Lyrics} or @code{ChordNames}. 

Old:

@example
   \property Voice.autoBeaming = ##f
   \property Staff.TimeSignature \set #'style = #'C
@end example

New:

@example
   \set autoBeaming = ##f
   \override Staff.TimeSignature #'style = #'C
@end example


@item Tweaks  made with @code{\override} and @code{\revert} no longer
hide tweaks at higher context levels.

@item Melismata in lyrics are also properly handled  in the MIDI output.

@item The lilypond-book script has been rewritten.
It is shorter, cleaner and faster.  The special construct
@code{mbinclude} has been removed, plain @code{@@include} or
@code{\input} can be used now.

It now supports running convert-ly on the lilypond snippets,
@example
    lilypond-book --filter='convert-ly --from=2.0.0' my-book.tely
@end example

@item The @code{LyricsVoice} context has been removed. Lyrics should only
be constructed in @code{Lyrics}.

@item The @code{Thread} context has been removed. Note heads and rests
are now constructed at @code{Voice} level.

@item  Harmonic notes can now be entered as

@example
   <c' g''\harmonic>
@end example

@item Drum notation is now supported  as a regular feature:
percussion may be entered in @code{\drums} mode, and printed or
performed in a @code{DrumStaff} context:

@example
  \score @{
    \drums \new DrumStaff @{ hihat4 cowbell8 @}
  @}
@end example

@item The automatic staff changer was internally rewritten. As a
result, the syntax has been simplified as well:

@example
  \autochange @var{the-music}
@end example

@item The ergonomic syntax of @code{\markup} now has an equivalent in
Scheme. The @code{markup*} macro creates such objects; the following
two markup commands are equivalent:
@example
 f4^#(markup* #:raise 0.2 #:bold "hi")
 f4^\markup @{ \raise #0.2 \bold hi  @}
@end example

@item Voice names, for vocal lines, have been added. They are similar
to instrument names.   They can be set by defining @code{vocalName}
and @code{vocNam}.

@item Safe mode has been reinstated for lilypond.
When lilypond is invoked with @code{--safe-mode}, @TeX{} and
PostScript file output is disallowed, and lilypond-bin is invoked with
@code{--safe-mode}, the user's Guile expressions are evaluated in a
safe environment and file inclusion is not allowed.

Warning: this does not protect against denial-of-service attacks using
Guile, @TeX{} or PostScript.

(This feature is still experimental.)

@item There is now a Scheme macro for defining markup
commands. Special mark-up commands can be defined in user-files too.

@item Many fixes for dimension scaling have been made,
resulting in correct results for scores that mix staves in different
sizes.

@item Improved robustness when  layout properties are accidentally removed.  

@item A  more cleanly constructed part combiner has been installed.
It is more robust and less buggy. The part-combiner can be used with
@example
  \partcombine @var{mus1} @var{mus2}
@end example

@noindent
See @file{input/regression/new-part-combine.ly} for an example.

@item Formatting of rehearsal marks has been improved. The @code{\mark}
command now only does automatic incrementing for marks specified as
integer. For example, @code{\mark #1} will print an A in the default
style.  See @file{input/regression/rehearsal-mark-letter.ly},
@file{input/regression/rehearsal-mark-number.ly}.

@item Formatting of ottava brackets has been much improved.

@item Objects in the output can now be documented: the following fragment
 boxes the note head, and adds the text ``heads or tails?'' three
spaces below the box.

@example
       \context Voice \applyoutput #(add-balloon-text
                                     'NoteHead "heads, or tails?"
                                     '(0 . -3))

      
       c8
@end example


@item Default staff sizes are now scalable. There are two new mechanisms for
setting staff sizes. Both are demonstrated in this fragment:

@example
  #(set-global-staff-size 15)
  \paper @{
    #(paper-set-staff-size (* 15 pt))
  @}   
@end example


@noindent
Both have the same effect on the global layout of a piece. Similarly,
the paper size may be changed as follows

@example
  #(set-default-paper-size "a4")
  \paper @{
    #(set-paper-size "a4")
  @}
@end example 


@item Warnings for bar check errors are more cleverly printed.  This
makes @code{barCheckSynchronize} superfluous, so it is now switched
off by default.

Warning: this will cause problems in scores that use bar checks to
shorten measures.

@item The black note head was made a little rounder, which causes a less
 frantic graphic impression.

@item
A more concise syntax for checking octaves was introduced. A note may
be followed by @code{=}@var{quotes} which indicates what its absolute
octave should be.  In the following example, 

@example
  \relative c'' @{ c='' b=' d,='' @}        
@end example

@noindent
the d will generate a warning, because a d'' is
expected, but a d' is found.
@c @code adds ` ', very confusing.

@item There is a new mechanism for putting lyrics to melodies.
With this mechanism, @code{Lyrics} lines can be put melodies
individually, allowing for different melismatic sections in every
@code{Lyrics}. See @file{input/regression/lyric-combine-new.ly}.

@item Bar lines may now be dotted.


@item The documentation now has links to a wiki, where everyone can
add personal comments to the manual.

@item Contexts may now be changed locally for an isolated music
expression. For example,
  
@example
  \new Voice \with @{
     \consists "Pitch_squash_engraver"
  @} @{
    c'4
  @}
@end example

@item The syntax for changing staffs has changed. The keyword
@code{\change} should now be used, e.g.

@example
  \change Staff = up
@end example 

@item Features of spanner contexts, like @code{Staff}, can now be changed
  using @code{\set}, eg.

@example
  \new Staff @{
       \override Staff.StaffSymbol #'line-count = #4
       c'4
  @} 
@end example

@noindent
puts a quarter note C on a staff with 4 lines.  


@item Multi measure rests are now truly centered between the
clefs/barlines of the staff, their position is independent of symbols
on the other staffs.

@item Collision resolution for dots in chords has been improved greatly. 

@item
Spacing following barlines was improved for widely stretched lines.

@item
Lyric hyphens and extenders now conform to standard typesetting
practice.

@item
Lyrics are now aligned under note heads conforming to engraving
standards. The responsible code has been rewritten, and is drastically
simpler from the previous version. To aid this rewrite, the syntactic
function of the extender line ( __ ) has been changed: it is now
attached to the lyric syllable.

@item
When redefining a context, the associated identifier is also
updated. For example, after reading 

@example
 \translator @{
        \ScoreContext
        autoBeaming = ##f
 @}
@end example

@noindent
the definition of @code{ScoreContext} is updated to include the changed
setting.


@item
The weight of the stafflines is now heavier at smaller staff sizes.
The font has been modified to match this look: at smaller sizes, the
font is heavier and the note heads are more rounded.

@item Processing scores is now done while parsing the file. New
Scheme functions give more flexibility: for example, it is now possible
interpret a score, collecting synchronized musical events in a list, and
manipulate that information using inline Scheme. For an example, see
@file{input/no-notation/recording.ly}.

@item Font sizes can now truly be scaled continuously: the  @code{font-size}
is similar to the old @code{font-relative-size}, but may be set to
fractional values; the closest design size will be scaled to achieve
the desired size. As a side-effect, there are now no longer
limitations in using smaller fonts (eg. grace notes) at small staff
sizes.

@item Stem tips are now also engraved with rounded corners.

@item
The placement of accidentals on chords and ledgered notes is improved.

@end itemize


@unnumbered New features in 2.0 since 1.8

@itemize

@item
Crescendos can now be drawn dotted or dashed.

@item
Quarter tones are now supported. They are entered by suffixing
@code{ih} for a half-sharp and @code{eh} for a half-flat. Hence, the
following is an ascending list of pitches:

@example
  ceses ceseh ces ceh c cih cis cisih cisis
@end example

@item
The following constructs have been removed from the syntax:

@example
  \duration #SCHEME-DURATION
  \pitch #SCHEME-PITCH
  \outputproperty @var{func} @var{symbol} = @var{value}
@end example

For @code{\outputproperty}, the following may be substituted:

@example
   \applyoutput #(outputproperty-compatibility @var{func}
                  @var{symbol} @var{value}) 
@end example 

@item
Clefs may now be transposed arbitrarily, for example

@example
  \clef "G_8"
  \clef "G_15"
  \clef "G_9"
@end example


@item
The syntax for chords and simultaneous music have changed.
Chords are entered as

@example
   <@var{pitches}>
@end example

while simultaneous music is entered as

@example
   <<@var{..music list..}>>
@end example

In effect, the meanings of both have been swapped relative to their 1.8
definition.  The syntax for lists in @code{\markup} has changed
alongside, but figured bass mode was not  changed, i.e.:

@example
  \markup @{ \center <@var{..list of markups..}> @}
  \figure @{ <@var{figures}> @}
@end example

As chords the more often used than simultaneous music, this change will
save keystrokes.

@item
Each music expression can now be tagged, to make different printed
versions from the same music expression.  In the following example,
we see two versions of a piece of music, one for the full score, and
one with cue notes for the instrumental part:

@example
<< \tag #'part <<
     @{ c4 f2 g4 @}      % in the part, we have cue-notes  
     \\ R1 >>
  \tag #'score R1  % in the score: only a rest
>>
@end example
 
The same can be applied to articulations, texts, etc.: they are
made by prepending

@example
        -\tag #@var{your-tags} 
@end example

to an articulation, for example, 

@example
        c4-\tag #'with-fingerings -4 -\tag #'with-strings \6
@end example

This defines a note, which has a conditional fingering and a
string-number indication.

@item
The settings for chord-fingering are more flexible. You can specify a
list where fingerings may be placed, eg.

@example
        \property Voice.fingeringOrientations = #'(left down)
@end example

This will put the fingering for the lowest note below the chord, and the
rest to the left.

@item
The script previously known as @code{ly2dvi} has been renamed to
@code{lilypond}. The binary itself is now installed as
@code{lilypond-bin}.

@item
Markup text (ie. general text formatting) may now be used for lyrics too. 

@item
Two new commands for grace notes have been added, @code{\acciaccatura}
and @code{\appoggiatura},

@example
  \appoggiatura f8 e4
  \acciaccatura g8 f4
@end example

Both reflect the traditional meanings of acciaccatura and appogiatura,
and both insert insert a slur from the first grace note to the main
note.

@item 
Layout options for grace notes are now stored in a context property,
and may now be set separately from musical content.

@item
The @code{\new} command will create a context with a unique
name automatically. Hence, for multi-staff scores, it is no longer
necessary to invent arbitrary context names. For example, a two-staff
score may be created by

@example
  \simultaneous @{
    \new Staff @{ @var{notes for 1st staff} @}
    \new Staff @{ @var{notes for 2nd staff} @}
  @}
@end example



@item
Octave checks make octave errors easier to correct.
The syntax is 

@example
  \octave @var{pitch}
@end example

This checks that @var{pitch} (without octave) yields @var{pitch} (with
octave) in \relative mode. If not, a warning is printed, and the
octave is corrected.

@item
All articulations must now be entered postfix. For example,

@example
        c8[( d8]) 
@end example

@noindent
is a pair of beamed slurred eighth notes.

@item
The definition of @code{\relative} has been simplified.  Octaves are
now always propagated in the order that music is entered. In the
following example,  

@example
  PRE
  \repeat "unfold" 3  BODY \alternative @{ ALT1 ALT2 @}
  POST
@end example

@noindent
the octave of BODY is based on PRE, the starting octave of ALT1 on
BODY, the starting octave of ALT2 on ALT1, and the starting octave of
POST on ALT2.

The same mechanism is used for all other music expressions, except the
chord. Backwards compatibility is retained through a special program option,
which is set through 

@example
  #(ly:set-option 'old-relative)
@end example

@item
Windows users can double click a @code{.ly} file to process and view
it automagically through the new @code{lily-wins} frontend.

@end itemize




@unnumbered New features in 1.8 since 1.6

@itemize

@item 
The chord entry code has been completely rewritten. It is now
cleaner and more flexible.

@item 
A new syntax has been added for text entry.  This syntax is more
friendly than the old mechanism, and it is implemented in a more
robust and modular way. For more information, refer to the section on
"Text markup" in the notation manual.

@item 
The integration of the input language and Scheme has been made deeper:
you can now use LilyPond identifiers in Scheme, and use Scheme
expressions instead of LilyPond identifiers.

@item 
The internal representation of music has been cleaned up completely
and converted to Scheme data structures.  The representation may be
exported as XML.

@item 
A new uniform postfix syntax for articulation has been introduced.
A beamed slurred pair of eighth notes can be entered as

@example
        c8-[-( d8-]-) 
@end example

In version 2.0, postfix syntax will be the only syntax
available, and the dashes will become optional.

This will simplify the language: all articulations can be entered as
postfix, in any order.

@item 
A new syntax has been added for chords:


@example
        << PITCHES >>
@end example 

It is not necessary to update files to this syntax, but it will be for
using LilyPond version 2.0.  In version 2.0, this syntax will be
changed to

@example
        < PITCHES >
@end example

for chords, and

@example
        \simultaneous @{ .. @} 
@end example

for simultaneous music.

To convert your files from <PITCHES> to <<PITCHES>>, use the script
included in @file{buildscripts/convert-new-chords.py}.

This change was introduced for the following reasons

@itemize @bullet
@item
It solves the "start score with chord" problem, where you have to
  state \context Voice explicitly when a chord was the start of a
  Staff or Score.
@item
With the new syntax, it is possible to distinguish between
  articulations (or fingerings) which are for a single chord note,
  and which are for the entire chord. This allows for per-note
  fingerings, and is more logical on the whole.
@end itemize

@item 
User code may now be executed during interpreting.  The syntax for
this code is

@example
        \applycontext #SCHEME-FUNCTION
@end example

@item 
User code may now be executed on arbitrary grobs during interpreting.
The syntax for this feature is

@example
        \applyoutput #SCHEME-FUNCTION
@end example

@noindent
SCHEME-FUNCTION takes a single argument, and is called for every grob
that is created in the current context.

@item 
New algorithms for chord-name formatting have been installed. They
can be tuned and have ergonomic syntax for entering exceptions.

@item 
Texts may now be put on multimeasure rests, e.g.

@example
        R1*20^\markup @{ "GP" @}
@end example

@item
Ancient notation now prints ligatures in Gregorian square neumes
notation, roughly following the typographical style of the Liber
hymnarius of Solesmes, published in 1983.  Ligatures are still printed
without the proper line breaking and horizontal spacing.

@item 
Glissandi can now be printed using the zigzag style.

@item 
LilyPond can now print clusters. The syntax is

@example
        \apply #notes-to-clusters @{ NOTE NOTE .. @}
@end example

@item
For irregular meters, beat grouping marks can be printed. The
syntax for this is

@example
        #(set-time-signature 7 8 '(3 2 2))
@end example


@item 
Nested horizontal brackets for music analysis can now be printed:

@example
        NOTE-\startGroup
                ..
        NOTE-\stopGroup
@end example


@item  Ottava brackets are now fully supported as a feature.  The syntax
is

@example
        #(set-octavation 1)
@end example


@item  Metronome markings are printed when a \tempo command is processed.



@item Fingerings can be put on chords horizontally.



@item  The appearance of various glyphs has been fine-tuned.



@item  Different types of percent style repeats may now be nested.



@item  The emacs support has been extended.


@item 
The manual has been completely revised and extended.

@end itemize

@unnumbered New features in 1.6 since 1.4


@itemize @bullet

@item
Support for figured bass and tablature.

@item
Completely rewritten beam formatting: provides much better output
now.


@item
Completely revised and improved music font.


@item
Completely rewritten MIDI import support.

@item
Completely rewritten grace note support. Practically speaking this
means that grace notes can be slurred to normal normal notes.


@item
Improved accidental handling and formatting: styles for producing
cautionaries may vary, and complex collisions between accidentals of a
chord are handled much better.

@item
Better spacing: both globally and locally. This includes subtle
details like optical stem spacing.

@item
More support for ancient notation: mensural ligatures, ambitus
(pitch range) of voices, more shapes, etc.

@item
More support for piano notation: bracket pedals, directed arpeggios,
arpeggio brackets.

@item
Easier music polyphonic music entry.

@item
More extensibility, many speedups and bugfixes

@item
The manual has been thoroughly revised.

@item
Development is now hosted at http://savannah.gnu.org, and sources
can be downloaded through anonymous CVS.

@item
Support for windows: LilyPond is part of the cygwin distribution,
which comes with a user-friendly installer.

@end itemize