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>
38 #include <c-strcase.h>
42 #include "libguile/_scm.h"
43 #include "libguile/async.h"
44 #include "libguile/deprecation.h"
45 #include "libguile/eval.h"
46 #include "libguile/fports.h" /* direct access for seek and truncate */
47 #include "libguile/goops.h"
48 #include "libguile/smob.h"
49 #include "libguile/chars.h"
50 #include "libguile/dynwind.h"
52 #include "libguile/keywords.h"
53 #include "libguile/hashtab.h"
54 #include "libguile/root.h"
55 #include "libguile/strings.h"
56 #include "libguile/mallocs.h"
57 #include "libguile/validate.h"
58 #include "libguile/ports.h"
59 #include "libguile/ports-internal.h"
60 #include "libguile/vectors.h"
61 #include "libguile/weaks.h"
62 #include "libguile/fluids.h"
63 #include "libguile/eq.h"
64 #include "libguile/alist.h"
78 #ifdef HAVE_SYS_IOCTL_H
79 #include <sys/ioctl.h>
82 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
83 already, but have this code here in case that wasn't so in past versions,
84 or perhaps to help other minimal DOS environments.
86 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
87 might be possibilities if we've got other systems without ftruncate. */
89 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
90 #define ftruncate(fd, size) chsize (fd, size)
92 #define HAVE_FTRUNCATE 1
96 /* The port kind table --- a dynamically resized array of port types. */
99 /* scm_ptobs scm_numptob
100 * implement a dynamically resized array of ptob records.
101 * Indexes into this table are used when generating type
102 * tags for smobjects (if you know a tag you can get an index and conversely).
104 scm_t_ptob_descriptor
*scm_ptobs
= NULL
;
105 long scm_numptob
= 0;
107 /* GC marker for a port with stream of SCM type. */
109 scm_markstream (SCM ptr
)
112 openp
= SCM_CELL_WORD_0 (ptr
) & SCM_OPN
;
114 return SCM_PACK (SCM_STREAM (ptr
));
120 * We choose to use an interface similar to the smob interface with
121 * fill_input and write as standard fields, passed to the port
122 * type constructor, and optional fields set by setters.
126 flush_port_default (SCM port SCM_UNUSED
)
131 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
136 scm_make_port_type (char *name
,
137 int (*fill_input
) (SCM port
),
138 void (*write
) (SCM port
, const void *data
, size_t size
))
141 if (SCM_I_MAX_PORT_TYPE_COUNT
- 1 <= scm_numptob
)
143 SCM_CRITICAL_SECTION_START
;
144 tmp
= (char *) scm_gc_realloc ((char *) scm_ptobs
,
145 scm_numptob
* sizeof (scm_t_ptob_descriptor
),
147 * sizeof (scm_t_ptob_descriptor
),
151 scm_ptobs
= (scm_t_ptob_descriptor
*) tmp
;
153 scm_ptobs
[scm_numptob
].name
= name
;
154 scm_ptobs
[scm_numptob
].mark
= 0;
155 scm_ptobs
[scm_numptob
].free
= NULL
;
156 scm_ptobs
[scm_numptob
].print
= scm_port_print
;
157 scm_ptobs
[scm_numptob
].equalp
= 0;
158 scm_ptobs
[scm_numptob
].close
= 0;
160 scm_ptobs
[scm_numptob
].write
= write
;
161 scm_ptobs
[scm_numptob
].flush
= flush_port_default
;
163 scm_ptobs
[scm_numptob
].end_input
= end_input_default
;
164 scm_ptobs
[scm_numptob
].fill_input
= fill_input
;
165 scm_ptobs
[scm_numptob
].input_waiting
= 0;
167 scm_ptobs
[scm_numptob
].seek
= 0;
168 scm_ptobs
[scm_numptob
].truncate
= 0;
172 SCM_CRITICAL_SECTION_END
;
176 scm_memory_error ("scm_make_port_type");
178 /* Make a class object if Goops is present */
179 if (SCM_UNPACK (scm_port_class
[0]) != 0)
180 scm_make_port_classes (scm_numptob
- 1, SCM_PTOBNAME (scm_numptob
- 1));
181 return scm_tc7_port
+ (scm_numptob
- 1) * 256;
185 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
187 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].mark
= mark
;
191 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
193 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].free
= free
;
197 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
198 scm_print_state
*pstate
))
200 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].print
= print
;
204 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
206 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].equalp
= equalp
;
210 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
212 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].flush
= flush
;
216 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
218 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].end_input
= end_input
;
222 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
224 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].close
= close
;
228 scm_set_port_seek (scm_t_bits tc
,
229 scm_t_off (*seek
) (SCM
, scm_t_off
, int))
231 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].seek
= seek
;
235 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
237 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].truncate
= truncate
;
241 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
243 scm_ptobs
[SCM_TC2PTOBNUM (tc
)].input_waiting
= input_waiting
;
247 scm_i_set_pending_eof (SCM port
)
249 SCM_PORT_GET_INTERNAL (port
)->pending_eof
= 1;
253 scm_i_clear_pending_eof (SCM port
)
255 SCM_PORT_GET_INTERNAL (port
)->pending_eof
= 0;
258 SCM_DEFINE (scm_i_port_property
, "%port-property", 2, 0, 0,
260 "Return the property of @var{port} associated with @var{key}.")
261 #define FUNC_NAME s_scm_i_port_property
263 SCM_VALIDATE_OPPORT (1, port
);
264 return scm_assq_ref (SCM_PORT_GET_INTERNAL (port
)->alist
, key
);
268 SCM_DEFINE (scm_i_set_port_property_x
, "%set-port-property!", 3, 0, 0,
269 (SCM port
, SCM key
, SCM value
),
270 "Set the property of @var{port} associated with @var{key} to @var{value}.")
271 #define FUNC_NAME s_scm_i_set_port_property_x
273 scm_t_port_internal
*pti
;
275 SCM_VALIDATE_OPPORT (1, port
);
276 pti
= SCM_PORT_GET_INTERNAL (port
);
277 pti
->alist
= scm_assq_set_x (pti
->alist
, key
, value
);
278 return SCM_UNSPECIFIED
;
284 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
286 "Return @code{#t} if a character is ready on input @var{port}\n"
287 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
288 "@code{#t} then the next @code{read-char} operation on\n"
289 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
290 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
292 "@code{char-ready?} exists to make it possible for a\n"
293 "program to accept characters from interactive ports without\n"
294 "getting stuck waiting for input. Any input editors associated\n"
295 "with such ports must make sure that characters whose existence\n"
296 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
297 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
298 "a port at end of file would be indistinguishable from an\n"
299 "interactive port that has no ready characters.")
300 #define FUNC_NAME s_scm_char_ready_p
304 if (SCM_UNBNDP (port
))
305 port
= scm_current_input_port ();
306 /* It's possible to close the current input port, so validate even in
308 SCM_VALIDATE_OPINPORT (1, port
);
310 pt
= SCM_PTAB_ENTRY (port
);
312 /* if the current read buffer is filled, or the
313 last pushed-back char has been read and the saved buffer is
314 filled, result is true. */
315 if (pt
->read_pos
< pt
->read_end
316 || (pt
->read_buf
== pt
->putback_buf
317 && pt
->saved_read_pos
< pt
->saved_read_end
))
321 scm_t_ptob_descriptor
*ptob
= &scm_ptobs
[SCM_PTOBNUM (port
)];
323 if (ptob
->input_waiting
)
324 return scm_from_bool(ptob
->input_waiting (port
));
331 /* Move up to READ_LEN bytes from PORT's putback and/or read buffers
332 into memory starting at DEST. Return the number of bytes moved.
333 PORT's line/column numbers are left unchanged. */
335 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
337 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
338 size_t bytes_read
= 0;
339 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
343 memcpy (dest
, pt
->read_pos
, from_buf
);
344 pt
->read_pos
+= from_buf
;
345 bytes_read
+= from_buf
;
346 read_len
-= from_buf
;
350 /* if putback was active, try the real input buffer too. */
351 if (pt
->read_buf
== pt
->putback_buf
)
353 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
356 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
357 pt
->saved_read_pos
+= from_buf
;
358 bytes_read
+= from_buf
;
365 /* Clear a port's read buffers, returning the contents. */
366 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
368 "This procedure clears a port's input buffers, similar\n"
369 "to the way that force-output clears the output buffer. The\n"
370 "contents of the buffers are returned as a single string, e.g.,\n"
373 "(define p (open-input-file ...))\n"
374 "(drain-input p) => empty string, nothing buffered yet.\n"
375 "(unread-char (read-char p) p)\n"
376 "(drain-input p) => initial chars from p, up to the buffer size.\n"
378 "Draining the buffers may be useful for cleanly finishing\n"
379 "buffered I/O so that the file descriptor can be used directly\n"
380 "for further input.")
381 #define FUNC_NAME s_scm_drain_input
388 SCM_VALIDATE_OPINPORT (1, port
);
389 pt
= SCM_PTAB_ENTRY (port
);
391 count
= pt
->read_end
- pt
->read_pos
;
392 if (pt
->read_buf
== pt
->putback_buf
)
393 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
397 result
= scm_i_make_string (count
, &data
, 0);
398 scm_take_from_input_buffers (port
, data
, count
);
401 result
= scm_nullstr
;
408 /* Standard ports --- current input, output, error, and more(!). */
410 static SCM cur_inport_fluid
= SCM_BOOL_F
;
411 static SCM cur_outport_fluid
= SCM_BOOL_F
;
412 static SCM cur_errport_fluid
= SCM_BOOL_F
;
413 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
415 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
417 "Return the current input port. This is the default port used\n"
418 "by many input procedures. Initially, @code{current-input-port}\n"
419 "returns the @dfn{standard input} in Unix and C terminology.")
420 #define FUNC_NAME s_scm_current_input_port
422 if (scm_is_true (cur_inport_fluid
))
423 return scm_fluid_ref (cur_inport_fluid
);
429 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
431 "Return the current output port. This is the default port used\n"
432 "by many output procedures. Initially,\n"
433 "@code{current-output-port} returns the @dfn{standard output} in\n"
434 "Unix and C terminology.")
435 #define FUNC_NAME s_scm_current_output_port
437 if (scm_is_true (cur_outport_fluid
))
438 return scm_fluid_ref (cur_outport_fluid
);
444 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
446 "Return the port to which errors and warnings should be sent (the\n"
447 "@dfn{standard error} in Unix and C terminology).")
448 #define FUNC_NAME s_scm_current_error_port
450 if (scm_is_true (cur_errport_fluid
))
451 return scm_fluid_ref (cur_errport_fluid
);
458 scm_current_warning_port (void)
460 static SCM cwp_var
= SCM_UNDEFINED
;
461 static scm_i_pthread_mutex_t cwp_var_mutex
462 = SCM_I_PTHREAD_MUTEX_INITIALIZER
;
464 scm_i_scm_pthread_mutex_lock (&cwp_var_mutex
);
465 if (SCM_UNBNDP (cwp_var
))
466 cwp_var
= scm_c_private_variable ("guile", "current-warning-port");
467 scm_i_pthread_mutex_unlock (&cwp_var_mutex
);
469 return scm_call_0 (scm_variable_ref (cwp_var
));
472 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
474 "Return the current-load-port.\n"
475 "The load port is used internally by @code{primitive-load}.")
476 #define FUNC_NAME s_scm_current_load_port
478 return scm_fluid_ref (cur_loadport_fluid
);
482 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
484 "@deffnx {Scheme Procedure} set-current-output-port port\n"
485 "@deffnx {Scheme Procedure} set-current-error-port port\n"
486 "Change the ports returned by @code{current-input-port},\n"
487 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
488 "so that they use the supplied @var{port} for input or output.")
489 #define FUNC_NAME s_scm_set_current_input_port
491 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
492 SCM_VALIDATE_OPINPORT (1, port
);
493 scm_fluid_set_x (cur_inport_fluid
, port
);
499 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
501 "Set the current default output port to @var{port}.")
502 #define FUNC_NAME s_scm_set_current_output_port
504 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
505 port
= SCM_COERCE_OUTPORT (port
);
506 SCM_VALIDATE_OPOUTPORT (1, port
);
507 scm_fluid_set_x (cur_outport_fluid
, port
);
513 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
515 "Set the current default error port to @var{port}.")
516 #define FUNC_NAME s_scm_set_current_error_port
518 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
519 port
= SCM_COERCE_OUTPORT (port
);
520 SCM_VALIDATE_OPOUTPORT (1, port
);
521 scm_fluid_set_x (cur_errport_fluid
, port
);
528 scm_set_current_warning_port (SCM port
)
530 static SCM cwp_var
= SCM_BOOL_F
;
532 if (scm_is_false (cwp_var
))
533 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
535 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
540 scm_dynwind_current_input_port (SCM port
)
541 #define FUNC_NAME NULL
543 SCM_VALIDATE_OPINPORT (1, port
);
544 scm_dynwind_fluid (cur_inport_fluid
, port
);
549 scm_dynwind_current_output_port (SCM port
)
550 #define FUNC_NAME NULL
552 port
= SCM_COERCE_OUTPORT (port
);
553 SCM_VALIDATE_OPOUTPORT (1, port
);
554 scm_dynwind_fluid (cur_outport_fluid
, port
);
559 scm_dynwind_current_error_port (SCM port
)
560 #define FUNC_NAME NULL
562 port
= SCM_COERCE_OUTPORT (port
);
563 SCM_VALIDATE_OPOUTPORT (1, port
);
564 scm_dynwind_fluid (cur_errport_fluid
, port
);
569 scm_i_dynwind_current_load_port (SCM port
)
571 scm_dynwind_fluid (cur_loadport_fluid
, port
);
575 /* The port table --- an array of pointers to ports. */
578 We need a global registry of ports to flush them all at exit, and to
579 get all the ports matching a file descriptor.
581 SCM scm_i_port_weak_hash
;
583 scm_i_pthread_mutex_t scm_i_port_table_mutex
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
586 /* Port finalization. */
589 static void finalize_port (void *, void *);
591 /* Register a finalizer for PORT. */
592 static SCM_C_INLINE_KEYWORD
void
593 register_finalizer_for_port (SCM port
)
595 /* Register a finalizer for PORT so that its
596 type's `free' function gets called. */
597 scm_i_set_finalizer (SCM2PTR (port
), finalize_port
, NULL
);
600 /* Finalize the object (a port) pointed to by PTR. */
602 finalize_port (void *ptr
, void *data
)
605 SCM port
= PTR2SCM (ptr
);
607 if (!SCM_PORTP (port
))
610 if (SCM_OPENP (port
))
612 if (SCM_REVEALED (port
) > 0)
613 /* Keep "revealed" ports alive and re-register a finalizer. */
614 register_finalizer_for_port (port
);
617 port_type
= SCM_TC2PTOBNUM (SCM_CELL_TYPE (port
));
618 if (port_type
>= scm_numptob
)
621 if (scm_ptobs
[port_type
].free
)
622 /* Yes, I really do mean `.free' rather than `.close'. `.close'
623 is for explicit `close-port' by user. */
624 scm_ptobs
[port_type
].free (port
);
626 SCM_SETSTREAM (port
, 0);
627 SCM_CLR_PORT_OPEN_FLAG (port
);
629 scm_gc_ports_collected
++;
638 /* This function is not and should not be thread safe. */
640 scm_new_port_table_entry (scm_t_bits tag
)
641 #define FUNC_NAME "scm_new_port_table_entry"
644 We initialize the cell to empty, this is in case scm_gc_calloc
645 triggers GC ; we don't want the GC to scan a half-finished Z.
648 SCM z
= scm_cons (SCM_EOL
, SCM_EOL
);
649 scm_t_port
*entry
= scm_gc_typed_calloc (scm_t_port
);
650 scm_t_port_internal
*pti
= scm_gc_typed_calloc (scm_t_port_internal
);
651 const char *encoding
;
653 entry
->file_name
= SCM_BOOL_F
;
654 entry
->rw_active
= SCM_PORT_NEITHER
;
657 /* Initialize this port with the thread's current default
659 encoding
= scm_i_default_port_encoding ();
660 entry
->ilseq_handler
= scm_i_default_port_conversion_handler ();
661 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
662 if (encoding
&& c_strcasecmp (encoding
, "UTF-8") == 0)
663 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
665 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
666 pti
->iconv_descriptors
= NULL
;
668 pti
->at_stream_start_for_bom_read
= 1;
669 pti
->at_stream_start_for_bom_write
= 1;
671 /* XXX These fields are not what they seem. They have been
672 repurposed, but cannot safely be renamed in 2.0 without breaking
673 ABI compatibility. This will be cleaned up in 2.2. */
674 entry
->input_cd
= pti
; /* XXX pointer to the internal port structure */
675 entry
->output_cd
= NULL
; /* XXX unused */
677 pti
->pending_eof
= 0;
678 pti
->alist
= SCM_EOL
;
680 SCM_SET_CELL_TYPE (z
, tag
);
681 SCM_SETPTAB_ENTRY (z
, entry
);
683 scm_hashq_set_x (scm_i_port_weak_hash
, z
, SCM_BOOL_F
);
685 /* For each new port, register a finalizer so that it port type's free
686 function can be invoked eventually. */
687 register_finalizer_for_port (z
);
693 #if SCM_ENABLE_DEPRECATED==1
695 scm_add_to_port_table (SCM port
)
700 scm_c_issue_deprecation_warning ("scm_add_to_port_table is deprecated.");
702 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex
);
703 z
= scm_new_port_table_entry (scm_tc7_port
);
704 pt
= SCM_PTAB_ENTRY(z
);
706 SCM_SETCAR (z
, SCM_EOL
);
707 SCM_SETCDR (z
, SCM_EOL
);
708 SCM_SETPTAB_ENTRY (port
, pt
);
709 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
716 /* Remove a port from the table and destroy it. */
718 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
721 scm_i_remove_port (SCM port
)
722 #define FUNC_NAME "scm_remove_port"
725 scm_t_port_internal
*pti
;
727 scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex
);
729 p
= SCM_PTAB_ENTRY (port
);
730 pti
= SCM_PORT_GET_INTERNAL (port
);
731 scm_port_non_buffer (p
);
732 p
->putback_buf
= NULL
;
733 p
->putback_buf_size
= 0;
735 if (pti
->iconv_descriptors
)
737 close_iconv_descriptors (pti
->iconv_descriptors
);
738 pti
->iconv_descriptors
= NULL
;
741 SCM_SETPTAB_ENTRY (port
, 0);
743 scm_hashq_remove_x (scm_i_port_weak_hash
, port
);
745 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
750 /* Functions for debugging. */
752 SCM_DEFINE (scm_pt_size
, "pt-size", 0, 0, 0,
754 "Return the number of ports in the port table. @code{pt-size}\n"
755 "is only included in @code{--enable-guile-debug} builds.")
756 #define FUNC_NAME s_scm_pt_size
758 return scm_from_int (SCM_HASHTABLE_N_ITEMS (scm_i_port_weak_hash
));
764 scm_port_non_buffer (scm_t_port
*pt
)
766 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
767 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
768 pt
->read_buf_size
= pt
->write_buf_size
= 1;
769 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
773 /* Revealed counts --- an oddity inherited from SCSH. */
775 /* Find a port in the table and return its revealed count.
776 Also used by the garbage collector.
780 scm_revealed_count (SCM port
)
782 return SCM_REVEALED(port
);
787 /* Return the revealed count for a port. */
789 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
791 "Return the revealed count for @var{port}.")
792 #define FUNC_NAME s_scm_port_revealed
794 port
= SCM_COERCE_OUTPORT (port
);
795 SCM_VALIDATE_OPENPORT (1, port
);
796 return scm_from_int (scm_revealed_count (port
));
800 /* Set the revealed count for a port. */
801 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
802 (SCM port
, SCM rcount
),
803 "Sets the revealed count for a port to a given value.\n"
804 "The return value is unspecified.")
805 #define FUNC_NAME s_scm_set_port_revealed_x
807 port
= SCM_COERCE_OUTPORT (port
);
808 SCM_VALIDATE_OPENPORT (1, port
);
809 SCM_REVEALED (port
) = scm_to_int (rcount
);
810 return SCM_UNSPECIFIED
;
816 /* Retrieving a port's mode. */
818 /* Return the flags that characterize a port based on the mode
819 * string used to open a file for that port.
821 * See PORT FLAGS in scm.h
825 scm_i_mode_bits_n (SCM modes
)
828 | (scm_i_string_contains_char (modes
, 'r')
829 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
830 | (scm_i_string_contains_char (modes
, 'w')
831 || scm_i_string_contains_char (modes
, 'a')
832 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
833 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
834 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
838 scm_mode_bits (char *modes
)
840 return scm_i_mode_bits (scm_from_locale_string (modes
));
844 scm_i_mode_bits (SCM modes
)
848 if (!scm_is_string (modes
))
849 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
851 bits
= scm_i_mode_bits_n (modes
);
852 scm_remember_upto_here_1 (modes
);
856 /* Return the mode flags from an open port.
857 * Some modes such as "append" are only used when opening
858 * a file and are not returned here. */
860 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
862 "Return the port modes associated with the open port @var{port}.\n"
863 "These will not necessarily be identical to the modes used when\n"
864 "the port was opened, since modes such as \"append\" which are\n"
865 "used only during port creation are not retained.")
866 #define FUNC_NAME s_scm_port_mode
871 port
= SCM_COERCE_OUTPORT (port
);
872 SCM_VALIDATE_OPPORT (1, port
);
873 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
874 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
875 strcpy (modes
, "r+");
879 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
881 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
883 return scm_from_locale_string (modes
);
892 * Call the close operation on a port object.
893 * see also scm_close.
895 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
897 "Close the specified port object. Return @code{#t} if it\n"
898 "successfully closes a port or @code{#f} if it was already\n"
899 "closed. An exception may be raised if an error occurs, for\n"
900 "example when flushing buffered output. See also @ref{Ports and\n"
901 "File Descriptors, close}, for a procedure which can close file\n"
903 #define FUNC_NAME s_scm_close_port
908 port
= SCM_COERCE_OUTPORT (port
);
910 SCM_VALIDATE_PORT (1, port
);
911 if (SCM_CLOSEDP (port
))
913 i
= SCM_PTOBNUM (port
);
914 if (scm_ptobs
[i
].close
)
915 rv
= (scm_ptobs
[i
].close
) (port
);
918 scm_i_remove_port (port
);
919 SCM_CLR_PORT_OPEN_FLAG (port
);
920 return scm_from_bool (rv
>= 0);
924 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
926 "Close the specified input port object. The routine has no effect if\n"
927 "the file has already been closed. An exception may be raised if an\n"
928 "error occurs. The value returned is unspecified.\n\n"
929 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
930 "which can close file descriptors.")
931 #define FUNC_NAME s_scm_close_input_port
933 SCM_VALIDATE_INPUT_PORT (1, port
);
934 scm_close_port (port
);
935 return SCM_UNSPECIFIED
;
939 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
941 "Close the specified output port object. The routine has no effect if\n"
942 "the file has already been closed. An exception may be raised if an\n"
943 "error occurs. The value returned is unspecified.\n\n"
944 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
945 "which can close file descriptors.")
946 #define FUNC_NAME s_scm_close_output_port
948 port
= SCM_COERCE_OUTPORT (port
);
949 SCM_VALIDATE_OUTPUT_PORT (1, port
);
950 scm_close_port (port
);
951 return SCM_UNSPECIFIED
;
956 collect_keys (void *unused
, SCM key
, SCM value
, SCM result
)
958 return scm_cons (key
, result
);
962 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
966 /* Copy out the port table as a list so that we get strong references
967 to all the values. */
968 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex
);
969 ports
= scm_internal_hash_fold (collect_keys
, NULL
,
970 SCM_EOL
, scm_i_port_weak_hash
);
971 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
973 for (; scm_is_pair (ports
); ports
= scm_cdr (ports
))
975 SCM p
= scm_car (ports
);
981 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
983 "Apply @var{proc} to each port in the Guile port table\n"
984 "in turn. The return value is unspecified. More specifically,\n"
985 "@var{proc} is applied exactly once to every port that exists\n"
986 "in the system at the time @code{port-for-each} is invoked.\n"
987 "Changes to the port table while @code{port-for-each} is running\n"
988 "have no effect as far as @code{port-for-each} is concerned.")
989 #define FUNC_NAME s_scm_port_for_each
993 SCM_VALIDATE_PROC (1, proc
);
995 /* Copy out the port table as a list so that we get strong references
996 to all the values. */
997 scm_i_pthread_mutex_lock (&scm_i_port_table_mutex
);
998 ports
= scm_internal_hash_fold (collect_keys
, NULL
,
999 SCM_EOL
, scm_i_port_weak_hash
);
1000 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
1002 for (; scm_is_pair (ports
); ports
= scm_cdr (ports
))
1003 if (SCM_PORTP (SCM_CAR (ports
)))
1004 scm_call_1 (proc
, SCM_CAR (ports
));
1006 return SCM_UNSPECIFIED
;
1012 /* Utter miscellany. Gosh, we should clean this up some time. */
1014 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
1016 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
1017 "@code{#f}. Any object satisfying this predicate also satisfies\n"
1019 #define FUNC_NAME s_scm_input_port_p
1021 return scm_from_bool (SCM_INPUT_PORT_P (x
));
1025 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
1027 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
1028 "@code{#f}. Any object satisfying this predicate also satisfies\n"
1030 #define FUNC_NAME s_scm_output_port_p
1032 x
= SCM_COERCE_OUTPORT (x
);
1033 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
1037 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
1039 "Return a boolean indicating whether @var{x} is a port.\n"
1040 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
1042 #define FUNC_NAME s_scm_port_p
1044 return scm_from_bool (SCM_PORTP (x
));
1048 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
1050 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
1052 #define FUNC_NAME s_scm_port_closed_p
1054 SCM_VALIDATE_PORT (1, port
);
1055 return scm_from_bool (!SCM_OPPORTP (port
));
1059 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
1061 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
1062 "return @code{#f}.")
1063 #define FUNC_NAME s_scm_eof_object_p
1065 return scm_from_bool(SCM_EOF_OBJECT_P (x
));
1069 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
1071 "Flush the specified output port, or the current output port if @var{port}\n"
1072 "is omitted. The current output buffer contents are passed to the\n"
1073 "underlying port implementation (e.g., in the case of fports, the\n"
1074 "data will be written to the file and the output buffer will be cleared.)\n"
1075 "It has no effect on an unbuffered port.\n\n"
1076 "The return value is unspecified.")
1077 #define FUNC_NAME s_scm_force_output
1079 if (SCM_UNBNDP (port
))
1080 port
= scm_current_output_port ();
1083 port
= SCM_COERCE_OUTPORT (port
);
1084 SCM_VALIDATE_OPOUTPORT (1, port
);
1087 return SCM_UNSPECIFIED
;
1093 flush_output_port (void *closure
, SCM port
)
1095 if (SCM_OPOUTPORTP (port
))
1099 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
1101 "Equivalent to calling @code{force-output} on\n"
1102 "all open output ports. The return value is unspecified.")
1103 #define FUNC_NAME s_scm_flush_all_ports
1105 scm_c_port_for_each (&flush_output_port
, NULL
);
1106 return SCM_UNSPECIFIED
;
1110 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1112 "Return the next character available from @var{port}, updating\n"
1113 "@var{port} to point to the following character. If no more\n"
1114 "characters are available, the end-of-file object is returned.\n"
1116 "When @var{port}'s data cannot be decoded according to its\n"
1117 "character encoding, a @code{decoding-error} is raised and\n"
1118 "@var{port} points past the erroneous byte sequence.\n")
1119 #define FUNC_NAME s_scm_read_char
1122 if (SCM_UNBNDP (port
))
1123 port
= scm_current_input_port ();
1124 SCM_VALIDATE_OPINPORT (1, port
);
1125 c
= scm_getc (port
);
1128 return SCM_MAKE_CHAR (c
);
1132 /* Update the line and column number of PORT after consumption of C. */
1134 update_port_lf (scm_t_wchar c
, SCM port
)
1159 #define SCM_MBCHAR_BUF_SIZE (4)
1161 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1162 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1164 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1166 scm_t_wchar codepoint
;
1168 if (utf8_buf
[0] <= 0x7f)
1171 codepoint
= utf8_buf
[0];
1173 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1176 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1177 | (utf8_buf
[1] & 0x3f);
1179 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1182 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1183 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1184 | (utf8_buf
[2] & 0x3f);
1189 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1190 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1191 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1192 | (utf8_buf
[3] & 0x3f);
1198 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1199 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1200 representation, and set *LEN to the length in bytes. Return
1201 `EILSEQ' on error. */
1203 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1204 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1206 #define ASSERT_NOT_EOF(b) \
1207 if (SCM_UNLIKELY ((b) == EOF)) \
1209 #define CONSUME_PEEKED_BYTE() \
1216 pt
= SCM_PTAB_ENTRY (port
);
1218 byte
= scm_get_byte_or_eof (port
);
1225 buf
[0] = (scm_t_uint8
) byte
;
1230 *codepoint
= buf
[0];
1231 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1234 byte
= scm_peek_byte_or_eof (port
);
1235 ASSERT_NOT_EOF (byte
);
1237 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1240 CONSUME_PEEKED_BYTE ();
1241 buf
[1] = (scm_t_uint8
) byte
;
1244 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1247 else if ((buf
[0] & 0xf0) == 0xe0)
1250 byte
= scm_peek_byte_or_eof (port
);
1251 ASSERT_NOT_EOF (byte
);
1253 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1254 || (buf
[0] == 0xe0 && byte
< 0xa0)
1255 || (buf
[0] == 0xed && byte
> 0x9f)))
1258 CONSUME_PEEKED_BYTE ();
1259 buf
[1] = (scm_t_uint8
) byte
;
1262 byte
= scm_peek_byte_or_eof (port
);
1263 ASSERT_NOT_EOF (byte
);
1265 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1268 CONSUME_PEEKED_BYTE ();
1269 buf
[2] = (scm_t_uint8
) byte
;
1272 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1273 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1276 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1279 byte
= scm_peek_byte_or_eof (port
);
1280 ASSERT_NOT_EOF (byte
);
1282 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1283 || (buf
[0] == 0xf0 && byte
< 0x90)
1284 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1287 CONSUME_PEEKED_BYTE ();
1288 buf
[1] = (scm_t_uint8
) byte
;
1291 byte
= scm_peek_byte_or_eof (port
);
1292 ASSERT_NOT_EOF (byte
);
1294 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1297 CONSUME_PEEKED_BYTE ();
1298 buf
[2] = (scm_t_uint8
) byte
;
1301 byte
= scm_peek_byte_or_eof (port
);
1302 ASSERT_NOT_EOF (byte
);
1304 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1307 CONSUME_PEEKED_BYTE ();
1308 buf
[3] = (scm_t_uint8
) byte
;
1311 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1312 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1313 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1322 /* Here we could choose the consume the faulty byte when it's not a
1323 valid starting byte, but it's not a requirement. What Section 3.9
1324 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1325 would otherwise be a valid starting byte. */
1329 #undef CONSUME_PEEKED_BYTE
1330 #undef ASSERT_NOT_EOF
1333 /* Likewise, read a byte sequence from PORT, passing it through its
1334 input conversion descriptor. */
1336 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1337 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1339 scm_t_iconv_descriptors
*id
;
1340 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1341 size_t input_size
= 0;
1343 id
= scm_i_port_iconv_descriptors (port
, SCM_PORT_READ
);
1348 char *input
, *output
;
1349 size_t input_left
, output_left
, done
;
1351 byte_read
= scm_get_byte_or_eof (port
);
1352 if (SCM_UNLIKELY (byte_read
== EOF
))
1354 if (SCM_LIKELY (input_size
== 0))
1356 *codepoint
= (scm_t_wchar
) EOF
;
1362 /* EOF found in the middle of a multibyte character. */
1363 scm_i_set_pending_eof (port
);
1368 buf
[input_size
++] = byte_read
;
1371 input_left
= input_size
;
1372 output
= (char *) utf8_buf
;
1373 output_left
= sizeof (utf8_buf
);
1375 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1377 if (done
== (size_t) -1)
1380 if (SCM_LIKELY (err
== EINVAL
))
1381 /* The input byte sequence did not form a complete
1382 character. Read another byte and try again. */
1389 size_t output_size
= sizeof (utf8_buf
) - output_left
;
1390 if (SCM_LIKELY (output_size
> 0))
1392 /* iconv generated output. Convert the UTF8_BUF sequence
1393 to a Unicode code point. */
1394 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1400 /* iconv consumed some bytes without producing any output.
1401 Most likely this means that a Unicode byte-order mark
1402 (BOM) was consumed, which should not be included in the
1403 returned buf. Shift any remaining bytes to the beginning
1404 of buf, and continue the loop. */
1405 memmove (buf
, input
, input_left
);
1406 input_size
= input_left
;
1413 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1414 with the byte representation of the codepoint in PORT's encoding, and
1415 set *LEN to the length in bytes of that representation. Return 0 on
1416 success and an errno value on error. */
1418 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1419 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1422 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1423 scm_t_port_internal
*pti
= SCM_PORT_GET_INTERNAL (port
);
1425 if (pti
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1426 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1428 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1430 if (SCM_LIKELY (err
== 0))
1432 if (SCM_UNLIKELY (pti
->at_stream_start_for_bom_read
))
1434 /* Record that we're no longer at stream start. */
1435 pti
->at_stream_start_for_bom_read
= 0;
1437 pti
->at_stream_start_for_bom_write
= 0;
1439 /* If we just read a BOM in an encoding that recognizes them,
1440 then silently consume it and read another code point. */
1442 (*codepoint
== SCM_UNICODE_BOM
1443 && (pti
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
1444 || c_strcasecmp (pt
->encoding
, "UTF-16") == 0
1445 || c_strcasecmp (pt
->encoding
, "UTF-32") == 0)))
1446 return get_codepoint (port
, codepoint
, buf
, len
);
1448 update_port_lf (*codepoint
, port
);
1450 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1454 update_port_lf (*codepoint
, port
);
1460 /* Read a codepoint from PORT and return it. */
1463 #define FUNC_NAME "scm_getc"
1467 scm_t_wchar codepoint
;
1468 char buf
[SCM_MBCHAR_BUF_SIZE
];
1470 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1471 if (SCM_UNLIKELY (err
!= 0))
1472 /* At this point PORT should point past the invalid encoding, as per
1473 R6RS-lib Section 8.2.4. */
1474 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1480 /* this should only be called when the read buffer is empty. it
1481 tries to refill the read buffer. it returns the first char from
1482 the port, which is either EOF or *(pt->read_pos). */
1484 scm_i_fill_input (SCM port
)
1486 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1487 scm_t_port_internal
*pti
= SCM_PORT_GET_INTERNAL (port
);
1489 assert (pt
->read_pos
== pt
->read_end
);
1491 if (pti
->pending_eof
)
1493 pti
->pending_eof
= 0;
1497 if (pt
->read_buf
== pt
->putback_buf
)
1499 /* finished reading put-back chars. */
1500 pt
->read_buf
= pt
->saved_read_buf
;
1501 pt
->read_pos
= pt
->saved_read_pos
;
1502 pt
->read_end
= pt
->saved_read_end
;
1503 pt
->read_buf_size
= pt
->saved_read_buf_size
;
1504 if (pt
->read_pos
< pt
->read_end
)
1505 return *(pt
->read_pos
);
1507 return scm_ptobs
[SCM_PTOBNUM (port
)].fill_input (port
);
1511 scm_fill_input (SCM port
)
1513 return scm_i_fill_input (port
);
1516 /* Slow-path fallback for 'scm_get_byte_or_eof' in inline.h */
1518 scm_slow_get_byte_or_eof (SCM port
)
1520 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1522 if (pt
->rw_active
== SCM_PORT_WRITE
)
1526 pt
->rw_active
= SCM_PORT_READ
;
1528 if (pt
->read_pos
>= pt
->read_end
)
1530 if (SCM_UNLIKELY (scm_i_fill_input (port
) == EOF
))
1534 return *pt
->read_pos
++;
1537 /* Slow-path fallback for 'scm_peek_byte_or_eof' in inline.h */
1539 scm_slow_peek_byte_or_eof (SCM port
)
1541 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1543 if (pt
->rw_active
== SCM_PORT_WRITE
)
1547 pt
->rw_active
= SCM_PORT_READ
;
1549 if (pt
->read_pos
>= pt
->read_end
)
1551 if (SCM_UNLIKELY (scm_i_fill_input (port
) == EOF
))
1553 scm_i_set_pending_eof (port
);
1558 return *pt
->read_pos
;
1564 * This function differs from scm_c_write; it updates port line and
1568 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
1570 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1571 scm_t_ptob_descriptor
*ptob
= &scm_ptobs
[SCM_PTOBNUM (port
)];
1573 if (pt
->rw_active
== SCM_PORT_READ
)
1574 scm_end_input (port
);
1576 ptob
->write (port
, ptr
, size
);
1578 for (; size
; ptr
++, size
--)
1579 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
1582 pt
->rw_active
= SCM_PORT_WRITE
;
1585 /* Write STR to PORT from START inclusive to END exclusive. */
1587 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
1589 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1591 if (pt
->rw_active
== SCM_PORT_READ
)
1592 scm_end_input (port
);
1594 if (end
== (size_t) -1)
1595 end
= scm_i_string_length (str
);
1597 scm_i_display_substring (str
, start
, end
, port
);
1600 pt
->rw_active
= SCM_PORT_WRITE
;
1605 * Used by an application to read arbitrary number of bytes from an
1606 * SCM port. Same semantics as libc read, except that scm_c_read only
1607 * returns less than SIZE bytes if at end-of-file.
1609 * Warning: Doesn't update port line and column counts! */
1611 /* This structure, and the following swap_buffer function, are used
1612 for temporarily swapping a port's own read buffer, and the buffer
1613 that the caller of scm_c_read provides. */
1614 struct port_and_swap_buffer
1617 unsigned char *buffer
;
1622 swap_buffer (void *data
)
1624 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1625 unsigned char *old_buf
= psb
->pt
->read_buf
;
1626 size_t old_size
= psb
->pt
->read_buf_size
;
1628 /* Make the port use (buffer, size) from the struct. */
1629 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1630 psb
->pt
->read_buf_size
= psb
->size
;
1632 /* Save the port's old (buffer, size) in the struct. */
1633 psb
->buffer
= old_buf
;
1634 psb
->size
= old_size
;
1638 scm_c_read (SCM port
, void *buffer
, size_t size
)
1639 #define FUNC_NAME "scm_c_read"
1642 size_t n_read
= 0, n_available
;
1643 struct port_and_swap_buffer psb
;
1645 SCM_VALIDATE_OPINPORT (1, port
);
1647 pt
= SCM_PTAB_ENTRY (port
);
1648 if (pt
->rw_active
== SCM_PORT_WRITE
)
1649 scm_ptobs
[SCM_PTOBNUM (port
)].flush (port
);
1652 pt
->rw_active
= SCM_PORT_READ
;
1654 /* Take bytes first from the port's read buffer. */
1655 if (pt
->read_pos
< pt
->read_end
)
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
;
1665 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1669 /* Now we will call scm_i_fill_input repeatedly until we have read the
1670 requested number of bytes. (Note that a single scm_i_fill_input
1671 call does not guarantee to fill the whole of the port's read
1673 if (pt
->read_buf_size
<= 1 &&
1674 (pt
->encoding
== NULL
1675 || c_strcasecmp (pt
->encoding
, "ISO-8859-1") == 0))
1677 /* The port that we are reading from is unbuffered - i.e. does
1678 not have its own persistent buffer - but we have a buffer,
1679 provided by our caller, that is the right size for the data
1680 that is wanted. For the following scm_i_fill_input calls,
1681 therefore, we use the buffer in hand as the port's read
1684 We need to make sure that the port's normal (1 byte) buffer
1685 is reinstated in case one of the scm_i_fill_input () calls
1686 throws an exception; we use the scm_dynwind_* API to achieve
1689 A consequence of this optimization is that the fill_input
1690 functions can't unget characters. That'll push data to the
1691 pushback buffer instead of this psb buffer. */
1693 unsigned char *pback
= pt
->putback_buf
;
1696 psb
.buffer
= buffer
;
1698 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1699 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1700 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1702 /* Call scm_i_fill_input until we have all the bytes that we need,
1704 while (pt
->read_buf_size
&& (scm_i_fill_input (port
) != EOF
))
1706 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1707 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1710 if (pback
!= pt
->putback_buf
1711 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1712 scm_misc_error (FUNC_NAME
,
1713 "scm_c_read must not call a fill function that pushes "
1714 "back characters onto an unbuffered port", SCM_EOL
);
1716 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1718 /* Reinstate the port's normal buffer. */
1723 /* The port has its own buffer. It is important that we use it,
1724 even if it happens to be smaller than our caller's buffer, so
1725 that a custom port implementation's entry points (in
1726 particular, fill_input) can rely on the buffer always being
1727 the same as they first set up. */
1728 while (size
&& (scm_i_fill_input (port
) != EOF
))
1730 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1731 memcpy (buffer
, pt
->read_pos
, n_available
);
1732 buffer
= (char *) buffer
+ n_available
;
1733 pt
->read_pos
+= n_available
;
1734 n_read
+= n_available
;
1735 size
-= n_available
;
1745 * Used by an application to write arbitrary number of bytes to an SCM
1746 * port. Similar semantics as libc write. However, unlike libc
1747 * write, scm_c_write writes the requested number of bytes and has no
1750 * Warning: Doesn't update port line and column counts!
1754 scm_c_write (SCM port
, const void *ptr
, size_t size
)
1755 #define FUNC_NAME "scm_c_write"
1758 scm_t_ptob_descriptor
*ptob
;
1760 SCM_VALIDATE_OPOUTPORT (1, port
);
1762 pt
= SCM_PTAB_ENTRY (port
);
1763 ptob
= &scm_ptobs
[SCM_PTOBNUM (port
)];
1765 if (pt
->rw_active
== SCM_PORT_READ
)
1766 scm_end_input (port
);
1768 ptob
->write (port
, ptr
, size
);
1771 pt
->rw_active
= SCM_PORT_WRITE
;
1776 scm_flush (SCM port
)
1778 long i
= SCM_PTOBNUM (port
);
1780 (scm_ptobs
[i
].flush
) (port
);
1784 scm_end_input (SCM port
)
1787 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1789 scm_i_clear_pending_eof (port
);
1790 if (pt
->read_buf
== pt
->putback_buf
)
1792 offset
= pt
->read_end
- pt
->read_pos
;
1793 pt
->read_buf
= pt
->saved_read_buf
;
1794 pt
->read_pos
= pt
->saved_read_pos
;
1795 pt
->read_end
= pt
->saved_read_end
;
1796 pt
->read_buf_size
= pt
->saved_read_buf_size
;
1801 scm_ptobs
[SCM_PTOBNUM (port
)].end_input (port
, offset
);
1808 scm_i_unget_bytes (const unsigned char *buf
, size_t len
, SCM port
)
1809 #define FUNC_NAME "scm_unget_bytes"
1811 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1812 size_t old_len
, new_len
;
1814 scm_i_clear_pending_eof (port
);
1816 if (pt
->read_buf
!= pt
->putback_buf
)
1817 /* switch to the put-back buffer. */
1819 if (pt
->putback_buf
== NULL
)
1821 pt
->putback_buf_size
= (len
> SCM_INITIAL_PUTBACK_BUF_SIZE
1822 ? len
: SCM_INITIAL_PUTBACK_BUF_SIZE
);
1824 = (unsigned char *) scm_gc_malloc_pointerless
1825 (pt
->putback_buf_size
, "putback buffer");
1828 pt
->saved_read_buf
= pt
->read_buf
;
1829 pt
->saved_read_pos
= pt
->read_pos
;
1830 pt
->saved_read_end
= pt
->read_end
;
1831 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1833 /* Put read_pos at the end of the buffer, so that ungets will not
1834 have to shift the buffer contents each time. */
1835 pt
->read_buf
= pt
->putback_buf
;
1836 pt
->read_pos
= pt
->read_end
= pt
->putback_buf
+ pt
->putback_buf_size
;
1837 pt
->read_buf_size
= pt
->putback_buf_size
;
1840 old_len
= pt
->read_end
- pt
->read_pos
;
1841 new_len
= old_len
+ len
;
1843 if (new_len
> pt
->read_buf_size
)
1844 /* The putback buffer needs to be enlarged. */
1846 size_t new_buf_size
;
1847 unsigned char *new_buf
, *new_end
, *new_pos
;
1849 new_buf_size
= pt
->read_buf_size
* 2;
1850 if (new_buf_size
< new_len
)
1851 new_buf_size
= new_len
;
1853 new_buf
= (unsigned char *)
1854 scm_gc_malloc_pointerless (new_buf_size
, "putback buffer");
1856 /* Put the bytes at the end of the buffer, so that future
1857 ungets won't need to shift the buffer. */
1858 new_end
= new_buf
+ new_buf_size
;
1859 new_pos
= new_end
- old_len
;
1860 memcpy (new_pos
, pt
->read_pos
, old_len
);
1862 pt
->read_buf
= pt
->putback_buf
= new_buf
;
1863 pt
->read_pos
= new_pos
;
1864 pt
->read_end
= new_end
;
1865 pt
->read_buf_size
= pt
->putback_buf_size
= new_buf_size
;
1867 else if (pt
->read_buf
+ len
< pt
->read_pos
)
1868 /* If needed, shift the existing buffer contents up.
1869 This should not happen unless some external code
1870 manipulates the putback buffer pointers. */
1872 unsigned char *new_end
= pt
->read_buf
+ pt
->read_buf_size
;
1873 unsigned char *new_pos
= new_end
- old_len
;
1875 memmove (new_pos
, pt
->read_pos
, old_len
);
1876 pt
->read_pos
= new_pos
;
1877 pt
->read_end
= new_end
;
1880 /* Move read_pos back and copy the bytes there. */
1881 pt
->read_pos
-= len
;
1882 memcpy (pt
->read_buf
+ (pt
->read_pos
- pt
->read_buf
), buf
, len
);
1884 if (pt
->rw_active
== SCM_PORT_WRITE
)
1888 pt
->rw_active
= SCM_PORT_READ
;
1893 scm_unget_bytes (const unsigned char *buf
, size_t len
, SCM port
)
1895 scm_i_unget_bytes (buf
, len
, port
);
1899 scm_unget_byte (int c
, SCM port
)
1904 scm_i_unget_bytes (&byte
, 1, port
);
1908 scm_ungetc (scm_t_wchar c
, SCM port
)
1909 #define FUNC_NAME "scm_ungetc"
1911 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1913 char result_buf
[10];
1914 const char *encoding
;
1917 if (pt
->encoding
!= NULL
)
1918 encoding
= pt
->encoding
;
1920 encoding
= "ISO-8859-1";
1922 len
= sizeof (result_buf
);
1923 result
= u32_conv_to_encoding (encoding
,
1924 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1925 (uint32_t *) &c
, 1, NULL
,
1928 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1929 scm_encoding_error (FUNC_NAME
, errno
,
1930 "conversion to port encoding failed",
1931 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1933 scm_i_unget_bytes ((unsigned char *) result
, len
, port
);
1935 if (SCM_UNLIKELY (result
!= result_buf
))
1940 /* What should col be in this case?
1941 * We'll leave it at -1.
1943 SCM_LINUM (port
) -= 1;
1952 scm_ungets (const char *s
, int n
, SCM port
)
1954 /* This is simple minded and inefficient, but unreading strings is
1955 * probably not a common operation, and remember that line and
1956 * column numbers have to be handled...
1958 * Please feel free to write an optimized version!
1961 scm_ungetc (s
[n
], port
);
1965 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1967 "Return the next character available from @var{port},\n"
1968 "@emph{without} updating @var{port} to point to the following\n"
1969 "character. If no more characters are available, the\n"
1970 "end-of-file object is returned.\n"
1972 "The value returned by\n"
1973 "a call to @code{peek-char} is the same as the value that would\n"
1974 "have been returned by a call to @code{read-char} on the same\n"
1975 "port. The only difference is that the very next call to\n"
1976 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1977 "return the value returned by the preceding call to\n"
1978 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1979 "an interactive port will hang waiting for input whenever a call\n"
1980 "to @code{read-char} would have hung.\n"
1982 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1983 "if such a situation occurs. However, unlike with @code{read-char},\n"
1984 "@var{port} still points at the beginning of the erroneous byte\n"
1985 "sequence when the error is raised.\n")
1986 #define FUNC_NAME s_scm_peek_char
1991 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1995 if (SCM_UNBNDP (port
))
1996 port
= scm_current_input_port ();
1997 SCM_VALIDATE_OPINPORT (1, port
);
1999 column
= SCM_COL (port
);
2000 line
= SCM_LINUM (port
);
2002 err
= get_codepoint (port
, &c
, bytes
, &len
);
2004 scm_i_unget_bytes ((unsigned char *) bytes
, len
, port
);
2006 SCM_COL (port
) = column
;
2007 SCM_LINUM (port
) = line
;
2009 if (SCM_UNLIKELY (err
!= 0))
2011 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
2013 /* Shouldn't happen since `catch' always aborts to prompt. */
2014 result
= SCM_BOOL_F
;
2018 scm_i_set_pending_eof (port
);
2019 result
= SCM_EOF_VAL
;
2022 result
= SCM_MAKE_CHAR (c
);
2028 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2029 (SCM cobj
, SCM port
),
2030 "Place character @var{cobj} in @var{port} so that it will be\n"
2031 "read by the next read operation. If called multiple times, the\n"
2032 "unread characters will be read again in last-in first-out\n"
2033 "order. If @var{port} is not supplied, the current input port\n"
2035 #define FUNC_NAME s_scm_unread_char
2039 SCM_VALIDATE_CHAR (1, cobj
);
2040 if (SCM_UNBNDP (port
))
2041 port
= scm_current_input_port ();
2042 SCM_VALIDATE_OPINPORT (2, port
);
2044 c
= SCM_CHAR (cobj
);
2046 scm_ungetc (c
, port
);
2051 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2052 (SCM str
, SCM port
),
2053 "Place the string @var{str} in @var{port} so that its characters will be\n"
2054 "read in subsequent read operations. If called multiple times, the\n"
2055 "unread characters will be read again in last-in first-out order. If\n"
2056 "@var{port} is not supplied, the current-input-port is used.")
2057 #define FUNC_NAME s_scm_unread_string
2060 SCM_VALIDATE_STRING (1, str
);
2061 if (SCM_UNBNDP (port
))
2062 port
= scm_current_input_port ();
2063 SCM_VALIDATE_OPINPORT (2, port
);
2065 n
= scm_i_string_length (str
);
2068 scm_ungetc (scm_i_string_ref (str
, n
), port
);
2074 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2075 (SCM fd_port
, SCM offset
, SCM whence
),
2076 "Sets the current position of @var{fd_port} to the integer\n"
2077 "@var{offset}, which is interpreted according to the value of\n"
2080 "One of the following variables should be supplied for\n"
2082 "@defvar SEEK_SET\n"
2083 "Seek from the beginning of the file.\n"
2085 "@defvar SEEK_CUR\n"
2086 "Seek from the current position.\n"
2088 "@defvar SEEK_END\n"
2089 "Seek from the end of the file.\n"
2091 "If @var{fd_port} is a file descriptor, the underlying system\n"
2092 "call is @code{lseek}. @var{port} may be a string port.\n"
2094 "The value returned is the new position in the file. This means\n"
2095 "that the current position of a port can be obtained using:\n"
2097 "(seek port 0 SEEK_CUR)\n"
2099 #define FUNC_NAME s_scm_seek
2103 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2105 how
= scm_to_int (whence
);
2106 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2107 SCM_OUT_OF_RANGE (3, whence
);
2109 if (SCM_OPPORTP (fd_port
))
2111 scm_t_port_internal
*pti
= SCM_PORT_GET_INTERNAL (fd_port
);
2112 scm_t_ptob_descriptor
*ptob
= scm_ptobs
+ SCM_PTOBNUM (fd_port
);
2113 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2114 off_t_or_off64_t rv
;
2117 SCM_MISC_ERROR ("port is not seekable",
2118 scm_cons (fd_port
, SCM_EOL
));
2120 rv
= ptob
->seek (fd_port
, off
, how
);
2122 /* Set stream-start flags according to new position. */
2123 pti
->at_stream_start_for_bom_read
= (rv
== 0);
2124 pti
->at_stream_start_for_bom_write
= (rv
== 0);
2126 scm_i_clear_pending_eof (fd_port
);
2128 return scm_from_off_t_or_off64_t (rv
);
2130 else /* file descriptor?. */
2132 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2133 off_t_or_off64_t rv
;
2134 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2137 return scm_from_off_t_or_off64_t (rv
);
2146 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2147 doesn't have the filename version truncate(), hence this code. */
2148 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2150 truncate (const char *file
, off_t length
)
2154 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2158 ret
= ftruncate (fdes
, length
);
2161 int save_errno
= errno
;
2167 return close (fdes
);
2169 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2171 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2172 (SCM object
, SCM length
),
2173 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2174 "can be a filename string, a port object, or an integer file\n"
2176 "The return value is unspecified.\n"
2178 "For a port or file descriptor @var{length} can be omitted, in\n"
2179 "which case the file is truncated at the current position (per\n"
2180 "@code{ftell} above).\n"
2182 "On most systems a file can be extended by giving a length\n"
2183 "greater than the current size, but this is not mandatory in the\n"
2185 #define FUNC_NAME s_scm_truncate_file
2189 /* "object" can be a port, fdes or filename.
2191 Negative "length" makes no sense, but it's left to truncate() or
2192 ftruncate() to give back an error for that (normally EINVAL).
2195 if (SCM_UNBNDP (length
))
2197 /* must supply length if object is a filename. */
2198 if (scm_is_string (object
))
2199 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2201 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2204 object
= SCM_COERCE_OUTPORT (object
);
2205 if (scm_is_integer (object
))
2207 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2208 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2211 else if (SCM_OPOUTPORTP (object
))
2213 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2214 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2215 scm_t_ptob_descriptor
*ptob
= scm_ptobs
+ SCM_PTOBNUM (object
);
2217 if (!ptob
->truncate
)
2218 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2220 scm_i_clear_pending_eof (object
);
2221 if (pt
->rw_active
== SCM_PORT_READ
)
2222 scm_end_input (object
);
2223 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2224 ptob
->flush (object
);
2226 ptob
->truncate (object
, c_length
);
2231 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2232 char *str
= scm_to_locale_string (object
);
2234 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2241 return SCM_UNSPECIFIED
;
2245 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2247 "Return the current line number for @var{port}.\n"
2249 "The first line of a file is 0. But you might want to add 1\n"
2250 "when printing line numbers, since starting from 1 is\n"
2251 "traditional in error messages, and likely to be more natural to\n"
2253 #define FUNC_NAME s_scm_port_line
2255 port
= SCM_COERCE_OUTPORT (port
);
2256 SCM_VALIDATE_OPENPORT (1, port
);
2257 return scm_from_long (SCM_LINUM (port
));
2261 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2262 (SCM port
, SCM line
),
2263 "Set the current line number for @var{port} to @var{line}. The\n"
2264 "first line of a file is 0.")
2265 #define FUNC_NAME s_scm_set_port_line_x
2267 port
= SCM_COERCE_OUTPORT (port
);
2268 SCM_VALIDATE_OPENPORT (1, port
);
2269 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2270 return SCM_UNSPECIFIED
;
2274 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2276 "Return the current column number of @var{port}.\n"
2277 "If the number is\n"
2278 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2279 "- i.e. the first character of the first line is line 0, column 0.\n"
2280 "(However, when you display a file position, for example in an error\n"
2281 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2282 "because lines and column numbers traditionally start with 1, and that is\n"
2283 "what non-programmers will find most natural.)")
2284 #define FUNC_NAME s_scm_port_column
2286 port
= SCM_COERCE_OUTPORT (port
);
2287 SCM_VALIDATE_OPENPORT (1, port
);
2288 return scm_from_int (SCM_COL (port
));
2292 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2293 (SCM port
, SCM column
),
2294 "Set the current column of @var{port}. Before reading the first\n"
2295 "character on a line the column should be 0.")
2296 #define FUNC_NAME s_scm_set_port_column_x
2298 port
= SCM_COERCE_OUTPORT (port
);
2299 SCM_VALIDATE_OPENPORT (1, port
);
2300 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2301 return SCM_UNSPECIFIED
;
2305 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2307 "Return the filename associated with @var{port}, or @code{#f}\n"
2308 "if no filename is associated with the port.")
2309 #define FUNC_NAME s_scm_port_filename
2311 port
= SCM_COERCE_OUTPORT (port
);
2312 SCM_VALIDATE_OPENPORT (1, port
);
2313 return SCM_FILENAME (port
);
2317 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2318 (SCM port
, SCM filename
),
2319 "Change the filename associated with @var{port}, using the current input\n"
2320 "port if none is specified. Note that this does not change the port's\n"
2321 "source of data, but only the value that is returned by\n"
2322 "@code{port-filename} and reported in diagnostic output.")
2323 #define FUNC_NAME s_scm_set_port_filename_x
2325 port
= SCM_COERCE_OUTPORT (port
);
2326 SCM_VALIDATE_OPENPORT (1, port
);
2327 /* We allow the user to set the filename to whatever he likes. */
2328 SCM_SET_FILENAME (port
, filename
);
2329 return SCM_UNSPECIFIED
;
2333 /* A fluid specifying the default encoding for newly created ports. If it is
2334 a string, that is the encoding. If it is #f, it is in the "native"
2335 (Latin-1) encoding. */
2336 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
2338 static int scm_port_encoding_init
= 0;
2340 /* Use ENCODING as the default encoding for future ports. */
2342 scm_i_set_default_port_encoding (const char *encoding
)
2344 if (!scm_port_encoding_init
2345 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
2346 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
2349 if (encoding
== NULL
2350 || c_strcasecmp (encoding
, "ASCII") == 0
2351 || c_strcasecmp (encoding
, "ANSI_X3.4-1968") == 0
2352 || c_strcasecmp (encoding
, "ISO-8859-1") == 0)
2353 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
2355 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
2356 scm_from_locale_string (encoding
));
2359 /* Return the name of the default encoding for newly created ports; a
2360 return value of NULL means "ISO-8859-1". */
2362 scm_i_default_port_encoding (void)
2364 if (!scm_port_encoding_init
)
2366 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
2372 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
2373 if (!scm_is_string (encoding
))
2376 return scm_i_string_chars (encoding
);
2380 /* If the next LEN bytes from PORT are equal to those in BYTES, then
2381 return 1, else return 0. Leave the port position unchanged. */
2383 looking_at_bytes (SCM port
, const unsigned char *bytes
, int len
)
2385 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2388 while (i
< len
&& scm_peek_byte_or_eof (port
) == bytes
[i
])
2393 scm_i_unget_bytes (bytes
, i
, port
);
2397 static const unsigned char scm_utf8_bom
[3] = {0xEF, 0xBB, 0xBF};
2398 static const unsigned char scm_utf16be_bom
[2] = {0xFE, 0xFF};
2399 static const unsigned char scm_utf16le_bom
[2] = {0xFF, 0xFE};
2400 static const unsigned char scm_utf32be_bom
[4] = {0x00, 0x00, 0xFE, 0xFF};
2401 static const unsigned char scm_utf32le_bom
[4] = {0xFF, 0xFE, 0x00, 0x00};
2403 /* Decide what byte order to use for a UTF-16 port. Return "UTF-16BE"
2404 or "UTF-16LE". MODE must be either SCM_PORT_READ or SCM_PORT_WRITE,
2405 and specifies which operation is about to be done. The MODE
2406 determines how we will decide the byte order. We deliberately avoid
2407 reading from the port unless the user is about to do so. If the user
2408 is about to read, then we look for a BOM, and if present, we use it
2409 to determine the byte order. Otherwise we choose big endian, as
2410 recommended by the Unicode Standard. Note that the BOM (if any) is
2411 not consumed here. */
2413 decide_utf16_encoding (SCM port
, scm_t_port_rw_active mode
)
2415 if (mode
== SCM_PORT_READ
2416 && SCM_PORT_GET_INTERNAL (port
)->at_stream_start_for_bom_read
2417 && looking_at_bytes (port
, scm_utf16le_bom
, sizeof scm_utf16le_bom
))
2423 /* Decide what byte order to use for a UTF-32 port. Return "UTF-32BE"
2424 or "UTF-32LE". See the comment above 'decide_utf16_encoding' for
2427 decide_utf32_encoding (SCM port
, scm_t_port_rw_active mode
)
2429 if (mode
== SCM_PORT_READ
2430 && SCM_PORT_GET_INTERNAL (port
)->at_stream_start_for_bom_read
2431 && looking_at_bytes (port
, scm_utf32le_bom
, sizeof scm_utf32le_bom
))
2438 finalize_iconv_descriptors (void *ptr
, void *data
)
2440 close_iconv_descriptors (ptr
);
2443 static scm_t_iconv_descriptors
*
2444 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
2446 scm_t_iconv_descriptors
*id
;
2447 iconv_t input_cd
, output_cd
;
2449 input_cd
= (iconv_t
) -1;
2450 output_cd
= (iconv_t
) -1;
2453 /* Open an input iconv conversion descriptor, from ENCODING
2454 to UTF-8. We choose UTF-8, not UTF-32, because iconv
2455 implementations can typically convert from anything to
2456 UTF-8, but not to UTF-32 (see
2457 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
2459 /* Assume opening an iconv descriptor causes about 16 KB of
2461 scm_gc_register_allocation (16 * 1024);
2463 input_cd
= iconv_open ("UTF-8", encoding
);
2464 if (input_cd
== (iconv_t
) -1)
2465 goto invalid_encoding
;
2470 /* Assume opening an iconv descriptor causes about 16 KB of
2472 scm_gc_register_allocation (16 * 1024);
2474 output_cd
= iconv_open (encoding
, "UTF-8");
2475 if (output_cd
== (iconv_t
) -1)
2477 if (input_cd
!= (iconv_t
) -1)
2478 iconv_close (input_cd
);
2479 goto invalid_encoding
;
2483 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
2484 id
->input_cd
= input_cd
;
2485 id
->output_cd
= output_cd
;
2487 /* Register a finalizer to close the descriptors. */
2488 scm_i_set_finalizer (id
, finalize_iconv_descriptors
, NULL
);
2495 err
= scm_from_locale_string (encoding
);
2496 scm_misc_error ("open_iconv_descriptors",
2497 "invalid or unknown character encoding ~s",
2503 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
2505 if (id
->input_cd
!= (iconv_t
) -1)
2506 iconv_close (id
->input_cd
);
2507 if (id
->output_cd
!= (iconv_t
) -1)
2508 iconv_close (id
->output_cd
);
2509 id
->input_cd
= (void *) -1;
2510 id
->output_cd
= (void *) -1;
2513 /* Return the iconv_descriptors, initializing them if necessary. MODE
2514 must be either SCM_PORT_READ or SCM_PORT_WRITE, and specifies which
2515 operation is about to be done. We deliberately avoid reading from
2516 the port unless the user was about to do so. */
2517 scm_t_iconv_descriptors
*
2518 scm_i_port_iconv_descriptors (SCM port
, scm_t_port_rw_active mode
)
2520 scm_t_port_internal
*pti
= SCM_PORT_GET_INTERNAL (port
);
2522 assert (pti
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
2524 if (!pti
->iconv_descriptors
)
2526 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2527 const char *precise_encoding
;
2530 pt
->encoding
= "ISO-8859-1";
2532 /* If the specified encoding is UTF-16 or UTF-32, then make
2533 that more precise by deciding what byte order to use. */
2534 if (c_strcasecmp (pt
->encoding
, "UTF-16") == 0)
2535 precise_encoding
= decide_utf16_encoding (port
, mode
);
2536 else if (c_strcasecmp (pt
->encoding
, "UTF-32") == 0)
2537 precise_encoding
= decide_utf32_encoding (port
, mode
);
2539 precise_encoding
= pt
->encoding
;
2541 pti
->iconv_descriptors
=
2542 open_iconv_descriptors (precise_encoding
,
2543 SCM_INPUT_PORT_P (port
),
2544 SCM_OUTPUT_PORT_P (port
));
2547 return pti
->iconv_descriptors
;
2551 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
2554 scm_t_port_internal
*pti
;
2555 scm_t_iconv_descriptors
*prev
;
2557 /* Set the character encoding for this port. */
2558 pt
= SCM_PTAB_ENTRY (port
);
2559 pti
= SCM_PORT_GET_INTERNAL (port
);
2560 prev
= pti
->iconv_descriptors
;
2562 /* In order to handle cases where the encoding changes mid-stream
2563 (e.g. within an HTTP stream, or within a file that is composed of
2564 segments with different encodings), we consider this to be "stream
2565 start" for purposes of BOM handling, regardless of our actual file
2567 pti
->at_stream_start_for_bom_read
= 1;
2568 pti
->at_stream_start_for_bom_write
= 1;
2570 if (encoding
== NULL
)
2571 encoding
= "ISO-8859-1";
2573 /* If ENCODING is UTF-8, then no conversion descriptor is opened
2574 because we do I/O ourselves. This saves 100+ KiB for each
2576 pt
->encoding
= scm_gc_strdup (encoding
, "port");
2577 if (c_strcasecmp (encoding
, "UTF-8") == 0)
2578 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
2580 pti
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
2582 pti
->iconv_descriptors
= NULL
;
2584 close_iconv_descriptors (prev
);
2587 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
2589 "Returns, as a string, the character encoding that @var{port}\n"
2590 "uses to interpret its input and output.\n")
2591 #define FUNC_NAME s_scm_port_encoding
2596 SCM_VALIDATE_PORT (1, port
);
2598 pt
= SCM_PTAB_ENTRY (port
);
2601 return scm_from_locale_string (pt
->encoding
);
2607 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
2608 (SCM port
, SCM enc
),
2609 "Sets the character encoding that will be used to interpret all\n"
2610 "port I/O. New ports are created with the encoding\n"
2611 "appropriate for the current locale if @code{setlocale} has \n"
2612 "been called or ISO-8859-1 otherwise\n"
2613 "and this procedure can be used to modify that encoding.\n")
2614 #define FUNC_NAME s_scm_set_port_encoding_x
2618 SCM_VALIDATE_PORT (1, port
);
2619 SCM_VALIDATE_STRING (2, enc
);
2621 enc_str
= scm_to_locale_string (enc
);
2622 scm_i_set_port_encoding_x (port
, enc_str
);
2625 return SCM_UNSPECIFIED
;
2630 /* A fluid specifying the default conversion handler for newly created
2631 ports. Its value should be one of the symbols below. */
2632 SCM_VARIABLE (default_conversion_strategy_var
,
2633 "%default-port-conversion-strategy");
2635 /* Whether the above fluid is initialized. */
2636 static int scm_conversion_strategy_init
= 0;
2638 /* The possible conversion strategies. */
2639 SCM_SYMBOL (sym_error
, "error");
2640 SCM_SYMBOL (sym_substitute
, "substitute");
2641 SCM_SYMBOL (sym_escape
, "escape");
2643 /* Return the default failed encoding conversion policy for new created
2645 scm_t_string_failed_conversion_handler
2646 scm_i_default_port_conversion_handler (void)
2648 scm_t_string_failed_conversion_handler handler
;
2650 if (!scm_conversion_strategy_init
2651 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
2652 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
2657 fluid
= SCM_VARIABLE_REF (default_conversion_strategy_var
);
2658 value
= scm_fluid_ref (fluid
);
2660 if (scm_is_eq (sym_substitute
, value
))
2661 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
2662 else if (scm_is_eq (sym_escape
, value
))
2663 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
2665 /* Default to 'error also when the fluid's value is not one of
2666 the valid symbols. */
2667 handler
= SCM_FAILED_CONVERSION_ERROR
;
2673 /* Use HANDLER as the default conversion strategy for future ports. */
2675 scm_i_set_default_port_conversion_handler (scm_t_string_failed_conversion_handler
2680 if (!scm_conversion_strategy_init
2681 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
2682 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
2687 case SCM_FAILED_CONVERSION_ERROR
:
2688 strategy
= sym_error
;
2691 case SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
:
2692 strategy
= sym_escape
;
2695 case SCM_FAILED_CONVERSION_QUESTION_MARK
:
2696 strategy
= sym_substitute
;
2703 scm_fluid_set_x (SCM_VARIABLE_REF (default_conversion_strategy_var
),
2707 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
2708 1, 0, 0, (SCM port
),
2709 "Returns the behavior of the port when handling a character that\n"
2710 "is not representable in the port's current encoding.\n"
2711 "It returns the symbol @code{error} if unrepresentable characters\n"
2712 "should cause exceptions, @code{substitute} if the port should\n"
2713 "try to replace unrepresentable characters with question marks or\n"
2714 "approximate characters, or @code{escape} if unrepresentable\n"
2715 "characters should be converted to string escapes.\n"
2717 "If @var{port} is @code{#f}, then the current default behavior\n"
2718 "will be returned. New ports will have this default behavior\n"
2719 "when they are created.\n")
2720 #define FUNC_NAME s_scm_port_conversion_strategy
2722 scm_t_string_failed_conversion_handler h
;
2724 if (scm_is_false (port
))
2725 h
= scm_i_default_port_conversion_handler ();
2730 SCM_VALIDATE_OPPORT (1, port
);
2731 pt
= SCM_PTAB_ENTRY (port
);
2733 h
= pt
->ilseq_handler
;
2736 if (h
== SCM_FAILED_CONVERSION_ERROR
)
2737 return scm_from_latin1_symbol ("error");
2738 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
2739 return scm_from_latin1_symbol ("substitute");
2740 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
2741 return scm_from_latin1_symbol ("escape");
2745 /* Never gets here. */
2746 return SCM_UNDEFINED
;
2750 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
2752 (SCM port
, SCM sym
),
2753 "Sets the behavior of the interpreter when outputting a character\n"
2754 "that is not representable in the port's current encoding.\n"
2755 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
2756 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
2757 "when an unconvertible character is encountered. If it is\n"
2758 "@code{'substitute}, then unconvertible characters will \n"
2759 "be replaced with approximate characters, or with question marks\n"
2760 "if no approximately correct character is available.\n"
2761 "If it is @code{'escape},\n"
2762 "it will appear as a hex escape when output.\n"
2764 "If @var{port} is an open port, the conversion error behavior\n"
2765 "is set for that port. If it is @code{#f}, it is set as the\n"
2766 "default behavior for any future ports that get created in\n"
2768 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
2770 scm_t_string_failed_conversion_handler handler
;
2772 if (scm_is_eq (sym
, sym_error
))
2773 handler
= SCM_FAILED_CONVERSION_ERROR
;
2774 else if (scm_is_eq (sym
, sym_substitute
))
2775 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
2776 else if (scm_is_eq (sym
, sym_escape
))
2777 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
2779 SCM_MISC_ERROR ("unknown conversion strategy ~s", scm_list_1 (sym
));
2781 if (scm_is_false (port
))
2782 scm_i_set_default_port_conversion_handler (handler
);
2785 SCM_VALIDATE_OPPORT (1, port
);
2786 SCM_PTAB_ENTRY (port
)->ilseq_handler
= handler
;
2789 return SCM_UNSPECIFIED
;
2796 scm_print_port_mode (SCM exp
, SCM port
)
2798 scm_puts (SCM_CLOSEDP (exp
)
2800 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2801 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2804 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2811 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2813 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2816 scm_puts ("#<", port
);
2817 scm_print_port_mode (exp
, port
);
2818 scm_puts (type
, port
);
2819 scm_putc (' ', port
);
2820 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2821 scm_putc ('>', port
);
2829 scm_t_bits scm_tc16_void_port
= 0;
2831 static int fill_input_void_port (SCM port SCM_UNUSED
)
2837 write_void_port (SCM port SCM_UNUSED
,
2838 const void *data SCM_UNUSED
,
2839 size_t size SCM_UNUSED
)
2844 scm_i_void_port (long mode_bits
)
2846 scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex
);
2848 SCM answer
= scm_new_port_table_entry (scm_tc16_void_port
);
2849 scm_t_port
* pt
= SCM_PTAB_ENTRY(answer
);
2851 scm_port_non_buffer (pt
);
2853 SCM_SETSTREAM (answer
, 0);
2854 SCM_SET_CELL_TYPE (answer
, scm_tc16_void_port
| mode_bits
);
2855 scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex
);
2861 scm_void_port (char *mode_str
)
2863 return scm_i_void_port (scm_mode_bits (mode_str
));
2866 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2868 "Create and return a new void port. A void port acts like\n"
2869 "@file{/dev/null}. The @var{mode} argument\n"
2870 "specifies the input/output modes for this port: see the\n"
2871 "documentation for @code{open-file} in @ref{File Ports}.")
2872 #define FUNC_NAME s_scm_sys_make_void_port
2874 return scm_i_void_port (scm_i_mode_bits (mode
));
2879 /* Initialization. */
2884 /* lseek() symbols. */
2885 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2886 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2887 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2889 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2892 cur_inport_fluid
= scm_make_fluid ();
2893 cur_outport_fluid
= scm_make_fluid ();
2894 cur_errport_fluid
= scm_make_fluid ();
2895 cur_loadport_fluid
= scm_make_fluid ();
2897 scm_i_port_weak_hash
= scm_make_weak_key_hash_table (SCM_I_MAKINUM(31));
2899 #include "libguile/ports.x"
2901 /* Use Latin-1 as the default port encoding. */
2902 SCM_VARIABLE_SET (default_port_encoding_var
,
2903 scm_make_fluid_with_default (SCM_BOOL_F
));
2904 scm_port_encoding_init
= 1;
2906 SCM_VARIABLE_SET (default_conversion_strategy_var
,
2907 scm_make_fluid_with_default (sym_substitute
));
2908 scm_conversion_strategy_init
= 1;
2910 /* These bindings are used when boot-9 turns `current-input-port' et
2911 al into parameters. They are then removed from the guile module. */
2912 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2913 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2914 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);