X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=lily%2Flexer.ll;h=c2d8b6baeff1273520aa7a68eb32492ad286ec3e;hb=4d188a039a5b45e3431f764e4754af20e16072d9;hp=707f84bd83727369deed3f303108f07cd9342668;hpb=34e5b21b1af28e4d787ffaf758d4cd4b5b503053;p=lilypond.git diff --git a/lily/lexer.ll b/lily/lexer.ll index 707f84bd83..c2d8b6baef 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--2011 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 @@ -47,7 +55,6 @@ using namespace std; #include "context-def.hh" #include "duration.hh" -#include "identifier-smob.hh" #include "international.hh" #include "interval.hh" #include "lily-guile.hh" @@ -87,11 +94,9 @@ bool is_valid_version (string s); yy_push_state (lyric_quote);\ yylval.string = new string -#define yylval \ - (*(YYSTYPE*)lexval_) +#define yylval (*lexval_) -#define yylloc \ - (*(YYLTYPE*)lexloc_) +#define yylloc (*lexloc_) #define YY_USER_ACTION add_lexed_char (YYLeng ()); @@ -126,6 +131,22 @@ 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] @@ -133,8 +154,9 @@ AN {AA}|{N} ANY_CHAR (.|\n) PUNCT [?!:'`] ACCENT \\[`'"^] +SPECIAL_CHAR [&@] NATIONAL [\001-\006\021-\027\031\036] -TEX {AA}|-|{PUNCT}|{ACCENT}|{NATIONAL} +TEX {AA}|-|{PUNCT}|{ACCENT}|{NATIONAL}|{SPECIAL_CHAR} DASHED_WORD {A}({AN}|-)* DASHED_KEY_WORD \\{DASHED_WORD} @@ -173,9 +195,10 @@ 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 = scm_cdar (extra_tokens_); + extra_tokens_ = scm_cdr (extra_tokens_); + if (scm_is_null (extra_tokens_)) yy_pop_state (); return type; @@ -185,9 +208,10 @@ BOM_UTF8 \357\273\277 /* Generate a token without swallowing anything */ /* 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 = scm_cdar (extra_tokens_); + extra_tokens_ = scm_cdr (extra_tokens_); + if (scm_is_null (extra_tokens_)) yy_pop_state (); return type; @@ -209,12 +233,15 @@ 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}+ { @@ -237,7 +264,7 @@ 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 (); @@ -251,7 +278,7 @@ BOM_UTF8 \357\273\277 } \"[^"]*\" { - string s (YYText () + 1); + string s (YYText_utf8 () + 1); s = s.substr (0, s.rfind ('\"')); yy_pop_state (); @@ -286,9 +313,10 @@ BOM_UTF8 \357\273\277 } { [^\%]* { + (void) YYText_utf8 (); } \%*[^}%]* { - + (void) YYText_utf8 (); } "%"+"}" { yy_pop_state (); @@ -309,15 +337,15 @@ 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; + string s = YYText_utf8 () + 1; strip_trailing_white (s); if (s.length () && (s[s.length () - 1] == ';')) s = s.substr (0, s.length () - 1); @@ -326,7 +354,7 @@ 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 ()); @@ -335,7 +363,34 @@ 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); } @@ -344,6 +399,10 @@ BOM_UTF8 \357\273\277 yylval.scm = scm_from_locale_string (s); return RESTNAME; } +q { + return CHORD_REPETITION; +} + R { return MULTI_MEASURE_REST; } @@ -355,10 +414,7 @@ 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++) { @@ -366,20 +422,30 @@ BOM_UTF8 \357\273\277 } char_count_stack_.back () += n; - if (unpack_identifier (sval) != SCM_UNDEFINED) + yylval.scm = 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 (); } + char_count_stack_.back () += n; - for (size_t i = 0; i < pending_string_includes_.size (); i++) - new_input ("", pending_string_includes_[i], - parser_->sources_); - pending_string_includes_.clear (); + sval = eval_scm (sval); - yylval.scm = sval; - return SCM_TOKEN; + int token = scan_scm_id (sval); + if (!scm_is_eq (yylval.scm, SCM_UNSPECIFIED)) + return token; } + { \<\< { return DOUBLE_ANGLE_OPEN; @@ -412,11 +478,11 @@ 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 ()); @@ -424,7 +490,7 @@ BOM_UTF8 \357\273\277 } {UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval.scm = scm_c_read_string (YYText ()); return UNSIGNED; } {E_UNSIGNED} { @@ -438,7 +504,7 @@ BOM_UTF8 \357\273\277 *yylval.string += to_string (escaped_char (YYText ()[1])); } [^\\""]+ { - *yylval.string += YYText (); + *yylval.string += YYText_utf8 (); } \" { @@ -450,7 +516,7 @@ BOM_UTF8 \357\273\277 delete sp; return is_lyric_state () ? LYRICS_STRING : STRING; } - . { + \\ { *yylval.string += YYText (); } } @@ -464,20 +530,20 @@ BOM_UTF8 \357\273\277 return FRACTION; } {UNSIGNED}/\/[^0-9] { // backup rule - yylval.i = String_convert::dec2int (string (YYText ())); + yylval.scm = scm_c_read_string (YYText ()); return UNSIGNED; } {UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval.scm = scm_c_read_string (YYText ()); 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 == "--") @@ -494,27 +560,27 @@ 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 ()); return FRACTION; } {UNSIGNED}/\/[^0-9] { // backup rule - yylval.i = String_convert::dec2int (string (YYText ())); + yylval.scm = scm_c_read_string (YYText ()); return UNSIGNED; } {UNSIGNED}/\/ | // backup rule {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval.scm = scm_c_read_string (YYText ()); return UNSIGNED; } - { @@ -533,7 +599,7 @@ BOM_UTF8 \357\273\277 return CHORD_CARET; } . { - return YYText ()[0]; + return YYText ()[0]; // ALPHAWORD catches all multibyte. } } @@ -543,7 +609,7 @@ 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); @@ -588,15 +654,15 @@ BOM_UTF8 \357\273\277 else if (predicate == ly_lily_module_constant ("markup?")) push_extra_token(EXPECT_MARKUP); else - push_extra_token(EXPECT_SCM); + push_extra_token(EXPECT_SCM, predicate); } return token_type; } [{}] { return YYText ()[0]; } - [^#{}\"\\ \t\n\r\f]+ { - string s (YYText ()); + [^$#{}\"\\ \t\n\r\f]+ { + string s (YYText_utf8 ()); char c = s[s.length () - 1]; /* brace open is for not confusing dumb tools. */ @@ -609,7 +675,7 @@ BOM_UTF8 \357\273\277 return STRING; } . { - return YYText()[0]; + return YYText()[0]; // Above is catchall for multibyte } } @@ -638,21 +704,16 @@ 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); } } -{UNSIGNED} | // backup rule {REAL} { - Real r; - int cnv = sscanf (YYText (), "%lf", &r); - assert (cnv == 1); - (void) cnv; - - yylval.scm = scm_from_double (r); + yylval.scm = scm_c_read_string (YYText ()); return REAL; } -\. { // backup rule @@ -661,7 +722,7 @@ BOM_UTF8 \357\273\277 } {UNSIGNED} { - yylval.i = String_convert::dec2int (string (YYText ())); + yylval.scm = scm_c_read_string (YYText ()); return UNSIGNED; } @@ -710,10 +771,10 @@ 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. } %% @@ -721,15 +782,15 @@ 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 @@ -773,10 +834,20 @@ 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 @@ -801,6 +872,20 @@ 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.scm = ly_string2scm (str); + + return STRING; +} + +int +Lily_lexer::scan_scm_id (SCM sid) +{ if (is_music_function (sid)) { int funtype = SCM_FUNCTION; @@ -817,6 +902,8 @@ Lily_lexer::scan_escaped_word (string 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"); @@ -824,35 +911,33 @@ Lily_lexer::scan_escaped_word (string 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); + + if (scm_is_pair (cs)) + { + optional = SCM_CDR (cs); + cs = SCM_CAR (cs); + } - if (cs == ly_music_p_proc) - push_extra_token (EXPECT_MUSIC); - else if (cs == Pitch_type_p_proc) + 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"); + 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; } - - if (sid != SCM_UNDEFINED) - { - yylval.scm = sid; - return identifier_type (sid); - } - - string msg (_f ("unknown escaped string: `\\%s'", str)); - LexerError (msg.c_str ()); - - yylval.scm = ly_string2scm (str); - - return STRING; + yylval.scm = sid; + return identifier_type (sid); } int @@ -877,11 +962,7 @@ 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; } @@ -919,6 +1000,136 @@ Lily_lexer::is_figure_state () const return get_state () == figures; } +SCM +Lily_lexer::eval_scm (SCM readerdata) +{ + 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; + } + 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 @@ -972,30 +1183,18 @@ is_valid_version (string s) /* - substitute _ and \, + substitute _ */ string lyric_fudge (string s) { - char *chars = string_copy (s); + size_t i=0; - 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); - } - - return s; + while ((i = s.find ('_', i)) != string::npos) + { + s[i++] = ' '; + } + return s; } /*