source: src/Parser/lex.ll@ 7305915

ADT aaron-thesis arm-eh ast-experimental cleanup-dtors ctor deferred_resn demangler enum forall-pointer-decay gc_noraii jacob/cs343-translation jenkins-sandbox memory new-ast new-ast-unique-expr new-env no_list persistent-indexer pthread-emulation qualifiedEnum resolv-new with_gc
Last change on this file since 7305915 was c6b1105, checked in by Peter A. Buhr <pabuhr@…>, 9 years ago

add reference operator, labels cannot be 0/1, small changes to tests

  • Property mode set to 100644
File size: 15.3 KB
Line 
1/*
2 * Cforall Version 1.0.0 Copyright (C) 2015 University of Waterloo
3 *
4 * The contents of this file are covered under the licence agreement in the
5 * file "LICENCE" distributed with Cforall.
6 *
7 * lex.l --
8 *
9 * Author : Peter A. Buhr
10 * Created On : Sat Sep 22 08:58:10 2001
11 * Last Modified By : Peter A. Buhr
12 * Last Modified On : Thu Jun 23 07:29:30 2016
13 * Update Count : 457
14 */
15
16%option yylineno
17%option nounput
18
19%{
20// This lexer assumes the program has been preprocessed by cpp. Hence, all user level preprocessor directive have been
21// performed and removed from the source. The only exceptions are preprocessor directives passed to the compiler (e.g.,
22// line-number directives) and C/C++ style comments, which are ignored.
23
24//**************************** Includes and Defines ****************************
25
26#include <string>
27
28#include "lex.h"
29#include "parser.h" // YACC generated definitions based on C++ grammar
30#include "ParseNode.h"
31#include "TypedefTable.h"
32
33char *yyfilename;
34std::string *strtext; // accumulate parts of character and string constant value
35
36#define RETURN_LOCN(x) yylval.tok.loc.file = yyfilename; yylval.tok.loc.line = yylineno; return( x )
37#define RETURN_VAL(x) yylval.tok.str = new std::string( yytext ); RETURN_LOCN( x )
38#define RETURN_CHAR(x) yylval.tok.str = NULL; RETURN_LOCN( x )
39#define RETURN_STR(x) yylval.tok.str = strtext; RETURN_LOCN( x )
40
41#define WHITE_RETURN(x) // do nothing
42#define NEWLINE_RETURN() WHITE_RETURN( '\n' )
43#define ASCIIOP_RETURN() RETURN_CHAR( (int)yytext[0] ) // single character operator
44#define NAMEDOP_RETURN(x) RETURN_VAL( x ) // multichar operator, with a name
45#define NUMERIC_RETURN(x) rm_underscore(); RETURN_VAL( x ) // numeric constant
46#define KEYWORD_RETURN(x) RETURN_CHAR( x ) // keyword
47#define IDENTIFIER_RETURN() RETURN_VAL( typedefTable.isKind( yytext ) )
48#define ATTRIBUTE_RETURN() RETURN_VAL( ATTR_IDENTIFIER )
49
50void rm_underscore() {
51 // Remove underscores in numeric constant by copying the non-underscore characters to the front of the string.
52 yyleng = 0;
53 for ( int i = 0; yytext[i] != '\0'; i += 1 ) {
54 if ( yytext[i] != '_' ) {
55 yytext[yyleng] = yytext[i];
56 yyleng += 1;
57 } // if
58 } // for
59 yytext[yyleng] = '\0';
60}
61
62%}
63
64octal [0-7]
65nonzero [1-9]
66decimal [0-9]
67hex [0-9a-fA-F]
68universal_char "\\"((u"_"?{hex_quad})|(U"_"?{hex_quad}{2}))
69
70 // identifier, GCC: $ in identifier
71identifier ([a-zA-Z_$]|{universal_char})([0-9a-zA-Z_$]|{universal_char})*
72
73 // quoted identifier
74quoted_identifier "`"{identifier}"`"
75
76 // attribute identifier, GCC: $ in identifier
77attr_identifier "@"{identifier}
78
79 // numeric constants, CFA: '_' in constant
80hex_quad {hex}("_"?{hex}){3}
81integer_suffix "_"?(([uU][lL]?)|([uU]("ll"|"LL")?)|([lL][uU]?)|("ll"|"LL")[uU]?)
82
83octal_digits ({octal})|({octal}({octal}|"_")*{octal})
84octal_prefix "0""_"?
85octal_constant (("0")|({octal_prefix}{octal_digits})){integer_suffix}?
86
87nonzero_digits ({nonzero})|({nonzero}({decimal}|"_")*{decimal})
88decimal_constant {nonzero_digits}{integer_suffix}?
89
90hex_digits ({hex})|({hex}({hex}|"_")*{hex})
91hex_prefix "0"[xX]"_"?
92hex_constant {hex_prefix}{hex_digits}{integer_suffix}?
93
94decimal_digits ({decimal})|({decimal}({decimal}|"_")*{decimal})
95fractional_constant ({decimal_digits}?"."{decimal_digits})|({decimal_digits}".")
96exponent "_"?[eE]"_"?[+-]?{decimal_digits}
97 // GCC: D (double), DL (long double) and iI (imaginary) suffixes
98floating_suffix "_"?([fFdDlL][iI]?|"DL"|[iI][lLfFdD]?)
99 //floating_suffix "_"?([fFdD]|[lL]|[D][L])|([iI][lLfFdD])|([lLfFdD][iI]))
100floating_constant (({fractional_constant}{exponent}?)|({decimal_digits}{exponent})){floating_suffix}?
101
102binary_exponent "_"?[pP]"_"?[+-]?{decimal_digits}
103hex_fractional_constant ({hex_digits}?"."{hex_digits})|({hex_digits}".")
104hex_floating_constant {hex_prefix}(({hex_fractional_constant}{binary_exponent})|({hex_digits}{binary_exponent})){floating_suffix}?
105
106 // character escape sequence, GCC: \e => esc character
107simple_escape "\\"[abefnrtv'"?\\]
108 // ' stop highlighting
109octal_escape "\\"{octal}("_"?{octal}){0,2}
110hex_escape "\\""x""_"?{hex_digits}
111escape_seq {simple_escape}|{octal_escape}|{hex_escape}|{universal_char}
112cwide_prefix "L"|"U"|"u"
113swide_prefix {cwide_prefix}|"u8"
114
115 // display/white-space characters
116h_tab [\011]
117form_feed [\014]
118v_tab [\013]
119c_return [\015]
120h_white [ ]|{h_tab}
121
122 // overloadable operators
123op_unary_only "~"|"!"
124op_unary_binary "+"|"-"|"*"
125op_unary_pre_post "++"|"--"
126op_unary {op_unary_only}|{op_unary_binary}|{op_unary_pre_post}
127
128op_binary_only "/"|"%"|"^"|"&"|"|"|"<"|">"|"="|"=="|"!="|"<<"|">>"|"<="|">="|"+="|"-="|"*="|"/="|"%="|"&="|"|="|"^="|"<<="|">>="
129op_binary_over {op_unary_binary}|{op_binary_only}
130 // op_binary_not_over "?"|"->"|"."|"&&"|"||"|"@="
131 // operator {op_unary_pre_post}|{op_binary_over}|{op_binary_not_over}
132
133%x COMMENT
134%x BKQUOTE
135%x QUOTE
136%x STRING
137
138%%
139 /* line directives */
140^{h_white}*"#"{h_white}*[0-9]+{h_white}*["][^"\n]+["].*"\n" {
141 /* " stop highlighting */
142 char *end_num;
143 char *begin_string, *end_string;
144 char *filename;
145 long lineno, length;
146 lineno = strtol( yytext + 1, &end_num, 0 );
147 begin_string = strchr( end_num, '"' );
148 if ( begin_string ) {
149 end_string = strchr( begin_string + 1, '"' );
150 if ( end_string ) {
151 length = end_string - begin_string - 1;
152 filename = new char[ length + 1 ];
153 memcpy( filename, begin_string + 1, length );
154 filename[ length ] = '\0';
155 //std::cout << "file " << filename << " line " << lineno << std::endl;
156 yylineno = lineno;
157 yyfilename = filename;
158 } // if
159 } // if
160}
161
162 /* ignore preprocessor directives (for now) */
163^{h_white}*"#"[^\n]*"\n" ;
164
165 /* ignore C style comments (ALSO HANDLED BY CPP) */
166"/*" { BEGIN COMMENT; }
167<COMMENT>.|\n ;
168<COMMENT>"*/" { BEGIN 0; }
169
170 /* ignore C++ style comments (ALSO HANDLED BY CPP) */
171"//"[^\n]*"\n" ;
172
173 /* ignore whitespace */
174{h_white}+ { WHITE_RETURN(' '); }
175({v_tab}|{c_return}|{form_feed})+ { WHITE_RETURN(' '); }
176({h_white}|{v_tab}|{c_return}|{form_feed})*"\n" { NEWLINE_RETURN(); }
177
178 /* keywords */
179_Alignas { KEYWORD_RETURN(ALIGNAS); } // C11
180_Alignof { KEYWORD_RETURN(ALIGNOF); } // C11
181__alignof { KEYWORD_RETURN(ALIGNOF); } // GCC
182__alignof__ { KEYWORD_RETURN(ALIGNOF); } // GCC
183asm { KEYWORD_RETURN(ASM); }
184__asm { KEYWORD_RETURN(ASM); } // GCC
185__asm__ { KEYWORD_RETURN(ASM); } // GCC
186_At { KEYWORD_RETURN(AT); } // CFA
187_Atomic { KEYWORD_RETURN(ATOMIC); } // C11
188__attribute { KEYWORD_RETURN(ATTRIBUTE); } // GCC
189__attribute__ { KEYWORD_RETURN(ATTRIBUTE); } // GCC
190auto { KEYWORD_RETURN(AUTO); }
191_Bool { KEYWORD_RETURN(BOOL); } // C99
192break { KEYWORD_RETURN(BREAK); }
193case { KEYWORD_RETURN(CASE); }
194catch { KEYWORD_RETURN(CATCH); } // CFA
195catchResume { KEYWORD_RETURN(CATCHRESUME); } // CFA
196char { KEYWORD_RETURN(CHAR); }
197choose { KEYWORD_RETURN(CHOOSE); } // CFA
198_Complex { KEYWORD_RETURN(COMPLEX); } // C99
199__complex { KEYWORD_RETURN(COMPLEX); } // GCC
200__complex__ { KEYWORD_RETURN(COMPLEX); } // GCC
201const { KEYWORD_RETURN(CONST); }
202__const { KEYWORD_RETURN(CONST); } // GCC
203__const__ { KEYWORD_RETURN(CONST); } // GCC
204continue { KEYWORD_RETURN(CONTINUE); }
205default { KEYWORD_RETURN(DEFAULT); }
206disable { KEYWORD_RETURN(DISABLE); } // CFA
207do { KEYWORD_RETURN(DO); }
208double { KEYWORD_RETURN(DOUBLE); }
209dtype { KEYWORD_RETURN(DTYPE); } // CFA
210else { KEYWORD_RETURN(ELSE); }
211enable { KEYWORD_RETURN(ENABLE); } // CFA
212enum { KEYWORD_RETURN(ENUM); }
213__extension__ { KEYWORD_RETURN(EXTENSION); } // GCC
214extern { KEYWORD_RETURN(EXTERN); }
215fallthru { KEYWORD_RETURN(FALLTHRU); } // CFA
216finally { KEYWORD_RETURN(FINALLY); } // CFA
217float { KEYWORD_RETURN(FLOAT); }
218__float128 { KEYWORD_RETURN(FLOAT); } // GCC
219for { KEYWORD_RETURN(FOR); }
220forall { KEYWORD_RETURN(FORALL); } // CFA
221fortran { KEYWORD_RETURN(FORTRAN); }
222ftype { KEYWORD_RETURN(FTYPE); } // CFA
223_Generic { KEYWORD_RETURN(GENERIC); } // C11
224goto { KEYWORD_RETURN(GOTO); }
225if { KEYWORD_RETURN(IF); }
226_Imaginary { KEYWORD_RETURN(IMAGINARY); } // C99
227__imag { KEYWORD_RETURN(IMAGINARY); } // GCC
228__imag__ { KEYWORD_RETURN(IMAGINARY); } // GCC
229inline { KEYWORD_RETURN(INLINE); } // C99
230__inline { KEYWORD_RETURN(INLINE); } // GCC
231__inline__ { KEYWORD_RETURN(INLINE); } // GCC
232int { KEYWORD_RETURN(INT); }
233__int128 { KEYWORD_RETURN(INT); } // GCC
234__label__ { KEYWORD_RETURN(LABEL); } // GCC
235long { KEYWORD_RETURN(LONG); }
236lvalue { KEYWORD_RETURN(LVALUE); } // CFA
237_Noreturn { KEYWORD_RETURN(NORETURN); } // C11
238__builtin_offsetof { KEYWORD_RETURN(OFFSETOF); } // GCC
239otype { KEYWORD_RETURN(OTYPE); } // CFA
240register { KEYWORD_RETURN(REGISTER); }
241restrict { KEYWORD_RETURN(RESTRICT); } // C99
242__restrict { KEYWORD_RETURN(RESTRICT); } // GCC
243__restrict__ { KEYWORD_RETURN(RESTRICT); } // GCC
244return { KEYWORD_RETURN(RETURN); }
245short { KEYWORD_RETURN(SHORT); }
246signed { KEYWORD_RETURN(SIGNED); }
247__signed { KEYWORD_RETURN(SIGNED); } // GCC
248__signed__ { KEYWORD_RETURN(SIGNED); } // GCC
249sizeof { KEYWORD_RETURN(SIZEOF); }
250static { KEYWORD_RETURN(STATIC); }
251_Static_assert { KEYWORD_RETURN(STATICASSERT); } // C11
252struct { KEYWORD_RETURN(STRUCT); }
253switch { KEYWORD_RETURN(SWITCH); }
254_Thread_local { KEYWORD_RETURN(THREADLOCAL); } // C11
255throw { KEYWORD_RETURN(THROW); } // CFA
256throwResume { KEYWORD_RETURN(THROWRESUME); } // CFA
257trait { KEYWORD_RETURN(TRAIT); } // CFA
258try { KEYWORD_RETURN(TRY); } // CFA
259typedef { KEYWORD_RETURN(TYPEDEF); }
260typeof { KEYWORD_RETURN(TYPEOF); } // GCC
261__typeof { KEYWORD_RETURN(TYPEOF); } // GCC
262__typeof__ { KEYWORD_RETURN(TYPEOF); } // GCC
263union { KEYWORD_RETURN(UNION); }
264unsigned { KEYWORD_RETURN(UNSIGNED); }
265__builtin_va_list { KEYWORD_RETURN(VALIST); } // GCC
266void { KEYWORD_RETURN(VOID); }
267volatile { KEYWORD_RETURN(VOLATILE); }
268__volatile { KEYWORD_RETURN(VOLATILE); } // GCC
269__volatile__ { KEYWORD_RETURN(VOLATILE); } // GCC
270while { KEYWORD_RETURN(WHILE); }
271
272 /* identifier */
273{identifier} { IDENTIFIER_RETURN(); }
274{attr_identifier} { ATTRIBUTE_RETURN(); }
275"`" { BEGIN BKQUOTE; }
276<BKQUOTE>{identifier} { IDENTIFIER_RETURN(); }
277<BKQUOTE>"`" { BEGIN 0; }
278
279 /* numeric constants */
280"0" { NUMERIC_RETURN(ZERO); } // CFA
281"1" { NUMERIC_RETURN(ONE); } // CFA
282{decimal_constant} { NUMERIC_RETURN(INTEGERconstant); }
283{octal_constant} { NUMERIC_RETURN(INTEGERconstant); }
284{hex_constant} { NUMERIC_RETURN(INTEGERconstant); }
285{floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
286{hex_floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
287
288 /* character constant, allows empty value */
289({cwide_prefix}[_]?)?['] { BEGIN QUOTE; rm_underscore(); strtext = new std::string; *strtext += std::string( yytext ); }
290<QUOTE>[^'\\\n]* { *strtext += std::string( yytext ); }
291<QUOTE>['\n] { BEGIN 0; *strtext += std::string( yytext); RETURN_STR(CHARACTERconstant); }
292 /* ' stop highlighting */
293
294 /* string constant */
295({swide_prefix}[_]?)?["] { BEGIN STRING; rm_underscore(); strtext = new std::string; *strtext += std::string( yytext ); }
296<STRING>[^"\\\n]* { *strtext += std::string( yytext ); }
297<STRING>["\n] { BEGIN 0; *strtext += std::string( yytext ); RETURN_STR(STRINGliteral); }
298 /* " stop highlighting */
299
300 /* common character/string constant */
301<QUOTE,STRING>{escape_seq} { rm_underscore(); *strtext += std::string( yytext ); }
302<QUOTE,STRING>"\\"{h_white}*"\n" {} // continuation (ALSO HANDLED BY CPP)
303<QUOTE,STRING>"\\" { *strtext += std::string( yytext ); } // unknown escape character
304
305 /* punctuation */
306"[" { ASCIIOP_RETURN(); }
307"]" { ASCIIOP_RETURN(); }
308"(" { ASCIIOP_RETURN(); }
309")" { ASCIIOP_RETURN(); }
310"{" { ASCIIOP_RETURN(); }
311"}" { ASCIIOP_RETURN(); }
312"," { ASCIIOP_RETURN(); } // also operator
313":" { ASCIIOP_RETURN(); }
314";" { ASCIIOP_RETURN(); }
315"." { ASCIIOP_RETURN(); } // also operator
316"..." { NAMEDOP_RETURN(ELLIPSIS); }
317
318 /* alternative C99 brackets, "<:" & "<:<:" handled by preprocessor */
319"<:" { RETURN_VAL('['); }
320":>" { RETURN_VAL(']'); }
321"<%" { RETURN_VAL('{'); }
322"%>" { RETURN_VAL('}'); }
323
324 /* operators */
325"!" { ASCIIOP_RETURN(); }
326"+" { ASCIIOP_RETURN(); }
327"-" { ASCIIOP_RETURN(); }
328"*" { ASCIIOP_RETURN(); }
329"/" { ASCIIOP_RETURN(); }
330"%" { ASCIIOP_RETURN(); }
331"^" { ASCIIOP_RETURN(); }
332"~" { ASCIIOP_RETURN(); }
333"&" { ASCIIOP_RETURN(); }
334"|" { ASCIIOP_RETURN(); }
335"<" { ASCIIOP_RETURN(); }
336">" { ASCIIOP_RETURN(); }
337"=" { ASCIIOP_RETURN(); }
338"?" { ASCIIOP_RETURN(); }
339
340"++" { NAMEDOP_RETURN(ICR); }
341"--" { NAMEDOP_RETURN(DECR); }
342"==" { NAMEDOP_RETURN(EQ); }
343"!=" { NAMEDOP_RETURN(NE); }
344"<<" { NAMEDOP_RETURN(LS); }
345">>" { NAMEDOP_RETURN(RS); }
346"<=" { NAMEDOP_RETURN(LE); }
347">=" { NAMEDOP_RETURN(GE); }
348"&&" { NAMEDOP_RETURN(ANDAND); }
349"||" { NAMEDOP_RETURN(OROR); }
350"->" { NAMEDOP_RETURN(ARROW); }
351"+=" { NAMEDOP_RETURN(PLUSassign); }
352"-=" { NAMEDOP_RETURN(MINUSassign); }
353"*=" { NAMEDOP_RETURN(MULTassign); }
354"/=" { NAMEDOP_RETURN(DIVassign); }
355"%=" { NAMEDOP_RETURN(MODassign); }
356"&=" { NAMEDOP_RETURN(ANDassign); }
357"|=" { NAMEDOP_RETURN(ORassign); }
358"^=" { NAMEDOP_RETURN(ERassign); }
359"<<=" { NAMEDOP_RETURN(LSassign); }
360">>=" { NAMEDOP_RETURN(RSassign); }
361
362"@=" { NAMEDOP_RETURN(ATassign); }
363
364 /* CFA, operator identifier */
365{op_unary}"?" { IDENTIFIER_RETURN(); } // unary
366"?"({op_unary_pre_post}|"()"|"[?]"|"{}") { IDENTIFIER_RETURN(); }
367"^?{}" { IDENTIFIER_RETURN(); }
368"?"{op_binary_over}"?" { IDENTIFIER_RETURN(); } // binary
369 /*
370 This rule handles ambiguous cases with operator identifiers, e.g., "int *?*?()", where the string "*?*?"
371 can be lexed as "*"/"?*?" or "*?"/"*?". Since it is common practise to put a unary operator juxtaposed
372 to an identifier, e.g., "*i", users will be annoyed if they cannot do this with respect to operator
373 identifiers. Even with this special hack, there are 5 general cases that cannot be handled. The first
374 case is for the function-call identifier "?()":
375
376 int * ?()(); // declaration: space required after '*'
377 * ?()(); // expression: space required after '*'
378
379 Without the space, the string "*?()" is ambiguous without N character look ahead; it requires scanning
380 ahead to determine if there is a '(', which is the start of an argument/parameter list.
381
382 The 4 remaining cases occur in expressions:
383
384 i++?i:0; // space required before '?'
385 i--?i:0; // space required before '?'
386 i?++i:0; // space required after '?'
387 i?--i:0; // space required after '?'
388
389 In the first two cases, the string "i++?" is ambiguous, where this string can be lexed as "i"/"++?" or
390 "i++"/"?"; it requires scanning ahead to determine if there is a '(', which is the start of an argument
391 list. In the second two cases, the string "?++x" is ambiguous, where this string can be lexed as
392 "?++"/"x" or "?"/"++x"; it requires scanning ahead to determine if there is a '(', which is the start of
393 an argument list.
394 */
395{op_unary}"?"({op_unary_pre_post}|"[?]"|{op_binary_over}"?") {
396 // 1 or 2 character unary operator ?
397 int i = yytext[1] == '?' ? 1 : 2;
398 yyless( i ); // put back characters up to first '?'
399 if ( i > 1 ) {
400 NAMEDOP_RETURN( yytext[0] == '+' ? ICR : DECR );
401 } else {
402 ASCIIOP_RETURN();
403 } // if
404}
405
406 /* unknown characters */
407. { printf("unknown character(s):\"%s\" on line %d\n", yytext, yylineno); }
408
409%%
410
411// Local Variables: //
412// mode: c++ //
413// tab-width: 4 //
414// compile-command: "make install" //
415// End: //
Note: See TracBrowser for help on using the repository browser.