source: src/Parser/lex.ll@ cd623a4

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 string with_gc
Last change on this file since cd623a4 was cd623a4, checked in by Peter A. Buhr <pabuhr@…>, 10 years ago

constant types, second attempt
  • Property mode set to 100644
File size: 15.0 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 : Sun Jun 7 07:14:16 2015
13 * Update Count : 374
14 */
15
16%option yylineno
17
18%{
19// This lexer assumes the program has been preprocessed by cpp. Hence, all user level preprocessor directive
20// have been performed and removed from the source. The only exceptions are preprocessor directives passed to
21// the compiler (e.g., line-number directives) and C/C++ style comments, which are ignored.
22
23//**************************** Includes and Defines ****************************
24
25#include <string>
26
27#include "lex.h"
28#include "ParseNode.h"
29#include "parser.h" // YACC generated definitions based on C++ grammar
30
31char *yyfilename;
32std::string *strtext; // accumulate parts of character and string constant value
33
34#define WHITE_RETURN(x) // do nothing
35#define NEWLINE_RETURN() WHITE_RETURN('\n')
36#define RETURN_VAL(x) yylval.tok.str = new std::string( yytext ); \
37 yylval.tok.loc.file = yyfilename; \
38 yylval.tok.loc.line = yylineno; \
39 return(x)
40#define RETURN_STR(x) yylval.tok.str = strtext; \
41 yylval.tok.loc.file = yyfilename; \
42 yylval.tok.loc.line = yylineno; \
43 return(x)
44
45#define KEYWORD_RETURN(x) RETURN_VAL(x) // keyword
46#define IDENTIFIER_RETURN() RETURN_VAL((typedefTable.isIdentifier(yytext) ? IDENTIFIER : typedefTable.isTypedef(yytext) ? TYPEDEFname : TYPEGENname))
47//#define ATTRIBUTE_RETURN() RETURN_VAL((typedefTable.isIdentifier(yytext) ? ATTR_IDENTIFIER : typedefTable.isTypedef(yytext) ? ATTR_TYPEDEFname : ATTR_TYPEGENname))
48#define ATTRIBUTE_RETURN() RETURN_VAL(ATTR_IDENTIFIER)
49
50#define ASCIIOP_RETURN() RETURN_VAL((int)yytext[0]) // single character operator
51#define NAMEDOP_RETURN(x) RETURN_VAL(x) // multichar operator, with a name
52
53#define NUMERIC_RETURN(x) rm_underscore(); RETURN_VAL(x) // numeric constant
54
55void rm_underscore() {
56 // remove underscores in numeric constant
57 int j = 0;
58 for ( int i = 0; yytext[i] != '\0'; i += 1 ) {
59 if ( yytext[i] != '_' ) {
60 yytext[j] = yytext[i];
61 j += 1;
62 } // if
63 } // for
64 yyleng = j;
65 yytext[yyleng] = '\0';
66}
67
68%}
69
70octal [0-7]
71nonzero [1-9]
72decimal [0-9]
73hex [0-9a-fA-F]
74universal_char "\\"((u"_"?{hex_quad})|(U"_"?{hex_quad}{2}))
75
76 // identifier, GCC: $ in identifier
77identifier ([a-zA-Z_$]|{universal_char})([0-9a-zA-Z_$]|{universal_char})*
78
79 // quoted identifier
80quoted_identifier "`"{identifier}"`"
81
82 // attribute identifier, GCC: $ in identifier
83attr_identifier "@"{identifier}
84
85 // numeric constants, CFA: '_' in constant
86hex_quad {hex}("_"?{hex}){3}
87integer_suffix "_"?(([uU][lL]?)|([uU]("ll"|"LL")?)|([lL][uU]?)|("ll"|"LL")[uU]?)
88
89octal_digits ({octal})|({octal}({octal}|"_")*{octal})
90octal_prefix "0""_"?
91octal_constant (("0")|({octal_prefix}{octal_digits})){integer_suffix}?
92
93nonzero_digits ({nonzero})|({nonzero}({decimal}|"_")*{decimal})
94decimal_constant {nonzero_digits}{integer_suffix}?
95
96hex_digits ({hex})|({hex}({hex}|"_")*{hex})
97hex_prefix "0"[xX]"_"?
98hex_constant {hex_prefix}{hex_digits}{integer_suffix}?
99
100decimal_digits ({decimal})|({decimal}({decimal}|"_")*{decimal})
101fractional_constant ({decimal_digits}?"."{decimal_digits})|({decimal_digits}".")
102exponent "_"?[eE]"_"?[+-]?{decimal_digits}
103floating_suffix "_"?[flFL]
104floating_constant (({fractional_constant}{exponent}?)|({decimal_digits}{exponent})){floating_suffix}?
105
106binary_exponent "_"?[pP]"_"?[+-]?{decimal_digits}
107hex_fractional_constant ({hex_digits}?"."{hex_digits})|({hex_digits}".")
108hex_floating_constant {hex_prefix}(({hex_fractional_constant}{binary_exponent})|({hex_digits}{binary_exponent})){floating_suffix}?
109
110 // character escape sequence, GCC: \e => esc character
111simple_escape "\\"[abefnrtv'"?\\]
112 // ' stop highlighting
113octal_escape "\\"{octal}("_"?{octal}){0,2}
114hex_escape "\\""x""_"?{hex_digits}
115escape_seq {simple_escape}|{octal_escape}|{hex_escape}|{universal_char}
116cwide_prefix "L"|"U"|"u"
117swide_prefix {cwide_prefix}|"u8"
118
119 // display/white-space characters
120h_tab [\011]
121form_feed [\014]
122v_tab [\013]
123c_return [\015]
124h_white [ ]|{h_tab}
125
126 // operators
127op_unary_only "~"|"!"
128op_unary_binary "+"|"-"|"*"
129op_unary_pre_post "++"|"--"
130op_unary {op_unary_only}|{op_unary_binary}|{op_unary_pre_post}
131
132op_binary_only "/"|"%"|"^"|"&"|"|"|"<"|">"|"="|"=="|"!="|"<<"|">>"|"<="|">="|"+="|"-="|"*="|"/="|"%="|"&="|"|="|"^="|"<<="|">>="
133op_binary_over {op_unary_binary}|{op_binary_only}
134op_binary_not_over "?"|"->"|"&&"|"||"
135operator {op_unary_pre_post}|{op_binary_over}|{op_binary_not_over}
136
137%x COMMENT
138%x BKQUOTE
139%x QUOTE
140%x STRING
141
142%%
143 /* line directives */
144^{h_white}*"#"{h_white}*[0-9]+{h_white}*["][^"\n]+["][^\n]*"\n" {
145 /* " stop highlighting */
146 char *end_num;
147 char *begin_string, *end_string;
148 char *filename;
149 long lineno, length;
150 lineno = strtol( yytext + 1, &end_num, 0 );
151 begin_string = strchr( end_num, '"' );
152 if ( begin_string ) {
153 end_string = strchr( begin_string + 1, '"' );
154 if ( end_string ) {
155 length = end_string - begin_string - 1;
156 filename = new char[ length + 1 ];
157 memcpy( filename, begin_string + 1, length );
158 filename[ length ] = '\0';
159 //std::cout << "file " << filename << " line " << lineno << std::endl;
160 yylineno = lineno;
161 yyfilename = filename;
162 } // if
163 } // if
164}
165
166 /* ignore preprocessor directives (for now) */
167^{h_white}*"#"[^\n]*"\n" ;
168
169 /* ignore C style comments (ALSO HANDLED BY CPP) */
170"/*" { BEGIN COMMENT; }
171<COMMENT>.|\n ;
172<COMMENT>"*/" { BEGIN 0; }
173
174 /* ignore C++ style comments (ALSO HANDLED BY CPP) */
175"//"[^\n]*"\n" ;
176
177 /* ignore whitespace */
178{h_white}+ { WHITE_RETURN(' '); }
179({v_tab}|{c_return}|{form_feed})+ { WHITE_RETURN(' '); }
180({h_white}|{v_tab}|{c_return}|{form_feed})*"\n" { NEWLINE_RETURN(); }
181
182 /* keywords */
183_Alignas { KEYWORD_RETURN(ALIGNAS); } // C11
184_Alignof { KEYWORD_RETURN(ALIGNOF); } // C11
185__alignof { KEYWORD_RETURN(ALIGNOF); } // GCC
186__alignof__ { KEYWORD_RETURN(ALIGNOF); } // GCC
187asm { KEYWORD_RETURN(ASM); }
188__asm { KEYWORD_RETURN(ASM); } // GCC
189__asm__ { KEYWORD_RETURN(ASM); } // GCC
190_Atomic { KEYWORD_RETURN(ATOMIC); } // C11
191__attribute { KEYWORD_RETURN(ATTRIBUTE); } // GCC
192__attribute__ { KEYWORD_RETURN(ATTRIBUTE); } // GCC
193auto { KEYWORD_RETURN(AUTO); }
194_Bool { KEYWORD_RETURN(BOOL); } // C99
195break { KEYWORD_RETURN(BREAK); }
196case { KEYWORD_RETURN(CASE); }
197catch { KEYWORD_RETURN(CATCH); } // CFA
198char { KEYWORD_RETURN(CHAR); }
199choose { KEYWORD_RETURN(CHOOSE); } // CFA
200_Complex { KEYWORD_RETURN(COMPLEX); } // C99
201__complex { KEYWORD_RETURN(COMPLEX); } // GCC
202__complex__ { KEYWORD_RETURN(COMPLEX); } // GCC
203const { KEYWORD_RETURN(CONST); }
204__const { KEYWORD_RETURN(CONST); } // GCC
205__const__ { KEYWORD_RETURN(CONST); } // GCC
206context { KEYWORD_RETURN(CONTEXT); } // CFA
207continue { KEYWORD_RETURN(CONTINUE); }
208default { KEYWORD_RETURN(DEFAULT); }
209do { KEYWORD_RETURN(DO); }
210double { KEYWORD_RETURN(DOUBLE); }
211dtype { KEYWORD_RETURN(DTYPE); } // CFA
212else { KEYWORD_RETURN(ELSE); }
213enum { KEYWORD_RETURN(ENUM); }
214__extension__ { KEYWORD_RETURN(EXTENSION); } // GCC
215extern { KEYWORD_RETURN(EXTERN); }
216fallthru { KEYWORD_RETURN(FALLTHRU); } // CFA
217finally { KEYWORD_RETURN(FINALLY); } // CFA
218float { KEYWORD_RETURN(FLOAT); }
219__float128 { KEYWORD_RETURN(FLOAT); } // GCC
220for { KEYWORD_RETURN(FOR); }
221forall { KEYWORD_RETURN(FORALL); } // CFA
222fortran { KEYWORD_RETURN(FORTRAN); }
223ftype { KEYWORD_RETURN(FTYPE); } // CFA
224_Generic { KEYWORD_RETURN(GENERIC); } // C11
225goto { KEYWORD_RETURN(GOTO); }
226if { KEYWORD_RETURN(IF); }
227_Imaginary { KEYWORD_RETURN(IMAGINARY); } // C99
228__imag { KEYWORD_RETURN(IMAGINARY); } // GCC
229__imag__ { KEYWORD_RETURN(IMAGINARY); } // GCC
230inline { KEYWORD_RETURN(INLINE); } // C99
231__inline { KEYWORD_RETURN(INLINE); } // GCC
232__inline__ { KEYWORD_RETURN(INLINE); } // GCC
233int { KEYWORD_RETURN(INT); }
234__int128 { KEYWORD_RETURN(INT); } // GCC
235__label__ { KEYWORD_RETURN(LABEL); } // GCC
236long { KEYWORD_RETURN(LONG); }
237lvalue { KEYWORD_RETURN(LVALUE); } // CFA
238_Noreturn { KEYWORD_RETURN(NORETURN); } // C11
239register { KEYWORD_RETURN(REGISTER); }
240restrict { KEYWORD_RETURN(RESTRICT); } // C99
241__restrict { KEYWORD_RETURN(RESTRICT); } // GCC
242__restrict__ { KEYWORD_RETURN(RESTRICT); } // GCC
243return { KEYWORD_RETURN(RETURN); }
244short { KEYWORD_RETURN(SHORT); }
245signed { KEYWORD_RETURN(SIGNED); }
246__signed { KEYWORD_RETURN(SIGNED); } // GCC
247__signed__ { KEYWORD_RETURN(SIGNED); } // GCC
248sizeof { KEYWORD_RETURN(SIZEOF); }
249static { KEYWORD_RETURN(STATIC); }
250_Static_assert { KEYWORD_RETURN(STATICASSERT); } // C11
251struct { KEYWORD_RETURN(STRUCT); }
252switch { KEYWORD_RETURN(SWITCH); }
253_Thread_local { KEYWORD_RETURN(THREADLOCAL); } // C11
254throw { KEYWORD_RETURN(THROW); } // CFA
255try { KEYWORD_RETURN(TRY); } // CFA
256type { KEYWORD_RETURN(TYPE); } // CFA
257typedef { KEYWORD_RETURN(TYPEDEF); }
258typeof { KEYWORD_RETURN(TYPEOF); } // GCC
259__typeof { KEYWORD_RETURN(TYPEOF); } // GCC
260__typeof__ { KEYWORD_RETURN(TYPEOF); } // GCC
261union { KEYWORD_RETURN(UNION); }
262unsigned { KEYWORD_RETURN(UNSIGNED); }
263void { KEYWORD_RETURN(VOID); }
264volatile { KEYWORD_RETURN(VOLATILE); }
265__volatile { KEYWORD_RETURN(VOLATILE); } // GCC
266__volatile__ { KEYWORD_RETURN(VOLATILE); } // GCC
267while { KEYWORD_RETURN(WHILE); }
268
269 /* identifier */
270{identifier} { IDENTIFIER_RETURN(); }
271{attr_identifier} { ATTRIBUTE_RETURN(); }
272"`" { BEGIN BKQUOTE; }
273<BKQUOTE>{identifier} { IDENTIFIER_RETURN(); }
274<BKQUOTE>"`" { BEGIN 0; }
275
276 /* numeric constants */
277"0" { NUMERIC_RETURN(ZERO); } // CFA
278"1" { NUMERIC_RETURN(ONE); } // CFA
279{decimal_constant} { NUMERIC_RETURN(INTEGERconstant); }
280{octal_constant} { NUMERIC_RETURN(INTEGERconstant); }
281{hex_constant} { NUMERIC_RETURN(INTEGERconstant); }
282{floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
283{hex_floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
284
285 /* character constant, allows empty value */
286({cwide_prefix}[_]?)?['] { BEGIN QUOTE; rm_underscore(); strtext = new std::string; *strtext += std::string( yytext ); }
287<QUOTE>[^'\\\n]* { *strtext += std::string( yytext ); }
288<QUOTE>['\n] { BEGIN 0; *strtext += std::string( yytext); RETURN_STR(CHARACTERconstant); }
289 /* ' stop highlighting */
290
291 /* string constant */
292({swide_prefix}[_]?)?["] { BEGIN STRING; rm_underscore(); strtext = new std::string; *strtext += std::string( yytext ); }
293<STRING>[^"\\\n]* { *strtext += std::string( yytext ); }
294<STRING>["\n] { BEGIN 0; *strtext += std::string( yytext ); RETURN_STR(STRINGliteral); }
295 /* " stop highlighting */
296
297 /* common character/string constant */
298<QUOTE,STRING>{escape_seq} { rm_underscore(); *strtext += std::string( yytext ); }
299<QUOTE,STRING>"\\"{h_white}*"\n" {} // continuation (ALSO HANDLED BY CPP)
300<QUOTE,STRING>"\\" { *strtext += std::string( yytext ); } // unknown escape character
301
302 /* punctuation */
303"[" { ASCIIOP_RETURN(); }
304"]" { ASCIIOP_RETURN(); }
305"(" { ASCIIOP_RETURN(); }
306")" { ASCIIOP_RETURN(); }
307"{" { ASCIIOP_RETURN(); }
308"}" { ASCIIOP_RETURN(); }
309"," { ASCIIOP_RETURN(); } // also operator
310":" { ASCIIOP_RETURN(); }
311";" { ASCIIOP_RETURN(); }
312"." { ASCIIOP_RETURN(); } // also operator
313"..." { NAMEDOP_RETURN(ELLIPSIS); }
314
315 /* alternative C99 brackets, "<:" & "<:<:" handled by preprocessor */
316"<:" { RETURN_VAL('['); }
317":>" { RETURN_VAL(']'); }
318"<%" { RETURN_VAL('{'); }
319"%>" { RETURN_VAL('}'); }
320
321 /* operators */
322"!" { ASCIIOP_RETURN(); }
323"+" { ASCIIOP_RETURN(); }
324"-" { ASCIIOP_RETURN(); }
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
337"++" { NAMEDOP_RETURN(ICR); }
338"--" { NAMEDOP_RETURN(DECR); }
339"==" { NAMEDOP_RETURN(EQ); }
340"!=" { NAMEDOP_RETURN(NE); }
341"<<" { NAMEDOP_RETURN(LS); }
342">>" { NAMEDOP_RETURN(RS); }
343"<=" { NAMEDOP_RETURN(LE); }
344">=" { NAMEDOP_RETURN(GE); }
345"&&" { NAMEDOP_RETURN(ANDAND); }
346"||" { NAMEDOP_RETURN(OROR); }
347"->" { NAMEDOP_RETURN(ARROW); }
348"+=" { NAMEDOP_RETURN(PLUSassign); }
349"-=" { NAMEDOP_RETURN(MINUSassign); }
350"*=" { NAMEDOP_RETURN(MULTassign); }
351"/=" { NAMEDOP_RETURN(DIVassign); }
352"%=" { NAMEDOP_RETURN(MODassign); }
353"&=" { NAMEDOP_RETURN(ANDassign); }
354"|=" { NAMEDOP_RETURN(ORassign); }
355"^=" { NAMEDOP_RETURN(ERassign); }
356"<<=" { NAMEDOP_RETURN(LSassign); }
357">>=" { NAMEDOP_RETURN(RSassign); }
358
359 /* CFA, operator identifier */
360{op_unary}"?" { IDENTIFIER_RETURN(); } // unary
361"?"({op_unary_pre_post}|"()"|"[?]"|"{}") { IDENTIFIER_RETURN(); }
362"?"{op_binary_over}"?" { IDENTIFIER_RETURN(); } // binary
363 /*
364 This rule handles ambiguous cases with operator identifiers, e.g., "int *?*?()", where the string "*?*?"
365 can be lexed as "*"/"?*?" or "*?"/"*?". Since it is common practise to put a unary operator juxtaposed
366 to an identifier, e.g., "*i", users will be annoyed if they cannot do this with respect to operator
367 identifiers. Even with this special hack, there are 5 general cases that cannot be handled. The first
368 case is for the function-call identifier "?()":
369
370 int * ?()(); // declaration: space required after '*'
371 * ?()(); // expression: space required after '*'
372
373 Without the space, the string "*?()" is ambiguous without N character look ahead; it requires scanning
374 ahead to determine if there is a '(', which is the start of an argument/parameter list.
375
376 The 4 remaining cases occur in expressions:
377
378 i++?i:0; // space required before '?'
379 i--?i:0; // space required before '?'
380 i?++i:0; // space required after '?'
381 i?--i:0; // space required after '?'
382
383 In the first two cases, the string "i++?" is ambiguous, where this string can be lexed as "i"/"++?" or
384 "i++"/"?"; it requires scanning ahead to determine if there is a '(', which is the start of an argument
385 list. In the second two cases, the string "?++x" is ambiguous, where this string can be lexed as
386 "?++"/"x" or "?"/"++x"; it requires scanning ahead to determine if there is a '(', which is the start of
387 an argument list.
388 */
389{op_unary}"?"(({op_unary_pre_post}|"[?]")|({op_binary_over}"?")) {
390 // 1 or 2 character unary operator ?
391 int i = yytext[1] == '?' ? 1 : 2;
392 yyless( i ); // put back characters up to first '?'
393 if ( i > 1 ) {
394 NAMEDOP_RETURN( yytext[0] == '+' ? ICR : DECR );
395 } else {
396 ASCIIOP_RETURN();
397 } // if
398}
399
400 /* unknown characters */
401. { printf("unknown character(s):\"%s\" on line %d\n", yytext, yylineno); }
402
403%%
404
405// Local Variables: //
406// fill-column: 110 //
407// tab-width: 4 //
408// mode: c++ //
409// compile-command: "make install" //
410// End: //
Note: See TracBrowser for help on using the repository browser.