X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=lily%2Flexer.ll;h=afb881f2b8f370e3d2afbc87f744517be2c44281;hb=9d3a999b3ae3502633858c50185c0180949d405c;hp=1777398f8ab3b2caafbceaad6c6f055960cce823;hpb=b82a74e8a67881d8dbe11bdc9c09dbe853236370;p=lilypond.git diff --git a/lily/lexer.ll b/lily/lexer.ll index 1777398f8a..afb881f2b8 100644 --- a/lily/lexer.ll +++ b/lily/lexer.ll @@ -2,7 +2,7 @@ /* This file is part of LilyPond, the GNU music typesetter. - Copyright (C) 1996--2011 Han-Wen Nienhuys + Copyright (C) 1996--2012 Han-Wen Nienhuys Jan Nieuwenhuizen LilyPond is free software: you can redistribute it and/or modify @@ -131,16 +131,31 @@ 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}|{SPECIAL_CHAR} +TEX {AA}|-|{PUNCT}|{NATIONAL}|{SPECIAL_CHAR} DASHED_WORD {A}({AN}|-)* DASHED_KEY_WORD \\{DASHED_WORD} @@ -217,12 +232,15 @@ BOM_UTF8 \357\273\277 yy_push_state (longcomment); } %[^{\n\r][^\n\r]*[\n\r] { + (void) YYText_utf8 (); } %[^{\n\r] { // backup rule + (void) YYText_utf8 (); } %[\n\r] { } %[^{\n\r][^\n\r]* { + (void) YYText_utf8 (); } {WHITE}+ { @@ -245,7 +263,7 @@ 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 (); @@ -258,8 +276,8 @@ BOM_UTF8 \357\273\277 } -\"[^"]*\" { - string s (YYText () + 1); +\"[^""]*\" { + string s (YYText_utf8 () + 1); s = s.substr (0, s.rfind ('\"')); yy_pop_state (); @@ -294,9 +312,10 @@ BOM_UTF8 \357\273\277 } { [^\%]* { + (void) YYText_utf8 (); } \%*[^}%]* { - + (void) YYText_utf8 (); } "%"+"}" { yy_pop_state (); @@ -318,14 +337,14 @@ BOM_UTF8 \357\273\277 yy_push_state (incl); } \"[^""]*\" { /* got the include file name */ - string s (YYText ()+1); + 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; + string s = YYText_utf8 () + 1; strip_trailing_white (s); if (s.length () && (s[s.length () - 1] == ';')) s = s.substr (0, s.length () - 1); @@ -379,6 +398,10 @@ BOM_UTF8 \357\273\277 yylval.scm = scm_from_locale_string (s); return RESTNAME; } +q { + return CHORD_REPETITION; +} + R { return MULTI_MEASURE_REST; } @@ -415,11 +438,11 @@ BOM_UTF8 \357\273\277 } char_count_stack_.back () += n; - sval = eval_scm (sval); - + sval = eval_scm (sval, '$'); + int token = scan_scm_id (sval); if (!scm_is_eq (yylval.scm, SCM_UNSPECIFIED)) - return token; + return token; } { @@ -454,11 +477,11 @@ BOM_UTF8 \357\273\277 { {ALPHAWORD} { - return scan_bare_word (YYText ()); + return scan_bare_word (YYText_utf8 ()); } {NOTECOMMAND} { - return scan_escaped_word (YYText () + 1); + return scan_escaped_word (YYText_utf8 () + 1); } {FRACTION} { yylval.scm = scan_fraction (YYText ()); @@ -480,7 +503,7 @@ BOM_UTF8 \357\273\277 *yylval.string += to_string (escaped_char (YYText ()[1])); } [^\\""]+ { - *yylval.string += YYText (); + *yylval.string += YYText_utf8 (); } \" { @@ -492,7 +515,7 @@ BOM_UTF8 \357\273\277 delete sp; return is_lyric_state () ? LYRICS_STRING : STRING; } - . { + \\ { *yylval.string += YYText (); } } @@ -515,11 +538,11 @@ BOM_UTF8 \357\273\277 return UNSIGNED; } {NOTECOMMAND} { - return scan_escaped_word (YYText () + 1); + return scan_escaped_word (YYText_utf8 () + 1); } {LYRICS} { /* ugr. This sux. */ - string s (YYText ()); + string s (YYText_utf8 ()); if (s == "__") return yylval.i = EXTENDER; if (s == "--") @@ -536,15 +559,15 @@ BOM_UTF8 \357\273\277 return LYRICS_STRING; } . { - return YYText ()[0]; + return YYText ()[0]; // LYRICS already catches all multibytes. } } { {ALPHAWORD} { - return scan_bare_word (YYText ()); + return scan_bare_word (YYText_utf8 ()); } {NOTECOMMAND} { - return scan_escaped_word (YYText () + 1); + return scan_escaped_word (YYText_utf8 () + 1); } {FRACTION} { yylval.scm = scan_fraction (YYText ()); @@ -575,7 +598,7 @@ BOM_UTF8 \357\273\277 return CHORD_CARET; } . { - return YYText ()[0]; + return YYText ()[0]; // ALPHAWORD catches all multibyte. } } @@ -585,7 +608,7 @@ BOM_UTF8 \357\273\277 return SCORE; } {MARKUPCOMMAND} { - string str (YYText () + 1); + string str (YYText_utf8 () + 1); int token_type = MARKUP_FUNCTION; SCM s = lookup_markup_command (str); @@ -638,7 +661,7 @@ BOM_UTF8 \357\273\277 return YYText ()[0]; } [^$#{}\"\\ \t\n\r\f]+ { - string s (YYText ()); + string s (YYText_utf8 ()); char c = s[s.length () - 1]; /* brace open is for not confusing dumb tools. */ @@ -651,7 +674,7 @@ BOM_UTF8 \357\273\277 return STRING; } . { - return YYText()[0]; + return YYText()[0]; // Above is catchall for multibyte } } @@ -680,13 +703,18 @@ BOM_UTF8 \357\273\277 { {DASHED_WORD} { - return scan_bare_word (YYText ()); + return scan_bare_word (YYText_utf8 ()); } {DASHED_KEY_WORD} { - return scan_escaped_word (YYText () + 1); + return scan_escaped_word (YYText_utf8 () + 1); } } +{FRACTION} { + yylval.scm = scan_fraction (YYText ()); + return FRACTION; +} + -{UNSIGNED} | // backup rule {REAL} { yylval.scm = scm_c_read_string (YYText ()); @@ -697,6 +725,7 @@ BOM_UTF8 \357\273\277 return REAL; } +{UNSIGNED}/\/ | // backup rule {UNSIGNED} { yylval.scm = scm_c_read_string (YYText ()); return UNSIGNED; @@ -747,10 +776,10 @@ 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]; + return '%'; // Better not return half a utf8 character. } %% @@ -810,10 +839,20 @@ Lily_lexer::push_note_state (SCM tab) 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 @@ -856,9 +895,9 @@ Lily_lexer::scan_scm_id (SCM sid) { int funtype = SCM_FUNCTION; - yylval.scm = get_music_function_transform (sid); + yylval.scm = sid; - SCM s = scm_object_property (yylval.scm, ly_symbol2scm ("music-function-signature")); + SCM s = get_music_function_signature (sid); SCM cs = scm_car (s); if (scm_is_pair (cs)) @@ -928,11 +967,7 @@ Lily_lexer::scan_bare_word (string str) 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 = ly_string2scm (str); return STRING; } @@ -970,8 +1005,15 @@ 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 '$', they get checked for their type and get +// pushed as a corresponding *_IDENTIFIER token. Since the latter +// tampers with yylval, it can only be done from the lexer itself, so +// this function is private. + SCM -Lily_lexer::eval_scm (SCM readerdata) +Lily_lexer::eval_scm (SCM readerdata, char extra_token) { SCM sval = SCM_UNDEFINED; @@ -988,9 +1030,143 @@ Lily_lexer::eval_scm (SCM readerdata) error_level_ = 1; return SCM_UNSPECIFIED; } + + if (extra_token && SCM_VALUESP (sval)) + { + sval = scm_struct_ref (sval, SCM_INUM0); + + if (scm_is_pair (sval)) { + for (SCM v = scm_reverse (scm_cdr (sval)); + scm_is_pair (v); + v = scm_cdr (v)) + { + int token; + switch (extra_token) { + case '$': + token = scan_scm_id (scm_car (v)); + if (!scm_is_eq (yylval.scm, SCM_UNSPECIFIED)) + push_extra_token (token, yylval.scm); + 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; +} /* @@ -1046,30 +1222,18 @@ 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; - - 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); - } + size_t i=0; - return s; + while ((i = s.find ('_', i)) != string::npos) + { + s[i++] = ' '; + } + return s; } /*