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 struct do_free_data data
;
583 SCM_CLR_PORT_OPEN_FLAG (port
);
585 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
588 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
589 scm_handle_by_message_noexit
, NULL
);
591 scm_gc_ports_collected
++;
599 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
600 const char *encoding
,
601 scm_t_string_failed_conversion_handler handler
,
606 scm_t_ptob_descriptor
*ptob
;
608 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
609 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
611 ret
= scm_words (tag
| mode_bits
, 3);
612 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
613 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
615 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
616 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
618 entry
->file_name
= SCM_BOOL_F
;
619 entry
->rw_active
= SCM_PORT_NEITHER
;
621 entry
->stream
= stream
;
622 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
623 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
624 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
626 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
627 entry
->ilseq_handler
= handler
;
628 entry
->iconv_descriptors
= NULL
;
630 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
631 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
633 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
634 register_finalizer_for_port (ret
);
640 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
642 return scm_c_make_port_with_encoding (tag
, mode_bits
,
643 scm_i_default_port_encoding (),
644 scm_i_get_conversion_strategy (SCM_BOOL_F
),
649 scm_new_port_table_entry (scm_t_bits tag
)
651 return scm_c_make_port (tag
, 0, 0);
658 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
660 "Return a boolean indicating whether @var{x} is a port.\n"
661 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
663 #define FUNC_NAME s_scm_port_p
665 return scm_from_bool (SCM_PORTP (x
));
669 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
671 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
672 "@code{#f}. Any object satisfying this predicate also satisfies\n"
674 #define FUNC_NAME s_scm_input_port_p
676 return scm_from_bool (SCM_INPUT_PORT_P (x
));
680 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
682 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
683 "@code{#f}. Any object satisfying this predicate also satisfies\n"
685 #define FUNC_NAME s_scm_output_port_p
687 x
= SCM_COERCE_OUTPORT (x
);
688 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
692 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
694 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
696 #define FUNC_NAME s_scm_port_closed_p
698 SCM_VALIDATE_PORT (1, port
);
699 return scm_from_bool (!SCM_OPPORTP (port
));
703 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
705 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
707 #define FUNC_NAME s_scm_eof_object_p
709 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
718 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
721 * Call the close operation on a port object.
722 * see also scm_close.
724 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
726 "Close the specified port object. Return @code{#t} if it\n"
727 "successfully closes a port or @code{#f} if it was already\n"
728 "closed. An exception may be raised if an error occurs, for\n"
729 "example when flushing buffered output. See also @ref{Ports and\n"
730 "File Descriptors, close}, for a procedure which can close file\n"
732 #define FUNC_NAME s_scm_close_port
737 port
= SCM_COERCE_OUTPORT (port
);
739 SCM_VALIDATE_PORT (1, port
);
740 if (SCM_CLOSEDP (port
))
743 p
= SCM_PTAB_ENTRY (port
);
744 SCM_CLR_PORT_OPEN_FLAG (port
);
746 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
747 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
749 if (SCM_PORT_DESCRIPTOR (port
)->close
)
750 /* Note! This may throw an exception. Anything after this point
751 should be resilient to non-local exits. */
752 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
756 if (p
->iconv_descriptors
)
758 /* If we don't get here, the iconv_descriptors finalizer will
760 close_iconv_descriptors (p
->iconv_descriptors
);
761 p
->iconv_descriptors
= NULL
;
764 return scm_from_bool (rv
>= 0);
768 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
770 "Close the specified input port object. The routine has no effect if\n"
771 "the file has already been closed. An exception may be raised if an\n"
772 "error occurs. The value returned is unspecified.\n\n"
773 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
774 "which can close file descriptors.")
775 #define FUNC_NAME s_scm_close_input_port
777 SCM_VALIDATE_INPUT_PORT (1, port
);
778 scm_close_port (port
);
779 return SCM_UNSPECIFIED
;
783 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
785 "Close the specified output port object. The routine has no effect if\n"
786 "the file has already been closed. An exception may be raised if an\n"
787 "error occurs. The value returned is unspecified.\n\n"
788 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
789 "which can close file descriptors.")
790 #define FUNC_NAME s_scm_close_output_port
792 port
= SCM_COERCE_OUTPORT (port
);
793 SCM_VALIDATE_OUTPUT_PORT (1, port
);
794 scm_close_port (port
);
795 return SCM_UNSPECIFIED
;
802 /* Encoding characters to byte streams, and decoding byte streams to
805 /* A fluid specifying the default encoding for newly created ports. If it is
806 a string, that is the encoding. If it is #f, it is in the "native"
807 (Latin-1) encoding. */
808 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
810 static int scm_port_encoding_init
= 0;
812 /* Use ENCODING as the default encoding for future ports. */
814 scm_i_set_default_port_encoding (const char *encoding
)
816 if (!scm_port_encoding_init
817 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
818 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
822 || !strcmp (encoding
, "ASCII")
823 || !strcmp (encoding
, "ANSI_X3.4-1968")
824 || !strcmp (encoding
, "ISO-8859-1"))
825 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
827 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
828 scm_from_locale_string (encoding
));
831 /* Return the name of the default encoding for newly created ports; a
832 return value of NULL means "ISO-8859-1". */
834 scm_i_default_port_encoding (void)
836 if (!scm_port_encoding_init
)
838 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
844 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
845 if (!scm_is_string (encoding
))
848 return scm_i_string_chars (encoding
);
853 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
855 close_iconv_descriptors (ptr
);
858 static scm_t_iconv_descriptors
*
859 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
861 scm_t_iconv_descriptors
*id
;
862 iconv_t input_cd
, output_cd
;
864 input_cd
= (iconv_t
) -1;
865 output_cd
= (iconv_t
) -1;
869 /* Open an input iconv conversion descriptor, from ENCODING
870 to UTF-8. We choose UTF-8, not UTF-32, because iconv
871 implementations can typically convert from anything to
872 UTF-8, but not to UTF-32 (see
873 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
875 /* Assume opening an iconv descriptor causes about 16 KB of
877 scm_gc_register_allocation (16 * 1024);
880 input_cd
= iconv_open ("UTF-8", encoding
);
881 scm_i_unlock_iconv ();
882 if (input_cd
== (iconv_t
) -1)
883 goto invalid_encoding
;
888 /* Assume opening an iconv descriptor causes about 16 KB of
890 scm_gc_register_allocation (16 * 1024);
893 output_cd
= iconv_open (encoding
, "UTF-8");
894 scm_i_unlock_iconv ();
895 if (output_cd
== (iconv_t
) -1)
898 if (input_cd
!= (iconv_t
) -1)
899 iconv_close (input_cd
);
900 scm_i_unlock_iconv ();
901 goto invalid_encoding
;
905 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
906 id
->input_cd
= input_cd
;
907 id
->output_cd
= output_cd
;
910 GC_finalization_proc prev_finalizer
;
911 GC_PTR prev_finalization_data
;
913 /* Register a finalizer to close the descriptors. */
914 GC_REGISTER_FINALIZER_NO_ORDER (id
, finalize_iconv_descriptors
, 0,
915 &prev_finalizer
, &prev_finalization_data
);
923 err
= scm_from_locale_string (encoding
);
924 scm_misc_error ("open_iconv_descriptors",
925 "invalid or unknown character encoding ~s",
931 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
934 if (id
->input_cd
!= (iconv_t
) -1)
935 iconv_close (id
->input_cd
);
936 if (id
->output_cd
!= (iconv_t
) -1)
937 iconv_close (id
->output_cd
);
938 scm_i_unlock_iconv ();
939 id
->input_cd
= (void *) -1;
940 id
->output_cd
= (void *) -1;
943 scm_t_iconv_descriptors
*
944 scm_i_port_iconv_descriptors (SCM port
)
948 pt
= SCM_PTAB_ENTRY (port
);
950 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
952 if (!pt
->iconv_descriptors
)
955 pt
->encoding
= "ISO-8859-1";
956 pt
->iconv_descriptors
=
957 open_iconv_descriptors (pt
->encoding
,
958 SCM_INPUT_PORT_P (port
),
959 SCM_OUTPUT_PORT_P (port
));
962 return pt
->iconv_descriptors
;
966 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
969 scm_t_iconv_descriptors
*prev
;
971 /* Set the character encoding for this port. */
972 pt
= SCM_PTAB_ENTRY (port
);
973 prev
= pt
->iconv_descriptors
;
975 if (encoding
== NULL
)
976 encoding
= "ISO-8859-1";
978 if (strcmp (encoding
, "UTF-8") == 0)
980 pt
->encoding
= "UTF-8";
981 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
982 pt
->iconv_descriptors
= NULL
;
986 /* Open descriptors before mutating the port. */
987 pt
->iconv_descriptors
=
988 open_iconv_descriptors (encoding
,
989 SCM_INPUT_PORT_P (port
),
990 SCM_OUTPUT_PORT_P (port
));
991 pt
->encoding
= scm_gc_strdup (encoding
, "port");
992 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
996 close_iconv_descriptors (prev
);
999 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
1001 "Returns, as a string, the character encoding that @var{port}\n"
1002 "uses to interpret its input and output.\n")
1003 #define FUNC_NAME s_scm_port_encoding
1008 SCM_VALIDATE_PORT (1, port
);
1010 pt
= SCM_PTAB_ENTRY (port
);
1013 return scm_from_locale_string (pt
->encoding
);
1019 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1020 (SCM port
, SCM enc
),
1021 "Sets the character encoding that will be used to interpret all\n"
1022 "port I/O. New ports are created with the encoding\n"
1023 "appropriate for the current locale if @code{setlocale} has \n"
1024 "been called or ISO-8859-1 otherwise\n"
1025 "and this procedure can be used to modify that encoding.\n")
1026 #define FUNC_NAME s_scm_set_port_encoding_x
1030 SCM_VALIDATE_PORT (1, port
);
1031 SCM_VALIDATE_STRING (2, enc
);
1033 enc_str
= scm_to_locale_string (enc
);
1034 scm_i_set_port_encoding_x (port
, enc_str
);
1037 return SCM_UNSPECIFIED
;
1042 /* This determines how conversions handle unconvertible characters. */
1043 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
1044 static int scm_conversion_strategy_init
= 0;
1046 scm_t_string_failed_conversion_handler
1047 scm_i_get_conversion_strategy (SCM port
)
1051 if (scm_is_false (port
))
1053 if (!scm_conversion_strategy_init
1054 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1055 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1058 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
1059 if (scm_is_false (encoding
))
1060 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1062 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
1068 pt
= SCM_PTAB_ENTRY (port
);
1069 return pt
->ilseq_handler
;
1075 scm_i_set_conversion_strategy_x (SCM port
,
1076 scm_t_string_failed_conversion_handler handler
)
1081 strategy
= scm_from_int ((int) handler
);
1083 if (scm_is_false (port
))
1085 /* Set the default encoding for future ports. */
1086 if (!scm_conversion_strategy_init
1087 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1088 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1090 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1094 /* Set the character encoding for this port. */
1095 pt
= SCM_PTAB_ENTRY (port
);
1096 pt
->ilseq_handler
= handler
;
1100 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1101 1, 0, 0, (SCM port
),
1102 "Returns the behavior of the port when handling a character that\n"
1103 "is not representable in the port's current encoding.\n"
1104 "It returns the symbol @code{error} if unrepresentable characters\n"
1105 "should cause exceptions, @code{substitute} if the port should\n"
1106 "try to replace unrepresentable characters with question marks or\n"
1107 "approximate characters, or @code{escape} if unrepresentable\n"
1108 "characters should be converted to string escapes.\n"
1110 "If @var{port} is @code{#f}, then the current default behavior\n"
1111 "will be returned. New ports will have this default behavior\n"
1112 "when they are created.\n")
1113 #define FUNC_NAME s_scm_port_conversion_strategy
1115 scm_t_string_failed_conversion_handler h
;
1117 SCM_VALIDATE_OPPORT (1, port
);
1119 if (!scm_is_false (port
))
1121 SCM_VALIDATE_OPPORT (1, port
);
1124 h
= scm_i_get_conversion_strategy (port
);
1125 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1126 return scm_from_latin1_symbol ("error");
1127 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1128 return scm_from_latin1_symbol ("substitute");
1129 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1130 return scm_from_latin1_symbol ("escape");
1134 /* Never gets here. */
1135 return SCM_UNDEFINED
;
1139 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1141 (SCM port
, SCM sym
),
1142 "Sets the behavior of the interpreter when outputting a character\n"
1143 "that is not representable in the port's current encoding.\n"
1144 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1145 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1146 "when an unconvertible character is encountered. If it is\n"
1147 "@code{'substitute}, then unconvertible characters will \n"
1148 "be replaced with approximate characters, or with question marks\n"
1149 "if no approximately correct character is available.\n"
1150 "If it is @code{'escape},\n"
1151 "it will appear as a hex escape when output.\n"
1153 "If @var{port} is an open port, the conversion error behavior\n"
1154 "is set for that port. If it is @code{#f}, it is set as the\n"
1155 "default behavior for any future ports that get created in\n"
1157 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1163 if (!scm_is_false (port
))
1165 SCM_VALIDATE_OPPORT (1, port
);
1168 err
= scm_from_latin1_symbol ("error");
1169 if (scm_is_true (scm_eqv_p (sym
, err
)))
1171 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1172 return SCM_UNSPECIFIED
;
1175 qm
= scm_from_latin1_symbol ("substitute");
1176 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1178 scm_i_set_conversion_strategy_x (port
,
1179 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1180 return SCM_UNSPECIFIED
;
1183 esc
= scm_from_latin1_symbol ("escape");
1184 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1186 scm_i_set_conversion_strategy_x (port
,
1187 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1188 return SCM_UNSPECIFIED
;
1191 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1193 return SCM_UNSPECIFIED
;
1200 /* The port lock. */
1203 lock_port (void *mutex
)
1205 scm_i_pthread_mutex_lock (mutex
);
1209 unlock_port (void *mutex
)
1211 scm_i_pthread_mutex_unlock (mutex
);
1215 scm_dynwind_lock_port (SCM port
)
1216 #define FUNC_NAME "dynwind-lock-port"
1218 scm_i_pthread_mutex_t
*lock
;
1219 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1220 scm_c_lock_port (port
, &lock
);
1223 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1224 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1235 scm_get_byte_or_eof (SCM port
)
1237 scm_i_pthread_mutex_t
*lock
;
1240 scm_c_lock_port (port
, &lock
);
1241 ret
= scm_get_byte_or_eof_unlocked (port
);
1243 scm_i_pthread_mutex_unlock (lock
);
1249 scm_peek_byte_or_eof (SCM port
)
1251 scm_i_pthread_mutex_t
*lock
;
1254 scm_c_lock_port (port
, &lock
);
1255 ret
= scm_peek_byte_or_eof_unlocked (port
);
1257 scm_i_pthread_mutex_unlock (lock
);
1264 * Used by an application to read arbitrary number of bytes from an
1265 * SCM port. Same semantics as libc read, except that scm_c_read only
1266 * returns less than SIZE bytes if at end-of-file.
1268 * Warning: Doesn't update port line and column counts! */
1270 /* This structure, and the following swap_buffer function, are used
1271 for temporarily swapping a port's own read buffer, and the buffer
1272 that the caller of scm_c_read provides. */
1273 struct port_and_swap_buffer
1276 unsigned char *buffer
;
1281 swap_buffer (void *data
)
1283 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1284 unsigned char *old_buf
= psb
->pt
->read_buf
;
1285 size_t old_size
= psb
->pt
->read_buf_size
;
1287 /* Make the port use (buffer, size) from the struct. */
1288 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1289 psb
->pt
->read_buf_size
= psb
->size
;
1291 /* Save the port's old (buffer, size) in the struct. */
1292 psb
->buffer
= old_buf
;
1293 psb
->size
= old_size
;
1297 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1298 #define FUNC_NAME "scm_c_read"
1301 size_t n_read
= 0, n_available
;
1302 struct port_and_swap_buffer psb
;
1304 SCM_VALIDATE_OPINPORT (1, port
);
1306 pt
= SCM_PTAB_ENTRY (port
);
1307 if (pt
->rw_active
== SCM_PORT_WRITE
)
1308 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1311 pt
->rw_active
= SCM_PORT_READ
;
1313 /* Take bytes first from the port's read buffer. */
1314 if (pt
->read_pos
< pt
->read_end
)
1316 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1317 memcpy (buffer
, pt
->read_pos
, n_available
);
1318 buffer
= (char *) buffer
+ n_available
;
1319 pt
->read_pos
+= n_available
;
1320 n_read
+= n_available
;
1321 size
-= n_available
;
1324 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1328 /* Now we will call scm_fill_input repeatedly until we have read the
1329 requested number of bytes. (Note that a single scm_fill_input
1330 call does not guarantee to fill the whole of the port's read
1332 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1334 /* The port that we are reading from is unbuffered - i.e. does
1335 not have its own persistent buffer - but we have a buffer,
1336 provided by our caller, that is the right size for the data
1337 that is wanted. For the following scm_fill_input calls,
1338 therefore, we use the buffer in hand as the port's read
1341 We need to make sure that the port's normal (1 byte) buffer
1342 is reinstated in case one of the scm_fill_input () calls
1343 throws an exception; we use the scm_dynwind_* API to achieve
1346 A consequence of this optimization is that the fill_input
1347 functions can't unget characters. That'll push data to the
1348 pushback buffer instead of this psb buffer. */
1350 unsigned char *pback
= pt
->putback_buf
;
1353 psb
.buffer
= buffer
;
1355 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1356 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1357 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1359 /* Call scm_fill_input until we have all the bytes that we need,
1361 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1363 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1364 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1367 if (pback
!= pt
->putback_buf
1368 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1369 scm_misc_error (FUNC_NAME
,
1370 "scm_c_read must not call a fill function that pushes "
1371 "back characters onto an unbuffered port", SCM_EOL
);
1373 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1375 /* Reinstate the port's normal buffer. */
1380 /* The port has its own buffer. It is important that we use it,
1381 even if it happens to be smaller than our caller's buffer, so
1382 that a custom port implementation's entry points (in
1383 particular, fill_input) can rely on the buffer always being
1384 the same as they first set up. */
1385 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1387 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1388 memcpy (buffer
, pt
->read_pos
, n_available
);
1389 buffer
= (char *) buffer
+ n_available
;
1390 pt
->read_pos
+= n_available
;
1391 n_read
+= n_available
;
1392 size
-= n_available
;
1401 scm_c_read (SCM port
, void *buffer
, size_t size
)
1403 scm_i_pthread_mutex_t
*lock
;
1406 scm_c_lock_port (port
, &lock
);
1407 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1409 scm_i_pthread_mutex_unlock (lock
);
1415 /* Update the line and column number of PORT after consumption of C. */
1417 update_port_lf (scm_t_wchar c
, SCM port
)
1442 #define SCM_MBCHAR_BUF_SIZE (4)
1444 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1445 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1447 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1449 scm_t_wchar codepoint
;
1451 if (utf8_buf
[0] <= 0x7f)
1454 codepoint
= utf8_buf
[0];
1456 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1459 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1460 | (utf8_buf
[1] & 0x3f);
1462 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1465 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1466 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1467 | (utf8_buf
[2] & 0x3f);
1472 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1473 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1474 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1475 | (utf8_buf
[3] & 0x3f);
1481 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1482 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1483 representation, and set *LEN to the length in bytes. Return
1484 `EILSEQ' on error. */
1486 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1487 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1489 #define ASSERT_NOT_EOF(b) \
1490 if (SCM_UNLIKELY ((b) == EOF)) \
1492 #define CONSUME_PEEKED_BYTE() \
1499 pt
= SCM_PTAB_ENTRY (port
);
1501 byte
= scm_get_byte_or_eof_unlocked (port
);
1508 buf
[0] = (scm_t_uint8
) byte
;
1513 *codepoint
= buf
[0];
1514 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1517 byte
= scm_peek_byte_or_eof_unlocked (port
);
1518 ASSERT_NOT_EOF (byte
);
1520 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1523 CONSUME_PEEKED_BYTE ();
1524 buf
[1] = (scm_t_uint8
) byte
;
1527 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1530 else if ((buf
[0] & 0xf0) == 0xe0)
1533 byte
= scm_peek_byte_or_eof_unlocked (port
);
1534 ASSERT_NOT_EOF (byte
);
1536 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1537 || (buf
[0] == 0xe0 && byte
< 0xa0)
1538 || (buf
[0] == 0xed && byte
> 0x9f)))
1541 CONSUME_PEEKED_BYTE ();
1542 buf
[1] = (scm_t_uint8
) byte
;
1545 byte
= scm_peek_byte_or_eof_unlocked (port
);
1546 ASSERT_NOT_EOF (byte
);
1548 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1551 CONSUME_PEEKED_BYTE ();
1552 buf
[2] = (scm_t_uint8
) byte
;
1555 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1556 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1559 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1562 byte
= scm_peek_byte_or_eof_unlocked (port
);
1563 ASSERT_NOT_EOF (byte
);
1565 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1566 || (buf
[0] == 0xf0 && byte
< 0x90)
1567 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1570 CONSUME_PEEKED_BYTE ();
1571 buf
[1] = (scm_t_uint8
) byte
;
1574 byte
= scm_peek_byte_or_eof_unlocked (port
);
1575 ASSERT_NOT_EOF (byte
);
1577 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1580 CONSUME_PEEKED_BYTE ();
1581 buf
[2] = (scm_t_uint8
) byte
;
1584 byte
= scm_peek_byte_or_eof_unlocked (port
);
1585 ASSERT_NOT_EOF (byte
);
1587 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1590 CONSUME_PEEKED_BYTE ();
1591 buf
[3] = (scm_t_uint8
) byte
;
1594 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1595 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1596 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1605 /* Here we could choose the consume the faulty byte when it's not a
1606 valid starting byte, but it's not a requirement. What Section 3.9
1607 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1608 would otherwise be a valid starting byte. */
1612 #undef CONSUME_PEEKED_BYTE
1613 #undef ASSERT_NOT_EOF
1616 /* Likewise, read a byte sequence from PORT, passing it through its
1617 input conversion descriptor. */
1619 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1620 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1622 scm_t_iconv_descriptors
*id
;
1624 size_t bytes_consumed
, output_size
;
1626 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1628 id
= scm_i_port_iconv_descriptors (port
);
1630 for (output_size
= 0, output
= (char *) utf8_buf
,
1631 bytes_consumed
= 0, err
= 0;
1632 err
== 0 && output_size
== 0
1633 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1637 size_t input_left
, output_left
, done
;
1639 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1640 if (byte_read
== EOF
)
1642 if (bytes_consumed
== 0)
1644 *codepoint
= (scm_t_wchar
) EOF
;
1652 buf
[bytes_consumed
] = byte_read
;
1655 input_left
= bytes_consumed
+ 1;
1656 output_left
= sizeof (utf8_buf
);
1658 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1659 if (done
== (size_t) -1)
1663 /* Missing input: keep trying. */
1667 output_size
= sizeof (utf8_buf
) - output_left
;
1670 if (SCM_UNLIKELY (output_size
== 0))
1671 /* An unterminated sequence. */
1673 else if (SCM_LIKELY (err
== 0))
1675 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1676 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1677 *len
= bytes_consumed
;
1683 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1684 with the byte representation of the codepoint in PORT's encoding, and
1685 set *LEN to the length in bytes of that representation. Return 0 on
1686 success and an errno value on error. */
1688 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1689 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1692 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1694 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1695 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1697 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1699 if (SCM_LIKELY (err
== 0))
1700 update_port_lf (*codepoint
, port
);
1701 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1705 update_port_lf (*codepoint
, port
);
1711 /* Read a codepoint from PORT and return it. */
1713 scm_getc_unlocked (SCM port
)
1714 #define FUNC_NAME "scm_getc"
1718 scm_t_wchar codepoint
;
1719 char buf
[SCM_MBCHAR_BUF_SIZE
];
1721 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1722 if (SCM_UNLIKELY (err
!= 0))
1723 /* At this point PORT should point past the invalid encoding, as per
1724 R6RS-lib Section 8.2.4. */
1725 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1734 scm_i_pthread_mutex_t
*lock
;
1737 scm_c_lock_port (port
, &lock
);
1738 ret
= scm_getc_unlocked (port
);
1740 scm_i_pthread_mutex_unlock (lock
);
1746 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1748 "Return the next character available from @var{port}, updating\n"
1749 "@var{port} to point to the following character. If no more\n"
1750 "characters are available, the end-of-file object is returned.\n"
1752 "When @var{port}'s data cannot be decoded according to its\n"
1753 "character encoding, a @code{decoding-error} is raised and\n"
1754 "@var{port} points past the erroneous byte sequence.\n")
1755 #define FUNC_NAME s_scm_read_char
1758 if (SCM_UNBNDP (port
))
1759 port
= scm_current_input_port ();
1760 SCM_VALIDATE_OPINPORT (1, port
);
1761 c
= scm_getc_unlocked (port
);
1764 return SCM_MAKE_CHAR (c
);
1774 scm_unget_byte_unlocked (int c
, SCM port
)
1775 #define FUNC_NAME "scm_unget_byte"
1777 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1779 if (pt
->read_buf
== pt
->putback_buf
)
1780 /* already using the put-back buffer. */
1782 /* enlarge putback_buf if necessary. */
1783 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1784 && pt
->read_buf
== pt
->read_pos
)
1786 size_t new_size
= pt
->read_buf_size
* 2;
1787 unsigned char *tmp
= (unsigned char *)
1788 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1791 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1792 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1793 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1796 /* shift any existing bytes to buffer + 1. */
1797 if (pt
->read_pos
== pt
->read_end
)
1798 pt
->read_end
= pt
->read_buf
+ 1;
1799 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1801 int count
= pt
->read_end
- pt
->read_pos
;
1803 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1804 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1807 pt
->read_pos
= pt
->read_buf
;
1810 /* switch to the put-back buffer. */
1812 if (pt
->putback_buf
== NULL
)
1815 = (unsigned char *) scm_gc_malloc_pointerless
1816 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1817 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1820 pt
->saved_read_buf
= pt
->read_buf
;
1821 pt
->saved_read_pos
= pt
->read_pos
;
1822 pt
->saved_read_end
= pt
->read_end
;
1823 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1825 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1826 pt
->read_end
= pt
->read_buf
+ 1;
1827 pt
->read_buf_size
= pt
->putback_buf_size
;
1833 pt
->rw_active
= SCM_PORT_READ
;
1838 scm_unget_byte (int c
, SCM port
)
1840 scm_i_pthread_mutex_t
*lock
;
1841 scm_c_lock_port (port
, &lock
);
1842 scm_unget_byte_unlocked (c
, port
);
1844 scm_i_pthread_mutex_unlock (lock
);
1849 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1850 #define FUNC_NAME "scm_ungetc"
1852 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1854 char result_buf
[10];
1855 const char *encoding
;
1859 if (pt
->encoding
!= NULL
)
1860 encoding
= pt
->encoding
;
1862 encoding
= "ISO-8859-1";
1864 len
= sizeof (result_buf
);
1865 scm_i_lock_iconv ();
1866 result
= u32_conv_to_encoding (encoding
,
1867 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1868 (uint32_t *) &c
, 1, NULL
,
1870 scm_i_unlock_iconv ();
1872 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1873 scm_encoding_error (FUNC_NAME
, errno
,
1874 "conversion to port encoding failed",
1875 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1877 for (i
= len
- 1; i
>= 0; i
--)
1878 scm_unget_byte_unlocked (result
[i
], port
);
1880 if (SCM_UNLIKELY (result
!= result_buf
))
1885 /* What should col be in this case?
1886 * We'll leave it at -1.
1888 SCM_LINUM (port
) -= 1;
1896 scm_ungetc (scm_t_wchar c
, SCM port
)
1898 scm_i_pthread_mutex_t
*lock
;
1899 scm_c_lock_port (port
, &lock
);
1900 scm_ungetc_unlocked (c
, port
);
1902 scm_i_pthread_mutex_unlock (lock
);
1907 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1909 /* This is simple minded and inefficient, but unreading strings is
1910 * probably not a common operation, and remember that line and
1911 * column numbers have to be handled...
1913 * Please feel free to write an optimized version!
1916 scm_ungetc_unlocked (s
[n
], port
);
1920 scm_ungets (const char *s
, int n
, SCM port
)
1922 scm_i_pthread_mutex_t
*lock
;
1923 scm_c_lock_port (port
, &lock
);
1924 scm_ungets_unlocked (s
, n
, port
);
1926 scm_i_pthread_mutex_unlock (lock
);
1930 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1932 "Return the next character available from @var{port},\n"
1933 "@emph{without} updating @var{port} to point to the following\n"
1934 "character. If no more characters are available, the\n"
1935 "end-of-file object is returned.\n"
1937 "The value returned by\n"
1938 "a call to @code{peek-char} is the same as the value that would\n"
1939 "have been returned by a call to @code{read-char} on the same\n"
1940 "port. The only difference is that the very next call to\n"
1941 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1942 "return the value returned by the preceding call to\n"
1943 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1944 "an interactive port will hang waiting for input whenever a call\n"
1945 "to @code{read-char} would have hung.\n"
1947 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1948 "if such a situation occurs. However, unlike with @code{read-char},\n"
1949 "@var{port} still points at the beginning of the erroneous byte\n"
1950 "sequence when the error is raised.\n")
1951 #define FUNC_NAME s_scm_peek_char
1956 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1957 long column
, line
, i
;
1960 if (SCM_UNBNDP (port
))
1961 port
= scm_current_input_port ();
1962 SCM_VALIDATE_OPINPORT (1, port
);
1964 column
= SCM_COL (port
);
1965 line
= SCM_LINUM (port
);
1967 err
= get_codepoint (port
, &c
, bytes
, &len
);
1969 for (i
= len
- 1; i
>= 0; i
--)
1970 scm_unget_byte_unlocked (bytes
[i
], port
);
1972 SCM_COL (port
) = column
;
1973 SCM_LINUM (port
) = line
;
1975 if (SCM_UNLIKELY (err
!= 0))
1977 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1979 /* Shouldn't happen since `catch' always aborts to prompt. */
1980 result
= SCM_BOOL_F
;
1983 result
= SCM_EOF_VAL
;
1985 result
= SCM_MAKE_CHAR (c
);
1991 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
1992 (SCM cobj
, SCM port
),
1993 "Place character @var{cobj} in @var{port} so that it will be\n"
1994 "read by the next read operation. If called multiple times, the\n"
1995 "unread characters will be read again in last-in first-out\n"
1996 "order. If @var{port} is not supplied, the current input port\n"
1998 #define FUNC_NAME s_scm_unread_char
2002 SCM_VALIDATE_CHAR (1, cobj
);
2003 if (SCM_UNBNDP (port
))
2004 port
= scm_current_input_port ();
2005 SCM_VALIDATE_OPINPORT (2, port
);
2007 c
= SCM_CHAR (cobj
);
2009 scm_ungetc_unlocked (c
, port
);
2014 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2015 (SCM str
, SCM port
),
2016 "Place the string @var{str} in @var{port} so that its characters will be\n"
2017 "read in subsequent read operations. If called multiple times, the\n"
2018 "unread characters will be read again in last-in first-out order. If\n"
2019 "@var{port} is not supplied, the current-input-port is used.")
2020 #define FUNC_NAME s_scm_unread_string
2023 SCM_VALIDATE_STRING (1, str
);
2024 if (SCM_UNBNDP (port
))
2025 port
= scm_current_input_port ();
2026 SCM_VALIDATE_OPINPORT (2, port
);
2028 n
= scm_i_string_length (str
);
2031 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2040 /* Manipulating the buffers. */
2042 /* This routine does not take any locks, as it is usually called as part
2043 of a port implementation. */
2045 scm_port_non_buffer (scm_t_port
*pt
)
2047 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2048 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2049 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2050 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2053 /* this should only be called when the read buffer is empty. it
2054 tries to refill the read buffer. it returns the first char from
2055 the port, which is either EOF or *(pt->read_pos). */
2057 scm_fill_input_unlocked (SCM port
)
2059 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2061 assert (pt
->read_pos
== pt
->read_end
);
2063 if (pt
->read_buf
== pt
->putback_buf
)
2065 /* finished reading put-back chars. */
2066 pt
->read_buf
= pt
->saved_read_buf
;
2067 pt
->read_pos
= pt
->saved_read_pos
;
2068 pt
->read_end
= pt
->saved_read_end
;
2069 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2070 if (pt
->read_pos
< pt
->read_end
)
2071 return *(pt
->read_pos
);
2073 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2077 scm_fill_input (SCM port
)
2079 scm_i_pthread_mutex_t
*lock
;
2082 scm_c_lock_port (port
, &lock
);
2083 ret
= scm_fill_input_unlocked (port
);
2085 scm_i_pthread_mutex_unlock (lock
);
2091 /* move up to read_len chars from port's putback and/or read buffers
2092 into memory starting at dest. returns the number of chars moved. */
2094 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2096 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2097 size_t chars_read
= 0;
2098 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2102 memcpy (dest
, pt
->read_pos
, from_buf
);
2103 pt
->read_pos
+= from_buf
;
2104 chars_read
+= from_buf
;
2105 read_len
-= from_buf
;
2109 /* if putback was active, try the real input buffer too. */
2110 if (pt
->read_buf
== pt
->putback_buf
)
2112 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2115 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2116 pt
->saved_read_pos
+= from_buf
;
2117 chars_read
+= from_buf
;
2123 /* Clear a port's read buffers, returning the contents. */
2124 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2126 "This procedure clears a port's input buffers, similar\n"
2127 "to the way that force-output clears the output buffer. The\n"
2128 "contents of the buffers are returned as a single string, e.g.,\n"
2131 "(define p (open-input-file ...))\n"
2132 "(drain-input p) => empty string, nothing buffered yet.\n"
2133 "(unread-char (read-char p) p)\n"
2134 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2136 "Draining the buffers may be useful for cleanly finishing\n"
2137 "buffered I/O so that the file descriptor can be used directly\n"
2138 "for further input.")
2139 #define FUNC_NAME s_scm_drain_input
2146 SCM_VALIDATE_OPINPORT (1, port
);
2147 pt
= SCM_PTAB_ENTRY (port
);
2149 count
= pt
->read_end
- pt
->read_pos
;
2150 if (pt
->read_buf
== pt
->putback_buf
)
2151 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2155 result
= scm_i_make_string (count
, &data
, 0);
2156 scm_take_from_input_buffers (port
, data
, count
);
2159 result
= scm_nullstr
;
2166 scm_end_input_unlocked (SCM port
)
2169 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2171 if (pt
->read_buf
== pt
->putback_buf
)
2173 offset
= pt
->read_end
- pt
->read_pos
;
2174 pt
->read_buf
= pt
->saved_read_buf
;
2175 pt
->read_pos
= pt
->saved_read_pos
;
2176 pt
->read_end
= pt
->saved_read_end
;
2177 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2182 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2186 scm_end_input (SCM port
)
2188 scm_i_pthread_mutex_t
*lock
;
2189 scm_c_lock_port (port
, &lock
);
2190 scm_end_input_unlocked (port
);
2192 scm_i_pthread_mutex_unlock (lock
);
2196 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2198 "Flush the specified output port, or the current output port if @var{port}\n"
2199 "is omitted. The current output buffer contents are passed to the\n"
2200 "underlying port implementation (e.g., in the case of fports, the\n"
2201 "data will be written to the file and the output buffer will be cleared.)\n"
2202 "It has no effect on an unbuffered port.\n\n"
2203 "The return value is unspecified.")
2204 #define FUNC_NAME s_scm_force_output
2206 if (SCM_UNBNDP (port
))
2207 port
= scm_current_output_port ();
2210 port
= SCM_COERCE_OUTPORT (port
);
2211 SCM_VALIDATE_OPOUTPORT (1, port
);
2213 scm_flush_unlocked (port
);
2214 return SCM_UNSPECIFIED
;
2219 scm_flush_unlocked (SCM port
)
2221 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2225 scm_flush (SCM port
)
2227 scm_i_pthread_mutex_t
*lock
;
2228 scm_c_lock_port (port
, &lock
);
2229 scm_flush_unlocked (port
);
2231 scm_i_pthread_mutex_unlock (lock
);
2241 scm_putc (char c
, SCM port
)
2243 scm_i_pthread_mutex_t
*lock
;
2244 scm_c_lock_port (port
, &lock
);
2245 scm_putc_unlocked (c
, port
);
2247 scm_i_pthread_mutex_unlock (lock
);
2252 scm_puts (const char *s
, SCM port
)
2254 scm_i_pthread_mutex_t
*lock
;
2255 scm_c_lock_port (port
, &lock
);
2256 scm_puts_unlocked (s
, port
);
2258 scm_i_pthread_mutex_unlock (lock
);
2264 * Used by an application to write arbitrary number of bytes to an SCM
2265 * port. Similar semantics as libc write. However, unlike libc
2266 * write, scm_c_write writes the requested number of bytes and has no
2269 * Warning: Doesn't update port line and column counts!
2272 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2273 #define FUNC_NAME "scm_c_write"
2276 scm_t_ptob_descriptor
*ptob
;
2278 SCM_VALIDATE_OPOUTPORT (1, port
);
2280 pt
= SCM_PTAB_ENTRY (port
);
2281 ptob
= SCM_PORT_DESCRIPTOR (port
);
2283 if (pt
->rw_active
== SCM_PORT_READ
)
2284 scm_end_input_unlocked (port
);
2286 ptob
->write (port
, ptr
, size
);
2289 pt
->rw_active
= SCM_PORT_WRITE
;
2294 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2296 scm_i_pthread_mutex_t
*lock
;
2297 scm_c_lock_port (port
, &lock
);
2298 scm_c_write_unlocked (port
, ptr
, size
);
2300 scm_i_pthread_mutex_unlock (lock
);
2306 * This function differs from scm_c_write; it updates port line and
2309 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2311 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2312 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2314 if (pt
->rw_active
== SCM_PORT_READ
)
2315 scm_end_input_unlocked (port
);
2317 ptob
->write (port
, ptr
, size
);
2319 for (; size
; ptr
++, size
--)
2320 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2323 pt
->rw_active
= SCM_PORT_WRITE
;
2327 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2329 scm_i_pthread_mutex_t
*lock
;
2330 scm_c_lock_port (port
, &lock
);
2331 scm_lfwrite_unlocked (ptr
, size
, port
);
2333 scm_i_pthread_mutex_unlock (lock
);
2337 /* Write STR to PORT from START inclusive to END exclusive. */
2339 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2341 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2343 if (pt
->rw_active
== SCM_PORT_READ
)
2344 scm_end_input_unlocked (port
);
2346 if (end
== (size_t) -1)
2347 end
= scm_i_string_length (str
);
2349 scm_display (scm_c_substring (str
, start
, end
), port
);
2352 pt
->rw_active
= SCM_PORT_WRITE
;
2358 /* Querying and setting positions, and character availability. */
2360 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2362 "Return @code{#t} if a character is ready on input @var{port}\n"
2363 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2364 "@code{#t} then the next @code{read-char} operation on\n"
2365 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2366 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2368 "@code{char-ready?} exists to make it possible for a\n"
2369 "program to accept characters from interactive ports without\n"
2370 "getting stuck waiting for input. Any input editors associated\n"
2371 "with such ports must make sure that characters whose existence\n"
2372 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2373 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2374 "a port at end of file would be indistinguishable from an\n"
2375 "interactive port that has no ready characters.")
2376 #define FUNC_NAME s_scm_char_ready_p
2380 if (SCM_UNBNDP (port
))
2381 port
= scm_current_input_port ();
2382 /* It's possible to close the current input port, so validate even in
2384 SCM_VALIDATE_OPINPORT (1, port
);
2386 pt
= SCM_PTAB_ENTRY (port
);
2388 /* if the current read buffer is filled, or the
2389 last pushed-back char has been read and the saved buffer is
2390 filled, result is true. */
2391 if (pt
->read_pos
< pt
->read_end
2392 || (pt
->read_buf
== pt
->putback_buf
2393 && pt
->saved_read_pos
< pt
->saved_read_end
))
2397 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2399 if (ptob
->input_waiting
)
2400 return scm_from_bool(ptob
->input_waiting (port
));
2407 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2408 (SCM fd_port
, SCM offset
, SCM whence
),
2409 "Sets the current position of @var{fd_port} to the integer\n"
2410 "@var{offset}, which is interpreted according to the value of\n"
2413 "One of the following variables should be supplied for\n"
2415 "@defvar SEEK_SET\n"
2416 "Seek from the beginning of the file.\n"
2418 "@defvar SEEK_CUR\n"
2419 "Seek from the current position.\n"
2421 "@defvar SEEK_END\n"
2422 "Seek from the end of the file.\n"
2424 "If @var{fd_port} is a file descriptor, the underlying system\n"
2425 "call is @code{lseek}. @var{port} may be a string port.\n"
2427 "The value returned is the new position in the file. This means\n"
2428 "that the current position of a port can be obtained using:\n"
2430 "(seek port 0 SEEK_CUR)\n"
2432 #define FUNC_NAME s_scm_seek
2436 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2438 how
= scm_to_int (whence
);
2439 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2440 SCM_OUT_OF_RANGE (3, whence
);
2442 if (SCM_OPPORTP (fd_port
))
2444 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2445 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2446 off_t_or_off64_t rv
;
2449 SCM_MISC_ERROR ("port is not seekable",
2450 scm_cons (fd_port
, SCM_EOL
));
2452 rv
= ptob
->seek (fd_port
, off
, how
);
2453 return scm_from_off_t_or_off64_t (rv
);
2455 else /* file descriptor?. */
2457 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2458 off_t_or_off64_t rv
;
2459 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2462 return scm_from_off_t_or_off64_t (rv
);
2471 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2472 doesn't have the filename version truncate(), hence this code. */
2473 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2475 truncate (const char *file
, off_t length
)
2479 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2483 ret
= ftruncate (fdes
, length
);
2486 int save_errno
= errno
;
2492 return close (fdes
);
2494 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2496 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2497 (SCM object
, SCM length
),
2498 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2499 "can be a filename string, a port object, or an integer file\n"
2501 "The return value is unspecified.\n"
2503 "For a port or file descriptor @var{length} can be omitted, in\n"
2504 "which case the file is truncated at the current position (per\n"
2505 "@code{ftell} above).\n"
2507 "On most systems a file can be extended by giving a length\n"
2508 "greater than the current size, but this is not mandatory in the\n"
2510 #define FUNC_NAME s_scm_truncate_file
2514 /* "object" can be a port, fdes or filename.
2516 Negative "length" makes no sense, but it's left to truncate() or
2517 ftruncate() to give back an error for that (normally EINVAL).
2520 if (SCM_UNBNDP (length
))
2522 /* must supply length if object is a filename. */
2523 if (scm_is_string (object
))
2524 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2526 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2529 object
= SCM_COERCE_OUTPORT (object
);
2530 if (scm_is_integer (object
))
2532 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2533 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2536 else if (SCM_OPOUTPORTP (object
))
2538 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2539 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2540 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2542 if (!ptob
->truncate
)
2543 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2544 if (pt
->rw_active
== SCM_PORT_READ
)
2545 scm_end_input_unlocked (object
);
2546 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2547 ptob
->flush (object
);
2549 ptob
->truncate (object
, c_length
);
2554 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2555 char *str
= scm_to_locale_string (object
);
2557 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2564 return SCM_UNSPECIFIED
;
2568 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2570 "Return the current line number for @var{port}.\n"
2572 "The first line of a file is 0. But you might want to add 1\n"
2573 "when printing line numbers, since starting from 1 is\n"
2574 "traditional in error messages, and likely to be more natural to\n"
2576 #define FUNC_NAME s_scm_port_line
2578 port
= SCM_COERCE_OUTPORT (port
);
2579 SCM_VALIDATE_OPENPORT (1, port
);
2580 return scm_from_long (SCM_LINUM (port
));
2584 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2585 (SCM port
, SCM line
),
2586 "Set the current line number for @var{port} to @var{line}. The\n"
2587 "first line of a file is 0.")
2588 #define FUNC_NAME s_scm_set_port_line_x
2590 port
= SCM_COERCE_OUTPORT (port
);
2591 SCM_VALIDATE_OPENPORT (1, port
);
2592 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2593 return SCM_UNSPECIFIED
;
2597 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2599 "Return the current column number of @var{port}.\n"
2600 "If the number is\n"
2601 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2602 "- i.e. the first character of the first line is line 0, column 0.\n"
2603 "(However, when you display a file position, for example in an error\n"
2604 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2605 "because lines and column numbers traditionally start with 1, and that is\n"
2606 "what non-programmers will find most natural.)")
2607 #define FUNC_NAME s_scm_port_column
2609 port
= SCM_COERCE_OUTPORT (port
);
2610 SCM_VALIDATE_OPENPORT (1, port
);
2611 return scm_from_int (SCM_COL (port
));
2615 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2616 (SCM port
, SCM column
),
2617 "Set the current column of @var{port}. Before reading the first\n"
2618 "character on a line the column should be 0.")
2619 #define FUNC_NAME s_scm_set_port_column_x
2621 port
= SCM_COERCE_OUTPORT (port
);
2622 SCM_VALIDATE_OPENPORT (1, port
);
2623 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2624 return SCM_UNSPECIFIED
;
2628 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2630 "Return the filename associated with @var{port}, or @code{#f}\n"
2631 "if no filename is associated with the port.")
2632 #define FUNC_NAME s_scm_port_filename
2634 port
= SCM_COERCE_OUTPORT (port
);
2635 SCM_VALIDATE_OPENPORT (1, port
);
2636 return SCM_FILENAME (port
);
2640 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2641 (SCM port
, SCM filename
),
2642 "Change the filename associated with @var{port}, using the current input\n"
2643 "port if none is specified. Note that this does not change the port's\n"
2644 "source of data, but only the value that is returned by\n"
2645 "@code{port-filename} and reported in diagnostic output.")
2646 #define FUNC_NAME s_scm_set_port_filename_x
2648 port
= SCM_COERCE_OUTPORT (port
);
2649 SCM_VALIDATE_OPENPORT (1, port
);
2650 /* We allow the user to set the filename to whatever he likes. */
2651 SCM_SET_FILENAME (port
, filename
);
2652 return SCM_UNSPECIFIED
;
2659 /* Implementation helpers for port printing functions. */
2662 scm_print_port_mode (SCM exp
, SCM port
)
2664 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2666 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2667 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2670 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2677 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2679 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2682 scm_puts_unlocked ("#<", port
);
2683 scm_print_port_mode (exp
, port
);
2684 scm_puts_unlocked (type
, port
);
2685 scm_putc_unlocked (' ', port
);
2686 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2687 scm_putc_unlocked ('>', port
);
2694 /* Iterating over all ports. */
2696 struct for_each_data
2698 void (*proc
) (void *data
, SCM p
);
2703 for_each_trampoline (void *data
, SCM port
, SCM result
)
2705 struct for_each_data
*d
= data
;
2707 d
->proc (d
->data
, port
);
2713 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2715 struct for_each_data d
;
2720 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2721 scm_i_port_weak_set
);
2725 scm_for_each_trampoline (void *data
, SCM port
)
2727 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2730 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2732 "Apply @var{proc} to each port in the Guile port table\n"
2733 "in turn. The return value is unspecified. More specifically,\n"
2734 "@var{proc} is applied exactly once to every port that exists\n"
2735 "in the system at the time @code{port-for-each} is invoked.\n"
2736 "Changes to the port table while @code{port-for-each} is running\n"
2737 "have no effect as far as @code{port-for-each} is concerned.")
2738 #define FUNC_NAME s_scm_port_for_each
2740 SCM_VALIDATE_PROC (1, proc
);
2742 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2744 return SCM_UNSPECIFIED
;
2749 flush_output_port (void *closure
, SCM port
)
2751 if (SCM_OPOUTPORTP (port
))
2752 scm_flush_unlocked (port
);
2755 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2757 "Equivalent to calling @code{force-output} on\n"
2758 "all open output ports. The return value is unspecified.")
2759 #define FUNC_NAME s_scm_flush_all_ports
2761 scm_c_port_for_each (&flush_output_port
, NULL
);
2762 return SCM_UNSPECIFIED
;
2771 scm_t_bits scm_tc16_void_port
= 0;
2773 static int fill_input_void_port (SCM port SCM_UNUSED
)
2779 write_void_port (SCM port SCM_UNUSED
,
2780 const void *data SCM_UNUSED
,
2781 size_t size SCM_UNUSED
)
2786 scm_i_void_port (long mode_bits
)
2790 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2792 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2798 scm_void_port (char *mode_str
)
2800 return scm_i_void_port (scm_mode_bits (mode_str
));
2803 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2805 "Create and return a new void port. A void port acts like\n"
2806 "@file{/dev/null}. The @var{mode} argument\n"
2807 "specifies the input/output modes for this port: see the\n"
2808 "documentation for @code{open-file} in @ref{File Ports}.")
2809 #define FUNC_NAME s_scm_sys_make_void_port
2811 return scm_i_void_port (scm_i_mode_bits (mode
));
2818 /* Initialization. */
2823 /* lseek() symbols. */
2824 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2825 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2826 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2828 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2831 cur_inport_fluid
= scm_make_fluid ();
2832 cur_outport_fluid
= scm_make_fluid ();
2833 cur_errport_fluid
= scm_make_fluid ();
2834 cur_loadport_fluid
= scm_make_fluid ();
2836 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2838 #include "libguile/ports.x"
2840 /* Use Latin-1 as the default port encoding. */
2841 SCM_VARIABLE_SET (default_port_encoding_var
,
2842 scm_make_fluid_with_default (SCM_BOOL_F
));
2843 scm_port_encoding_init
= 1;
2845 SCM_VARIABLE_SET (scm_conversion_strategy
,
2846 scm_make_fluid_with_default
2847 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2848 scm_conversion_strategy_init
= 1;
2850 /* These bindings are used when boot-9 turns `current-input-port' et
2851 al into parameters. They are then removed from the guile module. */
2852 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2853 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2854 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);