Context Navigation

lex.ll @ 0a2168f

ADTaaron-thesisarm-ehast-experimentalcleanup-dtorsdeferred_resndemanglerenumforall-pointer-decayjacob/cs343-translationjenkins-sandboxnew-astnew-ast-unique-exprnew-envno_listpersistent-indexerpthread-emulationqualifiedEnumresolv-newwith_gc

Last change on this file since 0a2168f was 0a2168f, checked in by Peter A. Buhr <pabuhr@…>, 6 years ago
add gcc binary constants
Property mode set to `100644`
File size: 18.3 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.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 : Sat Mar 3 18:38:16 2018
13	* Update Count : 640
14	*/
15
16	%option yylineno
17	%option noyywrap
18	%option nounput
19
20	%{
21	// The 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	unsigned int column = 0; // position of the end of the last token parsed
28	#define YY_USER_ACTION yylloc.first_line = yylineno; yylloc.first_column = column; column += yyleng; yylloc.last_column = column; yylloc.last_line = yylineno; yylloc.filename = yyfilename ? yyfilename : ""; // trigger before each matching rule's action
29
30	#include <string>
31	#include <cstdio> // FILENAME_MAX
32	using namespace std;
33
34	#include "ParseNode.h"
35	#include "TypedefTable.h"
36
37	char *yyfilename;
38	string *strtext; // accumulate parts of character and string constant value
39
40	#define RETURN_LOCN(x) yylval.tok.loc.file = yyfilename; yylval.tok.loc.line = yylineno; return( x )
41	#define RETURN_VAL(x) yylval.tok.str = new string( yytext ); RETURN_LOCN( x )
42	#define RETURN_CHAR(x) yylval.tok.str = nullptr; RETURN_LOCN( x )
43	#define RETURN_STR(x) yylval.tok.str = strtext; RETURN_LOCN( x )
44
45	#define WHITE_RETURN(x) // do nothing
46	#define NEWLINE_RETURN() column = 0; WHITE_RETURN( '\n' )
47	#define ASCIIOP_RETURN() RETURN_CHAR( (int)yytext[0] ) // single character operator
48	#define NAMEDOP_RETURN(x) RETURN_CHAR( x ) // multichar operator, with a name
49	#define NUMERIC_RETURN(x) rm_underscore(); RETURN_VAL( x ) // numeric constant
50	#define KEYWORD_RETURN(x) RETURN_CHAR( x ) // keyword
51	#define QKEYWORD_RETURN(x) typedefTable.isKind( yytext ); RETURN_VAL(x); // quasi-keyword
52	#define IDENTIFIER_RETURN() RETURN_VAL( typedefTable.isKind( yytext ) )
53	#define ATTRIBUTE_RETURN() RETURN_VAL( ATTR_IDENTIFIER )
54
55	void rm_underscore() {
56	// Remove underscores in numeric constant by copying the non-underscore characters to the front of the string.
57	yyleng = 0;
58	for ( int i = 0; yytext[i] != '\0'; i += 1 ) {
59	if ( yytext[i] == '`' ) {
60	// copy user suffix
61	for ( ; yytext[i] != '\0'; i += 1 ) {
62	yytext[yyleng] = yytext[i];
63	yyleng += 1;
64	} // for
65	break;
66	} // if
67	if ( yytext[i] != '_' ) {
68	yytext[yyleng] = yytext[i];
69	yyleng += 1;
70	} // if
71	} // for
72	yytext[yyleng] = '\0';
73	}
74
75	// Stop warning due to incorrectly generated flex code.
76	#pragma GCC diagnostic ignored "-Wsign-compare"
77	%}
78
79	binary [0-1]
80	octal [0-7]
81	nonzero [1-9]
82	decimal [0-9]
83	hex [0-9a-fA-F]
84	universal_char "\\"((u"_"?{hex_quad})\|(U"_"?{hex_quad}{2}))
85
86	// identifier, GCC: $ in identifier
87	identifier ([a-zA-Z_$]\|{universal_char})([0-9a-zA-Z_$]\|{universal_char})*
88
89	// attribute identifier, GCC: $ in identifier
90	attr_identifier "@"{identifier}
91
92	user_suffix_opt ("`"{identifier})?
93
94	// numeric constants, CFA: '_' in constant
95	hex_quad {hex}("_"?{hex}){3}
96	size_opt (8\|16\|32\|64\|128)?
97	length ("ll"\|"LL"\|[lL]{size_opt})\|("hh"\|"HH"\|[hH])
98	integer_suffix_opt ("_"?(([uU]({length}?[iI]?)\|([iI]{length}))\|([iI]({length}?[uU]?)\|([uU]{length}))\|({length}([iI]?[uU]?)\|([uU][iI]))\|[zZ]))?{user_suffix_opt}
99
100	octal_digits ({octal})\|({octal}({octal}\|"_")*{octal})
101	octal_prefix "0""_"?
102	octal_constant (("0")\|({octal_prefix}{octal_digits})){integer_suffix_opt}
103
104	nonzero_digits ({nonzero})\|({nonzero}({decimal}\|"_")*{decimal})
105	decimal_constant {nonzero_digits}{integer_suffix_opt}
106
107	binary_digits ({binary})\|({binary}({binary}\|"_")*{binary})
108	binary_prefix "0"[bB]"_"?
109	binary_constant {binary_prefix}{binary_digits}{integer_suffix_opt}
110
111	hex_digits ({hex})\|({hex}({hex}\|"_")*{hex})
112	hex_prefix "0"[xX]"_"?
113	hex_constant {hex_prefix}{hex_digits}{integer_suffix_opt}
114
115	// GCC: D (double) and iI (imaginary) suffixes, and DL (long double)
116	exponent "_"?[eE]"_"?[+-]?{decimal_digits}
117	floating_size 32\|64\|80\|128
118	floating_length ([fFdDlL]\|[lL]{floating_size})
119	floating_suffix ({floating_length}?[iI]?)\|([iI]{floating_length})
120	floating_suffix_opt ("_"?({floating_suffix}\|"DL"))?{user_suffix_opt}
121	decimal_digits ({decimal})\|({decimal}({decimal}\|"_")*{decimal})
122	floating_decimal {decimal_digits}"."{exponent}?{floating_suffix_opt}
123	floating_fraction "."{decimal_digits}{exponent}?{floating_suffix_opt}
124	floating_constant ({decimal_digits}{exponent}{floating_suffix_opt})\|({decimal_digits}{floating_fraction})
125
126	binary_exponent "_"?[pP]"_"?[+-]?{decimal_digits}
127	hex_floating_suffix_opt ("_"?({floating_suffix}))?{user_suffix_opt}
128	hex_floating_fraction ({hex_digits}?"."{hex_digits})\|({hex_digits}".")
129	hex_floating_constant {hex_prefix}(({hex_floating_fraction}{binary_exponent})\|({hex_digits}{binary_exponent})){hex_floating_suffix_opt}
130
131	// character escape sequence, GCC: \e => esc character
132	simple_escape "\\"[abefnrtv'"?\\]
133	// ' stop editor highlighting
134	octal_escape "\\"{octal}("_"?{octal}){0,2}
135	hex_escape "\\""x""_"?{hex_digits}
136	escape_seq {simple_escape}\|{octal_escape}\|{hex_escape}\|{universal_char}
137	cwide_prefix "L"\|"U"\|"u"
138	swide_prefix {cwide_prefix}\|"u8"
139
140	// display/white-space characters
141	h_tab [\011]
142	form_feed [\014]
143	v_tab [\013]
144	c_return [\015]
145	h_white [ ]\|{h_tab}
146
147	// overloadable operators
148	op_unary_only "~"\|"!"
149	op_unary_binary "+"\|"-"\|"*"
150	op_unary_pre_post "++"\|"--"
151	op_unary {op_unary_only}\|{op_unary_binary}\|{op_unary_pre_post}
152
153	op_binary_only "/"\|"%"\|"\\"\|"^"\|"&"\|"\|"\|"<"\|">"\|"="\|"=="\|"!="\|"<<"\|">>"\|"<="\|">="\|"+="\|"-="\|"*="\|"/="\|"%="\|"\\="\|"&="\|"\|="\|"^="\|"<<="\|">>="
154	op_binary_over {op_unary_binary}\|{op_binary_only}
155	// op_binary_not_over "?"\|"->"\|"."\|"&&"\|"\|\|"\|"@="
156	// operator {op_unary_pre_post}\|{op_binary_over}\|{op_binary_not_over}
157
158	%x COMMENT
159	%x BKQUOTE
160	%x QUOTE
161	%x STRING
162
163	%%
164	/* line directives */
165	^{h_white}"#"{h_white}[0-9]+{h_white}["][^"\n]+["]."\n" {
166	/* " stop editor highlighting */
167	static char filename[FILENAME_MAX]; // temporarily store current source-file name
168	char *end_num;
169	char begin_string, end_string;
170	long lineno, length;
171	lineno = strtol( yytext + 1, &end_num, 0 );
172	begin_string = strchr( end_num, '"' );
173	if ( begin_string ) { // file name ?
174	end_string = strchr( begin_string + 1, '"' ); // look for ending delimiter
175	assert( end_string ); // closing quote ?
176	length = end_string - begin_string - 1; // file-name length without quotes or sentinel
177	assert( length < FILENAME_MAX ); // room for sentinel ?
178	memcpy( &filename, begin_string + 1, length ); // copy file name from yytext
179	filename[ length ] = '\0'; // terminate string with sentinel
180	//cout << "file " << filename << " line " << lineno << endl;
181	yylineno = lineno;
182	yyfilename = filename;
183	} // if
184	}
185
186	/* ignore preprocessor directives (for now) */
187	^{h_white}"#"[^\n]"\n" ;
188
189	/* ignore C style comments (ALSO HANDLED BY CPP) */
190	"/*" { BEGIN COMMENT; }
191	<COMMENT>.\|\n ;
192	<COMMENT>"*/" { BEGIN 0; }
193
194	/* ignore C++ style comments (ALSO HANDLED BY CPP) */
195	"//"[^\n]*"\n" ;
196
197	/* ignore whitespace */
198	{h_white}+ { WHITE_RETURN(' '); }
199	({v_tab}\|{c_return}\|{form_feed})+ { WHITE_RETURN(' '); }
200	({h_white}\|{v_tab}\|{c_return}\|{form_feed})*"\n" { NEWLINE_RETURN(); }
201
202	/* keywords */
203	_Alignas { KEYWORD_RETURN(ALIGNAS); } // C11
204	_Alignof { KEYWORD_RETURN(ALIGNOF); } // C11
205	__alignof { KEYWORD_RETURN(ALIGNOF); } // GCC
206	__alignof__ { KEYWORD_RETURN(ALIGNOF); } // GCC
207	asm { KEYWORD_RETURN(ASM); }
208	__asm { KEYWORD_RETURN(ASM); } // GCC
209	__asm__ { KEYWORD_RETURN(ASM); } // GCC
210	_At { KEYWORD_RETURN(AT); } // CFA
211	_Atomic { KEYWORD_RETURN(ATOMIC); } // C11
212	__attribute { KEYWORD_RETURN(ATTRIBUTE); } // GCC
213	__attribute__ { KEYWORD_RETURN(ATTRIBUTE); } // GCC
214	auto { KEYWORD_RETURN(AUTO); }
215	_Bool { KEYWORD_RETURN(BOOL); } // C99
216	break { KEYWORD_RETURN(BREAK); }
217	case { KEYWORD_RETURN(CASE); }
218	catch { KEYWORD_RETURN(CATCH); } // CFA
219	catchResume { KEYWORD_RETURN(CATCHRESUME); } // CFA
220	char { KEYWORD_RETURN(CHAR); }
221	choose { KEYWORD_RETURN(CHOOSE); } // CFA
222	_Complex { KEYWORD_RETURN(COMPLEX); } // C99
223	__complex { KEYWORD_RETURN(COMPLEX); } // GCC
224	__complex__ { KEYWORD_RETURN(COMPLEX); } // GCC
225	const { KEYWORD_RETURN(CONST); }
226	__const { KEYWORD_RETURN(CONST); } // GCC
227	__const__ { KEYWORD_RETURN(CONST); } // GCC
228	continue { KEYWORD_RETURN(CONTINUE); }
229	coroutine { KEYWORD_RETURN(COROUTINE); } // CFA
230	default { KEYWORD_RETURN(DEFAULT); }
231	disable { KEYWORD_RETURN(DISABLE); } // CFA
232	do { KEYWORD_RETURN(DO); }
233	double { KEYWORD_RETURN(DOUBLE); }
234	dtype { KEYWORD_RETURN(DTYPE); } // CFA
235	else { KEYWORD_RETURN(ELSE); }
236	enable { KEYWORD_RETURN(ENABLE); } // CFA
237	enum { KEYWORD_RETURN(ENUM); }
238	__extension__ { KEYWORD_RETURN(EXTENSION); } // GCC
239	exception { KEYWORD_RETURN(EXCEPTION); } // CFA
240	extern { KEYWORD_RETURN(EXTERN); }
241	fallthru { KEYWORD_RETURN(FALLTHRU); } // CFA
242	fallthrough { KEYWORD_RETURN(FALLTHROUGH); } // CFA
243	finally { KEYWORD_RETURN(FINALLY); } // CFA
244	float { KEYWORD_RETURN(FLOAT); }
245	__float80 { KEYWORD_RETURN(FLOAT80); } // GCC
246	float80 { KEYWORD_RETURN(FLOAT80); } // GCC
247	__float128 { KEYWORD_RETURN(FLOAT128); } // GCC
248	float128 { KEYWORD_RETURN(FLOAT128); } // GCC
249	for { KEYWORD_RETURN(FOR); }
250	forall { KEYWORD_RETURN(FORALL); } // CFA
251	fortran { KEYWORD_RETURN(FORTRAN); }
252	ftype { KEYWORD_RETURN(FTYPE); } // CFA
253	_Generic { KEYWORD_RETURN(GENERIC); } // C11
254	goto { KEYWORD_RETURN(GOTO); }
255	if { KEYWORD_RETURN(IF); }
256	_Imaginary { KEYWORD_RETURN(IMAGINARY); } // C99
257	__imag { KEYWORD_RETURN(IMAGINARY); } // GCC
258	__imag__ { KEYWORD_RETURN(IMAGINARY); } // GCC
259	inline { KEYWORD_RETURN(INLINE); } // C99
260	__inline { KEYWORD_RETURN(INLINE); } // GCC
261	__inline__ { KEYWORD_RETURN(INLINE); } // GCC
262	int { KEYWORD_RETURN(INT); }
263	__int128 { KEYWORD_RETURN(INT128); } // GCC
264	int128 { KEYWORD_RETURN(INT128); } // GCC
265	__label__ { KEYWORD_RETURN(LABEL); } // GCC
266	long { KEYWORD_RETURN(LONG); }
267	monitor { KEYWORD_RETURN(MONITOR); } // CFA
268	mutex { KEYWORD_RETURN(MUTEX); } // CFA
269	_Noreturn { KEYWORD_RETURN(NORETURN); } // C11
270	__builtin_offsetof { KEYWORD_RETURN(OFFSETOF); } // GCC
271	one_t { NUMERIC_RETURN(ONE_T); } // CFA
272	otype { KEYWORD_RETURN(OTYPE); } // CFA
273	register { KEYWORD_RETURN(REGISTER); }
274	restrict { KEYWORD_RETURN(RESTRICT); } // C99
275	__restrict { KEYWORD_RETURN(RESTRICT); } // GCC
276	__restrict__ { KEYWORD_RETURN(RESTRICT); } // GCC
277	return { KEYWORD_RETURN(RETURN); }
278	short { KEYWORD_RETURN(SHORT); }
279	signed { KEYWORD_RETURN(SIGNED); }
280	__signed { KEYWORD_RETURN(SIGNED); } // GCC
281	__signed__ { KEYWORD_RETURN(SIGNED); } // GCC
282	sizeof { KEYWORD_RETURN(SIZEOF); }
283	static { KEYWORD_RETURN(STATIC); }
284	_Static_assert { KEYWORD_RETURN(STATICASSERT); } // C11
285	struct { KEYWORD_RETURN(STRUCT); }
286	switch { KEYWORD_RETURN(SWITCH); }
287	thread { KEYWORD_RETURN(THREAD); } // C11
288	_Thread_local { KEYWORD_RETURN(THREADLOCAL); } // C11
289	throw { KEYWORD_RETURN(THROW); } // CFA
290	throwResume { KEYWORD_RETURN(THROWRESUME); } // CFA
291	timeout { QKEYWORD_RETURN(TIMEOUT); } // CFA
292	trait { KEYWORD_RETURN(TRAIT); } // CFA
293	try { KEYWORD_RETURN(TRY); } // CFA
294	ttype { KEYWORD_RETURN(TTYPE); } // CFA
295	typedef { KEYWORD_RETURN(TYPEDEF); }
296	typeof { KEYWORD_RETURN(TYPEOF); } // GCC
297	__typeof { KEYWORD_RETURN(TYPEOF); } // GCC
298	__typeof__ { KEYWORD_RETURN(TYPEOF); } // GCC
299	union { KEYWORD_RETURN(UNION); }
300	unsigned { KEYWORD_RETURN(UNSIGNED); }
301	__builtin_va_list { KEYWORD_RETURN(VALIST); } // GCC
302	virtual { KEYWORD_RETURN(VIRTUAL); } // CFA
303	void { KEYWORD_RETURN(VOID); }
304	volatile { KEYWORD_RETURN(VOLATILE); }
305	__volatile { KEYWORD_RETURN(VOLATILE); } // GCC
306	__volatile__ { KEYWORD_RETURN(VOLATILE); } // GCC
307	waitfor { KEYWORD_RETURN(WAITFOR); }
308	or { QKEYWORD_RETURN(WOR); } // CFA
309	when { KEYWORD_RETURN(WHEN); }
310	while { KEYWORD_RETURN(WHILE); }
311	with { KEYWORD_RETURN(WITH); } // CFA
312	zero_t { NUMERIC_RETURN(ZERO_T); } // CFA
313
314	/* identifier */
315	{identifier} { IDENTIFIER_RETURN(); }
316	{attr_identifier} { ATTRIBUTE_RETURN(); }
317	"`" { BEGIN BKQUOTE; }
318	<BKQUOTE>{identifier} { IDENTIFIER_RETURN(); }
319	<BKQUOTE>"`" { BEGIN 0; }
320
321	/* numeric constants */
322	{binary_constant} { NUMERIC_RETURN(INTEGERconstant); }
323	{octal_constant} { NUMERIC_RETURN(INTEGERconstant); }
324	{decimal_constant} { NUMERIC_RETURN(INTEGERconstant); }
325	{hex_constant} { NUMERIC_RETURN(INTEGERconstant); }
326	{floating_decimal} { NUMERIC_RETURN(FLOATING_DECIMALconstant); } // must appear before floating_constant
327	{floating_fraction} { NUMERIC_RETURN(FLOATING_FRACTIONconstant); } // must appear before floating_constant
328	{floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
329	{hex_floating_constant} { NUMERIC_RETURN(FLOATINGconstant); }
330
331	/* character constant, allows empty value */
332	({cwide_prefix}[_]?)?['] { BEGIN QUOTE; rm_underscore(); strtext = new string( yytext, yyleng ); }
333	<QUOTE>[^'\\\n]* { strtext->append( yytext, yyleng ); }
334	<QUOTE>['\n]{user_suffix_opt} { BEGIN 0; strtext->append( yytext, yyleng ); RETURN_STR(CHARACTERconstant); }
335	/* ' stop editor highlighting */
336
337	/* string constant */
338	({swide_prefix}[_]?)?["] { BEGIN STRING; rm_underscore(); strtext = new string( yytext, yyleng ); }
339	<STRING>[^"\\\n]* { strtext->append( yytext, yyleng ); }
340	<STRING>["\n]{user_suffix_opt} { BEGIN 0; strtext->append( yytext, yyleng ); RETURN_STR(STRINGliteral); }
341	/* " stop editor highlighting */
342
343	/* common character/string constant */
344	<QUOTE,STRING>{escape_seq} { rm_underscore(); strtext->append( yytext, yyleng ); }
345	<QUOTE,STRING>"\\"{h_white}*"\n" {} // continuation (ALSO HANDLED BY CPP)
346	<QUOTE,STRING>"\\" { strtext->append( yytext, yyleng ); } // unknown escape character
347
348	/* punctuation */
349	"@" { ASCIIOP_RETURN(); }
350	"[" { ASCIIOP_RETURN(); }
351	"]" { ASCIIOP_RETURN(); }
352	"(" { ASCIIOP_RETURN(); }
353	")" { ASCIIOP_RETURN(); }
354	"{" { ASCIIOP_RETURN(); }
355	"}" { ASCIIOP_RETURN(); }
356	"," { ASCIIOP_RETURN(); } // also operator
357	":" { ASCIIOP_RETURN(); }
358	";" { ASCIIOP_RETURN(); }
359	"." { ASCIIOP_RETURN(); } // also operator
360	"..." { NAMEDOP_RETURN(ELLIPSIS); }
361
362	/* alternative C99 brackets, "<:" & "<:<:" handled by preprocessor */
363	"<:" { RETURN_VAL('['); }
364	":>" { RETURN_VAL(']'); }
365	"<%" { RETURN_VAL('{'); }
366	"%>" { RETURN_VAL('}'); }
367
368	/* operators */
369	"!" { ASCIIOP_RETURN(); }
370	"+" { ASCIIOP_RETURN(); }
371	"-" { ASCIIOP_RETURN(); }
372	"*" { ASCIIOP_RETURN(); }
373	"\\" { ASCIIOP_RETURN(); } // CFA, exponentiation
374	"/" { ASCIIOP_RETURN(); }
375	"%" { ASCIIOP_RETURN(); }
376	"^" { ASCIIOP_RETURN(); }
377	"~" { ASCIIOP_RETURN(); }
378	"&" { ASCIIOP_RETURN(); }
379	"\|" { ASCIIOP_RETURN(); }
380	"<" { ASCIIOP_RETURN(); }
381	">" { ASCIIOP_RETURN(); }
382	"=" { ASCIIOP_RETURN(); }
383	"?" { ASCIIOP_RETURN(); }
384
385	"++" { NAMEDOP_RETURN(ICR); }
386	"--" { NAMEDOP_RETURN(DECR); }
387	"==" { NAMEDOP_RETURN(EQ); }
388	"!=" { NAMEDOP_RETURN(NE); }
389	"<<" { NAMEDOP_RETURN(LS); }
390	">>" { NAMEDOP_RETURN(RS); }
391	"<=" { NAMEDOP_RETURN(LE); }
392	">=" { NAMEDOP_RETURN(GE); }
393	"&&" { NAMEDOP_RETURN(ANDAND); }
394	"\|\|" { NAMEDOP_RETURN(OROR); }
395	"->" { NAMEDOP_RETURN(ARROW); }
396	"+=" { NAMEDOP_RETURN(PLUSassign); }
397	"-=" { NAMEDOP_RETURN(MINUSassign); }
398	"\\=" { NAMEDOP_RETURN(EXPassign); } // CFA, exponentiation
399	"*=" { NAMEDOP_RETURN(MULTassign); }
400	"/=" { NAMEDOP_RETURN(DIVassign); }
401	"%=" { NAMEDOP_RETURN(MODassign); }
402	"&=" { NAMEDOP_RETURN(ANDassign); }
403	"\|=" { NAMEDOP_RETURN(ORassign); }
404	"^=" { NAMEDOP_RETURN(ERassign); }
405	"<<=" { NAMEDOP_RETURN(LSassign); }
406	">>=" { NAMEDOP_RETURN(RSassign); }
407
408	"@=" { NAMEDOP_RETURN(ATassign); } // CFA
409
410	/* CFA, operator identifier */
411	{op_unary}"?" { IDENTIFIER_RETURN(); } // unary
412	"?"({op_unary_pre_post}\|"()"\|"[?]"\|"{}") { IDENTIFIER_RETURN(); }
413	"^?{}" { IDENTIFIER_RETURN(); }
414	"?`"{identifier} { IDENTIFIER_RETURN(); } // unit operator
415	"?"{op_binary_over}"?" { IDENTIFIER_RETURN(); } // binary
416	/*
417	This rule handles ambiguous cases with operator identifiers, e.g., "int ??()", where the string "??" can be
418	lexed as "?"/"?" or ""/"??". Since it is common practise to put a unary operator juxtaposed to an identifier,
419	e.g., "*i", users will be annoyed if they cannot do this with respect to operator identifiers. Therefore, there is
420	a lexical look-ahead for the second case, with backtracking to return the leading unary operator and then
421	reparsing the trailing operator identifier. Otherwise a space is needed between the unary operator and operator
422	identifier to disambiguate this common case.
423
424	A similar issue occurs with the dereference, *?(...), and routine-call, ?()(...) identifiers. The ambiguity
425	occurs when the deference operator has no parameters, ?() and ?()(...), requiring arbitrary whitespace
426	look-ahead for the routine-call parameter-list to disambiguate. However, the dereference operator must have a
427	parameter/argument to dereference ?(...). Hence, always interpreting the string ?() as * ?() does not preclude
428	any meaningful program.
429
430	The remaining cases are with the increment/decrement operators and conditional expression:
431
432	i++? ...(...);
433	i?++ ...(...);
434
435	requiring arbitrary whitespace look-ahead for the operator parameter-list, even though that interpretation is an
436	incorrect expression (juxtaposed identifiers). Therefore, it is necessary to disambiguate these cases with a
437	space:
438
439	i++ ? i : 0;
440	i? ++i : 0;
441	*/
442	{op_unary}"?"({op_unary_pre_post}\|"()"\|"[?]"\|{op_binary_over}"?") {
443	// 1 or 2 character unary operator ?
444	int i = yytext[1] == '?' ? 1 : 2;
445	yyless( i ); // put back characters up to first '?'
446	if ( i > 1 ) {
447	NAMEDOP_RETURN( yytext[0] == '+' ? ICR : DECR );
448	} else {
449	ASCIIOP_RETURN();
450	} // if
451	}
452
453	/* unknown character */
454	. { yyerror( "unknown character" ); }
455
456	%%
457	// ----end of lexer----
458
459	void yyerror( const char * errmsg ) {
460	cout << (yyfilename ? yyfilename : "unknown file") << ':' << yylineno << ':' << column - yyleng + 1
461	<< ": " << ErrorHelpers::error_str() << errmsg << " at token \"" << (yytext[0] == '\0' ? "EOF" : yytext) << '"' << endl;
462	}
463
464	// Local Variables: //
465	// mode: c++ //
466	// tab-width: 4 //
467	// compile-command: "make install" //
468	// End: //

Note: See TracBrowser for help on using the repository browser.

Download in other formats:

Original Format