1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011, 2012 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/vectors.h"
59 #include "libguile/weak-set.h"
60 #include "libguile/fluids.h"
61 #include "libguile/eq.h"
75 #ifdef HAVE_SYS_IOCTL_H
76 #include <sys/ioctl.h>
79 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
80 already, but have this code here in case that wasn't so in past versions,
81 or perhaps to help other minimal DOS environments.
83 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
84 might be possibilities if we've got other systems without ftruncate. */
86 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
87 #define ftruncate(fd, size) chsize (fd, size)
89 #define HAVE_FTRUNCATE 1
93 /* The port kind table --- a dynamically resized array of port types. */
96 /* scm_ptobs scm_numptob
97 * implement a dynamically resized array of ptob records.
98 * Indexes into this table are used when generating type
99 * tags for smobjects (if you know a tag you can get an index and conversely).
101 static scm_t_ptob_descriptor
**scm_ptobs
= NULL
;
102 static long scm_numptob
= 0; /* Number of port types. */
103 static long scm_ptobs_size
= 0; /* Number of slots in the port type
105 static scm_i_pthread_mutex_t scm_ptobs_lock
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
106 SCM_PTHREAD_ATFORK_LOCK_STATIC_MUTEX (scm_ptobs_lock
);
109 scm_c_num_port_types (void)
113 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
115 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
120 scm_t_ptob_descriptor
*
121 scm_c_port_type_ref (long ptobnum
)
123 scm_t_ptob_descriptor
*ret
= NULL
;
125 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
127 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
128 ret
= scm_ptobs
[ptobnum
];
130 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
133 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
139 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
143 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
145 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
147 if (scm_numptob
== scm_ptobs_size
)
149 unsigned long old_size
= scm_ptobs_size
;
150 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
152 /* Currently there are only 9 predefined port types, so one
153 resize will cover it. */
154 scm_ptobs_size
= old_size
+ 10;
156 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
157 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
159 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
162 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
166 scm_ptobs
[ret
] = desc
;
169 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
172 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
178 * We choose to use an interface similar to the smob interface with
179 * fill_input and write as standard fields, passed to the port
180 * type constructor, and optional fields set by setters.
184 flush_port_default (SCM port SCM_UNUSED
)
189 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
194 scm_make_port_type (char *name
,
195 int (*fill_input
) (SCM port
),
196 void (*write
) (SCM port
, const void *data
, size_t size
))
198 scm_t_ptob_descriptor
*desc
;
201 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
202 memset (desc
, 0, sizeof (*desc
));
205 desc
->print
= scm_port_print
;
207 desc
->flush
= flush_port_default
;
208 desc
->end_input
= end_input_default
;
209 desc
->fill_input
= fill_input
;
211 ptobnum
= scm_c_port_type_add_x (desc
);
213 /* Make a class object if GOOPS is present. */
214 if (SCM_UNPACK (scm_port_class
[0]) != 0)
215 scm_make_port_classes (ptobnum
, name
);
217 return scm_tc7_port
+ ptobnum
* 256;
221 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
223 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
227 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
229 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
233 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
234 scm_print_state
*pstate
))
236 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
240 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
242 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
246 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
248 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
252 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
254 scm_t_ptob_descriptor
*ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
));
256 ptob
->flags
|= SCM_PORT_TYPE_HAS_FLUSH
;
260 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
262 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
266 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
268 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
272 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
274 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
278 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
280 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
285 /* Standard ports --- current input, output, error, and more(!). */
287 static SCM cur_inport_fluid
= SCM_BOOL_F
;
288 static SCM cur_outport_fluid
= SCM_BOOL_F
;
289 static SCM cur_errport_fluid
= SCM_BOOL_F
;
290 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
292 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
294 "Return the current input port. This is the default port used\n"
295 "by many input procedures. Initially, @code{current-input-port}\n"
296 "returns the @dfn{standard input} in Unix and C terminology.")
297 #define FUNC_NAME s_scm_current_input_port
299 if (scm_is_true (cur_inport_fluid
))
300 return scm_fluid_ref (cur_inport_fluid
);
306 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
308 "Return the current output port. This is the default port used\n"
309 "by many output procedures. Initially,\n"
310 "@code{current-output-port} returns the @dfn{standard output} in\n"
311 "Unix and C terminology.")
312 #define FUNC_NAME s_scm_current_output_port
314 if (scm_is_true (cur_outport_fluid
))
315 return scm_fluid_ref (cur_outport_fluid
);
321 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
323 "Return the port to which errors and warnings should be sent (the\n"
324 "@dfn{standard error} in Unix and C terminology).")
325 #define FUNC_NAME s_scm_current_error_port
327 if (scm_is_true (cur_errport_fluid
))
328 return scm_fluid_ref (cur_errport_fluid
);
335 scm_current_warning_port (void)
337 static SCM cwp_var
= SCM_BOOL_F
;
339 if (scm_is_false (cwp_var
))
340 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
342 return scm_call_0 (scm_variable_ref (cwp_var
));
345 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
347 "Return the current-load-port.\n"
348 "The load port is used internally by @code{primitive-load}.")
349 #define FUNC_NAME s_scm_current_load_port
351 return scm_fluid_ref (cur_loadport_fluid
);
355 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
357 "@deffnx {Scheme Procedure} set-current-output-port port\n"
358 "@deffnx {Scheme Procedure} set-current-error-port port\n"
359 "Change the ports returned by @code{current-input-port},\n"
360 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
361 "so that they use the supplied @var{port} for input or output.")
362 #define FUNC_NAME s_scm_set_current_input_port
364 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
365 SCM_VALIDATE_OPINPORT (1, port
);
366 scm_fluid_set_x (cur_inport_fluid
, port
);
372 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
374 "Set the current default output port to @var{port}.")
375 #define FUNC_NAME s_scm_set_current_output_port
377 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
378 port
= SCM_COERCE_OUTPORT (port
);
379 SCM_VALIDATE_OPOUTPORT (1, port
);
380 scm_fluid_set_x (cur_outport_fluid
, port
);
386 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
388 "Set the current default error port to @var{port}.")
389 #define FUNC_NAME s_scm_set_current_error_port
391 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
392 port
= SCM_COERCE_OUTPORT (port
);
393 SCM_VALIDATE_OPOUTPORT (1, port
);
394 scm_fluid_set_x (cur_errport_fluid
, port
);
401 scm_set_current_warning_port (SCM port
)
403 static SCM cwp_var
= SCM_BOOL_F
;
405 if (scm_is_false (cwp_var
))
406 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
408 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
413 scm_dynwind_current_input_port (SCM port
)
414 #define FUNC_NAME NULL
416 SCM_VALIDATE_OPINPORT (1, port
);
417 scm_dynwind_fluid (cur_inport_fluid
, port
);
422 scm_dynwind_current_output_port (SCM port
)
423 #define FUNC_NAME NULL
425 port
= SCM_COERCE_OUTPORT (port
);
426 SCM_VALIDATE_OPOUTPORT (1, port
);
427 scm_dynwind_fluid (cur_outport_fluid
, port
);
432 scm_dynwind_current_error_port (SCM port
)
433 #define FUNC_NAME NULL
435 port
= SCM_COERCE_OUTPORT (port
);
436 SCM_VALIDATE_OPOUTPORT (1, port
);
437 scm_dynwind_fluid (cur_errport_fluid
, port
);
442 scm_i_dynwind_current_load_port (SCM port
)
444 scm_dynwind_fluid (cur_loadport_fluid
, port
);
450 /* Retrieving a port's mode. */
452 /* Return the flags that characterize a port based on the mode
453 * string used to open a file for that port.
455 * See PORT FLAGS in scm.h
459 scm_i_mode_bits_n (SCM modes
)
462 | (scm_i_string_contains_char (modes
, 'r')
463 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
464 | (scm_i_string_contains_char (modes
, 'w')
465 || scm_i_string_contains_char (modes
, 'a')
466 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
467 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
468 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
472 scm_mode_bits (char *modes
)
474 return scm_i_mode_bits (scm_from_locale_string (modes
));
478 scm_i_mode_bits (SCM modes
)
482 if (!scm_is_string (modes
))
483 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
485 bits
= scm_i_mode_bits_n (modes
);
486 scm_remember_upto_here_1 (modes
);
490 /* Return the mode flags from an open port.
491 * Some modes such as "append" are only used when opening
492 * a file and are not returned here. */
494 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
496 "Return the port modes associated with the open port @var{port}.\n"
497 "These will not necessarily be identical to the modes used when\n"
498 "the port was opened, since modes such as \"append\" which are\n"
499 "used only during port creation are not retained.")
500 #define FUNC_NAME s_scm_port_mode
505 port
= SCM_COERCE_OUTPORT (port
);
506 SCM_VALIDATE_OPPORT (1, port
);
507 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
508 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
509 strcpy (modes
, "r+");
513 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
515 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
518 return scm_from_latin1_string (modes
);
525 /* The port table --- a weak set of all ports.
527 We need a global registry of ports to flush them all at exit, and to
528 get all the ports matching a file descriptor. */
529 SCM scm_i_port_weak_set
;
534 /* Port finalization. */
536 static void finalize_port (GC_PTR
, GC_PTR
);
538 /* Register a finalizer for PORT. */
539 static SCM_C_INLINE_KEYWORD
void
540 register_finalizer_for_port (SCM port
)
542 GC_finalization_proc prev_finalizer
;
543 GC_PTR prev_finalization_data
;
545 /* Register a finalizer for PORT so that its iconv CDs get freed and
546 optionally its type's `free' function gets called. */
547 GC_REGISTER_FINALIZER_NO_ORDER (SCM2PTR (port
), finalize_port
, 0,
549 &prev_finalization_data
);
554 scm_t_ptob_descriptor
*ptob
;
559 do_free (void *body_data
)
561 struct do_free_data
*data
= body_data
;
563 /* `close' is for explicit `close-port' by user. `free' is for this
564 purpose: ports collected by the GC. */
565 data
->ptob
->free (data
->port
);
570 /* Finalize the object (a port) pointed to by PTR. */
572 finalize_port (GC_PTR ptr
, GC_PTR data
)
574 SCM port
= SCM_PACK_POINTER (ptr
);
576 if (!SCM_PORTP (port
))
579 if (SCM_OPENP (port
))
581 if (SCM_REVEALED (port
) > 0)
582 /* Keep "revealed" ports alive and re-register a finalizer. */
583 register_finalizer_for_port (port
);
586 struct do_free_data data
;
588 SCM_CLR_PORT_OPEN_FLAG (port
);
590 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
593 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
594 scm_handle_by_message_noexit
, NULL
);
596 scm_gc_ports_collected
++;
605 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
606 const char *encoding
,
607 scm_t_string_failed_conversion_handler handler
,
612 scm_t_ptob_descriptor
*ptob
;
614 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
615 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
617 ret
= scm_words (tag
| mode_bits
, 3);
618 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
619 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
621 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
622 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
624 entry
->file_name
= SCM_BOOL_F
;
625 entry
->rw_active
= SCM_PORT_NEITHER
;
627 entry
->stream
= stream
;
628 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
629 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
630 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
632 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
633 entry
->ilseq_handler
= handler
;
634 entry
->iconv_descriptors
= NULL
;
636 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
637 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
639 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
640 register_finalizer_for_port (ret
);
646 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
648 return scm_c_make_port_with_encoding (tag
, mode_bits
,
649 scm_i_default_port_encoding (),
650 scm_i_get_conversion_strategy (SCM_BOOL_F
),
655 scm_new_port_table_entry (scm_t_bits tag
)
657 return scm_c_make_port (tag
, 0, 0);
664 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
666 "Return a boolean indicating whether @var{x} is a port.\n"
667 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
669 #define FUNC_NAME s_scm_port_p
671 return scm_from_bool (SCM_PORTP (x
));
675 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
677 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
678 "@code{#f}. Any object satisfying this predicate also satisfies\n"
680 #define FUNC_NAME s_scm_input_port_p
682 return scm_from_bool (SCM_INPUT_PORT_P (x
));
686 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
688 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
689 "@code{#f}. Any object satisfying this predicate also satisfies\n"
691 #define FUNC_NAME s_scm_output_port_p
693 x
= SCM_COERCE_OUTPORT (x
);
694 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
698 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
700 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
702 #define FUNC_NAME s_scm_port_closed_p
704 SCM_VALIDATE_PORT (1, port
);
705 return scm_from_bool (!SCM_OPPORTP (port
));
709 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
711 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
713 #define FUNC_NAME s_scm_eof_object_p
715 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
724 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
727 * Call the close operation on a port object.
728 * see also scm_close.
730 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
732 "Close the specified port object. Return @code{#t} if it\n"
733 "successfully closes a port or @code{#f} if it was already\n"
734 "closed. An exception may be raised if an error occurs, for\n"
735 "example when flushing buffered output. See also @ref{Ports and\n"
736 "File Descriptors, close}, for a procedure which can close file\n"
738 #define FUNC_NAME s_scm_close_port
743 port
= SCM_COERCE_OUTPORT (port
);
745 SCM_VALIDATE_PORT (1, port
);
746 if (SCM_CLOSEDP (port
))
749 p
= SCM_PTAB_ENTRY (port
);
750 SCM_CLR_PORT_OPEN_FLAG (port
);
752 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
753 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
755 if (SCM_PORT_DESCRIPTOR (port
)->close
)
756 /* Note! This may throw an exception. Anything after this point
757 should be resilient to non-local exits. */
758 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
762 if (p
->iconv_descriptors
)
764 /* If we don't get here, the iconv_descriptors finalizer will
766 close_iconv_descriptors (p
->iconv_descriptors
);
767 p
->iconv_descriptors
= NULL
;
770 return scm_from_bool (rv
>= 0);
774 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
776 "Close the specified input port object. The routine has no effect if\n"
777 "the file has already been closed. An exception may be raised if an\n"
778 "error occurs. The value returned is unspecified.\n\n"
779 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
780 "which can close file descriptors.")
781 #define FUNC_NAME s_scm_close_input_port
783 SCM_VALIDATE_INPUT_PORT (1, port
);
784 scm_close_port (port
);
785 return SCM_UNSPECIFIED
;
789 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
791 "Close the specified output port object. The routine has no effect if\n"
792 "the file has already been closed. An exception may be raised if an\n"
793 "error occurs. The value returned is unspecified.\n\n"
794 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
795 "which can close file descriptors.")
796 #define FUNC_NAME s_scm_close_output_port
798 port
= SCM_COERCE_OUTPORT (port
);
799 SCM_VALIDATE_OUTPUT_PORT (1, port
);
800 scm_close_port (port
);
801 return SCM_UNSPECIFIED
;
808 /* Encoding characters to byte streams, and decoding byte streams to
811 /* A fluid specifying the default encoding for newly created ports. If it is
812 a string, that is the encoding. If it is #f, it is in the "native"
813 (Latin-1) encoding. */
814 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
816 static int scm_port_encoding_init
= 0;
818 /* Use ENCODING as the default encoding for future ports. */
820 scm_i_set_default_port_encoding (const char *encoding
)
822 if (!scm_port_encoding_init
823 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
824 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
828 || !strcmp (encoding
, "ASCII")
829 || !strcmp (encoding
, "ANSI_X3.4-1968")
830 || !strcmp (encoding
, "ISO-8859-1"))
831 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
833 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
834 scm_from_locale_string (encoding
));
837 /* Return the name of the default encoding for newly created ports; a
838 return value of NULL means "ISO-8859-1". */
840 scm_i_default_port_encoding (void)
842 if (!scm_port_encoding_init
)
844 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
850 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
851 if (!scm_is_string (encoding
))
854 return scm_i_string_chars (encoding
);
859 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
861 close_iconv_descriptors (ptr
);
864 static scm_t_iconv_descriptors
*
865 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
867 scm_t_iconv_descriptors
*id
;
868 iconv_t input_cd
, output_cd
;
870 input_cd
= (iconv_t
) -1;
871 output_cd
= (iconv_t
) -1;
875 /* Open an input iconv conversion descriptor, from ENCODING
876 to UTF-8. We choose UTF-8, not UTF-32, because iconv
877 implementations can typically convert from anything to
878 UTF-8, but not to UTF-32 (see
879 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
881 /* Assume opening an iconv descriptor causes about 16 KB of
883 scm_gc_register_allocation (16 * 1024);
886 input_cd
= iconv_open ("UTF-8", encoding
);
887 scm_i_unlock_iconv ();
888 if (input_cd
== (iconv_t
) -1)
889 goto invalid_encoding
;
894 /* Assume opening an iconv descriptor causes about 16 KB of
896 scm_gc_register_allocation (16 * 1024);
899 output_cd
= iconv_open (encoding
, "UTF-8");
900 scm_i_unlock_iconv ();
901 if (output_cd
== (iconv_t
) -1)
904 if (input_cd
!= (iconv_t
) -1)
905 iconv_close (input_cd
);
906 scm_i_unlock_iconv ();
907 goto invalid_encoding
;
911 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
912 id
->input_cd
= input_cd
;
913 id
->output_cd
= output_cd
;
916 GC_finalization_proc prev_finalizer
;
917 GC_PTR prev_finalization_data
;
919 /* Register a finalizer to close the descriptors. */
920 GC_REGISTER_FINALIZER_NO_ORDER (id
, finalize_iconv_descriptors
, 0,
921 &prev_finalizer
, &prev_finalization_data
);
929 err
= scm_from_locale_string (encoding
);
930 scm_misc_error ("open_iconv_descriptors",
931 "invalid or unknown character encoding ~s",
937 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
940 if (id
->input_cd
!= (iconv_t
) -1)
941 iconv_close (id
->input_cd
);
942 if (id
->output_cd
!= (iconv_t
) -1)
943 iconv_close (id
->output_cd
);
944 scm_i_unlock_iconv ();
945 id
->input_cd
= (void *) -1;
946 id
->output_cd
= (void *) -1;
949 scm_t_iconv_descriptors
*
950 scm_i_port_iconv_descriptors (SCM port
)
954 pt
= SCM_PTAB_ENTRY (port
);
956 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
958 if (!pt
->iconv_descriptors
)
961 pt
->encoding
= "ISO-8859-1";
962 pt
->iconv_descriptors
=
963 open_iconv_descriptors (pt
->encoding
,
964 SCM_INPUT_PORT_P (port
),
965 SCM_OUTPUT_PORT_P (port
));
968 return pt
->iconv_descriptors
;
972 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
975 scm_t_iconv_descriptors
*prev
;
977 /* Set the character encoding for this port. */
978 pt
= SCM_PTAB_ENTRY (port
);
979 prev
= pt
->iconv_descriptors
;
981 if (encoding
== NULL
)
982 encoding
= "ISO-8859-1";
984 if (strcmp (encoding
, "UTF-8") == 0)
986 pt
->encoding
= "UTF-8";
987 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
988 pt
->iconv_descriptors
= NULL
;
992 /* Open descriptors before mutating the port. */
993 pt
->iconv_descriptors
=
994 open_iconv_descriptors (encoding
,
995 SCM_INPUT_PORT_P (port
),
996 SCM_OUTPUT_PORT_P (port
));
997 pt
->encoding
= scm_gc_strdup (encoding
, "port");
998 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
1002 close_iconv_descriptors (prev
);
1005 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
1007 "Returns, as a string, the character encoding that @var{port}\n"
1008 "uses to interpret its input and output.\n")
1009 #define FUNC_NAME s_scm_port_encoding
1014 SCM_VALIDATE_PORT (1, port
);
1016 pt
= SCM_PTAB_ENTRY (port
);
1019 return scm_from_locale_string (pt
->encoding
);
1025 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1026 (SCM port
, SCM enc
),
1027 "Sets the character encoding that will be used to interpret all\n"
1028 "port I/O. New ports are created with the encoding\n"
1029 "appropriate for the current locale if @code{setlocale} has \n"
1030 "been called or ISO-8859-1 otherwise\n"
1031 "and this procedure can be used to modify that encoding.\n")
1032 #define FUNC_NAME s_scm_set_port_encoding_x
1036 SCM_VALIDATE_PORT (1, port
);
1037 SCM_VALIDATE_STRING (2, enc
);
1039 enc_str
= scm_to_locale_string (enc
);
1040 scm_i_set_port_encoding_x (port
, enc_str
);
1043 return SCM_UNSPECIFIED
;
1048 /* This determines how conversions handle unconvertible characters. */
1049 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
1050 static int scm_conversion_strategy_init
= 0;
1052 scm_t_string_failed_conversion_handler
1053 scm_i_get_conversion_strategy (SCM port
)
1057 if (scm_is_false (port
))
1059 if (!scm_conversion_strategy_init
1060 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1061 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1064 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
1065 if (scm_is_false (encoding
))
1066 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1068 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
1074 pt
= SCM_PTAB_ENTRY (port
);
1075 return pt
->ilseq_handler
;
1081 scm_i_set_conversion_strategy_x (SCM port
,
1082 scm_t_string_failed_conversion_handler handler
)
1087 strategy
= scm_from_int ((int) handler
);
1089 if (scm_is_false (port
))
1091 /* Set the default encoding for future ports. */
1092 if (!scm_conversion_strategy_init
1093 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1094 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1096 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1100 /* Set the character encoding for this port. */
1101 pt
= SCM_PTAB_ENTRY (port
);
1102 pt
->ilseq_handler
= handler
;
1106 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1107 1, 0, 0, (SCM port
),
1108 "Returns the behavior of the port when handling a character that\n"
1109 "is not representable in the port's current encoding.\n"
1110 "It returns the symbol @code{error} if unrepresentable characters\n"
1111 "should cause exceptions, @code{substitute} if the port should\n"
1112 "try to replace unrepresentable characters with question marks or\n"
1113 "approximate characters, or @code{escape} if unrepresentable\n"
1114 "characters should be converted to string escapes.\n"
1116 "If @var{port} is @code{#f}, then the current default behavior\n"
1117 "will be returned. New ports will have this default behavior\n"
1118 "when they are created.\n")
1119 #define FUNC_NAME s_scm_port_conversion_strategy
1121 scm_t_string_failed_conversion_handler h
;
1123 SCM_VALIDATE_OPPORT (1, port
);
1125 if (!scm_is_false (port
))
1127 SCM_VALIDATE_OPPORT (1, port
);
1130 h
= scm_i_get_conversion_strategy (port
);
1131 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1132 return scm_from_latin1_symbol ("error");
1133 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1134 return scm_from_latin1_symbol ("substitute");
1135 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1136 return scm_from_latin1_symbol ("escape");
1140 /* Never gets here. */
1141 return SCM_UNDEFINED
;
1145 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1147 (SCM port
, SCM sym
),
1148 "Sets the behavior of the interpreter when outputting a character\n"
1149 "that is not representable in the port's current encoding.\n"
1150 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1151 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1152 "when an unconvertible character is encountered. If it is\n"
1153 "@code{'substitute}, then unconvertible characters will \n"
1154 "be replaced with approximate characters, or with question marks\n"
1155 "if no approximately correct character is available.\n"
1156 "If it is @code{'escape},\n"
1157 "it will appear as a hex escape when output.\n"
1159 "If @var{port} is an open port, the conversion error behavior\n"
1160 "is set for that port. If it is @code{#f}, it is set as the\n"
1161 "default behavior for any future ports that get created in\n"
1163 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1169 if (!scm_is_false (port
))
1171 SCM_VALIDATE_OPPORT (1, port
);
1174 err
= scm_from_latin1_symbol ("error");
1175 if (scm_is_true (scm_eqv_p (sym
, err
)))
1177 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1178 return SCM_UNSPECIFIED
;
1181 qm
= scm_from_latin1_symbol ("substitute");
1182 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1184 scm_i_set_conversion_strategy_x (port
,
1185 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1186 return SCM_UNSPECIFIED
;
1189 esc
= scm_from_latin1_symbol ("escape");
1190 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1192 scm_i_set_conversion_strategy_x (port
,
1193 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1194 return SCM_UNSPECIFIED
;
1197 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1199 return SCM_UNSPECIFIED
;
1206 /* The port lock. */
1209 lock_port (void *mutex
)
1211 scm_i_pthread_mutex_lock (mutex
);
1215 unlock_port (void *mutex
)
1217 scm_i_pthread_mutex_unlock (mutex
);
1221 scm_dynwind_lock_port (SCM port
)
1222 #define FUNC_NAME "dynwind-lock-port"
1224 scm_i_pthread_mutex_t
*lock
;
1225 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1226 scm_c_lock_port (port
, &lock
);
1229 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1230 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1238 /* Revealed counts --- an oddity inherited from SCSH. */
1240 /* Find a port in the table and return its revealed count.
1241 Also used by the garbage collector.
1244 scm_revealed_count (SCM port
)
1246 scm_i_pthread_mutex_t
*lock
;
1249 scm_c_lock_port (port
, &lock
);
1250 ret
= SCM_REVEALED (port
);
1252 scm_i_pthread_mutex_unlock (lock
);
1257 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
1259 "Return the revealed count for @var{port}.")
1260 #define FUNC_NAME s_scm_port_revealed
1262 port
= SCM_COERCE_OUTPORT (port
);
1263 SCM_VALIDATE_OPENPORT (1, port
);
1264 return scm_from_int (scm_revealed_count (port
));
1268 /* Set the revealed count for a port. */
1269 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
1270 (SCM port
, SCM rcount
),
1271 "Sets the revealed count for a port to a given value.\n"
1272 "The return value is unspecified.")
1273 #define FUNC_NAME s_scm_set_port_revealed_x
1276 scm_i_pthread_mutex_t
*lock
;
1278 /* FIXME: It doesn't make sense to manipulate revealed counts on ports
1279 without a free function. */
1281 port
= SCM_COERCE_OUTPORT (port
);
1282 SCM_VALIDATE_OPENPORT (1, port
);
1283 r
= scm_to_int (rcount
);
1284 scm_c_lock_port (port
, &lock
);
1285 SCM_REVEALED (port
) = r
;
1287 scm_i_pthread_mutex_unlock (lock
);
1288 return SCM_UNSPECIFIED
;
1292 /* Set the revealed count for a port. */
1293 SCM_DEFINE (scm_adjust_port_revealed_x
, "adjust-port-revealed!", 2, 0, 0,
1294 (SCM port
, SCM addend
),
1295 "Add @var{addend} to the revealed count of @var{port}.\n"
1296 "The return value is unspecified.")
1297 #define FUNC_NAME s_scm_adjust_port_revealed_x
1299 scm_i_pthread_mutex_t
*lock
;
1301 port
= SCM_COERCE_OUTPORT (port
);
1302 SCM_VALIDATE_OPENPORT (1, port
);
1303 a
= scm_to_int (addend
);
1304 scm_c_lock_port (port
, &lock
);
1305 SCM_REVEALED (port
) += a
;
1307 scm_i_pthread_mutex_unlock (lock
);
1308 return SCM_UNSPECIFIED
;
1318 scm_get_byte_or_eof (SCM port
)
1320 scm_i_pthread_mutex_t
*lock
;
1323 scm_c_lock_port (port
, &lock
);
1324 ret
= scm_get_byte_or_eof_unlocked (port
);
1326 scm_i_pthread_mutex_unlock (lock
);
1332 scm_peek_byte_or_eof (SCM port
)
1334 scm_i_pthread_mutex_t
*lock
;
1337 scm_c_lock_port (port
, &lock
);
1338 ret
= scm_peek_byte_or_eof_unlocked (port
);
1340 scm_i_pthread_mutex_unlock (lock
);
1347 * Used by an application to read arbitrary number of bytes from an
1348 * SCM port. Same semantics as libc read, except that scm_c_read only
1349 * returns less than SIZE bytes if at end-of-file.
1351 * Warning: Doesn't update port line and column counts! */
1353 /* This structure, and the following swap_buffer function, are used
1354 for temporarily swapping a port's own read buffer, and the buffer
1355 that the caller of scm_c_read provides. */
1356 struct port_and_swap_buffer
1359 unsigned char *buffer
;
1364 swap_buffer (void *data
)
1366 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1367 unsigned char *old_buf
= psb
->pt
->read_buf
;
1368 size_t old_size
= psb
->pt
->read_buf_size
;
1370 /* Make the port use (buffer, size) from the struct. */
1371 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1372 psb
->pt
->read_buf_size
= psb
->size
;
1374 /* Save the port's old (buffer, size) in the struct. */
1375 psb
->buffer
= old_buf
;
1376 psb
->size
= old_size
;
1380 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1381 #define FUNC_NAME "scm_c_read"
1384 size_t n_read
= 0, n_available
;
1385 struct port_and_swap_buffer psb
;
1387 SCM_VALIDATE_OPINPORT (1, port
);
1389 pt
= SCM_PTAB_ENTRY (port
);
1390 if (pt
->rw_active
== SCM_PORT_WRITE
)
1391 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1394 pt
->rw_active
= SCM_PORT_READ
;
1396 /* Take bytes first from the port's read buffer. */
1397 if (pt
->read_pos
< pt
->read_end
)
1399 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1400 memcpy (buffer
, pt
->read_pos
, n_available
);
1401 buffer
= (char *) buffer
+ n_available
;
1402 pt
->read_pos
+= n_available
;
1403 n_read
+= n_available
;
1404 size
-= n_available
;
1407 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1411 /* Now we will call scm_fill_input repeatedly until we have read the
1412 requested number of bytes. (Note that a single scm_fill_input
1413 call does not guarantee to fill the whole of the port's read
1415 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1417 /* The port that we are reading from is unbuffered - i.e. does
1418 not have its own persistent buffer - but we have a buffer,
1419 provided by our caller, that is the right size for the data
1420 that is wanted. For the following scm_fill_input calls,
1421 therefore, we use the buffer in hand as the port's read
1424 We need to make sure that the port's normal (1 byte) buffer
1425 is reinstated in case one of the scm_fill_input () calls
1426 throws an exception; we use the scm_dynwind_* API to achieve
1429 A consequence of this optimization is that the fill_input
1430 functions can't unget characters. That'll push data to the
1431 pushback buffer instead of this psb buffer. */
1433 unsigned char *pback
= pt
->putback_buf
;
1436 psb
.buffer
= buffer
;
1438 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1439 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1440 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1442 /* Call scm_fill_input until we have all the bytes that we need,
1444 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1446 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1447 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1450 if (pback
!= pt
->putback_buf
1451 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1452 scm_misc_error (FUNC_NAME
,
1453 "scm_c_read must not call a fill function that pushes "
1454 "back characters onto an unbuffered port", SCM_EOL
);
1456 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1458 /* Reinstate the port's normal buffer. */
1463 /* The port has its own buffer. It is important that we use it,
1464 even if it happens to be smaller than our caller's buffer, so
1465 that a custom port implementation's entry points (in
1466 particular, fill_input) can rely on the buffer always being
1467 the same as they first set up. */
1468 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1470 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1471 memcpy (buffer
, pt
->read_pos
, n_available
);
1472 buffer
= (char *) buffer
+ n_available
;
1473 pt
->read_pos
+= n_available
;
1474 n_read
+= n_available
;
1475 size
-= n_available
;
1484 scm_c_read (SCM port
, void *buffer
, size_t size
)
1486 scm_i_pthread_mutex_t
*lock
;
1489 scm_c_lock_port (port
, &lock
);
1490 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1492 scm_i_pthread_mutex_unlock (lock
);
1498 /* Update the line and column number of PORT after consumption of C. */
1500 update_port_lf (scm_t_wchar c
, SCM port
)
1525 #define SCM_MBCHAR_BUF_SIZE (4)
1527 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1528 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1530 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1532 scm_t_wchar codepoint
;
1534 if (utf8_buf
[0] <= 0x7f)
1537 codepoint
= utf8_buf
[0];
1539 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1542 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1543 | (utf8_buf
[1] & 0x3f);
1545 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1548 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1549 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1550 | (utf8_buf
[2] & 0x3f);
1555 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1556 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1557 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1558 | (utf8_buf
[3] & 0x3f);
1564 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1565 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1566 representation, and set *LEN to the length in bytes. Return
1567 `EILSEQ' on error. */
1569 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1570 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1572 #define ASSERT_NOT_EOF(b) \
1573 if (SCM_UNLIKELY ((b) == EOF)) \
1575 #define CONSUME_PEEKED_BYTE() \
1582 pt
= SCM_PTAB_ENTRY (port
);
1584 byte
= scm_get_byte_or_eof_unlocked (port
);
1591 buf
[0] = (scm_t_uint8
) byte
;
1596 *codepoint
= buf
[0];
1597 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1600 byte
= scm_peek_byte_or_eof_unlocked (port
);
1601 ASSERT_NOT_EOF (byte
);
1603 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1606 CONSUME_PEEKED_BYTE ();
1607 buf
[1] = (scm_t_uint8
) byte
;
1610 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1613 else if ((buf
[0] & 0xf0) == 0xe0)
1616 byte
= scm_peek_byte_or_eof_unlocked (port
);
1617 ASSERT_NOT_EOF (byte
);
1619 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1620 || (buf
[0] == 0xe0 && byte
< 0xa0)
1621 || (buf
[0] == 0xed && byte
> 0x9f)))
1624 CONSUME_PEEKED_BYTE ();
1625 buf
[1] = (scm_t_uint8
) byte
;
1628 byte
= scm_peek_byte_or_eof_unlocked (port
);
1629 ASSERT_NOT_EOF (byte
);
1631 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1634 CONSUME_PEEKED_BYTE ();
1635 buf
[2] = (scm_t_uint8
) byte
;
1638 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1639 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1642 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1645 byte
= scm_peek_byte_or_eof_unlocked (port
);
1646 ASSERT_NOT_EOF (byte
);
1648 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1649 || (buf
[0] == 0xf0 && byte
< 0x90)
1650 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1653 CONSUME_PEEKED_BYTE ();
1654 buf
[1] = (scm_t_uint8
) byte
;
1657 byte
= scm_peek_byte_or_eof_unlocked (port
);
1658 ASSERT_NOT_EOF (byte
);
1660 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1663 CONSUME_PEEKED_BYTE ();
1664 buf
[2] = (scm_t_uint8
) byte
;
1667 byte
= scm_peek_byte_or_eof_unlocked (port
);
1668 ASSERT_NOT_EOF (byte
);
1670 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1673 CONSUME_PEEKED_BYTE ();
1674 buf
[3] = (scm_t_uint8
) byte
;
1677 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1678 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1679 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1688 /* Here we could choose the consume the faulty byte when it's not a
1689 valid starting byte, but it's not a requirement. What Section 3.9
1690 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1691 would otherwise be a valid starting byte. */
1695 #undef CONSUME_PEEKED_BYTE
1696 #undef ASSERT_NOT_EOF
1699 /* Likewise, read a byte sequence from PORT, passing it through its
1700 input conversion descriptor. */
1702 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1703 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1705 scm_t_iconv_descriptors
*id
;
1707 size_t bytes_consumed
, output_size
;
1709 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1711 id
= scm_i_port_iconv_descriptors (port
);
1713 for (output_size
= 0, output
= (char *) utf8_buf
,
1714 bytes_consumed
= 0, err
= 0;
1715 err
== 0 && output_size
== 0
1716 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1720 size_t input_left
, output_left
, done
;
1722 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1723 if (byte_read
== EOF
)
1725 if (bytes_consumed
== 0)
1727 *codepoint
= (scm_t_wchar
) EOF
;
1735 buf
[bytes_consumed
] = byte_read
;
1738 input_left
= bytes_consumed
+ 1;
1739 output_left
= sizeof (utf8_buf
);
1741 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1742 if (done
== (size_t) -1)
1746 /* Missing input: keep trying. */
1750 output_size
= sizeof (utf8_buf
) - output_left
;
1753 if (SCM_UNLIKELY (output_size
== 0))
1754 /* An unterminated sequence. */
1756 else if (SCM_LIKELY (err
== 0))
1758 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1759 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1760 *len
= bytes_consumed
;
1766 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1767 with the byte representation of the codepoint in PORT's encoding, and
1768 set *LEN to the length in bytes of that representation. Return 0 on
1769 success and an errno value on error. */
1771 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1772 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1775 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1777 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1778 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1780 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1782 if (SCM_LIKELY (err
== 0))
1783 update_port_lf (*codepoint
, port
);
1784 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1788 update_port_lf (*codepoint
, port
);
1794 /* Read a codepoint from PORT and return it. */
1796 scm_getc_unlocked (SCM port
)
1797 #define FUNC_NAME "scm_getc"
1801 scm_t_wchar codepoint
;
1802 char buf
[SCM_MBCHAR_BUF_SIZE
];
1804 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1805 if (SCM_UNLIKELY (err
!= 0))
1806 /* At this point PORT should point past the invalid encoding, as per
1807 R6RS-lib Section 8.2.4. */
1808 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1817 scm_i_pthread_mutex_t
*lock
;
1820 scm_c_lock_port (port
, &lock
);
1821 ret
= scm_getc_unlocked (port
);
1823 scm_i_pthread_mutex_unlock (lock
);
1829 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1831 "Return the next character available from @var{port}, updating\n"
1832 "@var{port} to point to the following character. If no more\n"
1833 "characters are available, the end-of-file object is returned.\n"
1835 "When @var{port}'s data cannot be decoded according to its\n"
1836 "character encoding, a @code{decoding-error} is raised and\n"
1837 "@var{port} points past the erroneous byte sequence.\n")
1838 #define FUNC_NAME s_scm_read_char
1841 if (SCM_UNBNDP (port
))
1842 port
= scm_current_input_port ();
1843 SCM_VALIDATE_OPINPORT (1, port
);
1844 c
= scm_getc_unlocked (port
);
1847 return SCM_MAKE_CHAR (c
);
1857 scm_unget_byte_unlocked (int c
, SCM port
)
1858 #define FUNC_NAME "scm_unget_byte"
1860 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1862 if (pt
->read_buf
== pt
->putback_buf
)
1863 /* already using the put-back buffer. */
1865 /* enlarge putback_buf if necessary. */
1866 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1867 && pt
->read_buf
== pt
->read_pos
)
1869 size_t new_size
= pt
->read_buf_size
* 2;
1870 unsigned char *tmp
= (unsigned char *)
1871 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1874 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1875 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1876 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1879 /* shift any existing bytes to buffer + 1. */
1880 if (pt
->read_pos
== pt
->read_end
)
1881 pt
->read_end
= pt
->read_buf
+ 1;
1882 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1884 int count
= pt
->read_end
- pt
->read_pos
;
1886 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1887 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1890 pt
->read_pos
= pt
->read_buf
;
1893 /* switch to the put-back buffer. */
1895 if (pt
->putback_buf
== NULL
)
1898 = (unsigned char *) scm_gc_malloc_pointerless
1899 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1900 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1903 pt
->saved_read_buf
= pt
->read_buf
;
1904 pt
->saved_read_pos
= pt
->read_pos
;
1905 pt
->saved_read_end
= pt
->read_end
;
1906 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1908 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1909 pt
->read_end
= pt
->read_buf
+ 1;
1910 pt
->read_buf_size
= pt
->putback_buf_size
;
1916 pt
->rw_active
= SCM_PORT_READ
;
1921 scm_unget_byte (int c
, SCM port
)
1923 scm_i_pthread_mutex_t
*lock
;
1924 scm_c_lock_port (port
, &lock
);
1925 scm_unget_byte_unlocked (c
, port
);
1927 scm_i_pthread_mutex_unlock (lock
);
1932 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1933 #define FUNC_NAME "scm_ungetc"
1935 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1937 char result_buf
[10];
1938 const char *encoding
;
1942 if (pt
->encoding
!= NULL
)
1943 encoding
= pt
->encoding
;
1945 encoding
= "ISO-8859-1";
1947 len
= sizeof (result_buf
);
1948 scm_i_lock_iconv ();
1949 result
= u32_conv_to_encoding (encoding
,
1950 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1951 (uint32_t *) &c
, 1, NULL
,
1953 scm_i_unlock_iconv ();
1955 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1956 scm_encoding_error (FUNC_NAME
, errno
,
1957 "conversion to port encoding failed",
1958 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1960 for (i
= len
- 1; i
>= 0; i
--)
1961 scm_unget_byte_unlocked (result
[i
], port
);
1963 if (SCM_UNLIKELY (result
!= result_buf
))
1968 /* What should col be in this case?
1969 * We'll leave it at -1.
1971 SCM_LINUM (port
) -= 1;
1979 scm_ungetc (scm_t_wchar c
, SCM port
)
1981 scm_i_pthread_mutex_t
*lock
;
1982 scm_c_lock_port (port
, &lock
);
1983 scm_ungetc_unlocked (c
, port
);
1985 scm_i_pthread_mutex_unlock (lock
);
1990 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1992 /* This is simple minded and inefficient, but unreading strings is
1993 * probably not a common operation, and remember that line and
1994 * column numbers have to be handled...
1996 * Please feel free to write an optimized version!
1999 scm_ungetc_unlocked (s
[n
], port
);
2003 scm_ungets (const char *s
, int n
, SCM port
)
2005 scm_i_pthread_mutex_t
*lock
;
2006 scm_c_lock_port (port
, &lock
);
2007 scm_ungets_unlocked (s
, n
, port
);
2009 scm_i_pthread_mutex_unlock (lock
);
2013 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
2015 "Return the next character available from @var{port},\n"
2016 "@emph{without} updating @var{port} to point to the following\n"
2017 "character. If no more characters are available, the\n"
2018 "end-of-file object is returned.\n"
2020 "The value returned by\n"
2021 "a call to @code{peek-char} is the same as the value that would\n"
2022 "have been returned by a call to @code{read-char} on the same\n"
2023 "port. The only difference is that the very next call to\n"
2024 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
2025 "return the value returned by the preceding call to\n"
2026 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
2027 "an interactive port will hang waiting for input whenever a call\n"
2028 "to @code{read-char} would have hung.\n"
2030 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
2031 "if such a situation occurs. However, unlike with @code{read-char},\n"
2032 "@var{port} still points at the beginning of the erroneous byte\n"
2033 "sequence when the error is raised.\n")
2034 #define FUNC_NAME s_scm_peek_char
2039 char bytes
[SCM_MBCHAR_BUF_SIZE
];
2040 long column
, line
, i
;
2043 if (SCM_UNBNDP (port
))
2044 port
= scm_current_input_port ();
2045 SCM_VALIDATE_OPINPORT (1, port
);
2047 column
= SCM_COL (port
);
2048 line
= SCM_LINUM (port
);
2050 err
= get_codepoint (port
, &c
, bytes
, &len
);
2052 for (i
= len
- 1; i
>= 0; i
--)
2053 scm_unget_byte_unlocked (bytes
[i
], port
);
2055 SCM_COL (port
) = column
;
2056 SCM_LINUM (port
) = line
;
2058 if (SCM_UNLIKELY (err
!= 0))
2060 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
2062 /* Shouldn't happen since `catch' always aborts to prompt. */
2063 result
= SCM_BOOL_F
;
2066 result
= SCM_EOF_VAL
;
2068 result
= SCM_MAKE_CHAR (c
);
2074 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2075 (SCM cobj
, SCM port
),
2076 "Place character @var{cobj} in @var{port} so that it will be\n"
2077 "read by the next read operation. If called multiple times, the\n"
2078 "unread characters will be read again in last-in first-out\n"
2079 "order. If @var{port} is not supplied, the current input port\n"
2081 #define FUNC_NAME s_scm_unread_char
2085 SCM_VALIDATE_CHAR (1, cobj
);
2086 if (SCM_UNBNDP (port
))
2087 port
= scm_current_input_port ();
2088 SCM_VALIDATE_OPINPORT (2, port
);
2090 c
= SCM_CHAR (cobj
);
2092 scm_ungetc_unlocked (c
, port
);
2097 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2098 (SCM str
, SCM port
),
2099 "Place the string @var{str} in @var{port} so that its characters will be\n"
2100 "read in subsequent read operations. If called multiple times, the\n"
2101 "unread characters will be read again in last-in first-out order. If\n"
2102 "@var{port} is not supplied, the current-input-port is used.")
2103 #define FUNC_NAME s_scm_unread_string
2106 SCM_VALIDATE_STRING (1, str
);
2107 if (SCM_UNBNDP (port
))
2108 port
= scm_current_input_port ();
2109 SCM_VALIDATE_OPINPORT (2, port
);
2111 n
= scm_i_string_length (str
);
2114 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2123 /* Manipulating the buffers. */
2125 /* This routine does not take any locks, as it is usually called as part
2126 of a port implementation. */
2128 scm_port_non_buffer (scm_t_port
*pt
)
2130 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2131 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2132 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2133 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2136 /* this should only be called when the read buffer is empty. it
2137 tries to refill the read buffer. it returns the first char from
2138 the port, which is either EOF or *(pt->read_pos). */
2140 scm_fill_input_unlocked (SCM port
)
2142 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2144 assert (pt
->read_pos
== pt
->read_end
);
2146 if (pt
->read_buf
== pt
->putback_buf
)
2148 /* finished reading put-back chars. */
2149 pt
->read_buf
= pt
->saved_read_buf
;
2150 pt
->read_pos
= pt
->saved_read_pos
;
2151 pt
->read_end
= pt
->saved_read_end
;
2152 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2153 if (pt
->read_pos
< pt
->read_end
)
2154 return *(pt
->read_pos
);
2156 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2160 scm_fill_input (SCM port
)
2162 scm_i_pthread_mutex_t
*lock
;
2165 scm_c_lock_port (port
, &lock
);
2166 ret
= scm_fill_input_unlocked (port
);
2168 scm_i_pthread_mutex_unlock (lock
);
2174 /* move up to read_len chars from port's putback and/or read buffers
2175 into memory starting at dest. returns the number of chars moved. */
2177 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2179 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2180 size_t chars_read
= 0;
2181 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2185 memcpy (dest
, pt
->read_pos
, from_buf
);
2186 pt
->read_pos
+= from_buf
;
2187 chars_read
+= from_buf
;
2188 read_len
-= from_buf
;
2192 /* if putback was active, try the real input buffer too. */
2193 if (pt
->read_buf
== pt
->putback_buf
)
2195 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2198 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2199 pt
->saved_read_pos
+= from_buf
;
2200 chars_read
+= from_buf
;
2206 /* Clear a port's read buffers, returning the contents. */
2207 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2209 "This procedure clears a port's input buffers, similar\n"
2210 "to the way that force-output clears the output buffer. The\n"
2211 "contents of the buffers are returned as a single string, e.g.,\n"
2214 "(define p (open-input-file ...))\n"
2215 "(drain-input p) => empty string, nothing buffered yet.\n"
2216 "(unread-char (read-char p) p)\n"
2217 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2219 "Draining the buffers may be useful for cleanly finishing\n"
2220 "buffered I/O so that the file descriptor can be used directly\n"
2221 "for further input.")
2222 #define FUNC_NAME s_scm_drain_input
2229 SCM_VALIDATE_OPINPORT (1, port
);
2230 pt
= SCM_PTAB_ENTRY (port
);
2232 count
= pt
->read_end
- pt
->read_pos
;
2233 if (pt
->read_buf
== pt
->putback_buf
)
2234 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2238 result
= scm_i_make_string (count
, &data
, 0);
2239 scm_take_from_input_buffers (port
, data
, count
);
2242 result
= scm_nullstr
;
2249 scm_end_input_unlocked (SCM port
)
2252 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2254 if (pt
->read_buf
== pt
->putback_buf
)
2256 offset
= pt
->read_end
- pt
->read_pos
;
2257 pt
->read_buf
= pt
->saved_read_buf
;
2258 pt
->read_pos
= pt
->saved_read_pos
;
2259 pt
->read_end
= pt
->saved_read_end
;
2260 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2265 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2269 scm_end_input (SCM port
)
2271 scm_i_pthread_mutex_t
*lock
;
2272 scm_c_lock_port (port
, &lock
);
2273 scm_end_input_unlocked (port
);
2275 scm_i_pthread_mutex_unlock (lock
);
2279 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2281 "Flush the specified output port, or the current output port if @var{port}\n"
2282 "is omitted. The current output buffer contents are passed to the\n"
2283 "underlying port implementation (e.g., in the case of fports, the\n"
2284 "data will be written to the file and the output buffer will be cleared.)\n"
2285 "It has no effect on an unbuffered port.\n\n"
2286 "The return value is unspecified.")
2287 #define FUNC_NAME s_scm_force_output
2289 if (SCM_UNBNDP (port
))
2290 port
= scm_current_output_port ();
2293 port
= SCM_COERCE_OUTPORT (port
);
2294 SCM_VALIDATE_OPOUTPORT (1, port
);
2296 scm_flush_unlocked (port
);
2297 return SCM_UNSPECIFIED
;
2302 scm_flush_unlocked (SCM port
)
2304 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2308 scm_flush (SCM port
)
2310 scm_i_pthread_mutex_t
*lock
;
2311 scm_c_lock_port (port
, &lock
);
2312 scm_flush_unlocked (port
);
2314 scm_i_pthread_mutex_unlock (lock
);
2324 scm_putc (char c
, SCM port
)
2326 scm_i_pthread_mutex_t
*lock
;
2327 scm_c_lock_port (port
, &lock
);
2328 scm_putc_unlocked (c
, port
);
2330 scm_i_pthread_mutex_unlock (lock
);
2335 scm_puts (const char *s
, SCM port
)
2337 scm_i_pthread_mutex_t
*lock
;
2338 scm_c_lock_port (port
, &lock
);
2339 scm_puts_unlocked (s
, port
);
2341 scm_i_pthread_mutex_unlock (lock
);
2347 * Used by an application to write arbitrary number of bytes to an SCM
2348 * port. Similar semantics as libc write. However, unlike libc
2349 * write, scm_c_write writes the requested number of bytes and has no
2352 * Warning: Doesn't update port line and column counts!
2355 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2356 #define FUNC_NAME "scm_c_write"
2359 scm_t_ptob_descriptor
*ptob
;
2361 SCM_VALIDATE_OPOUTPORT (1, port
);
2363 pt
= SCM_PTAB_ENTRY (port
);
2364 ptob
= SCM_PORT_DESCRIPTOR (port
);
2366 if (pt
->rw_active
== SCM_PORT_READ
)
2367 scm_end_input_unlocked (port
);
2369 ptob
->write (port
, ptr
, size
);
2372 pt
->rw_active
= SCM_PORT_WRITE
;
2377 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2379 scm_i_pthread_mutex_t
*lock
;
2380 scm_c_lock_port (port
, &lock
);
2381 scm_c_write_unlocked (port
, ptr
, size
);
2383 scm_i_pthread_mutex_unlock (lock
);
2389 * This function differs from scm_c_write; it updates port line and
2392 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2394 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2395 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2397 if (pt
->rw_active
== SCM_PORT_READ
)
2398 scm_end_input_unlocked (port
);
2400 ptob
->write (port
, ptr
, size
);
2402 for (; size
; ptr
++, size
--)
2403 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2406 pt
->rw_active
= SCM_PORT_WRITE
;
2410 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2412 scm_i_pthread_mutex_t
*lock
;
2413 scm_c_lock_port (port
, &lock
);
2414 scm_lfwrite_unlocked (ptr
, size
, port
);
2416 scm_i_pthread_mutex_unlock (lock
);
2420 /* Write STR to PORT from START inclusive to END exclusive. */
2422 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2424 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2426 if (pt
->rw_active
== SCM_PORT_READ
)
2427 scm_end_input_unlocked (port
);
2429 if (end
== (size_t) -1)
2430 end
= scm_i_string_length (str
);
2432 scm_display (scm_c_substring (str
, start
, end
), port
);
2435 pt
->rw_active
= SCM_PORT_WRITE
;
2441 /* Querying and setting positions, and character availability. */
2443 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2445 "Return @code{#t} if a character is ready on input @var{port}\n"
2446 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2447 "@code{#t} then the next @code{read-char} operation on\n"
2448 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2449 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2451 "@code{char-ready?} exists to make it possible for a\n"
2452 "program to accept characters from interactive ports without\n"
2453 "getting stuck waiting for input. Any input editors associated\n"
2454 "with such ports must make sure that characters whose existence\n"
2455 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2456 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2457 "a port at end of file would be indistinguishable from an\n"
2458 "interactive port that has no ready characters.")
2459 #define FUNC_NAME s_scm_char_ready_p
2463 if (SCM_UNBNDP (port
))
2464 port
= scm_current_input_port ();
2465 /* It's possible to close the current input port, so validate even in
2467 SCM_VALIDATE_OPINPORT (1, port
);
2469 pt
= SCM_PTAB_ENTRY (port
);
2471 /* if the current read buffer is filled, or the
2472 last pushed-back char has been read and the saved buffer is
2473 filled, result is true. */
2474 if (pt
->read_pos
< pt
->read_end
2475 || (pt
->read_buf
== pt
->putback_buf
2476 && pt
->saved_read_pos
< pt
->saved_read_end
))
2480 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2482 if (ptob
->input_waiting
)
2483 return scm_from_bool(ptob
->input_waiting (port
));
2490 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2491 (SCM fd_port
, SCM offset
, SCM whence
),
2492 "Sets the current position of @var{fd_port} to the integer\n"
2493 "@var{offset}, which is interpreted according to the value of\n"
2496 "One of the following variables should be supplied for\n"
2498 "@defvar SEEK_SET\n"
2499 "Seek from the beginning of the file.\n"
2501 "@defvar SEEK_CUR\n"
2502 "Seek from the current position.\n"
2504 "@defvar SEEK_END\n"
2505 "Seek from the end of the file.\n"
2507 "If @var{fd_port} is a file descriptor, the underlying system\n"
2508 "call is @code{lseek}. @var{port} may be a string port.\n"
2510 "The value returned is the new position in the file. This means\n"
2511 "that the current position of a port can be obtained using:\n"
2513 "(seek port 0 SEEK_CUR)\n"
2515 #define FUNC_NAME s_scm_seek
2519 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2521 how
= scm_to_int (whence
);
2522 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2523 SCM_OUT_OF_RANGE (3, whence
);
2525 if (SCM_OPPORTP (fd_port
))
2527 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2528 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2529 off_t_or_off64_t rv
;
2532 SCM_MISC_ERROR ("port is not seekable",
2533 scm_cons (fd_port
, SCM_EOL
));
2535 rv
= ptob
->seek (fd_port
, off
, how
);
2536 return scm_from_off_t_or_off64_t (rv
);
2538 else /* file descriptor?. */
2540 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2541 off_t_or_off64_t rv
;
2542 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2545 return scm_from_off_t_or_off64_t (rv
);
2554 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2555 doesn't have the filename version truncate(), hence this code. */
2556 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2558 truncate (const char *file
, off_t length
)
2562 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2566 ret
= ftruncate (fdes
, length
);
2569 int save_errno
= errno
;
2575 return close (fdes
);
2577 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2579 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2580 (SCM object
, SCM length
),
2581 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2582 "can be a filename string, a port object, or an integer file\n"
2584 "The return value is unspecified.\n"
2586 "For a port or file descriptor @var{length} can be omitted, in\n"
2587 "which case the file is truncated at the current position (per\n"
2588 "@code{ftell} above).\n"
2590 "On most systems a file can be extended by giving a length\n"
2591 "greater than the current size, but this is not mandatory in the\n"
2593 #define FUNC_NAME s_scm_truncate_file
2597 /* "object" can be a port, fdes or filename.
2599 Negative "length" makes no sense, but it's left to truncate() or
2600 ftruncate() to give back an error for that (normally EINVAL).
2603 if (SCM_UNBNDP (length
))
2605 /* must supply length if object is a filename. */
2606 if (scm_is_string (object
))
2607 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2609 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2612 object
= SCM_COERCE_OUTPORT (object
);
2613 if (scm_is_integer (object
))
2615 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2616 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2619 else if (SCM_OPOUTPORTP (object
))
2621 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2622 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2623 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2625 if (!ptob
->truncate
)
2626 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2627 if (pt
->rw_active
== SCM_PORT_READ
)
2628 scm_end_input_unlocked (object
);
2629 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2630 ptob
->flush (object
);
2632 ptob
->truncate (object
, c_length
);
2637 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2638 char *str
= scm_to_locale_string (object
);
2640 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2647 return SCM_UNSPECIFIED
;
2651 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2653 "Return the current line number for @var{port}.\n"
2655 "The first line of a file is 0. But you might want to add 1\n"
2656 "when printing line numbers, since starting from 1 is\n"
2657 "traditional in error messages, and likely to be more natural to\n"
2659 #define FUNC_NAME s_scm_port_line
2661 port
= SCM_COERCE_OUTPORT (port
);
2662 SCM_VALIDATE_OPENPORT (1, port
);
2663 return scm_from_long (SCM_LINUM (port
));
2667 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2668 (SCM port
, SCM line
),
2669 "Set the current line number for @var{port} to @var{line}. The\n"
2670 "first line of a file is 0.")
2671 #define FUNC_NAME s_scm_set_port_line_x
2673 port
= SCM_COERCE_OUTPORT (port
);
2674 SCM_VALIDATE_OPENPORT (1, port
);
2675 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2676 return SCM_UNSPECIFIED
;
2680 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2682 "Return the current column number of @var{port}.\n"
2683 "If the number is\n"
2684 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2685 "- i.e. the first character of the first line is line 0, column 0.\n"
2686 "(However, when you display a file position, for example in an error\n"
2687 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2688 "because lines and column numbers traditionally start with 1, and that is\n"
2689 "what non-programmers will find most natural.)")
2690 #define FUNC_NAME s_scm_port_column
2692 port
= SCM_COERCE_OUTPORT (port
);
2693 SCM_VALIDATE_OPENPORT (1, port
);
2694 return scm_from_int (SCM_COL (port
));
2698 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2699 (SCM port
, SCM column
),
2700 "Set the current column of @var{port}. Before reading the first\n"
2701 "character on a line the column should be 0.")
2702 #define FUNC_NAME s_scm_set_port_column_x
2704 port
= SCM_COERCE_OUTPORT (port
);
2705 SCM_VALIDATE_OPENPORT (1, port
);
2706 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2707 return SCM_UNSPECIFIED
;
2711 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2713 "Return the filename associated with @var{port}, or @code{#f}\n"
2714 "if no filename is associated with the port.")
2715 #define FUNC_NAME s_scm_port_filename
2717 port
= SCM_COERCE_OUTPORT (port
);
2718 SCM_VALIDATE_OPENPORT (1, port
);
2719 return SCM_FILENAME (port
);
2723 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2724 (SCM port
, SCM filename
),
2725 "Change the filename associated with @var{port}, using the current input\n"
2726 "port if none is specified. Note that this does not change the port's\n"
2727 "source of data, but only the value that is returned by\n"
2728 "@code{port-filename} and reported in diagnostic output.")
2729 #define FUNC_NAME s_scm_set_port_filename_x
2731 port
= SCM_COERCE_OUTPORT (port
);
2732 SCM_VALIDATE_OPENPORT (1, port
);
2733 /* We allow the user to set the filename to whatever he likes. */
2734 SCM_SET_FILENAME (port
, filename
);
2735 return SCM_UNSPECIFIED
;
2742 /* Implementation helpers for port printing functions. */
2745 scm_print_port_mode (SCM exp
, SCM port
)
2747 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2749 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2750 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2753 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2760 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2762 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2765 scm_puts_unlocked ("#<", port
);
2766 scm_print_port_mode (exp
, port
);
2767 scm_puts_unlocked (type
, port
);
2768 scm_putc_unlocked (' ', port
);
2769 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2770 scm_putc_unlocked ('>', port
);
2777 /* Iterating over all ports. */
2779 struct for_each_data
2781 void (*proc
) (void *data
, SCM p
);
2786 for_each_trampoline (void *data
, SCM port
, SCM result
)
2788 struct for_each_data
*d
= data
;
2790 d
->proc (d
->data
, port
);
2796 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2798 struct for_each_data d
;
2803 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2804 scm_i_port_weak_set
);
2808 scm_for_each_trampoline (void *data
, SCM port
)
2810 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2813 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2815 "Apply @var{proc} to each port in the Guile port table\n"
2816 "in turn. The return value is unspecified. More specifically,\n"
2817 "@var{proc} is applied exactly once to every port that exists\n"
2818 "in the system at the time @code{port-for-each} is invoked.\n"
2819 "Changes to the port table while @code{port-for-each} is running\n"
2820 "have no effect as far as @code{port-for-each} is concerned.")
2821 #define FUNC_NAME s_scm_port_for_each
2823 SCM_VALIDATE_PROC (1, proc
);
2825 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2827 return SCM_UNSPECIFIED
;
2832 flush_output_port (void *closure
, SCM port
)
2834 if (SCM_OPOUTPORTP (port
))
2835 scm_flush_unlocked (port
);
2838 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2840 "Equivalent to calling @code{force-output} on\n"
2841 "all open output ports. The return value is unspecified.")
2842 #define FUNC_NAME s_scm_flush_all_ports
2844 scm_c_port_for_each (&flush_output_port
, NULL
);
2845 return SCM_UNSPECIFIED
;
2854 scm_t_bits scm_tc16_void_port
= 0;
2856 static int fill_input_void_port (SCM port SCM_UNUSED
)
2862 write_void_port (SCM port SCM_UNUSED
,
2863 const void *data SCM_UNUSED
,
2864 size_t size SCM_UNUSED
)
2869 scm_i_void_port (long mode_bits
)
2873 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2875 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2881 scm_void_port (char *mode_str
)
2883 return scm_i_void_port (scm_mode_bits (mode_str
));
2886 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2888 "Create and return a new void port. A void port acts like\n"
2889 "@file{/dev/null}. The @var{mode} argument\n"
2890 "specifies the input/output modes for this port: see the\n"
2891 "documentation for @code{open-file} in @ref{File Ports}.")
2892 #define FUNC_NAME s_scm_sys_make_void_port
2894 return scm_i_void_port (scm_i_mode_bits (mode
));
2901 /* Initialization. */
2906 /* lseek() symbols. */
2907 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2908 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2909 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2911 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2914 cur_inport_fluid
= scm_make_fluid ();
2915 cur_outport_fluid
= scm_make_fluid ();
2916 cur_errport_fluid
= scm_make_fluid ();
2917 cur_loadport_fluid
= scm_make_fluid ();
2919 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2921 #include "libguile/ports.x"
2923 /* Use Latin-1 as the default port encoding. */
2924 SCM_VARIABLE_SET (default_port_encoding_var
,
2925 scm_make_fluid_with_default (SCM_BOOL_F
));
2926 scm_port_encoding_init
= 1;
2928 SCM_VARIABLE_SET (scm_conversion_strategy
,
2929 scm_make_fluid_with_default
2930 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2931 scm_conversion_strategy_init
= 1;
2933 /* These bindings are used when boot-9 turns `current-input-port' et
2934 al into parameters. They are then removed from the guile module. */
2935 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2936 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2937 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);