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 | 41 | |
1bbd0b84 | 42 | |
0f2d19dd JB |
43 | \f |
44 | ||
0f2d19dd | 45 | #include <stdio.h> |
a0599745 MD |
46 | #include "libguile/_scm.h" |
47 | #include "libguile/chars.h" | |
48 | #include "libguile/eval.h" | |
49 | #include "libguile/unif.h" | |
50 | #include "libguile/keywords.h" | |
51 | #include "libguile/alist.h" | |
52 | #include "libguile/srcprop.h" | |
53 | #include "libguile/hashtab.h" | |
54 | #include "libguile/hash.h" | |
55 | #include "libguile/ports.h" | |
56 | #include "libguile/root.h" | |
57 | #include "libguile/strings.h" | |
ba1b2226 | 58 | #include "libguile/strports.h" |
a0599745 | 59 | #include "libguile/vectors.h" |
a0599745 | 60 | #include "libguile/validate.h" |
ba1b2226 | 61 | |
a0599745 | 62 | #include "libguile/read.h" |
0f2d19dd JB |
63 | |
64 | \f | |
65 | ||
c7733771 GH |
66 | SCM_SYMBOL (scm_keyword_prefix, "prefix"); |
67 | ||
92c2555f | 68 | scm_t_option scm_read_opts[] = { |
b7ff98dd MD |
69 | { SCM_OPTION_BOOLEAN, "copy", 0, |
70 | "Copy source code expressions." }, | |
ac74fc22 | 71 | { SCM_OPTION_BOOLEAN, "positions", 0, |
deca31e1 GH |
72 | "Record positions of source code expressions." }, |
73 | { SCM_OPTION_BOOLEAN, "case-insensitive", 0, | |
c7733771 | 74 | "Convert symbols to lower case."}, |
f1267706 | 75 | { SCM_OPTION_SCM, "keywords", SCM_UNPACK (SCM_BOOL_F), |
c96d76b8 | 76 | "Style of keyword recognition: #f or 'prefix."} |
a16f6fe7 MD |
77 | }; |
78 | ||
39e8f371 HWN |
79 | /* |
80 | Give meaningful error messages for errors | |
81 | ||
82 | We use the format | |
83 | ||
ba1b2226 | 84 | FILE:LINE:COL: MESSAGE |
39e8f371 HWN |
85 | This happened in .... |
86 | ||
87 | This is not standard GNU format, but the test-suite likes the real | |
88 | message to be in front. | |
89 | ||
39e8f371 HWN |
90 | */ |
91 | ||
92 | ||
ba1b2226 HWN |
93 | static void |
94 | scm_input_error(char const * function, | |
95 | SCM port, const char * message, SCM arg) | |
96 | { | |
97 | char *fn = SCM_STRINGP (SCM_FILENAME(port)) | |
98 | ? SCM_STRING_CHARS(SCM_FILENAME(port)) | |
99 | : "#<unknown port>"; | |
100 | ||
101 | SCM string_port = scm_open_output_string (); | |
102 | SCM string = SCM_EOL; | |
103 | scm_simple_format (string_port, | |
104 | scm_makfrom0str ("~A:~S:~S: ~A"), | |
105 | scm_list_4 (scm_makfrom0str (fn), | |
106 | scm_int2num (SCM_LINUM (port) + 1), | |
107 | scm_int2num (SCM_COL (port) + 1), | |
108 | scm_makfrom0str (message))); | |
39e8f371 | 109 | |
ba1b2226 HWN |
110 | |
111 | string = scm_get_output_string (string_port); | |
112 | scm_close_output_port (string_port); | |
113 | scm_error_scm (scm_str2symbol ("read-error"), | |
114 | scm_makfrom0str (function), | |
115 | string, | |
116 | SCM_EOL, | |
117 | SCM_BOOL_F); | |
118 | } | |
39e8f371 HWN |
119 | |
120 | ||
a1ec6916 | 121 | SCM_DEFINE (scm_read_options, "read-options-interface", 0, 1, 0, |
1bbd0b84 | 122 | (SCM setting), |
dc7fa443 MG |
123 | "Option interface for the read options. Instead of using\n" |
124 | "this procedure directly, use the procedures @code{read-enable},\n" | |
3939e9df | 125 | "@code{read-disable}, @code{read-set!} and @code{read-options}.") |
1bbd0b84 | 126 | #define FUNC_NAME s_scm_read_options |
a16f6fe7 | 127 | { |
b7ff98dd MD |
128 | SCM ans = scm_options (setting, |
129 | scm_read_opts, | |
130 | SCM_N_READ_OPTIONS, | |
1bbd0b84 | 131 | FUNC_NAME); |
b7ff98dd MD |
132 | if (SCM_COPY_SOURCE_P) |
133 | SCM_RECORD_POSITIONS_P = 1; | |
a16f6fe7 MD |
134 | return ans; |
135 | } | |
1bbd0b84 | 136 | #undef FUNC_NAME |
a16f6fe7 | 137 | |
14de3b42 GH |
138 | /* An association list mapping extra hash characters to procedures. */ |
139 | static SCM *scm_read_hash_procedures; | |
deca31e1 | 140 | |
a1ec6916 | 141 | SCM_DEFINE (scm_read, "read", 0, 1, 0, |
1bbd0b84 | 142 | (SCM port), |
dc7fa443 MG |
143 | "Read an s-expression from the input port @var{port}, or from\n" |
144 | "the current input port if @var{port} is not specified.\n" | |
145 | "Any whitespace before the next token is discarded.") | |
1bbd0b84 | 146 | #define FUNC_NAME s_scm_read |
0f2d19dd JB |
147 | { |
148 | int c; | |
09a4f039 | 149 | SCM tok_buf, copy; |
0f2d19dd JB |
150 | |
151 | if (SCM_UNBNDP (port)) | |
152 | port = scm_cur_inp; | |
34d19ef6 | 153 | SCM_VALIDATE_OPINPORT (1, port); |
0f2d19dd | 154 | |
0f2d19dd JB |
155 | c = scm_flush_ws (port, (char *) NULL); |
156 | if (EOF == c) | |
157 | return SCM_EOF_VAL; | |
b7f3516f | 158 | scm_ungetc (c, port); |
0f2d19dd | 159 | |
be54b15d | 160 | tok_buf = scm_allocate_string (30); |
deca31e1 | 161 | return scm_lreadr (&tok_buf, port, ©); |
0f2d19dd | 162 | } |
1bbd0b84 | 163 | #undef FUNC_NAME |
0f2d19dd JB |
164 | |
165 | ||
1cc91f1b | 166 | |
0f2d19dd | 167 | char * |
6e8d25a6 | 168 | scm_grow_tok_buf (SCM *tok_buf) |
0f2d19dd | 169 | { |
1be6b49c | 170 | size_t oldlen = SCM_STRING_LENGTH (*tok_buf); |
be54b15d | 171 | SCM newstr = scm_allocate_string (2 * oldlen); |
1be6b49c | 172 | size_t i; |
94115ae3 DH |
173 | |
174 | for (i = 0; i != oldlen; ++i) | |
175 | SCM_STRING_CHARS (newstr) [i] = SCM_STRING_CHARS (*tok_buf) [i]; | |
176 | ||
177 | *tok_buf = newstr; | |
178 | return SCM_STRING_CHARS (newstr); | |
0f2d19dd JB |
179 | } |
180 | ||
181 | ||
1cc91f1b | 182 | |
0f2d19dd | 183 | int |
6e8d25a6 | 184 | scm_flush_ws (SCM port, const char *eoferr) |
0f2d19dd JB |
185 | { |
186 | register int c; | |
187 | while (1) | |
b7f3516f | 188 | switch (c = scm_getc (port)) |
0f2d19dd JB |
189 | { |
190 | case EOF: | |
191 | goteof: | |
192 | if (eoferr) | |
d156d3b7 | 193 | { |
0f8ae50a HWN |
194 | scm_input_error (eoferr, |
195 | port, | |
196 | "end of file", | |
197 | SCM_EOL); | |
d156d3b7 | 198 | } |
0f2d19dd JB |
199 | return c; |
200 | case ';': | |
201 | lp: | |
b7f3516f | 202 | switch (c = scm_getc (port)) |
0f2d19dd JB |
203 | { |
204 | case EOF: | |
205 | goto goteof; | |
206 | default: | |
207 | goto lp; | |
208 | case SCM_LINE_INCREMENTORS: | |
209 | break; | |
210 | } | |
211 | break; | |
212 | case SCM_LINE_INCREMENTORS: | |
0f2d19dd | 213 | case SCM_SINGLE_SPACES: |
0f2d19dd | 214 | case '\t': |
0f2d19dd JB |
215 | break; |
216 | default: | |
217 | return c; | |
218 | } | |
219 | } | |
220 | ||
221 | ||
1cc91f1b | 222 | |
0f2d19dd | 223 | int |
6e8d25a6 | 224 | scm_casei_streq (char *s1, char *s2) |
0f2d19dd JB |
225 | { |
226 | while (*s1 && *s2) | |
227 | if (scm_downcase((int)*s1) != scm_downcase((int)*s2)) | |
228 | return 0; | |
229 | else | |
230 | { | |
231 | ++s1; | |
232 | ++s2; | |
233 | } | |
234 | return !(*s1 || *s2); | |
235 | } | |
236 | ||
237 | ||
09a4f039 MD |
238 | /* recsexpr is used when recording expressions |
239 | * constructed by read:sharp. | |
240 | */ | |
604d4dd9 JB |
241 | #ifndef DEBUG_EXTENSIONS |
242 | #define recsexpr(obj, line, column, filename) (obj) | |
243 | #else | |
09a4f039 | 244 | static SCM |
1be6b49c | 245 | recsexpr (SCM obj, long line, int column, SCM filename) |
09a4f039 | 246 | { |
fee7ef83 | 247 | if (!SCM_CONSP(obj)) { |
09a4f039 | 248 | return obj; |
fee7ef83 | 249 | } else { |
09a4f039 MD |
250 | SCM tmp = obj, copy; |
251 | /* If this sexpr is visible in the read:sharp source, we want to | |
252 | keep that information, so only record non-constant cons cells | |
253 | which haven't previously been read by the reader. */ | |
254 | if (SCM_FALSEP (scm_whash_lookup (scm_source_whash, obj))) | |
255 | { | |
256 | if (SCM_COPY_SOURCE_P) | |
257 | { | |
258 | copy = scm_cons (recsexpr (SCM_CAR (obj), line, column, filename), | |
259 | SCM_UNDEFINED); | |
0c95b57d | 260 | while ((tmp = SCM_CDR (tmp)) && SCM_CONSP (tmp)) |
a6c64c3c MD |
261 | { |
262 | SCM_SETCDR (copy, scm_cons (recsexpr (SCM_CAR (tmp), | |
263 | line, | |
264 | column, | |
265 | filename), | |
266 | SCM_UNDEFINED)); | |
267 | copy = SCM_CDR (copy); | |
268 | } | |
269 | SCM_SETCDR (copy, tmp); | |
09a4f039 MD |
270 | } |
271 | else | |
272 | { | |
273 | recsexpr (SCM_CAR (obj), line, column, filename); | |
0c95b57d | 274 | while ((tmp = SCM_CDR (tmp)) && SCM_CONSP (tmp)) |
09a4f039 MD |
275 | recsexpr (SCM_CAR (tmp), line, column, filename); |
276 | copy = SCM_UNDEFINED; | |
277 | } | |
278 | scm_whash_insert (scm_source_whash, | |
279 | obj, | |
280 | scm_make_srcprops (line, | |
281 | column, | |
282 | filename, | |
283 | copy, | |
284 | SCM_EOL)); | |
285 | } | |
286 | return obj; | |
287 | } | |
288 | } | |
604d4dd9 | 289 | #endif |
f9c68a47 JB |
290 | |
291 | /* Consume an SCSH-style block comment. Assume that we've already | |
f9731264 JB |
292 | read the initial `#!', and eat characters until we get a |
293 | newline/exclamation-point/sharp-sign/newline sequence. */ | |
f9c68a47 JB |
294 | |
295 | static void | |
6e8d25a6 | 296 | skip_scsh_block_comment (SCM port) |
db4b4ca6 | 297 | #define FUNC_NAME "skip_scsh_block_comment" |
f9c68a47 | 298 | { |
f9731264 JB |
299 | /* Is this portable? Dear God, spare me from the non-eight-bit |
300 | characters. But is it tasteful? */ | |
301 | long history = 0; | |
f9c68a47 JB |
302 | |
303 | for (;;) | |
304 | { | |
b7f3516f | 305 | int c = scm_getc (port); |
f9c68a47 JB |
306 | |
307 | if (c == EOF) | |
db4b4ca6 | 308 | SCM_MISC_ERROR ("unterminated `#! ... !#' comment", SCM_EOL); |
f9731264 | 309 | history = ((history << 8) | (c & 0xff)) & 0xffffffff; |
f9c68a47 | 310 | |
f9731264 JB |
311 | /* Were the last four characters read "\n!#\n"? */ |
312 | if (history == (('\n' << 24) | ('!' << 16) | ('#' << 8) | '\n')) | |
313 | return; | |
f9c68a47 JB |
314 | } |
315 | } | |
db4b4ca6 DH |
316 | #undef FUNC_NAME |
317 | ||
f9c68a47 | 318 | |
1bbd0b84 | 319 | static SCM scm_get_hash_procedure(int c); |
f9c68a47 | 320 | |
09a4f039 | 321 | static char s_list[]="list"; |
1cc91f1b | 322 | |
0f2d19dd | 323 | SCM |
34d19ef6 | 324 | scm_lreadr (SCM *tok_buf, SCM port, SCM *copy) |
db4b4ca6 | 325 | #define FUNC_NAME "scm_lreadr" |
0f2d19dd JB |
326 | { |
327 | int c; | |
1be6b49c | 328 | size_t j; |
0f2d19dd | 329 | SCM p; |
deca31e1 | 330 | |
b858464a | 331 | tryagain: |
1bbd0b84 | 332 | c = scm_flush_ws (port, s_scm_read); |
b858464a | 333 | tryagain_no_flush_ws: |
0f2d19dd JB |
334 | switch (c) |
335 | { | |
336 | case EOF: | |
337 | return SCM_EOF_VAL; | |
338 | ||
339 | case '(': | |
09a4f039 | 340 | return SCM_RECORD_POSITIONS_P |
b858464a | 341 | ? scm_lreadrecparen (tok_buf, port, s_list, copy) |
c6a040a8 | 342 | : scm_lreadparen (tok_buf, port, s_list, copy SCM_ELISP_CLOSE); |
0f2d19dd | 343 | case ')': |
ba1b2226 | 344 | scm_input_error (FUNC_NAME, port,"unexpected \")\"", SCM_EOL); |
0f2d19dd JB |
345 | goto tryagain; |
346 | ||
c6a040a8 NJ |
347 | #ifdef SCM_ELISP_READ_EXTENSIONS |
348 | case '[': | |
349 | p = scm_lreadparen (tok_buf, port, "vector", copy, ']'); | |
350 | return SCM_NULLP (p) ? scm_nullvect : scm_vector (p); | |
351 | #endif | |
0f2d19dd | 352 | case '\'': |
92e5aa0e | 353 | p = scm_sym_quote; |
09a4f039 | 354 | goto recquote; |
0f2d19dd | 355 | case '`': |
92e5aa0e | 356 | p = scm_sym_quasiquote; |
09a4f039 | 357 | goto recquote; |
0f2d19dd | 358 | case ',': |
b7f3516f | 359 | c = scm_getc (port); |
0f2d19dd | 360 | if ('@' == c) |
92e5aa0e | 361 | p = scm_sym_uq_splicing; |
0f2d19dd JB |
362 | else |
363 | { | |
b7f3516f | 364 | scm_ungetc (c, port); |
92e5aa0e | 365 | p = scm_sym_unquote; |
0f2d19dd | 366 | } |
09a4f039 MD |
367 | recquote: |
368 | p = scm_cons2 (p, | |
deca31e1 | 369 | scm_lreadr (tok_buf, port, copy), |
09a4f039 MD |
370 | SCM_EOL); |
371 | if (SCM_RECORD_POSITIONS_P) | |
372 | scm_whash_insert (scm_source_whash, | |
373 | p, | |
374 | scm_make_srcprops (SCM_LINUM (port), | |
375 | SCM_COL (port) - 1, | |
376 | SCM_FILENAME (port), | |
377 | SCM_COPY_SOURCE_P | |
378 | ? (*copy = scm_cons2 (SCM_CAR (p), | |
379 | SCM_CAR (SCM_CDR (p)), | |
380 | SCM_EOL)) | |
381 | : SCM_UNDEFINED, | |
382 | SCM_EOL)); | |
383 | return p; | |
0f2d19dd | 384 | case '#': |
b7f3516f | 385 | c = scm_getc (port); |
b858464a MG |
386 | |
387 | { | |
388 | /* Check for user-defined hash procedure first, to allow | |
389 | overriding of builtin hash read syntaxes. */ | |
390 | SCM sharp = scm_get_hash_procedure (c); | |
391 | if (!SCM_FALSEP (sharp)) | |
392 | { | |
393 | int line = SCM_LINUM (port); | |
394 | int column = SCM_COL (port) - 2; | |
395 | SCM got; | |
396 | ||
397 | got = scm_call_2 (sharp, SCM_MAKE_CHAR (c), port); | |
398 | if (SCM_EQ_P (got, SCM_UNSPECIFIED)) | |
89759084 | 399 | goto handle_sharp; |
b858464a MG |
400 | if (SCM_RECORD_POSITIONS_P) |
401 | return *copy = recsexpr (got, line, column, | |
402 | SCM_FILENAME (port)); | |
403 | else | |
404 | return got; | |
405 | } | |
406 | } | |
89759084 | 407 | handle_sharp: |
0f2d19dd JB |
408 | switch (c) |
409 | { | |
410 | case '(': | |
c6a040a8 | 411 | p = scm_lreadparen (tok_buf, port, "vector", copy SCM_ELISP_CLOSE); |
0f2d19dd JB |
412 | return SCM_NULLP (p) ? scm_nullvect : scm_vector (p); |
413 | ||
414 | case 't': | |
415 | case 'T': | |
416 | return SCM_BOOL_T; | |
417 | case 'f': | |
418 | case 'F': | |
419 | return SCM_BOOL_F; | |
420 | ||
421 | case 'b': | |
422 | case 'B': | |
423 | case 'o': | |
424 | case 'O': | |
425 | case 'd': | |
426 | case 'D': | |
427 | case 'x': | |
428 | case 'X': | |
429 | case 'i': | |
430 | case 'I': | |
431 | case 'e': | |
432 | case 'E': | |
b7f3516f | 433 | scm_ungetc (c, port); |
0f2d19dd JB |
434 | c = '#'; |
435 | goto num; | |
436 | ||
f9c68a47 JB |
437 | case '!': |
438 | /* start of a shell script. Parse as a block comment, | |
439 | terminated by !#, just like SCSH. */ | |
440 | skip_scsh_block_comment (port); | |
b6356af7 MV |
441 | /* EOF is not an error here */ |
442 | c = scm_flush_ws (port, (char *)NULL); | |
443 | goto tryagain_no_flush_ws; | |
f9c68a47 | 444 | |
971d1e36 | 445 | #if SCM_HAVE_ARRAYS |
0f2d19dd | 446 | case '*': |
deca31e1 | 447 | j = scm_read_token (c, tok_buf, port, 0); |
405aaef9 | 448 | p = scm_istr2bve (SCM_STRING_CHARS (*tok_buf) + 1, (long) (j - 1)); |
36284627 | 449 | if (!SCM_FALSEP (p)) |
0f2d19dd JB |
450 | return p; |
451 | else | |
452 | goto unkshrp; | |
afe5177e | 453 | #endif |
0f2d19dd JB |
454 | |
455 | case '{': | |
deca31e1 | 456 | j = scm_read_token (c, tok_buf, port, 1); |
38ae064c | 457 | return scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
0f2d19dd JB |
458 | |
459 | case '\\': | |
b7f3516f | 460 | c = scm_getc (port); |
deca31e1 | 461 | j = scm_read_token (c, tok_buf, port, 0); |
0f2d19dd | 462 | if (j == 1) |
7866a09b | 463 | return SCM_MAKE_CHAR (c); |
0f2d19dd JB |
464 | if (c >= '0' && c < '8') |
465 | { | |
3c9a524f DH |
466 | /* Dirk:FIXME:: This type of character syntax is not R5RS |
467 | * compliant. Further, it should be verified that the constant | |
468 | * does only consist of octal digits. Finally, it should be | |
469 | * checked whether the resulting fixnum is in the range of | |
470 | * characters. */ | |
471 | p = scm_i_mem2number (SCM_STRING_CHARS (*tok_buf), j, 8); | |
472 | if (SCM_INUMP (p)) | |
7866a09b | 473 | return SCM_MAKE_CHAR (SCM_INUM (p)); |
0f2d19dd JB |
474 | } |
475 | for (c = 0; c < scm_n_charnames; c++) | |
476 | if (scm_charnames[c] | |
405aaef9 | 477 | && (scm_casei_streq (scm_charnames[c], SCM_STRING_CHARS (*tok_buf)))) |
7866a09b | 478 | return SCM_MAKE_CHAR (scm_charnums[c]); |
ba1b2226 | 479 | scm_input_error (FUNC_NAME, port, "unknown # object", SCM_EOL); |
0f2d19dd | 480 | |
50a095f1 JB |
481 | /* #:SYMBOL is a syntax for keywords supported in all contexts. */ |
482 | case ':': | |
483 | j = scm_read_token ('-', tok_buf, port, 0); | |
38ae064c DH |
484 | p = scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
485 | return scm_make_keyword_from_dash_symbol (p); | |
0f2d19dd JB |
486 | |
487 | default: | |
488 | callshrp: | |
deca31e1 GH |
489 | { |
490 | SCM sharp = scm_get_hash_procedure (c); | |
491 | ||
36284627 | 492 | if (!SCM_FALSEP (sharp)) |
deca31e1 GH |
493 | { |
494 | int line = SCM_LINUM (port); | |
495 | int column = SCM_COL (port) - 2; | |
496 | SCM got; | |
497 | ||
fdc28395 | 498 | got = scm_call_2 (sharp, SCM_MAKE_CHAR (c), port); |
54778cd3 | 499 | if (SCM_EQ_P (got, SCM_UNSPECIFIED)) |
deca31e1 GH |
500 | goto unkshrp; |
501 | if (SCM_RECORD_POSITIONS_P) | |
502 | return *copy = recsexpr (got, line, column, | |
503 | SCM_FILENAME (port)); | |
504 | else | |
505 | return got; | |
506 | } | |
507 | } | |
03bc4386 | 508 | unkshrp: |
ba1b2226 | 509 | scm_input_error (FUNC_NAME, port, "Unknown # object: ~S", |
39e8f371 | 510 | scm_list_1 (SCM_MAKE_CHAR (c))); |
0f2d19dd JB |
511 | } |
512 | ||
513 | case '"': | |
514 | j = 0; | |
b7f3516f | 515 | while ('"' != (c = scm_getc (port))) |
0f2d19dd | 516 | { |
b3fcac34 | 517 | if (c == EOF) |
ba1b2226 | 518 | scm_input_error (FUNC_NAME, port, "end of file in string constant", SCM_EOL); |
0f2d19dd | 519 | |
94115ae3 | 520 | while (j + 2 >= SCM_STRING_LENGTH (*tok_buf)) |
0f2d19dd JB |
521 | scm_grow_tok_buf (tok_buf); |
522 | ||
523 | if (c == '\\') | |
b7f3516f | 524 | switch (c = scm_getc (port)) |
0f2d19dd JB |
525 | { |
526 | case '\n': | |
527 | continue; | |
528 | case '0': | |
529 | c = '\0'; | |
530 | break; | |
531 | case 'f': | |
532 | c = '\f'; | |
533 | break; | |
534 | case 'n': | |
535 | c = '\n'; | |
536 | break; | |
537 | case 'r': | |
538 | c = '\r'; | |
539 | break; | |
540 | case 't': | |
541 | c = '\t'; | |
542 | break; | |
543 | case 'a': | |
544 | c = '\007'; | |
545 | break; | |
546 | case 'v': | |
547 | c = '\v'; | |
548 | break; | |
549 | } | |
405aaef9 | 550 | SCM_STRING_CHARS (*tok_buf)[j] = c; |
b7f3516f | 551 | ++j; |
0f2d19dd JB |
552 | } |
553 | if (j == 0) | |
554 | return scm_nullstr; | |
405aaef9 | 555 | SCM_STRING_CHARS (*tok_buf)[j] = 0; |
36284627 | 556 | return scm_mem2string (SCM_STRING_CHARS (*tok_buf), j); |
0f2d19dd | 557 | |
3c9a524f DH |
558 | case '0': case '1': case '2': case '3': case '4': |
559 | case '5': case '6': case '7': case '8': case '9': | |
0f2d19dd JB |
560 | case '.': |
561 | case '-': | |
562 | case '+': | |
563 | num: | |
3c9a524f DH |
564 | j = scm_read_token (c, tok_buf, port, 0); |
565 | if (j == 1 && (c == '+' || c == '-')) | |
566 | /* Shortcut: Detected symbol '+ or '- */ | |
567 | goto tok; | |
568 | ||
569 | p = scm_i_mem2number (SCM_STRING_CHARS (*tok_buf), j, 10); | |
570 | if (!SCM_FALSEP (p)) | |
571 | return p; | |
572 | if (c == '#') | |
573 | { | |
574 | if ((j == 2) && (scm_getc (port) == '(')) | |
575 | { | |
576 | scm_ungetc ('(', port); | |
577 | c = SCM_STRING_CHARS (*tok_buf)[1]; | |
578 | goto callshrp; | |
579 | } | |
ba1b2226 | 580 | scm_input_error (FUNC_NAME, port, "unknown # object", SCM_EOL); |
3c9a524f DH |
581 | } |
582 | goto tok; | |
0f2d19dd JB |
583 | |
584 | case ':': | |
fee7ef83 | 585 | if (SCM_EQ_P (SCM_PACK (SCM_KEYWORD_STYLE), scm_keyword_prefix)) |
c7733771 GH |
586 | { |
587 | j = scm_read_token ('-', tok_buf, port, 0); | |
38ae064c DH |
588 | p = scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
589 | return scm_make_keyword_from_dash_symbol (p); | |
c7733771 GH |
590 | } |
591 | /* fallthrough */ | |
0f2d19dd | 592 | default: |
deca31e1 | 593 | j = scm_read_token (c, tok_buf, port, 0); |
0f2d19dd JB |
594 | /* fallthrough */ |
595 | ||
596 | tok: | |
38ae064c | 597 | return scm_mem2symbol (SCM_STRING_CHARS (*tok_buf), j); |
0f2d19dd JB |
598 | } |
599 | } | |
db4b4ca6 DH |
600 | #undef FUNC_NAME |
601 | ||
0f2d19dd JB |
602 | |
603 | #ifdef _UNICOS | |
604 | _Pragma ("noopt"); /* # pragma _CRI noopt */ | |
605 | #endif | |
1cc91f1b | 606 | |
1be6b49c | 607 | size_t |
6e8d25a6 | 608 | scm_read_token (int ic, SCM *tok_buf, SCM port, int weird) |
0f2d19dd | 609 | { |
1be6b49c | 610 | register size_t j; |
0f2d19dd JB |
611 | register int c; |
612 | register char *p; | |
613 | ||
deca31e1 | 614 | c = (SCM_CASE_INSENSITIVE_P ? scm_downcase(ic) : ic); |
405aaef9 | 615 | p = SCM_STRING_CHARS (*tok_buf); |
0f2d19dd JB |
616 | |
617 | if (weird) | |
618 | j = 0; | |
619 | else | |
620 | { | |
621 | j = 0; | |
94115ae3 | 622 | while (j + 2 >= SCM_STRING_LENGTH (*tok_buf)) |
0f2d19dd | 623 | p = scm_grow_tok_buf (tok_buf); |
b7f3516f TT |
624 | p[j] = c; |
625 | ++j; | |
0f2d19dd JB |
626 | } |
627 | ||
628 | while (1) | |
629 | { | |
94115ae3 | 630 | while (j + 2 >= SCM_STRING_LENGTH (*tok_buf)) |
0f2d19dd | 631 | p = scm_grow_tok_buf (tok_buf); |
b7f3516f | 632 | c = scm_getc (port); |
0f2d19dd JB |
633 | switch (c) |
634 | { | |
635 | case '(': | |
636 | case ')': | |
c6a040a8 NJ |
637 | #ifdef SCM_ELISP_READ_EXTENSIONS |
638 | case '[': | |
639 | case ']': | |
640 | #endif | |
0f2d19dd JB |
641 | case '"': |
642 | case ';': | |
643 | case SCM_WHITE_SPACES: | |
644 | case SCM_LINE_INCREMENTORS: | |
645 | if (weird) | |
646 | goto default_case; | |
647 | ||
b7f3516f | 648 | scm_ungetc (c, port); |
0f2d19dd JB |
649 | case EOF: |
650 | eof_case: | |
651 | p[j] = 0; | |
652 | return j; | |
653 | case '\\': | |
654 | if (!weird) | |
655 | goto default_case; | |
656 | else | |
657 | { | |
b7f3516f | 658 | c = scm_getc (port); |
0f2d19dd JB |
659 | if (c == EOF) |
660 | goto eof_case; | |
661 | else | |
662 | goto default_case; | |
663 | } | |
664 | case '}': | |
665 | if (!weird) | |
666 | goto default_case; | |
667 | ||
b7f3516f | 668 | c = scm_getc (port); |
0f2d19dd JB |
669 | if (c == '#') |
670 | { | |
671 | p[j] = 0; | |
672 | return j; | |
673 | } | |
674 | else | |
675 | { | |
b7f3516f | 676 | scm_ungetc (c, port); |
0f2d19dd JB |
677 | c = '}'; |
678 | goto default_case; | |
679 | } | |
680 | ||
681 | default: | |
682 | default_case: | |
683 | { | |
deca31e1 | 684 | c = (SCM_CASE_INSENSITIVE_P ? scm_downcase(c) : c); |
b7f3516f TT |
685 | p[j] = c; |
686 | ++j; | |
0f2d19dd JB |
687 | } |
688 | ||
689 | } | |
690 | } | |
691 | } | |
1cc91f1b | 692 | |
0f2d19dd JB |
693 | #ifdef _UNICOS |
694 | _Pragma ("opt"); /* # pragma _CRI opt */ | |
695 | #endif | |
696 | ||
0f2d19dd | 697 | SCM |
c6a040a8 NJ |
698 | scm_lreadparen (SCM *tok_buf, SCM port, char *name, SCM *copy |
699 | #ifdef SCM_ELISP_READ_EXTENSIONS | |
700 | , char term_char | |
701 | #else | |
702 | #define term_char ')' | |
703 | #endif | |
704 | ) | |
db4b4ca6 | 705 | #define FUNC_NAME "scm_lreadparen" |
0f2d19dd JB |
706 | { |
707 | SCM tmp; | |
708 | SCM tl; | |
709 | SCM ans; | |
710 | int c; | |
711 | ||
712 | c = scm_flush_ws (port, name); | |
c6a040a8 | 713 | if (term_char == c) |
0f2d19dd | 714 | return SCM_EOL; |
b7f3516f | 715 | scm_ungetc (c, port); |
54778cd3 | 716 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
0f2d19dd | 717 | { |
deca31e1 | 718 | ans = scm_lreadr (tok_buf, port, copy); |
0f2d19dd | 719 | closeit: |
c6a040a8 | 720 | if (term_char != (c = scm_flush_ws (port, name))) |
ba1b2226 | 721 | scm_input_error (FUNC_NAME, port, "missing close paren", SCM_EOL); |
0f2d19dd JB |
722 | return ans; |
723 | } | |
724 | ans = tl = scm_cons (tmp, SCM_EOL); | |
c6a040a8 | 725 | while (term_char != (c = scm_flush_ws (port, name))) |
0f2d19dd | 726 | { |
b7f3516f | 727 | scm_ungetc (c, port); |
54778cd3 | 728 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
0f2d19dd | 729 | { |
deca31e1 | 730 | SCM_SETCDR (tl, scm_lreadr (tok_buf, port, copy)); |
0f2d19dd JB |
731 | goto closeit; |
732 | } | |
a6c64c3c MD |
733 | SCM_SETCDR (tl, scm_cons (tmp, SCM_EOL)); |
734 | tl = SCM_CDR (tl); | |
0f2d19dd JB |
735 | } |
736 | return ans; | |
737 | } | |
db4b4ca6 | 738 | #undef FUNC_NAME |
c6a040a8 NJ |
739 | #ifndef SCM_ELISP_READ_EXTENSIONS |
740 | #undef term_char | |
741 | #endif | |
0f2d19dd | 742 | |
1cc91f1b | 743 | |
09a4f039 | 744 | SCM |
6e8d25a6 | 745 | scm_lreadrecparen (SCM *tok_buf, SCM port, char *name, SCM *copy) |
db4b4ca6 | 746 | #define FUNC_NAME "scm_lreadrecparen" |
09a4f039 MD |
747 | { |
748 | register int c; | |
749 | register SCM tmp; | |
4dc2435a JB |
750 | register SCM tl, tl2 = SCM_EOL; |
751 | SCM ans, ans2 = SCM_EOL; | |
09a4f039 MD |
752 | /* Need to capture line and column numbers here. */ |
753 | int line = SCM_LINUM (port); | |
754 | int column = SCM_COL (port) - 1; | |
755 | ||
756 | c = scm_flush_ws (port, name); | |
757 | if (')' == c) | |
758 | return SCM_EOL; | |
b7f3516f | 759 | scm_ungetc (c, port); |
54778cd3 | 760 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
09a4f039 | 761 | { |
deca31e1 | 762 | ans = scm_lreadr (tok_buf, port, copy); |
09a4f039 | 763 | if (')' != (c = scm_flush_ws (port, name))) |
ba1b2226 | 764 | scm_input_error (FUNC_NAME, port, "missing close paren", SCM_EOL); |
09a4f039 MD |
765 | return ans; |
766 | } | |
767 | /* Build the head of the list structure. */ | |
768 | ans = tl = scm_cons (tmp, SCM_EOL); | |
769 | if (SCM_COPY_SOURCE_P) | |
0c95b57d | 770 | ans2 = tl2 = scm_cons (SCM_CONSP (tmp) |
09a4f039 MD |
771 | ? *copy |
772 | : tmp, | |
773 | SCM_EOL); | |
774 | while (')' != (c = scm_flush_ws (port, name))) | |
775 | { | |
62850ef3 DH |
776 | SCM new_tail; |
777 | ||
b7f3516f | 778 | scm_ungetc (c, port); |
54778cd3 | 779 | if (SCM_EQ_P (scm_sym_dot, (tmp = scm_lreadr (tok_buf, port, copy)))) |
09a4f039 | 780 | { |
deca31e1 | 781 | SCM_SETCDR (tl, tmp = scm_lreadr (tok_buf, port, copy)); |
09a4f039 | 782 | if (SCM_COPY_SOURCE_P) |
0c95b57d | 783 | SCM_SETCDR (tl2, scm_cons (SCM_CONSP (tmp) |
09a4f039 MD |
784 | ? *copy |
785 | : tmp, | |
786 | SCM_EOL)); | |
787 | if (')' != (c = scm_flush_ws (port, name))) | |
ba1b2226 | 788 | scm_input_error (FUNC_NAME, port, "missing close paren", SCM_EOL); |
09a4f039 MD |
789 | goto exit; |
790 | } | |
62850ef3 DH |
791 | |
792 | new_tail = scm_cons (tmp, SCM_EOL); | |
793 | SCM_SETCDR (tl, new_tail); | |
794 | tl = new_tail; | |
795 | ||
09a4f039 | 796 | if (SCM_COPY_SOURCE_P) |
62850ef3 DH |
797 | { |
798 | SCM new_tail2 = scm_cons (SCM_CONSP (tmp) ? *copy : tmp, SCM_EOL); | |
799 | SCM_SETCDR (tl2, new_tail2); | |
800 | tl2 = new_tail2; | |
801 | } | |
09a4f039 MD |
802 | } |
803 | exit: | |
804 | scm_whash_insert (scm_source_whash, | |
805 | ans, | |
806 | scm_make_srcprops (line, | |
807 | column, | |
808 | SCM_FILENAME (port), | |
809 | SCM_COPY_SOURCE_P | |
810 | ? *copy = ans2 | |
811 | : SCM_UNDEFINED, | |
812 | SCM_EOL)); | |
813 | return ans; | |
814 | } | |
db4b4ca6 | 815 | #undef FUNC_NAME |
09a4f039 | 816 | |
0f2d19dd JB |
817 | |
818 | \f | |
819 | ||
14de3b42 GH |
820 | /* Manipulate the read-hash-procedures alist. This could be written in |
821 | Scheme, but maybe it will also be used by C code during initialisation. */ | |
a1ec6916 | 822 | SCM_DEFINE (scm_read_hash_extend, "read-hash-extend", 2, 0, 0, |
1bbd0b84 | 823 | (SCM chr, SCM proc), |
dc7fa443 MG |
824 | "Install the procedure @var{proc} for reading expressions\n" |
825 | "starting with the character sequence @code{#} and @var{chr}.\n" | |
826 | "@var{proc} will be called with two arguments: the character\n" | |
827 | "@var{chr} and the port to read further data from. The object\n" | |
828 | "returned will be the return value of @code{read}.") | |
1bbd0b84 | 829 | #define FUNC_NAME s_scm_read_hash_extend |
deca31e1 | 830 | { |
fed9c9a2 GH |
831 | SCM this; |
832 | SCM prev; | |
833 | ||
36284627 DH |
834 | SCM_VALIDATE_CHAR (1, chr); |
835 | SCM_ASSERT (SCM_FALSEP (proc) | |
836 | || SCM_EQ_P (scm_procedure_p (proc), SCM_BOOL_T), | |
837 | proc, SCM_ARG2, FUNC_NAME); | |
fed9c9a2 | 838 | |
14de3b42 GH |
839 | /* Check if chr is already in the alist. */ |
840 | this = *scm_read_hash_procedures; | |
841 | prev = SCM_BOOL_F; | |
fed9c9a2 GH |
842 | while (1) |
843 | { | |
844 | if (SCM_NULLP (this)) | |
845 | { | |
846 | /* not found, so add it to the beginning. */ | |
36284627 | 847 | if (!SCM_FALSEP (proc)) |
fed9c9a2 | 848 | { |
14de3b42 GH |
849 | *scm_read_hash_procedures = |
850 | scm_cons (scm_cons (chr, proc), *scm_read_hash_procedures); | |
fed9c9a2 GH |
851 | } |
852 | break; | |
853 | } | |
54778cd3 | 854 | if (SCM_EQ_P (chr, SCM_CAAR (this))) |
fed9c9a2 GH |
855 | { |
856 | /* already in the alist. */ | |
857 | if (SCM_FALSEP (proc)) | |
14de3b42 GH |
858 | { |
859 | /* remove it. */ | |
54778cd3 | 860 | if (SCM_FALSEP (prev)) |
14de3b42 GH |
861 | { |
862 | *scm_read_hash_procedures = | |
863 | SCM_CDR (*scm_read_hash_procedures); | |
864 | } | |
865 | else | |
866 | scm_set_cdr_x (prev, SCM_CDR (this)); | |
867 | } | |
fed9c9a2 | 868 | else |
14de3b42 GH |
869 | { |
870 | /* replace it. */ | |
871 | scm_set_cdr_x (SCM_CAR (this), proc); | |
872 | } | |
fed9c9a2 GH |
873 | break; |
874 | } | |
875 | prev = this; | |
876 | this = SCM_CDR (this); | |
877 | } | |
deca31e1 | 878 | |
deca31e1 GH |
879 | return SCM_UNSPECIFIED; |
880 | } | |
1bbd0b84 | 881 | #undef FUNC_NAME |
0f2d19dd | 882 | |
deca31e1 GH |
883 | /* Recover the read-hash procedure corresponding to char c. */ |
884 | static SCM | |
6e8d25a6 | 885 | scm_get_hash_procedure (int c) |
deca31e1 | 886 | { |
14de3b42 | 887 | SCM rest = *scm_read_hash_procedures; |
fed9c9a2 | 888 | |
deca31e1 GH |
889 | while (1) |
890 | { | |
891 | if (SCM_NULLP (rest)) | |
892 | return SCM_BOOL_F; | |
893 | ||
7866a09b | 894 | if (SCM_CHAR (SCM_CAAR (rest)) == c) |
deca31e1 GH |
895 | return SCM_CDAR (rest); |
896 | ||
897 | rest = SCM_CDR (rest); | |
898 | } | |
899 | } | |
1cc91f1b | 900 | |
0f2d19dd JB |
901 | void |
902 | scm_init_read () | |
0f2d19dd | 903 | { |
14de3b42 | 904 | scm_read_hash_procedures = |
86d31dfe | 905 | SCM_VARIABLE_LOC (scm_c_define ("read-hash-procedures", SCM_EOL)); |
fed9c9a2 | 906 | |
b7ff98dd | 907 | scm_init_opts (scm_read_options, scm_read_opts, SCM_N_READ_OPTIONS); |
a0599745 | 908 | #include "libguile/read.x" |
0f2d19dd | 909 | } |
89e00824 ML |
910 | |
911 | /* | |
912 | Local Variables: | |
913 | c-file-style: "gnu" | |
914 | End: | |
915 | */ |