Commit | Line | Data |
---|---|---|
be54b15d | 1 | /* Copyright (C) 1995,1996,1997,1999,2000,2001 Free Software Foundation, Inc. |
0f2d19dd JB |
2 | * |
3 | * This program is free software; you can redistribute it and/or modify | |
4 | * it under the terms of the GNU General Public License as published by | |
5 | * the Free Software Foundation; either version 2, or (at your option) | |
6 | * any later version. | |
7 | * | |
8 | * This program is distributed in the hope that it will be useful, | |
9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
11 | * GNU General Public License for more details. | |
12 | * | |
13 | * You should have received a copy of the GNU General Public License | |
14 | * along with this software; see the file COPYING. If not, write to | |
82892bed JB |
15 | * the Free Software Foundation, Inc., 59 Temple Place, Suite 330, |
16 | * Boston, MA 02111-1307 USA | |
0f2d19dd JB |
17 | * |
18 | * As a special exception, the Free Software Foundation gives permission | |
19 | * for additional uses of the text contained in its release of GUILE. | |
20 | * | |
21 | * The exception is that, if you link the GUILE library with other files | |
22 | * to produce an executable, this does not by itself cause the | |
23 | * resulting executable to be covered by the GNU General Public License. | |
24 | * Your use of that executable is in no way restricted on account of | |
25 | * linking the GUILE library code into it. | |
26 | * | |
27 | * This exception does not however invalidate any other reasons why | |
28 | * the executable file might be covered by the GNU General Public License. | |
29 | * | |
30 | * This exception applies only to the code released by the | |
31 | * Free Software Foundation under the name GUILE. If you copy | |
32 | * code from other Free Software Foundation releases into a copy of | |
33 | * GUILE, as the General Public License permits, the exception does | |
34 | * not apply to the code that you add in this way. To avoid misleading | |
35 | * anyone as to the status of such modified files, you must delete | |
36 | * this exception notice from them. | |
37 | * | |
38 | * If you write modifications of your own for GUILE, it is your choice | |
39 | * whether to permit this exception to apply to your modifications. | |
82892bed | 40 | * If you do not wish that, delete this exception notice. */ |
1bbd0b84 GB |
41 | |
42 | /* Software engineering face-lift by Greg J. Badros, 11-Dec-1999, | |
43 | gjb@cs.washington.edu, http://www.cs.washington.edu/homes/gjb */ | |
44 | ||
0f2d19dd JB |
45 | \f |
46 | ||
0f2d19dd | 47 | #include <stdio.h> |
a0599745 MD |
48 | #include "libguile/_scm.h" |
49 | #include "libguile/chars.h" | |
50 | #include "libguile/eval.h" | |
51 | #include "libguile/unif.h" | |
52 | #include "libguile/keywords.h" | |
53 | #include "libguile/alist.h" | |
54 | #include "libguile/srcprop.h" | |
55 | #include "libguile/hashtab.h" | |
56 | #include "libguile/hash.h" | |
57 | #include "libguile/ports.h" | |
58 | #include "libguile/root.h" | |
59 | #include "libguile/strings.h" | |
60 | #include "libguile/vectors.h" | |
61 | ||
62 | #include "libguile/validate.h" | |
63 | #include "libguile/read.h" | |
0f2d19dd JB |
64 | |
65 | \f | |
66 | ||
c7733771 GH |
67 | SCM_SYMBOL (scm_keyword_prefix, "prefix"); |
68 | ||
92c2555f | 69 | scm_t_option scm_read_opts[] = { |
b7ff98dd MD |
70 | { SCM_OPTION_BOOLEAN, "copy", 0, |
71 | "Copy source code expressions." }, | |
ac74fc22 | 72 | { SCM_OPTION_BOOLEAN, "positions", 0, |
deca31e1 GH |
73 | "Record positions of source code expressions." }, |
74 | { SCM_OPTION_BOOLEAN, "case-insensitive", 0, | |
c7733771 | 75 | "Convert symbols to lower case."}, |
f1267706 | 76 | { SCM_OPTION_SCM, "keywords", SCM_UNPACK (SCM_BOOL_F), |
c7733771 | 77 | "Style of keyword recognition: #f or 'prefix"} |
a16f6fe7 MD |
78 | }; |
79 | ||
a1ec6916 | 80 | SCM_DEFINE (scm_read_options, "read-options-interface", 0, 1, 0, |
1bbd0b84 | 81 | (SCM setting), |
dc7fa443 MG |
82 | "Option interface for the read options. Instead of using\n" |
83 | "this procedure directly, use the procedures @code{read-enable},\n" | |
84 | "@code{read-disable}, @code{read-set!} and @var{read-options}.") | |
1bbd0b84 | 85 | #define FUNC_NAME s_scm_read_options |
a16f6fe7 | 86 | { |
b7ff98dd MD |
87 | SCM ans = scm_options (setting, |
88 | scm_read_opts, | |
89 | SCM_N_READ_OPTIONS, | |
1bbd0b84 | 90 | FUNC_NAME); |
b7ff98dd MD |
91 | if (SCM_COPY_SOURCE_P) |
92 | SCM_RECORD_POSITIONS_P = 1; | |
a16f6fe7 MD |
93 | return ans; |
94 | } | |
1bbd0b84 | 95 | #undef FUNC_NAME |
a16f6fe7 | 96 | |
14de3b42 GH |
97 | /* An association list mapping extra hash characters to procedures. */ |
98 | static SCM *scm_read_hash_procedures; | |
deca31e1 | 99 | |
a1ec6916 | 100 | SCM_DEFINE (scm_read, "read", 0, 1, 0, |
1bbd0b84 | 101 | (SCM port), |
dc7fa443 MG |
102 | "Read an s-expression from the input port @var{port}, or from\n" |
103 | "the current input port if @var{port} is not specified.\n" | |
104 | "Any whitespace before the next token is discarded.") | |
1bbd0b84 | 105 | #define FUNC_NAME s_scm_read |
0f2d19dd JB |
106 | { |
107 | int c; | |
09a4f039 | 108 | SCM tok_buf, copy; |
0f2d19dd JB |
109 | |
110 | if (SCM_UNBNDP (port)) | |
111 | port = scm_cur_inp; | |
3b3b36dd | 112 | SCM_VALIDATE_OPINPORT (1,port); |
0f2d19dd | 113 | |
0f2d19dd JB |
114 | c = scm_flush_ws (port, (char *) NULL); |
115 | if (EOF == c) | |
116 | return SCM_EOF_VAL; | |
b7f3516f | 117 | scm_ungetc (c, port); |
0f2d19dd | 118 | |
be54b15d | 119 | tok_buf = scm_allocate_string (30); |
deca31e1 | 120 | return scm_lreadr (&tok_buf, port, ©); |
0f2d19dd | 121 | } |
1bbd0b84 | 122 | #undef FUNC_NAME |
0f2d19dd JB |
123 | |
124 | ||
1cc91f1b | 125 | |
0f2d19dd | 126 | char * |
6e8d25a6 | 127 | scm_grow_tok_buf (SCM *tok_buf) |
0f2d19dd | 128 | { |
1be6b49c | 129 | size_t oldlen = SCM_STRING_LENGTH (*tok_buf); |
be54b15d | 130 | SCM newstr = scm_allocate_string (2 * oldlen); |
1be6b49c | 131 | size_t i; |
94115ae3 DH |
132 | |
133 | for (i = 0; i != oldlen; ++i) | |
134 | SCM_STRING_CHARS (newstr) [i] = SCM_STRING_CHARS (*tok_buf) [i]; | |
135 | ||
136 | *tok_buf = newstr; | |
137 | return SCM_STRING_CHARS (newstr); | |
0f2d19dd JB |
138 | } |
139 | ||
140 | ||
1cc91f1b | 141 | |
0f2d19dd | 142 | int |
6e8d25a6 | 143 | scm_flush_ws (SCM port, const char *eoferr) |
0f2d19dd JB |
144 | { |
145 | register int c; | |
146 | while (1) | |
b7f3516f | 147 | switch (c = scm_getc (port)) |
0f2d19dd JB |
148 | { |
149 | case EOF: | |
150 | goteof: | |
151 | if (eoferr) | |
d156d3b7 | 152 | { |
d1ca2c64 | 153 | if (!SCM_FALSEP (SCM_FILENAME (port))) |
d156d3b7 MV |
154 | scm_misc_error (eoferr, |
155 | "end of file in ~A", | |
156 | SCM_LIST1 (SCM_FILENAME (port))); | |
157 | else | |
158 | scm_misc_error (eoferr, "end of file", SCM_EOL); | |
159 | } | |
0f2d19dd JB |
160 | return c; |
161 | case ';': | |
162 | lp: | |
b7f3516f | 163 | switch (c = scm_getc (port)) |
0f2d19dd JB |
164 | { |
165 | case EOF: | |
166 | goto goteof; | |
167 | default: | |
168 | goto lp; | |
169 | case SCM_LINE_INCREMENTORS: | |
170 | break; | |
171 | } | |
172 | break; | |
173 | case SCM_LINE_INCREMENTORS: | |
0f2d19dd | 174 | case SCM_SINGLE_SPACES: |
0f2d19dd | 175 | case '\t': |
0f2d19dd JB |
176 | break; |
177 | default: | |
178 | return c; | |
179 | } | |
180 | } | |
181 | ||
182 | ||
1cc91f1b | 183 | |
0f2d19dd | 184 | int |
6e8d25a6 | 185 | scm_casei_streq (char *s1, char *s2) |
0f2d19dd JB |
186 | { |
187 | while (*s1 && *s2) | |
188 | if (scm_downcase((int)*s1) != scm_downcase((int)*s2)) | |
189 | return 0; | |
190 | else | |
191 | { | |
192 | ++s1; | |
193 | ++s2; | |
194 | } | |
195 | return !(*s1 || *s2); | |
196 | } | |
197 | ||
198 | ||
09a4f039 MD |
199 | /* recsexpr is used when recording expressions |
200 | * constructed by read:sharp. | |
201 | */ | |
604d4dd9 JB |
202 | #ifndef DEBUG_EXTENSIONS |
203 | #define recsexpr(obj, line, column, filename) (obj) | |
204 | #else | |
09a4f039 | 205 | static SCM |
1be6b49c | 206 | recsexpr (SCM obj, long line, int column, SCM filename) |
09a4f039 | 207 | { |
fee7ef83 | 208 | if (!SCM_CONSP(obj)) { |
09a4f039 | 209 | return obj; |
fee7ef83 | 210 | } else { |
09a4f039 MD |
211 | SCM tmp = obj, copy; |
212 | /* If this sexpr is visible in the read:sharp source, we want to | |
213 | keep that information, so only record non-constant cons cells | |
214 | which haven't previously been read by the reader. */ | |
215 | if (SCM_FALSEP (scm_whash_lookup (scm_source_whash, obj))) | |
216 | { | |
217 | if (SCM_COPY_SOURCE_P) | |
218 | { | |
219 | copy = scm_cons (recsexpr (SCM_CAR (obj), line, column, filename), | |
220 | SCM_UNDEFINED); | |
0c95b57d | 221 | while ((tmp = SCM_CDR (tmp)) && SCM_CONSP (tmp)) |
a6c64c3c MD |
222 | { |
223 | SCM_SETCDR (copy, scm_cons (recsexpr (SCM_CAR (tmp), | |
224 | line, | |
225 | column, | |
226 | filename), | |
227 | SCM_UNDEFINED)); | |
228 | copy = SCM_CDR (copy); | |
229 | } | |
230 | SCM_SETCDR (copy, tmp); | |
09a4f039 MD |
231 | } |
232 | else | |
233 | { | |
234 | recsexpr (SCM_CAR (obj), line, column, filename); | |
0c95b57d | 235 | while ((tmp = SCM_CDR (tmp)) && SCM_CONSP (tmp)) |
09a4f039 MD |
236 | recsexpr (SCM_CAR (tmp), line, column, filename); |
237 | copy = SCM_UNDEFINED; | |
238 | } | |
239 | scm_whash_insert (scm_source_whash, | |
240 | obj, | |
241 | scm_make_srcprops (line, | |
242 | column, | |
243 | filename, | |
244 | copy, | |
245 | SCM_EOL)); | |
246 | } | |
247 | return obj; | |
248 | } | |
249 | } | |
604d4dd9 | 250 | #endif |
f9c68a47 JB |
251 | |
252 | /* Consume an SCSH-style block comment. Assume that we've already | |
f9731264 JB |
253 | read the initial `#!', and eat characters until we get a |
254 | newline/exclamation-point/sharp-sign/newline sequence. */ | |
f9c68a47 JB |
255 | |
256 | static void | |
6e8d25a6 | 257 | skip_scsh_block_comment (SCM port) |
db4b4ca6 | 258 | #define FUNC_NAME "skip_scsh_block_comment" |
f9c68a47 | 259 | { |
f9731264 JB |
260 | /* Is this portable? Dear God, spare me from the non-eight-bit |
261 | characters. But is it tasteful? */ | |
262 | long history = 0; | |
f9c68a47 JB |
263 | |
264 | for (;;) | |
265 | { | |
b7f3516f | 266 | int c = scm_getc (port); |
f9c68a47 JB |
267 | |
268 | if (c == EOF) | |
db4b4ca6 | 269 | SCM_MISC_ERROR ("unterminated `#! ... !#' comment", SCM_EOL); |
f9731264 | 270 | history = ((history << 8) | (c & 0xff)) & 0xffffffff; |
f9c68a47 | 271 | |
f9731264 JB |
272 | /* Were the last four characters read "\n!#\n"? */ |
273 | if (history == (('\n' << 24) | ('!' << 16) | ('#' << 8) | '\n')) | |
274 | return; | |
f9c68a47 JB |
275 | } |
276 | } | |
db4b4ca6 DH |
277 | #undef FUNC_NAME |
278 | ||
f9c68a47 | 279 | |
1bbd0b84 | 280 | static SCM scm_get_hash_procedure(int c); |
f9c68a47 | 281 | |
09a4f039 | 282 | static char s_list[]="list"; |
1cc91f1b | 283 | |
0f2d19dd | 284 | SCM |
1bbd0b84 | 285 | scm_lreadr (SCM *tok_buf,SCM port,SCM *copy) |
db4b4ca6 | 286 | #define FUNC_NAME "scm_lreadr" |
0f2d19dd JB |
287 | { |
288 | int c; | |
1be6b49c | 289 | size_t j; |
0f2d19dd | 290 | SCM p; |
deca31e1 | 291 | |
b858464a | 292 | tryagain: |
1bbd0b84 | 293 | c = scm_flush_ws (port, s_scm_read); |
b858464a | 294 | tryagain_no_flush_ws: |
0f2d19dd JB |
295 | switch (c) |
296 | { | |
297 | case EOF: | |
298 | return SCM_EOF_VAL; | |
299 | ||
300 | case '(': | |
09a4f039 | 301 | return SCM_RECORD_POSITIONS_P |
b858464a MG |
302 | ? scm_lreadrecparen (tok_buf, port, s_list, copy) |
303 | : scm_lreadparen (tok_buf, port, s_list, copy); | |
0f2d19dd | 304 | case ')': |
db4b4ca6 | 305 | SCM_MISC_ERROR ("unexpected \")\"", SCM_EOL); |
0f2d19dd JB |
306 | goto tryagain; |
307 | ||
308 | case '\'': | |
92e5aa0e | 309 | p = scm_sym_quote; |
09a4f039 | 310 | goto recquote; |
0f2d19dd | 311 | case '`': |
92e5aa0e | 312 | p = scm_sym_quasiquote; |
09a4f039 | 313 | goto recquote; |
0f2d19dd | 314 | case ',': |
b7f3516f | 315 | c = scm_getc (port); |
0f2d19dd | 316 | if ('@' == c) |
92e5aa0e | 317 | p = scm_sym_uq_splicing; |
0f2d19dd JB |
318 | else |
319 | { | |
b7f3516f | 320 | scm_ungetc (c, port); |
92e5aa0e | 321 | p = scm_sym_unquote; |
0f2d19dd | 322 | } |
09a4f039 MD |
323 | recquote: |
324 | p = scm_cons2 (p, | |
deca31e1 | 325 | scm_lreadr (tok_buf, port, copy), |
09a4f039 MD |
326 | SCM_EOL); |
327 | if (SCM_RECORD_POSITIONS_P) | |
328 | scm_whash_insert (scm_source_whash, | |
329 | p, | |
330 | scm_make_srcprops (SCM_LINUM (port), | |
331 | SCM_COL (port) - 1, | |
332 | SCM_FILENAME (port), | |
333 | SCM_COPY_SOURCE_P | |
334 | ? (*copy = scm_cons2 (SCM_CAR (p), | |
335 | SCM_CAR (SCM_CDR (p)), | |
336 | SCM_EOL)) | |
337 | : SCM_UNDEFINED, | |
338 | SCM_EOL)); | |
339 | return p; | |
0f2d19dd | 340 | case '#': |
b7f3516f | 341 | c = scm_getc (port); |
b858464a MG |
342 | |
343 | { | |
344 | /* Check for user-defined hash procedure first, to allow | |
345 | overriding of builtin hash read syntaxes. */ | |
346 | SCM sharp = scm_get_hash_procedure (c); | |
347 | if (!SCM_FALSEP (sharp)) | |
348 | { | |
349 | int line = SCM_LINUM (port); | |
350 | int column = SCM_COL (port) - 2; | |
351 | SCM got; | |
352 | ||
353 | got = scm_call_2 (sharp, SCM_MAKE_CHAR (c), port); | |
354 | if (SCM_EQ_P (got, SCM_UNSPECIFIED)) | |
355 | goto unkshrp; | |
356 | if (SCM_RECORD_POSITIONS_P) | |
357 | return *copy = recsexpr (got, line, column, | |
358 | SCM_FILENAME (port)); | |
359 | else | |
360 | return got; | |
361 | } | |
362 | } | |
0f2d19dd JB |
363 | switch (c) |
364 | { | |
365 | case '(': | |
deca31e1 | 366 | p = scm_lreadparen (tok_buf, port, "vector", copy); |
0f2d19dd JB |
367 | return SCM_NULLP (p) ? scm_nullvect : scm_vector (p); |
368 | ||
369 | case 't': | |
370 | case 'T': | |
371 | return SCM_BOOL_T; | |
372 | case 'f': | |
373 | case 'F': | |
374 | return SCM_BOOL_F; | |
375 | ||
376 | case 'b': | |
377 | case 'B': | |
378 | case 'o': | |
379 | case 'O': | |
380 | case 'd': | |
381 | case 'D': | |
382 | case 'x': | |
383 | case 'X': | |
384 | case 'i': | |
385 | case 'I': | |
386 | case 'e': | |
387 | case 'E': | |
b7f3516f | 388 | scm_ungetc (c, port); |
0f2d19dd JB |
389 | c = '#'; |
390 | goto num; | |
391 | ||
f9c68a47 JB |
392 | case '!': |
393 | /* start of a shell script. Parse as a block comment, | |
394 | terminated by !#, just like SCSH. */ | |
395 | skip_scsh_block_comment (port); | |
b6356af7 MV |
396 | /* EOF is not an error here */ |
397 | c = scm_flush_ws (port, (char *)NULL); | |
398 | goto tryagain_no_flush_ws; | |
f9c68a47 | 399 | |
afe5177e | 400 | #ifdef HAVE_ARRAYS |
0f2d19dd | 401 | case '*': |
deca31e1 | 402 | j = scm_read_token (c, tok_buf, port, 0); |
405aaef9 | 403 | p = scm_istr2bve (SCM_STRING_CHARS (*tok_buf) + 1, (long) (j - 1)); |
36284627 | 404 | if (!SCM_FALSEP (p)) |
0f2d19dd JB |
405 | return p; |
406 | else | |
407 | goto unkshrp; | |
afe5177e | 408 | #endif |
0f2d19dd JB |
409 | |
410 | case '{': | |
deca31e1 | 411 | j = scm_read_token (c, tok_buf, port, 1); |
38ae064c | 412 | return scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
0f2d19dd JB |
413 | |
414 | case '\\': | |
b7f3516f | 415 | c = scm_getc (port); |
deca31e1 | 416 | j = scm_read_token (c, tok_buf, port, 0); |
0f2d19dd | 417 | if (j == 1) |
7866a09b | 418 | return SCM_MAKE_CHAR (c); |
0f2d19dd JB |
419 | if (c >= '0' && c < '8') |
420 | { | |
405aaef9 | 421 | p = scm_istr2int (SCM_STRING_CHARS (*tok_buf), (long) j, 8); |
36284627 | 422 | if (!SCM_FALSEP (p)) |
7866a09b | 423 | return SCM_MAKE_CHAR (SCM_INUM (p)); |
0f2d19dd JB |
424 | } |
425 | for (c = 0; c < scm_n_charnames; c++) | |
426 | if (scm_charnames[c] | |
405aaef9 | 427 | && (scm_casei_streq (scm_charnames[c], SCM_STRING_CHARS (*tok_buf)))) |
7866a09b | 428 | return SCM_MAKE_CHAR (scm_charnums[c]); |
db4b4ca6 | 429 | SCM_MISC_ERROR ("unknown # object", SCM_EOL); |
0f2d19dd | 430 | |
50a095f1 JB |
431 | /* #:SYMBOL is a syntax for keywords supported in all contexts. */ |
432 | case ':': | |
433 | j = scm_read_token ('-', tok_buf, port, 0); | |
38ae064c DH |
434 | p = scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
435 | return scm_make_keyword_from_dash_symbol (p); | |
0f2d19dd JB |
436 | |
437 | default: | |
438 | callshrp: | |
deca31e1 GH |
439 | { |
440 | SCM sharp = scm_get_hash_procedure (c); | |
441 | ||
36284627 | 442 | if (!SCM_FALSEP (sharp)) |
deca31e1 GH |
443 | { |
444 | int line = SCM_LINUM (port); | |
445 | int column = SCM_COL (port) - 2; | |
446 | SCM got; | |
447 | ||
fdc28395 | 448 | got = scm_call_2 (sharp, SCM_MAKE_CHAR (c), port); |
54778cd3 | 449 | if (SCM_EQ_P (got, SCM_UNSPECIFIED)) |
deca31e1 GH |
450 | goto unkshrp; |
451 | if (SCM_RECORD_POSITIONS_P) | |
452 | return *copy = recsexpr (got, line, column, | |
453 | SCM_FILENAME (port)); | |
454 | else | |
455 | return got; | |
456 | } | |
457 | } | |
03bc4386 | 458 | unkshrp: |
b858464a MG |
459 | scm_misc_error (s_scm_read, "Unknown # object: ~S", |
460 | SCM_LIST1 (SCM_MAKE_CHAR (c))); | |
0f2d19dd JB |
461 | } |
462 | ||
463 | case '"': | |
464 | j = 0; | |
b7f3516f | 465 | while ('"' != (c = scm_getc (port))) |
0f2d19dd | 466 | { |
b3fcac34 DH |
467 | if (c == EOF) |
468 | SCM_MISC_ERROR ("end of file in string constant", SCM_EOL); | |
0f2d19dd | 469 | |
94115ae3 | 470 | while (j + 2 >= SCM_STRING_LENGTH (*tok_buf)) |
0f2d19dd JB |
471 | scm_grow_tok_buf (tok_buf); |
472 | ||
473 | if (c == '\\') | |
b7f3516f | 474 | switch (c = scm_getc (port)) |
0f2d19dd JB |
475 | { |
476 | case '\n': | |
477 | continue; | |
478 | case '0': | |
479 | c = '\0'; | |
480 | break; | |
481 | case 'f': | |
482 | c = '\f'; | |
483 | break; | |
484 | case 'n': | |
485 | c = '\n'; | |
486 | break; | |
487 | case 'r': | |
488 | c = '\r'; | |
489 | break; | |
490 | case 't': | |
491 | c = '\t'; | |
492 | break; | |
493 | case 'a': | |
494 | c = '\007'; | |
495 | break; | |
496 | case 'v': | |
497 | c = '\v'; | |
498 | break; | |
499 | } | |
405aaef9 | 500 | SCM_STRING_CHARS (*tok_buf)[j] = c; |
b7f3516f | 501 | ++j; |
0f2d19dd JB |
502 | } |
503 | if (j == 0) | |
504 | return scm_nullstr; | |
405aaef9 | 505 | SCM_STRING_CHARS (*tok_buf)[j] = 0; |
36284627 | 506 | return scm_mem2string (SCM_STRING_CHARS (*tok_buf), j); |
0f2d19dd | 507 | |
b858464a | 508 | case'0':case '1':case '2':case '3':case '4': |
0f2d19dd JB |
509 | case '5':case '6':case '7':case '8':case '9': |
510 | case '.': | |
511 | case '-': | |
512 | case '+': | |
513 | num: | |
b858464a MG |
514 | j = scm_read_token (c, tok_buf, port, 0); |
515 | p = scm_istring2number (SCM_STRING_CHARS (*tok_buf), (long) j, 10L); | |
516 | if (!SCM_FALSEP (p)) | |
517 | return p; | |
518 | if (c == '#') | |
519 | { | |
520 | if ((j == 2) && (scm_getc (port) == '(')) | |
521 | { | |
522 | scm_ungetc ('(', port); | |
523 | c = SCM_STRING_CHARS (*tok_buf)[1]; | |
524 | goto callshrp; | |
525 | } | |
526 | SCM_MISC_ERROR ("unknown # object", SCM_EOL); | |
527 | } | |
528 | goto tok; | |
0f2d19dd JB |
529 | |
530 | case ':': | |
fee7ef83 | 531 | if (SCM_EQ_P (SCM_PACK (SCM_KEYWORD_STYLE), scm_keyword_prefix)) |
c7733771 GH |
532 | { |
533 | j = scm_read_token ('-', tok_buf, port, 0); | |
38ae064c DH |
534 | p = scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
535 | return scm_make_keyword_from_dash_symbol (p); | |
c7733771 GH |
536 | } |
537 | /* fallthrough */ | |
0f2d19dd | 538 | default: |
deca31e1 | 539 | j = scm_read_token (c, tok_buf, port, 0); |
0f2d19dd JB |
540 | /* fallthrough */ |
541 | ||
542 | tok: | |
38ae064c | 543 | return scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
0f2d19dd JB |
544 | } |
545 | } | |
db4b4ca6 DH |
546 | #undef FUNC_NAME |
547 | ||
0f2d19dd JB |
548 | |
549 | #ifdef _UNICOS | |
550 | _Pragma ("noopt"); /* # pragma _CRI noopt */ | |
551 | #endif | |
1cc91f1b | 552 | |
1be6b49c | 553 | size_t |
6e8d25a6 | 554 | scm_read_token (int ic, SCM *tok_buf, SCM port, int weird) |
0f2d19dd | 555 | { |
1be6b49c | 556 | register size_t j; |
0f2d19dd JB |
557 | register int c; |
558 | register char *p; | |
559 | ||
deca31e1 | 560 | c = (SCM_CASE_INSENSITIVE_P ? scm_downcase(ic) : ic); |
405aaef9 | 561 | p = SCM_STRING_CHARS (*tok_buf); |
0f2d19dd JB |
562 | |
563 | if (weird) | |
564 | j = 0; | |
565 | else | |
566 | { | |
567 | j = 0; | |
94115ae3 | 568 | while (j + 2 >= SCM_STRING_LENGTH (*tok_buf)) |
0f2d19dd | 569 | p = scm_grow_tok_buf (tok_buf); |
b7f3516f TT |
570 | p[j] = c; |
571 | ++j; | |
0f2d19dd JB |
572 | } |
573 | ||
574 | while (1) | |
575 | { | |
94115ae3 | 576 | while (j + 2 >= SCM_STRING_LENGTH (*tok_buf)) |
0f2d19dd | 577 | p = scm_grow_tok_buf (tok_buf); |
b7f3516f | 578 | c = scm_getc (port); |
0f2d19dd JB |
579 | switch (c) |
580 | { | |
581 | case '(': | |
582 | case ')': | |
583 | case '"': | |
584 | case ';': | |
585 | case SCM_WHITE_SPACES: | |
586 | case SCM_LINE_INCREMENTORS: | |
587 | if (weird) | |
588 | goto default_case; | |
589 | ||
b7f3516f | 590 | scm_ungetc (c, port); |
0f2d19dd JB |
591 | case EOF: |
592 | eof_case: | |
593 | p[j] = 0; | |
594 | return j; | |
595 | case '\\': | |
596 | if (!weird) | |
597 | goto default_case; | |
598 | else | |
599 | { | |
b7f3516f | 600 | c = scm_getc (port); |
0f2d19dd JB |
601 | if (c == EOF) |
602 | goto eof_case; | |
603 | else | |
604 | goto default_case; | |
605 | } | |
606 | case '}': | |
607 | if (!weird) | |
608 | goto default_case; | |
609 | ||
b7f3516f | 610 | c = scm_getc (port); |
0f2d19dd JB |
611 | if (c == '#') |
612 | { | |
613 | p[j] = 0; | |
614 | return j; | |
615 | } | |
616 | else | |
617 | { | |
b7f3516f | 618 | scm_ungetc (c, port); |
0f2d19dd JB |
619 | c = '}'; |
620 | goto default_case; | |
621 | } | |
622 | ||
623 | default: | |
624 | default_case: | |
625 | { | |
deca31e1 | 626 | c = (SCM_CASE_INSENSITIVE_P ? scm_downcase(c) : c); |
b7f3516f TT |
627 | p[j] = c; |
628 | ++j; | |
0f2d19dd JB |
629 | } |
630 | ||
631 | } | |
632 | } | |
633 | } | |
1cc91f1b | 634 | |
0f2d19dd JB |
635 | #ifdef _UNICOS |
636 | _Pragma ("opt"); /* # pragma _CRI opt */ | |
637 | #endif | |
638 | ||
0f2d19dd | 639 | SCM |
6e8d25a6 | 640 | scm_lreadparen (SCM *tok_buf, SCM port, char *name, SCM *copy) |
db4b4ca6 | 641 | #define FUNC_NAME "scm_lreadparen" |
0f2d19dd JB |
642 | { |
643 | SCM tmp; | |
644 | SCM tl; | |
645 | SCM ans; | |
646 | int c; | |
647 | ||
648 | c = scm_flush_ws (port, name); | |
649 | if (')' == c) | |
650 | return SCM_EOL; | |
b7f3516f | 651 | scm_ungetc (c, port); |
54778cd3 | 652 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
0f2d19dd | 653 | { |
deca31e1 | 654 | ans = scm_lreadr (tok_buf, port, copy); |
0f2d19dd JB |
655 | closeit: |
656 | if (')' != (c = scm_flush_ws (port, name))) | |
db4b4ca6 | 657 | SCM_MISC_ERROR ("missing close paren", SCM_EOL); |
0f2d19dd JB |
658 | return ans; |
659 | } | |
660 | ans = tl = scm_cons (tmp, SCM_EOL); | |
661 | while (')' != (c = scm_flush_ws (port, name))) | |
662 | { | |
b7f3516f | 663 | scm_ungetc (c, port); |
54778cd3 | 664 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
0f2d19dd | 665 | { |
deca31e1 | 666 | SCM_SETCDR (tl, scm_lreadr (tok_buf, port, copy)); |
0f2d19dd JB |
667 | goto closeit; |
668 | } | |
a6c64c3c MD |
669 | SCM_SETCDR (tl, scm_cons (tmp, SCM_EOL)); |
670 | tl = SCM_CDR (tl); | |
0f2d19dd JB |
671 | } |
672 | return ans; | |
673 | } | |
db4b4ca6 | 674 | #undef FUNC_NAME |
0f2d19dd | 675 | |
1cc91f1b | 676 | |
09a4f039 | 677 | SCM |
6e8d25a6 | 678 | scm_lreadrecparen (SCM *tok_buf, SCM port, char *name, SCM *copy) |
db4b4ca6 | 679 | #define FUNC_NAME "scm_lreadrecparen" |
09a4f039 MD |
680 | { |
681 | register int c; | |
682 | register SCM tmp; | |
4dc2435a JB |
683 | register SCM tl, tl2 = SCM_EOL; |
684 | SCM ans, ans2 = SCM_EOL; | |
09a4f039 MD |
685 | /* Need to capture line and column numbers here. */ |
686 | int line = SCM_LINUM (port); | |
687 | int column = SCM_COL (port) - 1; | |
688 | ||
689 | c = scm_flush_ws (port, name); | |
690 | if (')' == c) | |
691 | return SCM_EOL; | |
b7f3516f | 692 | scm_ungetc (c, port); |
54778cd3 | 693 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
09a4f039 | 694 | { |
deca31e1 | 695 | ans = scm_lreadr (tok_buf, port, copy); |
09a4f039 | 696 | if (')' != (c = scm_flush_ws (port, name))) |
db4b4ca6 | 697 | SCM_MISC_ERROR ("missing close paren", SCM_EOL); |
09a4f039 MD |
698 | return ans; |
699 | } | |
700 | /* Build the head of the list structure. */ | |
701 | ans = tl = scm_cons (tmp, SCM_EOL); | |
702 | if (SCM_COPY_SOURCE_P) | |
0c95b57d | 703 | ans2 = tl2 = scm_cons (SCM_CONSP (tmp) |
09a4f039 MD |
704 | ? *copy |
705 | : tmp, | |
706 | SCM_EOL); | |
707 | while (')' != (c = scm_flush_ws (port, name))) | |
708 | { | |
62850ef3 DH |
709 | SCM new_tail; |
710 | ||
b7f3516f | 711 | scm_ungetc (c, port); |
54778cd3 | 712 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
09a4f039 | 713 | { |
deca31e1 | 714 | SCM_SETCDR (tl, tmp = scm_lreadr (tok_buf, port, copy)); |
09a4f039 | 715 | if (SCM_COPY_SOURCE_P) |
0c95b57d | 716 | SCM_SETCDR (tl2, scm_cons (SCM_CONSP (tmp) |
09a4f039 MD |
717 | ? *copy |
718 | : tmp, | |
719 | SCM_EOL)); | |
720 | if (')' != (c = scm_flush_ws (port, name))) | |
db4b4ca6 | 721 | SCM_MISC_ERROR ("missing close paren", SCM_EOL); |
09a4f039 MD |
722 | goto exit; |
723 | } | |
62850ef3 DH |
724 | |
725 | new_tail = scm_cons (tmp, SCM_EOL); | |
726 | SCM_SETCDR (tl, new_tail); | |
727 | tl = new_tail; | |
728 | ||
09a4f039 | 729 | if (SCM_COPY_SOURCE_P) |
62850ef3 DH |
730 | { |
731 | SCM new_tail2 = scm_cons (SCM_CONSP (tmp) ? *copy : tmp, SCM_EOL); | |
732 | SCM_SETCDR (tl2, new_tail2); | |
733 | tl2 = new_tail2; | |
734 | } | |
09a4f039 MD |
735 | } |
736 | exit: | |
737 | scm_whash_insert (scm_source_whash, | |
738 | ans, | |
739 | scm_make_srcprops (line, | |
740 | column, | |
741 | SCM_FILENAME (port), | |
742 | SCM_COPY_SOURCE_P | |
743 | ? *copy = ans2 | |
744 | : SCM_UNDEFINED, | |
745 | SCM_EOL)); | |
746 | return ans; | |
747 | } | |
db4b4ca6 | 748 | #undef FUNC_NAME |
09a4f039 | 749 | |
0f2d19dd JB |
750 | |
751 | \f | |
752 | ||
14de3b42 GH |
753 | /* Manipulate the read-hash-procedures alist. This could be written in |
754 | Scheme, but maybe it will also be used by C code during initialisation. */ | |
a1ec6916 | 755 | SCM_DEFINE (scm_read_hash_extend, "read-hash-extend", 2, 0, 0, |
1bbd0b84 | 756 | (SCM chr, SCM proc), |
dc7fa443 MG |
757 | "Install the procedure @var{proc} for reading expressions\n" |
758 | "starting with the character sequence @code{#} and @var{chr}.\n" | |
759 | "@var{proc} will be called with two arguments: the character\n" | |
760 | "@var{chr} and the port to read further data from. The object\n" | |
761 | "returned will be the return value of @code{read}.") | |
1bbd0b84 | 762 | #define FUNC_NAME s_scm_read_hash_extend |
deca31e1 | 763 | { |
fed9c9a2 GH |
764 | SCM this; |
765 | SCM prev; | |
766 | ||
36284627 DH |
767 | SCM_VALIDATE_CHAR (1, chr); |
768 | SCM_ASSERT (SCM_FALSEP (proc) | |
769 | || SCM_EQ_P (scm_procedure_p (proc), SCM_BOOL_T), | |
770 | proc, SCM_ARG2, FUNC_NAME); | |
fed9c9a2 | 771 | |
14de3b42 GH |
772 | /* Check if chr is already in the alist. */ |
773 | this = *scm_read_hash_procedures; | |
774 | prev = SCM_BOOL_F; | |
fed9c9a2 GH |
775 | while (1) |
776 | { | |
777 | if (SCM_NULLP (this)) | |
778 | { | |
779 | /* not found, so add it to the beginning. */ | |
36284627 | 780 | if (!SCM_FALSEP (proc)) |
fed9c9a2 | 781 | { |
14de3b42 GH |
782 | *scm_read_hash_procedures = |
783 | scm_cons (scm_cons (chr, proc), *scm_read_hash_procedures); | |
fed9c9a2 GH |
784 | } |
785 | break; | |
786 | } | |
54778cd3 | 787 | if (SCM_EQ_P (chr, SCM_CAAR (this))) |
fed9c9a2 GH |
788 | { |
789 | /* already in the alist. */ | |
790 | if (SCM_FALSEP (proc)) | |
14de3b42 GH |
791 | { |
792 | /* remove it. */ | |
54778cd3 | 793 | if (SCM_FALSEP (prev)) |
14de3b42 GH |
794 | { |
795 | *scm_read_hash_procedures = | |
796 | SCM_CDR (*scm_read_hash_procedures); | |
797 | } | |
798 | else | |
799 | scm_set_cdr_x (prev, SCM_CDR (this)); | |
800 | } | |
fed9c9a2 | 801 | else |
14de3b42 GH |
802 | { |
803 | /* replace it. */ | |
804 | scm_set_cdr_x (SCM_CAR (this), proc); | |
805 | } | |
fed9c9a2 GH |
806 | break; |
807 | } | |
808 | prev = this; | |
809 | this = SCM_CDR (this); | |
810 | } | |
deca31e1 | 811 | |
deca31e1 GH |
812 | return SCM_UNSPECIFIED; |
813 | } | |
1bbd0b84 | 814 | #undef FUNC_NAME |
0f2d19dd | 815 | |
deca31e1 GH |
816 | /* Recover the read-hash procedure corresponding to char c. */ |
817 | static SCM | |
6e8d25a6 | 818 | scm_get_hash_procedure (int c) |
deca31e1 | 819 | { |
14de3b42 | 820 | SCM rest = *scm_read_hash_procedures; |
fed9c9a2 | 821 | |
deca31e1 GH |
822 | while (1) |
823 | { | |
824 | if (SCM_NULLP (rest)) | |
825 | return SCM_BOOL_F; | |
826 | ||
7866a09b | 827 | if (SCM_CHAR (SCM_CAAR (rest)) == c) |
deca31e1 GH |
828 | return SCM_CDAR (rest); |
829 | ||
830 | rest = SCM_CDR (rest); | |
831 | } | |
832 | } | |
1cc91f1b | 833 | |
0f2d19dd JB |
834 | void |
835 | scm_init_read () | |
0f2d19dd | 836 | { |
14de3b42 | 837 | scm_read_hash_procedures = |
86d31dfe | 838 | SCM_VARIABLE_LOC (scm_c_define ("read-hash-procedures", SCM_EOL)); |
fed9c9a2 | 839 | |
b7ff98dd | 840 | scm_init_opts (scm_read_options, scm_read_opts, SCM_N_READ_OPTIONS); |
8dc9439f | 841 | #ifndef SCM_MAGIC_SNARFER |
a0599745 | 842 | #include "libguile/read.x" |
8dc9439f | 843 | #endif |
0f2d19dd | 844 | } |
89e00824 ML |
845 | |
846 | /* | |
847 | Local Variables: | |
848 | c-file-style: "gnu" | |
849 | End: | |
850 | */ |