1 /* -*- Mode: C; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
2 *
3 * ***** BEGIN LICENSE BLOCK *****
4 * Version: MPL 1.1/GPL 2.0/LGPL 2.1
5 *
6 * The contents of this file are subject to the Mozilla Public License Version
7 * 1.1 (the "License"); you may not use this file except in compliance with
8 * the License. You may obtain a copy of the License at
9 * http://www.mozilla.org/MPL/
10 *
11 * Software distributed under the License is distributed on an "AS IS" basis,
12 * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
13 * for the specific language governing rights and limitations under the
14 * License.
15 *
16 * The Original Code is Mozilla Communicator client code, released
17 * March 31, 1998.
18 *
19 * The Initial Developer of the Original Code is
20 * Netscape Communications Corporation.
21 * Portions created by the Initial Developer are Copyright (C) 1998
22 * the Initial Developer. All Rights Reserved.
23 *
24 * Contributor(s):
25 *
26 * Alternatively, the contents of this file may be used under the terms of
27 * either of the GNU General Public License Version 2 or later (the "GPL"),
28 * or the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
29 * in which case the provisions of the GPL or the LGPL are applicable instead
30 * of those above. If you wish to allow use of your version of this file only
31 * under the terms of either the GPL or the LGPL, and not to allow others to
32 * use your version of this file under the terms of the MPL, indicate your
33 * decision by deleting the provisions above and replace them with the notice
34 * and other provisions required by the GPL or the LGPL. If you do not delete
35 * the provisions above, a recipient may use your version of this file under
36 * the terms of any one of the MPL, the GPL or the LGPL.
37 *
38 * ***** END LICENSE BLOCK ***** */
40 #ifndef jsscan_h___
41 #define jsscan_h___
42 /*
43 * JS lexical scanner interface.
44 */
45 #include <stddef.h>
46 #include <stdio.h>
47 #include "jsopcode.h"
48 #include "jsprvtd.h"
49 #include "jspubtd.h"
51 JS_BEGIN_EXTERN_C
53 typedef enum JSTokenType {
54 TOK_ERROR = -1, /* well-known as the only code < EOF */
55 TOK_EOF = 0, /* end of file */
56 TOK_EOL = 1, /* end of line */
57 TOK_SEMI = 2, /* semicolon */
58 TOK_COMMA = 3, /* comma operator */
59 TOK_ASSIGN = 4, /* assignment ops (= += -= etc.) */
60 TOK_HOOK = 5, TOK_COLON = 6, /* conditional (?:) */
61 TOK_OR = 7, /* logical or (||) */
62 TOK_AND = 8, /* logical and (&&) */
63 TOK_BITOR = 9, /* bitwise-or (|) */
64 TOK_BITXOR = 10, /* bitwise-xor (^) */
65 TOK_BITAND = 11, /* bitwise-and (&) */
66 TOK_EQOP = 12, /* equality ops (== !=) */
67 TOK_RELOP = 13, /* relational ops (< <= > >=) */
68 TOK_SHOP = 14, /* shift ops (<< >> >>>) */
69 TOK_PLUS = 15, /* plus */
70 TOK_MINUS = 16, /* minus */
71 TOK_STAR = 17, TOK_DIVOP = 18, /* multiply/divide ops (* / %) */
72 TOK_UNARYOP = 19, /* unary prefix operator */
73 TOK_INC = 20, TOK_DEC = 21, /* increment/decrement (++ --) */
74 TOK_DOT = 22, /* member operator (.) */
75 TOK_LB = 23, TOK_RB = 24, /* left and right brackets */
76 TOK_LC = 25, TOK_RC = 26, /* left and right curlies (braces) */
77 TOK_LP = 27, TOK_RP = 28, /* left and right parentheses */
78 TOK_NAME = 29, /* identifier */
79 TOK_NUMBER = 30, /* numeric constant */
80 TOK_STRING = 31, /* string constant */
81 TOK_OBJECT = 32, /* RegExp or other object constant */
82 TOK_PRIMARY = 33, /* true, false, null, this, super */
83 TOK_FUNCTION = 34, /* function keyword */
84 TOK_EXPORT = 35, /* export keyword */
85 TOK_IMPORT = 36, /* import keyword */
86 TOK_IF = 37, /* if keyword */
87 TOK_ELSE = 38, /* else keyword */
88 TOK_SWITCH = 39, /* switch keyword */
89 TOK_CASE = 40, /* case keyword */
90 TOK_DEFAULT = 41, /* default keyword */
91 TOK_WHILE = 42, /* while keyword */
92 TOK_DO = 43, /* do keyword */
93 TOK_FOR = 44, /* for keyword */
94 TOK_BREAK = 45, /* break keyword */
95 TOK_CONTINUE = 46, /* continue keyword */
96 TOK_IN = 47, /* in keyword */
97 TOK_VAR = 48, /* var keyword */
98 TOK_WITH = 49, /* with keyword */
99 TOK_RETURN = 50, /* return keyword */
100 TOK_NEW = 51, /* new keyword */
101 TOK_DELETE = 52, /* delete keyword */
102 TOK_DEFSHARP = 53, /* #n= for object/array initializers */
103 TOK_USESHARP = 54, /* #n# for object/array initializers */
104 TOK_TRY = 55, /* try keyword */
105 TOK_CATCH = 56, /* catch keyword */
106 TOK_FINALLY = 57, /* finally keyword */
107 TOK_THROW = 58, /* throw keyword */
108 TOK_INSTANCEOF = 59, /* instanceof keyword */
109 TOK_DEBUGGER = 60, /* debugger keyword */
110 TOK_XMLSTAGO = 61, /* XML start tag open (<) */
111 TOK_XMLETAGO = 62, /* XML end tag open (</) */
112 TOK_XMLPTAGC = 63, /* XML point tag close (/>) */
113 TOK_XMLTAGC = 64, /* XML start or end tag close (>) */
114 TOK_XMLNAME = 65, /* XML start-tag non-final fragment */
115 TOK_XMLATTR = 66, /* XML quoted attribute value */
116 TOK_XMLSPACE = 67, /* XML whitespace */
117 TOK_XMLTEXT = 68, /* XML text */
118 TOK_XMLCOMMENT = 69, /* XML comment */
119 TOK_XMLCDATA = 70, /* XML CDATA section */
120 TOK_XMLPI = 71, /* XML processing instruction */
121 TOK_AT = 72, /* XML attribute op (@) */
122 TOK_DBLCOLON = 73, /* namespace qualified name op (::) */
123 TOK_ANYNAME = 74, /* XML AnyName singleton (*) */
124 TOK_DBLDOT = 75, /* XML descendant op (..) */
125 TOK_FILTER = 76, /* XML filtering predicate op (.()) */
126 TOK_XMLELEM = 77, /* XML element node type (no token) */
127 TOK_XMLLIST = 78, /* XML list node type (no token) */
128 TOK_RESERVED, /* reserved keywords */
129 TOK_LIMIT /* domain size */
130 } JSTokenType;
132 #define IS_PRIMARY_TOKEN(tt) \
133 ((uintN)((tt) - TOK_NAME) <= (uintN)(TOK_PRIMARY - TOK_NAME))
135 #define TOKEN_TYPE_IS_XML(tt) \
136 (tt == TOK_AT || tt == TOK_DBLCOLON || tt == TOK_ANYNAME)
138 struct JSStringBuffer {
139 jschar *base;
140 jschar *limit; /* length limit for quick bounds check */
141 jschar *ptr; /* slot for next non-NUL char to store */
142 void *data;
143 JSBool (*grow)(JSStringBuffer *sb, size_t newlength);
144 void (*free)(JSStringBuffer *sb);
145 };
147 #define STRING_BUFFER_ERROR_BASE ((jschar *) 1)
148 #define STRING_BUFFER_OK(sb) ((sb)->base != STRING_BUFFER_ERROR_BASE)
149 #define STRING_BUFFER_OFFSET(sb) ((sb)->ptr -(sb)->base)
151 extern void
152 js_InitStringBuffer(JSStringBuffer *sb);
154 extern void
155 js_FinishStringBuffer(JSStringBuffer *sb);
157 extern void
158 js_AppendChar(JSStringBuffer *sb, jschar c);
160 extern void
161 js_RepeatChar(JSStringBuffer *sb, jschar c, uintN count);
163 extern void
164 js_AppendCString(JSStringBuffer *sb, const char *asciiz);
166 extern void
167 js_AppendJSString(JSStringBuffer *sb, JSString *str);
169 struct JSTokenPtr {
170 uint16 index; /* index of char in physical line */
171 uint16 lineno; /* physical line number */
172 };
174 struct JSTokenPos {
175 JSTokenPtr begin; /* first character and line of token */
176 JSTokenPtr end; /* index 1 past last char, last line */
177 };
179 struct JSToken {
180 JSTokenType type; /* char value or above enumerator */
181 JSTokenPos pos; /* token position in file */
182 jschar *ptr; /* beginning of token in line buffer */
183 union {
184 struct { /* non-numeric literal */
185 JSOp op; /* operator, for minimal parser */
186 JSAtom *atom; /* atom table entry */
187 } s;
188 struct { /* atom pair, for XML PIs */
189 JSAtom *atom2; /* auxiliary atom table entry */
190 JSAtom *atom; /* main atom table entry */
191 } p;
192 jsdouble dval; /* floating point number */
193 } u;
194 };
196 #define t_op u.s.op
197 #define t_atom u.s.atom
198 #define t_atom2 u.p.atom2
199 #define t_dval u.dval
201 typedef struct JSTokenBuf {
202 jschar *base; /* base of line or stream buffer */
203 jschar *limit; /* limit for quick bounds check */
204 jschar *ptr; /* next char to get, or slot to use */
205 } JSTokenBuf;
207 #define JS_LINE_LIMIT 256 /* logical line buffer size limit --
208 physical line length is unlimited */
209 #define NTOKENS 4 /* 1 current + 2 lookahead, rounded */
210 #define NTOKENS_MASK (NTOKENS-1) /* to power of 2 to avoid divmod by 3 */
212 struct JSTokenStream {
213 JSToken tokens[NTOKENS];/* circular token buffer */
214 uintN cursor; /* index of last parsed token */
215 uintN lookahead; /* count of lookahead tokens */
216 uintN lineno; /* current line number */
217 uintN ungetpos; /* next free char slot in ungetbuf */
218 jschar ungetbuf[6]; /* at most 6, for \uXXXX lookahead */
219 uintN flags; /* flags -- see below */
220 ptrdiff_t linelen; /* physical linebuf segment length */
221 ptrdiff_t linepos; /* linebuf offset in physical line */
222 JSTokenBuf linebuf; /* line buffer for diagnostics */
223 JSTokenBuf userbuf; /* user input buffer if !file */
224 JSStringBuffer tokenbuf; /* current token string buffer */
225 const char *filename; /* input filename or null */
226 FILE *file; /* stdio stream if reading from file */
227 JSPrincipals *principals; /* principals associated with source */
228 JSSourceHandler listener; /* callback for source; eg debugger */
229 void *listenerData; /* listener 'this' data */
230 void *listenerTSData;/* listener data for this TokenStream */
231 jschar *saveEOL; /* save next end of line in userbuf, to
232 optimize for very long lines */
233 };
235 #define CURRENT_TOKEN(ts) ((ts)->tokens[(ts)->cursor])
236 #define ON_CURRENT_LINE(ts,pos) ((uint16)(ts)->lineno == (pos).end.lineno)
238 /* JSTokenStream flags */
239 #define TSF_ERROR 0x01 /* fatal error while compiling */
240 #define TSF_EOF 0x02 /* hit end of file */
241 #define TSF_NEWLINES 0x04 /* tokenize newlines */
242 #define TSF_OPERAND 0x08 /* looking for operand, not operator */
243 #define TSF_NLFLAG 0x20 /* last linebuf ended with \n */
244 #define TSF_CRFLAG 0x40 /* linebuf would have ended with \r */
245 #define TSF_DIRTYLINE 0x80 /* non-whitespace since start of line */
246 #define TSF_OWNFILENAME 0x100 /* ts->filename is malloc'd */
247 #define TSF_XMLTAGMODE 0x200 /* scanning within an XML tag in E4X */
248 #define TSF_XMLTEXTMODE 0x400 /* scanning XMLText terminal from E4X */
249 #define TSF_XMLONLYMODE 0x800 /* don't scan {expr} within text/tag */
251 /* Flag indicating unexpected end of input, i.e. TOK_EOF not at top-level. */
252 #define TSF_UNEXPECTED_EOF 0x1000
254 /*
255 * To handle the hard case of contiguous HTML comments, we want to clear the
256 * TSF_DIRTYINPUT flag at the end of each such comment. But we'd rather not
257 * scan for --> within every //-style comment unless we have to. So we set
258 * TSF_IN_HTML_COMMENT when a <!-- is scanned as an HTML begin-comment, and
259 * clear it (and TSF_DIRTYINPUT) when we scan --> either on a clean line, or
260 * only if (ts->flags & TSF_IN_HTML_COMMENT), in a //-style comment.
261 *
262 * This still works as before given a malformed comment hiding hack such as:
263 *
264 * <script>
265 * <!-- comment hiding hack #1
266 * code goes here
267 * // --> oops, markup for script-unaware browsers goes here!
268 * </script>
269 *
270 * It does not cope with malformed comment hiding hacks where --> is hidden
271 * by C-style comments, or on a dirty line. Such cases are already broken.
272 */
273 #define TSF_IN_HTML_COMMENT 0x2000
275 /* Unicode separators that are treated as line terminators, in addition to \n, \r */
276 #define LINE_SEPARATOR 0x2028
277 #define PARA_SEPARATOR 0x2029
279 /*
280 * Create a new token stream, either from an input buffer or from a file.
281 * Return null on file-open or memory-allocation failure.
282 *
283 * NB: All of js_New{,Buffer,File}TokenStream() return a pointer to transient
284 * memory in the current context's temp pool. This memory is deallocated via
285 * JS_ARENA_RELEASE() after parsing is finished.
286 */
287 extern JSTokenStream *
288 js_NewTokenStream(JSContext *cx, const jschar *base, size_t length,
289 const char *filename, uintN lineno, JSPrincipals *principals);
291 extern JS_FRIEND_API(JSTokenStream *)
292 js_NewBufferTokenStream(JSContext *cx, const jschar *base, size_t length);
294 extern JS_FRIEND_API(JSTokenStream *)
295 js_NewFileTokenStream(JSContext *cx, const char *filename, FILE *defaultfp);
297 extern JS_FRIEND_API(JSBool)
298 js_CloseTokenStream(JSContext *cx, JSTokenStream *ts);
300 extern JS_FRIEND_API(int)
301 js_fgets(char *buf, int size, FILE *file);
303 /*
304 * Initialize the scanner, installing JS keywords into cx's global scope.
305 */
306 extern JSBool
307 js_InitScanner(JSContext *cx);
309 /*
310 * Friend-exported API entry point to call a mapping function on each reserved
311 * identifier in the scanner's keyword table.
312 */
313 extern JS_FRIEND_API(void)
314 js_MapKeywords(void (*mapfun)(const char *));
316 /*
317 * Report a compile-time error by its number, using ts or cg to show context.
318 * Return true for a warning, false for an error.
319 */
320 extern JSBool
321 js_ReportCompileErrorNumber(JSContext *cx, void *handle, uintN flags,
322 uintN errorNumber, ...);
324 extern JSBool
325 js_ReportCompileErrorNumberUC(JSContext *cx, void *handle, uintN flags,
326 uintN errorNumber, ...);
328 /* Steal some JSREPORT_* bits (see jsapi.h) to tell handle's type. */
329 #define JSREPORT_HANDLE 0x300
330 #define JSREPORT_TS 0x000
331 #define JSREPORT_CG 0x100
332 #define JSREPORT_PN 0x200
334 /*
335 * Look ahead one token and return its type.
336 */
337 extern JSTokenType
338 js_PeekToken(JSContext *cx, JSTokenStream *ts);
340 extern JSTokenType
341 js_PeekTokenSameLine(JSContext *cx, JSTokenStream *ts);
343 /*
344 * Get the next token from ts.
345 */
346 extern JSTokenType
347 js_GetToken(JSContext *cx, JSTokenStream *ts);
349 /*
350 * Push back the last scanned token onto ts.
351 */
352 extern void
353 js_UngetToken(JSTokenStream *ts);
355 /*
356 * Get the next token from ts if its type is tt.
357 */
358 extern JSBool
359 js_MatchToken(JSContext *cx, JSTokenStream *ts, JSTokenType tt);
361 JS_END_EXTERN_C
363 #endif /* jsscan_h___ */