X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=lily%2Flexer.ll;h=0d3f7599d899a09aab95f6800960c4e790ae6535;hb=44b7603963e7b9d49bc53c53d584c7fad46f31da;hp=98b6e0f3d7f978e83f8cc75aca42c71b87730e65;hpb=bc95f4434f760d41191341ab4508b2064eb19025;p=lilypond.git diff --git a/lily/lexer.ll b/lily/lexer.ll index 98b6e0f3d7..0d3f7599d8 100644 --- a/lily/lexer.ll +++ b/lily/lexer.ll @@ -1,8 +1,8 @@ -%{ // -*-Fundamental-*- +%{ // -*- mode: c++; c-file-style: "linux" -*- /* This file is part of LilyPond, the GNU music typesetter. - Copyright (C) 1996--2010 Han-Wen Nienhuys + Copyright (C) 1996--2012 Han-Wen Nienhuys Jan Nieuwenhuizen LilyPond is free software: you can redistribute it and/or modify @@ -19,6 +19,14 @@ along with LilyPond. If not, see . */ +/* Mode and indentation are at best a rough approximation based on TAB + * formatting (reasonable for compatibility with unspecific editor + * modes as Flex modes are hard to find) and need manual correction + * frequently. Without a reasonably dependable way of formatting a + * Flex file sensibly, there is little point in trying to fix the + * inconsistent state of indentation. + */ + /* backup rules @@ -46,17 +54,19 @@ using namespace std; #include "context-def.hh" -#include "identifier-smob.hh" +#include "duration.hh" #include "international.hh" #include "interval.hh" #include "lily-guile.hh" #include "lily-lexer.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" @@ -76,19 +86,19 @@ SCM lookup_markup_list_command (string s); bool is_valid_version (string s); -#define start_quote() \ - yy_push_state (quote);\ - yylval.string = new string +#define start_quote() do { \ + yy_push_state (quote); \ + yylval = SCM_EOL; \ + } while (0) -#define start_lyric_quote() \ - yy_push_state (lyric_quote);\ - yylval.string = new string +#define start_lyric_quote() do { \ + yy_push_state (lyric_quote); \ + yylval = SCM_EOL; \ + } while (0) -#define yylval \ - (*(YYSTYPE*)lexval) +#define yylval (*lexval_) -#define yylloc \ - (*(YYLTYPE*)lexloc) +#define yylloc (*lexloc_) #define YY_USER_ACTION add_lexed_char (YYLeng ()); @@ -123,35 +133,42 @@ SCM (* scm_parse_error_handler) (void *); %x sourcefilename %x version +/* The strategy concerning multibyte characters is to accept them but + * call YYText_utf8 for patterns that might contain them, in order to + * get a single code path responsible for flagging non-UTF-8 input: + * Patterns for accepting only valid UTF-8 without backing up are + * really hard to do and complex, and if nice error messages are + * wanted, one would need patterns catching the invalid input as well. + * + * Since editors and operating environments don't necessarily behave + * reasonably in the presence of mixed encodings, we flag encoding + * errors also in identifiers, comments, and strings where it would be + * conceivable to just transparently work with the byte string. But + * the whole point of caring about UTF-8 in here at all is too avoid + * stranger errors later when input passes into backends or log files + * or console output or error messages. + */ + A [a-zA-Z\200-\377] AA {A}|_ N [0-9] -AN {AA}|{N} ANY_CHAR (.|\n) -PUNCT [?!:'`] -ACCENT \\[`'"^] +PUNCT [][()?!:'`] +SPECIAL_CHAR [&@] NATIONAL [\001-\006\021-\027\031\036] -TEX {AA}|-|{PUNCT}|{ACCENT}|{NATIONAL} -WORD {A}{AN}* -DASHED_WORD {A}({AN}|-)* -DASHED_KEY_WORD \\{DASHED_WORD} - - +TEX {AA}|-|{PUNCT}|{NATIONAL}|{SPECIAL_CHAR} +WORD {A}([-_]{A}|{A})* +COMMAND \\{WORD} -ALPHAWORD {A}+ -DIGIT {N} UNSIGNED {N}+ E_UNSIGNED \\{N}+ FRACTION {N}+\/{N}+ INT -?{UNSIGNED} REAL ({INT}\.{N}*)|(-?\.{N}+) -KEYWORD \\{WORD} WHITE [ \n\t\f\r] HORIZONTALWHITE [ \t] BLACK [^ \n\t\f\r] RESTNAME [rs] -NOTECOMMAND \\{A}+ -MARKUPCOMMAND \\({A}|[-_])+ LYRICS ({AA}|{TEX})[^0-9 \t\n\r\f]* ESCAPED [nt\\'"] EXTENDER __ @@ -162,7 +179,7 @@ BOM_UTF8 \357\273\277 <*>\r { - // windows-suck-suck-suck + // swallow and ignore carriage returns } {ANY_CHAR} { @@ -173,9 +190,23 @@ BOM_UTF8 \357\273\277 yyless (0); /* produce requested token */ - int type = extra_token_types_.back (); - extra_token_types_.pop_back (); - if (extra_token_types_.empty ()) + int type = scm_to_int (scm_caar (extra_tokens_)); + yylval = scm_cdar (extra_tokens_); + extra_tokens_ = scm_cdr (extra_tokens_); + if (scm_is_null (extra_tokens_)) + yy_pop_state (); + + return type; +} + +<> { + /* Generate a token without swallowing anything */ + + /* produce requested token */ + int type = scm_to_int (scm_caar (extra_tokens_)); + yylval = scm_cdar (extra_tokens_); + extra_tokens_ = scm_cdr (extra_tokens_); + if (scm_is_null (extra_tokens_)) yy_pop_state (); return type; @@ -184,26 +215,28 @@ BOM_UTF8 \357\273\277 /* Use the trailing context feature. Otherwise, the BOM will not be found if the file starts with an identifier definition. */ {BOM_UTF8}/.* { - if (this->lexloc->line_number () != 1 || this->lexloc->column_number () != 0) + if (this->lexloc_->line_number () != 1 || this->lexloc_->column_number () != 0) { - LexerError (_ ("stray UTF-8 BOM encountered").c_str ()); - exit (1); + LexerWarning (_ ("stray UTF-8 BOM encountered").c_str ()); + // exit (1); } - if (be_verbose_global) - message (_ ("Skipping UTF-8 BOM")); + debug_output (_ ("Skipping UTF-8 BOM")); } { "%{" { yy_push_state (longcomment); } - %[^{\n\r].*[\n\r] { + %[^{\n\r][^\n\r]*[\n\r] { + (void) YYText_utf8 (); } %[^{\n\r] { // backup rule + (void) YYText_utf8 (); } %[\n\r] { } - %[^{\n\r].* { + %[^{\n\r][^\n\r]* { + (void) YYText_utf8 (); } {WHITE}+ { @@ -226,7 +259,7 @@ BOM_UTF8 \357\273\277 yy_push_state (sourcefileline); } \"[^"]*\" { /* got the version number */ - string s (YYText () + 1); + string s (YYText_utf8 () + 1); s = s.substr (0, s.rfind ('\"')); yy_pop_state (); @@ -234,13 +267,13 @@ BOM_UTF8 \357\273\277 SCM top_scope = scm_car (scm_last_pair (scopes_)); scm_module_define (top_scope, ly_symbol2scm ("version-seen"), SCM_BOOL_T); - if (!is_valid_version (s)) + if (!is_valid_version (s)) { + yylval = SCM_UNSPECIFIED; return INVALID; - - + } } -\"[^"]*\" { - string s (YYText () + 1); +\"[^""]*\" { + string s (YYText_utf8 () + 1); s = s.substr (0, s.rfind ('\"')); yy_pop_state (); @@ -261,33 +294,28 @@ BOM_UTF8 \357\273\277 this->here_input ().get_source_file ()->set_line (here_input ().start (), i); } -. { +{ANY_CHAR} { LexerError (_ ("quoted string expected after \\version").c_str ()); yy_pop_state (); } -. { +{ANY_CHAR} { LexerError (_ ("quoted string expected after \\sourcefilename").c_str ()); yy_pop_state (); } -. { +{ANY_CHAR} { LexerError (_ ("integer expected after \\sourcefileline").c_str ()); yy_pop_state (); } { [^\%]* { + (void) YYText_utf8 (); } \%*[^}%]* { - + (void) YYText_utf8 (); } "%"+"}" { yy_pop_state (); } - <> { - LexerError (_ ("EOF found inside a comment").c_str ()); - 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 - } } @@ -304,15 +332,15 @@ BOM_UTF8 \357\273\277 \\include { yy_push_state (incl); } -\"[^"]*\" { /* got the include file name */ - string s (YYText ()+1); +\"[^""]*\" { /* got the include file name */ + string s (YYText_utf8 ()+1); s = s.substr (0, s.rfind ('"')); new_input (s, sources_); yy_pop_state (); } -\\{BLACK}*{WHITE} { /* got the include identifier */ - string s = YYText () + 1; +\\{BLACK}*{WHITE}? { /* got the include identifier */ + string s = YYText_utf8 () + 1; strip_trailing_white (s); if (s.length () && (s[s.length () - 1] == ';')) s = s.substr (0, s.length () - 1); @@ -321,7 +349,7 @@ BOM_UTF8 \357\273\277 if (scm_is_string (sid)) { new_input (ly_scm2string (sid), sources_); yy_pop_state (); - } else { + } else { string msg (_f ("wrong or undefined identifier: `%s'", s )); LexerError (msg.c_str ()); @@ -330,16 +358,59 @@ BOM_UTF8 \357\273\277 scm_display (sid, err); } } -\"[^"]* { // backup rule +(\$|#) { // scm for the filename + int n = 0; + Input hi = here_input(); + hi.step_forward (); + SCM sval = ly_parse_scm (hi.start (), &n, hi, + be_safe_global && is_main_input_, parser_); + sval = eval_scm (sval); + + for (int i = 0; i < n; i++) + { + yyinput (); + } + char_count_stack_.back () += n; + + if (scm_is_string (sval)) { + new_input (ly_scm2string (sval), sources_); + yy_pop_state (); + } else { + LexerError (_ ("string expected after \\include").c_str ()); + if (sval != SCM_UNDEFINED) { + SCM err = scm_current_error_port (); + scm_puts ("This value was found instead: ", err); + scm_display (sval, err); + } + } +} + +\"[^""]* { // backup rule error (_ ("end quote missing")); exit (1); } + + /* Flex picks the longest matching pattern including trailing + * contexts. Without the backup pattern, r-. does not trigger the + * {RESTNAME} rule but rather the {WORD}/[-_] rule coming later, + * needed for avoiding backup states. + */ + +{RESTNAME}/[-_] | // pseudo backup rule {RESTNAME} { char const *s = YYText (); - yylval.scm = scm_from_locale_string (s); + yylval = scm_from_locale_string (s); return RESTNAME; } +q/[-_] | // pseudo backup rule +q { + yylval = SCM_UNSPECIFIED; + return CHORD_REPETITION; +} + +R/[-_] | // pseudo backup rule R { + yylval = SCM_UNSPECIFIED; return MULTI_MEASURE_REST; } # { //embedded scm @@ -350,10 +421,7 @@ BOM_UTF8 \357\273\277 be_safe_global && is_main_input_, parser_); if (sval == SCM_UNDEFINED) - { - sval = SCM_UNSPECIFIED; error_level_ = 1; - } for (int i = 0; i < n; i++) { @@ -361,91 +429,117 @@ BOM_UTF8 \357\273\277 } char_count_stack_.back () += n; - if (unpack_identifier (sval) != SCM_UNDEFINED) + yylval = sval; + return SCM_TOKEN; +} + +\$ { //immediate scm + int n = 0; + Input hi = here_input(); + hi.step_forward (); + SCM sval = ly_parse_scm (hi.start (), &n, hi, + be_safe_global && is_main_input_, parser_); + + for (int i = 0; i < n; i++) { - yylval.scm = unpack_identifier(sval); - return identifier_type (yylval.scm); + yyinput (); } - - yylval.scm = sval; - return SCM_TOKEN; + char_count_stack_.back () += n; + + sval = eval_scm (sval, '$'); + + int token = scan_scm_id (sval); + if (!scm_is_eq (yylval, SCM_UNSPECIFIED)) + return token; } + { \<\< { + yylval = SCM_UNSPECIFIED; return DOUBLE_ANGLE_OPEN; } \>\> { + yylval = SCM_UNSPECIFIED; return DOUBLE_ANGLE_CLOSE; } } { \< { + yylval = SCM_UNSPECIFIED; return ANGLE_OPEN; } \> { + yylval = SCM_UNSPECIFIED; return ANGLE_CLOSE; } } { _ { + yylval = SCM_UNSPECIFIED; return FIGURE_SPACE; } \> { + yylval = SCM_UNSPECIFIED; return FIGURE_CLOSE; } \< { + yylval = SCM_UNSPECIFIED; return FIGURE_OPEN; } } { - {ALPHAWORD} { - return scan_bare_word (YYText ()); + {WORD}/[-_] | // backup rule + {WORD} { + return scan_bare_word (YYText_utf8 ()); } - {NOTECOMMAND} { - return scan_escaped_word (YYText () + 1); + {COMMAND}/[-_] | // backup rule + {COMMAND} { + return scan_escaped_word (YYText_utf8 () + 1); } {FRACTION} { - yylval.scm = scan_fraction (YYText ()); + yylval = scan_fraction (YYText ()); return FRACTION; } - - {DIGIT} { - yylval.i = String_convert::dec2int (string (YYText ())); - return DIGIT; - } + {UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval = scm_c_read_string (YYText ()); return UNSIGNED; } {E_UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText () +1)); + yylval = scm_c_read_string (YYText () + 1); return E_UNSIGNED; } } { \\{ESCAPED} { - *yylval.string += to_string (escaped_char (YYText ()[1])); + char c = escaped_char (YYText ()[1]); + yylval = scm_cons (scm_from_locale_stringn (&c, 1), + yylval); } - [^\\"]+ { - *yylval.string += YYText (); + [^\\""]+ { + yylval = scm_cons (scm_from_locale_string (YYText_utf8 ()), + yylval); } \" { yy_pop_state (); /* yylval is union. Must remember STRING before setting SCM*/ - string *sp = yylval.string; - yylval.scm = ly_string2scm (*sp); - delete sp; + + yylval = scm_string_concatenate_reverse (yylval, + SCM_UNDEFINED, + SCM_UNDEFINED); + return is_lyric_state () ? LYRICS_STRING : STRING; } - . { - *yylval.string += YYText (); + \\ { + yylval = scm_cons (scm_from_locale_string (YYText ()), + yylval); } } @@ -454,163 +548,177 @@ BOM_UTF8 \357\273\277 start_lyric_quote (); } {FRACTION} { - yylval.scm = scan_fraction (YYText ()); + yylval = scan_fraction (YYText ()); return FRACTION; } + {UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval = scm_c_read_string (YYText ()); return UNSIGNED; } - {NOTECOMMAND} { - return scan_escaped_word (YYText () + 1); + {COMMAND}/[-_] | // backup rule + {COMMAND} { + return scan_escaped_word (YYText_utf8 () + 1); } {LYRICS} { /* ugr. This sux. */ - string s (YYText ()); + string s (YYText_utf8 ()); + yylval = SCM_UNSPECIFIED; if (s == "__") - return yylval.i = EXTENDER; + return EXTENDER; if (s == "--") - return yylval.i = HYPHEN; + return HYPHEN; s = lyric_fudge (s); char c = s[s.length () - 1]; 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 = ly_string2scm (s); + yylval = ly_string2scm (s); return LYRICS_STRING; } . { - return YYText ()[0]; + yylval = SCM_UNSPECIFIED; + return YYText ()[0]; // LYRICS already catches all multibytes. } } { - {ALPHAWORD} { - return scan_bare_word (YYText ()); + {WORD}/[-_] | // backup rule + {WORD} { + return scan_bare_word (YYText_utf8 ()); } - {NOTECOMMAND} { - return scan_escaped_word (YYText () + 1); + {COMMAND}/[-_] | // backup rule + {COMMAND} { + return scan_escaped_word (YYText_utf8 () + 1); } {FRACTION} { - yylval.scm = scan_fraction (YYText ()); + yylval = scan_fraction (YYText ()); return FRACTION; } + {UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval = scm_c_read_string (YYText ()); return UNSIGNED; } - { + yylval = SCM_UNSPECIFIED; return CHORD_MINUS; } : { + yylval = SCM_UNSPECIFIED; return CHORD_COLON; } \/\+ { + yylval = SCM_UNSPECIFIED; return CHORD_BASS; } \/ { + yylval = SCM_UNSPECIFIED; return CHORD_SLASH; } \^ { + yylval = SCM_UNSPECIFIED; return CHORD_CARET; } . { - return YYText ()[0]; + yylval = SCM_UNSPECIFIED; + return YYText ()[0]; // WORD catches all multibyte. } } { \\score { + yylval = SCM_UNSPECIFIED; return SCORE; } - {MARKUPCOMMAND} { - string str (YYText () + 1); + {COMMAND}/[-_] | // backup rule + {COMMAND} { + string str (YYText_utf8 () + 1); + + int token_type = MARKUP_FUNCTION; SCM s = lookup_markup_command (str); - SCM s2 = lookup_markup_list_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-markup2-markup3")) - return MARKUP_HEAD_SCM0_SCM1_MARKUP2_MARKUP3; - else if (tag == ly_symbol2scm ("scheme0-markup1-markup2")) - return MARKUP_HEAD_SCM0_MARKUP1_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 if (scm_is_pair (s2) && scm_is_symbol (scm_cdr (s2))) { - yylval.scm = scm_car(s2); - SCM tag = scm_cdr(s2); - if (tag == ly_symbol2scm("empty")) - return MARKUP_LIST_HEAD_EMPTY; - else if (tag == ly_symbol2scm ("scheme0")) - return MARKUP_LIST_HEAD_SCM0; - else if (tag == ly_symbol2scm ("markup-list0")) - return MARKUP_LIST_HEAD_LIST0; - else if (tag == ly_symbol2scm ("scheme0-markup-list1")) - return MARKUP_LIST_HEAD_SCM0_LIST1; - else if (tag == ly_symbol2scm ("scheme0-scheme1-markup-list2")) - return MARKUP_LIST_HEAD_SCM0_SCM1_LIST2; - else { - programming_error ("no parser tag defined for this markup list 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_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, predicate); + } + return token_type; } [{}] { + yylval = SCM_UNSPECIFIED; return YYText ()[0]; } - [^#{}"\\ \t\n\r\f]+ { - string s (YYText ()); - - char c = s[s.length () - 1]; - /* brace open is for not confusing dumb tools. */ - if (c == '{' || c == '}') - here_input ().warning ( - _ ("Brace found at end of markup. Did you forget a space?")); - yylval.scm = ly_string2scm (s); - + [^$#{}\"\\ \t\n\r\f]+ { + string s (YYText_utf8 ()); + yylval = ly_string2scm (s); return STRING; } . { - return YYText()[0]; + yylval = SCM_UNSPECIFIED; + return YYText()[0]; // Above is catchall for multibyte } } -<*><> { - if (is_main_input_) +<> { + LexerError (_ ("EOF found inside a comment").c_str ()); + is_main_input_ = false; // should be safe , can't have \include in --safe. + yylval = SCM_UNSPECIFIED; + if (!close_input ()) + yyterminate (); // can't move this, since it actually rets a YY_NULL + } + +<> { + yylval = SCM_UNSPECIFIED; + if (is_main_input_) { /* 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 ()) + if (!close_input () || !is_main_input_) /* Returns YY_NULL */ yyterminate (); } @@ -620,51 +728,52 @@ BOM_UTF8 \357\273\277 } { - {DASHED_WORD} { - return scan_bare_word (YYText ()); + {WORD}/[-_] | // backup rule + {WORD} { + return scan_bare_word (YYText_utf8 ()); } - {DASHED_KEY_WORD} { - return scan_escaped_word (YYText () + 1); + {COMMAND}/[-_] | // backup rule + {COMMAND} { + return scan_escaped_word (YYText_utf8 () + 1); } } -{WORD} { - return scan_bare_word (YYText ()); -} -{KEYWORD} { - return scan_escaped_word (YYText () + 1); +{FRACTION} { + yylval = scan_fraction (YYText ()); + return FRACTION; } -{REAL} { - Real r; - int cnv = sscanf (YYText (), "%lf", &r); - assert (cnv == 1); - (void) cnv; - yylval.scm = scm_from_double (r); +-{UNSIGNED} | // backup rule +{REAL} { + yylval = scm_c_read_string (YYText ()); return REAL; } +{UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval = scm_c_read_string (YYText ()); return UNSIGNED; } [{}] { - + yylval = SCM_UNSPECIFIED; return YYText ()[0]; } -[*:=] { - char c = YYText ()[0]; - return c; +-/\. | // backup rule +[*:=] { + yylval = SCM_UNSPECIFIED; + return YYText ()[0]; } . { + yylval = SCM_UNSPECIFIED; return YYText ()[0]; } \\. { + yylval = SCM_UNSPECIFIED; char c = YYText ()[1]; switch (c) { @@ -694,10 +803,11 @@ BOM_UTF8 \357\273\277 } } -<*>. { - string msg = _f ("invalid character: `%c'", YYText ()[0]); +<*>.[\200-\277]* { + string msg = _f ("invalid character: `%s'", YYText_utf8 ()); LexerError (msg.c_str ()); - return YYText ()[0]; + yylval = SCM_UNSPECIFIED; + return '%'; // Better not return half a utf8 character. } %% @@ -705,21 +815,25 @@ 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) +Lily_lexer::push_extra_token (int token_type, SCM scm) { - if (extra_token_types_.empty ()) + if (scm_is_null (extra_tokens_)) { if (YY_START != extratoken) hidden_state_ = YY_START; yy_push_state (extratoken); } - extra_token_types_.push_back (token_type); + extra_tokens_ = scm_acons (scm_from_int (token_type), scm, extra_tokens_); } void -Lily_lexer::push_chord_state (SCM tab) +Lily_lexer::push_chord_state (SCM alist) { - pitchname_tab_stack_ = scm_cons (tab, pitchname_tab_stack_); + SCM p = scm_assq (alist, pitchname_tab_stack_); + + if (scm_is_false (p)) + p = scm_cons (alist, alist_to_hashq (alist)); + pitchname_tab_stack_ = scm_cons (p, pitchname_tab_stack_); yy_push_state (chords); } @@ -748,25 +862,49 @@ Lily_lexer::push_markup_state () } void -Lily_lexer::push_note_state (SCM tab) +Lily_lexer::push_note_state (SCM alist) { - pitchname_tab_stack_ = scm_cons (tab, pitchname_tab_stack_); + bool extra = (YYSTATE == extratoken); + + SCM p = scm_assq (alist, pitchname_tab_stack_); + + if (extra) + yy_pop_state (); + + if (scm_is_false (p)) + p = scm_cons (alist, alist_to_hashq (alist)); + pitchname_tab_stack_ = scm_cons (p, pitchname_tab_stack_); yy_push_state (notes); + + if (extra) { + hidden_state_ = YYSTATE; + yy_push_state (extratoken); + } } void Lily_lexer::pop_state () { + bool extra = (YYSTATE == extratoken); + + if (extra) + yy_pop_state (); + if (YYSTATE == notes || YYSTATE == chords) pitchname_tab_stack_ = scm_cdr (pitchname_tab_stack_); yy_pop_state (); + + if (extra) { + hidden_state_ = YYSTATE; + yy_push_state (extratoken); + } } int Lily_lexer::identifier_type (SCM sid) { - int k = try_special_identifiers (&yylval.scm , sid); + int k = try_special_identifiers (&yylval , sid); return k >= 0 ? k : SCM_IDENTIFIER; } @@ -778,6 +916,7 @@ Lily_lexer::scan_escaped_word (string str) // SCM sym = ly_symbol2scm (str.c_str ()); + yylval = SCM_UNSPECIFIED; int i = lookup_keyword (str); if (i == MARKUP && is_lyric_state ()) return LYRIC_MARKUP; @@ -785,37 +924,72 @@ Lily_lexer::scan_escaped_word (string str) return i; SCM sid = lookup_identifier (str); + if (sid != SCM_UNDEFINED) + return scan_scm_id (sid); + + string msg (_f ("unknown escaped string: `\\%s'", str)); + LexerError (msg.c_str ()); + + yylval = ly_string2scm (str); + + return STRING; +} + +int +Lily_lexer::scan_scm_id (SCM sid) +{ if (is_music_function (sid)) { - yylval.scm = get_music_function_transform (sid); + int funtype = SCM_FUNCTION; - 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)) + yylval = sid; + + SCM s = get_music_function_signature (sid); + SCM cs = scm_car (s); + + if (scm_is_pair (cs)) { - if (scm_car (s) == ly_music_p_proc) - push_extra_token (EXPECT_MUSIC); - else if (scm_car (s) == ly_lily_module_constant ("markup?")) - push_extra_token (EXPECT_MARKUP); - else if (ly_is_procedure (scm_car (s))) - push_extra_token (EXPECT_SCM); - else programming_error ("Function parameter without type-checking predicate"); + cs = SCM_CAR (cs); } - return MUSIC_FUNCTION; - } - if (sid != SCM_UNDEFINED) - { - yylval.scm = sid; - return identifier_type (sid); - } - - string msg (_f ("unknown escaped string: `\\%s'", str)); - LexerError (msg.c_str ()); + if (scm_is_eq (cs, ly_lily_module_constant ("ly:music?"))) + funtype = MUSIC_FUNCTION; + else if (scm_is_eq (cs, ly_lily_module_constant ("ly:event?"))) + funtype = EVENT_FUNCTION; + else if (ly_is_procedure (cs)) + funtype = SCM_FUNCTION; + else programming_error ("Bad syntax function predicate"); - yylval.scm = ly_string2scm (str); + push_extra_token (EXPECT_NO_MORE_ARGS); + for (s = scm_cdr (s); scm_is_pair (s); s = scm_cdr (s)) + { + SCM optional = SCM_UNDEFINED; + cs = scm_car (s); - return STRING; + if (scm_is_pair (cs)) + { + optional = SCM_CDR (cs); + cs = SCM_CAR (cs); + } + + 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 (ly_is_procedure (cs)) + push_extra_token (EXPECT_SCM, cs); + else + { + programming_error ("Function parameter without type-checking predicate"); + continue; + } + if (!scm_is_eq (optional, SCM_UNDEFINED)) + push_extra_token (EXPECT_OPTIONAL, optional); + } + return funtype; + } + yylval = sid; + return identifier_type (sid); } int @@ -825,27 +999,23 @@ Lily_lexer::scan_bare_word (string str) if ((YYSTATE == notes) || (YYSTATE == chords)) { SCM handle = SCM_BOOL_F; if (scm_is_pair (pitchname_tab_stack_)) - handle = scm_hashq_get_handle (scm_car (pitchname_tab_stack_), sym); + handle = scm_hashq_get_handle (scm_cdar (pitchname_tab_stack_), sym); if (scm_is_pair (handle)) { - yylval.scm = scm_cdr (handle); - if (unsmob_pitch (yylval.scm)) + yylval = scm_cdr (handle); + if (unsmob_pitch (yylval)) return (YYSTATE == notes) ? NOTENAME_PITCH : TONICNAME_PITCH; - else if (scm_is_symbol (yylval.scm)) + else if (scm_is_symbol (yylval)) return DRUM_PITCH; } else if ((YYSTATE == chords) && (handle = scm_hashq_get_handle (chordmodifier_tab_, sym))!= SCM_BOOL_F) { - yylval.scm = scm_cdr (handle); + yylval = 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 = ly_string2scm (str); + yylval = ly_string2scm (str); return STRING; } @@ -882,6 +1052,170 @@ Lily_lexer::is_figure_state () const return get_state () == figures; } +// The extra_token parameter specifies how to convert multiple values +// into additional tokens. For '#', additional values get pushed as +// SCM_IDENTIFIER. For '$', they get checked for their type and get +// pushed as a corresponding *_IDENTIFIER token. Since the latter +// tampers with yylval, it can only be done from the lexer itself, so +// this function is private. + +SCM +Lily_lexer::eval_scm (SCM readerdata, char extra_token) +{ + SCM sval = SCM_UNDEFINED; + + if (!SCM_UNBNDP (readerdata)) + { + sval = ly_eval_scm (scm_car (readerdata), + *unsmob_input (scm_cdr (readerdata)), + be_safe_global && is_main_input_, + parser_); + } + + if (SCM_UNBNDP (sval)) + { + error_level_ = 1; + return SCM_UNSPECIFIED; + } + + if (extra_token && SCM_VALUESP (sval)) + { + sval = scm_struct_ref (sval, SCM_INUM0); + + if (scm_is_pair (sval)) { + for (SCM v = scm_reverse (scm_cdr (sval)); + scm_is_pair (v); + v = scm_cdr (v)) + { + int token; + switch (extra_token) { + case '$': + token = scan_scm_id (scm_car (v)); + if (!scm_is_eq (yylval, SCM_UNSPECIFIED)) + push_extra_token (token, yylval); + break; + case '#': + push_extra_token (SCM_IDENTIFIER, scm_car (v)); + break; + } + } + sval = scm_car (sval); + } else + sval = SCM_UNSPECIFIED; + } + + return sval; +} + +/* Check for valid UTF-8 that has no overlong or surrogate codes and + is in the range 0-0x10ffff */ + +const char * +Lily_lexer::YYText_utf8 () +{ + const char * const p = YYText (); + for (int i=0; p[i];) { + if ((p[i] & 0xff) < 0x80) { + ++i; + continue; + } + int oldi = i; // start of character + int more = 0; // # of followup bytes, 0 if bad + switch (p[i++] & 0xff) { + // 0xc0 and 0xc1 are overlong prefixes for + // 0x00-0x3f and 0x40-0x7f respectively, bad. + case 0xc2: // 0x80-0xbf + case 0xc3: // 0xc0-0xff + case 0xc4: // 0x100-0x13f + case 0xc5: // 0x140-0x17f + case 0xc6: // 0x180-0x1bf + case 0xc7: // 0x1c0-0x1ff + case 0xc8: // 0x200-0x23f + case 0xc9: // 0x240-0x27f + case 0xca: // 0x280-0x2bf + case 0xcb: // 0x2c0-0x2ff + case 0xcc: // 0x300-0x33f + case 0xcd: // 0x340-0x37f + case 0xce: // 0x380-0x3bf + case 0xcf: // 0x3c0-0x3ff + case 0xd0: // 0x400-0x43f + case 0xd1: // 0x440-0x47f + case 0xd2: // 0x480-0x4bf + case 0xd3: // 0x4c0-0x4ff + case 0xd4: // 0x500-0x53f + case 0xd5: // 0x540-0x57f + case 0xd6: // 0x580-0x5bf + case 0xd7: // 0x5c0-0x5ff + case 0xd8: // 0x600-0x63f + case 0xd9: // 0x640-0x67f + case 0xda: // 0x680-0x6bf + case 0xdb: // 0x6c0-0x6ff + case 0xdc: // 0x700-0x73f + case 0xdd: // 0x740-0x77f + case 0xde: // 0x780-0x7bf + case 0xdf: // 0x7c0-0x7ff + more = 1; // 2-byte sequences, 0x80-0x7ff + break; + case 0xe0: + // don't allow overlong sequences for 0-0x7ff + if ((p[i] & 0xff) < 0xa0) + break; + case 0xe1: // 0x1000-0x1fff + case 0xe2: // 0x2000-0x2fff + case 0xe3: // 0x3000-0x3fff + case 0xe4: // 0x4000-0x4fff + case 0xe5: // 0x5000-0x5fff + case 0xe6: // 0x6000-0x6fff + case 0xe7: // 0x7000-0x7fff + case 0xe8: // 0x8000-0x8fff + case 0xe9: // 0x9000-0x9fff + case 0xea: // 0xa000-0xafff + case 0xeb: // 0xb000-0xbfff + case 0xec: // 0xc000-0xcfff + more = 2; // 3-byte sequences, 0x7ff-0xcfff + break; + case 0xed: // 0xd000-0xdfff + // Don't allow surrogate codes 0xd800-0xdfff + if ((p[i] & 0xff) >= 0xa0) + break; + case 0xee: // 0xe000-0xefff + case 0xef: // 0xf000-0xffff + more = 2; // 3-byte sequences, + // 0xd000-0xd7ff, 0xe000-0xffff + break; + case 0xf0: + // don't allow overlong sequences for 0-0xffff + if ((p[i] & 0xff) < 0x90) + break; + case 0xf1: // 0x40000-0x7ffff + case 0xf2: // 0x80000-0xbffff + case 0xf3: // 0xc0000-0xfffff + more = 3; // 4-byte sequences, 0x10000-0xfffff + break; + case 0xf4: + // don't allow more than 0x10ffff + if ((p[i] & 0xff) >= 0x90) + break; + more = 3; // 4-byte sequence, 0x100000-0x10ffff + break; + } + if (more) { + // check that all continuation bytes are valid + do { + if ((p[i++] & 0xc0) != 0x80) + break; + } while (--more); + if (!more) + continue; + } + Input h = here_input (); + h.set (h.get_source_file (), h.start () + oldi, h.start () + i); + h.warning (_ ("non-UTF-8 input").c_str ()); + } + return p; +} + + /* urg, belong to string (_convert) and should be generalised @@ -935,30 +1269,18 @@ is_valid_version (string s) /* - substitute _ and \, + substitute _ */ string lyric_fudge (string s) { - char *chars = string_copy (s); - - for (char *p = chars; *p ; p++) - { - if (*p == '_' && (p == chars || *(p-1) != '\\')) - *p = ' '; - } - - s = string (chars); - delete[] chars; + size_t i=0; - ssize i = 0; - if ((i = s.find ("\\,")) != NPOS) // change "\," to TeX's "\c " - { - * (((char*)s.c_str ()) + i + 1) = 'c'; - s = s.substr (0, i + 2) + " " + s.substr (i - 2); - } - - return s; + while ((i = s.find ('_', i)) != string::npos) + { + s[i++] = ' '; + } + return s; } /*