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