]> git.donarmstrong.com Git - lilypond.git/blobdiff - lily/lexer.ll
Merge branch 'master' of git://git.savannah.gnu.org/lilypond.git
[lilypond.git] / lily / lexer.ll
index 178ea88e97eb9d070e22359aa0458f2487158fab..83a7940b6938dc634083fe9936ee389635713412 100644 (file)
@@ -1,13 +1,23 @@
 %{ // -*-Fundamental-*-
 /*
-  lexer.ll -- implement the Flex lexer
+  This file is part of LilyPond, the GNU music typesetter.
 
-  source file of the LilyPond music typesetter
+  Copyright (C) 1996--2011 Han-Wen Nienhuys <hanwen@xs4all.nl>
+                 Jan Nieuwenhuizen <janneke@gnu.org>
 
-  (c) 1996--2006 Han-Wen Nienhuys <hanwen@xs4all.nl>
-           Jan Nieuwenhuizen <janneke@gnu.org>
-*/
+  LilyPond is free software: you can redistribute it and/or modify
+  it under the terms of the GNU General Public License as published by
+  the Free Software Foundation, either version 3 of the License, or
+  (at your option) any later version.
+
+  LilyPond is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+  GNU General Public License for more details.
 
+  You should have received a copy of the GNU General Public License
+  along with LilyPond.  If not, see <http://www.gnu.org/licenses/>.
+*/
 
 /*
   backup rules
 using namespace std;
 
 #include "context-def.hh"
+#include "duration.hh"
 #include "identifier-smob.hh"
 #include "international.hh"
 #include "interval.hh"
 #include "lily-guile.hh"
 #include "lily-lexer.hh"
-#include "lilypond-input-version.hh"
+#include "lily-parser.hh"
+#include "lilypond-version.hh"
 #include "main.hh"
+#include "music.hh"
 #include "music-function.hh"
 #include "parse-scm.hh"
 #include "parser.hh"
+#include "pitch.hh"
 #include "source-file.hh"
 #include "std-string.hh"
 #include "string-convert.hh"
@@ -60,8 +74,8 @@ RH 7 fix (?)
 void strip_trailing_white (string&);
 void strip_leading_white (string&);
 string lyric_fudge (string s);
-int music_function_type (SCM);
 SCM lookup_markup_command (string s);
+SCM lookup_markup_list_command (string s);
 bool is_valid_version (string s);
 
 
@@ -74,17 +88,12 @@ bool is_valid_version (string s);
        yylval.string = new string
 
 #define yylval \
-       (*(YYSTYPE*)lexval)
+       (*(YYSTYPE*)lexval_)
 
 #define yylloc \
-       (*(YYLTYPE*)lexloc)
+       (*(YYLTYPE*)lexloc_)
 
 #define YY_USER_ACTION add_lexed_char (YYLeng ());
-/*
-
-LYRICS         ({AA}|{TEX})[^0-9 \t\n\f]*
-
-*/
 
 
 SCM scan_fraction (string);
@@ -103,6 +112,7 @@ SCM (* scm_parse_error_handler) (void *);
 %option never-interactive 
 %option warn
 
+%x extratoken
 %x chords
 %x figures
 %x incl
@@ -116,17 +126,21 @@ SCM (* scm_parse_error_handler) (void *);
 %x sourcefilename
 %x version
 
-A              [a-zA-Z]
+A              [a-zA-Z\200-\377]
 AA             {A}|_
 N              [0-9]
 AN             {AA}|{N}
+ANY_CHAR       (.|\n)
 PUNCT          [?!:'`]
 ACCENT         \\[`'"^]
-NATIONAL       [\001-\006\021-\027\031\036\200-\377]
+NATIONAL       [\001-\006\021-\027\031\036]
 TEX            {AA}|-|{PUNCT}|{ACCENT}|{NATIONAL}
 WORD           {A}{AN}*
 DASHED_WORD            {A}({AN}|-)*
 DASHED_KEY_WORD                \\{DASHED_WORD}
+
+
+
 ALPHAWORD      {A}+
 DIGIT          {N}
 UNSIGNED       {N}+
@@ -146,6 +160,7 @@ ESCAPED             [nt\\'"]
 EXTENDER       __
 HYPHEN         --
 BOM_UTF8       \357\273\277
+
 %%
 
 
@@ -153,8 +168,26 @@ BOM_UTF8   \357\273\277
        // windows-suck-suck-suck
 }
 
-<INITIAL,chords,lyrics,figures,notes>{BOM_UTF8} {
-  if (this->lexloc->line_number () != 1 || this->lexloc->column_number () != 0)
+<extratoken>{ANY_CHAR} {
+  /* Generate a token without swallowing anything */
+
+  /* First unswallow the eaten character */
+  add_lexed_char (-YYLeng ());
+  yyless (0);
+
+  /* produce requested token */
+  int type = extra_token_types_.back ();
+  extra_token_types_.pop_back ();
+  if (extra_token_types_.empty ())
+    yy_pop_state ();
+
+  return type;
+}
+
+   /* Use the trailing context feature. Otherwise, the BOM will not be
+      found if the file starts with an identifier definition. */
+<INITIAL,chords,lyrics,figures,notes>{BOM_UTF8}/.* {
+  if (this->lexloc_->line_number () != 1 || this->lexloc_->column_number () != 0)
     {
       LexerError (_ ("stray UTF-8 BOM encountered").c_str ());
       exit (1);
@@ -167,13 +200,13 @@ BOM_UTF8  \357\273\277
   "%{" {
        yy_push_state (longcomment);
   }
-  %[^{\n\r].*[\n\r]    {
+  %[^{\n\r][^\n\r]*[\n\r]      {
   }
   %[^{\n\r]    { // backup rule
   }
   %[\n\r]      {
   }
-  %[^{\n\r].*  {
+  %[^{\n\r][^\n\r]*    {
   }
   {WHITE}+     {
 
@@ -219,7 +252,7 @@ BOM_UTF8    \357\273\277
        progress_indication ("\n");
        scm_module_define (scm_car (scopes_),
                     ly_symbol2scm ("input-file-name"),
-                    scm_makfrom0str (s.c_str ()));
+                    ly_string2scm (s));
 
 }
 
@@ -227,7 +260,6 @@ BOM_UTF8    \357\273\277
        int i;
        sscanf (YYText (), "%d", &i);
 
-//     this->set_debug (1); 
        yy_pop_state ();
        this->here_input ().get_source_file ()->set_line (here_input ().start (), i);
 }
@@ -255,7 +287,7 @@ BOM_UTF8    \357\273\277
        }
        <<EOF>>         {
                LexerError (_ ("EOF found inside a comment").c_str ());
-               is_main_input_ = false;
+               is_main_input_ = false; // should be safe , can't have \include in --safe.
                if (! close_input ()) 
                  yyterminate (); // can't move this, since it actually rets a YY_NULL
        }
@@ -307,7 +339,7 @@ BOM_UTF8    \357\273\277
 }
 <chords,notes,figures>{RESTNAME}       {
        char const *s = YYText ();
-       yylval.scm = scm_makfrom0str (s);
+       yylval.scm = scm_from_locale_string (s);
        return RESTNAME;
 }
 <chords,notes,figures>R                {
@@ -318,7 +350,7 @@ BOM_UTF8    \357\273\277
        Input hi = here_input();
        hi.step_forward ();
        SCM sval = ly_parse_scm (hi.start (), &n, hi,
-               be_safe_global && is_main_input_);
+               be_safe_global && is_main_input_, parser_);
 
        if (sval == SCM_UNDEFINED)
        {
@@ -337,6 +369,11 @@ BOM_UTF8   \357\273\277
                yylval.scm = unpack_identifier(sval);
                return identifier_type (yylval.scm);
        }
+
+       for (size_t i = 0; i < pending_string_includes_.size (); i++)
+               new_input ("<included string>", pending_string_includes_[i],
+                          parser_->sources_);
+       pending_string_includes_.clear ();
                
        yylval.scm = sval;
        return SCM_TOKEN;
@@ -402,7 +439,7 @@ BOM_UTF8    \357\273\277
        \\{ESCAPED}     {
                *yylval.string += to_string (escaped_char (YYText ()[1]));
        }
-       [^\\"]+ {
+       [^\\""]+        {
                *yylval.string += YYText ();
        }
        \"      {
@@ -411,7 +448,7 @@ BOM_UTF8    \357\273\277
 
                /* yylval is union. Must remember STRING before setting SCM*/
                string *sp = yylval.string;
-               yylval.scm = scm_makfrom0str (sp->c_str ());
+               yylval.scm = ly_string2scm (*sp);
                delete sp;
                return is_lyric_state () ? LYRICS_STRING : STRING;
        }
@@ -448,7 +485,7 @@ BOM_UTF8    \357\273\277
                if (c == '{' ||  c == '}') // brace open is for not confusing dumb tools.
                        here_input ().warning (
                                _ ("Brace found at end of lyric.  Did you forget a space?"));
-               yylval.scm = scm_makfrom0str (s.c_str ());
+               yylval.scm = ly_string2scm (s);
 
 
                return LYRICS_STRING;
@@ -499,36 +536,53 @@ BOM_UTF8  \357\273\277
        }
        {MARKUPCOMMAND} {
                string str (YYText () + 1);
+
+                int token_type = MARKUP_FUNCTION;
                SCM s = lookup_markup_command (str);
 
-               if (scm_is_pair (s) && scm_is_symbol (scm_cdr (s)) ) {
-                       yylval.scm = scm_car(s);
-                       SCM tag = scm_cdr(s);
-                       if (tag == ly_symbol2scm("markup0"))
-                               return MARKUP_HEAD_MARKUP0;
-                       if (tag == ly_symbol2scm("empty"))
-                               return MARKUP_HEAD_EMPTY;
-                       else if (tag == ly_symbol2scm ("markup0-markup1"))
-                               return MARKUP_HEAD_MARKUP0_MARKUP1;
-                       else if (tag == ly_symbol2scm ("markup-list0"))
-                               return MARKUP_HEAD_LIST0;
-                       else if (tag == ly_symbol2scm ("scheme0"))
-                               return MARKUP_HEAD_SCM0;
-                       else if (tag == ly_symbol2scm ("scheme0-scheme1"))
-                               return MARKUP_HEAD_SCM0_SCM1;
-                       else if (tag == ly_symbol2scm ("scheme0-markup1"))
-                               return MARKUP_HEAD_SCM0_MARKUP1;
-                       else if (tag == ly_symbol2scm ("scheme0-scheme1-markup2"))
-                               return MARKUP_HEAD_SCM0_SCM1_MARKUP2;
-                       else if (tag == ly_symbol2scm ("scheme0-scheme1-scheme2"))
-                               return MARKUP_HEAD_SCM0_SCM1_SCM2;
-                       else {
-                               programming_error ("no parser tag defined for this markup signature"); 
-                               ly_display_scm (s);
-                               assert(false);
-                       }
-               } else
-                       return scan_escaped_word (str);
+               // lookup-markup-command returns a pair with the car
+               // being the function to call, and the cdr being the
+               // call signature specified to define-markup-command,
+               // a list of predicates.
+
+                if (!scm_is_pair (s)) {
+                 // If lookup-markup-command was not successful, we
+                 // try lookup-markup-list-command instead.
+                 // If this fails as well, we just scan and return
+                 // the escaped word.
+                 s = lookup_markup_list_command (str);
+                 if (scm_is_pair (s))
+                   token_type = MARKUP_LIST_FUNCTION;
+                 else
+                   return scan_escaped_word (str);
+                }
+
+               // If the list of predicates is, say,
+               // (number? number? markup?), then tokens
+               // EXPECT_MARKUP EXPECT_SCM EXPECT_SCM EXPECT_NO_MORE_ARGS
+               // will be generated.  Note that we have to push them
+               // in reverse order, so the first token pushed in the
+               // loop will be EXPECT_NO_MORE_ARGS.
+
+               yylval.scm = scm_car(s);
+
+               // yylval now contains the function to call as token
+               // value (for token type MARKUP_FUNCTION or
+               // MARKUP_LIST_FUNCTION).
+
+               push_extra_token(EXPECT_NO_MORE_ARGS);
+               s = scm_cdr(s);
+               for (; scm_is_pair(s); s = scm_cdr(s)) {
+                 SCM predicate = scm_car(s);
+
+                 if (predicate == ly_lily_module_constant ("markup-list?"))
+                   push_extra_token(EXPECT_MARKUP_LIST);
+                 else if (predicate == ly_lily_module_constant ("markup?"))
+                   push_extra_token(EXPECT_MARKUP);
+                 else
+                   push_extra_token(EXPECT_SCM);
+               }
+               return token_type;
        }
        [{}]    {
                return YYText ()[0];
@@ -541,7 +595,7 @@ BOM_UTF8    \357\273\277
                if (c == '{' ||  c == '}')
                        here_input ().warning (
                                _ ("Brace found at end of markup.  Did you forget a space?"));
-               yylval.scm = scm_makfrom0str (s.c_str ());
+               yylval.scm = ly_string2scm (s);
 
 
                return STRING;
@@ -554,7 +608,11 @@ BOM_UTF8   \357\273\277
 <*><<EOF>> {
        if (is_main_input_)
        {
-               is_main_input_ = false;
+               /* 2 = init.ly + current file.
+                  > because we're before closing, but is_main_input_ should
+                  reflect after.
+               */ 
+               is_main_input_ = include_stack_.size () > 2;
                if (!close_input ())
                /* Returns YY_NULL */
                        yyterminate ();
@@ -647,6 +705,20 @@ BOM_UTF8   \357\273\277
 
 %%
 
+/* Make the lexer generate a token of the given type as the next token. 
+ TODO: make it possible to define a value for the token as well */
+void
+Lily_lexer::push_extra_token (int token_type)
+{
+       if (extra_token_types_.empty ())
+       {
+               if (YY_START != extratoken)
+                       hidden_state_ = YY_START;
+               yy_push_state (extratoken);
+       }
+       extra_token_types_.push_back (token_type);
+}
+
 void
 Lily_lexer::push_chord_state (SCM tab)
 {
@@ -690,6 +762,7 @@ Lily_lexer::pop_state ()
 {
        if (YYSTATE == notes || YYSTATE == chords)
                pitchname_tab_stack_ = scm_cdr (pitchname_tab_stack_);
+
        yy_pop_state ();
 }
 
@@ -718,7 +791,26 @@ Lily_lexer::scan_escaped_word (string str)
        if (is_music_function (sid))
        {
                yylval.scm = get_music_function_transform (sid);
-               return music_function_type (yylval.scm);
+
+               SCM s = scm_object_property (yylval.scm, ly_symbol2scm ("music-function-signature"));
+               push_extra_token (EXPECT_NO_MORE_ARGS);
+               for (; scm_is_pair (s); s = scm_cdr (s))
+               {
+                       SCM cs = scm_car (s);
+                       
+                       if (cs == ly_music_p_proc)
+                               push_extra_token (EXPECT_MUSIC);
+                       else if (cs == Pitch_type_p_proc)
+                               push_extra_token (EXPECT_PITCH);
+                       else if (cs == Duration_type_p_proc)
+                               push_extra_token (EXPECT_DURATION);
+                       else if (cs == ly_lily_module_constant ("markup?"))
+                               push_extra_token (EXPECT_MARKUP);
+                       else if (ly_is_procedure (cs))
+                               push_extra_token (EXPECT_SCM);
+                       else programming_error ("Function parameter without type-checking predicate");
+               }
+               return MUSIC_FUNCTION;
        }
 
        if (sid != SCM_UNDEFINED)
@@ -730,7 +822,7 @@ Lily_lexer::scan_escaped_word (string str)
        string msg (_f ("unknown escaped string: `\\%s'", str));        
        LexerError (msg.c_str ());
 
-       yylval.scm = scm_makfrom0str (str.c_str ());
+       yylval.scm = ly_string2scm (str);
 
        return STRING;
 }
@@ -751,39 +843,52 @@ Lily_lexer::scan_bare_word (string str)
                        else if (scm_is_symbol (yylval.scm))
                            return DRUM_PITCH;
                }
-               else if ((handle = scm_hashq_get_handle (chordmodifier_tab_, sym))!= SCM_BOOL_F)
+               else if ((YYSTATE == chords)
+                       && (handle = scm_hashq_get_handle (chordmodifier_tab_, sym))!= SCM_BOOL_F)
                {
                    yylval.scm = scm_cdr (handle);
                    return CHORD_MODIFIER;
                }
+               if ((chord_repetition_.repetition_symbol_ != SCM_EOL)
+                   && to_boolean (scm_equal_p (chord_repetition_.repetition_symbol_, sym)))
+                       return CHORD_REPETITION;
        }
 
-       yylval.scm = scm_makfrom0str (str.c_str ());
+       yylval.scm = ly_string2scm (str);
        return STRING;
 }
 
+int
+Lily_lexer::get_state () const
+{
+       if (YY_START == extratoken)
+               return hidden_state_;
+       else
+               return YY_START;
+}
+
 bool
 Lily_lexer::is_note_state () const
 {
-       return YY_START == notes;
+       return get_state () == notes;
 }
 
 bool
 Lily_lexer::is_chord_state () const
 {
-       return YY_START == chords;
+       return get_state () == chords;
 }
 
 bool
 Lily_lexer::is_lyric_state () const
 {
-       return YY_START == lyrics;
+       return get_state () == lyrics;
 }
 
 bool
 Lily_lexer::is_figure_state () const
 {
-       return YY_START == figures;
+       return get_state () == figures;
 }
 
 /*
@@ -822,12 +927,18 @@ is_valid_version (string s)
 {
   Lilypond_version current ( MAJOR_VERSION "." MINOR_VERSION "." PATCH_LEVEL );
   Lilypond_version ver (s);
-  if (! ((ver >= oldest_version) && (ver <= current)))
+  if (int (ver) < oldest_version)
        {       
-               non_fatal_error (_f ("Incorrect lilypond version: %s (%s, %s)", ver.to_string (), oldest_version.to_string (), current.to_string ()));
-               non_fatal_error (_ ("Consider updating the input with the convert-ly script")); 
+               non_fatal_error (_f ("file too old: %s (oldest supported: %s)", ver.to_string (), oldest_version.to_string ()));
+               non_fatal_error (_ ("consider updating the input with the convert-ly script"));
                return false;
-    }
+       }
+
+  if (ver > current)
+       {
+               non_fatal_error (_f ("program too old: %s (file requires: %s)",  current.to_string (), ver.to_string ()));
+               return false;
+       }
   return true;
 }
        
@@ -878,60 +989,14 @@ SCM
 lookup_markup_command (string s)
 {
        SCM proc = ly_lily_module_constant ("lookup-markup-command");
-       return scm_call_1 (proc, scm_makfrom0str (s.c_str ()));
-}
-
-struct Parser_signature
-{
-       char *symbol;
-       int token_type;
-};
-static SCM signature_hash_table;
-
-static void init_signature_hash_table ()
-{
-       signature_hash_table = scm_gc_protect_object (scm_c_make_hash_table (31));
-       Parser_signature sigs[]  = {
-               {"scm", MUSIC_FUNCTION_SCM},
-               {"music", MUSIC_FUNCTION_MUSIC},
-               {"scm-music", MUSIC_FUNCTION_SCM_MUSIC},
-               {"scm-scm", MUSIC_FUNCTION_SCM_SCM},
-               {"music-music", MUSIC_FUNCTION_MUSIC_MUSIC},
-               {"scm-music-music", MUSIC_FUNCTION_SCM_MUSIC_MUSIC},
-               {"scm-scm-music-music", MUSIC_FUNCTION_SCM_SCM_MUSIC_MUSIC},
-               {"scm-scm-music", MUSIC_FUNCTION_SCM_SCM_MUSIC},
-               {"scm-scm-scm-music", MUSIC_FUNCTION_SCM_SCM_SCM_SCM_MUSIC},
-               {"scm-scm-scm-scm-music", MUSIC_FUNCTION_SCM_SCM_SCM_MUSIC},
-               {"scm-scm-scm", MUSIC_FUNCTION_SCM_SCM_SCM},
-               {"markup", MUSIC_FUNCTION_MARKUP},
-               {"markup-music", MUSIC_FUNCTION_MARKUP_MUSIC},
-               {"markup-markup", MUSIC_FUNCTION_MARKUP_MARKUP},
-               {"markup-music-music", MUSIC_FUNCTION_MARKUP_MUSIC_MUSIC},
-               {"markup-markup-music", MUSIC_FUNCTION_MARKUP_MARKUP_MUSIC},
-               {"noarg", MUSIC_FUNCTION},
-               {0,0}
-       };
-
-       for (int i = 0; sigs[i].symbol; i++)
-               scm_hashq_set_x (signature_hash_table, ly_symbol2scm (sigs[i].symbol),
-                                scm_from_int (sigs[i].token_type));
+       return scm_call_1 (proc, ly_string2scm (s));
 }
 
-int
-music_function_type (SCM func)
+SCM
+lookup_markup_list_command (string s)
 {
-       if (!signature_hash_table)
-               init_signature_hash_table ();
-
-       SCM type = scm_object_property (func, ly_symbol2scm ("music-function-signature-keyword"));
-       SCM token_type = scm_hashq_ref (signature_hash_table, type, SCM_BOOL_F);
-       if (!scm_is_number (token_type))
-               {
-               programming_error (_ ("can't find signature for music function"));
-               return MUSIC_FUNCTION_SCM;
-               }
-       
-       return scm_to_int (token_type);
+       SCM proc = ly_lily_module_constant ("lookup-markup-list-command");
+       return scm_call_1 (proc, ly_string2scm (s));
 }
 
 /* Shut up lexer warnings.  */