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
;
108 scm_c_num_port_types (void)
112 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
114 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
119 scm_t_ptob_descriptor
*
120 scm_c_port_type_ref (long ptobnum
)
122 scm_t_ptob_descriptor
*ret
= NULL
;
124 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
126 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
127 ret
= scm_ptobs
[ptobnum
];
129 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
132 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
138 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
142 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
144 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
146 if (scm_numptob
== scm_ptobs_size
)
148 unsigned long old_size
= scm_ptobs_size
;
149 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
151 /* Currently there are only 9 predefined port types, so one
152 resize will cover it. */
153 scm_ptobs_size
= old_size
+ 10;
155 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
156 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
158 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
161 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
165 scm_ptobs
[ret
] = desc
;
168 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
171 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
177 * We choose to use an interface similar to the smob interface with
178 * fill_input and write as standard fields, passed to the port
179 * type constructor, and optional fields set by setters.
183 flush_port_default (SCM port SCM_UNUSED
)
188 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
193 scm_make_port_type (char *name
,
194 int (*fill_input
) (SCM port
),
195 void (*write
) (SCM port
, const void *data
, size_t size
))
197 scm_t_ptob_descriptor
*desc
;
200 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
201 memset (desc
, 0, sizeof (*desc
));
204 desc
->print
= scm_port_print
;
206 desc
->flush
= flush_port_default
;
207 desc
->end_input
= end_input_default
;
208 desc
->fill_input
= fill_input
;
210 ptobnum
= scm_c_port_type_add_x (desc
);
212 /* Make a class object if GOOPS is present. */
213 if (SCM_UNPACK (scm_port_class
[0]) != 0)
214 scm_make_port_classes (ptobnum
, name
);
216 return scm_tc7_port
+ ptobnum
* 256;
220 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
222 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
226 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
228 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
232 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
233 scm_print_state
*pstate
))
235 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
239 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
241 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
245 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
247 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
251 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
253 scm_t_ptob_descriptor
*ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
));
255 ptob
->flags
|= SCM_PORT_TYPE_HAS_FLUSH
;
259 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
261 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
265 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
267 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
271 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
273 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
277 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
279 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
284 /* Standard ports --- current input, output, error, and more(!). */
286 static SCM cur_inport_fluid
= SCM_BOOL_F
;
287 static SCM cur_outport_fluid
= SCM_BOOL_F
;
288 static SCM cur_errport_fluid
= SCM_BOOL_F
;
289 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
291 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
293 "Return the current input port. This is the default port used\n"
294 "by many input procedures. Initially, @code{current-input-port}\n"
295 "returns the @dfn{standard input} in Unix and C terminology.")
296 #define FUNC_NAME s_scm_current_input_port
298 if (scm_is_true (cur_inport_fluid
))
299 return scm_fluid_ref (cur_inport_fluid
);
305 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
307 "Return the current output port. This is the default port used\n"
308 "by many output procedures. Initially,\n"
309 "@code{current-output-port} returns the @dfn{standard output} in\n"
310 "Unix and C terminology.")
311 #define FUNC_NAME s_scm_current_output_port
313 if (scm_is_true (cur_outport_fluid
))
314 return scm_fluid_ref (cur_outport_fluid
);
320 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
322 "Return the port to which errors and warnings should be sent (the\n"
323 "@dfn{standard error} in Unix and C terminology).")
324 #define FUNC_NAME s_scm_current_error_port
326 if (scm_is_true (cur_errport_fluid
))
327 return scm_fluid_ref (cur_errport_fluid
);
334 scm_current_warning_port (void)
336 static SCM cwp_var
= SCM_BOOL_F
;
338 if (scm_is_false (cwp_var
))
339 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
341 return scm_call_0 (scm_variable_ref (cwp_var
));
344 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
346 "Return the current-load-port.\n"
347 "The load port is used internally by @code{primitive-load}.")
348 #define FUNC_NAME s_scm_current_load_port
350 return scm_fluid_ref (cur_loadport_fluid
);
354 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
356 "@deffnx {Scheme Procedure} set-current-output-port port\n"
357 "@deffnx {Scheme Procedure} set-current-error-port port\n"
358 "Change the ports returned by @code{current-input-port},\n"
359 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
360 "so that they use the supplied @var{port} for input or output.")
361 #define FUNC_NAME s_scm_set_current_input_port
363 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
364 SCM_VALIDATE_OPINPORT (1, port
);
365 scm_fluid_set_x (cur_inport_fluid
, port
);
371 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
373 "Set the current default output port to @var{port}.")
374 #define FUNC_NAME s_scm_set_current_output_port
376 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
377 port
= SCM_COERCE_OUTPORT (port
);
378 SCM_VALIDATE_OPOUTPORT (1, port
);
379 scm_fluid_set_x (cur_outport_fluid
, port
);
385 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
387 "Set the current default error port to @var{port}.")
388 #define FUNC_NAME s_scm_set_current_error_port
390 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
391 port
= SCM_COERCE_OUTPORT (port
);
392 SCM_VALIDATE_OPOUTPORT (1, port
);
393 scm_fluid_set_x (cur_errport_fluid
, port
);
400 scm_set_current_warning_port (SCM port
)
402 static SCM cwp_var
= SCM_BOOL_F
;
404 if (scm_is_false (cwp_var
))
405 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
407 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
412 scm_dynwind_current_input_port (SCM port
)
413 #define FUNC_NAME NULL
415 SCM_VALIDATE_OPINPORT (1, port
);
416 scm_dynwind_fluid (cur_inport_fluid
, port
);
421 scm_dynwind_current_output_port (SCM port
)
422 #define FUNC_NAME NULL
424 port
= SCM_COERCE_OUTPORT (port
);
425 SCM_VALIDATE_OPOUTPORT (1, port
);
426 scm_dynwind_fluid (cur_outport_fluid
, port
);
431 scm_dynwind_current_error_port (SCM port
)
432 #define FUNC_NAME NULL
434 port
= SCM_COERCE_OUTPORT (port
);
435 SCM_VALIDATE_OPOUTPORT (1, port
);
436 scm_dynwind_fluid (cur_errport_fluid
, port
);
441 scm_i_dynwind_current_load_port (SCM port
)
443 scm_dynwind_fluid (cur_loadport_fluid
, port
);
449 /* Retrieving a port's mode. */
451 /* Return the flags that characterize a port based on the mode
452 * string used to open a file for that port.
454 * See PORT FLAGS in scm.h
458 scm_i_mode_bits_n (SCM modes
)
461 | (scm_i_string_contains_char (modes
, 'r')
462 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
463 | (scm_i_string_contains_char (modes
, 'w')
464 || scm_i_string_contains_char (modes
, 'a')
465 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
466 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
467 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
471 scm_mode_bits (char *modes
)
473 return scm_i_mode_bits (scm_from_locale_string (modes
));
477 scm_i_mode_bits (SCM modes
)
481 if (!scm_is_string (modes
))
482 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
484 bits
= scm_i_mode_bits_n (modes
);
485 scm_remember_upto_here_1 (modes
);
489 /* Return the mode flags from an open port.
490 * Some modes such as "append" are only used when opening
491 * a file and are not returned here. */
493 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
495 "Return the port modes associated with the open port @var{port}.\n"
496 "These will not necessarily be identical to the modes used when\n"
497 "the port was opened, since modes such as \"append\" which are\n"
498 "used only during port creation are not retained.")
499 #define FUNC_NAME s_scm_port_mode
504 port
= SCM_COERCE_OUTPORT (port
);
505 SCM_VALIDATE_OPPORT (1, port
);
506 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
507 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
508 strcpy (modes
, "r+");
512 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
514 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
517 return scm_from_latin1_string (modes
);
524 /* The port table --- a weak set of all ports.
526 We need a global registry of ports to flush them all at exit, and to
527 get all the ports matching a file descriptor. */
528 SCM scm_i_port_weak_set
;
533 /* Port finalization. */
537 scm_t_ptob_descriptor
*ptob
;
542 do_free (void *body_data
)
544 struct do_free_data
*data
= body_data
;
546 /* `close' is for explicit `close-port' by user. `free' is for this
547 purpose: ports collected by the GC. */
548 data
->ptob
->free (data
->port
);
553 /* Finalize the object (a port) pointed to by PTR. */
555 finalize_port (GC_PTR ptr
, GC_PTR data
)
557 SCM port
= SCM_PACK_POINTER (ptr
);
559 if (!SCM_PORTP (port
))
562 if (SCM_OPENP (port
))
564 struct do_free_data data
;
566 SCM_CLR_PORT_OPEN_FLAG (port
);
568 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
571 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
572 scm_handle_by_message_noexit
, NULL
);
574 scm_gc_ports_collected
++;
582 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
583 const char *encoding
,
584 scm_t_string_failed_conversion_handler handler
,
589 scm_t_ptob_descriptor
*ptob
;
591 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
592 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
594 ret
= scm_words (tag
| mode_bits
, 3);
595 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
596 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
598 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
599 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
601 entry
->file_name
= SCM_BOOL_F
;
602 entry
->rw_active
= SCM_PORT_NEITHER
;
604 entry
->stream
= stream
;
605 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
606 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
607 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
608 else if (!encoding
|| strcmp (encoding
, "ISO-8859-1") == 0)
609 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
611 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
612 entry
->ilseq_handler
= handler
;
613 entry
->iconv_descriptors
= NULL
;
615 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
616 scm_i_set_finalizer (SCM2PTR (ret
), finalize_port
, NULL
);
618 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
619 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
625 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
627 return scm_c_make_port_with_encoding (tag
, mode_bits
,
628 scm_i_default_port_encoding (),
629 scm_i_get_conversion_strategy (SCM_BOOL_F
),
634 scm_new_port_table_entry (scm_t_bits tag
)
636 return scm_c_make_port (tag
, 0, 0);
643 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
645 "Return a boolean indicating whether @var{x} is a port.\n"
646 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
648 #define FUNC_NAME s_scm_port_p
650 return scm_from_bool (SCM_PORTP (x
));
654 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
656 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
657 "@code{#f}. Any object satisfying this predicate also satisfies\n"
659 #define FUNC_NAME s_scm_input_port_p
661 return scm_from_bool (SCM_INPUT_PORT_P (x
));
665 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
667 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
668 "@code{#f}. Any object satisfying this predicate also satisfies\n"
670 #define FUNC_NAME s_scm_output_port_p
672 x
= SCM_COERCE_OUTPORT (x
);
673 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
677 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
679 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
681 #define FUNC_NAME s_scm_port_closed_p
683 SCM_VALIDATE_PORT (1, port
);
684 return scm_from_bool (!SCM_OPPORTP (port
));
688 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
690 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
692 #define FUNC_NAME s_scm_eof_object_p
694 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
703 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
706 * Call the close operation on a port object.
707 * see also scm_close.
709 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
711 "Close the specified port object. Return @code{#t} if it\n"
712 "successfully closes a port or @code{#f} if it was already\n"
713 "closed. An exception may be raised if an error occurs, for\n"
714 "example when flushing buffered output. See also @ref{Ports and\n"
715 "File Descriptors, close}, for a procedure which can close file\n"
717 #define FUNC_NAME s_scm_close_port
722 port
= SCM_COERCE_OUTPORT (port
);
724 SCM_VALIDATE_PORT (1, port
);
725 if (SCM_CLOSEDP (port
))
728 p
= SCM_PTAB_ENTRY (port
);
729 SCM_CLR_PORT_OPEN_FLAG (port
);
731 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
732 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
734 if (SCM_PORT_DESCRIPTOR (port
)->close
)
735 /* Note! This may throw an exception. Anything after this point
736 should be resilient to non-local exits. */
737 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
741 if (p
->iconv_descriptors
)
743 /* If we don't get here, the iconv_descriptors finalizer will
745 close_iconv_descriptors (p
->iconv_descriptors
);
746 p
->iconv_descriptors
= NULL
;
749 return scm_from_bool (rv
>= 0);
753 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
755 "Close the specified input port object. The routine has no effect if\n"
756 "the file has already been closed. An exception may be raised if an\n"
757 "error occurs. The value returned is unspecified.\n\n"
758 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
759 "which can close file descriptors.")
760 #define FUNC_NAME s_scm_close_input_port
762 SCM_VALIDATE_INPUT_PORT (1, port
);
763 scm_close_port (port
);
764 return SCM_UNSPECIFIED
;
768 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
770 "Close the specified output 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_output_port
777 port
= SCM_COERCE_OUTPORT (port
);
778 SCM_VALIDATE_OUTPUT_PORT (1, port
);
779 scm_close_port (port
);
780 return SCM_UNSPECIFIED
;
787 /* Encoding characters to byte streams, and decoding byte streams to
790 /* A fluid specifying the default encoding for newly created ports. If it is
791 a string, that is the encoding. If it is #f, it is in the "native"
792 (Latin-1) encoding. */
793 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
795 static int scm_port_encoding_init
= 0;
797 /* Use ENCODING as the default encoding for future ports. */
799 scm_i_set_default_port_encoding (const char *encoding
)
801 if (!scm_port_encoding_init
802 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
803 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
807 || !strcmp (encoding
, "ASCII")
808 || !strcmp (encoding
, "ANSI_X3.4-1968")
809 || !strcmp (encoding
, "ISO-8859-1"))
810 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
812 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
813 scm_from_locale_string (encoding
));
816 /* Return the name of the default encoding for newly created ports; a
817 return value of NULL means "ISO-8859-1". */
819 scm_i_default_port_encoding (void)
821 if (!scm_port_encoding_init
)
823 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
829 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
830 if (!scm_is_string (encoding
))
833 return scm_i_string_chars (encoding
);
838 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
840 close_iconv_descriptors (ptr
);
843 static scm_t_iconv_descriptors
*
844 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
846 scm_t_iconv_descriptors
*id
;
847 iconv_t input_cd
, output_cd
;
849 input_cd
= (iconv_t
) -1;
850 output_cd
= (iconv_t
) -1;
854 /* Open an input iconv conversion descriptor, from ENCODING
855 to UTF-8. We choose UTF-8, not UTF-32, because iconv
856 implementations can typically convert from anything to
857 UTF-8, but not to UTF-32 (see
858 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
860 /* Assume opening an iconv descriptor causes about 16 KB of
862 scm_gc_register_allocation (16 * 1024);
864 input_cd
= iconv_open ("UTF-8", encoding
);
865 if (input_cd
== (iconv_t
) -1)
866 goto invalid_encoding
;
871 /* Assume opening an iconv descriptor causes about 16 KB of
873 scm_gc_register_allocation (16 * 1024);
875 output_cd
= iconv_open (encoding
, "UTF-8");
876 if (output_cd
== (iconv_t
) -1)
878 if (input_cd
!= (iconv_t
) -1)
879 iconv_close (input_cd
);
880 goto invalid_encoding
;
884 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
885 id
->input_cd
= input_cd
;
886 id
->output_cd
= output_cd
;
888 /* Register a finalizer to close the descriptors. */
889 scm_i_set_finalizer (id
, finalize_iconv_descriptors
, NULL
);
896 err
= scm_from_locale_string (encoding
);
897 scm_misc_error ("open_iconv_descriptors",
898 "invalid or unknown character encoding ~s",
904 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
906 if (id
->input_cd
!= (iconv_t
) -1)
907 iconv_close (id
->input_cd
);
908 if (id
->output_cd
!= (iconv_t
) -1)
909 iconv_close (id
->output_cd
);
910 id
->input_cd
= (void *) -1;
911 id
->output_cd
= (void *) -1;
914 scm_t_iconv_descriptors
*
915 scm_i_port_iconv_descriptors (SCM port
)
919 pt
= SCM_PTAB_ENTRY (port
);
921 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
923 if (!pt
->iconv_descriptors
)
926 pt
->encoding
= "ISO-8859-1";
927 pt
->iconv_descriptors
=
928 open_iconv_descriptors (pt
->encoding
,
929 SCM_INPUT_PORT_P (port
),
930 SCM_OUTPUT_PORT_P (port
));
933 return pt
->iconv_descriptors
;
937 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
940 scm_t_iconv_descriptors
*prev
;
942 /* Set the character encoding for this port. */
943 pt
= SCM_PTAB_ENTRY (port
);
944 prev
= pt
->iconv_descriptors
;
946 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
948 pt
->encoding
= "UTF-8";
949 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
950 pt
->iconv_descriptors
= NULL
;
952 else if (!encoding
|| strcmp (encoding
, "ISO-8859-1") == 0)
954 pt
->encoding
= "ISO-8859-1";
955 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
956 pt
->iconv_descriptors
= NULL
;
960 /* Open descriptors before mutating the port. */
961 pt
->iconv_descriptors
=
962 open_iconv_descriptors (encoding
,
963 SCM_INPUT_PORT_P (port
),
964 SCM_OUTPUT_PORT_P (port
));
965 pt
->encoding
= scm_gc_strdup (encoding
, "port");
966 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
970 close_iconv_descriptors (prev
);
973 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
975 "Returns, as a string, the character encoding that @var{port}\n"
976 "uses to interpret its input and output.\n")
977 #define FUNC_NAME s_scm_port_encoding
982 SCM_VALIDATE_PORT (1, port
);
984 pt
= SCM_PTAB_ENTRY (port
);
987 return scm_from_locale_string (pt
->encoding
);
993 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
995 "Sets the character encoding that will be used to interpret all\n"
996 "port I/O. New ports are created with the encoding\n"
997 "appropriate for the current locale if @code{setlocale} has \n"
998 "been called or ISO-8859-1 otherwise\n"
999 "and this procedure can be used to modify that encoding.\n")
1000 #define FUNC_NAME s_scm_set_port_encoding_x
1004 SCM_VALIDATE_PORT (1, port
);
1005 SCM_VALIDATE_STRING (2, enc
);
1007 enc_str
= scm_to_locale_string (enc
);
1008 scm_i_set_port_encoding_x (port
, enc_str
);
1011 return SCM_UNSPECIFIED
;
1016 /* This determines how conversions handle unconvertible characters. */
1017 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
1018 static int scm_conversion_strategy_init
= 0;
1020 scm_t_string_failed_conversion_handler
1021 scm_i_get_conversion_strategy (SCM port
)
1025 if (scm_is_false (port
))
1027 if (!scm_conversion_strategy_init
1028 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1029 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1032 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
1033 if (scm_is_false (encoding
))
1034 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1036 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
1042 pt
= SCM_PTAB_ENTRY (port
);
1043 return pt
->ilseq_handler
;
1049 scm_i_set_conversion_strategy_x (SCM port
,
1050 scm_t_string_failed_conversion_handler handler
)
1055 strategy
= scm_from_int ((int) handler
);
1057 if (scm_is_false (port
))
1059 /* Set the default encoding for future ports. */
1060 if (!scm_conversion_strategy_init
1061 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1062 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1064 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1068 /* Set the character encoding for this port. */
1069 pt
= SCM_PTAB_ENTRY (port
);
1070 pt
->ilseq_handler
= handler
;
1074 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1075 1, 0, 0, (SCM port
),
1076 "Returns the behavior of the port when handling a character that\n"
1077 "is not representable in the port's current encoding.\n"
1078 "It returns the symbol @code{error} if unrepresentable characters\n"
1079 "should cause exceptions, @code{substitute} if the port should\n"
1080 "try to replace unrepresentable characters with question marks or\n"
1081 "approximate characters, or @code{escape} if unrepresentable\n"
1082 "characters should be converted to string escapes.\n"
1084 "If @var{port} is @code{#f}, then the current default behavior\n"
1085 "will be returned. New ports will have this default behavior\n"
1086 "when they are created.\n")
1087 #define FUNC_NAME s_scm_port_conversion_strategy
1089 scm_t_string_failed_conversion_handler h
;
1091 SCM_VALIDATE_OPPORT (1, port
);
1093 if (!scm_is_false (port
))
1095 SCM_VALIDATE_OPPORT (1, port
);
1098 h
= scm_i_get_conversion_strategy (port
);
1099 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1100 return scm_from_latin1_symbol ("error");
1101 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1102 return scm_from_latin1_symbol ("substitute");
1103 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1104 return scm_from_latin1_symbol ("escape");
1108 /* Never gets here. */
1109 return SCM_UNDEFINED
;
1113 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1115 (SCM port
, SCM sym
),
1116 "Sets the behavior of the interpreter when outputting a character\n"
1117 "that is not representable in the port's current encoding.\n"
1118 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1119 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1120 "when an unconvertible character is encountered. If it is\n"
1121 "@code{'substitute}, then unconvertible characters will \n"
1122 "be replaced with approximate characters, or with question marks\n"
1123 "if no approximately correct character is available.\n"
1124 "If it is @code{'escape},\n"
1125 "it will appear as a hex escape when output.\n"
1127 "If @var{port} is an open port, the conversion error behavior\n"
1128 "is set for that port. If it is @code{#f}, it is set as the\n"
1129 "default behavior for any future ports that get created in\n"
1131 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1137 if (!scm_is_false (port
))
1139 SCM_VALIDATE_OPPORT (1, port
);
1142 err
= scm_from_latin1_symbol ("error");
1143 if (scm_is_true (scm_eqv_p (sym
, err
)))
1145 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1146 return SCM_UNSPECIFIED
;
1149 qm
= scm_from_latin1_symbol ("substitute");
1150 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1152 scm_i_set_conversion_strategy_x (port
,
1153 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1154 return SCM_UNSPECIFIED
;
1157 esc
= scm_from_latin1_symbol ("escape");
1158 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1160 scm_i_set_conversion_strategy_x (port
,
1161 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1162 return SCM_UNSPECIFIED
;
1165 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1167 return SCM_UNSPECIFIED
;
1174 /* The port lock. */
1177 lock_port (void *mutex
)
1179 scm_i_pthread_mutex_lock (mutex
);
1183 unlock_port (void *mutex
)
1185 scm_i_pthread_mutex_unlock (mutex
);
1189 scm_dynwind_lock_port (SCM port
)
1190 #define FUNC_NAME "dynwind-lock-port"
1192 scm_i_pthread_mutex_t
*lock
;
1193 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1194 scm_c_lock_port (port
, &lock
);
1197 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1198 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1209 scm_get_byte_or_eof (SCM port
)
1211 scm_i_pthread_mutex_t
*lock
;
1214 scm_c_lock_port (port
, &lock
);
1215 ret
= scm_get_byte_or_eof_unlocked (port
);
1217 scm_i_pthread_mutex_unlock (lock
);
1223 scm_peek_byte_or_eof (SCM port
)
1225 scm_i_pthread_mutex_t
*lock
;
1228 scm_c_lock_port (port
, &lock
);
1229 ret
= scm_peek_byte_or_eof_unlocked (port
);
1231 scm_i_pthread_mutex_unlock (lock
);
1238 * Used by an application to read arbitrary number of bytes from an
1239 * SCM port. Same semantics as libc read, except that scm_c_read only
1240 * returns less than SIZE bytes if at end-of-file.
1242 * Warning: Doesn't update port line and column counts! */
1244 /* This structure, and the following swap_buffer function, are used
1245 for temporarily swapping a port's own read buffer, and the buffer
1246 that the caller of scm_c_read provides. */
1247 struct port_and_swap_buffer
1250 unsigned char *buffer
;
1255 swap_buffer (void *data
)
1257 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1258 unsigned char *old_buf
= psb
->pt
->read_buf
;
1259 size_t old_size
= psb
->pt
->read_buf_size
;
1261 /* Make the port use (buffer, size) from the struct. */
1262 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1263 psb
->pt
->read_buf_size
= psb
->size
;
1265 /* Save the port's old (buffer, size) in the struct. */
1266 psb
->buffer
= old_buf
;
1267 psb
->size
= old_size
;
1271 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1272 #define FUNC_NAME "scm_c_read"
1275 size_t n_read
= 0, n_available
;
1276 struct port_and_swap_buffer psb
;
1278 SCM_VALIDATE_OPINPORT (1, port
);
1280 pt
= SCM_PTAB_ENTRY (port
);
1281 if (pt
->rw_active
== SCM_PORT_WRITE
)
1282 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1285 pt
->rw_active
= SCM_PORT_READ
;
1287 /* Take bytes first from the port's read buffer. */
1288 if (pt
->read_pos
< pt
->read_end
)
1290 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1291 memcpy (buffer
, pt
->read_pos
, n_available
);
1292 buffer
= (char *) buffer
+ n_available
;
1293 pt
->read_pos
+= n_available
;
1294 n_read
+= n_available
;
1295 size
-= n_available
;
1298 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1302 /* Now we will call scm_fill_input repeatedly until we have read the
1303 requested number of bytes. (Note that a single scm_fill_input
1304 call does not guarantee to fill the whole of the port's read
1306 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1308 /* The port that we are reading from is unbuffered - i.e. does
1309 not have its own persistent buffer - but we have a buffer,
1310 provided by our caller, that is the right size for the data
1311 that is wanted. For the following scm_fill_input calls,
1312 therefore, we use the buffer in hand as the port's read
1315 We need to make sure that the port's normal (1 byte) buffer
1316 is reinstated in case one of the scm_fill_input () calls
1317 throws an exception; we use the scm_dynwind_* API to achieve
1320 A consequence of this optimization is that the fill_input
1321 functions can't unget characters. That'll push data to the
1322 pushback buffer instead of this psb buffer. */
1324 unsigned char *pback
= pt
->putback_buf
;
1327 psb
.buffer
= buffer
;
1329 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1330 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1331 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1333 /* Call scm_fill_input until we have all the bytes that we need,
1335 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1337 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1338 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1341 if (pback
!= pt
->putback_buf
1342 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1343 scm_misc_error (FUNC_NAME
,
1344 "scm_c_read must not call a fill function that pushes "
1345 "back characters onto an unbuffered port", SCM_EOL
);
1347 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1349 /* Reinstate the port's normal buffer. */
1354 /* The port has its own buffer. It is important that we use it,
1355 even if it happens to be smaller than our caller's buffer, so
1356 that a custom port implementation's entry points (in
1357 particular, fill_input) can rely on the buffer always being
1358 the same as they first set up. */
1359 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1361 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1362 memcpy (buffer
, pt
->read_pos
, n_available
);
1363 buffer
= (char *) buffer
+ n_available
;
1364 pt
->read_pos
+= n_available
;
1365 n_read
+= n_available
;
1366 size
-= n_available
;
1375 scm_c_read (SCM port
, void *buffer
, size_t size
)
1377 scm_i_pthread_mutex_t
*lock
;
1380 scm_c_lock_port (port
, &lock
);
1381 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1383 scm_i_pthread_mutex_unlock (lock
);
1389 /* Update the line and column number of PORT after consumption of C. */
1391 update_port_lf (scm_t_wchar c
, SCM port
)
1416 #define SCM_MBCHAR_BUF_SIZE (4)
1418 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1419 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1421 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1423 scm_t_wchar codepoint
;
1425 if (utf8_buf
[0] <= 0x7f)
1428 codepoint
= utf8_buf
[0];
1430 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1433 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1434 | (utf8_buf
[1] & 0x3f);
1436 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1439 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1440 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1441 | (utf8_buf
[2] & 0x3f);
1446 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1447 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1448 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1449 | (utf8_buf
[3] & 0x3f);
1455 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1456 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1457 representation, and set *LEN to the length in bytes. Return
1458 `EILSEQ' on error. */
1460 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1461 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1463 #define ASSERT_NOT_EOF(b) \
1464 if (SCM_UNLIKELY ((b) == EOF)) \
1466 #define CONSUME_PEEKED_BYTE() \
1473 pt
= SCM_PTAB_ENTRY (port
);
1475 byte
= scm_get_byte_or_eof_unlocked (port
);
1482 buf
[0] = (scm_t_uint8
) byte
;
1487 *codepoint
= buf
[0];
1488 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1491 byte
= scm_peek_byte_or_eof_unlocked (port
);
1492 ASSERT_NOT_EOF (byte
);
1494 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1497 CONSUME_PEEKED_BYTE ();
1498 buf
[1] = (scm_t_uint8
) byte
;
1501 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1504 else if ((buf
[0] & 0xf0) == 0xe0)
1507 byte
= scm_peek_byte_or_eof_unlocked (port
);
1508 ASSERT_NOT_EOF (byte
);
1510 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1511 || (buf
[0] == 0xe0 && byte
< 0xa0)
1512 || (buf
[0] == 0xed && byte
> 0x9f)))
1515 CONSUME_PEEKED_BYTE ();
1516 buf
[1] = (scm_t_uint8
) byte
;
1519 byte
= scm_peek_byte_or_eof_unlocked (port
);
1520 ASSERT_NOT_EOF (byte
);
1522 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1525 CONSUME_PEEKED_BYTE ();
1526 buf
[2] = (scm_t_uint8
) byte
;
1529 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1530 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1533 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1536 byte
= scm_peek_byte_or_eof_unlocked (port
);
1537 ASSERT_NOT_EOF (byte
);
1539 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1540 || (buf
[0] == 0xf0 && byte
< 0x90)
1541 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1544 CONSUME_PEEKED_BYTE ();
1545 buf
[1] = (scm_t_uint8
) byte
;
1548 byte
= scm_peek_byte_or_eof_unlocked (port
);
1549 ASSERT_NOT_EOF (byte
);
1551 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1554 CONSUME_PEEKED_BYTE ();
1555 buf
[2] = (scm_t_uint8
) byte
;
1558 byte
= scm_peek_byte_or_eof_unlocked (port
);
1559 ASSERT_NOT_EOF (byte
);
1561 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1564 CONSUME_PEEKED_BYTE ();
1565 buf
[3] = (scm_t_uint8
) byte
;
1568 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1569 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1570 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1579 /* Here we could choose the consume the faulty byte when it's not a
1580 valid starting byte, but it's not a requirement. What Section 3.9
1581 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1582 would otherwise be a valid starting byte. */
1586 #undef CONSUME_PEEKED_BYTE
1587 #undef ASSERT_NOT_EOF
1590 /* Read an ISO-8859-1 codepoint (a byte) from PORT. On success, return
1591 *0 and set CODEPOINT to the codepoint that was read, fill BUF with
1592 *its UTF-8 representation, and set *LEN to the length in bytes.
1593 *Return `EILSEQ' on error. */
1595 get_latin1_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1596 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1598 *codepoint
= scm_get_byte_or_eof_unlocked (port
);
1600 if (*codepoint
== EOF
)
1605 buf
[0] = *codepoint
;
1610 /* Likewise, read a byte sequence from PORT, passing it through its
1611 input conversion descriptor. */
1613 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1614 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1616 scm_t_iconv_descriptors
*id
;
1618 size_t bytes_consumed
, output_size
;
1620 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1622 id
= scm_i_port_iconv_descriptors (port
);
1624 for (output_size
= 0, output
= (char *) utf8_buf
,
1625 bytes_consumed
= 0, err
= 0;
1626 err
== 0 && output_size
== 0
1627 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1631 size_t input_left
, output_left
, done
;
1633 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1634 if (byte_read
== EOF
)
1636 if (bytes_consumed
== 0)
1638 *codepoint
= (scm_t_wchar
) EOF
;
1646 buf
[bytes_consumed
] = byte_read
;
1649 input_left
= bytes_consumed
+ 1;
1650 output_left
= sizeof (utf8_buf
);
1652 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1653 if (done
== (size_t) -1)
1657 /* Missing input: keep trying. */
1661 output_size
= sizeof (utf8_buf
) - output_left
;
1664 if (SCM_UNLIKELY (output_size
== 0))
1665 /* An unterminated sequence. */
1667 else if (SCM_LIKELY (err
== 0))
1669 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1670 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1671 *len
= bytes_consumed
;
1677 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1678 with the byte representation of the codepoint in PORT's encoding, and
1679 set *LEN to the length in bytes of that representation. Return 0 on
1680 success and an errno value on error. */
1682 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1683 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1686 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1688 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1689 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1690 else if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_LATIN1
)
1691 err
= get_latin1_codepoint (port
, codepoint
, buf
, len
);
1693 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1695 if (SCM_LIKELY (err
== 0))
1696 update_port_lf (*codepoint
, port
);
1697 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1701 update_port_lf (*codepoint
, port
);
1707 /* Read a codepoint from PORT and return it. */
1709 scm_getc_unlocked (SCM port
)
1710 #define FUNC_NAME "scm_getc"
1714 scm_t_wchar codepoint
;
1715 char buf
[SCM_MBCHAR_BUF_SIZE
];
1717 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1718 if (SCM_UNLIKELY (err
!= 0))
1719 /* At this point PORT should point past the invalid encoding, as per
1720 R6RS-lib Section 8.2.4. */
1721 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1730 scm_i_pthread_mutex_t
*lock
;
1733 scm_c_lock_port (port
, &lock
);
1734 ret
= scm_getc_unlocked (port
);
1736 scm_i_pthread_mutex_unlock (lock
);
1742 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1744 "Return the next character available from @var{port}, updating\n"
1745 "@var{port} to point to the following character. If no more\n"
1746 "characters are available, the end-of-file object is returned.\n"
1748 "When @var{port}'s data cannot be decoded according to its\n"
1749 "character encoding, a @code{decoding-error} is raised and\n"
1750 "@var{port} points past the erroneous byte sequence.\n")
1751 #define FUNC_NAME s_scm_read_char
1754 if (SCM_UNBNDP (port
))
1755 port
= scm_current_input_port ();
1756 SCM_VALIDATE_OPINPORT (1, port
);
1757 c
= scm_getc_unlocked (port
);
1760 return SCM_MAKE_CHAR (c
);
1770 scm_unget_byte_unlocked (int c
, SCM port
)
1771 #define FUNC_NAME "scm_unget_byte"
1773 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1775 if (pt
->read_buf
== pt
->putback_buf
)
1776 /* already using the put-back buffer. */
1778 /* enlarge putback_buf if necessary. */
1779 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1780 && pt
->read_buf
== pt
->read_pos
)
1782 size_t new_size
= pt
->read_buf_size
* 2;
1783 unsigned char *tmp
= (unsigned char *)
1784 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1787 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1788 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1789 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1792 /* shift any existing bytes to buffer + 1. */
1793 if (pt
->read_pos
== pt
->read_end
)
1794 pt
->read_end
= pt
->read_buf
+ 1;
1795 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1797 int count
= pt
->read_end
- pt
->read_pos
;
1799 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1800 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1803 pt
->read_pos
= pt
->read_buf
;
1806 /* switch to the put-back buffer. */
1808 if (pt
->putback_buf
== NULL
)
1811 = (unsigned char *) scm_gc_malloc_pointerless
1812 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1813 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1816 pt
->saved_read_buf
= pt
->read_buf
;
1817 pt
->saved_read_pos
= pt
->read_pos
;
1818 pt
->saved_read_end
= pt
->read_end
;
1819 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1821 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1822 pt
->read_end
= pt
->read_buf
+ 1;
1823 pt
->read_buf_size
= pt
->putback_buf_size
;
1829 pt
->rw_active
= SCM_PORT_READ
;
1834 scm_unget_byte (int c
, SCM port
)
1836 scm_i_pthread_mutex_t
*lock
;
1837 scm_c_lock_port (port
, &lock
);
1838 scm_unget_byte_unlocked (c
, port
);
1840 scm_i_pthread_mutex_unlock (lock
);
1845 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1846 #define FUNC_NAME "scm_ungetc"
1848 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1850 char result_buf
[10];
1851 const char *encoding
;
1855 if (pt
->encoding
!= NULL
)
1856 encoding
= pt
->encoding
;
1858 encoding
= "ISO-8859-1";
1860 len
= sizeof (result_buf
);
1861 result
= u32_conv_to_encoding (encoding
,
1862 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1863 (uint32_t *) &c
, 1, NULL
,
1866 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1867 scm_encoding_error (FUNC_NAME
, errno
,
1868 "conversion to port encoding failed",
1869 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1871 for (i
= len
- 1; i
>= 0; i
--)
1872 scm_unget_byte_unlocked (result
[i
], port
);
1874 if (SCM_UNLIKELY (result
!= result_buf
))
1879 /* What should col be in this case?
1880 * We'll leave it at -1.
1882 SCM_LINUM (port
) -= 1;
1890 scm_ungetc (scm_t_wchar c
, SCM port
)
1892 scm_i_pthread_mutex_t
*lock
;
1893 scm_c_lock_port (port
, &lock
);
1894 scm_ungetc_unlocked (c
, port
);
1896 scm_i_pthread_mutex_unlock (lock
);
1901 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1903 /* This is simple minded and inefficient, but unreading strings is
1904 * probably not a common operation, and remember that line and
1905 * column numbers have to be handled...
1907 * Please feel free to write an optimized version!
1910 scm_ungetc_unlocked (s
[n
], port
);
1914 scm_ungets (const char *s
, int n
, SCM port
)
1916 scm_i_pthread_mutex_t
*lock
;
1917 scm_c_lock_port (port
, &lock
);
1918 scm_ungets_unlocked (s
, n
, port
);
1920 scm_i_pthread_mutex_unlock (lock
);
1924 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1926 "Return the next character available from @var{port},\n"
1927 "@emph{without} updating @var{port} to point to the following\n"
1928 "character. If no more characters are available, the\n"
1929 "end-of-file object is returned.\n"
1931 "The value returned by\n"
1932 "a call to @code{peek-char} is the same as the value that would\n"
1933 "have been returned by a call to @code{read-char} on the same\n"
1934 "port. The only difference is that the very next call to\n"
1935 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1936 "return the value returned by the preceding call to\n"
1937 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1938 "an interactive port will hang waiting for input whenever a call\n"
1939 "to @code{read-char} would have hung.\n"
1941 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1942 "if such a situation occurs. However, unlike with @code{read-char},\n"
1943 "@var{port} still points at the beginning of the erroneous byte\n"
1944 "sequence when the error is raised.\n")
1945 #define FUNC_NAME s_scm_peek_char
1950 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1951 long column
, line
, i
;
1954 if (SCM_UNBNDP (port
))
1955 port
= scm_current_input_port ();
1956 SCM_VALIDATE_OPINPORT (1, port
);
1958 column
= SCM_COL (port
);
1959 line
= SCM_LINUM (port
);
1961 err
= get_codepoint (port
, &c
, bytes
, &len
);
1963 for (i
= len
- 1; i
>= 0; i
--)
1964 scm_unget_byte_unlocked (bytes
[i
], port
);
1966 SCM_COL (port
) = column
;
1967 SCM_LINUM (port
) = line
;
1969 if (SCM_UNLIKELY (err
!= 0))
1971 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1973 /* Shouldn't happen since `catch' always aborts to prompt. */
1974 result
= SCM_BOOL_F
;
1977 result
= SCM_EOF_VAL
;
1979 result
= SCM_MAKE_CHAR (c
);
1985 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
1986 (SCM cobj
, SCM port
),
1987 "Place character @var{cobj} in @var{port} so that it will be\n"
1988 "read by the next read operation. If called multiple times, the\n"
1989 "unread characters will be read again in last-in first-out\n"
1990 "order. If @var{port} is not supplied, the current input port\n"
1992 #define FUNC_NAME s_scm_unread_char
1996 SCM_VALIDATE_CHAR (1, cobj
);
1997 if (SCM_UNBNDP (port
))
1998 port
= scm_current_input_port ();
1999 SCM_VALIDATE_OPINPORT (2, port
);
2001 c
= SCM_CHAR (cobj
);
2003 scm_ungetc_unlocked (c
, port
);
2008 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2009 (SCM str
, SCM port
),
2010 "Place the string @var{str} in @var{port} so that its characters will be\n"
2011 "read in subsequent read operations. If called multiple times, the\n"
2012 "unread characters will be read again in last-in first-out order. If\n"
2013 "@var{port} is not supplied, the current-input-port is used.")
2014 #define FUNC_NAME s_scm_unread_string
2017 SCM_VALIDATE_STRING (1, str
);
2018 if (SCM_UNBNDP (port
))
2019 port
= scm_current_input_port ();
2020 SCM_VALIDATE_OPINPORT (2, port
);
2022 n
= scm_i_string_length (str
);
2025 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2034 /* Manipulating the buffers. */
2036 /* This routine does not take any locks, as it is usually called as part
2037 of a port implementation. */
2039 scm_port_non_buffer (scm_t_port
*pt
)
2041 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2042 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2043 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2044 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2047 /* this should only be called when the read buffer is empty. it
2048 tries to refill the read buffer. it returns the first char from
2049 the port, which is either EOF or *(pt->read_pos). */
2051 scm_fill_input_unlocked (SCM port
)
2053 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2055 assert (pt
->read_pos
== pt
->read_end
);
2057 if (pt
->read_buf
== pt
->putback_buf
)
2059 /* finished reading put-back chars. */
2060 pt
->read_buf
= pt
->saved_read_buf
;
2061 pt
->read_pos
= pt
->saved_read_pos
;
2062 pt
->read_end
= pt
->saved_read_end
;
2063 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2064 if (pt
->read_pos
< pt
->read_end
)
2065 return *(pt
->read_pos
);
2067 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2071 scm_fill_input (SCM port
)
2073 scm_i_pthread_mutex_t
*lock
;
2076 scm_c_lock_port (port
, &lock
);
2077 ret
= scm_fill_input_unlocked (port
);
2079 scm_i_pthread_mutex_unlock (lock
);
2085 /* move up to read_len chars from port's putback and/or read buffers
2086 into memory starting at dest. returns the number of chars moved. */
2088 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2090 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2091 size_t chars_read
= 0;
2092 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2096 memcpy (dest
, pt
->read_pos
, from_buf
);
2097 pt
->read_pos
+= from_buf
;
2098 chars_read
+= from_buf
;
2099 read_len
-= from_buf
;
2103 /* if putback was active, try the real input buffer too. */
2104 if (pt
->read_buf
== pt
->putback_buf
)
2106 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2109 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2110 pt
->saved_read_pos
+= from_buf
;
2111 chars_read
+= from_buf
;
2117 /* Clear a port's read buffers, returning the contents. */
2118 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2120 "This procedure clears a port's input buffers, similar\n"
2121 "to the way that force-output clears the output buffer. The\n"
2122 "contents of the buffers are returned as a single string, e.g.,\n"
2125 "(define p (open-input-file ...))\n"
2126 "(drain-input p) => empty string, nothing buffered yet.\n"
2127 "(unread-char (read-char p) p)\n"
2128 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2130 "Draining the buffers may be useful for cleanly finishing\n"
2131 "buffered I/O so that the file descriptor can be used directly\n"
2132 "for further input.")
2133 #define FUNC_NAME s_scm_drain_input
2140 SCM_VALIDATE_OPINPORT (1, port
);
2141 pt
= SCM_PTAB_ENTRY (port
);
2143 count
= pt
->read_end
- pt
->read_pos
;
2144 if (pt
->read_buf
== pt
->putback_buf
)
2145 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2149 result
= scm_i_make_string (count
, &data
, 0);
2150 scm_take_from_input_buffers (port
, data
, count
);
2153 result
= scm_nullstr
;
2160 scm_end_input_unlocked (SCM port
)
2163 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2165 if (pt
->read_buf
== pt
->putback_buf
)
2167 offset
= pt
->read_end
- pt
->read_pos
;
2168 pt
->read_buf
= pt
->saved_read_buf
;
2169 pt
->read_pos
= pt
->saved_read_pos
;
2170 pt
->read_end
= pt
->saved_read_end
;
2171 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2176 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2180 scm_end_input (SCM port
)
2182 scm_i_pthread_mutex_t
*lock
;
2183 scm_c_lock_port (port
, &lock
);
2184 scm_end_input_unlocked (port
);
2186 scm_i_pthread_mutex_unlock (lock
);
2190 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2192 "Flush the specified output port, or the current output port if @var{port}\n"
2193 "is omitted. The current output buffer contents are passed to the\n"
2194 "underlying port implementation (e.g., in the case of fports, the\n"
2195 "data will be written to the file and the output buffer will be cleared.)\n"
2196 "It has no effect on an unbuffered port.\n\n"
2197 "The return value is unspecified.")
2198 #define FUNC_NAME s_scm_force_output
2200 if (SCM_UNBNDP (port
))
2201 port
= scm_current_output_port ();
2204 port
= SCM_COERCE_OUTPORT (port
);
2205 SCM_VALIDATE_OPOUTPORT (1, port
);
2207 scm_flush_unlocked (port
);
2208 return SCM_UNSPECIFIED
;
2213 scm_flush_unlocked (SCM port
)
2215 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2219 scm_flush (SCM port
)
2221 scm_i_pthread_mutex_t
*lock
;
2222 scm_c_lock_port (port
, &lock
);
2223 scm_flush_unlocked (port
);
2225 scm_i_pthread_mutex_unlock (lock
);
2235 scm_putc (char c
, SCM port
)
2237 scm_i_pthread_mutex_t
*lock
;
2238 scm_c_lock_port (port
, &lock
);
2239 scm_putc_unlocked (c
, port
);
2241 scm_i_pthread_mutex_unlock (lock
);
2246 scm_puts (const char *s
, SCM port
)
2248 scm_i_pthread_mutex_t
*lock
;
2249 scm_c_lock_port (port
, &lock
);
2250 scm_puts_unlocked (s
, port
);
2252 scm_i_pthread_mutex_unlock (lock
);
2258 * Used by an application to write arbitrary number of bytes to an SCM
2259 * port. Similar semantics as libc write. However, unlike libc
2260 * write, scm_c_write writes the requested number of bytes and has no
2263 * Warning: Doesn't update port line and column counts!
2266 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2267 #define FUNC_NAME "scm_c_write"
2270 scm_t_ptob_descriptor
*ptob
;
2272 SCM_VALIDATE_OPOUTPORT (1, port
);
2274 pt
= SCM_PTAB_ENTRY (port
);
2275 ptob
= SCM_PORT_DESCRIPTOR (port
);
2277 if (pt
->rw_active
== SCM_PORT_READ
)
2278 scm_end_input_unlocked (port
);
2280 ptob
->write (port
, ptr
, size
);
2283 pt
->rw_active
= SCM_PORT_WRITE
;
2288 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2290 scm_i_pthread_mutex_t
*lock
;
2291 scm_c_lock_port (port
, &lock
);
2292 scm_c_write_unlocked (port
, ptr
, size
);
2294 scm_i_pthread_mutex_unlock (lock
);
2300 * This function differs from scm_c_write; it updates port line and
2303 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2305 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2306 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2308 if (pt
->rw_active
== SCM_PORT_READ
)
2309 scm_end_input_unlocked (port
);
2311 ptob
->write (port
, ptr
, size
);
2313 for (; size
; ptr
++, size
--)
2314 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2317 pt
->rw_active
= SCM_PORT_WRITE
;
2321 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2323 scm_i_pthread_mutex_t
*lock
;
2324 scm_c_lock_port (port
, &lock
);
2325 scm_lfwrite_unlocked (ptr
, size
, port
);
2327 scm_i_pthread_mutex_unlock (lock
);
2331 /* Write STR to PORT from START inclusive to END exclusive. */
2333 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2335 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2337 if (pt
->rw_active
== SCM_PORT_READ
)
2338 scm_end_input_unlocked (port
);
2340 if (end
== (size_t) -1)
2341 end
= scm_i_string_length (str
);
2343 scm_display (scm_c_substring (str
, start
, end
), port
);
2346 pt
->rw_active
= SCM_PORT_WRITE
;
2352 /* Querying and setting positions, and character availability. */
2354 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2356 "Return @code{#t} if a character is ready on input @var{port}\n"
2357 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2358 "@code{#t} then the next @code{read-char} operation on\n"
2359 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2360 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2362 "@code{char-ready?} exists to make it possible for a\n"
2363 "program to accept characters from interactive ports without\n"
2364 "getting stuck waiting for input. Any input editors associated\n"
2365 "with such ports must make sure that characters whose existence\n"
2366 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2367 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2368 "a port at end of file would be indistinguishable from an\n"
2369 "interactive port that has no ready characters.")
2370 #define FUNC_NAME s_scm_char_ready_p
2374 if (SCM_UNBNDP (port
))
2375 port
= scm_current_input_port ();
2376 /* It's possible to close the current input port, so validate even in
2378 SCM_VALIDATE_OPINPORT (1, port
);
2380 pt
= SCM_PTAB_ENTRY (port
);
2382 /* if the current read buffer is filled, or the
2383 last pushed-back char has been read and the saved buffer is
2384 filled, result is true. */
2385 if (pt
->read_pos
< pt
->read_end
2386 || (pt
->read_buf
== pt
->putback_buf
2387 && pt
->saved_read_pos
< pt
->saved_read_end
))
2391 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2393 if (ptob
->input_waiting
)
2394 return scm_from_bool(ptob
->input_waiting (port
));
2401 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2402 (SCM fd_port
, SCM offset
, SCM whence
),
2403 "Sets the current position of @var{fd_port} to the integer\n"
2404 "@var{offset}, which is interpreted according to the value of\n"
2407 "One of the following variables should be supplied for\n"
2409 "@defvar SEEK_SET\n"
2410 "Seek from the beginning of the file.\n"
2412 "@defvar SEEK_CUR\n"
2413 "Seek from the current position.\n"
2415 "@defvar SEEK_END\n"
2416 "Seek from the end of the file.\n"
2418 "If @var{fd_port} is a file descriptor, the underlying system\n"
2419 "call is @code{lseek}. @var{port} may be a string port.\n"
2421 "The value returned is the new position in the file. This means\n"
2422 "that the current position of a port can be obtained using:\n"
2424 "(seek port 0 SEEK_CUR)\n"
2426 #define FUNC_NAME s_scm_seek
2430 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2432 how
= scm_to_int (whence
);
2433 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2434 SCM_OUT_OF_RANGE (3, whence
);
2436 if (SCM_OPPORTP (fd_port
))
2438 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2439 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2440 off_t_or_off64_t rv
;
2443 SCM_MISC_ERROR ("port is not seekable",
2444 scm_cons (fd_port
, SCM_EOL
));
2446 rv
= ptob
->seek (fd_port
, off
, how
);
2447 return scm_from_off_t_or_off64_t (rv
);
2449 else /* file descriptor?. */
2451 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2452 off_t_or_off64_t rv
;
2453 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2456 return scm_from_off_t_or_off64_t (rv
);
2465 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2466 doesn't have the filename version truncate(), hence this code. */
2467 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2469 truncate (const char *file
, off_t length
)
2473 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2477 ret
= ftruncate (fdes
, length
);
2480 int save_errno
= errno
;
2486 return close (fdes
);
2488 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2490 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2491 (SCM object
, SCM length
),
2492 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2493 "can be a filename string, a port object, or an integer file\n"
2495 "The return value is unspecified.\n"
2497 "For a port or file descriptor @var{length} can be omitted, in\n"
2498 "which case the file is truncated at the current position (per\n"
2499 "@code{ftell} above).\n"
2501 "On most systems a file can be extended by giving a length\n"
2502 "greater than the current size, but this is not mandatory in the\n"
2504 #define FUNC_NAME s_scm_truncate_file
2508 /* "object" can be a port, fdes or filename.
2510 Negative "length" makes no sense, but it's left to truncate() or
2511 ftruncate() to give back an error for that (normally EINVAL).
2514 if (SCM_UNBNDP (length
))
2516 /* must supply length if object is a filename. */
2517 if (scm_is_string (object
))
2518 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2520 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2523 object
= SCM_COERCE_OUTPORT (object
);
2524 if (scm_is_integer (object
))
2526 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2527 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2530 else if (SCM_OPOUTPORTP (object
))
2532 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2533 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2534 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2536 if (!ptob
->truncate
)
2537 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2538 if (pt
->rw_active
== SCM_PORT_READ
)
2539 scm_end_input_unlocked (object
);
2540 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2541 ptob
->flush (object
);
2543 ptob
->truncate (object
, c_length
);
2548 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2549 char *str
= scm_to_locale_string (object
);
2551 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2558 return SCM_UNSPECIFIED
;
2562 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2564 "Return the current line number for @var{port}.\n"
2566 "The first line of a file is 0. But you might want to add 1\n"
2567 "when printing line numbers, since starting from 1 is\n"
2568 "traditional in error messages, and likely to be more natural to\n"
2570 #define FUNC_NAME s_scm_port_line
2572 port
= SCM_COERCE_OUTPORT (port
);
2573 SCM_VALIDATE_OPENPORT (1, port
);
2574 return scm_from_long (SCM_LINUM (port
));
2578 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2579 (SCM port
, SCM line
),
2580 "Set the current line number for @var{port} to @var{line}. The\n"
2581 "first line of a file is 0.")
2582 #define FUNC_NAME s_scm_set_port_line_x
2584 port
= SCM_COERCE_OUTPORT (port
);
2585 SCM_VALIDATE_OPENPORT (1, port
);
2586 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2587 return SCM_UNSPECIFIED
;
2591 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2593 "Return the current column number of @var{port}.\n"
2594 "If the number is\n"
2595 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2596 "- i.e. the first character of the first line is line 0, column 0.\n"
2597 "(However, when you display a file position, for example in an error\n"
2598 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2599 "because lines and column numbers traditionally start with 1, and that is\n"
2600 "what non-programmers will find most natural.)")
2601 #define FUNC_NAME s_scm_port_column
2603 port
= SCM_COERCE_OUTPORT (port
);
2604 SCM_VALIDATE_OPENPORT (1, port
);
2605 return scm_from_int (SCM_COL (port
));
2609 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2610 (SCM port
, SCM column
),
2611 "Set the current column of @var{port}. Before reading the first\n"
2612 "character on a line the column should be 0.")
2613 #define FUNC_NAME s_scm_set_port_column_x
2615 port
= SCM_COERCE_OUTPORT (port
);
2616 SCM_VALIDATE_OPENPORT (1, port
);
2617 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2618 return SCM_UNSPECIFIED
;
2622 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2624 "Return the filename associated with @var{port}, or @code{#f}\n"
2625 "if no filename is associated with the port.")
2626 #define FUNC_NAME s_scm_port_filename
2628 port
= SCM_COERCE_OUTPORT (port
);
2629 SCM_VALIDATE_OPENPORT (1, port
);
2630 return SCM_FILENAME (port
);
2634 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2635 (SCM port
, SCM filename
),
2636 "Change the filename associated with @var{port}, using the current input\n"
2637 "port if none is specified. Note that this does not change the port's\n"
2638 "source of data, but only the value that is returned by\n"
2639 "@code{port-filename} and reported in diagnostic output.")
2640 #define FUNC_NAME s_scm_set_port_filename_x
2642 port
= SCM_COERCE_OUTPORT (port
);
2643 SCM_VALIDATE_OPENPORT (1, port
);
2644 /* We allow the user to set the filename to whatever he likes. */
2645 SCM_SET_FILENAME (port
, filename
);
2646 return SCM_UNSPECIFIED
;
2653 /* Implementation helpers for port printing functions. */
2656 scm_print_port_mode (SCM exp
, SCM port
)
2658 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2660 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2661 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2664 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2671 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2673 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2676 scm_puts_unlocked ("#<", port
);
2677 scm_print_port_mode (exp
, port
);
2678 scm_puts_unlocked (type
, port
);
2679 scm_putc_unlocked (' ', port
);
2680 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2681 scm_putc_unlocked ('>', port
);
2688 /* Iterating over all ports. */
2690 struct for_each_data
2692 void (*proc
) (void *data
, SCM p
);
2697 for_each_trampoline (void *data
, SCM port
, SCM result
)
2699 struct for_each_data
*d
= data
;
2701 d
->proc (d
->data
, port
);
2707 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2709 struct for_each_data d
;
2714 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2715 scm_i_port_weak_set
);
2719 scm_for_each_trampoline (void *data
, SCM port
)
2721 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2724 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2726 "Apply @var{proc} to each port in the Guile port table\n"
2727 "in turn. The return value is unspecified. More specifically,\n"
2728 "@var{proc} is applied exactly once to every port that exists\n"
2729 "in the system at the time @code{port-for-each} is invoked.\n"
2730 "Changes to the port table while @code{port-for-each} is running\n"
2731 "have no effect as far as @code{port-for-each} is concerned.")
2732 #define FUNC_NAME s_scm_port_for_each
2734 SCM_VALIDATE_PROC (1, proc
);
2736 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2738 return SCM_UNSPECIFIED
;
2743 flush_output_port (void *closure
, SCM port
)
2745 if (SCM_OPOUTPORTP (port
))
2746 scm_flush_unlocked (port
);
2749 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2751 "Equivalent to calling @code{force-output} on\n"
2752 "all open output ports. The return value is unspecified.")
2753 #define FUNC_NAME s_scm_flush_all_ports
2755 scm_c_port_for_each (&flush_output_port
, NULL
);
2756 return SCM_UNSPECIFIED
;
2765 scm_t_bits scm_tc16_void_port
= 0;
2767 static int fill_input_void_port (SCM port SCM_UNUSED
)
2773 write_void_port (SCM port SCM_UNUSED
,
2774 const void *data SCM_UNUSED
,
2775 size_t size SCM_UNUSED
)
2780 scm_i_void_port (long mode_bits
)
2784 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2786 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2792 scm_void_port (char *mode_str
)
2794 return scm_i_void_port (scm_mode_bits (mode_str
));
2797 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2799 "Create and return a new void port. A void port acts like\n"
2800 "@file{/dev/null}. The @var{mode} argument\n"
2801 "specifies the input/output modes for this port: see the\n"
2802 "documentation for @code{open-file} in @ref{File Ports}.")
2803 #define FUNC_NAME s_scm_sys_make_void_port
2805 return scm_i_void_port (scm_i_mode_bits (mode
));
2812 /* Initialization. */
2817 /* lseek() symbols. */
2818 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2819 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2820 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2822 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2825 cur_inport_fluid
= scm_make_fluid ();
2826 cur_outport_fluid
= scm_make_fluid ();
2827 cur_errport_fluid
= scm_make_fluid ();
2828 cur_loadport_fluid
= scm_make_fluid ();
2830 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2832 #include "libguile/ports.x"
2834 /* Use Latin-1 as the default port encoding. */
2835 SCM_VARIABLE_SET (default_port_encoding_var
,
2836 scm_make_fluid_with_default (SCM_BOOL_F
));
2837 scm_port_encoding_init
= 1;
2839 SCM_VARIABLE_SET (scm_conversion_strategy
,
2840 scm_make_fluid_with_default
2841 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2842 scm_conversion_strategy_init
= 1;
2844 /* These bindings are used when boot-9 turns `current-input-port' et
2845 al into parameters. They are then removed from the guile module. */
2846 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2847 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2848 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);