Merge remote-tracking branch 'origin/stable-2.0'
[bpt/guile.git] / libguile / ports.c
1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
3 *
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 3 of
7 * the License, or (at your option) any later version.
8 *
9 * This library is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
13 *
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
17 * 02110-1301 USA
18 */
19
20
21 \f
22 /* Headers. */
23
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
25
26 #ifdef HAVE_CONFIG_H
27 # include <config.h>
28 #endif
29
30 #include <stdio.h>
31 #include <errno.h>
32 #include <fcntl.h> /* for chsize on mingw */
33 #include <assert.h>
34 #include <iconv.h>
35 #include <uniconv.h>
36 #include <unistr.h>
37 #include <striconveh.h>
38
39 #include <assert.h>
40
41 #include "libguile/_scm.h"
42 #include "libguile/async.h"
43 #include "libguile/deprecation.h"
44 #include "libguile/eval.h"
45 #include "libguile/fports.h" /* direct access for seek and truncate */
46 #include "libguile/goops.h"
47 #include "libguile/smob.h"
48 #include "libguile/chars.h"
49 #include "libguile/dynwind.h"
50
51 #include "libguile/keywords.h"
52 #include "libguile/hashtab.h"
53 #include "libguile/root.h"
54 #include "libguile/strings.h"
55 #include "libguile/mallocs.h"
56 #include "libguile/validate.h"
57 #include "libguile/ports.h"
58 #include "libguile/vectors.h"
59 #include "libguile/weaks.h"
60 #include "libguile/fluids.h"
61 #include "libguile/eq.h"
62
63 #ifdef HAVE_STRING_H
64 #include <string.h>
65 #endif
66
67 #ifdef HAVE_IO_H
68 #include <io.h>
69 #endif
70
71 #ifdef HAVE_UNISTD_H
72 #include <unistd.h>
73 #endif
74
75 #ifdef HAVE_SYS_IOCTL_H
76 #include <sys/ioctl.h>
77 #endif
78
79 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
80 already, but have this code here in case that wasn't so in past versions,
81 or perhaps to help other minimal DOS environments.
82
83 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
84 might be possibilities if we've got other systems without ftruncate. */
85
86 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
87 #define ftruncate(fd, size) chsize (fd, size)
88 #undef HAVE_FTRUNCATE
89 #define HAVE_FTRUNCATE 1
90 #endif
91
92 \f
93 /* The port kind table --- a dynamically resized array of port types. */
94
95
96 /* scm_ptobs scm_numptob
97 * implement a dynamically resized array of ptob records.
98 * Indexes into this table are used when generating type
99 * tags for smobjects (if you know a tag you can get an index and conversely).
100 */
101 scm_t_ptob_descriptor *scm_ptobs = NULL;
102 long scm_numptob = 0;
103
104 /* GC marker for a port with stream of SCM type. */
105 SCM
106 scm_markstream (SCM ptr)
107 {
108 int openp;
109 openp = SCM_CELL_WORD_0 (ptr) & SCM_OPN;
110 if (openp)
111 return SCM_PACK (SCM_STREAM (ptr));
112 else
113 return SCM_BOOL_F;
114 }
115
116 /*
117 * We choose to use an interface similar to the smob interface with
118 * fill_input and write as standard fields, passed to the port
119 * type constructor, and optional fields set by setters.
120 */
121
122 static void
123 flush_port_default (SCM port SCM_UNUSED)
124 {
125 }
126
127 static void
128 end_input_default (SCM port SCM_UNUSED, int offset SCM_UNUSED)
129 {
130 }
131
132 scm_t_bits
133 scm_make_port_type (char *name,
134 int (*fill_input) (SCM port),
135 void (*write) (SCM port, const void *data, size_t size))
136 {
137 char *tmp;
138 if (SCM_I_MAX_PORT_TYPE_COUNT - 1 <= scm_numptob)
139 goto ptoberr;
140 SCM_CRITICAL_SECTION_START;
141 tmp = (char *) scm_gc_realloc ((char *) scm_ptobs,
142 scm_numptob * sizeof (scm_t_ptob_descriptor),
143 (1 + scm_numptob)
144 * sizeof (scm_t_ptob_descriptor),
145 "port-type");
146 if (tmp)
147 {
148 scm_ptobs = (scm_t_ptob_descriptor *) tmp;
149
150 scm_ptobs[scm_numptob].name = name;
151 scm_ptobs[scm_numptob].mark = 0;
152 scm_ptobs[scm_numptob].free = NULL;
153 scm_ptobs[scm_numptob].print = scm_port_print;
154 scm_ptobs[scm_numptob].equalp = 0;
155 scm_ptobs[scm_numptob].close = 0;
156
157 scm_ptobs[scm_numptob].write = write;
158 scm_ptobs[scm_numptob].flush = flush_port_default;
159
160 scm_ptobs[scm_numptob].end_input = end_input_default;
161 scm_ptobs[scm_numptob].fill_input = fill_input;
162 scm_ptobs[scm_numptob].input_waiting = 0;
163
164 scm_ptobs[scm_numptob].seek = 0;
165 scm_ptobs[scm_numptob].truncate = 0;
166
167 scm_numptob++;
168 }
169 SCM_CRITICAL_SECTION_END;
170 if (!tmp)
171 {
172 ptoberr:
173 scm_memory_error ("scm_make_port_type");
174 }
175 /* Make a class object if Goops is present */
176 if (SCM_UNPACK (scm_port_class[0]) != 0)
177 scm_make_port_classes (scm_numptob - 1, SCM_PTOBNAME (scm_numptob - 1));
178 return scm_tc7_port + (scm_numptob - 1) * 256;
179 }
180
181 void
182 scm_set_port_mark (scm_t_bits tc, SCM (*mark) (SCM))
183 {
184 scm_ptobs[SCM_TC2PTOBNUM (tc)].mark = mark;
185 }
186
187 void
188 scm_set_port_free (scm_t_bits tc, size_t (*free) (SCM))
189 {
190 scm_ptobs[SCM_TC2PTOBNUM (tc)].free = free;
191 }
192
193 void
194 scm_set_port_print (scm_t_bits tc, int (*print) (SCM exp, SCM port,
195 scm_print_state *pstate))
196 {
197 scm_ptobs[SCM_TC2PTOBNUM (tc)].print = print;
198 }
199
200 void
201 scm_set_port_equalp (scm_t_bits tc, SCM (*equalp) (SCM, SCM))
202 {
203 scm_ptobs[SCM_TC2PTOBNUM (tc)].equalp = equalp;
204 }
205
206 void
207 scm_set_port_flush (scm_t_bits tc, void (*flush) (SCM port))
208 {
209 scm_ptobs[SCM_TC2PTOBNUM (tc)].flush = flush;
210 }
211
212 void
213 scm_set_port_end_input (scm_t_bits tc, void (*end_input) (SCM port, int offset))
214 {
215 scm_ptobs[SCM_TC2PTOBNUM (tc)].end_input = end_input;
216 }
217
218 void
219 scm_set_port_close (scm_t_bits tc, int (*close) (SCM))
220 {
221 scm_ptobs[SCM_TC2PTOBNUM (tc)].close = close;
222 }
223
224 void
225 scm_set_port_seek (scm_t_bits tc,
226 scm_t_off (*seek) (SCM, scm_t_off, int))
227 {
228 scm_ptobs[SCM_TC2PTOBNUM (tc)].seek = seek;
229 }
230
231 void
232 scm_set_port_truncate (scm_t_bits tc, void (*truncate) (SCM, scm_t_off))
233 {
234 scm_ptobs[SCM_TC2PTOBNUM (tc)].truncate = truncate;
235 }
236
237 void
238 scm_set_port_input_waiting (scm_t_bits tc, int (*input_waiting) (SCM))
239 {
240 scm_ptobs[SCM_TC2PTOBNUM (tc)].input_waiting = input_waiting;
241 }
242
243 \f
244
245 SCM_DEFINE (scm_char_ready_p, "char-ready?", 0, 1, 0,
246 (SCM port),
247 "Return @code{#t} if a character is ready on input @var{port}\n"
248 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
249 "@code{#t} then the next @code{read-char} operation on\n"
250 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
251 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
252 "\n"
253 "@code{char-ready?} exists to make it possible for a\n"
254 "program to accept characters from interactive ports without\n"
255 "getting stuck waiting for input. Any input editors associated\n"
256 "with such ports must make sure that characters whose existence\n"
257 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
258 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
259 "a port at end of file would be indistinguishable from an\n"
260 "interactive port that has no ready characters.")
261 #define FUNC_NAME s_scm_char_ready_p
262 {
263 scm_t_port *pt;
264
265 if (SCM_UNBNDP (port))
266 port = scm_current_input_port ();
267 /* It's possible to close the current input port, so validate even in
268 this case. */
269 SCM_VALIDATE_OPINPORT (1, port);
270
271 pt = SCM_PTAB_ENTRY (port);
272
273 /* if the current read buffer is filled, or the
274 last pushed-back char has been read and the saved buffer is
275 filled, result is true. */
276 if (pt->read_pos < pt->read_end
277 || (pt->read_buf == pt->putback_buf
278 && pt->saved_read_pos < pt->saved_read_end))
279 return SCM_BOOL_T;
280 else
281 {
282 scm_t_ptob_descriptor *ptob = &scm_ptobs[SCM_PTOBNUM (port)];
283
284 if (ptob->input_waiting)
285 return scm_from_bool(ptob->input_waiting (port));
286 else
287 return SCM_BOOL_T;
288 }
289 }
290 #undef FUNC_NAME
291
292 /* move up to read_len chars from port's putback and/or read buffers
293 into memory starting at dest. returns the number of chars moved. */
294 size_t scm_take_from_input_buffers (SCM port, char *dest, size_t read_len)
295 {
296 scm_t_port *pt = SCM_PTAB_ENTRY (port);
297 size_t chars_read = 0;
298 size_t from_buf = min (pt->read_end - pt->read_pos, read_len);
299
300 if (from_buf > 0)
301 {
302 memcpy (dest, pt->read_pos, from_buf);
303 pt->read_pos += from_buf;
304 chars_read += from_buf;
305 read_len -= from_buf;
306 dest += from_buf;
307 }
308
309 /* if putback was active, try the real input buffer too. */
310 if (pt->read_buf == pt->putback_buf)
311 {
312 from_buf = min (pt->saved_read_end - pt->saved_read_pos, read_len);
313 if (from_buf > 0)
314 {
315 memcpy (dest, pt->saved_read_pos, from_buf);
316 pt->saved_read_pos += from_buf;
317 chars_read += from_buf;
318 }
319 }
320 return chars_read;
321 }
322
323 /* Clear a port's read buffers, returning the contents. */
324 SCM_DEFINE (scm_drain_input, "drain-input", 1, 0, 0,
325 (SCM port),
326 "This procedure clears a port's input buffers, similar\n"
327 "to the way that force-output clears the output buffer. The\n"
328 "contents of the buffers are returned as a single string, e.g.,\n"
329 "\n"
330 "@lisp\n"
331 "(define p (open-input-file ...))\n"
332 "(drain-input p) => empty string, nothing buffered yet.\n"
333 "(unread-char (read-char p) p)\n"
334 "(drain-input p) => initial chars from p, up to the buffer size.\n"
335 "@end lisp\n\n"
336 "Draining the buffers may be useful for cleanly finishing\n"
337 "buffered I/O so that the file descriptor can be used directly\n"
338 "for further input.")
339 #define FUNC_NAME s_scm_drain_input
340 {
341 SCM result;
342 char *data;
343 scm_t_port *pt;
344 long count;
345
346 SCM_VALIDATE_OPINPORT (1, port);
347 pt = SCM_PTAB_ENTRY (port);
348
349 count = pt->read_end - pt->read_pos;
350 if (pt->read_buf == pt->putback_buf)
351 count += pt->saved_read_end - pt->saved_read_pos;
352
353 if (count)
354 {
355 result = scm_i_make_string (count, &data, 0);
356 scm_take_from_input_buffers (port, data, count);
357 }
358 else
359 result = scm_nullstr;
360
361 return result;
362 }
363 #undef FUNC_NAME
364
365 \f
366 /* Standard ports --- current input, output, error, and more(!). */
367
368 static SCM cur_inport_fluid = 0;
369 static SCM cur_outport_fluid = 0;
370 static SCM cur_errport_fluid = 0;
371 static SCM cur_loadport_fluid = 0;
372
373 SCM_DEFINE (scm_current_input_port, "current-input-port", 0, 0, 0,
374 (),
375 "Return the current input port. This is the default port used\n"
376 "by many input procedures. Initially, @code{current-input-port}\n"
377 "returns the @dfn{standard input} in Unix and C terminology.")
378 #define FUNC_NAME s_scm_current_input_port
379 {
380 if (cur_inport_fluid)
381 return scm_fluid_ref (cur_inport_fluid);
382 else
383 return SCM_BOOL_F;
384 }
385 #undef FUNC_NAME
386
387 SCM_DEFINE (scm_current_output_port, "current-output-port", 0, 0, 0,
388 (),
389 "Return the current output port. This is the default port used\n"
390 "by many output procedures. Initially,\n"
391 "@code{current-output-port} returns the @dfn{standard output} in\n"
392 "Unix and C terminology.")
393 #define FUNC_NAME s_scm_current_output_port
394 {
395 if (cur_outport_fluid)
396 return scm_fluid_ref (cur_outport_fluid);
397 else
398 return SCM_BOOL_F;
399 }
400 #undef FUNC_NAME
401
402 SCM_DEFINE (scm_current_error_port, "current-error-port", 0, 0, 0,
403 (),
404 "Return the port to which errors and warnings should be sent (the\n"
405 "@dfn{standard error} in Unix and C terminology).")
406 #define FUNC_NAME s_scm_current_error_port
407 {
408 if (cur_errport_fluid)
409 return scm_fluid_ref (cur_errport_fluid);
410 else
411 return SCM_BOOL_F;
412 }
413 #undef FUNC_NAME
414
415 SCM_DEFINE (scm_current_load_port, "current-load-port", 0, 0, 0,
416 (),
417 "Return the current-load-port.\n"
418 "The load port is used internally by @code{primitive-load}.")
419 #define FUNC_NAME s_scm_current_load_port
420 {
421 return scm_fluid_ref (cur_loadport_fluid);
422 }
423 #undef FUNC_NAME
424
425 SCM_DEFINE (scm_set_current_input_port, "set-current-input-port", 1, 0, 0,
426 (SCM port),
427 "@deffnx {Scheme Procedure} set-current-output-port port\n"
428 "@deffnx {Scheme Procedure} set-current-error-port port\n"
429 "Change the ports returned by @code{current-input-port},\n"
430 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
431 "so that they use the supplied @var{port} for input or output.")
432 #define FUNC_NAME s_scm_set_current_input_port
433 {
434 SCM oinp = scm_fluid_ref (cur_inport_fluid);
435 SCM_VALIDATE_OPINPORT (1, port);
436 scm_fluid_set_x (cur_inport_fluid, port);
437 return oinp;
438 }
439 #undef FUNC_NAME
440
441
442 SCM_DEFINE (scm_set_current_output_port, "set-current-output-port", 1, 0, 0,
443 (SCM port),
444 "Set the current default output port to @var{port}.")
445 #define FUNC_NAME s_scm_set_current_output_port
446 {
447 SCM ooutp = scm_fluid_ref (cur_outport_fluid);
448 port = SCM_COERCE_OUTPORT (port);
449 SCM_VALIDATE_OPOUTPORT (1, port);
450 scm_fluid_set_x (cur_outport_fluid, port);
451 return ooutp;
452 }
453 #undef FUNC_NAME
454
455
456 SCM_DEFINE (scm_set_current_error_port, "set-current-error-port", 1, 0, 0,
457 (SCM port),
458 "Set the current default error port to @var{port}.")
459 #define FUNC_NAME s_scm_set_current_error_port
460 {
461 SCM oerrp = scm_fluid_ref (cur_errport_fluid);
462 port = SCM_COERCE_OUTPORT (port);
463 SCM_VALIDATE_OPOUTPORT (1, port);
464 scm_fluid_set_x (cur_errport_fluid, port);
465 return oerrp;
466 }
467 #undef FUNC_NAME
468
469 void
470 scm_dynwind_current_input_port (SCM port)
471 #define FUNC_NAME NULL
472 {
473 SCM_VALIDATE_OPINPORT (1, port);
474 scm_dynwind_fluid (cur_inport_fluid, port);
475 }
476 #undef FUNC_NAME
477
478 void
479 scm_dynwind_current_output_port (SCM port)
480 #define FUNC_NAME NULL
481 {
482 port = SCM_COERCE_OUTPORT (port);
483 SCM_VALIDATE_OPOUTPORT (1, port);
484 scm_dynwind_fluid (cur_outport_fluid, port);
485 }
486 #undef FUNC_NAME
487
488 void
489 scm_dynwind_current_error_port (SCM port)
490 #define FUNC_NAME NULL
491 {
492 port = SCM_COERCE_OUTPORT (port);
493 SCM_VALIDATE_OPOUTPORT (1, port);
494 scm_dynwind_fluid (cur_errport_fluid, port);
495 }
496 #undef FUNC_NAME
497
498 void
499 scm_i_dynwind_current_load_port (SCM port)
500 {
501 scm_dynwind_fluid (cur_loadport_fluid, port);
502 }
503
504 \f
505 /* The port table --- an array of pointers to ports. */
506
507 /*
508 We need a global registry of ports to flush them all at exit, and to
509 get all the ports matching a file descriptor.
510 */
511 SCM scm_i_port_weak_hash;
512
513 scm_i_pthread_mutex_t scm_i_port_table_mutex = SCM_I_PTHREAD_MUTEX_INITIALIZER;
514
515 \f
516 /* Port finalization. */
517
518
519 static void finalize_port (GC_PTR, GC_PTR);
520
521 /* Register a finalizer for PORT. */
522 static SCM_C_INLINE_KEYWORD void
523 register_finalizer_for_port (SCM port)
524 {
525 GC_finalization_proc prev_finalizer;
526 GC_PTR prev_finalization_data;
527
528 /* Register a finalizer for PORT so that its iconv CDs get freed and
529 optionally its type's `free' function gets called. */
530 GC_REGISTER_FINALIZER_NO_ORDER (SCM2PTR (port), finalize_port, 0,
531 &prev_finalizer,
532 &prev_finalization_data);
533 }
534
535 /* Finalize the object (a port) pointed to by PTR. */
536 static void
537 finalize_port (GC_PTR ptr, GC_PTR data)
538 {
539 long port_type;
540 SCM port = PTR2SCM (ptr);
541
542 if (!SCM_PORTP (port))
543 abort ();
544
545 if (SCM_OPENP (port))
546 {
547 if (SCM_REVEALED (port) > 0)
548 /* Keep "revealed" ports alive and re-register a finalizer. */
549 register_finalizer_for_port (port);
550 else
551 {
552 scm_t_port *entry;
553
554 port_type = SCM_TC2PTOBNUM (SCM_CELL_TYPE (port));
555 if (port_type >= scm_numptob)
556 abort ();
557
558 if (scm_ptobs[port_type].free)
559 /* Yes, I really do mean `.free' rather than `.close'. `.close'
560 is for explicit `close-port' by user. */
561 scm_ptobs[port_type].free (port);
562
563 entry = SCM_PTAB_ENTRY (port);
564
565 if (entry->input_cd != (iconv_t) -1)
566 iconv_close (entry->input_cd);
567 if (entry->output_cd != (iconv_t) -1)
568 iconv_close (entry->output_cd);
569
570 SCM_SETSTREAM (port, 0);
571 SCM_CLR_PORT_OPEN_FLAG (port);
572
573 scm_gc_ports_collected++;
574 }
575 }
576 }
577
578
579
580 \f
581
582 /* This function is not and should not be thread safe. */
583 SCM
584 scm_new_port_table_entry (scm_t_bits tag)
585 #define FUNC_NAME "scm_new_port_table_entry"
586 {
587 /*
588 We initialize the cell to empty, this is in case scm_gc_calloc
589 triggers GC ; we don't want the GC to scan a half-finished Z.
590 */
591
592 SCM z = scm_cons (SCM_EOL, SCM_EOL);
593 scm_t_port *entry = (scm_t_port *) scm_gc_calloc (sizeof (scm_t_port), "port");
594 const char *enc;
595
596 entry->file_name = SCM_BOOL_F;
597 entry->rw_active = SCM_PORT_NEITHER;
598 entry->port = z;
599
600 /* Initialize this port with the thread's current default
601 encoding. */
602 enc = scm_i_default_port_encoding ();
603 entry->encoding = enc ? scm_gc_strdup (enc, "port") : NULL;
604
605 /* The conversion descriptors will be opened lazily. */
606 entry->input_cd = (iconv_t) -1;
607 entry->output_cd = (iconv_t) -1;
608
609 entry->ilseq_handler = scm_i_get_conversion_strategy (SCM_BOOL_F);
610
611 SCM_SET_CELL_TYPE (z, tag);
612 SCM_SETPTAB_ENTRY (z, entry);
613
614 scm_hashq_set_x (scm_i_port_weak_hash, z, SCM_BOOL_F);
615
616 /* For each new port, register a finalizer so that it port type's free
617 function can be invoked eventually. */
618 register_finalizer_for_port (z);
619
620 return z;
621 }
622 #undef FUNC_NAME
623
624 #if SCM_ENABLE_DEPRECATED==1
625 scm_t_port *
626 scm_add_to_port_table (SCM port)
627 {
628 SCM z;
629 scm_t_port * pt;
630
631 scm_c_issue_deprecation_warning ("scm_add_to_port_table is deprecated.");
632
633 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex);
634 z = scm_new_port_table_entry (scm_tc7_port);
635 pt = SCM_PTAB_ENTRY(z);
636 pt->port = port;
637 SCM_SETCAR (z, SCM_EOL);
638 SCM_SETCDR (z, SCM_EOL);
639 SCM_SETPTAB_ENTRY (port, pt);
640 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex);
641
642 return pt;
643 }
644 #endif
645
646
647 /* Remove a port from the table and destroy it. */
648
649 static void
650 scm_i_remove_port (SCM port)
651 #define FUNC_NAME "scm_remove_port"
652 {
653 scm_t_port *p;
654
655 scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex);
656
657 p = SCM_PTAB_ENTRY (port);
658 scm_port_non_buffer (p);
659 p->putback_buf = NULL;
660 p->putback_buf_size = 0;
661
662 if (p->input_cd != (iconv_t) -1)
663 {
664 iconv_close (p->input_cd);
665 p->input_cd = (iconv_t) -1;
666 }
667
668 if (p->output_cd != (iconv_t) -1)
669 {
670 iconv_close (p->output_cd);
671 p->output_cd = (iconv_t) -1;
672 }
673
674 SCM_SETPTAB_ENTRY (port, 0);
675
676 scm_hashq_remove_x (scm_i_port_weak_hash, port);
677
678 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex);
679 }
680 #undef FUNC_NAME
681
682
683 /* Functions for debugging. */
684 #ifdef GUILE_DEBUG
685 SCM_DEFINE (scm_pt_size, "pt-size", 0, 0, 0,
686 (),
687 "Return the number of ports in the port table. @code{pt-size}\n"
688 "is only included in @code{--enable-guile-debug} builds.")
689 #define FUNC_NAME s_scm_pt_size
690 {
691 return scm_from_int (SCM_HASHTABLE_N_ITEMS (scm_i_port_weak_hash));
692 }
693 #undef FUNC_NAME
694 #endif
695
696 void
697 scm_port_non_buffer (scm_t_port *pt)
698 {
699 pt->read_pos = pt->read_buf = pt->read_end = &pt->shortbuf;
700 pt->write_buf = pt->write_pos = &pt->shortbuf;
701 pt->read_buf_size = pt->write_buf_size = 1;
702 pt->write_end = pt->write_buf + pt->write_buf_size;
703 }
704
705 \f
706 /* Revealed counts --- an oddity inherited from SCSH. */
707
708 /* Find a port in the table and return its revealed count.
709 Also used by the garbage collector.
710 */
711
712 int
713 scm_revealed_count (SCM port)
714 {
715 return SCM_REVEALED(port);
716 }
717
718
719
720 /* Return the revealed count for a port. */
721
722 SCM_DEFINE (scm_port_revealed, "port-revealed", 1, 0, 0,
723 (SCM port),
724 "Return the revealed count for @var{port}.")
725 #define FUNC_NAME s_scm_port_revealed
726 {
727 port = SCM_COERCE_OUTPORT (port);
728 SCM_VALIDATE_OPENPORT (1, port);
729 return scm_from_int (scm_revealed_count (port));
730 }
731 #undef FUNC_NAME
732
733 /* Set the revealed count for a port. */
734 SCM_DEFINE (scm_set_port_revealed_x, "set-port-revealed!", 2, 0, 0,
735 (SCM port, SCM rcount),
736 "Sets the revealed count for a port to a given value.\n"
737 "The return value is unspecified.")
738 #define FUNC_NAME s_scm_set_port_revealed_x
739 {
740 port = SCM_COERCE_OUTPORT (port);
741 SCM_VALIDATE_OPENPORT (1, port);
742 SCM_REVEALED (port) = scm_to_int (rcount);
743 return SCM_UNSPECIFIED;
744 }
745 #undef FUNC_NAME
746
747
748 \f
749 /* Retrieving a port's mode. */
750
751 /* Return the flags that characterize a port based on the mode
752 * string used to open a file for that port.
753 *
754 * See PORT FLAGS in scm.h
755 */
756
757 static long
758 scm_i_mode_bits_n (SCM modes)
759 {
760 return (SCM_OPN
761 | (scm_i_string_contains_char (modes, 'r')
762 || scm_i_string_contains_char (modes, '+') ? SCM_RDNG : 0)
763 | (scm_i_string_contains_char (modes, 'w')
764 || scm_i_string_contains_char (modes, 'a')
765 || scm_i_string_contains_char (modes, '+') ? SCM_WRTNG : 0)
766 | (scm_i_string_contains_char (modes, '0') ? SCM_BUF0 : 0)
767 | (scm_i_string_contains_char (modes, 'l') ? SCM_BUFLINE : 0));
768 }
769
770 long
771 scm_mode_bits (char *modes)
772 {
773 return scm_i_mode_bits (scm_from_locale_string (modes));
774 }
775
776 long
777 scm_i_mode_bits (SCM modes)
778 {
779 long bits;
780
781 if (!scm_is_string (modes))
782 scm_wrong_type_arg_msg (NULL, 0, modes, "string");
783
784 bits = scm_i_mode_bits_n (modes);
785 scm_remember_upto_here_1 (modes);
786 return bits;
787 }
788
789 /* Return the mode flags from an open port.
790 * Some modes such as "append" are only used when opening
791 * a file and are not returned here. */
792
793 SCM_DEFINE (scm_port_mode, "port-mode", 1, 0, 0,
794 (SCM port),
795 "Return the port modes associated with the open port @var{port}.\n"
796 "These will not necessarily be identical to the modes used when\n"
797 "the port was opened, since modes such as \"append\" which are\n"
798 "used only during port creation are not retained.")
799 #define FUNC_NAME s_scm_port_mode
800 {
801 char modes[4];
802 modes[0] = '\0';
803
804 port = SCM_COERCE_OUTPORT (port);
805 SCM_VALIDATE_OPPORT (1, port);
806 if (SCM_CELL_WORD_0 (port) & SCM_RDNG) {
807 if (SCM_CELL_WORD_0 (port) & SCM_WRTNG)
808 strcpy (modes, "r+");
809 else
810 strcpy (modes, "r");
811 }
812 else if (SCM_CELL_WORD_0 (port) & SCM_WRTNG)
813 strcpy (modes, "w");
814 if (SCM_CELL_WORD_0 (port) & SCM_BUF0)
815 strcat (modes, "0");
816 return scm_from_locale_string (modes);
817 }
818 #undef FUNC_NAME
819
820
821 \f
822 /* Closing ports. */
823
824 /* scm_close_port
825 * Call the close operation on a port object.
826 * see also scm_close.
827 */
828 SCM_DEFINE (scm_close_port, "close-port", 1, 0, 0,
829 (SCM port),
830 "Close the specified port object. Return @code{#t} if it\n"
831 "successfully closes a port or @code{#f} if it was already\n"
832 "closed. An exception may be raised if an error occurs, for\n"
833 "example when flushing buffered output. See also @ref{Ports and\n"
834 "File Descriptors, close}, for a procedure which can close file\n"
835 "descriptors.")
836 #define FUNC_NAME s_scm_close_port
837 {
838 size_t i;
839 int rv;
840
841 port = SCM_COERCE_OUTPORT (port);
842
843 SCM_VALIDATE_PORT (1, port);
844 if (SCM_CLOSEDP (port))
845 return SCM_BOOL_F;
846 i = SCM_PTOBNUM (port);
847 if (scm_ptobs[i].close)
848 rv = (scm_ptobs[i].close) (port);
849 else
850 rv = 0;
851 scm_i_remove_port (port);
852 SCM_CLR_PORT_OPEN_FLAG (port);
853 return scm_from_bool (rv >= 0);
854 }
855 #undef FUNC_NAME
856
857 SCM_DEFINE (scm_close_input_port, "close-input-port", 1, 0, 0,
858 (SCM port),
859 "Close the specified input port object. The routine has no effect if\n"
860 "the file has already been closed. An exception may be raised if an\n"
861 "error occurs. The value returned is unspecified.\n\n"
862 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
863 "which can close file descriptors.")
864 #define FUNC_NAME s_scm_close_input_port
865 {
866 SCM_VALIDATE_INPUT_PORT (1, port);
867 scm_close_port (port);
868 return SCM_UNSPECIFIED;
869 }
870 #undef FUNC_NAME
871
872 SCM_DEFINE (scm_close_output_port, "close-output-port", 1, 0, 0,
873 (SCM port),
874 "Close the specified output port object. The routine has no effect if\n"
875 "the file has already been closed. An exception may be raised if an\n"
876 "error occurs. The value returned is unspecified.\n\n"
877 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
878 "which can close file descriptors.")
879 #define FUNC_NAME s_scm_close_output_port
880 {
881 port = SCM_COERCE_OUTPORT (port);
882 SCM_VALIDATE_OUTPUT_PORT (1, port);
883 scm_close_port (port);
884 return SCM_UNSPECIFIED;
885 }
886 #undef FUNC_NAME
887
888 static SCM
889 collect_keys (void *unused, SCM key, SCM value, SCM result)
890 {
891 return scm_cons (key, result);
892 }
893
894 void
895 scm_c_port_for_each (void (*proc)(void *data, SCM p), void *data)
896 {
897 SCM ports;
898
899 /* Copy out the port table as a list so that we get strong references
900 to all the values. */
901 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex);
902 ports = scm_internal_hash_fold (collect_keys, NULL,
903 SCM_EOL, scm_i_port_weak_hash);
904 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex);
905
906 for (; scm_is_pair (ports); ports = scm_cdr (ports))
907 {
908 SCM p = scm_car (ports);
909 if (SCM_PORTP (p))
910 proc (data, p);
911 }
912 }
913
914 SCM_DEFINE (scm_port_for_each, "port-for-each", 1, 0, 0,
915 (SCM proc),
916 "Apply @var{proc} to each port in the Guile port table\n"
917 "in turn. The return value is unspecified. More specifically,\n"
918 "@var{proc} is applied exactly once to every port that exists\n"
919 "in the system at the time @var{port-for-each} is invoked.\n"
920 "Changes to the port table while @var{port-for-each} is running\n"
921 "have no effect as far as @var{port-for-each} is concerned.")
922 #define FUNC_NAME s_scm_port_for_each
923 {
924 SCM_VALIDATE_PROC (1, proc);
925
926 scm_c_port_for_each ((void (*)(void*,SCM))scm_call_1, proc);
927 return SCM_UNSPECIFIED;
928 }
929 #undef FUNC_NAME
930
931
932 \f
933 /* Utter miscellany. Gosh, we should clean this up some time. */
934
935 SCM_DEFINE (scm_input_port_p, "input-port?", 1, 0, 0,
936 (SCM x),
937 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
938 "@code{#f}. Any object satisfying this predicate also satisfies\n"
939 "@code{port?}.")
940 #define FUNC_NAME s_scm_input_port_p
941 {
942 return scm_from_bool (SCM_INPUT_PORT_P (x));
943 }
944 #undef FUNC_NAME
945
946 SCM_DEFINE (scm_output_port_p, "output-port?", 1, 0, 0,
947 (SCM x),
948 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
949 "@code{#f}. Any object satisfying this predicate also satisfies\n"
950 "@code{port?}.")
951 #define FUNC_NAME s_scm_output_port_p
952 {
953 x = SCM_COERCE_OUTPORT (x);
954 return scm_from_bool (SCM_OUTPUT_PORT_P (x));
955 }
956 #undef FUNC_NAME
957
958 SCM_DEFINE (scm_port_p, "port?", 1, 0, 0,
959 (SCM x),
960 "Return a boolean indicating whether @var{x} is a port.\n"
961 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
962 "@var{x}))}.")
963 #define FUNC_NAME s_scm_port_p
964 {
965 return scm_from_bool (SCM_PORTP (x));
966 }
967 #undef FUNC_NAME
968
969 SCM_DEFINE (scm_port_closed_p, "port-closed?", 1, 0, 0,
970 (SCM port),
971 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
972 "open.")
973 #define FUNC_NAME s_scm_port_closed_p
974 {
975 SCM_VALIDATE_PORT (1, port);
976 return scm_from_bool (!SCM_OPPORTP (port));
977 }
978 #undef FUNC_NAME
979
980 SCM_DEFINE (scm_eof_object_p, "eof-object?", 1, 0, 0,
981 (SCM x),
982 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
983 "return @code{#f}.")
984 #define FUNC_NAME s_scm_eof_object_p
985 {
986 return scm_from_bool(SCM_EOF_OBJECT_P (x));
987 }
988 #undef FUNC_NAME
989
990 SCM_DEFINE (scm_force_output, "force-output", 0, 1, 0,
991 (SCM port),
992 "Flush the specified output port, or the current output port if @var{port}\n"
993 "is omitted. The current output buffer contents are passed to the\n"
994 "underlying port implementation (e.g., in the case of fports, the\n"
995 "data will be written to the file and the output buffer will be cleared.)\n"
996 "It has no effect on an unbuffered port.\n\n"
997 "The return value is unspecified.")
998 #define FUNC_NAME s_scm_force_output
999 {
1000 if (SCM_UNBNDP (port))
1001 port = scm_current_output_port ();
1002 else
1003 {
1004 port = SCM_COERCE_OUTPORT (port);
1005 SCM_VALIDATE_OPOUTPORT (1, port);
1006 }
1007 scm_flush (port);
1008 return SCM_UNSPECIFIED;
1009 }
1010 #undef FUNC_NAME
1011
1012
1013 static void
1014 flush_output_port (void *closure, SCM port)
1015 {
1016 if (SCM_OPOUTPORTP (port))
1017 scm_flush (port);
1018 }
1019
1020 SCM_DEFINE (scm_flush_all_ports, "flush-all-ports", 0, 0, 0,
1021 (),
1022 "Equivalent to calling @code{force-output} on\n"
1023 "all open output ports. The return value is unspecified.")
1024 #define FUNC_NAME s_scm_flush_all_ports
1025 {
1026 scm_c_port_for_each (&flush_output_port, NULL);
1027 return SCM_UNSPECIFIED;
1028 }
1029 #undef FUNC_NAME
1030
1031 SCM_DEFINE (scm_read_char, "read-char", 0, 1, 0,
1032 (SCM port),
1033 "Return the next character available from @var{port}, updating\n"
1034 "@var{port} to point to the following character. If no more\n"
1035 "characters are available, the end-of-file object is returned.\n"
1036 "\n"
1037 "When @var{port}'s data cannot be decoded according to its\n"
1038 "character encoding, a @code{decoding-error} is raised and\n"
1039 "@var{port} points past the erroneous byte sequence.\n")
1040 #define FUNC_NAME s_scm_read_char
1041 {
1042 scm_t_wchar c;
1043 if (SCM_UNBNDP (port))
1044 port = scm_current_input_port ();
1045 SCM_VALIDATE_OPINPORT (1, port);
1046 c = scm_getc (port);
1047 if (EOF == c)
1048 return SCM_EOF_VAL;
1049 return SCM_MAKE_CHAR (c);
1050 }
1051 #undef FUNC_NAME
1052
1053 /* Update the line and column number of PORT after consumption of C. */
1054 static inline void
1055 update_port_lf (scm_t_wchar c, SCM port)
1056 {
1057 switch (c)
1058 {
1059 case '\a':
1060 break;
1061 case '\b':
1062 SCM_DECCOL (port);
1063 break;
1064 case '\n':
1065 SCM_INCLINE (port);
1066 break;
1067 case '\r':
1068 SCM_ZEROCOL (port);
1069 break;
1070 case '\t':
1071 SCM_TABCOL (port);
1072 break;
1073 default:
1074 SCM_INCCOL (port);
1075 break;
1076 }
1077 }
1078
1079 #define SCM_MBCHAR_BUF_SIZE (4)
1080
1081 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1082 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1083 static scm_t_wchar
1084 utf8_to_codepoint (const scm_t_uint8 *utf8_buf, size_t size)
1085 {
1086 scm_t_wchar codepoint;
1087
1088 if (utf8_buf[0] <= 0x7f)
1089 {
1090 assert (size == 1);
1091 codepoint = utf8_buf[0];
1092 }
1093 else if ((utf8_buf[0] & 0xe0) == 0xc0)
1094 {
1095 assert (size == 2);
1096 codepoint = ((scm_t_wchar) utf8_buf[0] & 0x1f) << 6UL
1097 | (utf8_buf[1] & 0x3f);
1098 }
1099 else if ((utf8_buf[0] & 0xf0) == 0xe0)
1100 {
1101 assert (size == 3);
1102 codepoint = ((scm_t_wchar) utf8_buf[0] & 0x0f) << 12UL
1103 | ((scm_t_wchar) utf8_buf[1] & 0x3f) << 6UL
1104 | (utf8_buf[2] & 0x3f);
1105 }
1106 else
1107 {
1108 assert (size == 4);
1109 codepoint = ((scm_t_wchar) utf8_buf[0] & 0x07) << 18UL
1110 | ((scm_t_wchar) utf8_buf[1] & 0x3f) << 12UL
1111 | ((scm_t_wchar) utf8_buf[2] & 0x3f) << 6UL
1112 | (utf8_buf[3] & 0x3f);
1113 }
1114
1115 return codepoint;
1116 }
1117
1118 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1119 with the byte representation of the codepoint in PORT's encoding, and
1120 set *LEN to the length in bytes of that representation. Return 0 on
1121 success and an errno value on error. */
1122 static int
1123 get_codepoint (SCM port, scm_t_wchar *codepoint,
1124 char buf[SCM_MBCHAR_BUF_SIZE], size_t *len)
1125 {
1126 int err, byte_read;
1127 size_t bytes_consumed, output_size;
1128 char *output;
1129 scm_t_uint8 utf8_buf[SCM_MBCHAR_BUF_SIZE];
1130 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1131
1132 if (SCM_UNLIKELY (pt->input_cd == (iconv_t) -1))
1133 /* Initialize the conversion descriptors. */
1134 scm_i_set_port_encoding_x (port, pt->encoding);
1135
1136 for (output_size = 0, output = (char *) utf8_buf,
1137 bytes_consumed = 0, err = 0;
1138 err == 0 && output_size == 0
1139 && (bytes_consumed == 0 || byte_read != EOF);
1140 bytes_consumed++)
1141 {
1142 char *input;
1143 size_t input_left, output_left, done;
1144
1145 byte_read = scm_get_byte_or_eof (port);
1146 if (byte_read == EOF)
1147 {
1148 if (bytes_consumed == 0)
1149 {
1150 *codepoint = (scm_t_wchar) EOF;
1151 *len = 0;
1152 return 0;
1153 }
1154 else
1155 continue;
1156 }
1157
1158 buf[bytes_consumed] = byte_read;
1159
1160 input = buf;
1161 input_left = bytes_consumed + 1;
1162 output_left = sizeof (utf8_buf);
1163
1164 done = iconv (pt->input_cd, &input, &input_left,
1165 &output, &output_left);
1166 if (done == (size_t) -1)
1167 {
1168 err = errno;
1169 if (err == EINVAL)
1170 /* Missing input: keep trying. */
1171 err = 0;
1172 }
1173 else
1174 output_size = sizeof (utf8_buf) - output_left;
1175 }
1176
1177 if (SCM_UNLIKELY (output_size == 0))
1178 /* An unterminated sequence. */
1179 err = EILSEQ;
1180
1181 if (SCM_UNLIKELY (err != 0))
1182 {
1183 /* Reset the `iconv' state. */
1184 iconv (pt->input_cd, NULL, NULL, NULL, NULL);
1185
1186 if (pt->ilseq_handler == SCM_ICONVEH_QUESTION_MARK)
1187 {
1188 *codepoint = '?';
1189 err = 0;
1190 }
1191
1192 /* Fail when the strategy is SCM_ICONVEH_ERROR or
1193 SCM_ICONVEH_ESCAPE_SEQUENCE (the latter doesn't make sense for
1194 input encoding errors.) */
1195 }
1196 else
1197 {
1198 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1199 *codepoint = utf8_to_codepoint (utf8_buf, output_size);
1200 update_port_lf (*codepoint, port);
1201 }
1202
1203 *len = bytes_consumed;
1204
1205 return err;
1206 }
1207
1208 /* Read a codepoint from PORT and return it. */
1209 scm_t_wchar
1210 scm_getc (SCM port)
1211 #define FUNC_NAME "scm_getc"
1212 {
1213 int err;
1214 size_t len;
1215 scm_t_wchar codepoint;
1216 char buf[SCM_MBCHAR_BUF_SIZE];
1217
1218 err = get_codepoint (port, &codepoint, buf, &len);
1219 if (SCM_UNLIKELY (err != 0))
1220 /* At this point PORT should point past the invalid encoding, as per
1221 R6RS-lib Section 8.2.4. */
1222 scm_decoding_error (FUNC_NAME, err, "input decoding error", port);
1223
1224 return codepoint;
1225 }
1226 #undef FUNC_NAME
1227
1228 /* this should only be called when the read buffer is empty. it
1229 tries to refill the read buffer. it returns the first char from
1230 the port, which is either EOF or *(pt->read_pos). */
1231 int
1232 scm_fill_input (SCM port)
1233 {
1234 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1235
1236 assert (pt->read_pos == pt->read_end);
1237
1238 if (pt->read_buf == pt->putback_buf)
1239 {
1240 /* finished reading put-back chars. */
1241 pt->read_buf = pt->saved_read_buf;
1242 pt->read_pos = pt->saved_read_pos;
1243 pt->read_end = pt->saved_read_end;
1244 pt->read_buf_size = pt->saved_read_buf_size;
1245 if (pt->read_pos < pt->read_end)
1246 return *(pt->read_pos);
1247 }
1248 return scm_ptobs[SCM_PTOBNUM (port)].fill_input (port);
1249 }
1250
1251
1252 /* scm_lfwrite
1253 *
1254 * This function differs from scm_c_write; it updates port line and
1255 * column. */
1256
1257 void
1258 scm_lfwrite (const char *ptr, size_t size, SCM port)
1259 {
1260 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1261 scm_t_ptob_descriptor *ptob = &scm_ptobs[SCM_PTOBNUM (port)];
1262
1263 if (pt->rw_active == SCM_PORT_READ)
1264 scm_end_input (port);
1265
1266 ptob->write (port, ptr, size);
1267
1268 for (; size; ptr++, size--)
1269 update_port_lf ((scm_t_wchar) (unsigned char) *ptr, port);
1270
1271 if (pt->rw_random)
1272 pt->rw_active = SCM_PORT_WRITE;
1273 }
1274
1275 /* Write STR to PORT from START inclusive to END exclusive. */
1276 void
1277 scm_lfwrite_substr (SCM str, size_t start, size_t end, SCM port)
1278 {
1279 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1280
1281 if (pt->rw_active == SCM_PORT_READ)
1282 scm_end_input (port);
1283
1284 if (end == (size_t) -1)
1285 end = scm_i_string_length (str);
1286
1287 scm_display (scm_c_substring (str, start, end), port);
1288
1289 if (pt->rw_random)
1290 pt->rw_active = SCM_PORT_WRITE;
1291 }
1292
1293 /* scm_c_read
1294 *
1295 * Used by an application to read arbitrary number of bytes from an
1296 * SCM port. Same semantics as libc read, except that scm_c_read only
1297 * returns less than SIZE bytes if at end-of-file.
1298 *
1299 * Warning: Doesn't update port line and column counts! */
1300
1301 /* This structure, and the following swap_buffer function, are used
1302 for temporarily swapping a port's own read buffer, and the buffer
1303 that the caller of scm_c_read provides. */
1304 struct port_and_swap_buffer
1305 {
1306 scm_t_port *pt;
1307 unsigned char *buffer;
1308 size_t size;
1309 };
1310
1311 static void
1312 swap_buffer (void *data)
1313 {
1314 struct port_and_swap_buffer *psb = (struct port_and_swap_buffer *) data;
1315 unsigned char *old_buf = psb->pt->read_buf;
1316 size_t old_size = psb->pt->read_buf_size;
1317
1318 /* Make the port use (buffer, size) from the struct. */
1319 psb->pt->read_pos = psb->pt->read_buf = psb->pt->read_end = psb->buffer;
1320 psb->pt->read_buf_size = psb->size;
1321
1322 /* Save the port's old (buffer, size) in the struct. */
1323 psb->buffer = old_buf;
1324 psb->size = old_size;
1325 }
1326
1327 size_t
1328 scm_c_read (SCM port, void *buffer, size_t size)
1329 #define FUNC_NAME "scm_c_read"
1330 {
1331 scm_t_port *pt;
1332 size_t n_read = 0, n_available;
1333 struct port_and_swap_buffer psb;
1334
1335 SCM_VALIDATE_OPINPORT (1, port);
1336
1337 pt = SCM_PTAB_ENTRY (port);
1338 if (pt->rw_active == SCM_PORT_WRITE)
1339 scm_ptobs[SCM_PTOBNUM (port)].flush (port);
1340
1341 if (pt->rw_random)
1342 pt->rw_active = SCM_PORT_READ;
1343
1344 /* Take bytes first from the port's read buffer. */
1345 if (pt->read_pos < pt->read_end)
1346 {
1347 n_available = min (size, pt->read_end - pt->read_pos);
1348 memcpy (buffer, pt->read_pos, n_available);
1349 buffer = (char *) buffer + n_available;
1350 pt->read_pos += n_available;
1351 n_read += n_available;
1352 size -= n_available;
1353 }
1354
1355 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1356 if (size == 0)
1357 return n_read;
1358
1359 /* Now we will call scm_fill_input repeatedly until we have read the
1360 requested number of bytes. (Note that a single scm_fill_input
1361 call does not guarantee to fill the whole of the port's read
1362 buffer.) */
1363 if (pt->read_buf_size <= 1 && pt->encoding == NULL)
1364 {
1365 /* The port that we are reading from is unbuffered - i.e. does
1366 not have its own persistent buffer - but we have a buffer,
1367 provided by our caller, that is the right size for the data
1368 that is wanted. For the following scm_fill_input calls,
1369 therefore, we use the buffer in hand as the port's read
1370 buffer.
1371
1372 We need to make sure that the port's normal (1 byte) buffer
1373 is reinstated in case one of the scm_fill_input () calls
1374 throws an exception; we use the scm_dynwind_* API to achieve
1375 that.
1376
1377 A consequence of this optimization is that the fill_input
1378 functions can't unget characters. That'll push data to the
1379 pushback buffer instead of this psb buffer. */
1380 #if SCM_DEBUG == 1
1381 unsigned char *pback = pt->putback_buf;
1382 #endif
1383 psb.pt = pt;
1384 psb.buffer = buffer;
1385 psb.size = size;
1386 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE);
1387 scm_dynwind_rewind_handler (swap_buffer, &psb, SCM_F_WIND_EXPLICITLY);
1388 scm_dynwind_unwind_handler (swap_buffer, &psb, SCM_F_WIND_EXPLICITLY);
1389
1390 /* Call scm_fill_input until we have all the bytes that we need,
1391 or we hit EOF. */
1392 while (pt->read_buf_size && (scm_fill_input (port) != EOF))
1393 {
1394 pt->read_buf_size -= (pt->read_end - pt->read_pos);
1395 pt->read_pos = pt->read_buf = pt->read_end;
1396 }
1397 #if SCM_DEBUG == 1
1398 if (pback != pt->putback_buf
1399 || pt->read_buf - (unsigned char *) buffer < 0)
1400 scm_misc_error (FUNC_NAME,
1401 "scm_c_read must not call a fill function that pushes "
1402 "back characters onto an unbuffered port", SCM_EOL);
1403 #endif
1404 n_read += pt->read_buf - (unsigned char *) buffer;
1405
1406 /* Reinstate the port's normal buffer. */
1407 scm_dynwind_end ();
1408 }
1409 else
1410 {
1411 /* The port has its own buffer. It is important that we use it,
1412 even if it happens to be smaller than our caller's buffer, so
1413 that a custom port implementation's entry points (in
1414 particular, fill_input) can rely on the buffer always being
1415 the same as they first set up. */
1416 while (size && (scm_fill_input (port) != EOF))
1417 {
1418 n_available = min (size, pt->read_end - pt->read_pos);
1419 memcpy (buffer, pt->read_pos, n_available);
1420 buffer = (char *) buffer + n_available;
1421 pt->read_pos += n_available;
1422 n_read += n_available;
1423 size -= n_available;
1424 }
1425 }
1426
1427 return n_read;
1428 }
1429 #undef FUNC_NAME
1430
1431 /* scm_c_write
1432 *
1433 * Used by an application to write arbitrary number of bytes to an SCM
1434 * port. Similar semantics as libc write. However, unlike libc
1435 * write, scm_c_write writes the requested number of bytes and has no
1436 * return value.
1437 *
1438 * Warning: Doesn't update port line and column counts!
1439 */
1440
1441 void
1442 scm_c_write (SCM port, const void *ptr, size_t size)
1443 #define FUNC_NAME "scm_c_write"
1444 {
1445 scm_t_port *pt;
1446 scm_t_ptob_descriptor *ptob;
1447
1448 SCM_VALIDATE_OPOUTPORT (1, port);
1449
1450 pt = SCM_PTAB_ENTRY (port);
1451 ptob = &scm_ptobs[SCM_PTOBNUM (port)];
1452
1453 if (pt->rw_active == SCM_PORT_READ)
1454 scm_end_input (port);
1455
1456 ptob->write (port, ptr, size);
1457
1458 if (pt->rw_random)
1459 pt->rw_active = SCM_PORT_WRITE;
1460 }
1461 #undef FUNC_NAME
1462
1463 void
1464 scm_flush (SCM port)
1465 {
1466 long i = SCM_PTOBNUM (port);
1467 assert (i >= 0);
1468 (scm_ptobs[i].flush) (port);
1469 }
1470
1471 void
1472 scm_end_input (SCM port)
1473 {
1474 long offset;
1475 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1476
1477 if (pt->read_buf == pt->putback_buf)
1478 {
1479 offset = pt->read_end - pt->read_pos;
1480 pt->read_buf = pt->saved_read_buf;
1481 pt->read_pos = pt->saved_read_pos;
1482 pt->read_end = pt->saved_read_end;
1483 pt->read_buf_size = pt->saved_read_buf_size;
1484 }
1485 else
1486 offset = 0;
1487
1488 scm_ptobs[SCM_PTOBNUM (port)].end_input (port, offset);
1489 }
1490
1491 \f
1492
1493
1494 void
1495 scm_unget_byte (int c, SCM port)
1496 #define FUNC_NAME "scm_unget_byte"
1497 {
1498 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1499
1500 if (pt->read_buf == pt->putback_buf)
1501 /* already using the put-back buffer. */
1502 {
1503 /* enlarge putback_buf if necessary. */
1504 if (pt->read_end == pt->read_buf + pt->read_buf_size
1505 && pt->read_buf == pt->read_pos)
1506 {
1507 size_t new_size = pt->read_buf_size * 2;
1508 unsigned char *tmp = (unsigned char *)
1509 scm_gc_realloc (pt->putback_buf, pt->read_buf_size, new_size,
1510 "putback buffer");
1511
1512 pt->read_pos = pt->read_buf = pt->putback_buf = tmp;
1513 pt->read_end = pt->read_buf + pt->read_buf_size;
1514 pt->read_buf_size = pt->putback_buf_size = new_size;
1515 }
1516
1517 /* shift any existing bytes to buffer + 1. */
1518 if (pt->read_pos == pt->read_end)
1519 pt->read_end = pt->read_buf + 1;
1520 else if (pt->read_pos != pt->read_buf + 1)
1521 {
1522 int count = pt->read_end - pt->read_pos;
1523
1524 memmove (pt->read_buf + 1, pt->read_pos, count);
1525 pt->read_end = pt->read_buf + 1 + count;
1526 }
1527
1528 pt->read_pos = pt->read_buf;
1529 }
1530 else
1531 /* switch to the put-back buffer. */
1532 {
1533 if (pt->putback_buf == NULL)
1534 {
1535 pt->putback_buf
1536 = (unsigned char *) scm_gc_malloc_pointerless
1537 (SCM_INITIAL_PUTBACK_BUF_SIZE, "putback buffer");
1538 pt->putback_buf_size = SCM_INITIAL_PUTBACK_BUF_SIZE;
1539 }
1540
1541 pt->saved_read_buf = pt->read_buf;
1542 pt->saved_read_pos = pt->read_pos;
1543 pt->saved_read_end = pt->read_end;
1544 pt->saved_read_buf_size = pt->read_buf_size;
1545
1546 pt->read_pos = pt->read_buf = pt->putback_buf;
1547 pt->read_end = pt->read_buf + 1;
1548 pt->read_buf_size = pt->putback_buf_size;
1549 }
1550
1551 *pt->read_buf = c;
1552
1553 if (pt->rw_random)
1554 pt->rw_active = SCM_PORT_READ;
1555 }
1556 #undef FUNC_NAME
1557
1558 void
1559 scm_ungetc (scm_t_wchar c, SCM port)
1560 #define FUNC_NAME "scm_ungetc"
1561 {
1562 scm_t_port *pt = SCM_PTAB_ENTRY (port);
1563 char *result;
1564 char result_buf[10];
1565 const char *encoding;
1566 size_t len;
1567 int i;
1568
1569 if (pt->encoding != NULL)
1570 encoding = pt->encoding;
1571 else
1572 encoding = "ISO-8859-1";
1573
1574 len = sizeof (result_buf);
1575 result = u32_conv_to_encoding (encoding,
1576 (enum iconv_ilseq_handler) pt->ilseq_handler,
1577 (uint32_t *) &c, 1, NULL,
1578 result_buf, &len);
1579
1580 if (SCM_UNLIKELY (result == NULL || len == 0))
1581 scm_encoding_error (FUNC_NAME, errno,
1582 "conversion to port encoding failed",
1583 SCM_BOOL_F, SCM_MAKE_CHAR (c));
1584
1585 for (i = len - 1; i >= 0; i--)
1586 scm_unget_byte (result[i], port);
1587
1588 if (SCM_UNLIKELY (result != result_buf))
1589 free (result);
1590
1591 if (c == '\n')
1592 {
1593 /* What should col be in this case?
1594 * We'll leave it at -1.
1595 */
1596 SCM_LINUM (port) -= 1;
1597 }
1598 else
1599 SCM_COL(port) -= 1;
1600 }
1601 #undef FUNC_NAME
1602
1603
1604 void
1605 scm_ungets (const char *s, int n, SCM port)
1606 {
1607 /* This is simple minded and inefficient, but unreading strings is
1608 * probably not a common operation, and remember that line and
1609 * column numbers have to be handled...
1610 *
1611 * Please feel free to write an optimized version!
1612 */
1613 while (n--)
1614 scm_ungetc (s[n], port);
1615 }
1616
1617
1618 SCM_DEFINE (scm_peek_char, "peek-char", 0, 1, 0,
1619 (SCM port),
1620 "Return the next character available from @var{port},\n"
1621 "@emph{without} updating @var{port} to point to the following\n"
1622 "character. If no more characters are available, the\n"
1623 "end-of-file object is returned.\n"
1624 "\n"
1625 "The value returned by\n"
1626 "a call to @code{peek-char} is the same as the value that would\n"
1627 "have been returned by a call to @code{read-char} on the same\n"
1628 "port. The only difference is that the very next call to\n"
1629 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1630 "return the value returned by the preceding call to\n"
1631 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1632 "an interactive port will hang waiting for input whenever a call\n"
1633 "to @code{read-char} would have hung.\n"
1634 "\n"
1635 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1636 "if such a situation occurs. However, unlike with @code{read-char},\n"
1637 "@var{port} still points at the beginning of the erroneous byte\n"
1638 "sequence when the error is raised.\n")
1639 #define FUNC_NAME s_scm_peek_char
1640 {
1641 int err;
1642 SCM result;
1643 scm_t_wchar c;
1644 char bytes[SCM_MBCHAR_BUF_SIZE];
1645 long column, line, i;
1646 size_t len;
1647
1648 if (SCM_UNBNDP (port))
1649 port = scm_current_input_port ();
1650 SCM_VALIDATE_OPINPORT (1, port);
1651
1652 column = SCM_COL (port);
1653 line = SCM_LINUM (port);
1654
1655 err = get_codepoint (port, &c, bytes, &len);
1656
1657 for (i = len - 1; i >= 0; i--)
1658 scm_unget_byte (bytes[i], port);
1659
1660 SCM_COL (port) = column;
1661 SCM_LINUM (port) = line;
1662
1663 if (SCM_UNLIKELY (err != 0))
1664 {
1665 scm_decoding_error (FUNC_NAME, err, "input decoding error", port);
1666
1667 /* Shouldn't happen since `catch' always aborts to prompt. */
1668 result = SCM_BOOL_F;
1669 }
1670 else if (c == EOF)
1671 result = SCM_EOF_VAL;
1672 else
1673 result = SCM_MAKE_CHAR (c);
1674
1675 return result;
1676 }
1677 #undef FUNC_NAME
1678
1679 SCM_DEFINE (scm_unread_char, "unread-char", 1, 1, 0,
1680 (SCM cobj, SCM port),
1681 "Place @var{char} in @var{port} so that it will be read by the\n"
1682 "next read operation. If called multiple times, the unread characters\n"
1683 "will be read again in last-in first-out order. If @var{port} is\n"
1684 "not supplied, the current input port is used.")
1685 #define FUNC_NAME s_scm_unread_char
1686 {
1687 int c;
1688
1689 SCM_VALIDATE_CHAR (1, cobj);
1690 if (SCM_UNBNDP (port))
1691 port = scm_current_input_port ();
1692 SCM_VALIDATE_OPINPORT (2, port);
1693
1694 c = SCM_CHAR (cobj);
1695
1696 scm_ungetc (c, port);
1697 return cobj;
1698 }
1699 #undef FUNC_NAME
1700
1701 SCM_DEFINE (scm_unread_string, "unread-string", 2, 0, 0,
1702 (SCM str, SCM port),
1703 "Place the string @var{str} in @var{port} so that its characters will be\n"
1704 "read in subsequent read operations. If called multiple times, the\n"
1705 "unread characters will be read again in last-in first-out order. If\n"
1706 "@var{port} is not supplied, the current-input-port is used.")
1707 #define FUNC_NAME s_scm_unread_string
1708 {
1709 int n;
1710 SCM_VALIDATE_STRING (1, str);
1711 if (SCM_UNBNDP (port))
1712 port = scm_current_input_port ();
1713 SCM_VALIDATE_OPINPORT (2, port);
1714
1715 n = scm_i_string_length (str);
1716
1717 while (n--)
1718 scm_ungetc (scm_i_string_ref (str, n), port);
1719
1720 return str;
1721 }
1722 #undef FUNC_NAME
1723
1724 SCM_DEFINE (scm_seek, "seek", 3, 0, 0,
1725 (SCM fd_port, SCM offset, SCM whence),
1726 "Sets the current position of @var{fd/port} to the integer\n"
1727 "@var{offset}, which is interpreted according to the value of\n"
1728 "@var{whence}.\n"
1729 "\n"
1730 "One of the following variables should be supplied for\n"
1731 "@var{whence}:\n"
1732 "@defvar SEEK_SET\n"
1733 "Seek from the beginning of the file.\n"
1734 "@end defvar\n"
1735 "@defvar SEEK_CUR\n"
1736 "Seek from the current position.\n"
1737 "@end defvar\n"
1738 "@defvar SEEK_END\n"
1739 "Seek from the end of the file.\n"
1740 "@end defvar\n"
1741 "If @var{fd/port} is a file descriptor, the underlying system\n"
1742 "call is @code{lseek}. @var{port} may be a string port.\n"
1743 "\n"
1744 "The value returned is the new position in the file. This means\n"
1745 "that the current position of a port can be obtained using:\n"
1746 "@lisp\n"
1747 "(seek port 0 SEEK_CUR)\n"
1748 "@end lisp")
1749 #define FUNC_NAME s_scm_seek
1750 {
1751 int how;
1752
1753 fd_port = SCM_COERCE_OUTPORT (fd_port);
1754
1755 how = scm_to_int (whence);
1756 if (how != SEEK_SET && how != SEEK_CUR && how != SEEK_END)
1757 SCM_OUT_OF_RANGE (3, whence);
1758
1759 if (SCM_OPPORTP (fd_port))
1760 {
1761 scm_t_ptob_descriptor *ptob = scm_ptobs + SCM_PTOBNUM (fd_port);
1762 off_t_or_off64_t off = scm_to_off_t_or_off64_t (offset);
1763 off_t_or_off64_t rv;
1764
1765 if (!ptob->seek)
1766 SCM_MISC_ERROR ("port is not seekable",
1767 scm_cons (fd_port, SCM_EOL));
1768 else
1769 rv = ptob->seek (fd_port, off, how);
1770 return scm_from_off_t_or_off64_t (rv);
1771 }
1772 else /* file descriptor?. */
1773 {
1774 off_t_or_off64_t off = scm_to_off_t_or_off64_t (offset);
1775 off_t_or_off64_t rv;
1776 rv = lseek_or_lseek64 (scm_to_int (fd_port), off, how);
1777 if (rv == -1)
1778 SCM_SYSERROR;
1779 return scm_from_off_t_or_off64_t (rv);
1780 }
1781 }
1782 #undef FUNC_NAME
1783
1784 #ifndef O_BINARY
1785 #define O_BINARY 0
1786 #endif
1787
1788 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
1789 doesn't have the filename version truncate(), hence this code. */
1790 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
1791 static int
1792 truncate (const char *file, off_t length)
1793 {
1794 int ret, fdes;
1795
1796 fdes = open (file, O_BINARY | O_WRONLY);
1797 if (fdes == -1)
1798 return -1;
1799
1800 ret = ftruncate (fdes, length);
1801 if (ret == -1)
1802 {
1803 int save_errno = errno;
1804 close (fdes);
1805 errno = save_errno;
1806 return -1;
1807 }
1808
1809 return close (fdes);
1810 }
1811 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
1812
1813 SCM_DEFINE (scm_truncate_file, "truncate-file", 1, 1, 0,
1814 (SCM object, SCM length),
1815 "Truncate @var{file} to @var{length} bytes. @var{file} can be a\n"
1816 "filename string, a port object, or an integer file descriptor.\n"
1817 "The return value is unspecified.\n"
1818 "\n"
1819 "For a port or file descriptor @var{length} can be omitted, in\n"
1820 "which case the file is truncated at the current position (per\n"
1821 "@code{ftell} above).\n"
1822 "\n"
1823 "On most systems a file can be extended by giving a length\n"
1824 "greater than the current size, but this is not mandatory in the\n"
1825 "POSIX standard.")
1826 #define FUNC_NAME s_scm_truncate_file
1827 {
1828 int rv;
1829
1830 /* "object" can be a port, fdes or filename.
1831
1832 Negative "length" makes no sense, but it's left to truncate() or
1833 ftruncate() to give back an error for that (normally EINVAL).
1834 */
1835
1836 if (SCM_UNBNDP (length))
1837 {
1838 /* must supply length if object is a filename. */
1839 if (scm_is_string (object))
1840 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL);
1841
1842 length = scm_seek (object, SCM_INUM0, scm_from_int (SEEK_CUR));
1843 }
1844
1845 object = SCM_COERCE_OUTPORT (object);
1846 if (scm_is_integer (object))
1847 {
1848 off_t_or_off64_t c_length = scm_to_off_t_or_off64_t (length);
1849 SCM_SYSCALL (rv = ftruncate_or_ftruncate64 (scm_to_int (object),
1850 c_length));
1851 }
1852 else if (SCM_OPOUTPORTP (object))
1853 {
1854 off_t_or_off64_t c_length = scm_to_off_t_or_off64_t (length);
1855 scm_t_port *pt = SCM_PTAB_ENTRY (object);
1856 scm_t_ptob_descriptor *ptob = scm_ptobs + SCM_PTOBNUM (object);
1857
1858 if (!ptob->truncate)
1859 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL);
1860 if (pt->rw_active == SCM_PORT_READ)
1861 scm_end_input (object);
1862 else if (pt->rw_active == SCM_PORT_WRITE)
1863 ptob->flush (object);
1864
1865 ptob->truncate (object, c_length);
1866 rv = 0;
1867 }
1868 else
1869 {
1870 off_t_or_off64_t c_length = scm_to_off_t_or_off64_t (length);
1871 char *str = scm_to_locale_string (object);
1872 int eno;
1873 SCM_SYSCALL (rv = truncate_or_truncate64 (str, c_length));
1874 eno = errno;
1875 free (str);
1876 errno = eno;
1877 }
1878 if (rv == -1)
1879 SCM_SYSERROR;
1880 return SCM_UNSPECIFIED;
1881 }
1882 #undef FUNC_NAME
1883
1884 SCM_DEFINE (scm_port_line, "port-line", 1, 0, 0,
1885 (SCM port),
1886 "Return the current line number for @var{port}.\n"
1887 "\n"
1888 "The first line of a file is 0. But you might want to add 1\n"
1889 "when printing line numbers, since starting from 1 is\n"
1890 "traditional in error messages, and likely to be more natural to\n"
1891 "non-programmers.")
1892 #define FUNC_NAME s_scm_port_line
1893 {
1894 port = SCM_COERCE_OUTPORT (port);
1895 SCM_VALIDATE_OPENPORT (1, port);
1896 return scm_from_long (SCM_LINUM (port));
1897 }
1898 #undef FUNC_NAME
1899
1900 SCM_DEFINE (scm_set_port_line_x, "set-port-line!", 2, 0, 0,
1901 (SCM port, SCM line),
1902 "Set the current line number for @var{port} to @var{line}. The\n"
1903 "first line of a file is 0.")
1904 #define FUNC_NAME s_scm_set_port_line_x
1905 {
1906 port = SCM_COERCE_OUTPORT (port);
1907 SCM_VALIDATE_OPENPORT (1, port);
1908 SCM_PTAB_ENTRY (port)->line_number = scm_to_long (line);
1909 return SCM_UNSPECIFIED;
1910 }
1911 #undef FUNC_NAME
1912
1913 SCM_DEFINE (scm_port_column, "port-column", 1, 0, 0,
1914 (SCM port),
1915 "Return the current column number of @var{port}.\n"
1916 "If the number is\n"
1917 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
1918 "- i.e. the first character of the first line is line 0, column 0.\n"
1919 "(However, when you display a file position, for example in an error\n"
1920 "message, we recommend you add 1 to get 1-origin integers. This is\n"
1921 "because lines and column numbers traditionally start with 1, and that is\n"
1922 "what non-programmers will find most natural.)")
1923 #define FUNC_NAME s_scm_port_column
1924 {
1925 port = SCM_COERCE_OUTPORT (port);
1926 SCM_VALIDATE_OPENPORT (1, port);
1927 return scm_from_int (SCM_COL (port));
1928 }
1929 #undef FUNC_NAME
1930
1931 SCM_DEFINE (scm_set_port_column_x, "set-port-column!", 2, 0, 0,
1932 (SCM port, SCM column),
1933 "Set the current column of @var{port}. Before reading the first\n"
1934 "character on a line the column should be 0.")
1935 #define FUNC_NAME s_scm_set_port_column_x
1936 {
1937 port = SCM_COERCE_OUTPORT (port);
1938 SCM_VALIDATE_OPENPORT (1, port);
1939 SCM_PTAB_ENTRY (port)->column_number = scm_to_int (column);
1940 return SCM_UNSPECIFIED;
1941 }
1942 #undef FUNC_NAME
1943
1944 SCM_DEFINE (scm_port_filename, "port-filename", 1, 0, 0,
1945 (SCM port),
1946 "Return the filename associated with @var{port}, or @code{#f}\n"
1947 "if no filename is associated with the port.")
1948 #define FUNC_NAME s_scm_port_filename
1949 {
1950 port = SCM_COERCE_OUTPORT (port);
1951 SCM_VALIDATE_OPENPORT (1, port);
1952 return SCM_FILENAME (port);
1953 }
1954 #undef FUNC_NAME
1955
1956 SCM_DEFINE (scm_set_port_filename_x, "set-port-filename!", 2, 0, 0,
1957 (SCM port, SCM filename),
1958 "Change the filename associated with @var{port}, using the current input\n"
1959 "port if none is specified. Note that this does not change the port's\n"
1960 "source of data, but only the value that is returned by\n"
1961 "@code{port-filename} and reported in diagnostic output.")
1962 #define FUNC_NAME s_scm_set_port_filename_x
1963 {
1964 port = SCM_COERCE_OUTPORT (port);
1965 SCM_VALIDATE_OPENPORT (1, port);
1966 /* We allow the user to set the filename to whatever he likes. */
1967 SCM_SET_FILENAME (port, filename);
1968 return SCM_UNSPECIFIED;
1969 }
1970 #undef FUNC_NAME
1971
1972 /* A fluid specifying the default encoding for newly created ports. If it is
1973 a string, that is the encoding. If it is #f, it is in the "native"
1974 (Latin-1) encoding. */
1975 SCM_VARIABLE (default_port_encoding_var, "%default-port-encoding");
1976
1977 static int scm_port_encoding_init = 0;
1978
1979 /* Use ENCODING as the default encoding for future ports. */
1980 void
1981 scm_i_set_default_port_encoding (const char *encoding)
1982 {
1983 if (!scm_port_encoding_init
1984 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var)))
1985 scm_misc_error (NULL, "tried to set port encoding fluid before it is initialized",
1986 SCM_EOL);
1987
1988 if (encoding == NULL
1989 || !strcmp (encoding, "ASCII")
1990 || !strcmp (encoding, "ANSI_X3.4-1968")
1991 || !strcmp (encoding, "ISO-8859-1"))
1992 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var), SCM_BOOL_F);
1993 else
1994 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var),
1995 scm_from_locale_string (encoding));
1996 }
1997
1998 /* Return the name of the default encoding for newly created ports; a
1999 return value of NULL means "ISO-8859-1". */
2000 const char *
2001 scm_i_default_port_encoding (void)
2002 {
2003 if (!scm_port_encoding_init)
2004 return NULL;
2005 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var)))
2006 return NULL;
2007 else
2008 {
2009 SCM encoding;
2010
2011 encoding = scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var));
2012 if (!scm_is_string (encoding))
2013 return NULL;
2014 else
2015 return scm_i_string_chars (encoding);
2016 }
2017 }
2018
2019 void
2020 scm_i_set_port_encoding_x (SCM port, const char *encoding)
2021 {
2022 scm_t_port *pt;
2023 iconv_t new_input_cd, new_output_cd;
2024
2025 new_input_cd = (iconv_t) -1;
2026 new_output_cd = (iconv_t) -1;
2027
2028 /* Set the character encoding for this port. */
2029 pt = SCM_PTAB_ENTRY (port);
2030
2031 if (encoding == NULL)
2032 encoding = "ISO-8859-1";
2033
2034 pt->encoding = scm_gc_strdup (encoding, "port");
2035
2036 if (SCM_CELL_WORD_0 (port) & SCM_RDNG)
2037 {
2038 /* Open an input iconv conversion descriptor, from ENCODING
2039 to UTF-8. We choose UTF-8, not UTF-32, because iconv
2040 implementations can typically convert from anything to
2041 UTF-8, but not to UTF-32 (see
2042 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
2043 new_input_cd = iconv_open ("UTF-8", encoding);
2044 if (new_input_cd == (iconv_t) -1)
2045 goto invalid_encoding;
2046 }
2047
2048 if (SCM_CELL_WORD_0 (port) & SCM_WRTNG)
2049 {
2050 new_output_cd = iconv_open (encoding, "UTF-8");
2051 if (new_output_cd == (iconv_t) -1)
2052 {
2053 if (new_input_cd != (iconv_t) -1)
2054 iconv_close (new_input_cd);
2055 goto invalid_encoding;
2056 }
2057 }
2058
2059 if (pt->input_cd != (iconv_t) -1)
2060 iconv_close (pt->input_cd);
2061 if (pt->output_cd != (iconv_t) -1)
2062 iconv_close (pt->output_cd);
2063
2064 pt->input_cd = new_input_cd;
2065 pt->output_cd = new_output_cd;
2066
2067 return;
2068
2069 invalid_encoding:
2070 {
2071 SCM err;
2072 err = scm_from_locale_string (encoding);
2073 scm_misc_error ("scm_i_set_port_encoding_x",
2074 "invalid or unknown character encoding ~s",
2075 scm_list_1 (err));
2076 }
2077 }
2078
2079 SCM_DEFINE (scm_port_encoding, "port-encoding", 1, 0, 0,
2080 (SCM port),
2081 "Returns, as a string, the character encoding that @var{port}\n"
2082 "uses to interpret its input and output.\n")
2083 #define FUNC_NAME s_scm_port_encoding
2084 {
2085 scm_t_port *pt;
2086 const char *enc;
2087
2088 SCM_VALIDATE_PORT (1, port);
2089
2090 pt = SCM_PTAB_ENTRY (port);
2091 enc = pt->encoding;
2092 if (enc)
2093 return scm_from_locale_string (pt->encoding);
2094 else
2095 return SCM_BOOL_F;
2096 }
2097 #undef FUNC_NAME
2098
2099 SCM_DEFINE (scm_set_port_encoding_x, "set-port-encoding!", 2, 0, 0,
2100 (SCM port, SCM enc),
2101 "Sets the character encoding that will be used to interpret all\n"
2102 "port I/O. New ports are created with the encoding\n"
2103 "appropriate for the current locale if @code{setlocale} has \n"
2104 "been called or ISO-8859-1 otherwise\n"
2105 "and this procedure can be used to modify that encoding.\n")
2106 #define FUNC_NAME s_scm_set_port_encoding_x
2107 {
2108 char *enc_str;
2109
2110 SCM_VALIDATE_PORT (1, port);
2111 SCM_VALIDATE_STRING (2, enc);
2112
2113 enc_str = scm_to_locale_string (enc);
2114 scm_i_set_port_encoding_x (port, enc_str);
2115 free (enc_str);
2116
2117 return SCM_UNSPECIFIED;
2118 }
2119 #undef FUNC_NAME
2120
2121
2122 /* This determines how conversions handle unconvertible characters. */
2123 SCM_GLOBAL_VARIABLE (scm_conversion_strategy, "%port-conversion-strategy");
2124 static int scm_conversion_strategy_init = 0;
2125
2126 scm_t_string_failed_conversion_handler
2127 scm_i_get_conversion_strategy (SCM port)
2128 {
2129 SCM encoding;
2130
2131 if (scm_is_false (port))
2132 {
2133 if (!scm_conversion_strategy_init
2134 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy)))
2135 return SCM_FAILED_CONVERSION_QUESTION_MARK;
2136 else
2137 {
2138 encoding = scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy));
2139 if (scm_is_false (encoding))
2140 return SCM_FAILED_CONVERSION_QUESTION_MARK;
2141 else
2142 return (scm_t_string_failed_conversion_handler) scm_to_int (encoding);
2143 }
2144 }
2145 else
2146 {
2147 scm_t_port *pt;
2148 pt = SCM_PTAB_ENTRY (port);
2149 return pt->ilseq_handler;
2150 }
2151
2152 }
2153
2154 void
2155 scm_i_set_conversion_strategy_x (SCM port,
2156 scm_t_string_failed_conversion_handler handler)
2157 {
2158 SCM strategy;
2159 scm_t_port *pt;
2160
2161 strategy = scm_from_int ((int) handler);
2162
2163 if (scm_is_false (port))
2164 {
2165 /* Set the default encoding for future ports. */
2166 if (!scm_conversion_strategy
2167 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy)))
2168 scm_misc_error (NULL, "tried to set conversion strategy fluid before it is initialized",
2169 SCM_EOL);
2170 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy), strategy);
2171 }
2172 else
2173 {
2174 /* Set the character encoding for this port. */
2175 pt = SCM_PTAB_ENTRY (port);
2176 pt->ilseq_handler = handler;
2177 }
2178 }
2179
2180 SCM_DEFINE (scm_port_conversion_strategy, "port-conversion-strategy",
2181 1, 0, 0, (SCM port),
2182 "Returns the behavior of the port when handling a character that\n"
2183 "is not representable in the port's current encoding.\n"
2184 "It returns the symbol @code{error} if unrepresentable characters\n"
2185 "should cause exceptions, @code{substitute} if the port should\n"
2186 "try to replace unrepresentable characters with question marks or\n"
2187 "approximate characters, or @code{escape} if unrepresentable\n"
2188 "characters should be converted to string escapes.\n"
2189 "\n"
2190 "If @var{port} is @code{#f}, then the current default behavior\n"
2191 "will be returned. New ports will have this default behavior\n"
2192 "when they are created.\n")
2193 #define FUNC_NAME s_scm_port_conversion_strategy
2194 {
2195 scm_t_string_failed_conversion_handler h;
2196
2197 SCM_VALIDATE_OPPORT (1, port);
2198
2199 if (!scm_is_false (port))
2200 {
2201 SCM_VALIDATE_OPPORT (1, port);
2202 }
2203
2204 h = scm_i_get_conversion_strategy (port);
2205 if (h == SCM_FAILED_CONVERSION_ERROR)
2206 return scm_from_latin1_symbol ("error");
2207 else if (h == SCM_FAILED_CONVERSION_QUESTION_MARK)
2208 return scm_from_latin1_symbol ("substitute");
2209 else if (h == SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE)
2210 return scm_from_latin1_symbol ("escape");
2211 else
2212 abort ();
2213
2214 /* Never gets here. */
2215 return SCM_UNDEFINED;
2216 }
2217 #undef FUNC_NAME
2218
2219 SCM_DEFINE (scm_set_port_conversion_strategy_x, "set-port-conversion-strategy!",
2220 2, 0, 0,
2221 (SCM port, SCM sym),
2222 "Sets the behavior of the interpreter when outputting a character\n"
2223 "that is not representable in the port's current encoding.\n"
2224 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
2225 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
2226 "when an unconvertible character is encountered. If it is\n"
2227 "@code{'substitute}, then unconvertible characters will \n"
2228 "be replaced with approximate characters, or with question marks\n"
2229 "if no approximately correct character is available.\n"
2230 "If it is @code{'escape},\n"
2231 "it will appear as a hex escape when output.\n"
2232 "\n"
2233 "If @var{port} is an open port, the conversion error behavior\n"
2234 "is set for that port. If it is @code{#f}, it is set as the\n"
2235 "default behavior for any future ports that get created in\n"
2236 "this thread.\n")
2237 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
2238 {
2239 SCM err;
2240 SCM qm;
2241 SCM esc;
2242
2243 if (!scm_is_false (port))
2244 {
2245 SCM_VALIDATE_OPPORT (1, port);
2246 }
2247
2248 err = scm_from_latin1_symbol ("error");
2249 if (scm_is_true (scm_eqv_p (sym, err)))
2250 {
2251 scm_i_set_conversion_strategy_x (port, SCM_FAILED_CONVERSION_ERROR);
2252 return SCM_UNSPECIFIED;
2253 }
2254
2255 qm = scm_from_latin1_symbol ("substitute");
2256 if (scm_is_true (scm_eqv_p (sym, qm)))
2257 {
2258 scm_i_set_conversion_strategy_x (port,
2259 SCM_FAILED_CONVERSION_QUESTION_MARK);
2260 return SCM_UNSPECIFIED;
2261 }
2262
2263 esc = scm_from_latin1_symbol ("escape");
2264 if (scm_is_true (scm_eqv_p (sym, esc)))
2265 {
2266 scm_i_set_conversion_strategy_x (port,
2267 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE);
2268 return SCM_UNSPECIFIED;
2269 }
2270
2271 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym));
2272
2273 return SCM_UNSPECIFIED;
2274 }
2275 #undef FUNC_NAME
2276
2277
2278
2279 void
2280 scm_print_port_mode (SCM exp, SCM port)
2281 {
2282 scm_puts (SCM_CLOSEDP (exp)
2283 ? "closed: "
2284 : (SCM_RDNG & SCM_CELL_WORD_0 (exp)
2285 ? (SCM_WRTNG & SCM_CELL_WORD_0 (exp)
2286 ? "input-output: "
2287 : "input: ")
2288 : (SCM_WRTNG & SCM_CELL_WORD_0 (exp)
2289 ? "output: "
2290 : "bogus: ")),
2291 port);
2292 }
2293
2294 int
2295 scm_port_print (SCM exp, SCM port, scm_print_state *pstate SCM_UNUSED)
2296 {
2297 char *type = SCM_PTOBNAME (SCM_PTOBNUM (exp));
2298 if (!type)
2299 type = "port";
2300 scm_puts ("#<", port);
2301 scm_print_port_mode (exp, port);
2302 scm_puts (type, port);
2303 scm_putc (' ', port);
2304 scm_uintprint (SCM_CELL_WORD_1 (exp), 16, port);
2305 scm_putc ('>', port);
2306 return 1;
2307 }
2308
2309 \f
2310
2311 /* Void ports. */
2312
2313 scm_t_bits scm_tc16_void_port = 0;
2314
2315 static int fill_input_void_port (SCM port SCM_UNUSED)
2316 {
2317 return EOF;
2318 }
2319
2320 static void
2321 write_void_port (SCM port SCM_UNUSED,
2322 const void *data SCM_UNUSED,
2323 size_t size SCM_UNUSED)
2324 {
2325 }
2326
2327 static SCM
2328 scm_i_void_port (long mode_bits)
2329 {
2330 scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex);
2331 {
2332 SCM answer = scm_new_port_table_entry (scm_tc16_void_port);
2333 scm_t_port * pt = SCM_PTAB_ENTRY(answer);
2334
2335 scm_port_non_buffer (pt);
2336
2337 SCM_SETSTREAM (answer, 0);
2338 SCM_SET_CELL_TYPE (answer, scm_tc16_void_port | mode_bits);
2339 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex);
2340 return answer;
2341 }
2342 }
2343
2344 SCM
2345 scm_void_port (char *mode_str)
2346 {
2347 return scm_i_void_port (scm_mode_bits (mode_str));
2348 }
2349
2350 SCM_DEFINE (scm_sys_make_void_port, "%make-void-port", 1, 0, 0,
2351 (SCM mode),
2352 "Create and return a new void port. A void port acts like\n"
2353 "@file{/dev/null}. The @var{mode} argument\n"
2354 "specifies the input/output modes for this port: see the\n"
2355 "documentation for @code{open-file} in @ref{File Ports}.")
2356 #define FUNC_NAME s_scm_sys_make_void_port
2357 {
2358 return scm_i_void_port (scm_i_mode_bits (mode));
2359 }
2360 #undef FUNC_NAME
2361
2362 \f
2363 /* Initialization. */
2364
2365 void
2366 scm_init_ports ()
2367 {
2368 /* lseek() symbols. */
2369 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET));
2370 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR));
2371 scm_c_define ("SEEK_END", scm_from_int (SEEK_END));
2372
2373 scm_tc16_void_port = scm_make_port_type ("void", fill_input_void_port,
2374 write_void_port);
2375
2376 cur_inport_fluid = scm_make_fluid ();
2377 cur_outport_fluid = scm_make_fluid ();
2378 cur_errport_fluid = scm_make_fluid ();
2379 cur_loadport_fluid = scm_make_fluid ();
2380
2381 scm_i_port_weak_hash = scm_make_weak_key_hash_table (SCM_I_MAKINUM(31));
2382
2383 #include "libguile/ports.x"
2384
2385 /* Use Latin-1 as the default port encoding. */
2386 SCM_VARIABLE_SET (default_port_encoding_var, scm_make_fluid ());
2387 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var), SCM_BOOL_F);
2388 scm_port_encoding_init = 1;
2389
2390 SCM_VARIABLE_SET (scm_conversion_strategy, scm_make_fluid ());
2391 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy),
2392 scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK));
2393 scm_conversion_strategy_init = 1;
2394
2395 }
2396
2397 /*
2398 Local Variables:
2399 c-file-style: "gnu"
2400 End:
2401 */