Use INTERNAL_FIELD for symbols.
[bpt/emacs.git] / src / lread.c
... / ...
CommitLineData
1/* Lisp parsing and input streams.
2
3Copyright (C) 1985-1989, 1993-1995, 1997-2012 Free Software Foundation, Inc.
4
5This file is part of GNU Emacs.
6
7GNU Emacs is free software: you can redistribute it and/or modify
8it under the terms of the GNU General Public License as published by
9the Free Software Foundation, either version 3 of the License, or
10(at your option) any later version.
11
12GNU Emacs is distributed in the hope that it will be useful,
13but WITHOUT ANY WARRANTY; without even the implied warranty of
14MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15GNU General Public License for more details.
16
17You should have received a copy of the GNU General Public License
18along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */
19
20
21#include <config.h>
22#include <stdio.h>
23#include <sys/types.h>
24#include <sys/stat.h>
25#include <sys/file.h>
26#include <errno.h>
27#include <limits.h> /* For CHAR_BIT. */
28#include <setjmp.h>
29#include <stat-time.h>
30#include "lisp.h"
31#include "intervals.h"
32#include "character.h"
33#include "buffer.h"
34#include "charset.h"
35#include "coding.h"
36#include <epaths.h>
37#include "commands.h"
38#include "keyboard.h"
39#include "frame.h"
40#include "termhooks.h"
41#include "coding.h"
42#include "blockinput.h"
43
44#ifdef MSDOS
45#include "msdos.h"
46#endif
47
48#ifdef HAVE_NS
49#include "nsterm.h"
50#endif
51
52#include <unistd.h>
53#include <math.h>
54
55#ifdef HAVE_SETLOCALE
56#include <locale.h>
57#endif /* HAVE_SETLOCALE */
58
59#include <fcntl.h>
60
61#ifdef HAVE_FSEEKO
62#define file_offset off_t
63#define file_tell ftello
64#else
65#define file_offset long
66#define file_tell ftell
67#endif
68
69/* Hash table read constants. */
70static Lisp_Object Qhash_table, Qdata;
71static Lisp_Object Qtest, Qsize;
72static Lisp_Object Qweakness;
73static Lisp_Object Qrehash_size;
74static Lisp_Object Qrehash_threshold;
75
76static Lisp_Object Qread_char, Qget_file_char, Qcurrent_load_list;
77Lisp_Object Qstandard_input;
78Lisp_Object Qvariable_documentation;
79static Lisp_Object Qascii_character, Qload, Qload_file_name;
80Lisp_Object Qbackquote, Qcomma, Qcomma_at, Qcomma_dot, Qfunction;
81static Lisp_Object Qinhibit_file_name_operation;
82static Lisp_Object Qeval_buffer_list;
83static Lisp_Object Qlexical_binding;
84static Lisp_Object Qfile_truename, Qdo_after_load_evaluation; /* ACM 2006/5/16 */
85
86/* Used instead of Qget_file_char while loading *.elc files compiled
87 by Emacs 21 or older. */
88static Lisp_Object Qget_emacs_mule_file_char;
89
90static Lisp_Object Qload_force_doc_strings;
91
92extern Lisp_Object Qinternal_interpreter_environment;
93
94static Lisp_Object Qload_in_progress;
95
96/* The association list of objects read with the #n=object form.
97 Each member of the list has the form (n . object), and is used to
98 look up the object for the corresponding #n# construct.
99 It must be set to nil before all top-level calls to read0. */
100static Lisp_Object read_objects;
101
102/* Nonzero means READCHAR should read bytes one by one (not character)
103 when READCHARFUN is Qget_file_char or Qget_emacs_mule_file_char.
104 This is set to 1 by read1 temporarily while handling #@NUMBER. */
105static int load_each_byte;
106
107/* List of descriptors now open for Fload. */
108static Lisp_Object load_descriptor_list;
109
110/* File for get_file_char to read from. Use by load. */
111static FILE *instream;
112
113/* For use within read-from-string (this reader is non-reentrant!!) */
114static ptrdiff_t read_from_string_index;
115static ptrdiff_t read_from_string_index_byte;
116static ptrdiff_t read_from_string_limit;
117
118/* Number of characters read in the current call to Fread or
119 Fread_from_string. */
120static EMACS_INT readchar_count;
121
122/* This contains the last string skipped with #@. */
123static char *saved_doc_string;
124/* Length of buffer allocated in saved_doc_string. */
125static ptrdiff_t saved_doc_string_size;
126/* Length of actual data in saved_doc_string. */
127static ptrdiff_t saved_doc_string_length;
128/* This is the file position that string came from. */
129static file_offset saved_doc_string_position;
130
131/* This contains the previous string skipped with #@.
132 We copy it from saved_doc_string when a new string
133 is put in saved_doc_string. */
134static char *prev_saved_doc_string;
135/* Length of buffer allocated in prev_saved_doc_string. */
136static ptrdiff_t prev_saved_doc_string_size;
137/* Length of actual data in prev_saved_doc_string. */
138static ptrdiff_t prev_saved_doc_string_length;
139/* This is the file position that string came from. */
140static file_offset prev_saved_doc_string_position;
141
142/* Nonzero means inside a new-style backquote
143 with no surrounding parentheses.
144 Fread initializes this to zero, so we need not specbind it
145 or worry about what happens to it when there is an error. */
146static int new_backquote_flag;
147static Lisp_Object Qold_style_backquotes;
148
149/* A list of file names for files being loaded in Fload. Used to
150 check for recursive loads. */
151
152static Lisp_Object Vloads_in_progress;
153
154static int read_emacs_mule_char (int, int (*) (int, Lisp_Object),
155 Lisp_Object);
156
157static void readevalloop (Lisp_Object, FILE*, Lisp_Object, int,
158 Lisp_Object, Lisp_Object,
159 Lisp_Object, Lisp_Object);
160static Lisp_Object load_unwind (Lisp_Object);
161static Lisp_Object load_descriptor_unwind (Lisp_Object);
162\f
163/* Functions that read one byte from the current source READCHARFUN
164 or unreads one byte. If the integer argument C is -1, it returns
165 one read byte, or -1 when there's no more byte in the source. If C
166 is 0 or positive, it unreads C, and the return value is not
167 interesting. */
168
169static int readbyte_for_lambda (int, Lisp_Object);
170static int readbyte_from_file (int, Lisp_Object);
171static int readbyte_from_string (int, Lisp_Object);
172
173/* Handle unreading and rereading of characters.
174 Write READCHAR to read a character,
175 UNREAD(c) to unread c to be read again.
176
177 These macros correctly read/unread multibyte characters. */
178
179#define READCHAR readchar (readcharfun, NULL)
180#define UNREAD(c) unreadchar (readcharfun, c)
181
182/* Same as READCHAR but set *MULTIBYTE to the multibyteness of the source. */
183#define READCHAR_REPORT_MULTIBYTE(multibyte) readchar (readcharfun, multibyte)
184
185/* When READCHARFUN is Qget_file_char, Qget_emacs_mule_file_char,
186 Qlambda, or a cons, we use this to keep an unread character because
187 a file stream can't handle multibyte-char unreading. The value -1
188 means that there's no unread character. */
189static int unread_char;
190
191static int
192readchar (Lisp_Object readcharfun, int *multibyte)
193{
194 Lisp_Object tem;
195 register int c;
196 int (*readbyte) (int, Lisp_Object);
197 unsigned char buf[MAX_MULTIBYTE_LENGTH];
198 int i, len;
199 int emacs_mule_encoding = 0;
200
201 if (multibyte)
202 *multibyte = 0;
203
204 readchar_count++;
205
206 if (BUFFERP (readcharfun))
207 {
208 register struct buffer *inbuffer = XBUFFER (readcharfun);
209
210 ptrdiff_t pt_byte = BUF_PT_BYTE (inbuffer);
211
212 if (pt_byte >= BUF_ZV_BYTE (inbuffer))
213 return -1;
214
215 if (! NILP (BVAR (inbuffer, enable_multibyte_characters)))
216 {
217 /* Fetch the character code from the buffer. */
218 unsigned char *p = BUF_BYTE_ADDRESS (inbuffer, pt_byte);
219 BUF_INC_POS (inbuffer, pt_byte);
220 c = STRING_CHAR (p);
221 if (multibyte)
222 *multibyte = 1;
223 }
224 else
225 {
226 c = BUF_FETCH_BYTE (inbuffer, pt_byte);
227 if (! ASCII_BYTE_P (c))
228 c = BYTE8_TO_CHAR (c);
229 pt_byte++;
230 }
231 SET_BUF_PT_BOTH (inbuffer, BUF_PT (inbuffer) + 1, pt_byte);
232
233 return c;
234 }
235 if (MARKERP (readcharfun))
236 {
237 register struct buffer *inbuffer = XMARKER (readcharfun)->buffer;
238
239 ptrdiff_t bytepos = marker_byte_position (readcharfun);
240
241 if (bytepos >= BUF_ZV_BYTE (inbuffer))
242 return -1;
243
244 if (! NILP (BVAR (inbuffer, enable_multibyte_characters)))
245 {
246 /* Fetch the character code from the buffer. */
247 unsigned char *p = BUF_BYTE_ADDRESS (inbuffer, bytepos);
248 BUF_INC_POS (inbuffer, bytepos);
249 c = STRING_CHAR (p);
250 if (multibyte)
251 *multibyte = 1;
252 }
253 else
254 {
255 c = BUF_FETCH_BYTE (inbuffer, bytepos);
256 if (! ASCII_BYTE_P (c))
257 c = BYTE8_TO_CHAR (c);
258 bytepos++;
259 }
260
261 XMARKER (readcharfun)->bytepos = bytepos;
262 XMARKER (readcharfun)->charpos++;
263
264 return c;
265 }
266
267 if (EQ (readcharfun, Qlambda))
268 {
269 readbyte = readbyte_for_lambda;
270 goto read_multibyte;
271 }
272
273 if (EQ (readcharfun, Qget_file_char))
274 {
275 readbyte = readbyte_from_file;
276 goto read_multibyte;
277 }
278
279 if (STRINGP (readcharfun))
280 {
281 if (read_from_string_index >= read_from_string_limit)
282 c = -1;
283 else if (STRING_MULTIBYTE (readcharfun))
284 {
285 if (multibyte)
286 *multibyte = 1;
287 FETCH_STRING_CHAR_ADVANCE_NO_CHECK (c, readcharfun,
288 read_from_string_index,
289 read_from_string_index_byte);
290 }
291 else
292 {
293 c = SREF (readcharfun, read_from_string_index_byte);
294 read_from_string_index++;
295 read_from_string_index_byte++;
296 }
297 return c;
298 }
299
300 if (CONSP (readcharfun))
301 {
302 /* This is the case that read_vector is reading from a unibyte
303 string that contains a byte sequence previously skipped
304 because of #@NUMBER. The car part of readcharfun is that
305 string, and the cdr part is a value of readcharfun given to
306 read_vector. */
307 readbyte = readbyte_from_string;
308 if (EQ (XCDR (readcharfun), Qget_emacs_mule_file_char))
309 emacs_mule_encoding = 1;
310 goto read_multibyte;
311 }
312
313 if (EQ (readcharfun, Qget_emacs_mule_file_char))
314 {
315 readbyte = readbyte_from_file;
316 emacs_mule_encoding = 1;
317 goto read_multibyte;
318 }
319
320 tem = call0 (readcharfun);
321
322 if (NILP (tem))
323 return -1;
324 return XINT (tem);
325
326 read_multibyte:
327 if (unread_char >= 0)
328 {
329 c = unread_char;
330 unread_char = -1;
331 return c;
332 }
333 c = (*readbyte) (-1, readcharfun);
334 if (c < 0 || load_each_byte)
335 return c;
336 if (multibyte)
337 *multibyte = 1;
338 if (ASCII_BYTE_P (c))
339 return c;
340 if (emacs_mule_encoding)
341 return read_emacs_mule_char (c, readbyte, readcharfun);
342 i = 0;
343 buf[i++] = c;
344 len = BYTES_BY_CHAR_HEAD (c);
345 while (i < len)
346 {
347 c = (*readbyte) (-1, readcharfun);
348 if (c < 0 || ! TRAILING_CODE_P (c))
349 {
350 while (--i > 1)
351 (*readbyte) (buf[i], readcharfun);
352 return BYTE8_TO_CHAR (buf[0]);
353 }
354 buf[i++] = c;
355 }
356 return STRING_CHAR (buf);
357}
358
359/* Unread the character C in the way appropriate for the stream READCHARFUN.
360 If the stream is a user function, call it with the char as argument. */
361
362static void
363unreadchar (Lisp_Object readcharfun, int c)
364{
365 readchar_count--;
366 if (c == -1)
367 /* Don't back up the pointer if we're unreading the end-of-input mark,
368 since readchar didn't advance it when we read it. */
369 ;
370 else if (BUFFERP (readcharfun))
371 {
372 struct buffer *b = XBUFFER (readcharfun);
373 ptrdiff_t charpos = BUF_PT (b);
374 ptrdiff_t bytepos = BUF_PT_BYTE (b);
375
376 if (! NILP (BVAR (b, enable_multibyte_characters)))
377 BUF_DEC_POS (b, bytepos);
378 else
379 bytepos--;
380
381 SET_BUF_PT_BOTH (b, charpos - 1, bytepos);
382 }
383 else if (MARKERP (readcharfun))
384 {
385 struct buffer *b = XMARKER (readcharfun)->buffer;
386 ptrdiff_t bytepos = XMARKER (readcharfun)->bytepos;
387
388 XMARKER (readcharfun)->charpos--;
389 if (! NILP (BVAR (b, enable_multibyte_characters)))
390 BUF_DEC_POS (b, bytepos);
391 else
392 bytepos--;
393
394 XMARKER (readcharfun)->bytepos = bytepos;
395 }
396 else if (STRINGP (readcharfun))
397 {
398 read_from_string_index--;
399 read_from_string_index_byte
400 = string_char_to_byte (readcharfun, read_from_string_index);
401 }
402 else if (CONSP (readcharfun))
403 {
404 unread_char = c;
405 }
406 else if (EQ (readcharfun, Qlambda))
407 {
408 unread_char = c;
409 }
410 else if (EQ (readcharfun, Qget_file_char)
411 || EQ (readcharfun, Qget_emacs_mule_file_char))
412 {
413 if (load_each_byte)
414 {
415 BLOCK_INPUT;
416 ungetc (c, instream);
417 UNBLOCK_INPUT;
418 }
419 else
420 unread_char = c;
421 }
422 else
423 call1 (readcharfun, make_number (c));
424}
425
426static int
427readbyte_for_lambda (int c, Lisp_Object readcharfun)
428{
429 return read_bytecode_char (c >= 0);
430}
431
432
433static int
434readbyte_from_file (int c, Lisp_Object readcharfun)
435{
436 if (c >= 0)
437 {
438 BLOCK_INPUT;
439 ungetc (c, instream);
440 UNBLOCK_INPUT;
441 return 0;
442 }
443
444 BLOCK_INPUT;
445 c = getc (instream);
446
447#ifdef EINTR
448 /* Interrupted reads have been observed while reading over the network. */
449 while (c == EOF && ferror (instream) && errno == EINTR)
450 {
451 UNBLOCK_INPUT;
452 QUIT;
453 BLOCK_INPUT;
454 clearerr (instream);
455 c = getc (instream);
456 }
457#endif
458
459 UNBLOCK_INPUT;
460
461 return (c == EOF ? -1 : c);
462}
463
464static int
465readbyte_from_string (int c, Lisp_Object readcharfun)
466{
467 Lisp_Object string = XCAR (readcharfun);
468
469 if (c >= 0)
470 {
471 read_from_string_index--;
472 read_from_string_index_byte
473 = string_char_to_byte (string, read_from_string_index);
474 }
475
476 if (read_from_string_index >= read_from_string_limit)
477 c = -1;
478 else
479 FETCH_STRING_CHAR_ADVANCE (c, string,
480 read_from_string_index,
481 read_from_string_index_byte);
482 return c;
483}
484
485
486/* Read one non-ASCII character from INSTREAM. The character is
487 encoded in `emacs-mule' and the first byte is already read in
488 C. */
489
490static int
491read_emacs_mule_char (int c, int (*readbyte) (int, Lisp_Object), Lisp_Object readcharfun)
492{
493 /* Emacs-mule coding uses at most 4-byte for one character. */
494 unsigned char buf[4];
495 int len = emacs_mule_bytes[c];
496 struct charset *charset;
497 int i;
498 unsigned code;
499
500 if (len == 1)
501 /* C is not a valid leading-code of `emacs-mule'. */
502 return BYTE8_TO_CHAR (c);
503
504 i = 0;
505 buf[i++] = c;
506 while (i < len)
507 {
508 c = (*readbyte) (-1, readcharfun);
509 if (c < 0xA0)
510 {
511 while (--i > 1)
512 (*readbyte) (buf[i], readcharfun);
513 return BYTE8_TO_CHAR (buf[0]);
514 }
515 buf[i++] = c;
516 }
517
518 if (len == 2)
519 {
520 charset = CHARSET_FROM_ID (emacs_mule_charset[buf[0]]);
521 code = buf[1] & 0x7F;
522 }
523 else if (len == 3)
524 {
525 if (buf[0] == EMACS_MULE_LEADING_CODE_PRIVATE_11
526 || buf[0] == EMACS_MULE_LEADING_CODE_PRIVATE_12)
527 {
528 charset = CHARSET_FROM_ID (emacs_mule_charset[buf[1]]);
529 code = buf[2] & 0x7F;
530 }
531 else
532 {
533 charset = CHARSET_FROM_ID (emacs_mule_charset[buf[0]]);
534 code = ((buf[1] << 8) | buf[2]) & 0x7F7F;
535 }
536 }
537 else
538 {
539 charset = CHARSET_FROM_ID (emacs_mule_charset[buf[1]]);
540 code = ((buf[2] << 8) | buf[3]) & 0x7F7F;
541 }
542 c = DECODE_CHAR (charset, code);
543 if (c < 0)
544 Fsignal (Qinvalid_read_syntax,
545 Fcons (build_string ("invalid multibyte form"), Qnil));
546 return c;
547}
548
549
550static Lisp_Object read_internal_start (Lisp_Object, Lisp_Object,
551 Lisp_Object);
552static Lisp_Object read0 (Lisp_Object);
553static Lisp_Object read1 (Lisp_Object, int *, int);
554
555static Lisp_Object read_list (int, Lisp_Object);
556static Lisp_Object read_vector (Lisp_Object, int);
557
558static Lisp_Object substitute_object_recurse (Lisp_Object, Lisp_Object,
559 Lisp_Object);
560static void substitute_object_in_subtree (Lisp_Object,
561 Lisp_Object);
562static void substitute_in_interval (INTERVAL, Lisp_Object);
563
564\f
565/* Get a character from the tty. */
566
567/* Read input events until we get one that's acceptable for our purposes.
568
569 If NO_SWITCH_FRAME is non-zero, switch-frame events are stashed
570 until we get a character we like, and then stuffed into
571 unread_switch_frame.
572
573 If ASCII_REQUIRED is non-zero, we check function key events to see
574 if the unmodified version of the symbol has a Qascii_character
575 property, and use that character, if present.
576
577 If ERROR_NONASCII is non-zero, we signal an error if the input we
578 get isn't an ASCII character with modifiers. If it's zero but
579 ASCII_REQUIRED is non-zero, we just re-read until we get an ASCII
580 character.
581
582 If INPUT_METHOD is nonzero, we invoke the current input method
583 if the character warrants that.
584
585 If SECONDS is a number, we wait that many seconds for input, and
586 return Qnil if no input arrives within that time. */
587
588static Lisp_Object
589read_filtered_event (int no_switch_frame, int ascii_required,
590 int error_nonascii, int input_method, Lisp_Object seconds)
591{
592 Lisp_Object val, delayed_switch_frame;
593 EMACS_TIME end_time;
594
595#ifdef HAVE_WINDOW_SYSTEM
596 if (display_hourglass_p)
597 cancel_hourglass ();
598#endif
599
600 delayed_switch_frame = Qnil;
601
602 /* Compute timeout. */
603 if (NUMBERP (seconds))
604 {
605 double duration = extract_float (seconds);
606 EMACS_TIME wait_time = EMACS_TIME_FROM_DOUBLE (duration);
607 end_time = add_emacs_time (current_emacs_time (), wait_time);
608 }
609
610/* Read until we get an acceptable event. */
611 retry:
612 do
613 val = read_char (0, 0, 0, (input_method ? Qnil : Qt), 0,
614 NUMBERP (seconds) ? &end_time : NULL);
615 while (INTEGERP (val) && XINT (val) == -2); /* wrong_kboard_jmpbuf */
616
617 if (BUFFERP (val))
618 goto retry;
619
620 /* switch-frame events are put off until after the next ASCII
621 character. This is better than signaling an error just because
622 the last characters were typed to a separate minibuffer frame,
623 for example. Eventually, some code which can deal with
624 switch-frame events will read it and process it. */
625 if (no_switch_frame
626 && EVENT_HAS_PARAMETERS (val)
627 && EQ (EVENT_HEAD_KIND (EVENT_HEAD (val)), Qswitch_frame))
628 {
629 delayed_switch_frame = val;
630 goto retry;
631 }
632
633 if (ascii_required && !(NUMBERP (seconds) && NILP (val)))
634 {
635 /* Convert certain symbols to their ASCII equivalents. */
636 if (SYMBOLP (val))
637 {
638 Lisp_Object tem, tem1;
639 tem = Fget (val, Qevent_symbol_element_mask);
640 if (!NILP (tem))
641 {
642 tem1 = Fget (Fcar (tem), Qascii_character);
643 /* Merge this symbol's modifier bits
644 with the ASCII equivalent of its basic code. */
645 if (!NILP (tem1))
646 XSETFASTINT (val, XINT (tem1) | XINT (Fcar (Fcdr (tem))));
647 }
648 }
649
650 /* If we don't have a character now, deal with it appropriately. */
651 if (!INTEGERP (val))
652 {
653 if (error_nonascii)
654 {
655 Vunread_command_events = Fcons (val, Qnil);
656 error ("Non-character input-event");
657 }
658 else
659 goto retry;
660 }
661 }
662
663 if (! NILP (delayed_switch_frame))
664 unread_switch_frame = delayed_switch_frame;
665
666#if 0
667
668#ifdef HAVE_WINDOW_SYSTEM
669 if (display_hourglass_p)
670 start_hourglass ();
671#endif
672
673#endif
674
675 return val;
676}
677
678DEFUN ("read-char", Fread_char, Sread_char, 0, 3, 0,
679 doc: /* Read a character from the command input (keyboard or macro).
680It is returned as a number.
681If the character has modifiers, they are resolved and reflected to the
682character code if possible (e.g. C-SPC -> 0).
683
684If the user generates an event which is not a character (i.e. a mouse
685click or function key event), `read-char' signals an error. As an
686exception, switch-frame events are put off until non-character events
687can be read.
688If you want to read non-character events, or ignore them, call
689`read-event' or `read-char-exclusive' instead.
690
691If the optional argument PROMPT is non-nil, display that as a prompt.
692If the optional argument INHERIT-INPUT-METHOD is non-nil and some
693input method is turned on in the current buffer, that input method
694is used for reading a character.
695If the optional argument SECONDS is non-nil, it should be a number
696specifying the maximum number of seconds to wait for input. If no
697input arrives in that time, return nil. SECONDS may be a
698floating-point value. */)
699 (Lisp_Object prompt, Lisp_Object inherit_input_method, Lisp_Object seconds)
700{
701 Lisp_Object val;
702
703 if (! NILP (prompt))
704 message_with_string ("%s", prompt, 0);
705 val = read_filtered_event (1, 1, 1, ! NILP (inherit_input_method), seconds);
706
707 return (NILP (val) ? Qnil
708 : make_number (char_resolve_modifier_mask (XINT (val))));
709}
710
711DEFUN ("read-event", Fread_event, Sread_event, 0, 3, 0,
712 doc: /* Read an event object from the input stream.
713If the optional argument PROMPT is non-nil, display that as a prompt.
714If the optional argument INHERIT-INPUT-METHOD is non-nil and some
715input method is turned on in the current buffer, that input method
716is used for reading a character.
717If the optional argument SECONDS is non-nil, it should be a number
718specifying the maximum number of seconds to wait for input. If no
719input arrives in that time, return nil. SECONDS may be a
720floating-point value. */)
721 (Lisp_Object prompt, Lisp_Object inherit_input_method, Lisp_Object seconds)
722{
723 if (! NILP (prompt))
724 message_with_string ("%s", prompt, 0);
725 return read_filtered_event (0, 0, 0, ! NILP (inherit_input_method), seconds);
726}
727
728DEFUN ("read-char-exclusive", Fread_char_exclusive, Sread_char_exclusive, 0, 3, 0,
729 doc: /* Read a character from the command input (keyboard or macro).
730It is returned as a number. Non-character events are ignored.
731If the character has modifiers, they are resolved and reflected to the
732character code if possible (e.g. C-SPC -> 0).
733
734If the optional argument PROMPT is non-nil, display that as a prompt.
735If the optional argument INHERIT-INPUT-METHOD is non-nil and some
736input method is turned on in the current buffer, that input method
737is used for reading a character.
738If the optional argument SECONDS is non-nil, it should be a number
739specifying the maximum number of seconds to wait for input. If no
740input arrives in that time, return nil. SECONDS may be a
741floating-point value. */)
742 (Lisp_Object prompt, Lisp_Object inherit_input_method, Lisp_Object seconds)
743{
744 Lisp_Object val;
745
746 if (! NILP (prompt))
747 message_with_string ("%s", prompt, 0);
748
749 val = read_filtered_event (1, 1, 0, ! NILP (inherit_input_method), seconds);
750
751 return (NILP (val) ? Qnil
752 : make_number (char_resolve_modifier_mask (XINT (val))));
753}
754
755DEFUN ("get-file-char", Fget_file_char, Sget_file_char, 0, 0, 0,
756 doc: /* Don't use this yourself. */)
757 (void)
758{
759 register Lisp_Object val;
760 BLOCK_INPUT;
761 XSETINT (val, getc (instream));
762 UNBLOCK_INPUT;
763 return val;
764}
765
766
767\f
768
769/* Return true if the lisp code read using READCHARFUN defines a non-nil
770 `lexical-binding' file variable. After returning, the stream is
771 positioned following the first line, if it is a comment, otherwise
772 nothing is read. */
773
774static int
775lisp_file_lexically_bound_p (Lisp_Object readcharfun)
776{
777 int ch = READCHAR;
778 if (ch != ';')
779 /* The first line isn't a comment, just give up. */
780 {
781 UNREAD (ch);
782 return 0;
783 }
784 else
785 /* Look for an appropriate file-variable in the first line. */
786 {
787 int rv = 0;
788 enum {
789 NOMINAL, AFTER_FIRST_DASH, AFTER_ASTERIX,
790 } beg_end_state = NOMINAL;
791 int in_file_vars = 0;
792
793#define UPDATE_BEG_END_STATE(ch) \
794 if (beg_end_state == NOMINAL) \
795 beg_end_state = (ch == '-' ? AFTER_FIRST_DASH : NOMINAL); \
796 else if (beg_end_state == AFTER_FIRST_DASH) \
797 beg_end_state = (ch == '*' ? AFTER_ASTERIX : NOMINAL); \
798 else if (beg_end_state == AFTER_ASTERIX) \
799 { \
800 if (ch == '-') \
801 in_file_vars = !in_file_vars; \
802 beg_end_state = NOMINAL; \
803 }
804
805 /* Skip until we get to the file vars, if any. */
806 do
807 {
808 ch = READCHAR;
809 UPDATE_BEG_END_STATE (ch);
810 }
811 while (!in_file_vars && ch != '\n' && ch != EOF);
812
813 while (in_file_vars)
814 {
815 char var[100], val[100];
816 unsigned i;
817
818 ch = READCHAR;
819
820 /* Read a variable name. */
821 while (ch == ' ' || ch == '\t')
822 ch = READCHAR;
823
824 i = 0;
825 while (ch != ':' && ch != '\n' && ch != EOF && in_file_vars)
826 {
827 if (i < sizeof var - 1)
828 var[i++] = ch;
829 UPDATE_BEG_END_STATE (ch);
830 ch = READCHAR;
831 }
832
833 /* Stop scanning if no colon was found before end marker. */
834 if (!in_file_vars || ch == '\n' || ch == EOF)
835 break;
836
837 while (i > 0 && (var[i - 1] == ' ' || var[i - 1] == '\t'))
838 i--;
839 var[i] = '\0';
840
841 if (ch == ':')
842 {
843 /* Read a variable value. */
844 ch = READCHAR;
845
846 while (ch == ' ' || ch == '\t')
847 ch = READCHAR;
848
849 i = 0;
850 while (ch != ';' && ch != '\n' && ch != EOF && in_file_vars)
851 {
852 if (i < sizeof val - 1)
853 val[i++] = ch;
854 UPDATE_BEG_END_STATE (ch);
855 ch = READCHAR;
856 }
857 if (! in_file_vars)
858 /* The value was terminated by an end-marker, which remove. */
859 i -= 3;
860 while (i > 0 && (val[i - 1] == ' ' || val[i - 1] == '\t'))
861 i--;
862 val[i] = '\0';
863
864 if (strcmp (var, "lexical-binding") == 0)
865 /* This is it... */
866 {
867 rv = (strcmp (val, "nil") != 0);
868 break;
869 }
870 }
871 }
872
873 while (ch != '\n' && ch != EOF)
874 ch = READCHAR;
875
876 return rv;
877 }
878}
879\f
880/* Value is a version number of byte compiled code if the file
881 associated with file descriptor FD is a compiled Lisp file that's
882 safe to load. Only files compiled with Emacs are safe to load.
883 Files compiled with XEmacs can lead to a crash in Fbyte_code
884 because of an incompatible change in the byte compiler. */
885
886static int
887safe_to_load_p (int fd)
888{
889 char buf[512];
890 int nbytes, i;
891 int safe_p = 1;
892 int version = 1;
893
894 /* Read the first few bytes from the file, and look for a line
895 specifying the byte compiler version used. */
896 nbytes = emacs_read (fd, buf, sizeof buf - 1);
897 if (nbytes > 0)
898 {
899 buf[nbytes] = '\0';
900
901 /* Skip to the next newline, skipping over the initial `ELC'
902 with NUL bytes following it, but note the version. */
903 for (i = 0; i < nbytes && buf[i] != '\n'; ++i)
904 if (i == 4)
905 version = buf[i];
906
907 if (i >= nbytes
908 || fast_c_string_match_ignore_case (Vbytecomp_version_regexp,
909 buf + i, nbytes - i) < 0)
910 safe_p = 0;
911 }
912 if (safe_p)
913 safe_p = version;
914
915 lseek (fd, 0, SEEK_SET);
916 return safe_p;
917}
918
919
920/* Callback for record_unwind_protect. Restore the old load list OLD,
921 after loading a file successfully. */
922
923static Lisp_Object
924record_load_unwind (Lisp_Object old)
925{
926 return Vloads_in_progress = old;
927}
928
929/* This handler function is used via internal_condition_case_1. */
930
931static Lisp_Object
932load_error_handler (Lisp_Object data)
933{
934 return Qnil;
935}
936
937static Lisp_Object
938load_warn_old_style_backquotes (Lisp_Object file)
939{
940 if (!NILP (Vold_style_backquotes))
941 {
942 Lisp_Object args[2];
943 args[0] = build_string ("Loading `%s': old-style backquotes detected!");
944 args[1] = file;
945 Fmessage (2, args);
946 }
947 return Qnil;
948}
949
950DEFUN ("get-load-suffixes", Fget_load_suffixes, Sget_load_suffixes, 0, 0, 0,
951 doc: /* Return the suffixes that `load' should try if a suffix is \
952required.
953This uses the variables `load-suffixes' and `load-file-rep-suffixes'. */)
954 (void)
955{
956 Lisp_Object lst = Qnil, suffixes = Vload_suffixes, suffix, ext;
957 while (CONSP (suffixes))
958 {
959 Lisp_Object exts = Vload_file_rep_suffixes;
960 suffix = XCAR (suffixes);
961 suffixes = XCDR (suffixes);
962 while (CONSP (exts))
963 {
964 ext = XCAR (exts);
965 exts = XCDR (exts);
966 lst = Fcons (concat2 (suffix, ext), lst);
967 }
968 }
969 return Fnreverse (lst);
970}
971
972DEFUN ("load", Fload, Sload, 1, 5, 0,
973 doc: /* Execute a file of Lisp code named FILE.
974First try FILE with `.elc' appended, then try with `.el',
975then try FILE unmodified (the exact suffixes in the exact order are
976determined by `load-suffixes'). Environment variable references in
977FILE are replaced with their values by calling `substitute-in-file-name'.
978This function searches the directories in `load-path'.
979
980If optional second arg NOERROR is non-nil,
981report no error if FILE doesn't exist.
982Print messages at start and end of loading unless
983optional third arg NOMESSAGE is non-nil (but `force-load-messages'
984overrides that).
985If optional fourth arg NOSUFFIX is non-nil, don't try adding
986suffixes `.elc' or `.el' to the specified name FILE.
987If optional fifth arg MUST-SUFFIX is non-nil, insist on
988the suffix `.elc' or `.el'; don't accept just FILE unless
989it ends in one of those suffixes or includes a directory name.
990
991If this function fails to find a file, it may look for different
992representations of that file before trying another file.
993It does so by adding the non-empty suffixes in `load-file-rep-suffixes'
994to the file name. Emacs uses this feature mainly to find compressed
995versions of files when Auto Compression mode is enabled.
996
997The exact suffixes that this function tries out, in the exact order,
998are given by the value of the variable `load-file-rep-suffixes' if
999NOSUFFIX is non-nil and by the return value of the function
1000`get-load-suffixes' if MUST-SUFFIX is non-nil. If both NOSUFFIX and
1001MUST-SUFFIX are nil, this function first tries out the latter suffixes
1002and then the former.
1003
1004Loading a file records its definitions, and its `provide' and
1005`require' calls, in an element of `load-history' whose
1006car is the file name loaded. See `load-history'.
1007
1008While the file is in the process of being loaded, the variable
1009`load-in-progress' is non-nil and the variable `load-file-name'
1010is bound to the file's name.
1011
1012Return t if the file exists and loads successfully. */)
1013 (Lisp_Object file, Lisp_Object noerror, Lisp_Object nomessage, Lisp_Object nosuffix, Lisp_Object must_suffix)
1014{
1015 register FILE *stream;
1016 register int fd = -1;
1017 ptrdiff_t count = SPECPDL_INDEX ();
1018 struct gcpro gcpro1, gcpro2, gcpro3;
1019 Lisp_Object found, efound, hist_file_name;
1020 /* 1 means we printed the ".el is newer" message. */
1021 int newer = 0;
1022 /* 1 means we are loading a compiled file. */
1023 int compiled = 0;
1024 Lisp_Object handler;
1025 int safe_p = 1;
1026 const char *fmode = "r";
1027 Lisp_Object tmp[2];
1028 int version;
1029
1030#ifdef DOS_NT
1031 fmode = "rt";
1032#endif /* DOS_NT */
1033
1034 CHECK_STRING (file);
1035
1036 /* If file name is magic, call the handler. */
1037 /* This shouldn't be necessary any more now that `openp' handles it right.
1038 handler = Ffind_file_name_handler (file, Qload);
1039 if (!NILP (handler))
1040 return call5 (handler, Qload, file, noerror, nomessage, nosuffix); */
1041
1042 /* Do this after the handler to avoid
1043 the need to gcpro noerror, nomessage and nosuffix.
1044 (Below here, we care only whether they are nil or not.)
1045 The presence of this call is the result of a historical accident:
1046 it used to be in every file-operation and when it got removed
1047 everywhere, it accidentally stayed here. Since then, enough people
1048 supposedly have things like (load "$PROJECT/foo.el") in their .emacs
1049 that it seemed risky to remove. */
1050 if (! NILP (noerror))
1051 {
1052 file = internal_condition_case_1 (Fsubstitute_in_file_name, file,
1053 Qt, load_error_handler);
1054 if (NILP (file))
1055 return Qnil;
1056 }
1057 else
1058 file = Fsubstitute_in_file_name (file);
1059
1060
1061 /* Avoid weird lossage with null string as arg,
1062 since it would try to load a directory as a Lisp file. */
1063 if (SBYTES (file) > 0)
1064 {
1065 ptrdiff_t size = SBYTES (file);
1066
1067 found = Qnil;
1068 GCPRO2 (file, found);
1069
1070 if (! NILP (must_suffix))
1071 {
1072 /* Don't insist on adding a suffix if FILE already ends with one. */
1073 if (size > 3
1074 && !strcmp (SSDATA (file) + size - 3, ".el"))
1075 must_suffix = Qnil;
1076 else if (size > 4
1077 && !strcmp (SSDATA (file) + size - 4, ".elc"))
1078 must_suffix = Qnil;
1079 /* Don't insist on adding a suffix
1080 if the argument includes a directory name. */
1081 else if (! NILP (Ffile_name_directory (file)))
1082 must_suffix = Qnil;
1083 }
1084
1085 fd = openp (Vload_path, file,
1086 (!NILP (nosuffix) ? Qnil
1087 : !NILP (must_suffix) ? Fget_load_suffixes ()
1088 : Fappend (2, (tmp[0] = Fget_load_suffixes (),
1089 tmp[1] = Vload_file_rep_suffixes,
1090 tmp))),
1091 &found, Qnil);
1092 UNGCPRO;
1093 }
1094
1095 if (fd == -1)
1096 {
1097 if (NILP (noerror))
1098 xsignal2 (Qfile_error, build_string ("Cannot open load file"), file);
1099 return Qnil;
1100 }
1101
1102 /* Tell startup.el whether or not we found the user's init file. */
1103 if (EQ (Qt, Vuser_init_file))
1104 Vuser_init_file = found;
1105
1106 /* If FD is -2, that means openp found a magic file. */
1107 if (fd == -2)
1108 {
1109 if (NILP (Fequal (found, file)))
1110 /* If FOUND is a different file name from FILE,
1111 find its handler even if we have already inhibited
1112 the `load' operation on FILE. */
1113 handler = Ffind_file_name_handler (found, Qt);
1114 else
1115 handler = Ffind_file_name_handler (found, Qload);
1116 if (! NILP (handler))
1117 return call5 (handler, Qload, found, noerror, nomessage, Qt);
1118#ifdef DOS_NT
1119 /* Tramp has to deal with semi-broken packages that prepend
1120 drive letters to remote files. For that reason, Tramp
1121 catches file operations that test for file existence, which
1122 makes openp think X:/foo.elc files are remote. However,
1123 Tramp does not catch `load' operations for such files, so we
1124 end up with a nil as the `load' handler above. If we would
1125 continue with fd = -2, we will behave wrongly, and in
1126 particular try reading a .elc file in the "rt" mode instead
1127 of "rb". See bug #9311 for the results. To work around
1128 this, we try to open the file locally, and go with that if it
1129 succeeds. */
1130 fd = emacs_open (SSDATA (ENCODE_FILE (found)), O_RDONLY, 0);
1131 if (fd == -1)
1132 fd = -2;
1133#endif
1134 }
1135
1136 /* Check if we're stuck in a recursive load cycle.
1137
1138 2000-09-21: It's not possible to just check for the file loaded
1139 being a member of Vloads_in_progress. This fails because of the
1140 way the byte compiler currently works; `provide's are not
1141 evaluated, see font-lock.el/jit-lock.el as an example. This
1142 leads to a certain amount of ``normal'' recursion.
1143
1144 Also, just loading a file recursively is not always an error in
1145 the general case; the second load may do something different. */
1146 {
1147 int load_count = 0;
1148 Lisp_Object tem;
1149 for (tem = Vloads_in_progress; CONSP (tem); tem = XCDR (tem))
1150 if (!NILP (Fequal (found, XCAR (tem))) && (++load_count > 3))
1151 {
1152 if (fd >= 0)
1153 emacs_close (fd);
1154 signal_error ("Recursive load", Fcons (found, Vloads_in_progress));
1155 }
1156 record_unwind_protect (record_load_unwind, Vloads_in_progress);
1157 Vloads_in_progress = Fcons (found, Vloads_in_progress);
1158 }
1159
1160 /* All loads are by default dynamic, unless the file itself specifies
1161 otherwise using a file-variable in the first line. This is bound here
1162 so that it takes effect whether or not we use
1163 Vload_source_file_function. */
1164 specbind (Qlexical_binding, Qnil);
1165
1166 /* Get the name for load-history. */
1167 hist_file_name = (! NILP (Vpurify_flag)
1168 ? Fconcat (2, (tmp[0] = Ffile_name_directory (file),
1169 tmp[1] = Ffile_name_nondirectory (found),
1170 tmp))
1171 : found) ;
1172
1173 version = -1;
1174
1175 /* Check for the presence of old-style quotes and warn about them. */
1176 specbind (Qold_style_backquotes, Qnil);
1177 record_unwind_protect (load_warn_old_style_backquotes, file);
1178
1179 if (!memcmp (SDATA (found) + SBYTES (found) - 4, ".elc", 4)
1180 || (fd >= 0 && (version = safe_to_load_p (fd)) > 0))
1181 /* Load .elc files directly, but not when they are
1182 remote and have no handler! */
1183 {
1184 if (fd != -2)
1185 {
1186 struct stat s1, s2;
1187 int result;
1188
1189 GCPRO3 (file, found, hist_file_name);
1190
1191 if (version < 0
1192 && ! (version = safe_to_load_p (fd)))
1193 {
1194 safe_p = 0;
1195 if (!load_dangerous_libraries)
1196 {
1197 if (fd >= 0)
1198 emacs_close (fd);
1199 error ("File `%s' was not compiled in Emacs",
1200 SDATA (found));
1201 }
1202 else if (!NILP (nomessage) && !force_load_messages)
1203 message_with_string ("File `%s' not compiled in Emacs", found, 1);
1204 }
1205
1206 compiled = 1;
1207
1208 efound = ENCODE_FILE (found);
1209
1210#ifdef DOS_NT
1211 fmode = "rb";
1212#endif /* DOS_NT */
1213 result = stat (SSDATA (efound), &s1);
1214 if (result == 0)
1215 {
1216 SSET (efound, SBYTES (efound) - 1, 0);
1217 result = stat (SSDATA (efound), &s2);
1218 SSET (efound, SBYTES (efound) - 1, 'c');
1219 }
1220
1221 if (result == 0
1222 && EMACS_TIME_LT (get_stat_mtime (&s1), get_stat_mtime (&s2)))
1223 {
1224 /* Make the progress messages mention that source is newer. */
1225 newer = 1;
1226
1227 /* If we won't print another message, mention this anyway. */
1228 if (!NILP (nomessage) && !force_load_messages)
1229 {
1230 Lisp_Object msg_file;
1231 msg_file = Fsubstring (found, make_number (0), make_number (-1));
1232 message_with_string ("Source file `%s' newer than byte-compiled file",
1233 msg_file, 1);
1234 }
1235 }
1236 UNGCPRO;
1237 }
1238 }
1239 else
1240 {
1241 /* We are loading a source file (*.el). */
1242 if (!NILP (Vload_source_file_function))
1243 {
1244 Lisp_Object val;
1245
1246 if (fd >= 0)
1247 emacs_close (fd);
1248 val = call4 (Vload_source_file_function, found, hist_file_name,
1249 NILP (noerror) ? Qnil : Qt,
1250 (NILP (nomessage) || force_load_messages) ? Qnil : Qt);
1251 return unbind_to (count, val);
1252 }
1253 }
1254
1255 GCPRO3 (file, found, hist_file_name);
1256
1257#ifdef WINDOWSNT
1258 efound = ENCODE_FILE (found);
1259 /* If we somehow got here with fd == -2, meaning the file is deemed
1260 to be remote, don't even try to reopen the file locally; just
1261 force a failure instead. */
1262 if (fd >= 0)
1263 {
1264 emacs_close (fd);
1265 stream = fopen (SSDATA (efound), fmode);
1266 }
1267 else
1268 stream = NULL;
1269#else /* not WINDOWSNT */
1270 stream = fdopen (fd, fmode);
1271#endif /* not WINDOWSNT */
1272 if (stream == 0)
1273 {
1274 emacs_close (fd);
1275 error ("Failure to create stdio stream for %s", SDATA (file));
1276 }
1277
1278 if (! NILP (Vpurify_flag))
1279 Vpreloaded_file_list = Fcons (Fpurecopy (file), Vpreloaded_file_list);
1280
1281 if (NILP (nomessage) || force_load_messages)
1282 {
1283 if (!safe_p)
1284 message_with_string ("Loading %s (compiled; note unsafe, not compiled in Emacs)...",
1285 file, 1);
1286 else if (!compiled)
1287 message_with_string ("Loading %s (source)...", file, 1);
1288 else if (newer)
1289 message_with_string ("Loading %s (compiled; note, source file is newer)...",
1290 file, 1);
1291 else /* The typical case; compiled file newer than source file. */
1292 message_with_string ("Loading %s...", file, 1);
1293 }
1294
1295 record_unwind_protect (load_unwind, make_save_value (stream, 0));
1296 record_unwind_protect (load_descriptor_unwind, load_descriptor_list);
1297 specbind (Qload_file_name, found);
1298 specbind (Qinhibit_file_name_operation, Qnil);
1299 load_descriptor_list
1300 = Fcons (make_number (fileno (stream)), load_descriptor_list);
1301 specbind (Qload_in_progress, Qt);
1302
1303 instream = stream;
1304 if (lisp_file_lexically_bound_p (Qget_file_char))
1305 Fset (Qlexical_binding, Qt);
1306
1307 if (! version || version >= 22)
1308 readevalloop (Qget_file_char, stream, hist_file_name,
1309 0, Qnil, Qnil, Qnil, Qnil);
1310 else
1311 {
1312 /* We can't handle a file which was compiled with
1313 byte-compile-dynamic by older version of Emacs. */
1314 specbind (Qload_force_doc_strings, Qt);
1315 readevalloop (Qget_emacs_mule_file_char, stream, hist_file_name,
1316 0, Qnil, Qnil, Qnil, Qnil);
1317 }
1318 unbind_to (count, Qnil);
1319
1320 /* Run any eval-after-load forms for this file. */
1321 if (!NILP (Ffboundp (Qdo_after_load_evaluation)))
1322 call1 (Qdo_after_load_evaluation, hist_file_name) ;
1323
1324 UNGCPRO;
1325
1326 xfree (saved_doc_string);
1327 saved_doc_string = 0;
1328 saved_doc_string_size = 0;
1329
1330 xfree (prev_saved_doc_string);
1331 prev_saved_doc_string = 0;
1332 prev_saved_doc_string_size = 0;
1333
1334 if (!noninteractive && (NILP (nomessage) || force_load_messages))
1335 {
1336 if (!safe_p)
1337 message_with_string ("Loading %s (compiled; note unsafe, not compiled in Emacs)...done",
1338 file, 1);
1339 else if (!compiled)
1340 message_with_string ("Loading %s (source)...done", file, 1);
1341 else if (newer)
1342 message_with_string ("Loading %s (compiled; note, source file is newer)...done",
1343 file, 1);
1344 else /* The typical case; compiled file newer than source file. */
1345 message_with_string ("Loading %s...done", file, 1);
1346 }
1347
1348 return Qt;
1349}
1350
1351static Lisp_Object
1352load_unwind (Lisp_Object arg) /* Used as unwind-protect function in load. */
1353{
1354 FILE *stream = (FILE *) XSAVE_VALUE (arg)->pointer;
1355 if (stream != NULL)
1356 {
1357 BLOCK_INPUT;
1358 fclose (stream);
1359 UNBLOCK_INPUT;
1360 }
1361 return Qnil;
1362}
1363
1364static Lisp_Object
1365load_descriptor_unwind (Lisp_Object oldlist)
1366{
1367 load_descriptor_list = oldlist;
1368 return Qnil;
1369}
1370
1371/* Close all descriptors in use for Floads.
1372 This is used when starting a subprocess. */
1373
1374void
1375close_load_descs (void)
1376{
1377#ifndef WINDOWSNT
1378 Lisp_Object tail;
1379 for (tail = load_descriptor_list; CONSP (tail); tail = XCDR (tail))
1380 emacs_close (XFASTINT (XCAR (tail)));
1381#endif
1382}
1383\f
1384static int
1385complete_filename_p (Lisp_Object pathname)
1386{
1387 register const unsigned char *s = SDATA (pathname);
1388 return (IS_DIRECTORY_SEP (s[0])
1389 || (SCHARS (pathname) > 2
1390 && IS_DEVICE_SEP (s[1]) && IS_DIRECTORY_SEP (s[2])));
1391}
1392
1393DEFUN ("locate-file-internal", Flocate_file_internal, Slocate_file_internal, 2, 4, 0,
1394 doc: /* Search for FILENAME through PATH.
1395Returns the file's name in absolute form, or nil if not found.
1396If SUFFIXES is non-nil, it should be a list of suffixes to append to
1397file name when searching.
1398If non-nil, PREDICATE is used instead of `file-readable-p'.
1399PREDICATE can also be an integer to pass to the access(2) function,
1400in which case file-name-handlers are ignored.
1401This function will normally skip directories, so if you want it to find
1402directories, make sure the PREDICATE function returns `dir-ok' for them. */)
1403 (Lisp_Object filename, Lisp_Object path, Lisp_Object suffixes, Lisp_Object predicate)
1404{
1405 Lisp_Object file;
1406 int fd = openp (path, filename, suffixes, &file, predicate);
1407 if (NILP (predicate) && fd > 0)
1408 close (fd);
1409 return file;
1410}
1411
1412static Lisp_Object Qdir_ok;
1413
1414/* Search for a file whose name is STR, looking in directories
1415 in the Lisp list PATH, and trying suffixes from SUFFIX.
1416 On success, returns a file descriptor. On failure, returns -1.
1417
1418 SUFFIXES is a list of strings containing possible suffixes.
1419 The empty suffix is automatically added if the list is empty.
1420
1421 PREDICATE non-nil means don't open the files,
1422 just look for one that satisfies the predicate. In this case,
1423 returns 1 on success. The predicate can be a lisp function or
1424 an integer to pass to `access' (in which case file-name-handlers
1425 are ignored).
1426
1427 If STOREPTR is nonzero, it points to a slot where the name of
1428 the file actually found should be stored as a Lisp string.
1429 nil is stored there on failure.
1430
1431 If the file we find is remote, return -2
1432 but store the found remote file name in *STOREPTR. */
1433
1434int
1435openp (Lisp_Object path, Lisp_Object str, Lisp_Object suffixes, Lisp_Object *storeptr, Lisp_Object predicate)
1436{
1437 register int fd;
1438 ptrdiff_t fn_size = 100;
1439 char buf[100];
1440 register char *fn = buf;
1441 int absolute = 0;
1442 ptrdiff_t want_length;
1443 Lisp_Object filename;
1444 struct stat st;
1445 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4, gcpro5, gcpro6;
1446 Lisp_Object string, tail, encoded_fn;
1447 ptrdiff_t max_suffix_len = 0;
1448
1449 CHECK_STRING (str);
1450
1451 for (tail = suffixes; CONSP (tail); tail = XCDR (tail))
1452 {
1453 CHECK_STRING_CAR (tail);
1454 max_suffix_len = max (max_suffix_len,
1455 SBYTES (XCAR (tail)));
1456 }
1457
1458 string = filename = encoded_fn = Qnil;
1459 GCPRO6 (str, string, filename, path, suffixes, encoded_fn);
1460
1461 if (storeptr)
1462 *storeptr = Qnil;
1463
1464 if (complete_filename_p (str))
1465 absolute = 1;
1466
1467 for (; CONSP (path); path = XCDR (path))
1468 {
1469 filename = Fexpand_file_name (str, XCAR (path));
1470 if (!complete_filename_p (filename))
1471 /* If there are non-absolute elts in PATH (eg "."). */
1472 /* Of course, this could conceivably lose if luser sets
1473 default-directory to be something non-absolute... */
1474 {
1475 filename = Fexpand_file_name (filename, BVAR (current_buffer, directory));
1476 if (!complete_filename_p (filename))
1477 /* Give up on this path element! */
1478 continue;
1479 }
1480
1481 /* Calculate maximum length of any filename made from
1482 this path element/specified file name and any possible suffix. */
1483 want_length = max_suffix_len + SBYTES (filename);
1484 if (fn_size <= want_length)
1485 fn = alloca (fn_size = 100 + want_length);
1486
1487 /* Loop over suffixes. */
1488 for (tail = NILP (suffixes) ? Fcons (empty_unibyte_string, Qnil) : suffixes;
1489 CONSP (tail); tail = XCDR (tail))
1490 {
1491 ptrdiff_t fnlen, lsuffix = SBYTES (XCAR (tail));
1492 Lisp_Object handler;
1493 int exists;
1494
1495 /* Concatenate path element/specified name with the suffix.
1496 If the directory starts with /:, remove that. */
1497 int prefixlen = ((SCHARS (filename) > 2
1498 && SREF (filename, 0) == '/'
1499 && SREF (filename, 1) == ':')
1500 ? 2 : 0);
1501 fnlen = SBYTES (filename) - prefixlen;
1502 memcpy (fn, SDATA (filename) + prefixlen, fnlen);
1503 memcpy (fn + fnlen, SDATA (XCAR (tail)), lsuffix + 1);
1504 fnlen += lsuffix;
1505 /* Check that the file exists and is not a directory. */
1506 /* We used to only check for handlers on non-absolute file names:
1507 if (absolute)
1508 handler = Qnil;
1509 else
1510 handler = Ffind_file_name_handler (filename, Qfile_exists_p);
1511 It's not clear why that was the case and it breaks things like
1512 (load "/bar.el") where the file is actually "/bar.el.gz". */
1513 string = make_string (fn, fnlen);
1514 handler = Ffind_file_name_handler (string, Qfile_exists_p);
1515 if ((!NILP (handler) || !NILP (predicate)) && !NATNUMP (predicate))
1516 {
1517 if (NILP (predicate))
1518 exists = !NILP (Ffile_readable_p (string));
1519 else
1520 {
1521 Lisp_Object tmp = call1 (predicate, string);
1522 exists = !NILP (tmp)
1523 && (EQ (tmp, Qdir_ok)
1524 || NILP (Ffile_directory_p (string)));
1525 }
1526
1527 if (exists)
1528 {
1529 /* We succeeded; return this descriptor and filename. */
1530 if (storeptr)
1531 *storeptr = string;
1532 UNGCPRO;
1533 return -2;
1534 }
1535 }
1536 else
1537 {
1538 const char *pfn;
1539
1540 encoded_fn = ENCODE_FILE (string);
1541 pfn = SSDATA (encoded_fn);
1542 exists = (stat (pfn, &st) == 0 && ! S_ISDIR (st.st_mode));
1543 if (exists)
1544 {
1545 /* Check that we can access or open it. */
1546 if (NATNUMP (predicate))
1547 fd = (((XFASTINT (predicate) & ~INT_MAX) == 0
1548 && access (pfn, XFASTINT (predicate)) == 0)
1549 ? 1 : -1);
1550 else
1551 fd = emacs_open (pfn, O_RDONLY, 0);
1552
1553 if (fd >= 0)
1554 {
1555 /* We succeeded; return this descriptor and filename. */
1556 if (storeptr)
1557 *storeptr = string;
1558 UNGCPRO;
1559 return fd;
1560 }
1561 }
1562 }
1563 }
1564 if (absolute)
1565 break;
1566 }
1567
1568 UNGCPRO;
1569 return -1;
1570}
1571
1572\f
1573/* Merge the list we've accumulated of globals from the current input source
1574 into the load_history variable. The details depend on whether
1575 the source has an associated file name or not.
1576
1577 FILENAME is the file name that we are loading from.
1578 ENTIRE is 1 if loading that entire file, 0 if evaluating part of it. */
1579
1580static void
1581build_load_history (Lisp_Object filename, int entire)
1582{
1583 register Lisp_Object tail, prev, newelt;
1584 register Lisp_Object tem, tem2;
1585 register int foundit = 0;
1586
1587 tail = Vload_history;
1588 prev = Qnil;
1589
1590 while (CONSP (tail))
1591 {
1592 tem = XCAR (tail);
1593
1594 /* Find the feature's previous assoc list... */
1595 if (!NILP (Fequal (filename, Fcar (tem))))
1596 {
1597 foundit = 1;
1598
1599 /* If we're loading the entire file, remove old data. */
1600 if (entire)
1601 {
1602 if (NILP (prev))
1603 Vload_history = XCDR (tail);
1604 else
1605 Fsetcdr (prev, XCDR (tail));
1606 }
1607
1608 /* Otherwise, cons on new symbols that are not already members. */
1609 else
1610 {
1611 tem2 = Vcurrent_load_list;
1612
1613 while (CONSP (tem2))
1614 {
1615 newelt = XCAR (tem2);
1616
1617 if (NILP (Fmember (newelt, tem)))
1618 Fsetcar (tail, Fcons (XCAR (tem),
1619 Fcons (newelt, XCDR (tem))));
1620
1621 tem2 = XCDR (tem2);
1622 QUIT;
1623 }
1624 }
1625 }
1626 else
1627 prev = tail;
1628 tail = XCDR (tail);
1629 QUIT;
1630 }
1631
1632 /* If we're loading an entire file, cons the new assoc onto the
1633 front of load-history, the most-recently-loaded position. Also
1634 do this if we didn't find an existing member for the file. */
1635 if (entire || !foundit)
1636 Vload_history = Fcons (Fnreverse (Vcurrent_load_list),
1637 Vload_history);
1638}
1639
1640static Lisp_Object
1641readevalloop_1 (Lisp_Object old)
1642{
1643 load_convert_to_unibyte = ! NILP (old);
1644 return Qnil;
1645}
1646
1647/* Signal an `end-of-file' error, if possible with file name
1648 information. */
1649
1650static _Noreturn void
1651end_of_file_error (void)
1652{
1653 if (STRINGP (Vload_file_name))
1654 xsignal1 (Qend_of_file, Vload_file_name);
1655
1656 xsignal0 (Qend_of_file);
1657}
1658
1659/* UNIBYTE specifies how to set load_convert_to_unibyte
1660 for this invocation.
1661 READFUN, if non-nil, is used instead of `read'.
1662
1663 START, END specify region to read in current buffer (from eval-region).
1664 If the input is not from a buffer, they must be nil. */
1665
1666static void
1667readevalloop (Lisp_Object readcharfun,
1668 FILE *stream,
1669 Lisp_Object sourcename,
1670 int printflag,
1671 Lisp_Object unibyte, Lisp_Object readfun,
1672 Lisp_Object start, Lisp_Object end)
1673{
1674 register int c;
1675 register Lisp_Object val;
1676 ptrdiff_t count = SPECPDL_INDEX ();
1677 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
1678 struct buffer *b = 0;
1679 int continue_reading_p;
1680 Lisp_Object lex_bound;
1681 /* Nonzero if reading an entire buffer. */
1682 int whole_buffer = 0;
1683 /* 1 on the first time around. */
1684 int first_sexp = 1;
1685
1686 if (MARKERP (readcharfun))
1687 {
1688 if (NILP (start))
1689 start = readcharfun;
1690 }
1691
1692 if (BUFFERP (readcharfun))
1693 b = XBUFFER (readcharfun);
1694 else if (MARKERP (readcharfun))
1695 b = XMARKER (readcharfun)->buffer;
1696
1697 /* We assume START is nil when input is not from a buffer. */
1698 if (! NILP (start) && !b)
1699 abort ();
1700
1701 specbind (Qstandard_input, readcharfun); /* GCPROs readcharfun. */
1702 specbind (Qcurrent_load_list, Qnil);
1703 record_unwind_protect (readevalloop_1, load_convert_to_unibyte ? Qt : Qnil);
1704 load_convert_to_unibyte = !NILP (unibyte);
1705
1706 /* If lexical binding is active (either because it was specified in
1707 the file's header, or via a buffer-local variable), create an empty
1708 lexical environment, otherwise, turn off lexical binding. */
1709 lex_bound = find_symbol_value (Qlexical_binding);
1710 specbind (Qinternal_interpreter_environment,
1711 NILP (lex_bound) || EQ (lex_bound, Qunbound)
1712 ? Qnil : Fcons (Qt, Qnil));
1713
1714 GCPRO4 (sourcename, readfun, start, end);
1715
1716 /* Try to ensure sourcename is a truename, except whilst preloading. */
1717 if (NILP (Vpurify_flag)
1718 && !NILP (sourcename) && !NILP (Ffile_name_absolute_p (sourcename))
1719 && !NILP (Ffboundp (Qfile_truename)))
1720 sourcename = call1 (Qfile_truename, sourcename) ;
1721
1722 LOADHIST_ATTACH (sourcename);
1723
1724 continue_reading_p = 1;
1725 while (continue_reading_p)
1726 {
1727 ptrdiff_t count1 = SPECPDL_INDEX ();
1728
1729 if (b != 0 && NILP (BVAR (b, name)))
1730 error ("Reading from killed buffer");
1731
1732 if (!NILP (start))
1733 {
1734 /* Switch to the buffer we are reading from. */
1735 record_unwind_protect (save_excursion_restore, save_excursion_save ());
1736 set_buffer_internal (b);
1737
1738 /* Save point in it. */
1739 record_unwind_protect (save_excursion_restore, save_excursion_save ());
1740 /* Save ZV in it. */
1741 record_unwind_protect (save_restriction_restore, save_restriction_save ());
1742 /* Those get unbound after we read one expression. */
1743
1744 /* Set point and ZV around stuff to be read. */
1745 Fgoto_char (start);
1746 if (!NILP (end))
1747 Fnarrow_to_region (make_number (BEGV), end);
1748
1749 /* Just for cleanliness, convert END to a marker
1750 if it is an integer. */
1751 if (INTEGERP (end))
1752 end = Fpoint_max_marker ();
1753 }
1754
1755 /* On the first cycle, we can easily test here
1756 whether we are reading the whole buffer. */
1757 if (b && first_sexp)
1758 whole_buffer = (PT == BEG && ZV == Z);
1759
1760 instream = stream;
1761 read_next:
1762 c = READCHAR;
1763 if (c == ';')
1764 {
1765 while ((c = READCHAR) != '\n' && c != -1);
1766 goto read_next;
1767 }
1768 if (c < 0)
1769 {
1770 unbind_to (count1, Qnil);
1771 break;
1772 }
1773
1774 /* Ignore whitespace here, so we can detect eof. */
1775 if (c == ' ' || c == '\t' || c == '\n' || c == '\f' || c == '\r'
1776 || c == 0xa0) /* NBSP */
1777 goto read_next;
1778
1779 if (!NILP (Vpurify_flag) && c == '(')
1780 {
1781 val = read_list (0, readcharfun);
1782 }
1783 else
1784 {
1785 UNREAD (c);
1786 read_objects = Qnil;
1787 if (!NILP (readfun))
1788 {
1789 val = call1 (readfun, readcharfun);
1790
1791 /* If READCHARFUN has set point to ZV, we should
1792 stop reading, even if the form read sets point
1793 to a different value when evaluated. */
1794 if (BUFFERP (readcharfun))
1795 {
1796 struct buffer *buf = XBUFFER (readcharfun);
1797 if (BUF_PT (buf) == BUF_ZV (buf))
1798 continue_reading_p = 0;
1799 }
1800 }
1801 else if (! NILP (Vload_read_function))
1802 val = call1 (Vload_read_function, readcharfun);
1803 else
1804 val = read_internal_start (readcharfun, Qnil, Qnil);
1805 }
1806
1807 if (!NILP (start) && continue_reading_p)
1808 start = Fpoint_marker ();
1809
1810 /* Restore saved point and BEGV. */
1811 unbind_to (count1, Qnil);
1812
1813 /* Now eval what we just read. */
1814 val = eval_sub (val);
1815
1816 if (printflag)
1817 {
1818 Vvalues = Fcons (val, Vvalues);
1819 if (EQ (Vstandard_output, Qt))
1820 Fprin1 (val, Qnil);
1821 else
1822 Fprint (val, Qnil);
1823 }
1824
1825 first_sexp = 0;
1826 }
1827
1828 build_load_history (sourcename,
1829 stream || whole_buffer);
1830
1831 UNGCPRO;
1832
1833 unbind_to (count, Qnil);
1834}
1835
1836DEFUN ("eval-buffer", Feval_buffer, Seval_buffer, 0, 5, "",
1837 doc: /* Execute the current buffer as Lisp code.
1838When called from a Lisp program (i.e., not interactively), this
1839function accepts up to five optional arguments:
1840BUFFER is the buffer to evaluate (nil means use current buffer).
1841PRINTFLAG controls printing of output:
1842 A value of nil means discard it; anything else is stream for print.
1843FILENAME specifies the file name to use for `load-history'.
1844UNIBYTE, if non-nil, specifies `load-convert-to-unibyte' for this
1845 invocation.
1846DO-ALLOW-PRINT, if non-nil, specifies that `print' and related
1847 functions should work normally even if PRINTFLAG is nil.
1848
1849This function preserves the position of point. */)
1850 (Lisp_Object buffer, Lisp_Object printflag, Lisp_Object filename, Lisp_Object unibyte, Lisp_Object do_allow_print)
1851{
1852 ptrdiff_t count = SPECPDL_INDEX ();
1853 Lisp_Object tem, buf;
1854
1855 if (NILP (buffer))
1856 buf = Fcurrent_buffer ();
1857 else
1858 buf = Fget_buffer (buffer);
1859 if (NILP (buf))
1860 error ("No such buffer");
1861
1862 if (NILP (printflag) && NILP (do_allow_print))
1863 tem = Qsymbolp;
1864 else
1865 tem = printflag;
1866
1867 if (NILP (filename))
1868 filename = BVAR (XBUFFER (buf), filename);
1869
1870 specbind (Qeval_buffer_list, Fcons (buf, Veval_buffer_list));
1871 specbind (Qstandard_output, tem);
1872 record_unwind_protect (save_excursion_restore, save_excursion_save ());
1873 BUF_TEMP_SET_PT (XBUFFER (buf), BUF_BEGV (XBUFFER (buf)));
1874 specbind (Qlexical_binding, lisp_file_lexically_bound_p (buf) ? Qt : Qnil);
1875 readevalloop (buf, 0, filename,
1876 !NILP (printflag), unibyte, Qnil, Qnil, Qnil);
1877 unbind_to (count, Qnil);
1878
1879 return Qnil;
1880}
1881
1882DEFUN ("eval-region", Feval_region, Seval_region, 2, 4, "r",
1883 doc: /* Execute the region as Lisp code.
1884When called from programs, expects two arguments,
1885giving starting and ending indices in the current buffer
1886of the text to be executed.
1887Programs can pass third argument PRINTFLAG which controls output:
1888A value of nil means discard it; anything else is stream for printing it.
1889Also the fourth argument READ-FUNCTION, if non-nil, is used
1890instead of `read' to read each expression. It gets one argument
1891which is the input stream for reading characters.
1892
1893This function does not move point. */)
1894 (Lisp_Object start, Lisp_Object end, Lisp_Object printflag, Lisp_Object read_function)
1895{
1896 /* FIXME: Do the eval-sexp-add-defvars dance! */
1897 ptrdiff_t count = SPECPDL_INDEX ();
1898 Lisp_Object tem, cbuf;
1899
1900 cbuf = Fcurrent_buffer ();
1901
1902 if (NILP (printflag))
1903 tem = Qsymbolp;
1904 else
1905 tem = printflag;
1906 specbind (Qstandard_output, tem);
1907 specbind (Qeval_buffer_list, Fcons (cbuf, Veval_buffer_list));
1908
1909 /* `readevalloop' calls functions which check the type of start and end. */
1910 readevalloop (cbuf, 0, BVAR (XBUFFER (cbuf), filename),
1911 !NILP (printflag), Qnil, read_function,
1912 start, end);
1913
1914 return unbind_to (count, Qnil);
1915}
1916
1917\f
1918DEFUN ("read", Fread, Sread, 0, 1, 0,
1919 doc: /* Read one Lisp expression as text from STREAM, return as Lisp object.
1920If STREAM is nil, use the value of `standard-input' (which see).
1921STREAM or the value of `standard-input' may be:
1922 a buffer (read from point and advance it)
1923 a marker (read from where it points and advance it)
1924 a function (call it with no arguments for each character,
1925 call it with a char as argument to push a char back)
1926 a string (takes text from string, starting at the beginning)
1927 t (read text line using minibuffer and use it, or read from
1928 standard input in batch mode). */)
1929 (Lisp_Object stream)
1930{
1931 if (NILP (stream))
1932 stream = Vstandard_input;
1933 if (EQ (stream, Qt))
1934 stream = Qread_char;
1935 if (EQ (stream, Qread_char))
1936 return Fread_minibuffer (build_string ("Lisp expression: "), Qnil);
1937
1938 return read_internal_start (stream, Qnil, Qnil);
1939}
1940
1941DEFUN ("read-from-string", Fread_from_string, Sread_from_string, 1, 3, 0,
1942 doc: /* Read one Lisp expression which is represented as text by STRING.
1943Returns a cons: (OBJECT-READ . FINAL-STRING-INDEX).
1944FINAL-STRING-INDEX is an integer giving the position of the next
1945 remaining character in STRING.
1946START and END optionally delimit a substring of STRING from which to read;
1947 they default to 0 and (length STRING) respectively. */)
1948 (Lisp_Object string, Lisp_Object start, Lisp_Object end)
1949{
1950 Lisp_Object ret;
1951 CHECK_STRING (string);
1952 /* `read_internal_start' sets `read_from_string_index'. */
1953 ret = read_internal_start (string, start, end);
1954 return Fcons (ret, make_number (read_from_string_index));
1955}
1956
1957/* Function to set up the global context we need in toplevel read
1958 calls. */
1959static Lisp_Object
1960read_internal_start (Lisp_Object stream, Lisp_Object start, Lisp_Object end)
1961/* `start', `end' only used when stream is a string. */
1962{
1963 Lisp_Object retval;
1964
1965 readchar_count = 0;
1966 new_backquote_flag = 0;
1967 read_objects = Qnil;
1968 if (EQ (Vread_with_symbol_positions, Qt)
1969 || EQ (Vread_with_symbol_positions, stream))
1970 Vread_symbol_positions_list = Qnil;
1971
1972 if (STRINGP (stream)
1973 || ((CONSP (stream) && STRINGP (XCAR (stream)))))
1974 {
1975 ptrdiff_t startval, endval;
1976 Lisp_Object string;
1977
1978 if (STRINGP (stream))
1979 string = stream;
1980 else
1981 string = XCAR (stream);
1982
1983 if (NILP (end))
1984 endval = SCHARS (string);
1985 else
1986 {
1987 CHECK_NUMBER (end);
1988 if (! (0 <= XINT (end) && XINT (end) <= SCHARS (string)))
1989 args_out_of_range (string, end);
1990 endval = XINT (end);
1991 }
1992
1993 if (NILP (start))
1994 startval = 0;
1995 else
1996 {
1997 CHECK_NUMBER (start);
1998 if (! (0 <= XINT (start) && XINT (start) <= endval))
1999 args_out_of_range (string, start);
2000 startval = XINT (start);
2001 }
2002 read_from_string_index = startval;
2003 read_from_string_index_byte = string_char_to_byte (string, startval);
2004 read_from_string_limit = endval;
2005 }
2006
2007 retval = read0 (stream);
2008 if (EQ (Vread_with_symbol_positions, Qt)
2009 || EQ (Vread_with_symbol_positions, stream))
2010 Vread_symbol_positions_list = Fnreverse (Vread_symbol_positions_list);
2011 return retval;
2012}
2013\f
2014
2015/* Signal Qinvalid_read_syntax error.
2016 S is error string of length N (if > 0) */
2017
2018static _Noreturn void
2019invalid_syntax (const char *s)
2020{
2021 xsignal1 (Qinvalid_read_syntax, build_string (s));
2022}
2023
2024
2025/* Use this for recursive reads, in contexts where internal tokens
2026 are not allowed. */
2027
2028static Lisp_Object
2029read0 (Lisp_Object readcharfun)
2030{
2031 register Lisp_Object val;
2032 int c;
2033
2034 val = read1 (readcharfun, &c, 0);
2035 if (!c)
2036 return val;
2037
2038 xsignal1 (Qinvalid_read_syntax,
2039 Fmake_string (make_number (1), make_number (c)));
2040}
2041\f
2042static ptrdiff_t read_buffer_size;
2043static char *read_buffer;
2044
2045/* Read a \-escape sequence, assuming we already read the `\'.
2046 If the escape sequence forces unibyte, return eight-bit char. */
2047
2048static int
2049read_escape (Lisp_Object readcharfun, int stringp)
2050{
2051 register int c = READCHAR;
2052 /* \u allows up to four hex digits, \U up to eight. Default to the
2053 behavior for \u, and change this value in the case that \U is seen. */
2054 int unicode_hex_count = 4;
2055
2056 switch (c)
2057 {
2058 case -1:
2059 end_of_file_error ();
2060
2061 case 'a':
2062 return '\007';
2063 case 'b':
2064 return '\b';
2065 case 'd':
2066 return 0177;
2067 case 'e':
2068 return 033;
2069 case 'f':
2070 return '\f';
2071 case 'n':
2072 return '\n';
2073 case 'r':
2074 return '\r';
2075 case 't':
2076 return '\t';
2077 case 'v':
2078 return '\v';
2079 case '\n':
2080 return -1;
2081 case ' ':
2082 if (stringp)
2083 return -1;
2084 return ' ';
2085
2086 case 'M':
2087 c = READCHAR;
2088 if (c != '-')
2089 error ("Invalid escape character syntax");
2090 c = READCHAR;
2091 if (c == '\\')
2092 c = read_escape (readcharfun, 0);
2093 return c | meta_modifier;
2094
2095 case 'S':
2096 c = READCHAR;
2097 if (c != '-')
2098 error ("Invalid escape character syntax");
2099 c = READCHAR;
2100 if (c == '\\')
2101 c = read_escape (readcharfun, 0);
2102 return c | shift_modifier;
2103
2104 case 'H':
2105 c = READCHAR;
2106 if (c != '-')
2107 error ("Invalid escape character syntax");
2108 c = READCHAR;
2109 if (c == '\\')
2110 c = read_escape (readcharfun, 0);
2111 return c | hyper_modifier;
2112
2113 case 'A':
2114 c = READCHAR;
2115 if (c != '-')
2116 error ("Invalid escape character syntax");
2117 c = READCHAR;
2118 if (c == '\\')
2119 c = read_escape (readcharfun, 0);
2120 return c | alt_modifier;
2121
2122 case 's':
2123 c = READCHAR;
2124 if (stringp || c != '-')
2125 {
2126 UNREAD (c);
2127 return ' ';
2128 }
2129 c = READCHAR;
2130 if (c == '\\')
2131 c = read_escape (readcharfun, 0);
2132 return c | super_modifier;
2133
2134 case 'C':
2135 c = READCHAR;
2136 if (c != '-')
2137 error ("Invalid escape character syntax");
2138 case '^':
2139 c = READCHAR;
2140 if (c == '\\')
2141 c = read_escape (readcharfun, 0);
2142 if ((c & ~CHAR_MODIFIER_MASK) == '?')
2143 return 0177 | (c & CHAR_MODIFIER_MASK);
2144 else if (! SINGLE_BYTE_CHAR_P ((c & ~CHAR_MODIFIER_MASK)))
2145 return c | ctrl_modifier;
2146 /* ASCII control chars are made from letters (both cases),
2147 as well as the non-letters within 0100...0137. */
2148 else if ((c & 0137) >= 0101 && (c & 0137) <= 0132)
2149 return (c & (037 | ~0177));
2150 else if ((c & 0177) >= 0100 && (c & 0177) <= 0137)
2151 return (c & (037 | ~0177));
2152 else
2153 return c | ctrl_modifier;
2154
2155 case '0':
2156 case '1':
2157 case '2':
2158 case '3':
2159 case '4':
2160 case '5':
2161 case '6':
2162 case '7':
2163 /* An octal escape, as in ANSI C. */
2164 {
2165 register int i = c - '0';
2166 register int count = 0;
2167 while (++count < 3)
2168 {
2169 if ((c = READCHAR) >= '0' && c <= '7')
2170 {
2171 i *= 8;
2172 i += c - '0';
2173 }
2174 else
2175 {
2176 UNREAD (c);
2177 break;
2178 }
2179 }
2180
2181 if (i >= 0x80 && i < 0x100)
2182 i = BYTE8_TO_CHAR (i);
2183 return i;
2184 }
2185
2186 case 'x':
2187 /* A hex escape, as in ANSI C. */
2188 {
2189 unsigned int i = 0;
2190 int count = 0;
2191 while (1)
2192 {
2193 c = READCHAR;
2194 if (c >= '0' && c <= '9')
2195 {
2196 i *= 16;
2197 i += c - '0';
2198 }
2199 else if ((c >= 'a' && c <= 'f')
2200 || (c >= 'A' && c <= 'F'))
2201 {
2202 i *= 16;
2203 if (c >= 'a' && c <= 'f')
2204 i += c - 'a' + 10;
2205 else
2206 i += c - 'A' + 10;
2207 }
2208 else
2209 {
2210 UNREAD (c);
2211 break;
2212 }
2213 /* Allow hex escapes as large as ?\xfffffff, because some
2214 packages use them to denote characters with modifiers. */
2215 if ((CHAR_META | (CHAR_META - 1)) < i)
2216 error ("Hex character out of range: \\x%x...", i);
2217 count += count < 3;
2218 }
2219
2220 if (count < 3 && i >= 0x80)
2221 return BYTE8_TO_CHAR (i);
2222 return i;
2223 }
2224
2225 case 'U':
2226 /* Post-Unicode-2.0: Up to eight hex chars. */
2227 unicode_hex_count = 8;
2228 case 'u':
2229
2230 /* A Unicode escape. We only permit them in strings and characters,
2231 not arbitrarily in the source code, as in some other languages. */
2232 {
2233 unsigned int i = 0;
2234 int count = 0;
2235
2236 while (++count <= unicode_hex_count)
2237 {
2238 c = READCHAR;
2239 /* `isdigit' and `isalpha' may be locale-specific, which we don't
2240 want. */
2241 if (c >= '0' && c <= '9') i = (i << 4) + (c - '0');
2242 else if (c >= 'a' && c <= 'f') i = (i << 4) + (c - 'a') + 10;
2243 else if (c >= 'A' && c <= 'F') i = (i << 4) + (c - 'A') + 10;
2244 else
2245 error ("Non-hex digit used for Unicode escape");
2246 }
2247 if (i > 0x10FFFF)
2248 error ("Non-Unicode character: 0x%x", i);
2249 return i;
2250 }
2251
2252 default:
2253 return c;
2254 }
2255}
2256
2257/* Return the digit that CHARACTER stands for in the given BASE.
2258 Return -1 if CHARACTER is out of range for BASE,
2259 and -2 if CHARACTER is not valid for any supported BASE. */
2260static inline int
2261digit_to_number (int character, int base)
2262{
2263 int digit;
2264
2265 if ('0' <= character && character <= '9')
2266 digit = character - '0';
2267 else if ('a' <= character && character <= 'z')
2268 digit = character - 'a' + 10;
2269 else if ('A' <= character && character <= 'Z')
2270 digit = character - 'A' + 10;
2271 else
2272 return -2;
2273
2274 return digit < base ? digit : -1;
2275}
2276
2277/* Read an integer in radix RADIX using READCHARFUN to read
2278 characters. RADIX must be in the interval [2..36]; if it isn't, a
2279 read error is signaled . Value is the integer read. Signals an
2280 error if encountering invalid read syntax or if RADIX is out of
2281 range. */
2282
2283static Lisp_Object
2284read_integer (Lisp_Object readcharfun, EMACS_INT radix)
2285{
2286 /* Room for sign, leading 0, other digits, trailing null byte.
2287 Also, room for invalid syntax diagnostic. */
2288 char buf[max (1 + 1 + sizeof (uintmax_t) * CHAR_BIT + 1,
2289 sizeof "integer, radix " + INT_STRLEN_BOUND (EMACS_INT))];
2290
2291 int valid = -1; /* 1 if valid, 0 if not, -1 if incomplete. */
2292
2293 if (radix < 2 || radix > 36)
2294 valid = 0;
2295 else
2296 {
2297 char *p = buf;
2298 int c, digit;
2299
2300 c = READCHAR;
2301 if (c == '-' || c == '+')
2302 {
2303 *p++ = c;
2304 c = READCHAR;
2305 }
2306
2307 if (c == '0')
2308 {
2309 *p++ = c;
2310 valid = 1;
2311
2312 /* Ignore redundant leading zeros, so the buffer doesn't
2313 fill up with them. */
2314 do
2315 c = READCHAR;
2316 while (c == '0');
2317 }
2318
2319 while (-1 <= (digit = digit_to_number (c, radix)))
2320 {
2321 if (digit == -1)
2322 valid = 0;
2323 if (valid < 0)
2324 valid = 1;
2325
2326 if (p < buf + sizeof buf - 1)
2327 *p++ = c;
2328 else
2329 valid = 0;
2330
2331 c = READCHAR;
2332 }
2333
2334 UNREAD (c);
2335 *p = '\0';
2336 }
2337
2338 if (! valid)
2339 {
2340 sprintf (buf, "integer, radix %"pI"d", radix);
2341 invalid_syntax (buf);
2342 }
2343
2344 return string_to_number (buf, radix, 0);
2345}
2346
2347
2348/* If the next token is ')' or ']' or '.', we store that character
2349 in *PCH and the return value is not interesting. Else, we store
2350 zero in *PCH and we read and return one lisp object.
2351
2352 FIRST_IN_LIST is nonzero if this is the first element of a list. */
2353
2354static Lisp_Object
2355read1 (register Lisp_Object readcharfun, int *pch, int first_in_list)
2356{
2357 register int c;
2358 unsigned uninterned_symbol = 0;
2359 int multibyte;
2360
2361 *pch = 0;
2362 load_each_byte = 0;
2363
2364 retry:
2365
2366 c = READCHAR_REPORT_MULTIBYTE (&multibyte);
2367 if (c < 0)
2368 end_of_file_error ();
2369
2370 switch (c)
2371 {
2372 case '(':
2373 return read_list (0, readcharfun);
2374
2375 case '[':
2376 return read_vector (readcharfun, 0);
2377
2378 case ')':
2379 case ']':
2380 {
2381 *pch = c;
2382 return Qnil;
2383 }
2384
2385 case '#':
2386 c = READCHAR;
2387 if (c == 's')
2388 {
2389 c = READCHAR;
2390 if (c == '(')
2391 {
2392 /* Accept extended format for hashtables (extensible to
2393 other types), e.g.
2394 #s(hash-table size 2 test equal data (k1 v1 k2 v2)) */
2395 Lisp_Object tmp = read_list (0, readcharfun);
2396 Lisp_Object head = CAR_SAFE (tmp);
2397 Lisp_Object data = Qnil;
2398 Lisp_Object val = Qnil;
2399 /* The size is 2 * number of allowed keywords to
2400 make-hash-table. */
2401 Lisp_Object params[10];
2402 Lisp_Object ht;
2403 Lisp_Object key = Qnil;
2404 int param_count = 0;
2405
2406 if (!EQ (head, Qhash_table))
2407 error ("Invalid extended read marker at head of #s list "
2408 "(only hash-table allowed)");
2409
2410 tmp = CDR_SAFE (tmp);
2411
2412 /* This is repetitive but fast and simple. */
2413 params[param_count] = QCsize;
2414 params[param_count + 1] = Fplist_get (tmp, Qsize);
2415 if (!NILP (params[param_count + 1]))
2416 param_count += 2;
2417
2418 params[param_count] = QCtest;
2419 params[param_count + 1] = Fplist_get (tmp, Qtest);
2420 if (!NILP (params[param_count + 1]))
2421 param_count += 2;
2422
2423 params[param_count] = QCweakness;
2424 params[param_count + 1] = Fplist_get (tmp, Qweakness);
2425 if (!NILP (params[param_count + 1]))
2426 param_count += 2;
2427
2428 params[param_count] = QCrehash_size;
2429 params[param_count + 1] = Fplist_get (tmp, Qrehash_size);
2430 if (!NILP (params[param_count + 1]))
2431 param_count += 2;
2432
2433 params[param_count] = QCrehash_threshold;
2434 params[param_count + 1] = Fplist_get (tmp, Qrehash_threshold);
2435 if (!NILP (params[param_count + 1]))
2436 param_count += 2;
2437
2438 /* This is the hashtable data. */
2439 data = Fplist_get (tmp, Qdata);
2440
2441 /* Now use params to make a new hashtable and fill it. */
2442 ht = Fmake_hash_table (param_count, params);
2443
2444 while (CONSP (data))
2445 {
2446 key = XCAR (data);
2447 data = XCDR (data);
2448 if (!CONSP (data))
2449 error ("Odd number of elements in hashtable data");
2450 val = XCAR (data);
2451 data = XCDR (data);
2452 Fputhash (key, val, ht);
2453 }
2454
2455 return ht;
2456 }
2457 UNREAD (c);
2458 invalid_syntax ("#");
2459 }
2460 if (c == '^')
2461 {
2462 c = READCHAR;
2463 if (c == '[')
2464 {
2465 Lisp_Object tmp;
2466 tmp = read_vector (readcharfun, 0);
2467 if (ASIZE (tmp) < CHAR_TABLE_STANDARD_SLOTS)
2468 error ("Invalid size char-table");
2469 XSETPVECTYPE (XVECTOR (tmp), PVEC_CHAR_TABLE);
2470 return tmp;
2471 }
2472 else if (c == '^')
2473 {
2474 c = READCHAR;
2475 if (c == '[')
2476 {
2477 Lisp_Object tmp;
2478 int depth;
2479 ptrdiff_t size;
2480
2481 tmp = read_vector (readcharfun, 0);
2482 size = ASIZE (tmp);
2483 if (size == 0)
2484 error ("Invalid size char-table");
2485 if (! RANGED_INTEGERP (1, AREF (tmp, 0), 3))
2486 error ("Invalid depth in char-table");
2487 depth = XINT (AREF (tmp, 0));
2488 if (chartab_size[depth] != size - 2)
2489 error ("Invalid size char-table");
2490 XSETPVECTYPE (XVECTOR (tmp), PVEC_SUB_CHAR_TABLE);
2491 return tmp;
2492 }
2493 invalid_syntax ("#^^");
2494 }
2495 invalid_syntax ("#^");
2496 }
2497 if (c == '&')
2498 {
2499 Lisp_Object length;
2500 length = read1 (readcharfun, pch, first_in_list);
2501 c = READCHAR;
2502 if (c == '"')
2503 {
2504 Lisp_Object tmp, val;
2505 EMACS_INT size_in_chars
2506 = ((XFASTINT (length) + BOOL_VECTOR_BITS_PER_CHAR - 1)
2507 / BOOL_VECTOR_BITS_PER_CHAR);
2508
2509 UNREAD (c);
2510 tmp = read1 (readcharfun, pch, first_in_list);
2511 if (STRING_MULTIBYTE (tmp)
2512 || (size_in_chars != SCHARS (tmp)
2513 /* We used to print 1 char too many
2514 when the number of bits was a multiple of 8.
2515 Accept such input in case it came from an old
2516 version. */
2517 && ! (XFASTINT (length)
2518 == (SCHARS (tmp) - 1) * BOOL_VECTOR_BITS_PER_CHAR)))
2519 invalid_syntax ("#&...");
2520
2521 val = Fmake_bool_vector (length, Qnil);
2522 memcpy (XBOOL_VECTOR (val)->data, SDATA (tmp), size_in_chars);
2523 /* Clear the extraneous bits in the last byte. */
2524 if (XINT (length) != size_in_chars * BOOL_VECTOR_BITS_PER_CHAR)
2525 XBOOL_VECTOR (val)->data[size_in_chars - 1]
2526 &= (1 << (XINT (length) % BOOL_VECTOR_BITS_PER_CHAR)) - 1;
2527 return val;
2528 }
2529 invalid_syntax ("#&...");
2530 }
2531 if (c == '[')
2532 {
2533 /* Accept compiled functions at read-time so that we don't have to
2534 build them using function calls. */
2535 Lisp_Object tmp;
2536 tmp = read_vector (readcharfun, 1);
2537 make_byte_code (XVECTOR (tmp));
2538 return tmp;
2539 }
2540 if (c == '(')
2541 {
2542 Lisp_Object tmp;
2543 struct gcpro gcpro1;
2544 int ch;
2545
2546 /* Read the string itself. */
2547 tmp = read1 (readcharfun, &ch, 0);
2548 if (ch != 0 || !STRINGP (tmp))
2549 invalid_syntax ("#");
2550 GCPRO1 (tmp);
2551 /* Read the intervals and their properties. */
2552 while (1)
2553 {
2554 Lisp_Object beg, end, plist;
2555
2556 beg = read1 (readcharfun, &ch, 0);
2557 end = plist = Qnil;
2558 if (ch == ')')
2559 break;
2560 if (ch == 0)
2561 end = read1 (readcharfun, &ch, 0);
2562 if (ch == 0)
2563 plist = read1 (readcharfun, &ch, 0);
2564 if (ch)
2565 invalid_syntax ("Invalid string property list");
2566 Fset_text_properties (beg, end, plist, tmp);
2567 }
2568 UNGCPRO;
2569 return tmp;
2570 }
2571
2572 /* #@NUMBER is used to skip NUMBER following characters.
2573 That's used in .elc files to skip over doc strings
2574 and function definitions. */
2575 if (c == '@')
2576 {
2577 enum { extra = 100 };
2578 ptrdiff_t i, nskip = 0;
2579
2580 load_each_byte = 1;
2581 /* Read a decimal integer. */
2582 while ((c = READCHAR) >= 0
2583 && c >= '0' && c <= '9')
2584 {
2585 if ((STRING_BYTES_BOUND - extra) / 10 <= nskip)
2586 string_overflow ();
2587 nskip *= 10;
2588 nskip += c - '0';
2589 }
2590 UNREAD (c);
2591
2592 if (load_force_doc_strings
2593 && (EQ (readcharfun, Qget_file_char)
2594 || EQ (readcharfun, Qget_emacs_mule_file_char)))
2595 {
2596 /* If we are supposed to force doc strings into core right now,
2597 record the last string that we skipped,
2598 and record where in the file it comes from. */
2599
2600 /* But first exchange saved_doc_string
2601 with prev_saved_doc_string, so we save two strings. */
2602 {
2603 char *temp = saved_doc_string;
2604 ptrdiff_t temp_size = saved_doc_string_size;
2605 file_offset temp_pos = saved_doc_string_position;
2606 ptrdiff_t temp_len = saved_doc_string_length;
2607
2608 saved_doc_string = prev_saved_doc_string;
2609 saved_doc_string_size = prev_saved_doc_string_size;
2610 saved_doc_string_position = prev_saved_doc_string_position;
2611 saved_doc_string_length = prev_saved_doc_string_length;
2612
2613 prev_saved_doc_string = temp;
2614 prev_saved_doc_string_size = temp_size;
2615 prev_saved_doc_string_position = temp_pos;
2616 prev_saved_doc_string_length = temp_len;
2617 }
2618
2619 if (saved_doc_string_size == 0)
2620 {
2621 saved_doc_string = xmalloc (nskip + extra);
2622 saved_doc_string_size = nskip + extra;
2623 }
2624 if (nskip > saved_doc_string_size)
2625 {
2626 saved_doc_string = xrealloc (saved_doc_string, nskip + extra);
2627 saved_doc_string_size = nskip + extra;
2628 }
2629
2630 saved_doc_string_position = file_tell (instream);
2631
2632 /* Copy that many characters into saved_doc_string. */
2633 for (i = 0; i < nskip && c >= 0; i++)
2634 saved_doc_string[i] = c = READCHAR;
2635
2636 saved_doc_string_length = i;
2637 }
2638 else
2639 {
2640 /* Skip that many characters. */
2641 for (i = 0; i < nskip && c >= 0; i++)
2642 c = READCHAR;
2643 }
2644
2645 load_each_byte = 0;
2646 goto retry;
2647 }
2648 if (c == '!')
2649 {
2650 /* #! appears at the beginning of an executable file.
2651 Skip the first line. */
2652 while (c != '\n' && c >= 0)
2653 c = READCHAR;
2654 goto retry;
2655 }
2656 if (c == '$')
2657 return Vload_file_name;
2658 if (c == '\'')
2659 return Fcons (Qfunction, Fcons (read0 (readcharfun), Qnil));
2660 /* #:foo is the uninterned symbol named foo. */
2661 if (c == ':')
2662 {
2663 uninterned_symbol = 1;
2664 c = READCHAR;
2665 if (!(c > 040
2666 && c != 0xa0 /* NBSP */
2667 && (c >= 0200
2668 || strchr ("\"';()[]#`,", c) == NULL)))
2669 {
2670 /* No symbol character follows, this is the empty
2671 symbol. */
2672 UNREAD (c);
2673 return Fmake_symbol (empty_unibyte_string);
2674 }
2675 goto read_symbol;
2676 }
2677 /* ## is the empty symbol. */
2678 if (c == '#')
2679 return Fintern (empty_unibyte_string, Qnil);
2680 /* Reader forms that can reuse previously read objects. */
2681 if (c >= '0' && c <= '9')
2682 {
2683 EMACS_INT n = 0;
2684 Lisp_Object tem;
2685
2686 /* Read a non-negative integer. */
2687 while (c >= '0' && c <= '9')
2688 {
2689 if (MOST_POSITIVE_FIXNUM / 10 < n
2690 || MOST_POSITIVE_FIXNUM < n * 10 + c - '0')
2691 n = MOST_POSITIVE_FIXNUM + 1;
2692 else
2693 n = n * 10 + c - '0';
2694 c = READCHAR;
2695 }
2696
2697 if (n <= MOST_POSITIVE_FIXNUM)
2698 {
2699 if (c == 'r' || c == 'R')
2700 return read_integer (readcharfun, n);
2701
2702 if (! NILP (Vread_circle))
2703 {
2704 /* #n=object returns object, but associates it with
2705 n for #n#. */
2706 if (c == '=')
2707 {
2708 /* Make a placeholder for #n# to use temporarily. */
2709 Lisp_Object placeholder;
2710 Lisp_Object cell;
2711
2712 placeholder = Fcons (Qnil, Qnil);
2713 cell = Fcons (make_number (n), placeholder);
2714 read_objects = Fcons (cell, read_objects);
2715
2716 /* Read the object itself. */
2717 tem = read0 (readcharfun);
2718
2719 /* Now put it everywhere the placeholder was... */
2720 substitute_object_in_subtree (tem, placeholder);
2721
2722 /* ...and #n# will use the real value from now on. */
2723 Fsetcdr (cell, tem);
2724
2725 return tem;
2726 }
2727
2728 /* #n# returns a previously read object. */
2729 if (c == '#')
2730 {
2731 tem = Fassq (make_number (n), read_objects);
2732 if (CONSP (tem))
2733 return XCDR (tem);
2734 }
2735 }
2736 }
2737 /* Fall through to error message. */
2738 }
2739 else if (c == 'x' || c == 'X')
2740 return read_integer (readcharfun, 16);
2741 else if (c == 'o' || c == 'O')
2742 return read_integer (readcharfun, 8);
2743 else if (c == 'b' || c == 'B')
2744 return read_integer (readcharfun, 2);
2745
2746 UNREAD (c);
2747 invalid_syntax ("#");
2748
2749 case ';':
2750 while ((c = READCHAR) >= 0 && c != '\n');
2751 goto retry;
2752
2753 case '\'':
2754 {
2755 return Fcons (Qquote, Fcons (read0 (readcharfun), Qnil));
2756 }
2757
2758 case '`':
2759 {
2760 int next_char = READCHAR;
2761 UNREAD (next_char);
2762 /* Transition from old-style to new-style:
2763 If we see "(`" it used to mean old-style, which usually works
2764 fine because ` should almost never appear in such a position
2765 for new-style. But occasionally we need "(`" to mean new
2766 style, so we try to distinguish the two by the fact that we
2767 can either write "( `foo" or "(` foo", where the first
2768 intends to use new-style whereas the second intends to use
2769 old-style. For Emacs-25, we should completely remove this
2770 first_in_list exception (old-style can still be obtained via
2771 "(\`" anyway). */
2772 if (!new_backquote_flag && first_in_list && next_char == ' ')
2773 {
2774 Vold_style_backquotes = Qt;
2775 goto default_label;
2776 }
2777 else
2778 {
2779 Lisp_Object value;
2780
2781 new_backquote_flag++;
2782 value = read0 (readcharfun);
2783 new_backquote_flag--;
2784
2785 return Fcons (Qbackquote, Fcons (value, Qnil));
2786 }
2787 }
2788 case ',':
2789 {
2790 int next_char = READCHAR;
2791 UNREAD (next_char);
2792 /* Transition from old-style to new-style:
2793 It used to be impossible to have a new-style , other than within
2794 a new-style `. This is sufficient when ` and , are used in the
2795 normal way, but ` and , can also appear in args to macros that
2796 will not interpret them in the usual way, in which case , may be
2797 used without any ` anywhere near.
2798 So we now use the same heuristic as for backquote: old-style
2799 unquotes are only recognized when first on a list, and when
2800 followed by a space.
2801 Because it's more difficult to peek 2 chars ahead, a new-style
2802 ,@ can still not be used outside of a `, unless it's in the middle
2803 of a list. */
2804 if (new_backquote_flag
2805 || !first_in_list
2806 || (next_char != ' ' && next_char != '@'))
2807 {
2808 Lisp_Object comma_type = Qnil;
2809 Lisp_Object value;
2810 int ch = READCHAR;
2811
2812 if (ch == '@')
2813 comma_type = Qcomma_at;
2814 else if (ch == '.')
2815 comma_type = Qcomma_dot;
2816 else
2817 {
2818 if (ch >= 0) UNREAD (ch);
2819 comma_type = Qcomma;
2820 }
2821
2822 value = read0 (readcharfun);
2823 return Fcons (comma_type, Fcons (value, Qnil));
2824 }
2825 else
2826 {
2827 Vold_style_backquotes = Qt;
2828 goto default_label;
2829 }
2830 }
2831 case '?':
2832 {
2833 int modifiers;
2834 int next_char;
2835 int ok;
2836
2837 c = READCHAR;
2838 if (c < 0)
2839 end_of_file_error ();
2840
2841 /* Accept `single space' syntax like (list ? x) where the
2842 whitespace character is SPC or TAB.
2843 Other literal whitespace like NL, CR, and FF are not accepted,
2844 as there are well-established escape sequences for these. */
2845 if (c == ' ' || c == '\t')
2846 return make_number (c);
2847
2848 if (c == '\\')
2849 c = read_escape (readcharfun, 0);
2850 modifiers = c & CHAR_MODIFIER_MASK;
2851 c &= ~CHAR_MODIFIER_MASK;
2852 if (CHAR_BYTE8_P (c))
2853 c = CHAR_TO_BYTE8 (c);
2854 c |= modifiers;
2855
2856 next_char = READCHAR;
2857 ok = (next_char <= 040
2858 || (next_char < 0200
2859 && strchr ("\"';()[]#?`,.", next_char) != NULL));
2860 UNREAD (next_char);
2861 if (ok)
2862 return make_number (c);
2863
2864 invalid_syntax ("?");
2865 }
2866
2867 case '"':
2868 {
2869 char *p = read_buffer;
2870 char *end = read_buffer + read_buffer_size;
2871 register int ch;
2872 /* Nonzero if we saw an escape sequence specifying
2873 a multibyte character. */
2874 int force_multibyte = 0;
2875 /* Nonzero if we saw an escape sequence specifying
2876 a single-byte character. */
2877 int force_singlebyte = 0;
2878 int cancel = 0;
2879 ptrdiff_t nchars = 0;
2880
2881 while ((ch = READCHAR) >= 0
2882 && ch != '\"')
2883 {
2884 if (end - p < MAX_MULTIBYTE_LENGTH)
2885 {
2886 ptrdiff_t offset = p - read_buffer;
2887 if (min (PTRDIFF_MAX, SIZE_MAX) / 2 < read_buffer_size)
2888 memory_full (SIZE_MAX);
2889 read_buffer = xrealloc (read_buffer, read_buffer_size * 2);
2890 read_buffer_size *= 2;
2891 p = read_buffer + offset;
2892 end = read_buffer + read_buffer_size;
2893 }
2894
2895 if (ch == '\\')
2896 {
2897 int modifiers;
2898
2899 ch = read_escape (readcharfun, 1);
2900
2901 /* CH is -1 if \ newline has just been seen. */
2902 if (ch == -1)
2903 {
2904 if (p == read_buffer)
2905 cancel = 1;
2906 continue;
2907 }
2908
2909 modifiers = ch & CHAR_MODIFIER_MASK;
2910 ch = ch & ~CHAR_MODIFIER_MASK;
2911
2912 if (CHAR_BYTE8_P (ch))
2913 force_singlebyte = 1;
2914 else if (! ASCII_CHAR_P (ch))
2915 force_multibyte = 1;
2916 else /* I.e. ASCII_CHAR_P (ch). */
2917 {
2918 /* Allow `\C- ' and `\C-?'. */
2919 if (modifiers == CHAR_CTL)
2920 {
2921 if (ch == ' ')
2922 ch = 0, modifiers = 0;
2923 else if (ch == '?')
2924 ch = 127, modifiers = 0;
2925 }
2926 if (modifiers & CHAR_SHIFT)
2927 {
2928 /* Shift modifier is valid only with [A-Za-z]. */
2929 if (ch >= 'A' && ch <= 'Z')
2930 modifiers &= ~CHAR_SHIFT;
2931 else if (ch >= 'a' && ch <= 'z')
2932 ch -= ('a' - 'A'), modifiers &= ~CHAR_SHIFT;
2933 }
2934
2935 if (modifiers & CHAR_META)
2936 {
2937 /* Move the meta bit to the right place for a
2938 string. */
2939 modifiers &= ~CHAR_META;
2940 ch = BYTE8_TO_CHAR (ch | 0x80);
2941 force_singlebyte = 1;
2942 }
2943 }
2944
2945 /* Any modifiers remaining are invalid. */
2946 if (modifiers)
2947 error ("Invalid modifier in string");
2948 p += CHAR_STRING (ch, (unsigned char *) p);
2949 }
2950 else
2951 {
2952 p += CHAR_STRING (ch, (unsigned char *) p);
2953 if (CHAR_BYTE8_P (ch))
2954 force_singlebyte = 1;
2955 else if (! ASCII_CHAR_P (ch))
2956 force_multibyte = 1;
2957 }
2958 nchars++;
2959 }
2960
2961 if (ch < 0)
2962 end_of_file_error ();
2963
2964 /* If purifying, and string starts with \ newline,
2965 return zero instead. This is for doc strings
2966 that we are really going to find in etc/DOC.nn.nn. */
2967 if (!NILP (Vpurify_flag) && NILP (Vdoc_file_name) && cancel)
2968 return make_number (0);
2969
2970 if (! force_multibyte && force_singlebyte)
2971 {
2972 /* READ_BUFFER contains raw 8-bit bytes and no multibyte
2973 forms. Convert it to unibyte. */
2974 nchars = str_as_unibyte ((unsigned char *) read_buffer,
2975 p - read_buffer);
2976 p = read_buffer + nchars;
2977 }
2978
2979 return make_specified_string (read_buffer, nchars, p - read_buffer,
2980 (force_multibyte
2981 || (p - read_buffer != nchars)));
2982 }
2983
2984 case '.':
2985 {
2986 int next_char = READCHAR;
2987 UNREAD (next_char);
2988
2989 if (next_char <= 040
2990 || (next_char < 0200
2991 && strchr ("\"';([#?`,", next_char) != NULL))
2992 {
2993 *pch = c;
2994 return Qnil;
2995 }
2996
2997 /* Otherwise, we fall through! Note that the atom-reading loop
2998 below will now loop at least once, assuring that we will not
2999 try to UNREAD two characters in a row. */
3000 }
3001 default:
3002 default_label:
3003 if (c <= 040) goto retry;
3004 if (c == 0xa0) /* NBSP */
3005 goto retry;
3006
3007 read_symbol:
3008 {
3009 char *p = read_buffer;
3010 int quoted = 0;
3011 EMACS_INT start_position = readchar_count - 1;
3012
3013 {
3014 char *end = read_buffer + read_buffer_size;
3015
3016 do
3017 {
3018 if (end - p < MAX_MULTIBYTE_LENGTH)
3019 {
3020 ptrdiff_t offset = p - read_buffer;
3021 if (min (PTRDIFF_MAX, SIZE_MAX) / 2 < read_buffer_size)
3022 memory_full (SIZE_MAX);
3023 read_buffer = xrealloc (read_buffer, read_buffer_size * 2);
3024 read_buffer_size *= 2;
3025 p = read_buffer + offset;
3026 end = read_buffer + read_buffer_size;
3027 }
3028
3029 if (c == '\\')
3030 {
3031 c = READCHAR;
3032 if (c == -1)
3033 end_of_file_error ();
3034 quoted = 1;
3035 }
3036
3037 if (multibyte)
3038 p += CHAR_STRING (c, (unsigned char *) p);
3039 else
3040 *p++ = c;
3041 c = READCHAR;
3042 }
3043 while (c > 040
3044 && c != 0xa0 /* NBSP */
3045 && (c >= 0200
3046 || strchr ("\"';()[]#`,", c) == NULL));
3047
3048 if (p == end)
3049 {
3050 ptrdiff_t offset = p - read_buffer;
3051 if (min (PTRDIFF_MAX, SIZE_MAX) / 2 < read_buffer_size)
3052 memory_full (SIZE_MAX);
3053 read_buffer = xrealloc (read_buffer, read_buffer_size * 2);
3054 read_buffer_size *= 2;
3055 p = read_buffer + offset;
3056 end = read_buffer + read_buffer_size;
3057 }
3058 *p = 0;
3059 UNREAD (c);
3060 }
3061
3062 if (!quoted && !uninterned_symbol)
3063 {
3064 Lisp_Object result = string_to_number (read_buffer, 10, 0);
3065 if (! NILP (result))
3066 return result;
3067 }
3068 {
3069 Lisp_Object name, result;
3070 ptrdiff_t nbytes = p - read_buffer;
3071 ptrdiff_t nchars
3072 = (multibyte
3073 ? multibyte_chars_in_text ((unsigned char *) read_buffer,
3074 nbytes)
3075 : nbytes);
3076
3077 name = ((uninterned_symbol && ! NILP (Vpurify_flag)
3078 ? make_pure_string : make_specified_string)
3079 (read_buffer, nchars, nbytes, multibyte));
3080 result = (uninterned_symbol ? Fmake_symbol (name)
3081 : Fintern (name, Qnil));
3082
3083 if (EQ (Vread_with_symbol_positions, Qt)
3084 || EQ (Vread_with_symbol_positions, readcharfun))
3085 Vread_symbol_positions_list
3086 = Fcons (Fcons (result, make_number (start_position)),
3087 Vread_symbol_positions_list);
3088 return result;
3089 }
3090 }
3091 }
3092}
3093\f
3094
3095/* List of nodes we've seen during substitute_object_in_subtree. */
3096static Lisp_Object seen_list;
3097
3098static void
3099substitute_object_in_subtree (Lisp_Object object, Lisp_Object placeholder)
3100{
3101 Lisp_Object check_object;
3102
3103 /* We haven't seen any objects when we start. */
3104 seen_list = Qnil;
3105
3106 /* Make all the substitutions. */
3107 check_object
3108 = substitute_object_recurse (object, placeholder, object);
3109
3110 /* Clear seen_list because we're done with it. */
3111 seen_list = Qnil;
3112
3113 /* The returned object here is expected to always eq the
3114 original. */
3115 if (!EQ (check_object, object))
3116 error ("Unexpected mutation error in reader");
3117}
3118
3119/* Feval doesn't get called from here, so no gc protection is needed. */
3120#define SUBSTITUTE(get_val, set_val) \
3121 do { \
3122 Lisp_Object old_value = get_val; \
3123 Lisp_Object true_value \
3124 = substitute_object_recurse (object, placeholder, \
3125 old_value); \
3126 \
3127 if (!EQ (old_value, true_value)) \
3128 { \
3129 set_val; \
3130 } \
3131 } while (0)
3132
3133static Lisp_Object
3134substitute_object_recurse (Lisp_Object object, Lisp_Object placeholder, Lisp_Object subtree)
3135{
3136 /* If we find the placeholder, return the target object. */
3137 if (EQ (placeholder, subtree))
3138 return object;
3139
3140 /* If we've been to this node before, don't explore it again. */
3141 if (!EQ (Qnil, Fmemq (subtree, seen_list)))
3142 return subtree;
3143
3144 /* If this node can be the entry point to a cycle, remember that
3145 we've seen it. It can only be such an entry point if it was made
3146 by #n=, which means that we can find it as a value in
3147 read_objects. */
3148 if (!EQ (Qnil, Frassq (subtree, read_objects)))
3149 seen_list = Fcons (subtree, seen_list);
3150
3151 /* Recurse according to subtree's type.
3152 Every branch must return a Lisp_Object. */
3153 switch (XTYPE (subtree))
3154 {
3155 case Lisp_Vectorlike:
3156 {
3157 ptrdiff_t i, length = 0;
3158 if (BOOL_VECTOR_P (subtree))
3159 return subtree; /* No sub-objects anyway. */
3160 else if (CHAR_TABLE_P (subtree) || SUB_CHAR_TABLE_P (subtree)
3161 || COMPILEDP (subtree))
3162 length = ASIZE (subtree) & PSEUDOVECTOR_SIZE_MASK;
3163 else if (VECTORP (subtree))
3164 length = ASIZE (subtree);
3165 else
3166 /* An unknown pseudovector may contain non-Lisp fields, so we
3167 can't just blindly traverse all its fields. We used to call
3168 `Flength' which signaled `sequencep', so I just preserved this
3169 behavior. */
3170 wrong_type_argument (Qsequencep, subtree);
3171
3172 for (i = 0; i < length; i++)
3173 SUBSTITUTE (AREF (subtree, i),
3174 ASET (subtree, i, true_value));
3175 return subtree;
3176 }
3177
3178 case Lisp_Cons:
3179 {
3180 SUBSTITUTE (XCAR (subtree),
3181 XSETCAR (subtree, true_value));
3182 SUBSTITUTE (XCDR (subtree),
3183 XSETCDR (subtree, true_value));
3184 return subtree;
3185 }
3186
3187 case Lisp_String:
3188 {
3189 /* Check for text properties in each interval.
3190 substitute_in_interval contains part of the logic. */
3191
3192 INTERVAL root_interval = STRING_INTERVALS (subtree);
3193 Lisp_Object arg = Fcons (object, placeholder);
3194
3195 traverse_intervals_noorder (root_interval,
3196 &substitute_in_interval, arg);
3197
3198 return subtree;
3199 }
3200
3201 /* Other types don't recurse any further. */
3202 default:
3203 return subtree;
3204 }
3205}
3206
3207/* Helper function for substitute_object_recurse. */
3208static void
3209substitute_in_interval (INTERVAL interval, Lisp_Object arg)
3210{
3211 Lisp_Object object = Fcar (arg);
3212 Lisp_Object placeholder = Fcdr (arg);
3213
3214 SUBSTITUTE (interval->plist, interval->plist = true_value);
3215}
3216
3217\f
3218#define LEAD_INT 1
3219#define DOT_CHAR 2
3220#define TRAIL_INT 4
3221#define E_EXP 16
3222
3223
3224/* Convert STRING to a number, assuming base BASE. Return a fixnum if CP has
3225 integer syntax and fits in a fixnum, else return the nearest float if CP has
3226 either floating point or integer syntax and BASE is 10, else return nil. If
3227 IGNORE_TRAILING is nonzero, consider just the longest prefix of CP that has
3228 valid floating point syntax. Signal an overflow if BASE is not 10 and the
3229 number has integer syntax but does not fit. */
3230
3231Lisp_Object
3232string_to_number (char const *string, int base, int ignore_trailing)
3233{
3234 int state;
3235 char const *cp = string;
3236 int leading_digit;
3237 int float_syntax = 0;
3238 double value = 0;
3239
3240 /* Compute NaN and infinities using a variable, to cope with compilers that
3241 think they are smarter than we are. */
3242 double zero = 0;
3243
3244 /* Negate the value ourselves. This treats 0, NaNs, and infinity properly on
3245 IEEE floating point hosts, and works around a formerly-common bug where
3246 atof ("-0.0") drops the sign. */
3247 int negative = *cp == '-';
3248
3249 int signedp = negative || *cp == '+';
3250 cp += signedp;
3251
3252 state = 0;
3253
3254 leading_digit = digit_to_number (*cp, base);
3255 if (0 <= leading_digit)
3256 {
3257 state |= LEAD_INT;
3258 do
3259 ++cp;
3260 while (0 <= digit_to_number (*cp, base));
3261 }
3262 if (*cp == '.')
3263 {
3264 state |= DOT_CHAR;
3265 cp++;
3266 }
3267
3268 if (base == 10)
3269 {
3270 if ('0' <= *cp && *cp <= '9')
3271 {
3272 state |= TRAIL_INT;
3273 do
3274 cp++;
3275 while ('0' <= *cp && *cp <= '9');
3276 }
3277 if (*cp == 'e' || *cp == 'E')
3278 {
3279 char const *ecp = cp;
3280 cp++;
3281 if (*cp == '+' || *cp == '-')
3282 cp++;
3283 if ('0' <= *cp && *cp <= '9')
3284 {
3285 state |= E_EXP;
3286 do
3287 cp++;
3288 while ('0' <= *cp && *cp <= '9');
3289 }
3290 else if (cp[-1] == '+'
3291 && cp[0] == 'I' && cp[1] == 'N' && cp[2] == 'F')
3292 {
3293 state |= E_EXP;
3294 cp += 3;
3295 value = 1.0 / zero;
3296 }
3297 else if (cp[-1] == '+'
3298 && cp[0] == 'N' && cp[1] == 'a' && cp[2] == 'N')
3299 {
3300 state |= E_EXP;
3301 cp += 3;
3302 value = zero / zero;
3303
3304 /* If that made a "negative" NaN, negate it. */
3305 {
3306 int i;
3307 union { double d; char c[sizeof (double)]; }
3308 u_data, u_minus_zero;
3309 u_data.d = value;
3310 u_minus_zero.d = -0.0;
3311 for (i = 0; i < sizeof (double); i++)
3312 if (u_data.c[i] & u_minus_zero.c[i])
3313 {
3314 value = -value;
3315 break;
3316 }
3317 }
3318 /* Now VALUE is a positive NaN. */
3319 }
3320 else
3321 cp = ecp;
3322 }
3323
3324 float_syntax = ((state & (DOT_CHAR|TRAIL_INT)) == (DOT_CHAR|TRAIL_INT)
3325 || state == (LEAD_INT|E_EXP));
3326 }
3327
3328 /* Return nil if the number uses invalid syntax. If IGNORE_TRAILING, accept
3329 any prefix that matches. Otherwise, the entire string must match. */
3330 if (! (ignore_trailing
3331 ? ((state & LEAD_INT) != 0 || float_syntax)
3332 : (!*cp && ((state & ~DOT_CHAR) == LEAD_INT || float_syntax))))
3333 return Qnil;
3334
3335 /* If the number uses integer and not float syntax, and is in C-language
3336 range, use its value, preferably as a fixnum. */
3337 if (0 <= leading_digit && ! float_syntax)
3338 {
3339 uintmax_t n;
3340
3341 /* Fast special case for single-digit integers. This also avoids a
3342 glitch when BASE is 16 and IGNORE_TRAILING is nonzero, because in that
3343 case some versions of strtoumax accept numbers like "0x1" that Emacs
3344 does not allow. */
3345 if (digit_to_number (string[signedp + 1], base) < 0)
3346 return make_number (negative ? -leading_digit : leading_digit);
3347
3348 errno = 0;
3349 n = strtoumax (string + signedp, NULL, base);
3350 if (errno == ERANGE)
3351 {
3352 /* Unfortunately there's no simple and accurate way to convert
3353 non-base-10 numbers that are out of C-language range. */
3354 if (base != 10)
3355 xsignal1 (Qoverflow_error, build_string (string));
3356 }
3357 else if (n <= (negative ? -MOST_NEGATIVE_FIXNUM : MOST_POSITIVE_FIXNUM))
3358 {
3359 EMACS_INT signed_n = n;
3360 return make_number (negative ? -signed_n : signed_n);
3361 }
3362 else
3363 value = n;
3364 }
3365
3366 /* Either the number uses float syntax, or it does not fit into a fixnum.
3367 Convert it from string to floating point, unless the value is already
3368 known because it is an infinity, a NAN, or its absolute value fits in
3369 uintmax_t. */
3370 if (! value)
3371 value = atof (string + signedp);
3372
3373 return make_float (negative ? -value : value);
3374}
3375
3376\f
3377static Lisp_Object
3378read_vector (Lisp_Object readcharfun, int bytecodeflag)
3379{
3380 ptrdiff_t i, size;
3381 register Lisp_Object *ptr;
3382 register Lisp_Object tem, item, vector;
3383 register struct Lisp_Cons *otem;
3384 Lisp_Object len;
3385
3386 tem = read_list (1, readcharfun);
3387 len = Flength (tem);
3388 vector = Fmake_vector (len, Qnil);
3389
3390 size = ASIZE (vector);
3391 ptr = XVECTOR (vector)->contents;
3392 for (i = 0; i < size; i++)
3393 {
3394 item = Fcar (tem);
3395 /* If `load-force-doc-strings' is t when reading a lazily-loaded
3396 bytecode object, the docstring containing the bytecode and
3397 constants values must be treated as unibyte and passed to
3398 Fread, to get the actual bytecode string and constants vector. */
3399 if (bytecodeflag && load_force_doc_strings)
3400 {
3401 if (i == COMPILED_BYTECODE)
3402 {
3403 if (!STRINGP (item))
3404 error ("Invalid byte code");
3405
3406 /* Delay handling the bytecode slot until we know whether
3407 it is lazily-loaded (we can tell by whether the
3408 constants slot is nil). */
3409 ptr[COMPILED_CONSTANTS] = item;
3410 item = Qnil;
3411 }
3412 else if (i == COMPILED_CONSTANTS)
3413 {
3414 Lisp_Object bytestr = ptr[COMPILED_CONSTANTS];
3415
3416 if (NILP (item))
3417 {
3418 /* Coerce string to unibyte (like string-as-unibyte,
3419 but without generating extra garbage and
3420 guaranteeing no change in the contents). */
3421 STRING_SET_CHARS (bytestr, SBYTES (bytestr));
3422 STRING_SET_UNIBYTE (bytestr);
3423
3424 item = Fread (Fcons (bytestr, readcharfun));
3425 if (!CONSP (item))
3426 error ("Invalid byte code");
3427
3428 otem = XCONS (item);
3429 bytestr = XCAR (item);
3430 item = XCDR (item);
3431 free_cons (otem);
3432 }
3433
3434 /* Now handle the bytecode slot. */
3435 ptr[COMPILED_BYTECODE] = bytestr;
3436 }
3437 else if (i == COMPILED_DOC_STRING
3438 && STRINGP (item)
3439 && ! STRING_MULTIBYTE (item))
3440 {
3441 if (EQ (readcharfun, Qget_emacs_mule_file_char))
3442 item = Fdecode_coding_string (item, Qemacs_mule, Qnil, Qnil);
3443 else
3444 item = Fstring_as_multibyte (item);
3445 }
3446 }
3447 ptr[i] = item;
3448 otem = XCONS (tem);
3449 tem = Fcdr (tem);
3450 free_cons (otem);
3451 }
3452 return vector;
3453}
3454
3455/* FLAG = 1 means check for ] to terminate rather than ) and . */
3456
3457static Lisp_Object
3458read_list (int flag, register Lisp_Object readcharfun)
3459{
3460 Lisp_Object val, tail;
3461 register Lisp_Object elt, tem;
3462 struct gcpro gcpro1, gcpro2;
3463 /* 0 is the normal case.
3464 1 means this list is a doc reference; replace it with the number 0.
3465 2 means this list is a doc reference; replace it with the doc string. */
3466 int doc_reference = 0;
3467
3468 /* Initialize this to 1 if we are reading a list. */
3469 int first_in_list = flag <= 0;
3470
3471 val = Qnil;
3472 tail = Qnil;
3473
3474 while (1)
3475 {
3476 int ch;
3477 GCPRO2 (val, tail);
3478 elt = read1 (readcharfun, &ch, first_in_list);
3479 UNGCPRO;
3480
3481 first_in_list = 0;
3482
3483 /* While building, if the list starts with #$, treat it specially. */
3484 if (EQ (elt, Vload_file_name)
3485 && ! NILP (elt)
3486 && !NILP (Vpurify_flag))
3487 {
3488 if (NILP (Vdoc_file_name))
3489 /* We have not yet called Snarf-documentation, so assume
3490 this file is described in the DOC-MM.NN file
3491 and Snarf-documentation will fill in the right value later.
3492 For now, replace the whole list with 0. */
3493 doc_reference = 1;
3494 else
3495 /* We have already called Snarf-documentation, so make a relative
3496 file name for this file, so it can be found properly
3497 in the installed Lisp directory.
3498 We don't use Fexpand_file_name because that would make
3499 the directory absolute now. */
3500 elt = concat2 (build_string ("../lisp/"),
3501 Ffile_name_nondirectory (elt));
3502 }
3503 else if (EQ (elt, Vload_file_name)
3504 && ! NILP (elt)
3505 && load_force_doc_strings)
3506 doc_reference = 2;
3507
3508 if (ch)
3509 {
3510 if (flag > 0)
3511 {
3512 if (ch == ']')
3513 return val;
3514 invalid_syntax (") or . in a vector");
3515 }
3516 if (ch == ')')
3517 return val;
3518 if (ch == '.')
3519 {
3520 GCPRO2 (val, tail);
3521 if (!NILP (tail))
3522 XSETCDR (tail, read0 (readcharfun));
3523 else
3524 val = read0 (readcharfun);
3525 read1 (readcharfun, &ch, 0);
3526 UNGCPRO;
3527 if (ch == ')')
3528 {
3529 if (doc_reference == 1)
3530 return make_number (0);
3531 if (doc_reference == 2)
3532 {
3533 /* Get a doc string from the file we are loading.
3534 If it's in saved_doc_string, get it from there.
3535
3536 Here, we don't know if the string is a
3537 bytecode string or a doc string. As a
3538 bytecode string must be unibyte, we always
3539 return a unibyte string. If it is actually a
3540 doc string, caller must make it
3541 multibyte. */
3542
3543 EMACS_INT pos = XINT (XCDR (val));
3544 /* Position is negative for user variables. */
3545 if (pos < 0) pos = -pos;
3546 if (pos >= saved_doc_string_position
3547 && pos < (saved_doc_string_position
3548 + saved_doc_string_length))
3549 {
3550 ptrdiff_t start = pos - saved_doc_string_position;
3551 ptrdiff_t from, to;
3552
3553 /* Process quoting with ^A,
3554 and find the end of the string,
3555 which is marked with ^_ (037). */
3556 for (from = start, to = start;
3557 saved_doc_string[from] != 037;)
3558 {
3559 int c = saved_doc_string[from++];
3560 if (c == 1)
3561 {
3562 c = saved_doc_string[from++];
3563 if (c == 1)
3564 saved_doc_string[to++] = c;
3565 else if (c == '0')
3566 saved_doc_string[to++] = 0;
3567 else if (c == '_')
3568 saved_doc_string[to++] = 037;
3569 }
3570 else
3571 saved_doc_string[to++] = c;
3572 }
3573
3574 return make_unibyte_string (saved_doc_string + start,
3575 to - start);
3576 }
3577 /* Look in prev_saved_doc_string the same way. */
3578 else if (pos >= prev_saved_doc_string_position
3579 && pos < (prev_saved_doc_string_position
3580 + prev_saved_doc_string_length))
3581 {
3582 ptrdiff_t start =
3583 pos - prev_saved_doc_string_position;
3584 ptrdiff_t from, to;
3585
3586 /* Process quoting with ^A,
3587 and find the end of the string,
3588 which is marked with ^_ (037). */
3589 for (from = start, to = start;
3590 prev_saved_doc_string[from] != 037;)
3591 {
3592 int c = prev_saved_doc_string[from++];
3593 if (c == 1)
3594 {
3595 c = prev_saved_doc_string[from++];
3596 if (c == 1)
3597 prev_saved_doc_string[to++] = c;
3598 else if (c == '0')
3599 prev_saved_doc_string[to++] = 0;
3600 else if (c == '_')
3601 prev_saved_doc_string[to++] = 037;
3602 }
3603 else
3604 prev_saved_doc_string[to++] = c;
3605 }
3606
3607 return make_unibyte_string (prev_saved_doc_string
3608 + start,
3609 to - start);
3610 }
3611 else
3612 return get_doc_string (val, 1, 0);
3613 }
3614
3615 return val;
3616 }
3617 invalid_syntax (". in wrong context");
3618 }
3619 invalid_syntax ("] in a list");
3620 }
3621 tem = Fcons (elt, Qnil);
3622 if (!NILP (tail))
3623 XSETCDR (tail, tem);
3624 else
3625 val = tem;
3626 tail = tem;
3627 }
3628}
3629\f
3630static Lisp_Object initial_obarray;
3631
3632/* `oblookup' stores the bucket number here, for the sake of Funintern. */
3633
3634static size_t oblookup_last_bucket_number;
3635
3636/* Get an error if OBARRAY is not an obarray.
3637 If it is one, return it. */
3638
3639Lisp_Object
3640check_obarray (Lisp_Object obarray)
3641{
3642 if (!VECTORP (obarray) || ASIZE (obarray) == 0)
3643 {
3644 /* If Vobarray is now invalid, force it to be valid. */
3645 if (EQ (Vobarray, obarray)) Vobarray = initial_obarray;
3646 wrong_type_argument (Qvectorp, obarray);
3647 }
3648 return obarray;
3649}
3650
3651/* Intern the C string STR: return a symbol with that name,
3652 interned in the current obarray. */
3653
3654Lisp_Object
3655intern_1 (const char *str, ptrdiff_t len)
3656{
3657 Lisp_Object obarray = check_obarray (Vobarray);
3658 Lisp_Object tem = oblookup (obarray, str, len, len);
3659
3660 return SYMBOLP (tem) ? tem : Fintern (make_string (str, len), obarray);
3661}
3662
3663Lisp_Object
3664intern_c_string_1 (const char *str, ptrdiff_t len)
3665{
3666 Lisp_Object obarray = check_obarray (Vobarray);
3667 Lisp_Object tem = oblookup (obarray, str, len, len);
3668
3669 if (SYMBOLP (tem))
3670 return tem;
3671
3672 if (NILP (Vpurify_flag))
3673 /* Creating a non-pure string from a string literal not
3674 implemented yet. We could just use make_string here and live
3675 with the extra copy. */
3676 abort ();
3677
3678 return Fintern (make_pure_c_string (str, len), obarray);
3679}
3680\f
3681DEFUN ("intern", Fintern, Sintern, 1, 2, 0,
3682 doc: /* Return the canonical symbol whose name is STRING.
3683If there is none, one is created by this function and returned.
3684A second optional argument specifies the obarray to use;
3685it defaults to the value of `obarray'. */)
3686 (Lisp_Object string, Lisp_Object obarray)
3687{
3688 register Lisp_Object tem, sym, *ptr;
3689
3690 if (NILP (obarray)) obarray = Vobarray;
3691 obarray = check_obarray (obarray);
3692
3693 CHECK_STRING (string);
3694
3695 tem = oblookup (obarray, SSDATA (string),
3696 SCHARS (string),
3697 SBYTES (string));
3698 if (!INTEGERP (tem))
3699 return tem;
3700
3701 if (!NILP (Vpurify_flag))
3702 string = Fpurecopy (string);
3703 sym = Fmake_symbol (string);
3704
3705 if (EQ (obarray, initial_obarray))
3706 XSYMBOL (sym)->interned = SYMBOL_INTERNED_IN_INITIAL_OBARRAY;
3707 else
3708 XSYMBOL (sym)->interned = SYMBOL_INTERNED;
3709
3710 if ((SREF (string, 0) == ':')
3711 && EQ (obarray, initial_obarray))
3712 {
3713 XSYMBOL (sym)->constant = 1;
3714 XSYMBOL (sym)->redirect = SYMBOL_PLAINVAL;
3715 SET_SYMBOL_VAL (XSYMBOL (sym), sym);
3716 }
3717
3718 ptr = &AREF (obarray, XINT(tem));
3719 if (SYMBOLP (*ptr))
3720 XSYMBOL (sym)->next = XSYMBOL (*ptr);
3721 else
3722 XSYMBOL (sym)->next = 0;
3723 *ptr = sym;
3724 return sym;
3725}
3726
3727DEFUN ("intern-soft", Fintern_soft, Sintern_soft, 1, 2, 0,
3728 doc: /* Return the canonical symbol named NAME, or nil if none exists.
3729NAME may be a string or a symbol. If it is a symbol, that exact
3730symbol is searched for.
3731A second optional argument specifies the obarray to use;
3732it defaults to the value of `obarray'. */)
3733 (Lisp_Object name, Lisp_Object obarray)
3734{
3735 register Lisp_Object tem, string;
3736
3737 if (NILP (obarray)) obarray = Vobarray;
3738 obarray = check_obarray (obarray);
3739
3740 if (!SYMBOLP (name))
3741 {
3742 CHECK_STRING (name);
3743 string = name;
3744 }
3745 else
3746 string = SYMBOL_NAME (name);
3747
3748 tem = oblookup (obarray, SSDATA (string), SCHARS (string), SBYTES (string));
3749 if (INTEGERP (tem) || (SYMBOLP (name) && !EQ (name, tem)))
3750 return Qnil;
3751 else
3752 return tem;
3753}
3754\f
3755DEFUN ("unintern", Funintern, Sunintern, 1, 2, 0,
3756 doc: /* Delete the symbol named NAME, if any, from OBARRAY.
3757The value is t if a symbol was found and deleted, nil otherwise.
3758NAME may be a string or a symbol. If it is a symbol, that symbol
3759is deleted, if it belongs to OBARRAY--no other symbol is deleted.
3760OBARRAY defaults to the value of the variable `obarray'. */)
3761 (Lisp_Object name, Lisp_Object obarray)
3762{
3763 register Lisp_Object string, tem;
3764 size_t hash;
3765
3766 if (NILP (obarray)) obarray = Vobarray;
3767 obarray = check_obarray (obarray);
3768
3769 if (SYMBOLP (name))
3770 string = SYMBOL_NAME (name);
3771 else
3772 {
3773 CHECK_STRING (name);
3774 string = name;
3775 }
3776
3777 tem = oblookup (obarray, SSDATA (string),
3778 SCHARS (string),
3779 SBYTES (string));
3780 if (INTEGERP (tem))
3781 return Qnil;
3782 /* If arg was a symbol, don't delete anything but that symbol itself. */
3783 if (SYMBOLP (name) && !EQ (name, tem))
3784 return Qnil;
3785
3786 /* There are plenty of other symbols which will screw up the Emacs
3787 session if we unintern them, as well as even more ways to use
3788 `setq' or `fset' or whatnot to make the Emacs session
3789 unusable. Let's not go down this silly road. --Stef */
3790 /* if (EQ (tem, Qnil) || EQ (tem, Qt))
3791 error ("Attempt to unintern t or nil"); */
3792
3793 XSYMBOL (tem)->interned = SYMBOL_UNINTERNED;
3794
3795 hash = oblookup_last_bucket_number;
3796
3797 if (EQ (AREF (obarray, hash), tem))
3798 {
3799 if (XSYMBOL (tem)->next)
3800 XSETSYMBOL (AREF (obarray, hash), XSYMBOL (tem)->next);
3801 else
3802 XSETINT (AREF (obarray, hash), 0);
3803 }
3804 else
3805 {
3806 Lisp_Object tail, following;
3807
3808 for (tail = AREF (obarray, hash);
3809 XSYMBOL (tail)->next;
3810 tail = following)
3811 {
3812 XSETSYMBOL (following, XSYMBOL (tail)->next);
3813 if (EQ (following, tem))
3814 {
3815 XSYMBOL (tail)->next = XSYMBOL (following)->next;
3816 break;
3817 }
3818 }
3819 }
3820
3821 return Qt;
3822}
3823\f
3824/* Return the symbol in OBARRAY whose names matches the string
3825 of SIZE characters (SIZE_BYTE bytes) at PTR.
3826 If there is no such symbol in OBARRAY, return nil.
3827
3828 Also store the bucket number in oblookup_last_bucket_number. */
3829
3830Lisp_Object
3831oblookup (Lisp_Object obarray, register const char *ptr, ptrdiff_t size, ptrdiff_t size_byte)
3832{
3833 size_t hash;
3834 size_t obsize;
3835 register Lisp_Object tail;
3836 Lisp_Object bucket, tem;
3837
3838 obarray = check_obarray (obarray);
3839 obsize = ASIZE (obarray);
3840
3841 /* This is sometimes needed in the middle of GC. */
3842 obsize &= ~ARRAY_MARK_FLAG;
3843 hash = hash_string (ptr, size_byte) % obsize;
3844 bucket = AREF (obarray, hash);
3845 oblookup_last_bucket_number = hash;
3846 if (EQ (bucket, make_number (0)))
3847 ;
3848 else if (!SYMBOLP (bucket))
3849 error ("Bad data in guts of obarray"); /* Like CADR error message. */
3850 else
3851 for (tail = bucket; ; XSETSYMBOL (tail, XSYMBOL (tail)->next))
3852 {
3853 if (SBYTES (SYMBOL_NAME (tail)) == size_byte
3854 && SCHARS (SYMBOL_NAME (tail)) == size
3855 && !memcmp (SDATA (SYMBOL_NAME (tail)), ptr, size_byte))
3856 return tail;
3857 else if (XSYMBOL (tail)->next == 0)
3858 break;
3859 }
3860 XSETINT (tem, hash);
3861 return tem;
3862}
3863\f
3864void
3865map_obarray (Lisp_Object obarray, void (*fn) (Lisp_Object, Lisp_Object), Lisp_Object arg)
3866{
3867 ptrdiff_t i;
3868 register Lisp_Object tail;
3869 CHECK_VECTOR (obarray);
3870 for (i = ASIZE (obarray) - 1; i >= 0; i--)
3871 {
3872 tail = AREF (obarray, i);
3873 if (SYMBOLP (tail))
3874 while (1)
3875 {
3876 (*fn) (tail, arg);
3877 if (XSYMBOL (tail)->next == 0)
3878 break;
3879 XSETSYMBOL (tail, XSYMBOL (tail)->next);
3880 }
3881 }
3882}
3883
3884static void
3885mapatoms_1 (Lisp_Object sym, Lisp_Object function)
3886{
3887 call1 (function, sym);
3888}
3889
3890DEFUN ("mapatoms", Fmapatoms, Smapatoms, 1, 2, 0,
3891 doc: /* Call FUNCTION on every symbol in OBARRAY.
3892OBARRAY defaults to the value of `obarray'. */)
3893 (Lisp_Object function, Lisp_Object obarray)
3894{
3895 if (NILP (obarray)) obarray = Vobarray;
3896 obarray = check_obarray (obarray);
3897
3898 map_obarray (obarray, mapatoms_1, function);
3899 return Qnil;
3900}
3901
3902#define OBARRAY_SIZE 1511
3903
3904void
3905init_obarray (void)
3906{
3907 Lisp_Object oblength;
3908 ptrdiff_t size = 100 + MAX_MULTIBYTE_LENGTH;
3909
3910 XSETFASTINT (oblength, OBARRAY_SIZE);
3911
3912 Vobarray = Fmake_vector (oblength, make_number (0));
3913 initial_obarray = Vobarray;
3914 staticpro (&initial_obarray);
3915
3916 Qunbound = Fmake_symbol (build_pure_c_string ("unbound"));
3917 /* Set temporary dummy values to Qnil and Vpurify_flag to satisfy the
3918 NILP (Vpurify_flag) check in intern_c_string. */
3919 Qnil = make_number (-1); Vpurify_flag = make_number (1);
3920 Qnil = intern_c_string ("nil");
3921
3922 /* Fmake_symbol inits fields of new symbols with Qunbound and Qnil,
3923 so those two need to be fixed manually. */
3924 SET_SYMBOL_VAL (XSYMBOL (Qunbound), Qunbound);
3925 SVAR (XSYMBOL (Qunbound), function) = Qunbound;
3926 SVAR (XSYMBOL (Qunbound), plist) = Qnil;
3927 /* XSYMBOL (Qnil)->function = Qunbound; */
3928 SET_SYMBOL_VAL (XSYMBOL (Qnil), Qnil);
3929 XSYMBOL (Qnil)->constant = 1;
3930 XSYMBOL (Qnil)->declared_special = 1;
3931 SVAR (XSYMBOL (Qnil), plist) = Qnil;
3932
3933 Qt = intern_c_string ("t");
3934 SET_SYMBOL_VAL (XSYMBOL (Qt), Qt);
3935 XSYMBOL (Qnil)->declared_special = 1;
3936 XSYMBOL (Qt)->constant = 1;
3937
3938 /* Qt is correct even if CANNOT_DUMP. loadup.el will set to nil at end. */
3939 Vpurify_flag = Qt;
3940
3941 DEFSYM (Qvariable_documentation, "variable-documentation");
3942
3943 read_buffer = xmalloc (size);
3944 read_buffer_size = size;
3945}
3946\f
3947void
3948defsubr (struct Lisp_Subr *sname)
3949{
3950 Lisp_Object sym;
3951 sym = intern_c_string (sname->symbol_name);
3952 XSETTYPED_PVECTYPE (sname, size, PVEC_SUBR);
3953 XSETSUBR (SVAR (XSYMBOL (sym), function), sname);
3954}
3955
3956#ifdef NOTDEF /* Use fset in subr.el now! */
3957void
3958defalias (struct Lisp_Subr *sname, char *string)
3959{
3960 Lisp_Object sym;
3961 sym = intern (string);
3962 XSETSUBR (XSYMBOL (sym)->function, sname);
3963}
3964#endif /* NOTDEF */
3965
3966/* Define an "integer variable"; a symbol whose value is forwarded to a
3967 C variable of type EMACS_INT. Sample call (with "xx" to fool make-docfile):
3968 DEFxxVAR_INT ("emacs-priority", &emacs_priority, "Documentation"); */
3969void
3970defvar_int (struct Lisp_Intfwd *i_fwd,
3971 const char *namestring, EMACS_INT *address)
3972{
3973 Lisp_Object sym;
3974 sym = intern_c_string (namestring);
3975 i_fwd->type = Lisp_Fwd_Int;
3976 i_fwd->intvar = address;
3977 XSYMBOL (sym)->declared_special = 1;
3978 XSYMBOL (sym)->redirect = SYMBOL_FORWARDED;
3979 SET_SYMBOL_FWD (XSYMBOL (sym), (union Lisp_Fwd *)i_fwd);
3980}
3981
3982/* Similar but define a variable whose value is t if address contains 1,
3983 nil if address contains 0. */
3984void
3985defvar_bool (struct Lisp_Boolfwd *b_fwd,
3986 const char *namestring, int *address)
3987{
3988 Lisp_Object sym;
3989 sym = intern_c_string (namestring);
3990 b_fwd->type = Lisp_Fwd_Bool;
3991 b_fwd->boolvar = address;
3992 XSYMBOL (sym)->declared_special = 1;
3993 XSYMBOL (sym)->redirect = SYMBOL_FORWARDED;
3994 SET_SYMBOL_FWD (XSYMBOL (sym), (union Lisp_Fwd *)b_fwd);
3995 Vbyte_boolean_vars = Fcons (sym, Vbyte_boolean_vars);
3996}
3997
3998/* Similar but define a variable whose value is the Lisp Object stored
3999 at address. Two versions: with and without gc-marking of the C
4000 variable. The nopro version is used when that variable will be
4001 gc-marked for some other reason, since marking the same slot twice
4002 can cause trouble with strings. */
4003void
4004defvar_lisp_nopro (struct Lisp_Objfwd *o_fwd,
4005 const char *namestring, Lisp_Object *address)
4006{
4007 Lisp_Object sym;
4008 sym = intern_c_string (namestring);
4009 o_fwd->type = Lisp_Fwd_Obj;
4010 o_fwd->objvar = address;
4011 XSYMBOL (sym)->declared_special = 1;
4012 XSYMBOL (sym)->redirect = SYMBOL_FORWARDED;
4013 SET_SYMBOL_FWD (XSYMBOL (sym), (union Lisp_Fwd *)o_fwd);
4014}
4015
4016void
4017defvar_lisp (struct Lisp_Objfwd *o_fwd,
4018 const char *namestring, Lisp_Object *address)
4019{
4020 defvar_lisp_nopro (o_fwd, namestring, address);
4021 staticpro (address);
4022}
4023
4024/* Similar but define a variable whose value is the Lisp Object stored
4025 at a particular offset in the current kboard object. */
4026
4027void
4028defvar_kboard (struct Lisp_Kboard_Objfwd *ko_fwd,
4029 const char *namestring, int offset)
4030{
4031 Lisp_Object sym;
4032 sym = intern_c_string (namestring);
4033 ko_fwd->type = Lisp_Fwd_Kboard_Obj;
4034 ko_fwd->offset = offset;
4035 XSYMBOL (sym)->declared_special = 1;
4036 XSYMBOL (sym)->redirect = SYMBOL_FORWARDED;
4037 SET_SYMBOL_FWD (XSYMBOL (sym), (union Lisp_Fwd *)ko_fwd);
4038}
4039\f
4040/* Check that the elements of Vload_path exist. */
4041
4042static void
4043load_path_check (void)
4044{
4045 Lisp_Object path_tail;
4046
4047 /* The only elements that might not exist are those from
4048 PATH_LOADSEARCH, EMACSLOADPATH. Anything else is only added if
4049 it exists. */
4050 for (path_tail = Vload_path; !NILP (path_tail); path_tail = XCDR (path_tail))
4051 {
4052 Lisp_Object dirfile;
4053 dirfile = Fcar (path_tail);
4054 if (STRINGP (dirfile))
4055 {
4056 dirfile = Fdirectory_file_name (dirfile);
4057 if (access (SSDATA (dirfile), 0) < 0)
4058 dir_warning ("Warning: Lisp directory `%s' does not exist.\n",
4059 XCAR (path_tail));
4060 }
4061 }
4062}
4063
4064/* Record the value of load-path used at the start of dumping
4065 so we can see if the site changed it later during dumping. */
4066static Lisp_Object dump_path;
4067
4068/* Compute the default Vload_path, with the following logic:
4069 If CANNOT_DUMP:
4070 use EMACSLOADPATH env-var if set; otherwise use PATH_LOADSEARCH,
4071 prepending PATH_SITELOADSEARCH unless --no-site-lisp.
4072 The remainder is what happens when dumping works:
4073 If purify-flag (ie dumping) just use PATH_DUMPLOADSEARCH.
4074 Otherwise use EMACSLOADPATH if set, else PATH_LOADSEARCH.
4075
4076 If !initialized, then just set both Vload_path and dump_path.
4077 If initialized, then if Vload_path != dump_path, do nothing.
4078 (Presumably the load-path has already been changed by something.
4079 This can only be from a site-load file during dumping,
4080 or because EMACSLOADPATH is set.)
4081 If Vinstallation_directory is not nil (ie, running uninstalled):
4082 If installation-dir/lisp exists and not already a member,
4083 we must be running uninstalled. Reset the load-path
4084 to just installation-dir/lisp. (The default PATH_LOADSEARCH
4085 refers to the eventual installation directories. Since we
4086 are not yet installed, we should not use them, even if they exist.)
4087 If installation-dir/lisp does not exist, just add dump_path at the
4088 end instead.
4089 Add installation-dir/leim (if exists and not already a member) at the front.
4090 Add installation-dir/site-lisp (if !no_site_lisp, and exists
4091 and not already a member) at the front.
4092 If installation-dir != source-dir (ie running an uninstalled,
4093 out-of-tree build) AND install-dir/src/Makefile exists BUT
4094 install-dir/src/Makefile.in does NOT exist (this is a sanity
4095 check), then repeat the above steps for source-dir/lisp,
4096 leim and site-lisp.
4097 Finally, add the site-lisp directories at the front (if !no_site_lisp).
4098*/
4099
4100void
4101init_lread (void)
4102{
4103 const char *normal;
4104
4105#ifdef CANNOT_DUMP
4106#ifdef HAVE_NS
4107 const char *loadpath = ns_load_path ();
4108#endif
4109
4110 normal = PATH_LOADSEARCH;
4111#ifdef HAVE_NS
4112 Vload_path = decode_env_path ("EMACSLOADPATH", loadpath ? loadpath : normal);
4113#else
4114 Vload_path = decode_env_path ("EMACSLOADPATH", normal);
4115#endif
4116
4117 load_path_check ();
4118
4119 /* FIXME CANNOT_DUMP platforms should get source-dir/lisp etc added
4120 to their load-path too, AFAICS. I don't think we can tell the
4121 difference between initialized and !initialized in this case,
4122 so we'll have to do it unconditionally when Vinstallation_directory
4123 is non-nil. */
4124 if (!no_site_lisp && !egetenv ("EMACSLOADPATH"))
4125 {
4126 Lisp_Object sitelisp;
4127 sitelisp = decode_env_path (0, PATH_SITELOADSEARCH);
4128 if (! NILP (sitelisp)) Vload_path = nconc2 (sitelisp, Vload_path);
4129 }
4130#else
4131 if (NILP (Vpurify_flag))
4132 {
4133 normal = PATH_LOADSEARCH;
4134 /* If the EMACSLOADPATH environment variable is set, use its value.
4135 This doesn't apply if we're dumping. */
4136 if (egetenv ("EMACSLOADPATH"))
4137 Vload_path = decode_env_path ("EMACSLOADPATH", normal);
4138 }
4139 else
4140 normal = PATH_DUMPLOADSEARCH;
4141
4142 /* In a dumped Emacs, we normally reset the value of Vload_path using
4143 PATH_LOADSEARCH, since the value that was dumped uses lisp/ in
4144 the source directory, instead of the path of the installed elisp
4145 libraries. However, if it appears that Vload_path has already been
4146 changed from the default that was saved before dumping, don't
4147 change it further. Changes can only be due to EMACSLOADPATH, or
4148 site-lisp files that were processed during dumping. */
4149 if (initialized)
4150 {
4151 if (NILP (Fequal (dump_path, Vload_path)))
4152 {
4153 /* Do not make any changes, just check the elements exist. */
4154 /* Note: --no-site-lisp is ignored.
4155 I don't know what to do about this. */
4156 load_path_check ();
4157 }
4158 else
4159 {
4160#ifdef HAVE_NS
4161 const char *loadpath = ns_load_path ();
4162 Vload_path = decode_env_path (0, loadpath ? loadpath : normal);
4163#else
4164 Vload_path = decode_env_path (0, normal);
4165#endif
4166 if (!NILP (Vinstallation_directory))
4167 {
4168 Lisp_Object tem, tem1;
4169
4170 /* Add to the path the lisp subdir of the installation
4171 dir, if it exists. Note: in out-of-tree builds,
4172 this directory is empty save for Makefile. */
4173 tem = Fexpand_file_name (build_string ("lisp"),
4174 Vinstallation_directory);
4175 tem1 = Ffile_exists_p (tem);
4176 if (!NILP (tem1))
4177 {
4178 if (NILP (Fmember (tem, Vload_path)))
4179 {
4180 /* We are running uninstalled. The default load-path
4181 points to the eventual installed lisp, leim
4182 directories. We should not use those now, even
4183 if they exist, so start over from a clean slate. */
4184 Vload_path = Fcons (tem, Qnil);
4185 }
4186 }
4187 else
4188 /* That dir doesn't exist, so add the build-time
4189 Lisp dirs instead. */
4190 Vload_path = nconc2 (Vload_path, dump_path);
4191
4192 /* Add leim under the installation dir, if it exists. */
4193 tem = Fexpand_file_name (build_string ("leim"),
4194 Vinstallation_directory);
4195 tem1 = Ffile_exists_p (tem);
4196 if (!NILP (tem1))
4197 {
4198 if (NILP (Fmember (tem, Vload_path)))
4199 Vload_path = Fcons (tem, Vload_path);
4200 }
4201
4202 /* Add site-lisp under the installation dir, if it exists. */
4203 if (!no_site_lisp)
4204 {
4205 tem = Fexpand_file_name (build_string ("site-lisp"),
4206 Vinstallation_directory);
4207 tem1 = Ffile_exists_p (tem);
4208 if (!NILP (tem1))
4209 {
4210 if (NILP (Fmember (tem, Vload_path)))
4211 Vload_path = Fcons (tem, Vload_path);
4212 }
4213 }
4214
4215 /* If Emacs was not built in the source directory,
4216 and it is run from where it was built, add to load-path
4217 the lisp, leim and site-lisp dirs under that directory. */
4218
4219 if (NILP (Fequal (Vinstallation_directory, Vsource_directory)))
4220 {
4221 Lisp_Object tem2;
4222
4223 tem = Fexpand_file_name (build_string ("src/Makefile"),
4224 Vinstallation_directory);
4225 tem1 = Ffile_exists_p (tem);
4226
4227 /* Don't be fooled if they moved the entire source tree
4228 AFTER dumping Emacs. If the build directory is indeed
4229 different from the source dir, src/Makefile.in and
4230 src/Makefile will not be found together. */
4231 tem = Fexpand_file_name (build_string ("src/Makefile.in"),
4232 Vinstallation_directory);
4233 tem2 = Ffile_exists_p (tem);
4234 if (!NILP (tem1) && NILP (tem2))
4235 {
4236 tem = Fexpand_file_name (build_string ("lisp"),
4237 Vsource_directory);
4238
4239 if (NILP (Fmember (tem, Vload_path)))
4240 Vload_path = Fcons (tem, Vload_path);
4241
4242 tem = Fexpand_file_name (build_string ("leim"),
4243 Vsource_directory);
4244
4245 if (NILP (Fmember (tem, Vload_path)))
4246 Vload_path = Fcons (tem, Vload_path);
4247
4248 if (!no_site_lisp)
4249 {
4250 tem = Fexpand_file_name (build_string ("site-lisp"),
4251 Vsource_directory);
4252
4253 if (NILP (Fmember (tem, Vload_path)))
4254 Vload_path = Fcons (tem, Vload_path);
4255 }
4256 }
4257 } /* Vinstallation_directory != Vsource_directory */
4258
4259 } /* if Vinstallation_directory */
4260
4261 /* Check before adding the site-lisp directories.
4262 The install should have created them, but they are not
4263 required, so no need to warn if they are absent.
4264 Or we might be running before installation. */
4265 load_path_check ();
4266
4267 /* Add the site-lisp directories at the front. */
4268 if (!no_site_lisp)
4269 {
4270 Lisp_Object sitelisp;
4271 sitelisp = decode_env_path (0, PATH_SITELOADSEARCH);
4272 if (! NILP (sitelisp)) Vload_path = nconc2 (sitelisp, Vload_path);
4273 }
4274 } /* if dump_path == Vload_path */
4275 }
4276 else /* !initialized */
4277 {
4278 /* NORMAL refers to PATH_DUMPLOADSEARCH, ie the lisp dir in the
4279 source directory. We used to add ../lisp (ie the lisp dir in
4280 the build directory) at the front here, but that caused trouble
4281 because it was copied from dump_path into Vload_path, above,
4282 when Vinstallation_directory was non-nil. It should not be
4283 necessary, since in out of tree builds lisp/ is empty, save
4284 for Makefile. */
4285 Vload_path = decode_env_path (0, normal);
4286 dump_path = Vload_path;
4287 /* No point calling load_path_check; load-path only contains essential
4288 elements from the source directory at this point. They cannot
4289 be missing unless something went extremely (and improbably)
4290 wrong, in which case the build will fail in obvious ways. */
4291 }
4292#endif /* CANNOT_DUMP */
4293
4294 Vvalues = Qnil;
4295
4296 load_in_progress = 0;
4297 Vload_file_name = Qnil;
4298
4299 load_descriptor_list = Qnil;
4300
4301 Vstandard_input = Qt;
4302 Vloads_in_progress = Qnil;
4303}
4304
4305/* Print a warning, using format string FORMAT, that directory DIRNAME
4306 does not exist. Print it on stderr and put it in *Messages*. */
4307
4308void
4309dir_warning (const char *format, Lisp_Object dirname)
4310{
4311 fprintf (stderr, format, SDATA (dirname));
4312
4313 /* Don't log the warning before we've initialized!! */
4314 if (initialized)
4315 {
4316 char *buffer;
4317 ptrdiff_t message_len;
4318 USE_SAFE_ALLOCA;
4319 SAFE_ALLOCA (buffer, char *,
4320 SBYTES (dirname) + strlen (format) - (sizeof "%s" - 1) + 1);
4321 message_len = esprintf (buffer, format, SDATA (dirname));
4322 message_dolog (buffer, message_len, 0, STRING_MULTIBYTE (dirname));
4323 SAFE_FREE ();
4324 }
4325}
4326
4327void
4328syms_of_lread (void)
4329{
4330 defsubr (&Sread);
4331 defsubr (&Sread_from_string);
4332 defsubr (&Sintern);
4333 defsubr (&Sintern_soft);
4334 defsubr (&Sunintern);
4335 defsubr (&Sget_load_suffixes);
4336 defsubr (&Sload);
4337 defsubr (&Seval_buffer);
4338 defsubr (&Seval_region);
4339 defsubr (&Sread_char);
4340 defsubr (&Sread_char_exclusive);
4341 defsubr (&Sread_event);
4342 defsubr (&Sget_file_char);
4343 defsubr (&Smapatoms);
4344 defsubr (&Slocate_file_internal);
4345
4346 DEFVAR_LISP ("obarray", Vobarray,
4347 doc: /* Symbol table for use by `intern' and `read'.
4348It is a vector whose length ought to be prime for best results.
4349The vector's contents don't make sense if examined from Lisp programs;
4350to find all the symbols in an obarray, use `mapatoms'. */);
4351
4352 DEFVAR_LISP ("values", Vvalues,
4353 doc: /* List of values of all expressions which were read, evaluated and printed.
4354 Order is reverse chronological. */);
4355 XSYMBOL (intern ("values"))->declared_special = 0;
4356
4357 DEFVAR_LISP ("standard-input", Vstandard_input,
4358 doc: /* Stream for read to get input from.
4359See documentation of `read' for possible values. */);
4360 Vstandard_input = Qt;
4361
4362 DEFVAR_LISP ("read-with-symbol-positions", Vread_with_symbol_positions,
4363 doc: /* If non-nil, add position of read symbols to `read-symbol-positions-list'.
4364
4365If this variable is a buffer, then only forms read from that buffer
4366will be added to `read-symbol-positions-list'.
4367If this variable is t, then all read forms will be added.
4368The effect of all other values other than nil are not currently
4369defined, although they may be in the future.
4370
4371The positions are relative to the last call to `read' or
4372`read-from-string'. It is probably a bad idea to set this variable at
4373the toplevel; bind it instead. */);
4374 Vread_with_symbol_positions = Qnil;
4375
4376 DEFVAR_LISP ("read-symbol-positions-list", Vread_symbol_positions_list,
4377 doc: /* A list mapping read symbols to their positions.
4378This variable is modified during calls to `read' or
4379`read-from-string', but only when `read-with-symbol-positions' is
4380non-nil.
4381
4382Each element of the list looks like (SYMBOL . CHAR-POSITION), where
4383CHAR-POSITION is an integer giving the offset of that occurrence of the
4384symbol from the position where `read' or `read-from-string' started.
4385
4386Note that a symbol will appear multiple times in this list, if it was
4387read multiple times. The list is in the same order as the symbols
4388were read in. */);
4389 Vread_symbol_positions_list = Qnil;
4390
4391 DEFVAR_LISP ("read-circle", Vread_circle,
4392 doc: /* Non-nil means read recursive structures using #N= and #N# syntax. */);
4393 Vread_circle = Qt;
4394
4395 DEFVAR_LISP ("load-path", Vload_path,
4396 doc: /* List of directories to search for files to load.
4397Each element is a string (directory name) or nil (try default directory).
4398Initialized based on EMACSLOADPATH environment variable, if any,
4399otherwise to default specified by file `epaths.h' when Emacs was built. */);
4400
4401 DEFVAR_LISP ("load-suffixes", Vload_suffixes,
4402 doc: /* List of suffixes for (compiled or source) Emacs Lisp files.
4403This list should not include the empty string.
4404`load' and related functions try to append these suffixes, in order,
4405to the specified file name if a Lisp suffix is allowed or required. */);
4406 Vload_suffixes = Fcons (build_pure_c_string (".elc"),
4407 Fcons (build_pure_c_string (".el"), Qnil));
4408 DEFVAR_LISP ("load-file-rep-suffixes", Vload_file_rep_suffixes,
4409 doc: /* List of suffixes that indicate representations of \
4410the same file.
4411This list should normally start with the empty string.
4412
4413Enabling Auto Compression mode appends the suffixes in
4414`jka-compr-load-suffixes' to this list and disabling Auto Compression
4415mode removes them again. `load' and related functions use this list to
4416determine whether they should look for compressed versions of a file
4417and, if so, which suffixes they should try to append to the file name
4418in order to do so. However, if you want to customize which suffixes
4419the loading functions recognize as compression suffixes, you should
4420customize `jka-compr-load-suffixes' rather than the present variable. */);
4421 Vload_file_rep_suffixes = Fcons (empty_unibyte_string, Qnil);
4422
4423 DEFVAR_BOOL ("load-in-progress", load_in_progress,
4424 doc: /* Non-nil if inside of `load'. */);
4425 DEFSYM (Qload_in_progress, "load-in-progress");
4426
4427 DEFVAR_LISP ("after-load-alist", Vafter_load_alist,
4428 doc: /* An alist of expressions to be evalled when particular files are loaded.
4429Each element looks like (REGEXP-OR-FEATURE FORMS...).
4430
4431REGEXP-OR-FEATURE is either a regular expression to match file names, or
4432a symbol \(a feature name).
4433
4434When `load' is run and the file-name argument matches an element's
4435REGEXP-OR-FEATURE, or when `provide' is run and provides the symbol
4436REGEXP-OR-FEATURE, the FORMS in the element are executed.
4437
4438An error in FORMS does not undo the load, but does prevent execution of
4439the rest of the FORMS. */);
4440 Vafter_load_alist = Qnil;
4441
4442 DEFVAR_LISP ("load-history", Vload_history,
4443 doc: /* Alist mapping loaded file names to symbols and features.
4444Each alist element should be a list (FILE-NAME ENTRIES...), where
4445FILE-NAME is the name of a file that has been loaded into Emacs.
4446The file name is absolute and true (i.e. it doesn't contain symlinks).
4447As an exception, one of the alist elements may have FILE-NAME nil,
4448for symbols and features not associated with any file.
4449
4450The remaining ENTRIES in the alist element describe the functions and
4451variables defined in that file, the features provided, and the
4452features required. Each entry has the form `(provide . FEATURE)',
4453`(require . FEATURE)', `(defun . FUNCTION)', `(autoload . SYMBOL)',
4454`(defface . SYMBOL)', or `(t . SYMBOL)'. Entries like `(t . SYMBOL)'
4455may precede a `(defun . FUNCTION)' entry, and means that SYMBOL was an
4456autoload before this file redefined it as a function. In addition,
4457entries may also be single symbols, which means that SYMBOL was
4458defined by `defvar' or `defconst'.
4459
4460During preloading, the file name recorded is relative to the main Lisp
4461directory. These file names are converted to absolute at startup. */);
4462 Vload_history = Qnil;
4463
4464 DEFVAR_LISP ("load-file-name", Vload_file_name,
4465 doc: /* Full name of file being loaded by `load'. */);
4466 Vload_file_name = Qnil;
4467
4468 DEFVAR_LISP ("user-init-file", Vuser_init_file,
4469 doc: /* File name, including directory, of user's initialization file.
4470If the file loaded had extension `.elc', and the corresponding source file
4471exists, this variable contains the name of source file, suitable for use
4472by functions like `custom-save-all' which edit the init file.
4473While Emacs loads and evaluates the init file, value is the real name
4474of the file, regardless of whether or not it has the `.elc' extension. */);
4475 Vuser_init_file = Qnil;
4476
4477 DEFVAR_LISP ("current-load-list", Vcurrent_load_list,
4478 doc: /* Used for internal purposes by `load'. */);
4479 Vcurrent_load_list = Qnil;
4480
4481 DEFVAR_LISP ("load-read-function", Vload_read_function,
4482 doc: /* Function used by `load' and `eval-region' for reading expressions.
4483The default is nil, which means use the function `read'. */);
4484 Vload_read_function = Qnil;
4485
4486 DEFVAR_LISP ("load-source-file-function", Vload_source_file_function,
4487 doc: /* Function called in `load' for loading an Emacs Lisp source file.
4488This function is for doing code conversion before reading the source file.
4489If nil, loading is done without any code conversion.
4490Arguments are FULLNAME, FILE, NOERROR, NOMESSAGE, where
4491 FULLNAME is the full name of FILE.
4492See `load' for the meaning of the remaining arguments. */);
4493 Vload_source_file_function = Qnil;
4494
4495 DEFVAR_BOOL ("load-force-doc-strings", load_force_doc_strings,
4496 doc: /* Non-nil means `load' should force-load all dynamic doc strings.
4497This is useful when the file being loaded is a temporary copy. */);
4498 load_force_doc_strings = 0;
4499
4500 DEFVAR_BOOL ("load-convert-to-unibyte", load_convert_to_unibyte,
4501 doc: /* Non-nil means `read' converts strings to unibyte whenever possible.
4502This is normally bound by `load' and `eval-buffer' to control `read',
4503and is not meant for users to change. */);
4504 load_convert_to_unibyte = 0;
4505
4506 DEFVAR_LISP ("source-directory", Vsource_directory,
4507 doc: /* Directory in which Emacs sources were found when Emacs was built.
4508You cannot count on them to still be there! */);
4509 Vsource_directory
4510 = Fexpand_file_name (build_string ("../"),
4511 Fcar (decode_env_path (0, PATH_DUMPLOADSEARCH)));
4512
4513 DEFVAR_LISP ("preloaded-file-list", Vpreloaded_file_list,
4514 doc: /* List of files that were preloaded (when dumping Emacs). */);
4515 Vpreloaded_file_list = Qnil;
4516
4517 DEFVAR_LISP ("byte-boolean-vars", Vbyte_boolean_vars,
4518 doc: /* List of all DEFVAR_BOOL variables, used by the byte code optimizer. */);
4519 Vbyte_boolean_vars = Qnil;
4520
4521 DEFVAR_BOOL ("load-dangerous-libraries", load_dangerous_libraries,
4522 doc: /* Non-nil means load dangerous compiled Lisp files.
4523Some versions of XEmacs use different byte codes than Emacs. These
4524incompatible byte codes can make Emacs crash when it tries to execute
4525them. */);
4526 load_dangerous_libraries = 0;
4527
4528 DEFVAR_BOOL ("force-load-messages", force_load_messages,
4529 doc: /* Non-nil means force printing messages when loading Lisp files.
4530This overrides the value of the NOMESSAGE argument to `load'. */);
4531 force_load_messages = 0;
4532
4533 DEFVAR_LISP ("bytecomp-version-regexp", Vbytecomp_version_regexp,
4534 doc: /* Regular expression matching safe to load compiled Lisp files.
4535When Emacs loads a compiled Lisp file, it reads the first 512 bytes
4536from the file, and matches them against this regular expression.
4537When the regular expression matches, the file is considered to be safe
4538to load. See also `load-dangerous-libraries'. */);
4539 Vbytecomp_version_regexp
4540 = build_pure_c_string ("^;;;.\\(in Emacs version\\|bytecomp version FSF\\)");
4541
4542 Qlexical_binding = intern ("lexical-binding");
4543 staticpro (&Qlexical_binding);
4544 DEFVAR_LISP ("lexical-binding", Vlexical_binding,
4545 doc: /* Whether to use lexical binding when evaluating code.
4546Non-nil means that the code in the current buffer should be evaluated
4547with lexical binding.
4548This variable is automatically set from the file variables of an
4549interpreted Lisp file read using `load'. Unlike other file local
4550variables, this must be set in the first line of a file. */);
4551 Fmake_variable_buffer_local (Qlexical_binding);
4552
4553 DEFVAR_LISP ("eval-buffer-list", Veval_buffer_list,
4554 doc: /* List of buffers being read from by calls to `eval-buffer' and `eval-region'. */);
4555 Veval_buffer_list = Qnil;
4556
4557 DEFVAR_LISP ("old-style-backquotes", Vold_style_backquotes,
4558 doc: /* Set to non-nil when `read' encounters an old-style backquote. */);
4559 Vold_style_backquotes = Qnil;
4560 DEFSYM (Qold_style_backquotes, "old-style-backquotes");
4561
4562 /* Vsource_directory was initialized in init_lread. */
4563
4564 load_descriptor_list = Qnil;
4565 staticpro (&load_descriptor_list);
4566
4567 DEFSYM (Qcurrent_load_list, "current-load-list");
4568 DEFSYM (Qstandard_input, "standard-input");
4569 DEFSYM (Qread_char, "read-char");
4570 DEFSYM (Qget_file_char, "get-file-char");
4571 DEFSYM (Qget_emacs_mule_file_char, "get-emacs-mule-file-char");
4572 DEFSYM (Qload_force_doc_strings, "load-force-doc-strings");
4573
4574 DEFSYM (Qbackquote, "`");
4575 DEFSYM (Qcomma, ",");
4576 DEFSYM (Qcomma_at, ",@");
4577 DEFSYM (Qcomma_dot, ",.");
4578
4579 DEFSYM (Qinhibit_file_name_operation, "inhibit-file-name-operation");
4580 DEFSYM (Qascii_character, "ascii-character");
4581 DEFSYM (Qfunction, "function");
4582 DEFSYM (Qload, "load");
4583 DEFSYM (Qload_file_name, "load-file-name");
4584 DEFSYM (Qeval_buffer_list, "eval-buffer-list");
4585 DEFSYM (Qfile_truename, "file-truename");
4586 DEFSYM (Qdir_ok, "dir-ok");
4587 DEFSYM (Qdo_after_load_evaluation, "do-after-load-evaluation");
4588
4589 staticpro (&dump_path);
4590
4591 staticpro (&read_objects);
4592 read_objects = Qnil;
4593 staticpro (&seen_list);
4594 seen_list = Qnil;
4595
4596 Vloads_in_progress = Qnil;
4597 staticpro (&Vloads_in_progress);
4598
4599 DEFSYM (Qhash_table, "hash-table");
4600 DEFSYM (Qdata, "data");
4601 DEFSYM (Qtest, "test");
4602 DEFSYM (Qsize, "size");
4603 DEFSYM (Qweakness, "weakness");
4604 DEFSYM (Qrehash_size, "rehash-size");
4605 DEFSYM (Qrehash_threshold, "rehash-threshold");
4606}