1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004, 2006,
2 * 2007, 2008, 2009, 2010, 2011, 2012, 2013 Free Software Foundation, Inc.
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.
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.
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
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
32 #include <fcntl.h> /* for chsize on mingw */
37 #include <striconveh.h>
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"
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/ports-internal.h"
59 #include "libguile/vectors.h"
60 #include "libguile/weaks.h"
61 #include "libguile/fluids.h"
62 #include "libguile/eq.h"
76 #ifdef HAVE_SYS_IOCTL_H
77 #include <sys/ioctl.h>
80 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
81 already, but have this code here in case that wasn't so in past versions,
82 or perhaps to help other minimal DOS environments.
84 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
85 might be possibilities if we've got other systems without ftruncate. */
87 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
88 #define ftruncate(fd, size) chsize (fd, size)
90 #define HAVE_FTRUNCATE 1
94 /* The port kind table --- a dynamically resized array of port types. */
97 /* scm_ptobs scm_numptob
98 * implement a dynamically resized array of ptob records.
99 * Indexes into this table are used when generating type
100 * tags for smobjects (if you know a tag you can get an index and conversely).
102 scm_t_ptob_descriptor
*scm_ptobs
= NULL
;
103 long scm_numptob
= 0;
105 /* GC marker for a port with stream of SCM type. */
107 scm_markstream (SCM ptr
)
110 openp
= SCM_CELL_WORD_0 (ptr
) & SCM_OPN
;
112 return SCM_PACK (SCM_STREAM (ptr
));
118 * We choose to use an interface similar to the smob interface with
119 * fill_input and write as standard fields, passed to the port
120 * type constructor, and optional fields set by setters.
124 flush_port_default (SCM port SCM_UNUSED
)
129 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
134 scm_make_port_type (char *name
,
135 int (*fill_input
) (SCM port
),
136 void (*write
) (SCM port
, const void *data
, size_t size
))
139 if (SCM_I_MAX_PORT_TYPE_COUNT
- 1 <= scm_numptob
)
141 SCM_CRITICAL_SECTION_START
;
142 tmp
= (char *) scm_gc_realloc ((char *) scm_ptobs
,
143 scm_numptob
* sizeof (scm_t_ptob_descriptor
),
145 * sizeof (scm_t_ptob_descriptor
),
149 scm_ptobs
= (scm_t_ptob_descriptor
*) tmp
;
151 scm_ptobs
[scm_numptob
].name
= name
;
152 scm_ptobs
[scm_numptob
].mark
= 0;
153 scm_ptobs
[scm_numptob
].free
= NULL
;
154 scm_ptobs
[scm_numptob
].print
= scm_port_print
;
155 scm_ptobs
[scm_numptob
].equalp
= 0;
156 scm_ptobs
[scm_numptob
].close
= 0;
158 scm_ptobs
[scm_numptob
].write
= write
;
159 scm_ptobs
[scm_numptob
].flush
= flush_port_default
;
161 scm_ptobs
[scm_numptob
].end_input
= end_input_default
;
162 scm_ptobs
[scm_numptob
].fill_input
= fill_input
;
163 scm_ptobs
[scm_numptob
].input_waiting
= 0;
165 scm_ptobs
[scm_numptob
].seek
= 0;
166 scm_ptobs
[scm_numptob
].truncate
= 0;
170 SCM_CRITICAL_SECTION_END
;
174 scm_memory_error ("scm_make_port_type");
176 /* Make a class object if Goops is present */
177 if (SCM_UNPACK (scm_port_class
[0]) != 0)
178 scm_make_port_classes (scm_numptob
- 1, SCM_PTOBNAME (scm_numptob
- 1));
179 return scm_tc7_port
+ (scm_numptob
- 1) * 256;
183 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
185 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].mark
= mark
;
189 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
191 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].free
= free
;
195 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
196 scm_print_state
*pstate
))
198 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].print
= print
;
202 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
204 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].equalp
= equalp
;
208 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
210 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].flush
= flush
;
214 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
216 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].end_input
= end_input
;
220 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
222 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].close
= close
;
226 scm_set_port_seek (scm_t_bits tc
,
227 scm_t_off (*seek
) (SCM
, scm_t_off
, int))
229 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].seek
= seek
;
233 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
235 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].truncate
= truncate
;
239 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
241 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].input_waiting
= input_waiting
;
245 scm_i_port_alist (SCM port
)
247 return SCM_PORT_GET_INTERNAL (port
)->alist
;
251 scm_i_set_port_alist_x (SCM port
, SCM alist
)
253 SCM_PORT_GET_INTERNAL (port
)->alist
= alist
;
258 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
260 "Return @code{#t} if a character is ready on input @var{port}\n"
261 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
262 "@code{#t} then the next @code{read-char} operation on\n"
263 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
264 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
266 "@code{char-ready?} exists to make it possible for a\n"
267 "program to accept characters from interactive ports without\n"
268 "getting stuck waiting for input. Any input editors associated\n"
269 "with such ports must make sure that characters whose existence\n"
270 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
271 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
272 "a port at end of file would be indistinguishable from an\n"
273 "interactive port that has no ready characters.")
274 #define FUNC_NAME s_scm_char_ready_p
278 if (SCM_UNBNDP (port
))
279 port
= scm_current_input_port ();
280 /* It's possible to close the current input port, so validate even in
282 SCM_VALIDATE_OPINPORT (1, port
);
284 pt
= SCM_PTAB_ENTRY (port
);
286 /* if the current read buffer is filled, or the
287 last pushed-back char has been read and the saved buffer is
288 filled, result is true. */
289 if (pt
->read_pos
< pt
->read_end
290 || (pt
->read_buf
== pt
->putback_buf
291 && pt
->saved_read_pos
< pt
->saved_read_end
))
295 scm_t_ptob_descriptor
*ptob
= &scm_ptobs
[SCM_PTOBNUM (port
)];
297 if (ptob
->input_waiting
)
298 return scm_from_bool(ptob
->input_waiting (port
));
305 /* Move up to READ_LEN bytes from PORT's putback and/or read buffers
306 into memory starting at DEST. Return the number of bytes moved.
307 PORT's line/column numbers are left unchanged. */
309 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
311 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
312 size_t bytes_read
= 0;
313 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
317 memcpy (dest
, pt
->read_pos
, from_buf
);
318 pt
->read_pos
+= from_buf
;
319 bytes_read
+= from_buf
;
320 read_len
-= from_buf
;
324 /* if putback was active, try the real input buffer too. */
325 if (pt
->read_buf
== pt
->putback_buf
)
327 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
330 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
331 pt
->saved_read_pos
+= from_buf
;
332 bytes_read
+= from_buf
;
339 /* Clear a port's read buffers, returning the contents. */
340 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
342 "This procedure clears a port's input buffers, similar\n"
343 "to the way that force-output clears the output buffer. The\n"
344 "contents of the buffers are returned as a single string, e.g.,\n"
347 "(define p (open-input-file ...))\n"
348 "(drain-input p) => empty string, nothing buffered yet.\n"
349 "(unread-char (read-char p) p)\n"
350 "(drain-input p) => initial chars from p, up to the buffer size.\n"
352 "Draining the buffers may be useful for cleanly finishing\n"
353 "buffered I/O so that the file descriptor can be used directly\n"
354 "for further input.")
355 #define FUNC_NAME s_scm_drain_input
362 SCM_VALIDATE_OPINPORT (1, port
);
363 pt
= SCM_PTAB_ENTRY (port
);
365 count
= pt
->read_end
- pt
->read_pos
;
366 if (pt
->read_buf
== pt
->putback_buf
)
367 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
371 result
= scm_i_make_string (count
, &data
, 0);
372 scm_take_from_input_buffers (port
, data
, count
);
375 result
= scm_nullstr
;
382 /* Standard ports --- current input, output, error, and more(!). */
384 static SCM cur_inport_fluid
= SCM_BOOL_F
;
385 static SCM cur_outport_fluid
= SCM_BOOL_F
;
386 static SCM cur_errport_fluid
= SCM_BOOL_F
;
387 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
389 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
391 "Return the current input port. This is the default port used\n"
392 "by many input procedures. Initially, @code{current-input-port}\n"
393 "returns the @dfn{standard input} in Unix and C terminology.")
394 #define FUNC_NAME s_scm_current_input_port
396 if (scm_is_true (cur_inport_fluid
))
397 return scm_fluid_ref (cur_inport_fluid
);
403 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
405 "Return the current output port. This is the default port used\n"
406 "by many output procedures. Initially,\n"
407 "@code{current-output-port} returns the @dfn{standard output} in\n"
408 "Unix and C terminology.")
409 #define FUNC_NAME s_scm_current_output_port
411 if (scm_is_true (cur_outport_fluid
))
412 return scm_fluid_ref (cur_outport_fluid
);
418 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
420 "Return the port to which errors and warnings should be sent (the\n"
421 "@dfn{standard error} in Unix and C terminology).")
422 #define FUNC_NAME s_scm_current_error_port
424 if (scm_is_true (cur_errport_fluid
))
425 return scm_fluid_ref (cur_errport_fluid
);
432 scm_current_warning_port (void)
434 static SCM cwp_var
= SCM_UNDEFINED
;
435 static scm_i_pthread_mutex_t cwp_var_mutex
436 = SCM_I_PTHREAD_MUTEX_INITIALIZER
;
438 scm_i_scm_pthread_mutex_lock (&cwp_var_mutex
);
439 if (SCM_UNBNDP (cwp_var
))
440 cwp_var
= scm_c_private_variable ("guile", "current-warning-port");
441 scm_i_pthread_mutex_unlock (&cwp_var_mutex
);
443 return scm_call_0 (scm_variable_ref (cwp_var
));
446 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
448 "Return the current-load-port.\n"
449 "The load port is used internally by @code{primitive-load}.")
450 #define FUNC_NAME s_scm_current_load_port
452 return scm_fluid_ref (cur_loadport_fluid
);
456 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
458 "@deffnx {Scheme Procedure} set-current-output-port port\n"
459 "@deffnx {Scheme Procedure} set-current-error-port port\n"
460 "Change the ports returned by @code{current-input-port},\n"
461 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
462 "so that they use the supplied @var{port} for input or output.")
463 #define FUNC_NAME s_scm_set_current_input_port
465 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
466 SCM_VALIDATE_OPINPORT (1, port
);
467 scm_fluid_set_x (cur_inport_fluid
, port
);
473 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
475 "Set the current default output port to @var{port}.")
476 #define FUNC_NAME s_scm_set_current_output_port
478 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
479 port
= SCM_COERCE_OUTPORT (port
);
480 SCM_VALIDATE_OPOUTPORT (1, port
);
481 scm_fluid_set_x (cur_outport_fluid
, port
);
487 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
489 "Set the current default error port to @var{port}.")
490 #define FUNC_NAME s_scm_set_current_error_port
492 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
493 port
= SCM_COERCE_OUTPORT (port
);
494 SCM_VALIDATE_OPOUTPORT (1, port
);
495 scm_fluid_set_x (cur_errport_fluid
, port
);
502 scm_set_current_warning_port (SCM port
)
504 static SCM cwp_var
= SCM_BOOL_F
;
506 if (scm_is_false (cwp_var
))
507 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
509 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
514 scm_dynwind_current_input_port (SCM port
)
515 #define FUNC_NAME NULL
517 SCM_VALIDATE_OPINPORT (1, port
);
518 scm_dynwind_fluid (cur_inport_fluid
, port
);
523 scm_dynwind_current_output_port (SCM port
)
524 #define FUNC_NAME NULL
526 port
= SCM_COERCE_OUTPORT (port
);
527 SCM_VALIDATE_OPOUTPORT (1, port
);
528 scm_dynwind_fluid (cur_outport_fluid
, port
);
533 scm_dynwind_current_error_port (SCM port
)
534 #define FUNC_NAME NULL
536 port
= SCM_COERCE_OUTPORT (port
);
537 SCM_VALIDATE_OPOUTPORT (1, port
);
538 scm_dynwind_fluid (cur_errport_fluid
, port
);
543 scm_i_dynwind_current_load_port (SCM port
)
545 scm_dynwind_fluid (cur_loadport_fluid
, port
);
549 /* The port table --- an array of pointers to ports. */
552 We need a global registry of ports to flush them all at exit, and to
553 get all the ports matching a file descriptor.
555 SCM scm_i_port_weak_hash
;
557 scm_i_pthread_mutex_t scm_i_port_table_mutex
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
560 /* Port finalization. */
563 static void finalize_port (void *, void *);
565 /* Register a finalizer for PORT. */
566 static SCM_C_INLINE_KEYWORD
void
567 register_finalizer_for_port (SCM port
)
569 /* Register a finalizer for PORT so that its
570 type's `free' function gets called. */
571 scm_i_set_finalizer (SCM2PTR (port
), finalize_port
, NULL
);
574 /* Finalize the object (a port) pointed to by PTR. */
576 finalize_port (void *ptr
, void *data
)
579 SCM port
= PTR2SCM (ptr
);
581 if (!SCM_PORTP (port
))
584 if (SCM_OPENP (port
))
586 if (SCM_REVEALED (port
) > 0)
587 /* Keep "revealed" ports alive and re-register a finalizer. */
588 register_finalizer_for_port (port
);
591 port_type
= SCM_TC2PTOBNUM (SCM_CELL_TYPE (port
));
592 if (port_type
>= scm_numptob
)
595 if (scm_ptobs
[port_type
].free
)
596 /* Yes, I really do mean `.free' rather than `.close'. `.close'
597 is for explicit `close-port' by user. */
598 scm_ptobs
[port_type
].free (port
);
600 SCM_SETSTREAM (port
, 0);
601 SCM_CLR_PORT_OPEN_FLAG (port
);
603 scm_gc_ports_collected
++;
612 /* This function is not and should not be thread safe. */
614 scm_new_port_table_entry (scm_t_bits tag
)
615 #define FUNC_NAME "scm_new_port_table_entry"
618 We initialize the cell to empty, this is in case scm_gc_calloc
619 triggers GC ; we don't want the GC to scan a half-finished Z.
622 SCM z
= scm_cons (SCM_EOL
, SCM_EOL
);
623 scm_t_port
*entry
= scm_gc_typed_calloc (scm_t_port
);
624 scm_t_port_internal
*pti
= scm_gc_typed_calloc (scm_t_port_internal
);
625 const char *encoding
;
627 entry
->file_name
= SCM_BOOL_F
;
628 entry
->rw_active
= SCM_PORT_NEITHER
;
631 /* Initialize this port with the thread's current default
633 encoding
= scm_i_default_port_encoding ();
634 entry
->ilseq_handler
= scm_i_default_port_conversion_handler ();
635 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
636 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
637 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
639 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
640 pti
->iconv_descriptors
= NULL
;
642 /* XXX These fields are not what they seem. They have been
643 repurposed, but cannot safely be renamed in 2.0 without breaking
644 ABI compatibility. This will be cleaned up in 2.2. */
645 entry
->input_cd
= pti
; /* XXX pointer to the internal port structure */
646 entry
->output_cd
= NULL
; /* XXX unused */
648 pti
->alist
= SCM_EOL
;
650 SCM_SET_CELL_TYPE (z
, tag
);
651 SCM_SETPTAB_ENTRY (z
, entry
);
653 scm_hashq_set_x (scm_i_port_weak_hash
, z
, SCM_BOOL_F
);
655 /* For each new port, register a finalizer so that it port type's free
656 function can be invoked eventually. */
657 register_finalizer_for_port (z
);
663 #if SCM_ENABLE_DEPRECATED==1
665 scm_add_to_port_table (SCM port
)
670 scm_c_issue_deprecation_warning ("scm_add_to_port_table is deprecated.");
672 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex
);
673 z
= scm_new_port_table_entry (scm_tc7_port
);
674 pt
= SCM_PTAB_ENTRY(z
);
676 SCM_SETCAR (z
, SCM_EOL
);
677 SCM_SETCDR (z
, SCM_EOL
);
678 SCM_SETPTAB_ENTRY (port
, pt
);
679 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
686 /* Remove a port from the table and destroy it. */
688 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
691 scm_i_remove_port (SCM port
)
692 #define FUNC_NAME "scm_remove_port"
695 scm_t_port_internal
*pti
;
697 scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex
);
699 p
= SCM_PTAB_ENTRY (port
);
700 pti
= SCM_PORT_GET_INTERNAL (port
);
701 scm_port_non_buffer (p
);
702 p
->putback_buf
= NULL
;
703 p
->putback_buf_size
= 0;
705 if (pti
->iconv_descriptors
)
707 close_iconv_descriptors (pti
->iconv_descriptors
);
708 pti
->iconv_descriptors
= NULL
;
711 SCM_SETPTAB_ENTRY (port
, 0);
713 scm_hashq_remove_x (scm_i_port_weak_hash
, port
);
715 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
720 /* Functions for debugging. */
722 SCM_DEFINE (scm_pt_size
, "pt-size", 0, 0, 0,
724 "Return the number of ports in the port table. @code{pt-size}\n"
725 "is only included in @code{--enable-guile-debug} builds.")
726 #define FUNC_NAME s_scm_pt_size
728 return scm_from_int (SCM_HASHTABLE_N_ITEMS (scm_i_port_weak_hash
));
734 scm_port_non_buffer (scm_t_port
*pt
)
736 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
737 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
738 pt
->read_buf_size
= pt
->write_buf_size
= 1;
739 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
743 /* Revealed counts --- an oddity inherited from SCSH. */
745 /* Find a port in the table and return its revealed count.
746 Also used by the garbage collector.
750 scm_revealed_count (SCM port
)
752 return SCM_REVEALED(port
);
757 /* Return the revealed count for a port. */
759 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
761 "Return the revealed count for @var{port}.")
762 #define FUNC_NAME s_scm_port_revealed
764 port
= SCM_COERCE_OUTPORT (port
);
765 SCM_VALIDATE_OPENPORT (1, port
);
766 return scm_from_int (scm_revealed_count (port
));
770 /* Set the revealed count for a port. */
771 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
772 (SCM port
, SCM rcount
),
773 "Sets the revealed count for a port to a given value.\n"
774 "The return value is unspecified.")
775 #define FUNC_NAME s_scm_set_port_revealed_x
777 port
= SCM_COERCE_OUTPORT (port
);
778 SCM_VALIDATE_OPENPORT (1, port
);
779 SCM_REVEALED (port
) = scm_to_int (rcount
);
780 return SCM_UNSPECIFIED
;
786 /* Retrieving a port's mode. */
788 /* Return the flags that characterize a port based on the mode
789 * string used to open a file for that port.
791 * See PORT FLAGS in scm.h
795 scm_i_mode_bits_n (SCM modes
)
798 | (scm_i_string_contains_char (modes
, 'r')
799 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
800 | (scm_i_string_contains_char (modes
, 'w')
801 || scm_i_string_contains_char (modes
, 'a')
802 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
803 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
804 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
808 scm_mode_bits (char *modes
)
810 return scm_i_mode_bits (scm_from_locale_string (modes
));
814 scm_i_mode_bits (SCM modes
)
818 if (!scm_is_string (modes
))
819 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
821 bits
= scm_i_mode_bits_n (modes
);
822 scm_remember_upto_here_1 (modes
);
826 /* Return the mode flags from an open port.
827 * Some modes such as "append" are only used when opening
828 * a file and are not returned here. */
830 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
832 "Return the port modes associated with the open port @var{port}.\n"
833 "These will not necessarily be identical to the modes used when\n"
834 "the port was opened, since modes such as \"append\" which are\n"
835 "used only during port creation are not retained.")
836 #define FUNC_NAME s_scm_port_mode
841 port
= SCM_COERCE_OUTPORT (port
);
842 SCM_VALIDATE_OPPORT (1, port
);
843 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
844 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
845 strcpy (modes
, "r+");
849 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
851 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
853 return scm_from_locale_string (modes
);
862 * Call the close operation on a port object.
863 * see also scm_close.
865 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
867 "Close the specified port object. Return @code{#t} if it\n"
868 "successfully closes a port or @code{#f} if it was already\n"
869 "closed. An exception may be raised if an error occurs, for\n"
870 "example when flushing buffered output. See also @ref{Ports and\n"
871 "File Descriptors, close}, for a procedure which can close file\n"
873 #define FUNC_NAME s_scm_close_port
878 port
= SCM_COERCE_OUTPORT (port
);
880 SCM_VALIDATE_PORT (1, port
);
881 if (SCM_CLOSEDP (port
))
883 i
= SCM_PTOBNUM (port
);
884 if (scm_ptobs
[i
].close
)
885 rv
= (scm_ptobs
[i
].close
) (port
);
888 scm_i_remove_port (port
);
889 SCM_CLR_PORT_OPEN_FLAG (port
);
890 return scm_from_bool (rv
>= 0);
894 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
896 "Close the specified input port object. The routine has no effect if\n"
897 "the file has already been closed. An exception may be raised if an\n"
898 "error occurs. The value returned is unspecified.\n\n"
899 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
900 "which can close file descriptors.")
901 #define FUNC_NAME s_scm_close_input_port
903 SCM_VALIDATE_INPUT_PORT (1, port
);
904 scm_close_port (port
);
905 return SCM_UNSPECIFIED
;
909 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
911 "Close the specified output port object. The routine has no effect if\n"
912 "the file has already been closed. An exception may be raised if an\n"
913 "error occurs. The value returned is unspecified.\n\n"
914 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
915 "which can close file descriptors.")
916 #define FUNC_NAME s_scm_close_output_port
918 port
= SCM_COERCE_OUTPORT (port
);
919 SCM_VALIDATE_OUTPUT_PORT (1, port
);
920 scm_close_port (port
);
921 return SCM_UNSPECIFIED
;
926 collect_keys (void *unused
, SCM key
, SCM value
, SCM result
)
928 return scm_cons (key
, result
);
932 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
936 /* Copy out the port table as a list so that we get strong references
937 to all the values. */
938 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex
);
939 ports
= scm_internal_hash_fold (collect_keys
, NULL
,
940 SCM_EOL
, scm_i_port_weak_hash
);
941 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
943 for (; scm_is_pair (ports
); ports
= scm_cdr (ports
))
945 SCM p
= scm_car (ports
);
951 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
953 "Apply @var{proc} to each port in the Guile port table\n"
954 "in turn. The return value is unspecified. More specifically,\n"
955 "@var{proc} is applied exactly once to every port that exists\n"
956 "in the system at the time @code{port-for-each} is invoked.\n"
957 "Changes to the port table while @code{port-for-each} is running\n"
958 "have no effect as far as @code{port-for-each} is concerned.")
959 #define FUNC_NAME s_scm_port_for_each
963 SCM_VALIDATE_PROC (1, proc
);
965 /* Copy out the port table as a list so that we get strong references
966 to all the values. */
967 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex
);
968 ports
= scm_internal_hash_fold (collect_keys
, NULL
,
969 SCM_EOL
, scm_i_port_weak_hash
);
970 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
972 for (; scm_is_pair (ports
); ports
= scm_cdr (ports
))
973 if (SCM_PORTP (SCM_CAR (ports
)))
974 scm_call_1 (proc
, SCM_CAR (ports
));
976 return SCM_UNSPECIFIED
;
982 /* Utter miscellany. Gosh, we should clean this up some time. */
984 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
986 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
987 "@code{#f}. Any object satisfying this predicate also satisfies\n"
989 #define FUNC_NAME s_scm_input_port_p
991 return scm_from_bool (SCM_INPUT_PORT_P (x
));
995 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
997 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
998 "@code{#f}. Any object satisfying this predicate also satisfies\n"
1000 #define FUNC_NAME s_scm_output_port_p
1002 x
= SCM_COERCE_OUTPORT (x
);
1003 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
1007 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
1009 "Return a boolean indicating whether @var{x} is a port.\n"
1010 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
1012 #define FUNC_NAME s_scm_port_p
1014 return scm_from_bool (SCM_PORTP (x
));
1018 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
1020 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
1022 #define FUNC_NAME s_scm_port_closed_p
1024 SCM_VALIDATE_PORT (1, port
);
1025 return scm_from_bool (!SCM_OPPORTP (port
));
1029 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
1031 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
1032 "return @code{#f}.")
1033 #define FUNC_NAME s_scm_eof_object_p
1035 return scm_from_bool(SCM_EOF_OBJECT_P (x
));
1039 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
1041 "Flush the specified output port, or the current output port if @var{port}\n"
1042 "is omitted. The current output buffer contents are passed to the\n"
1043 "underlying port implementation (e.g., in the case of fports, the\n"
1044 "data will be written to the file and the output buffer will be cleared.)\n"
1045 "It has no effect on an unbuffered port.\n\n"
1046 "The return value is unspecified.")
1047 #define FUNC_NAME s_scm_force_output
1049 if (SCM_UNBNDP (port
))
1050 port
= scm_current_output_port ();
1053 port
= SCM_COERCE_OUTPORT (port
);
1054 SCM_VALIDATE_OPOUTPORT (1, port
);
1057 return SCM_UNSPECIFIED
;
1063 flush_output_port (void *closure
, SCM port
)
1065 if (SCM_OPOUTPORTP (port
))
1069 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
1071 "Equivalent to calling @code{force-output} on\n"
1072 "all open output ports. The return value is unspecified.")
1073 #define FUNC_NAME s_scm_flush_all_ports
1075 scm_c_port_for_each (&flush_output_port
, NULL
);
1076 return SCM_UNSPECIFIED
;
1080 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1082 "Return the next character available from @var{port}, updating\n"
1083 "@var{port} to point to the following character. If no more\n"
1084 "characters are available, the end-of-file object is returned.\n"
1086 "When @var{port}'s data cannot be decoded according to its\n"
1087 "character encoding, a @code{decoding-error} is raised and\n"
1088 "@var{port} points past the erroneous byte sequence.\n")
1089 #define FUNC_NAME s_scm_read_char
1092 if (SCM_UNBNDP (port
))
1093 port
= scm_current_input_port ();
1094 SCM_VALIDATE_OPINPORT (1, port
);
1095 c
= scm_getc (port
);
1098 return SCM_MAKE_CHAR (c
);
1102 /* Update the line and column number of PORT after consumption of C. */
1104 update_port_lf (scm_t_wchar c
, SCM port
)
1129 #define SCM_MBCHAR_BUF_SIZE (4)
1131 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1132 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1134 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1136 scm_t_wchar codepoint
;
1138 if (utf8_buf
[0] <= 0x7f)
1141 codepoint
= utf8_buf
[0];
1143 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1146 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1147 | (utf8_buf
[1] & 0x3f);
1149 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1152 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1153 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1154 | (utf8_buf
[2] & 0x3f);
1159 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1160 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1161 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1162 | (utf8_buf
[3] & 0x3f);
1168 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1169 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1170 representation, and set *LEN to the length in bytes. Return
1171 `EILSEQ' on error. */
1173 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1174 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1176 #define ASSERT_NOT_EOF(b) \
1177 if (SCM_UNLIKELY ((b) == EOF)) \
1179 #define CONSUME_PEEKED_BYTE() \
1186 pt
= SCM_PTAB_ENTRY (port
);
1188 byte
= scm_get_byte_or_eof (port
);
1195 buf
[0] = (scm_t_uint8
) byte
;
1200 *codepoint
= buf
[0];
1201 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1204 byte
= scm_peek_byte_or_eof (port
);
1205 ASSERT_NOT_EOF (byte
);
1207 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1210 CONSUME_PEEKED_BYTE ();
1211 buf
[1] = (scm_t_uint8
) byte
;
1214 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1217 else if ((buf
[0] & 0xf0) == 0xe0)
1220 byte
= scm_peek_byte_or_eof (port
);
1221 ASSERT_NOT_EOF (byte
);
1223 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1224 || (buf
[0] == 0xe0 && byte
< 0xa0)
1225 || (buf
[0] == 0xed && byte
> 0x9f)))
1228 CONSUME_PEEKED_BYTE ();
1229 buf
[1] = (scm_t_uint8
) byte
;
1232 byte
= scm_peek_byte_or_eof (port
);
1233 ASSERT_NOT_EOF (byte
);
1235 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1238 CONSUME_PEEKED_BYTE ();
1239 buf
[2] = (scm_t_uint8
) byte
;
1242 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1243 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1246 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1249 byte
= scm_peek_byte_or_eof (port
);
1250 ASSERT_NOT_EOF (byte
);
1252 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1253 || (buf
[0] == 0xf0 && byte
< 0x90)
1254 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1257 CONSUME_PEEKED_BYTE ();
1258 buf
[1] = (scm_t_uint8
) byte
;
1261 byte
= scm_peek_byte_or_eof (port
);
1262 ASSERT_NOT_EOF (byte
);
1264 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1267 CONSUME_PEEKED_BYTE ();
1268 buf
[2] = (scm_t_uint8
) byte
;
1271 byte
= scm_peek_byte_or_eof (port
);
1272 ASSERT_NOT_EOF (byte
);
1274 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1277 CONSUME_PEEKED_BYTE ();
1278 buf
[3] = (scm_t_uint8
) byte
;
1281 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1282 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1283 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1292 /* Here we could choose the consume the faulty byte when it's not a
1293 valid starting byte, but it's not a requirement. What Section 3.9
1294 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1295 would otherwise be a valid starting byte. */
1299 #undef CONSUME_PEEKED_BYTE
1300 #undef ASSERT_NOT_EOF
1303 /* Likewise, read a byte sequence from PORT, passing it through its
1304 input conversion descriptor. */
1306 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1307 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1309 scm_t_iconv_descriptors
*id
;
1311 size_t bytes_consumed
, output_size
;
1313 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1315 id
= scm_i_port_iconv_descriptors (port
);
1317 for (output_size
= 0, output
= (char *) utf8_buf
,
1318 bytes_consumed
= 0, err
= 0;
1319 err
== 0 && output_size
== 0
1320 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1324 size_t input_left
, output_left
, done
;
1326 byte_read
= scm_get_byte_or_eof (port
);
1327 if (byte_read
== EOF
)
1329 if (bytes_consumed
== 0)
1331 *codepoint
= (scm_t_wchar
) EOF
;
1339 buf
[bytes_consumed
] = byte_read
;
1342 input_left
= bytes_consumed
+ 1;
1343 output_left
= sizeof (utf8_buf
);
1345 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1346 if (done
== (size_t) -1)
1350 /* Missing input: keep trying. */
1354 output_size
= sizeof (utf8_buf
) - output_left
;
1357 if (SCM_UNLIKELY (output_size
== 0))
1358 /* An unterminated sequence. */
1360 else if (SCM_LIKELY (err
== 0))
1362 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1363 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1364 *len
= bytes_consumed
;
1370 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1371 with the byte representation of the codepoint in PORT's encoding, and
1372 set *LEN to the length in bytes of that representation. Return 0 on
1373 success and an errno value on error. */
1375 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1376 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1379 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1380 scm_t_port_internal
*pti
= SCM_PORT_GET_INTERNAL (port
);
1382 if (pti
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1383 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1385 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1387 if (SCM_LIKELY (err
== 0))
1388 update_port_lf (*codepoint
, port
);
1389 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1393 update_port_lf (*codepoint
, port
);
1399 /* Read a codepoint from PORT and return it. */
1402 #define FUNC_NAME "scm_getc"
1406 scm_t_wchar codepoint
;
1407 char buf
[SCM_MBCHAR_BUF_SIZE
];
1409 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1410 if (SCM_UNLIKELY (err
!= 0))
1411 /* At this point PORT should point past the invalid encoding, as per
1412 R6RS-lib Section 8.2.4. */
1413 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1419 /* this should only be called when the read buffer is empty. it
1420 tries to refill the read buffer. it returns the first char from
1421 the port, which is either EOF or *(pt->read_pos). */
1423 scm_i_fill_input (SCM port
)
1425 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1427 assert (pt
->read_pos
== pt
->read_end
);
1429 if (pt
->read_buf
== pt
->putback_buf
)
1431 /* finished reading put-back chars. */
1432 pt
->read_buf
= pt
->saved_read_buf
;
1433 pt
->read_pos
= pt
->saved_read_pos
;
1434 pt
->read_end
= pt
->saved_read_end
;
1435 pt
->read_buf_size
= pt
->saved_read_buf_size
;
1436 if (pt
->read_pos
< pt
->read_end
)
1437 return *(pt
->read_pos
);
1439 return scm_ptobs
[SCM_PTOBNUM (port
)].fill_input (port
);
1443 scm_fill_input (SCM port
)
1445 return scm_i_fill_input (port
);
1448 /* Slow-path fallback for 'scm_get_byte_or_eof' in inline.h */
1450 scm_i_get_byte_or_eof (SCM port
)
1452 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1454 if (pt
->rw_active
== SCM_PORT_WRITE
)
1458 pt
->rw_active
= SCM_PORT_READ
;
1460 if (pt
->read_pos
>= pt
->read_end
)
1462 if (SCM_UNLIKELY (scm_i_fill_input (port
) == EOF
))
1466 return *pt
->read_pos
++;
1469 /* Slow-path fallback for 'scm_peek_byte_or_eof' in inline.h */
1471 scm_i_peek_byte_or_eof (SCM port
)
1473 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1475 if (pt
->rw_active
== SCM_PORT_WRITE
)
1479 pt
->rw_active
= SCM_PORT_READ
;
1481 if (pt
->read_pos
>= pt
->read_end
)
1483 if (SCM_UNLIKELY (scm_i_fill_input (port
) == EOF
))
1487 return *pt
->read_pos
;
1493 * This function differs from scm_c_write; it updates port line and
1497 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
1499 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1500 scm_t_ptob_descriptor
*ptob
= &scm_ptobs
[SCM_PTOBNUM (port
)];
1502 if (pt
->rw_active
== SCM_PORT_READ
)
1503 scm_end_input (port
);
1505 ptob
->write (port
, ptr
, size
);
1507 for (; size
; ptr
++, size
--)
1508 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
1511 pt
->rw_active
= SCM_PORT_WRITE
;
1514 /* Write STR to PORT from START inclusive to END exclusive. */
1516 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
1518 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1520 if (pt
->rw_active
== SCM_PORT_READ
)
1521 scm_end_input (port
);
1523 if (end
== (size_t) -1)
1524 end
= scm_i_string_length (str
);
1526 scm_i_display_substring (str
, start
, end
, port
);
1529 pt
->rw_active
= SCM_PORT_WRITE
;
1534 * Used by an application to read arbitrary number of bytes from an
1535 * SCM port. Same semantics as libc read, except that scm_c_read only
1536 * returns less than SIZE bytes if at end-of-file.
1538 * Warning: Doesn't update port line and column counts! */
1540 /* This structure, and the following swap_buffer function, are used
1541 for temporarily swapping a port's own read buffer, and the buffer
1542 that the caller of scm_c_read provides. */
1543 struct port_and_swap_buffer
1546 unsigned char *buffer
;
1551 swap_buffer (void *data
)
1553 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1554 unsigned char *old_buf
= psb
->pt
->read_buf
;
1555 size_t old_size
= psb
->pt
->read_buf_size
;
1557 /* Make the port use (buffer, size) from the struct. */
1558 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1559 psb
->pt
->read_buf_size
= psb
->size
;
1561 /* Save the port's old (buffer, size) in the struct. */
1562 psb
->buffer
= old_buf
;
1563 psb
->size
= old_size
;
1567 scm_c_read (SCM port
, void *buffer
, size_t size
)
1568 #define FUNC_NAME "scm_c_read"
1571 size_t n_read
= 0, n_available
;
1572 struct port_and_swap_buffer psb
;
1574 SCM_VALIDATE_OPINPORT (1, port
);
1576 pt
= SCM_PTAB_ENTRY (port
);
1577 if (pt
->rw_active
== SCM_PORT_WRITE
)
1578 scm_ptobs
[SCM_PTOBNUM (port
)].flush (port
);
1581 pt
->rw_active
= SCM_PORT_READ
;
1583 /* Take bytes first from the port's read buffer. */
1584 if (pt
->read_pos
< pt
->read_end
)
1586 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1587 memcpy (buffer
, pt
->read_pos
, n_available
);
1588 buffer
= (char *) buffer
+ n_available
;
1589 pt
->read_pos
+= n_available
;
1590 n_read
+= n_available
;
1591 size
-= n_available
;
1594 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1598 /* Now we will call scm_i_fill_input repeatedly until we have read the
1599 requested number of bytes. (Note that a single scm_i_fill_input
1600 call does not guarantee to fill the whole of the port's read
1602 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1604 /* The port that we are reading from is unbuffered - i.e. does
1605 not have its own persistent buffer - but we have a buffer,
1606 provided by our caller, that is the right size for the data
1607 that is wanted. For the following scm_i_fill_input calls,
1608 therefore, we use the buffer in hand as the port's read
1611 We need to make sure that the port's normal (1 byte) buffer
1612 is reinstated in case one of the scm_i_fill_input () calls
1613 throws an exception; we use the scm_dynwind_* API to achieve
1616 A consequence of this optimization is that the fill_input
1617 functions can't unget characters. That'll push data to the
1618 pushback buffer instead of this psb buffer. */
1620 unsigned char *pback
= pt
->putback_buf
;
1623 psb
.buffer
= buffer
;
1625 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1626 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1627 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1629 /* Call scm_i_fill_input until we have all the bytes that we need,
1631 while (pt
->read_buf_size
&& (scm_i_fill_input (port
) != EOF
))
1633 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1634 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1637 if (pback
!= pt
->putback_buf
1638 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1639 scm_misc_error (FUNC_NAME
,
1640 "scm_c_read must not call a fill function that pushes "
1641 "back characters onto an unbuffered port", SCM_EOL
);
1643 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1645 /* Reinstate the port's normal buffer. */
1650 /* The port has its own buffer. It is important that we use it,
1651 even if it happens to be smaller than our caller's buffer, so
1652 that a custom port implementation's entry points (in
1653 particular, fill_input) can rely on the buffer always being
1654 the same as they first set up. */
1655 while (size
&& (scm_i_fill_input (port
) != EOF
))
1657 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1658 memcpy (buffer
, pt
->read_pos
, n_available
);
1659 buffer
= (char *) buffer
+ n_available
;
1660 pt
->read_pos
+= n_available
;
1661 n_read
+= n_available
;
1662 size
-= n_available
;
1672 * Used by an application to write arbitrary number of bytes to an SCM
1673 * port. Similar semantics as libc write. However, unlike libc
1674 * write, scm_c_write writes the requested number of bytes and has no
1677 * Warning: Doesn't update port line and column counts!
1681 scm_c_write (SCM port
, const void *ptr
, size_t size
)
1682 #define FUNC_NAME "scm_c_write"
1685 scm_t_ptob_descriptor
*ptob
;
1687 SCM_VALIDATE_OPOUTPORT (1, port
);
1689 pt
= SCM_PTAB_ENTRY (port
);
1690 ptob
= &scm_ptobs
[SCM_PTOBNUM (port
)];
1692 if (pt
->rw_active
== SCM_PORT_READ
)
1693 scm_end_input (port
);
1695 ptob
->write (port
, ptr
, size
);
1698 pt
->rw_active
= SCM_PORT_WRITE
;
1703 scm_flush (SCM port
)
1705 long i
= SCM_PTOBNUM (port
);
1707 (scm_ptobs
[i
].flush
) (port
);
1711 scm_end_input (SCM port
)
1714 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1716 if (pt
->read_buf
== pt
->putback_buf
)
1718 offset
= pt
->read_end
- pt
->read_pos
;
1719 pt
->read_buf
= pt
->saved_read_buf
;
1720 pt
->read_pos
= pt
->saved_read_pos
;
1721 pt
->read_end
= pt
->saved_read_end
;
1722 pt
->read_buf_size
= pt
->saved_read_buf_size
;
1727 scm_ptobs
[SCM_PTOBNUM (port
)].end_input (port
, offset
);
1734 scm_unget_byte (int c
, SCM port
)
1735 #define FUNC_NAME "scm_unget_byte"
1737 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1739 if (pt
->read_buf
== pt
->putback_buf
)
1740 /* already using the put-back buffer. */
1742 /* enlarge putback_buf if necessary. */
1743 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1744 && pt
->read_buf
== pt
->read_pos
)
1746 size_t new_size
= pt
->read_buf_size
* 2;
1747 unsigned char *tmp
= (unsigned char *)
1748 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1751 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1752 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1753 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1756 /* shift any existing bytes to buffer + 1. */
1757 if (pt
->read_pos
== pt
->read_end
)
1758 pt
->read_end
= pt
->read_buf
+ 1;
1759 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1761 int count
= pt
->read_end
- pt
->read_pos
;
1763 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1764 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1767 pt
->read_pos
= pt
->read_buf
;
1770 /* switch to the put-back buffer. */
1772 if (pt
->putback_buf
== NULL
)
1775 = (unsigned char *) scm_gc_malloc_pointerless
1776 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1777 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1780 pt
->saved_read_buf
= pt
->read_buf
;
1781 pt
->saved_read_pos
= pt
->read_pos
;
1782 pt
->saved_read_end
= pt
->read_end
;
1783 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1785 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1786 pt
->read_end
= pt
->read_buf
+ 1;
1787 pt
->read_buf_size
= pt
->putback_buf_size
;
1793 pt
->rw_active
= SCM_PORT_READ
;
1798 scm_ungetc (scm_t_wchar c
, SCM port
)
1799 #define FUNC_NAME "scm_ungetc"
1801 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1803 char result_buf
[10];
1804 const char *encoding
;
1808 if (pt
->encoding
!= NULL
)
1809 encoding
= pt
->encoding
;
1811 encoding
= "ISO-8859-1";
1813 len
= sizeof (result_buf
);
1814 result
= u32_conv_to_encoding (encoding
,
1815 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1816 (uint32_t *) &c
, 1, NULL
,
1819 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1820 scm_encoding_error (FUNC_NAME
, errno
,
1821 "conversion to port encoding failed",
1822 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1824 for (i
= len
- 1; i
>= 0; i
--)
1825 scm_unget_byte (result
[i
], port
);
1827 if (SCM_UNLIKELY (result
!= result_buf
))
1832 /* What should col be in this case?
1833 * We'll leave it at -1.
1835 SCM_LINUM (port
) -= 1;
1844 scm_ungets (const char *s
, int n
, SCM port
)
1846 /* This is simple minded and inefficient, but unreading strings is
1847 * probably not a common operation, and remember that line and
1848 * column numbers have to be handled...
1850 * Please feel free to write an optimized version!
1853 scm_ungetc (s
[n
], port
);
1857 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1859 "Return the next character available from @var{port},\n"
1860 "@emph{without} updating @var{port} to point to the following\n"
1861 "character. If no more characters are available, the\n"
1862 "end-of-file object is returned.\n"
1864 "The value returned by\n"
1865 "a call to @code{peek-char} is the same as the value that would\n"
1866 "have been returned by a call to @code{read-char} on the same\n"
1867 "port. The only difference is that the very next call to\n"
1868 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1869 "return the value returned by the preceding call to\n"
1870 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1871 "an interactive port will hang waiting for input whenever a call\n"
1872 "to @code{read-char} would have hung.\n"
1874 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1875 "if such a situation occurs. However, unlike with @code{read-char},\n"
1876 "@var{port} still points at the beginning of the erroneous byte\n"
1877 "sequence when the error is raised.\n")
1878 #define FUNC_NAME s_scm_peek_char
1883 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1884 long column
, line
, i
;
1887 if (SCM_UNBNDP (port
))
1888 port
= scm_current_input_port ();
1889 SCM_VALIDATE_OPINPORT (1, port
);
1891 column
= SCM_COL (port
);
1892 line
= SCM_LINUM (port
);
1894 err
= get_codepoint (port
, &c
, bytes
, &len
);
1896 for (i
= len
- 1; i
>= 0; i
--)
1897 scm_unget_byte (bytes
[i
], port
);
1899 SCM_COL (port
) = column
;
1900 SCM_LINUM (port
) = line
;
1902 if (SCM_UNLIKELY (err
!= 0))
1904 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1906 /* Shouldn't happen since `catch' always aborts to prompt. */
1907 result
= SCM_BOOL_F
;
1910 result
= SCM_EOF_VAL
;
1912 result
= SCM_MAKE_CHAR (c
);
1918 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
1919 (SCM cobj
, SCM port
),
1920 "Place character @var{cobj} in @var{port} so that it will be\n"
1921 "read by the next read operation. If called multiple times, the\n"
1922 "unread characters will be read again in last-in first-out\n"
1923 "order. If @var{port} is not supplied, the current input port\n"
1925 #define FUNC_NAME s_scm_unread_char
1929 SCM_VALIDATE_CHAR (1, cobj
);
1930 if (SCM_UNBNDP (port
))
1931 port
= scm_current_input_port ();
1932 SCM_VALIDATE_OPINPORT (2, port
);
1934 c
= SCM_CHAR (cobj
);
1936 scm_ungetc (c
, port
);
1941 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
1942 (SCM str
, SCM port
),
1943 "Place the string @var{str} in @var{port} so that its characters will be\n"
1944 "read in subsequent read operations. If called multiple times, the\n"
1945 "unread characters will be read again in last-in first-out order. If\n"
1946 "@var{port} is not supplied, the current-input-port is used.")
1947 #define FUNC_NAME s_scm_unread_string
1950 SCM_VALIDATE_STRING (1, str
);
1951 if (SCM_UNBNDP (port
))
1952 port
= scm_current_input_port ();
1953 SCM_VALIDATE_OPINPORT (2, port
);
1955 n
= scm_i_string_length (str
);
1958 scm_ungetc (scm_i_string_ref (str
, n
), port
);
1964 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
1965 (SCM fd_port
, SCM offset
, SCM whence
),
1966 "Sets the current position of @var{fd_port} to the integer\n"
1967 "@var{offset}, which is interpreted according to the value of\n"
1970 "One of the following variables should be supplied for\n"
1972 "@defvar SEEK_SET\n"
1973 "Seek from the beginning of the file.\n"
1975 "@defvar SEEK_CUR\n"
1976 "Seek from the current position.\n"
1978 "@defvar SEEK_END\n"
1979 "Seek from the end of the file.\n"
1981 "If @var{fd_port} is a file descriptor, the underlying system\n"
1982 "call is @code{lseek}. @var{port} may be a string port.\n"
1984 "The value returned is the new position in the file. This means\n"
1985 "that the current position of a port can be obtained using:\n"
1987 "(seek port 0 SEEK_CUR)\n"
1989 #define FUNC_NAME s_scm_seek
1993 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
1995 how
= scm_to_int (whence
);
1996 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
1997 SCM_OUT_OF_RANGE (3, whence
);
1999 if (SCM_OPPORTP (fd_port
))
2001 scm_t_ptob_descriptor
*ptob
= scm_ptobs
+ SCM_PTOBNUM (fd_port
);
2002 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2003 off_t_or_off64_t rv
;
2006 SCM_MISC_ERROR ("port is not seekable",
2007 scm_cons (fd_port
, SCM_EOL
));
2009 rv
= ptob
->seek (fd_port
, off
, how
);
2010 return scm_from_off_t_or_off64_t (rv
);
2012 else /* file descriptor?. */
2014 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2015 off_t_or_off64_t rv
;
2016 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2019 return scm_from_off_t_or_off64_t (rv
);
2028 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2029 doesn't have the filename version truncate(), hence this code. */
2030 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2032 truncate (const char *file
, off_t length
)
2036 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2040 ret
= ftruncate (fdes
, length
);
2043 int save_errno
= errno
;
2049 return close (fdes
);
2051 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2053 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2054 (SCM object
, SCM length
),
2055 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2056 "can be a filename string, a port object, or an integer file\n"
2058 "The return value is unspecified.\n"
2060 "For a port or file descriptor @var{length} can be omitted, in\n"
2061 "which case the file is truncated at the current position (per\n"
2062 "@code{ftell} above).\n"
2064 "On most systems a file can be extended by giving a length\n"
2065 "greater than the current size, but this is not mandatory in the\n"
2067 #define FUNC_NAME s_scm_truncate_file
2071 /* "object" can be a port, fdes or filename.
2073 Negative "length" makes no sense, but it's left to truncate() or
2074 ftruncate() to give back an error for that (normally EINVAL).
2077 if (SCM_UNBNDP (length
))
2079 /* must supply length if object is a filename. */
2080 if (scm_is_string (object
))
2081 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2083 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2086 object
= SCM_COERCE_OUTPORT (object
);
2087 if (scm_is_integer (object
))
2089 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2090 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2093 else if (SCM_OPOUTPORTP (object
))
2095 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2096 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2097 scm_t_ptob_descriptor
*ptob
= scm_ptobs
+ SCM_PTOBNUM (object
);
2099 if (!ptob
->truncate
)
2100 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2101 if (pt
->rw_active
== SCM_PORT_READ
)
2102 scm_end_input (object
);
2103 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2104 ptob
->flush (object
);
2106 ptob
->truncate (object
, c_length
);
2111 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2112 char *str
= scm_to_locale_string (object
);
2114 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2121 return SCM_UNSPECIFIED
;
2125 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2127 "Return the current line number for @var{port}.\n"
2129 "The first line of a file is 0. But you might want to add 1\n"
2130 "when printing line numbers, since starting from 1 is\n"
2131 "traditional in error messages, and likely to be more natural to\n"
2133 #define FUNC_NAME s_scm_port_line
2135 port
= SCM_COERCE_OUTPORT (port
);
2136 SCM_VALIDATE_OPENPORT (1, port
);
2137 return scm_from_long (SCM_LINUM (port
));
2141 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2142 (SCM port
, SCM line
),
2143 "Set the current line number for @var{port} to @var{line}. The\n"
2144 "first line of a file is 0.")
2145 #define FUNC_NAME s_scm_set_port_line_x
2147 port
= SCM_COERCE_OUTPORT (port
);
2148 SCM_VALIDATE_OPENPORT (1, port
);
2149 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2150 return SCM_UNSPECIFIED
;
2154 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2156 "Return the current column number of @var{port}.\n"
2157 "If the number is\n"
2158 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2159 "- i.e. the first character of the first line is line 0, column 0.\n"
2160 "(However, when you display a file position, for example in an error\n"
2161 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2162 "because lines and column numbers traditionally start with 1, and that is\n"
2163 "what non-programmers will find most natural.)")
2164 #define FUNC_NAME s_scm_port_column
2166 port
= SCM_COERCE_OUTPORT (port
);
2167 SCM_VALIDATE_OPENPORT (1, port
);
2168 return scm_from_int (SCM_COL (port
));
2172 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2173 (SCM port
, SCM column
),
2174 "Set the current column of @var{port}. Before reading the first\n"
2175 "character on a line the column should be 0.")
2176 #define FUNC_NAME s_scm_set_port_column_x
2178 port
= SCM_COERCE_OUTPORT (port
);
2179 SCM_VALIDATE_OPENPORT (1, port
);
2180 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2181 return SCM_UNSPECIFIED
;
2185 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2187 "Return the filename associated with @var{port}, or @code{#f}\n"
2188 "if no filename is associated with the port.")
2189 #define FUNC_NAME s_scm_port_filename
2191 port
= SCM_COERCE_OUTPORT (port
);
2192 SCM_VALIDATE_OPENPORT (1, port
);
2193 return SCM_FILENAME (port
);
2197 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2198 (SCM port
, SCM filename
),
2199 "Change the filename associated with @var{port}, using the current input\n"
2200 "port if none is specified. Note that this does not change the port's\n"
2201 "source of data, but only the value that is returned by\n"
2202 "@code{port-filename} and reported in diagnostic output.")
2203 #define FUNC_NAME s_scm_set_port_filename_x
2205 port
= SCM_COERCE_OUTPORT (port
);
2206 SCM_VALIDATE_OPENPORT (1, port
);
2207 /* We allow the user to set the filename to whatever he likes. */
2208 SCM_SET_FILENAME (port
, filename
);
2209 return SCM_UNSPECIFIED
;
2213 /* A fluid specifying the default encoding for newly created ports. If it is
2214 a string, that is the encoding. If it is #f, it is in the "native"
2215 (Latin-1) encoding. */
2216 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
2218 static int scm_port_encoding_init
= 0;
2220 /* Use ENCODING as the default encoding for future ports. */
2222 scm_i_set_default_port_encoding (const char *encoding
)
2224 if (!scm_port_encoding_init
2225 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
2226 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
2229 if (encoding
== NULL
2230 || !strcmp (encoding
, "ASCII")
2231 || !strcmp (encoding
, "ANSI_X3.4-1968")
2232 || !strcmp (encoding
, "ISO-8859-1"))
2233 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
2235 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
2236 scm_from_locale_string (encoding
));
2239 /* Return the name of the default encoding for newly created ports; a
2240 return value of NULL means "ISO-8859-1". */
2242 scm_i_default_port_encoding (void)
2244 if (!scm_port_encoding_init
)
2246 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
2252 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
2253 if (!scm_is_string (encoding
))
2256 return scm_i_string_chars (encoding
);
2261 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
2263 close_iconv_descriptors (ptr
);
2266 static scm_t_iconv_descriptors
*
2267 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
2269 scm_t_iconv_descriptors
*id
;
2270 iconv_t input_cd
, output_cd
;
2272 input_cd
= (iconv_t
) -1;
2273 output_cd
= (iconv_t
) -1;
2276 /* Open an input iconv conversion descriptor, from ENCODING
2277 to UTF-8. We choose UTF-8, not UTF-32, because iconv
2278 implementations can typically convert from anything to
2279 UTF-8, but not to UTF-32 (see
2280 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
2282 /* Assume opening an iconv descriptor causes about 16 KB of
2284 scm_gc_register_allocation (16 * 1024);
2286 input_cd
= iconv_open ("UTF-8", encoding
);
2287 if (input_cd
== (iconv_t
) -1)
2288 goto invalid_encoding
;
2293 /* Assume opening an iconv descriptor causes about 16 KB of
2295 scm_gc_register_allocation (16 * 1024);
2297 output_cd
= iconv_open (encoding
, "UTF-8");
2298 if (output_cd
== (iconv_t
) -1)
2300 if (input_cd
!= (iconv_t
) -1)
2301 iconv_close (input_cd
);
2302 goto invalid_encoding
;
2306 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
2307 id
->input_cd
= input_cd
;
2308 id
->output_cd
= output_cd
;
2310 /* Register a finalizer to close the descriptors. */
2311 scm_i_set_finalizer (id
, finalize_iconv_descriptors
, NULL
);
2318 err
= scm_from_locale_string (encoding
);
2319 scm_misc_error ("open_iconv_descriptors",
2320 "invalid or unknown character encoding ~s",
2326 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
2328 if (id
->input_cd
!= (iconv_t
) -1)
2329 iconv_close (id
->input_cd
);
2330 if (id
->output_cd
!= (iconv_t
) -1)
2331 iconv_close (id
->output_cd
);
2332 id
->input_cd
= (void *) -1;
2333 id
->output_cd
= (void *) -1;
2336 scm_t_iconv_descriptors
*
2337 scm_i_port_iconv_descriptors (SCM port
)
2340 scm_t_port_internal
*pti
;
2342 pt
= SCM_PTAB_ENTRY (port
);
2343 pti
= SCM_PORT_GET_INTERNAL (port
);
2345 assert (pti
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
2347 if (!pti
->iconv_descriptors
)
2350 pt
->encoding
= "ISO-8859-1";
2351 pti
->iconv_descriptors
=
2352 open_iconv_descriptors (pt
->encoding
,
2353 SCM_INPUT_PORT_P (port
),
2354 SCM_OUTPUT_PORT_P (port
));
2357 return pti
->iconv_descriptors
;
2361 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
2364 scm_t_port_internal
*pti
;
2365 scm_t_iconv_descriptors
*prev
;
2367 /* Set the character encoding for this port. */
2368 pt
= SCM_PTAB_ENTRY (port
);
2369 pti
= SCM_PORT_GET_INTERNAL (port
);
2370 prev
= pti
->iconv_descriptors
;
2372 if (encoding
== NULL
)
2373 encoding
= "ISO-8859-1";
2375 /* If ENCODING is UTF-8, then no conversion descriptor is opened
2376 because we do I/O ourselves. This saves 100+ KiB for each
2378 if (strcmp (encoding
, "UTF-8") == 0)
2380 pt
->encoding
= "UTF-8";
2381 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
2382 pti
->iconv_descriptors
= NULL
;
2386 /* Open descriptors before mutating the port. */
2387 pti
->iconv_descriptors
=
2388 open_iconv_descriptors (encoding
,
2389 SCM_INPUT_PORT_P (port
),
2390 SCM_OUTPUT_PORT_P (port
));
2391 pt
->encoding
= scm_gc_strdup (encoding
, "port");
2392 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
2396 close_iconv_descriptors (prev
);
2399 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
2401 "Returns, as a string, the character encoding that @var{port}\n"
2402 "uses to interpret its input and output.\n")
2403 #define FUNC_NAME s_scm_port_encoding
2408 SCM_VALIDATE_PORT (1, port
);
2410 pt
= SCM_PTAB_ENTRY (port
);
2413 return scm_from_locale_string (pt
->encoding
);
2419 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
2420 (SCM port
, SCM enc
),
2421 "Sets the character encoding that will be used to interpret all\n"
2422 "port I/O. New ports are created with the encoding\n"
2423 "appropriate for the current locale if @code{setlocale} has \n"
2424 "been called or ISO-8859-1 otherwise\n"
2425 "and this procedure can be used to modify that encoding.\n")
2426 #define FUNC_NAME s_scm_set_port_encoding_x
2430 SCM_VALIDATE_PORT (1, port
);
2431 SCM_VALIDATE_STRING (2, enc
);
2433 enc_str
= scm_to_locale_string (enc
);
2434 scm_i_set_port_encoding_x (port
, enc_str
);
2437 return SCM_UNSPECIFIED
;
2442 /* A fluid specifying the default conversion handler for newly created
2443 ports. Its value should be one of the symbols below. */
2444 SCM_VARIABLE (default_conversion_strategy_var
,
2445 "%default-port-conversion-strategy");
2447 /* Whether the above fluid is initialized. */
2448 static int scm_conversion_strategy_init
= 0;
2450 /* The possible conversion strategies. */
2451 SCM_SYMBOL (sym_error
, "error");
2452 SCM_SYMBOL (sym_substitute
, "substitute");
2453 SCM_SYMBOL (sym_escape
, "escape");
2455 /* Return the default failed encoding conversion policy for new created
2457 scm_t_string_failed_conversion_handler
2458 scm_i_default_port_conversion_handler (void)
2460 scm_t_string_failed_conversion_handler handler
;
2462 if (!scm_conversion_strategy_init
2463 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
2464 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
2469 fluid
= SCM_VARIABLE_REF (default_conversion_strategy_var
);
2470 value
= scm_fluid_ref (fluid
);
2472 if (scm_is_eq (sym_substitute
, value
))
2473 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
2474 else if (scm_is_eq (sym_escape
, value
))
2475 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
2477 /* Default to 'error also when the fluid's value is not one of
2478 the valid symbols. */
2479 handler
= SCM_FAILED_CONVERSION_ERROR
;
2485 /* Use HANDLER as the default conversion strategy for future ports. */
2487 scm_i_set_default_port_conversion_handler (scm_t_string_failed_conversion_handler
2492 if (!scm_conversion_strategy_init
2493 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
2494 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
2499 case SCM_FAILED_CONVERSION_ERROR
:
2500 strategy
= sym_error
;
2503 case SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
:
2504 strategy
= sym_escape
;
2507 case SCM_FAILED_CONVERSION_QUESTION_MARK
:
2508 strategy
= sym_substitute
;
2515 scm_fluid_set_x (SCM_VARIABLE_REF (default_conversion_strategy_var
),
2519 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
2520 1, 0, 0, (SCM port
),
2521 "Returns the behavior of the port when handling a character that\n"
2522 "is not representable in the port's current encoding.\n"
2523 "It returns the symbol @code{error} if unrepresentable characters\n"
2524 "should cause exceptions, @code{substitute} if the port should\n"
2525 "try to replace unrepresentable characters with question marks or\n"
2526 "approximate characters, or @code{escape} if unrepresentable\n"
2527 "characters should be converted to string escapes.\n"
2529 "If @var{port} is @code{#f}, then the current default behavior\n"
2530 "will be returned. New ports will have this default behavior\n"
2531 "when they are created.\n")
2532 #define FUNC_NAME s_scm_port_conversion_strategy
2534 scm_t_string_failed_conversion_handler h
;
2536 if (scm_is_false (port
))
2537 h
= scm_i_default_port_conversion_handler ();
2542 SCM_VALIDATE_OPPORT (1, port
);
2543 pt
= SCM_PTAB_ENTRY (port
);
2545 h
= pt
->ilseq_handler
;
2548 if (h
== SCM_FAILED_CONVERSION_ERROR
)
2549 return scm_from_latin1_symbol ("error");
2550 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
2551 return scm_from_latin1_symbol ("substitute");
2552 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
2553 return scm_from_latin1_symbol ("escape");
2557 /* Never gets here. */
2558 return SCM_UNDEFINED
;
2562 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
2564 (SCM port
, SCM sym
),
2565 "Sets the behavior of the interpreter when outputting a character\n"
2566 "that is not representable in the port's current encoding.\n"
2567 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
2568 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
2569 "when an unconvertible character is encountered. If it is\n"
2570 "@code{'substitute}, then unconvertible characters will \n"
2571 "be replaced with approximate characters, or with question marks\n"
2572 "if no approximately correct character is available.\n"
2573 "If it is @code{'escape},\n"
2574 "it will appear as a hex escape when output.\n"
2576 "If @var{port} is an open port, the conversion error behavior\n"
2577 "is set for that port. If it is @code{#f}, it is set as the\n"
2578 "default behavior for any future ports that get created in\n"
2580 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
2582 scm_t_string_failed_conversion_handler handler
;
2584 if (scm_is_eq (sym
, sym_error
))
2585 handler
= SCM_FAILED_CONVERSION_ERROR
;
2586 else if (scm_is_eq (sym
, sym_substitute
))
2587 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
2588 else if (scm_is_eq (sym
, sym_escape
))
2589 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
2591 SCM_MISC_ERROR ("unknown conversion strategy ~s", scm_list_1 (sym
));
2593 if (scm_is_false (port
))
2594 scm_i_set_default_port_conversion_handler (handler
);
2597 SCM_VALIDATE_OPPORT (1, port
);
2598 SCM_PTAB_ENTRY (port
)->ilseq_handler
= handler
;
2601 return SCM_UNSPECIFIED
;
2608 scm_print_port_mode (SCM exp
, SCM port
)
2610 scm_puts (SCM_CLOSEDP (exp
)
2612 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2613 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2616 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2623 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2625 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2628 scm_puts ("#<", port
);
2629 scm_print_port_mode (exp
, port
);
2630 scm_puts (type
, port
);
2631 scm_putc (' ', port
);
2632 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2633 scm_putc ('>', port
);
2641 scm_t_bits scm_tc16_void_port
= 0;
2643 static int fill_input_void_port (SCM port SCM_UNUSED
)
2649 write_void_port (SCM port SCM_UNUSED
,
2650 const void *data SCM_UNUSED
,
2651 size_t size SCM_UNUSED
)
2656 scm_i_void_port (long mode_bits
)
2658 scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex
);
2660 SCM answer
= scm_new_port_table_entry (scm_tc16_void_port
);
2661 scm_t_port
* pt
= SCM_PTAB_ENTRY(answer
);
2663 scm_port_non_buffer (pt
);
2665 SCM_SETSTREAM (answer
, 0);
2666 SCM_SET_CELL_TYPE (answer
, scm_tc16_void_port
| mode_bits
);
2667 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
2673 scm_void_port (char *mode_str
)
2675 return scm_i_void_port (scm_mode_bits (mode_str
));
2678 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2680 "Create and return a new void port. A void port acts like\n"
2681 "@file{/dev/null}. The @var{mode} argument\n"
2682 "specifies the input/output modes for this port: see the\n"
2683 "documentation for @code{open-file} in @ref{File Ports}.")
2684 #define FUNC_NAME s_scm_sys_make_void_port
2686 return scm_i_void_port (scm_i_mode_bits (mode
));
2691 /* Initialization. */
2696 /* lseek() symbols. */
2697 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2698 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2699 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2701 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2704 cur_inport_fluid
= scm_make_fluid ();
2705 cur_outport_fluid
= scm_make_fluid ();
2706 cur_errport_fluid
= scm_make_fluid ();
2707 cur_loadport_fluid
= scm_make_fluid ();
2709 scm_i_port_weak_hash
= scm_make_weak_key_hash_table (SCM_I_MAKINUM(31));
2711 #include "libguile/ports.x"
2713 /* Use Latin-1 as the default port encoding. */
2714 SCM_VARIABLE_SET (default_port_encoding_var
,
2715 scm_make_fluid_with_default (SCM_BOOL_F
));
2716 scm_port_encoding_init
= 1;
2718 SCM_VARIABLE_SET (default_conversion_strategy_var
,
2719 scm_make_fluid_with_default (sym_substitute
));
2720 scm_conversion_strategy_init
= 1;
2722 /* These bindings are used when boot-9 turns `current-input-port' et
2723 al into parameters. They are then removed from the guile module. */
2724 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2725 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2726 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);