source: src/Parser/lex.ll@ 9f10c4b8

ADT aaron-thesis arm-eh ast-experimental cleanup-dtors deferred_resn demangler enum forall-pointer-decay jacob/cs343-translation jenkins-sandbox 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 9f10c4b8 was 9236060, checked in by Rob Schluntz <rschlunt@…>, 8 years ago

Merge branch 'master' into references

  • Property mode set to 100644
File size: 16.5 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.ll --
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 Jul 27 21:46:06 2017
13 * Update Count : 550
14 */
15
16%option yylineno
17%option noyywrap
18%option nounput
19
20%{
21// This lexer assumes the program has been preprocessed by cpp. Hence, all user level preprocessor directive have been
22// performed and removed from the source. The only exceptions are preprocessor directives passed to the compiler (e.g.,
23// line-number directives) and C/C++ style comments, which are ignored.
24
25//**************************** Includes and Defines ****************************
26
27#include <string>
28#include <cstdio> // FILENAME_MAX
29
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 = nullptr; 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_CHAR( 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// Stop warning due to incorrectly generated flex code.
63#pragma GCC diagnostic ignored "-Wsign-compare"
64%}
65
66octal [0-7]
67nonzero [1-9]
68decimal [0-9]
69hex [0-9a-fA-F]
70universal_char "\\"((u"_"?{hex_quad})|(U"_"?{hex_quad}{2}))
71
72 // identifier, GCC: $ in identifier
73identifier ([a-zA-Z_$]|{universal_char})([0-9a-zA-Z_$]|{universal_char})*
74
75 // attribute identifier, GCC: $ in identifier
76attr_identifier "@"{identifier}
77
78 // numeric constants, CFA: '_' in constant
79hex_quad {hex}("_"?{hex}){3}
80integer_suffix "_"?(([uU](("ll"|"LL"|[lL])[iI]|[iI]?("ll"|"LL"|[lL])?))|([iI](("ll"|"LL"|[lL])[uU]|[uU]?("ll"|"LL"|[lL])?))|(("ll"|"LL"|[lL])([iI][uU]|[uU]?[iI]?)))
81
82octal_digits ({octal})|({octal}({octal}|"_")*{octal})
83octal_prefix "0""_"?
84octal_constant (("0")|({octal_prefix}{octal_digits})){integer_suffix}?
85
86nonzero_digits ({nonzero})|({nonzero}({decimal}|"_")*{decimal})
87decimal_constant {nonzero_digits}{integer_suffix}?
88
89hex_digits ({hex})|({hex}({hex}|"_")*{hex})
90hex_prefix "0"[xX]"_"?
91hex_constant {hex_prefix}{hex_digits}{integer_suffix}?
92
93decimal_digits ({decimal})|({decimal}({decimal}|"_")*{decimal})
94real_decimal {decimal_digits}"."{exponent}?{floating_suffix}?
95real_fraction "."{decimal_digits}{exponent}?{floating_suffix}?
96real_constant {decimal_digits}{real_fraction}
97exponent "_"?[eE]"_"?[+-]?{decimal_digits}
98 // GCC: D (double) and iI (imaginary) suffixes, and DL (long double)
99floating_suffix "_"?([fFdDlL][iI]?|[iI][lLfFdD]?|"DL")
100floating_constant (({real_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 static char filename[FILENAME_MAX]; // temporarily store current source-file name
143 char *end_num;
144 char *begin_string, *end_string;
145 long lineno, length;
146 lineno = strtol( yytext + 1, &end_num, 0 );
147 begin_string = strchr( end_num, '"' );
148 if ( begin_string ) { // file name ?
149 end_string = strchr( begin_string + 1, '"' ); // look for ending delimiter
150 assert( end_string ); // closing quote ?
151 length = end_string - begin_string - 1; // file-name length without quotes or sentinel
152 assert( length < FILENAME_MAX ); // room for sentinel ?
153 memcpy( &filename, begin_string + 1, length ); // copy file name from yytext
154 filename[ length ] = '\0'; // terminate string with sentinel
155 //std::cout << "file " << filename << " line " << lineno << std::endl;
156 yylineno = lineno;
157 yyfilename = filename;
158 } // if
159}
160
161 /* ignore preprocessor directives (for now) */
162^{h_white}*"#"[^\n]*"\n" ;
163
164 /* ignore C style comments (ALSO HANDLED BY CPP) */
165"/*" { BEGIN COMMENT; }
166<COMMENT>.|\n ;
167<COMMENT>"*/" { BEGIN 0; }
168
169 /* ignore C++ style comments (ALSO HANDLED BY CPP) */
170"//"[^\n]*"\n" ;
171
172 /* ignore whitespace */
173{h_white}+ { WHITE_RETURN(' '); }
174({v_tab}|{c_return}|{form_feed})+ { WHITE_RETURN(' '); }
175({h_white}|{v_tab}|{c_return}|{form_feed})*"\n" { NEWLINE_RETURN(); }
176
177 /* keywords */
178_Alignas { KEYWORD_RETURN(ALIGNAS); } // C11
179_Alignof { KEYWORD_RETURN(ALIGNOF); } // C11
180__alignof { KEYWORD_RETURN(ALIGNOF); } // GCC
181__alignof__ { KEYWORD_RETURN(ALIGNOF); } // GCC
182asm { KEYWORD_RETURN(ASM); }
183__asm { KEYWORD_RETURN(ASM); } // GCC
184__asm__ { KEYWORD_RETURN(ASM); } // GCC
185_At { KEYWORD_RETURN(AT); } // CFA
186_Atomic { KEYWORD_RETURN(ATOMIC); } // C11
187__attribute { KEYWORD_RETURN(ATTRIBUTE); } // GCC
188__attribute__ { KEYWORD_RETURN(ATTRIBUTE); } // GCC
189auto { KEYWORD_RETURN(AUTO); }
190_Bool { KEYWORD_RETURN(BOOL); } // C99
191break { KEYWORD_RETURN(BREAK); }
192case { KEYWORD_RETURN(CASE); }
193catch { KEYWORD_RETURN(CATCH); } // CFA
194catchResume { KEYWORD_RETURN(CATCHRESUME); } // CFA
195char { KEYWORD_RETURN(CHAR); }
196choose { KEYWORD_RETURN(CHOOSE); } // CFA
197_Complex { KEYWORD_RETURN(COMPLEX); } // C99
198__complex { KEYWORD_RETURN(COMPLEX); } // GCC
199__complex__ { KEYWORD_RETURN(COMPLEX); } // GCC
200const { KEYWORD_RETURN(CONST); }
201__const { KEYWORD_RETURN(CONST); } // GCC
202__const__ { KEYWORD_RETURN(CONST); } // GCC
203continue { KEYWORD_RETURN(CONTINUE); }
204coroutine { KEYWORD_RETURN(COROUTINE); } // CFA
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); }
215fallthrough { KEYWORD_RETURN(FALLTHRU); } // CFA
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__int128_t { KEYWORD_RETURN(INT); } // GCC
236__label__ { KEYWORD_RETURN(LABEL); } // GCC
237long { KEYWORD_RETURN(LONG); }
238monitor { KEYWORD_RETURN(MONITOR); } // CFA
239mutex { KEYWORD_RETURN(MUTEX); } // CFA
240_Noreturn { KEYWORD_RETURN(NORETURN); } // C11
241__builtin_offsetof { KEYWORD_RETURN(OFFSETOF); } // GCC
242one_t { NUMERIC_RETURN(ONE_T); } // CFA
243otype { KEYWORD_RETURN(OTYPE); } // CFA
244register { KEYWORD_RETURN(REGISTER); }
245restrict { KEYWORD_RETURN(RESTRICT); } // C99
246__restrict { KEYWORD_RETURN(RESTRICT); } // GCC
247__restrict__ { KEYWORD_RETURN(RESTRICT); } // GCC
248return { KEYWORD_RETURN(RETURN); }
249short { KEYWORD_RETURN(SHORT); }
250signed { KEYWORD_RETURN(SIGNED); }
251__signed { KEYWORD_RETURN(SIGNED); } // GCC
252__signed__ { KEYWORD_RETURN(SIGNED); } // GCC
253sizeof { KEYWORD_RETURN(SIZEOF); }
254static { KEYWORD_RETURN(STATIC); }
255_Static_assert { KEYWORD_RETURN(STATICASSERT); } // C11
256struct { KEYWORD_RETURN(STRUCT); }
257switch { KEYWORD_RETURN(SWITCH); }
258thread { KEYWORD_RETURN(THREAD); } // C11
259_Thread_local { KEYWORD_RETURN(THREADLOCAL); } // C11
260throw { KEYWORD_RETURN(THROW); } // CFA
261throwResume { KEYWORD_RETURN(THROWRESUME); } // CFA
262trait { KEYWORD_RETURN(TRAIT); } // CFA
263try { KEYWORD_RETURN(TRY); } // CFA
264ttype { KEYWORD_RETURN(TTYPE); } // CFA
265typedef { KEYWORD_RETURN(TYPEDEF); }
266typeof { KEYWORD_RETURN(TYPEOF); } // GCC
267__typeof { KEYWORD_RETURN(TYPEOF); } // GCC
268__typeof__ { KEYWORD_RETURN(TYPEOF); } // GCC
269__uint128_t { KEYWORD_RETURN(INT); } // GCC
270union { KEYWORD_RETURN(UNION); }
271unsigned { KEYWORD_RETURN(UNSIGNED); }
272__builtin_va_list { KEYWORD_RETURN(VALIST); } // GCC
273virtual { KEYWORD_RETURN(VIRTUAL); } // CFA
274void { KEYWORD_RETURN(VOID); }
275volatile { KEYWORD_RETURN(VOLATILE); }
276__volatile { KEYWORD_RETURN(VOLATILE); } // GCC
277__volatile__ { KEYWORD_RETURN(VOLATILE); } // GCC
278while { KEYWORD_RETURN(WHILE); }
279with { KEYWORD_RETURN(WITH); } // CFA
280zero_t { NUMERIC_RETURN(ZERO_T); } // CFA
281
282 /* identifier */
283{identifier} { IDENTIFIER_RETURN(); }
284{attr_identifier} { ATTRIBUTE_RETURN(); }
285"`" { BEGIN BKQUOTE; }
286<BKQUOTE>{identifier} { IDENTIFIER_RETURN(); }
287<BKQUOTE>"`" { BEGIN 0; }
288
289 /* numeric constants */
290{decimal_constant} { NUMERIC_RETURN(INTEGERconstant); }
291{octal_constant} { NUMERIC_RETURN(INTEGERconstant); }
292{hex_constant} { NUMERIC_RETURN(INTEGERconstant); }
293{real_decimal} { NUMERIC_RETURN(REALDECIMALconstant); } // must appear before floating_constant
294{real_fraction} { NUMERIC_RETURN(REALFRACTIONconstant); } // must appear before floating_constant
295{floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
296{hex_floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
297
298 /* character constant, allows empty value */
299({cwide_prefix}[_]?)?['] { BEGIN QUOTE; rm_underscore(); strtext = new std::string( yytext, yyleng ); }
300<QUOTE>[^'\\\n]* { strtext->append( yytext, yyleng ); }
301<QUOTE>['\n] { BEGIN 0; strtext->append( yytext, yyleng ); RETURN_STR(CHARACTERconstant); }
302 /* ' stop highlighting */
303
304 /* string constant */
305({swide_prefix}[_]?)?["] { BEGIN STRING; rm_underscore(); strtext = new std::string( yytext, yyleng ); }
306<STRING>[^"\\\n]* { strtext->append( yytext, yyleng ); }
307<STRING>["\n] { BEGIN 0; strtext->append( yytext, yyleng ); RETURN_STR(STRINGliteral); }
308 /* " stop highlighting */
309
310 /* common character/string constant */
311<QUOTE,STRING>{escape_seq} { rm_underscore(); strtext->append( yytext, yyleng ); }
312<QUOTE,STRING>"\\"{h_white}*"\n" {} // continuation (ALSO HANDLED BY CPP)
313<QUOTE,STRING>"\\" { strtext->append( yytext, yyleng ); } // unknown escape character
314
315 /* punctuation */
316"@" { ASCIIOP_RETURN(); }
317"[" { ASCIIOP_RETURN(); }
318"]" { ASCIIOP_RETURN(); }
319"(" { ASCIIOP_RETURN(); }
320")" { ASCIIOP_RETURN(); }
321"{" { ASCIIOP_RETURN(); }
322"}" { ASCIIOP_RETURN(); }
323"," { ASCIIOP_RETURN(); } // also operator
324":" { ASCIIOP_RETURN(); }
325";" { ASCIIOP_RETURN(); }
326"." { ASCIIOP_RETURN(); } // also operator
327"..." { NAMEDOP_RETURN(ELLIPSIS); }
328
329 /* alternative C99 brackets, "<:" & "<:<:" handled by preprocessor */
330"<:" { RETURN_VAL('['); }
331":>" { RETURN_VAL(']'); }
332"<%" { RETURN_VAL('{'); }
333"%>" { RETURN_VAL('}'); }
334
335 /* operators */
336"!" { ASCIIOP_RETURN(); }
337"+" { ASCIIOP_RETURN(); }
338"-" { ASCIIOP_RETURN(); }
339"*" { ASCIIOP_RETURN(); }
340"\\" { ASCIIOP_RETURN(); } // CFA, exponentiation
341"/" { ASCIIOP_RETURN(); }
342"%" { ASCIIOP_RETURN(); }
343"^" { ASCIIOP_RETURN(); }
344"~" { ASCIIOP_RETURN(); }
345"&" { ASCIIOP_RETURN(); }
346"|" { ASCIIOP_RETURN(); }
347"<" { ASCIIOP_RETURN(); }
348">" { ASCIIOP_RETURN(); }
349"=" { ASCIIOP_RETURN(); }
350"?" { ASCIIOP_RETURN(); }
351
352"++" { NAMEDOP_RETURN(ICR); }
353"--" { NAMEDOP_RETURN(DECR); }
354"==" { NAMEDOP_RETURN(EQ); }
355"!=" { NAMEDOP_RETURN(NE); }
356"<<" { NAMEDOP_RETURN(LS); }
357">>" { NAMEDOP_RETURN(RS); }
358"<=" { NAMEDOP_RETURN(LE); }
359">=" { NAMEDOP_RETURN(GE); }
360"&&" { NAMEDOP_RETURN(ANDAND); }
361"||" { NAMEDOP_RETURN(OROR); }
362"->" { NAMEDOP_RETURN(ARROW); }
363"+=" { NAMEDOP_RETURN(PLUSassign); }
364"-=" { NAMEDOP_RETURN(MINUSassign); }
365"\\=" { NAMEDOP_RETURN(EXPassign); } // CFA, exponentiation
366"*=" { NAMEDOP_RETURN(MULTassign); }
367"/=" { NAMEDOP_RETURN(DIVassign); }
368"%=" { NAMEDOP_RETURN(MODassign); }
369"&=" { NAMEDOP_RETURN(ANDassign); }
370"|=" { NAMEDOP_RETURN(ORassign); }
371"^=" { NAMEDOP_RETURN(ERassign); }
372"<<=" { NAMEDOP_RETURN(LSassign); }
373">>=" { NAMEDOP_RETURN(RSassign); }
374
375"@=" { NAMEDOP_RETURN(ATassign); } // CFA
376
377 /* CFA, operator identifier */
378{op_unary}"?" { IDENTIFIER_RETURN(); } // unary
379"?"({op_unary_pre_post}|"()"|"[?]"|"{}") { IDENTIFIER_RETURN(); }
380"^?{}" { IDENTIFIER_RETURN(); }
381"?"{op_binary_over}"?" { IDENTIFIER_RETURN(); } // binary
382 /*
383 This rule handles ambiguous cases with operator identifiers, e.g., "int *?*?()", where the string "*?*?" can be
384 lexed as "*?"/"*?" or "*"/"?*?". Since it is common practise to put a unary operator juxtaposed to an identifier,
385 e.g., "*i", users will be annoyed if they cannot do this with respect to operator identifiers. Therefore, there is
386 a lexical look-ahead for the second case, with backtracking to return the leading unary operator and then
387 reparsing the trailing operator identifier. Otherwise a space is needed between the unary operator and operator
388 identifier to disambiguate this common case.
389
390 A similar issue occurs with the dereference, *?(...), and routine-call, ?()(...) identifiers. The ambiguity
391 occurs when the deference operator has no parameters, *?() and *?()(...), requiring arbitrary whitespace
392 look-ahead for the routine-call parameter-list to disambiguate. However, the dereference operator must have a
393 parameter/argument to dereference *?(...). Hence, always interpreting the string *?() as * ?() does not preclude
394 any meaningful program.
395
396 The remaining cases are with the increment/decrement operators and conditional expression:
397
398 i++? ...(...);
399 i?++ ...(...);
400
401 requiring arbitrary whitespace look-ahead for the operator parameter-list, even though that interpretation is an
402 incorrect expression (juxtaposed identifiers). Therefore, it is necessary to disambiguate these cases with a
403 space:
404
405 i++ ? i : 0;
406 i? ++i : 0;
407 */
408{op_unary}"?"({op_unary_pre_post}|"()"|"[?]"|{op_binary_over}"?") {
409 // 1 or 2 character unary operator ?
410 int i = yytext[1] == '?' ? 1 : 2;
411 yyless( i ); // put back characters up to first '?'
412 if ( i > 1 ) {
413 NAMEDOP_RETURN( yytext[0] == '+' ? ICR : DECR );
414 } else {
415 ASCIIOP_RETURN();
416 } // if
417}
418
419 /* unknown characters */
420. { printf("unknown character(s):\"%s\" on line %d\n", yytext, yylineno); }
421
422%%
423
424// Local Variables: //
425// mode: c++ //
426// tab-width: 4 //
427// compile-command: "make install" //
428// End: //
Note: See TracBrowser for help on using the repository browser.