1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011 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_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->flush
= flush
;
257 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
259 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
263 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
265 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
269 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
271 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
275 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
277 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
282 /* Standard ports --- current input, output, error, and more(!). */
284 static SCM cur_inport_fluid
= SCM_BOOL_F
;
285 static SCM cur_outport_fluid
= SCM_BOOL_F
;
286 static SCM cur_errport_fluid
= SCM_BOOL_F
;
287 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
289 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
291 "Return the current input port. This is the default port used\n"
292 "by many input procedures. Initially, @code{current-input-port}\n"
293 "returns the @dfn{standard input} in Unix and C terminology.")
294 #define FUNC_NAME s_scm_current_input_port
296 if (scm_is_true (cur_inport_fluid
))
297 return scm_fluid_ref (cur_inport_fluid
);
303 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
305 "Return the current output port. This is the default port used\n"
306 "by many output procedures. Initially,\n"
307 "@code{current-output-port} returns the @dfn{standard output} in\n"
308 "Unix and C terminology.")
309 #define FUNC_NAME s_scm_current_output_port
311 if (scm_is_true (cur_outport_fluid
))
312 return scm_fluid_ref (cur_outport_fluid
);
318 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
320 "Return the port to which errors and warnings should be sent (the\n"
321 "@dfn{standard error} in Unix and C terminology).")
322 #define FUNC_NAME s_scm_current_error_port
324 if (scm_is_true (cur_errport_fluid
))
325 return scm_fluid_ref (cur_errport_fluid
);
331 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
333 "Return the current-load-port.\n"
334 "The load port is used internally by @code{primitive-load}.")
335 #define FUNC_NAME s_scm_current_load_port
337 return scm_fluid_ref (cur_loadport_fluid
);
341 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
343 "@deffnx {Scheme Procedure} set-current-output-port port\n"
344 "@deffnx {Scheme Procedure} set-current-error-port port\n"
345 "Change the ports returned by @code{current-input-port},\n"
346 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
347 "so that they use the supplied @var{port} for input or output.")
348 #define FUNC_NAME s_scm_set_current_input_port
350 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
351 SCM_VALIDATE_OPINPORT (1, port
);
352 scm_fluid_set_x (cur_inport_fluid
, port
);
358 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
360 "Set the current default output port to @var{port}.")
361 #define FUNC_NAME s_scm_set_current_output_port
363 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
364 port
= SCM_COERCE_OUTPORT (port
);
365 SCM_VALIDATE_OPOUTPORT (1, port
);
366 scm_fluid_set_x (cur_outport_fluid
, port
);
372 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
374 "Set the current default error port to @var{port}.")
375 #define FUNC_NAME s_scm_set_current_error_port
377 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
378 port
= SCM_COERCE_OUTPORT (port
);
379 SCM_VALIDATE_OPOUTPORT (1, port
);
380 scm_fluid_set_x (cur_errport_fluid
, port
);
386 scm_dynwind_current_input_port (SCM port
)
387 #define FUNC_NAME NULL
389 SCM_VALIDATE_OPINPORT (1, port
);
390 scm_dynwind_fluid (cur_inport_fluid
, port
);
395 scm_dynwind_current_output_port (SCM port
)
396 #define FUNC_NAME NULL
398 port
= SCM_COERCE_OUTPORT (port
);
399 SCM_VALIDATE_OPOUTPORT (1, port
);
400 scm_dynwind_fluid (cur_outport_fluid
, port
);
405 scm_dynwind_current_error_port (SCM port
)
406 #define FUNC_NAME NULL
408 port
= SCM_COERCE_OUTPORT (port
);
409 SCM_VALIDATE_OPOUTPORT (1, port
);
410 scm_dynwind_fluid (cur_errport_fluid
, port
);
415 scm_i_dynwind_current_load_port (SCM port
)
417 scm_dynwind_fluid (cur_loadport_fluid
, port
);
423 /* Retrieving a port's mode. */
425 /* Return the flags that characterize a port based on the mode
426 * string used to open a file for that port.
428 * See PORT FLAGS in scm.h
432 scm_i_mode_bits_n (SCM modes
)
435 | (scm_i_string_contains_char (modes
, 'r')
436 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
437 | (scm_i_string_contains_char (modes
, 'w')
438 || scm_i_string_contains_char (modes
, 'a')
439 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
440 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
441 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
445 scm_mode_bits (char *modes
)
447 return scm_i_mode_bits (scm_from_locale_string (modes
));
451 scm_i_mode_bits (SCM modes
)
455 if (!scm_is_string (modes
))
456 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
458 bits
= scm_i_mode_bits_n (modes
);
459 scm_remember_upto_here_1 (modes
);
463 /* Return the mode flags from an open port.
464 * Some modes such as "append" are only used when opening
465 * a file and are not returned here. */
467 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
469 "Return the port modes associated with the open port @var{port}.\n"
470 "These will not necessarily be identical to the modes used when\n"
471 "the port was opened, since modes such as \"append\" which are\n"
472 "used only during port creation are not retained.")
473 #define FUNC_NAME s_scm_port_mode
478 port
= SCM_COERCE_OUTPORT (port
);
479 SCM_VALIDATE_OPPORT (1, port
);
480 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
481 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
482 strcpy (modes
, "r+");
486 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
488 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
491 return scm_from_latin1_string (modes
);
498 /* The port table --- a weak set of all ports.
500 We need a global registry of ports to flush them all at exit, and to
501 get all the ports matching a file descriptor. */
502 SCM scm_i_port_weak_set
;
507 /* Port finalization. */
509 static void finalize_port (GC_PTR
, GC_PTR
);
511 /* Register a finalizer for PORT. */
512 static SCM_C_INLINE_KEYWORD
void
513 register_finalizer_for_port (SCM port
)
515 GC_finalization_proc prev_finalizer
;
516 GC_PTR prev_finalization_data
;
518 /* Register a finalizer for PORT so that its iconv CDs get freed and
519 optionally its type's `free' function gets called. */
520 GC_REGISTER_FINALIZER_NO_ORDER (SCM_HEAP_OBJECT_BASE (port
),
523 &prev_finalization_data
);
526 /* Finalize the object (a port) pointed to by PTR. */
528 finalize_port (GC_PTR ptr
, GC_PTR data
)
530 SCM port
= SCM_PACK_POINTER (ptr
);
532 if (!SCM_PORTP (port
))
535 if (SCM_OPENP (port
))
537 if (SCM_REVEALED (port
) > 0)
538 /* Keep "revealed" ports alive and re-register a finalizer. */
539 register_finalizer_for_port (port
);
542 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
546 /* Yes, I really do mean `free' rather than `close'. `close'
547 is for explicit `close-port' by user. */
550 entry
= SCM_PTAB_ENTRY (port
);
552 if (entry
->input_cd
!= (iconv_t
) -1)
553 iconv_close (entry
->input_cd
);
554 if (entry
->output_cd
!= (iconv_t
) -1)
555 iconv_close (entry
->output_cd
);
557 SCM_SETSTREAM (port
, 0);
558 SCM_CLR_PORT_OPEN_FLAG (port
);
560 scm_gc_ports_collected
++;
569 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
570 const char *encoding
,
571 scm_t_string_failed_conversion_handler handler
,
576 scm_t_ptob_descriptor
*ptob
;
578 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
579 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
581 ret
= scm_words (tag
| mode_bits
, 3);
582 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
583 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
587 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
588 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
591 entry
->file_name
= SCM_BOOL_F
;
592 entry
->rw_active
= SCM_PORT_NEITHER
;
594 entry
->stream
= stream
;
595 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
596 /* The conversion descriptors will be opened lazily. */
597 entry
->input_cd
= (iconv_t
) -1;
598 entry
->output_cd
= (iconv_t
) -1;
599 entry
->ilseq_handler
= handler
;
601 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
603 /* For each new port, register a finalizer so that it port type's free
604 function can be invoked eventually. */
605 register_finalizer_for_port (ret
);
611 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
613 return scm_c_make_port_with_encoding (tag
, mode_bits
,
614 scm_i_default_port_encoding (),
615 scm_i_get_conversion_strategy (SCM_BOOL_F
),
620 scm_new_port_table_entry (scm_t_bits tag
)
622 return scm_c_make_port (tag
, 0, 0);
625 /* Remove a port from the table and destroy it. */
628 scm_i_remove_port (SCM port
)
629 #define FUNC_NAME "scm_remove_port"
633 p
= SCM_PTAB_ENTRY (port
);
634 scm_port_non_buffer (p
);
635 SCM_SETPTAB_ENTRY (port
, 0);
636 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
638 p
->putback_buf
= NULL
;
639 p
->putback_buf_size
= 0;
641 if (p
->input_cd
!= (iconv_t
) -1)
643 iconv_close (p
->input_cd
);
644 p
->input_cd
= (iconv_t
) -1;
647 if (p
->output_cd
!= (iconv_t
) -1)
649 iconv_close (p
->output_cd
);
650 p
->output_cd
= (iconv_t
) -1;
660 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
662 "Return a boolean indicating whether @var{x} is a port.\n"
663 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
665 #define FUNC_NAME s_scm_port_p
667 return scm_from_bool (SCM_PORTP (x
));
671 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
673 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
674 "@code{#f}. Any object satisfying this predicate also satisfies\n"
676 #define FUNC_NAME s_scm_input_port_p
678 return scm_from_bool (SCM_INPUT_PORT_P (x
));
682 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
684 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
685 "@code{#f}. Any object satisfying this predicate also satisfies\n"
687 #define FUNC_NAME s_scm_output_port_p
689 x
= SCM_COERCE_OUTPORT (x
);
690 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
694 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
696 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
698 #define FUNC_NAME s_scm_port_closed_p
700 SCM_VALIDATE_PORT (1, port
);
701 return scm_from_bool (!SCM_OPPORTP (port
));
705 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
707 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
709 #define FUNC_NAME s_scm_eof_object_p
711 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
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
736 port
= SCM_COERCE_OUTPORT (port
);
738 SCM_VALIDATE_PORT (1, port
);
739 if (SCM_CLOSEDP (port
))
741 if (SCM_PORT_DESCRIPTOR (port
)->close
)
742 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
745 scm_i_remove_port (port
);
746 SCM_CLR_PORT_OPEN_FLAG (port
);
747 return scm_from_bool (rv
>= 0);
751 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
753 "Close the specified input port object. The routine has no effect if\n"
754 "the file has already been closed. An exception may be raised if an\n"
755 "error occurs. The value returned is unspecified.\n\n"
756 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
757 "which can close file descriptors.")
758 #define FUNC_NAME s_scm_close_input_port
760 SCM_VALIDATE_INPUT_PORT (1, port
);
761 scm_close_port (port
);
762 return SCM_UNSPECIFIED
;
766 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
768 "Close the specified output port object. The routine has no effect if\n"
769 "the file has already been closed. An exception may be raised if an\n"
770 "error occurs. The value returned is unspecified.\n\n"
771 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
772 "which can close file descriptors.")
773 #define FUNC_NAME s_scm_close_output_port
775 port
= SCM_COERCE_OUTPORT (port
);
776 SCM_VALIDATE_OUTPUT_PORT (1, port
);
777 scm_close_port (port
);
778 return SCM_UNSPECIFIED
;
785 /* Encoding characters to byte streams, and decoding byte streams to
788 /* A fluid specifying the default encoding for newly created ports. If it is
789 a string, that is the encoding. If it is #f, it is in the "native"
790 (Latin-1) encoding. */
791 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
793 static int scm_port_encoding_init
= 0;
795 /* Use ENCODING as the default encoding for future ports. */
797 scm_i_set_default_port_encoding (const char *encoding
)
799 if (!scm_port_encoding_init
800 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
801 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
805 || !strcmp (encoding
, "ASCII")
806 || !strcmp (encoding
, "ANSI_X3.4-1968")
807 || !strcmp (encoding
, "ISO-8859-1"))
808 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
810 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
811 scm_from_locale_string (encoding
));
814 /* Return the name of the default encoding for newly created ports; a
815 return value of NULL means "ISO-8859-1". */
817 scm_i_default_port_encoding (void)
819 if (!scm_port_encoding_init
)
821 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
827 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
828 if (!scm_is_string (encoding
))
831 return scm_i_string_chars (encoding
);
836 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
839 iconv_t new_input_cd
, new_output_cd
;
841 new_input_cd
= (iconv_t
) -1;
842 new_output_cd
= (iconv_t
) -1;
844 /* Set the character encoding for this port. */
845 pt
= SCM_PTAB_ENTRY (port
);
847 if (encoding
== NULL
)
848 encoding
= "ISO-8859-1";
850 if (pt
->encoding
!= encoding
)
851 pt
->encoding
= scm_gc_strdup (encoding
, "port");
853 /* If ENCODING is UTF-8, then no conversion descriptor is opened
854 because we do I/O ourselves. This saves 100+ KiB for each
856 if (strcmp (encoding
, "UTF-8"))
858 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
)
860 /* Open an input iconv conversion descriptor, from ENCODING
861 to UTF-8. We choose UTF-8, not UTF-32, because iconv
862 implementations can typically convert from anything to
863 UTF-8, but not to UTF-32 (see
864 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
865 new_input_cd
= iconv_open ("UTF-8", encoding
);
866 if (new_input_cd
== (iconv_t
) -1)
867 goto invalid_encoding
;
870 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
872 new_output_cd
= iconv_open (encoding
, "UTF-8");
873 if (new_output_cd
== (iconv_t
) -1)
875 if (new_input_cd
!= (iconv_t
) -1)
876 iconv_close (new_input_cd
);
877 goto invalid_encoding
;
882 if (pt
->input_cd
!= (iconv_t
) -1)
883 iconv_close (pt
->input_cd
);
884 if (pt
->output_cd
!= (iconv_t
) -1)
885 iconv_close (pt
->output_cd
);
887 pt
->input_cd
= new_input_cd
;
888 pt
->output_cd
= new_output_cd
;
895 err
= scm_from_locale_string (encoding
);
896 scm_misc_error ("scm_i_set_port_encoding_x",
897 "invalid or unknown character encoding ~s",
902 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
904 "Returns, as a string, the character encoding that @var{port}\n"
905 "uses to interpret its input and output.\n")
906 #define FUNC_NAME s_scm_port_encoding
911 SCM_VALIDATE_PORT (1, port
);
913 pt
= SCM_PTAB_ENTRY (port
);
916 return scm_from_locale_string (pt
->encoding
);
922 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
924 "Sets the character encoding that will be used to interpret all\n"
925 "port I/O. New ports are created with the encoding\n"
926 "appropriate for the current locale if @code{setlocale} has \n"
927 "been called or ISO-8859-1 otherwise\n"
928 "and this procedure can be used to modify that encoding.\n")
929 #define FUNC_NAME s_scm_set_port_encoding_x
933 SCM_VALIDATE_PORT (1, port
);
934 SCM_VALIDATE_STRING (2, enc
);
936 enc_str
= scm_to_locale_string (enc
);
937 scm_i_set_port_encoding_x (port
, enc_str
);
940 return SCM_UNSPECIFIED
;
945 /* This determines how conversions handle unconvertible characters. */
946 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
947 static int scm_conversion_strategy_init
= 0;
949 scm_t_string_failed_conversion_handler
950 scm_i_get_conversion_strategy (SCM port
)
954 if (scm_is_false (port
))
956 if (!scm_conversion_strategy_init
957 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
958 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
961 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
962 if (scm_is_false (encoding
))
963 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
965 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
971 pt
= SCM_PTAB_ENTRY (port
);
972 return pt
->ilseq_handler
;
978 scm_i_set_conversion_strategy_x (SCM port
,
979 scm_t_string_failed_conversion_handler handler
)
984 strategy
= scm_from_int ((int) handler
);
986 if (scm_is_false (port
))
988 /* Set the default encoding for future ports. */
989 if (!scm_conversion_strategy_init
990 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
991 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
993 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
997 /* Set the character encoding for this port. */
998 pt
= SCM_PTAB_ENTRY (port
);
999 pt
->ilseq_handler
= handler
;
1003 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1004 1, 0, 0, (SCM port
),
1005 "Returns the behavior of the port when handling a character that\n"
1006 "is not representable in the port's current encoding.\n"
1007 "It returns the symbol @code{error} if unrepresentable characters\n"
1008 "should cause exceptions, @code{substitute} if the port should\n"
1009 "try to replace unrepresentable characters with question marks or\n"
1010 "approximate characters, or @code{escape} if unrepresentable\n"
1011 "characters should be converted to string escapes.\n"
1013 "If @var{port} is @code{#f}, then the current default behavior\n"
1014 "will be returned. New ports will have this default behavior\n"
1015 "when they are created.\n")
1016 #define FUNC_NAME s_scm_port_conversion_strategy
1018 scm_t_string_failed_conversion_handler h
;
1020 SCM_VALIDATE_OPPORT (1, port
);
1022 if (!scm_is_false (port
))
1024 SCM_VALIDATE_OPPORT (1, port
);
1027 h
= scm_i_get_conversion_strategy (port
);
1028 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1029 return scm_from_latin1_symbol ("error");
1030 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1031 return scm_from_latin1_symbol ("substitute");
1032 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1033 return scm_from_latin1_symbol ("escape");
1037 /* Never gets here. */
1038 return SCM_UNDEFINED
;
1042 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1044 (SCM port
, SCM sym
),
1045 "Sets the behavior of the interpreter when outputting a character\n"
1046 "that is not representable in the port's current encoding.\n"
1047 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1048 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1049 "when an unconvertible character is encountered. If it is\n"
1050 "@code{'substitute}, then unconvertible characters will \n"
1051 "be replaced with approximate characters, or with question marks\n"
1052 "if no approximately correct character is available.\n"
1053 "If it is @code{'escape},\n"
1054 "it will appear as a hex escape when output.\n"
1056 "If @var{port} is an open port, the conversion error behavior\n"
1057 "is set for that port. If it is @code{#f}, it is set as the\n"
1058 "default behavior for any future ports that get created in\n"
1060 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1066 if (!scm_is_false (port
))
1068 SCM_VALIDATE_OPPORT (1, port
);
1071 err
= scm_from_latin1_symbol ("error");
1072 if (scm_is_true (scm_eqv_p (sym
, err
)))
1074 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1075 return SCM_UNSPECIFIED
;
1078 qm
= scm_from_latin1_symbol ("substitute");
1079 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1081 scm_i_set_conversion_strategy_x (port
,
1082 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1083 return SCM_UNSPECIFIED
;
1086 esc
= scm_from_latin1_symbol ("escape");
1087 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1089 scm_i_set_conversion_strategy_x (port
,
1090 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1091 return SCM_UNSPECIFIED
;
1094 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1096 return SCM_UNSPECIFIED
;
1103 /* The port lock. */
1106 lock_port (void *mutex
)
1108 scm_i_pthread_mutex_lock (mutex
);
1112 unlock_port (void *mutex
)
1114 scm_i_pthread_mutex_unlock (mutex
);
1118 scm_dynwind_lock_port (SCM port
)
1119 #define FUNC_NAME "dynwind-lock-port"
1121 scm_i_pthread_mutex_t
*lock
;
1122 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1123 scm_c_lock_port (port
, &lock
);
1126 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1127 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1135 /* Revealed counts --- an oddity inherited from SCSH. */
1137 /* Find a port in the table and return its revealed count.
1138 Also used by the garbage collector.
1141 scm_revealed_count (SCM port
)
1143 scm_i_pthread_mutex_t
*lock
;
1146 scm_c_lock_port (port
, &lock
);
1147 ret
= SCM_REVEALED (port
);
1149 scm_i_pthread_mutex_unlock (lock
);
1154 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
1156 "Return the revealed count for @var{port}.")
1157 #define FUNC_NAME s_scm_port_revealed
1159 port
= SCM_COERCE_OUTPORT (port
);
1160 SCM_VALIDATE_OPENPORT (1, port
);
1161 return scm_from_int (scm_revealed_count (port
));
1165 /* Set the revealed count for a port. */
1166 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
1167 (SCM port
, SCM rcount
),
1168 "Sets the revealed count for a port to a given value.\n"
1169 "The return value is unspecified.")
1170 #define FUNC_NAME s_scm_set_port_revealed_x
1173 scm_i_pthread_mutex_t
*lock
;
1175 port
= SCM_COERCE_OUTPORT (port
);
1176 SCM_VALIDATE_OPENPORT (1, port
);
1177 r
= scm_to_int (rcount
);
1178 scm_c_lock_port (port
, &lock
);
1179 SCM_REVEALED (port
) = r
;
1181 scm_i_pthread_mutex_unlock (lock
);
1182 return SCM_UNSPECIFIED
;
1186 /* Set the revealed count for a port. */
1187 SCM_DEFINE (scm_adjust_port_revealed_x
, "adjust-port-revealed!", 2, 0, 0,
1188 (SCM port
, SCM addend
),
1189 "Add @var{addend} to the revealed count of @var{port}.\n"
1190 "The return value is unspecified.")
1191 #define FUNC_NAME s_scm_set_port_revealed_x
1193 scm_i_pthread_mutex_t
*lock
;
1195 port
= SCM_COERCE_OUTPORT (port
);
1196 SCM_VALIDATE_OPENPORT (1, port
);
1197 a
= scm_to_int (addend
);
1198 scm_c_lock_port (port
, &lock
);
1199 SCM_REVEALED (port
) += a
;
1201 scm_i_pthread_mutex_unlock (lock
);
1202 return SCM_UNSPECIFIED
;
1212 scm_get_byte_or_eof (SCM port
)
1214 scm_i_pthread_mutex_t
*lock
;
1217 scm_c_lock_port (port
, &lock
);
1218 ret
= scm_get_byte_or_eof_unlocked (port
);
1220 scm_i_pthread_mutex_unlock (lock
);
1226 scm_peek_byte_or_eof (SCM port
)
1228 scm_i_pthread_mutex_t
*lock
;
1231 scm_c_lock_port (port
, &lock
);
1232 ret
= scm_peek_byte_or_eof_unlocked (port
);
1234 scm_i_pthread_mutex_unlock (lock
);
1241 * Used by an application to read arbitrary number of bytes from an
1242 * SCM port. Same semantics as libc read, except that scm_c_read only
1243 * returns less than SIZE bytes if at end-of-file.
1245 * Warning: Doesn't update port line and column counts! */
1247 /* This structure, and the following swap_buffer function, are used
1248 for temporarily swapping a port's own read buffer, and the buffer
1249 that the caller of scm_c_read provides. */
1250 struct port_and_swap_buffer
1253 unsigned char *buffer
;
1258 swap_buffer (void *data
)
1260 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1261 unsigned char *old_buf
= psb
->pt
->read_buf
;
1262 size_t old_size
= psb
->pt
->read_buf_size
;
1264 /* Make the port use (buffer, size) from the struct. */
1265 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1266 psb
->pt
->read_buf_size
= psb
->size
;
1268 /* Save the port's old (buffer, size) in the struct. */
1269 psb
->buffer
= old_buf
;
1270 psb
->size
= old_size
;
1274 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1275 #define FUNC_NAME "scm_c_read"
1278 size_t n_read
= 0, n_available
;
1279 struct port_and_swap_buffer psb
;
1281 SCM_VALIDATE_OPINPORT (1, port
);
1283 pt
= SCM_PTAB_ENTRY (port
);
1284 if (pt
->rw_active
== SCM_PORT_WRITE
)
1285 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1288 pt
->rw_active
= SCM_PORT_READ
;
1290 /* Take bytes first from the port's read buffer. */
1291 if (pt
->read_pos
< pt
->read_end
)
1293 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1294 memcpy (buffer
, pt
->read_pos
, n_available
);
1295 buffer
= (char *) buffer
+ n_available
;
1296 pt
->read_pos
+= n_available
;
1297 n_read
+= n_available
;
1298 size
-= n_available
;
1301 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1305 /* Now we will call scm_fill_input repeatedly until we have read the
1306 requested number of bytes. (Note that a single scm_fill_input
1307 call does not guarantee to fill the whole of the port's read
1309 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1311 /* The port that we are reading from is unbuffered - i.e. does
1312 not have its own persistent buffer - but we have a buffer,
1313 provided by our caller, that is the right size for the data
1314 that is wanted. For the following scm_fill_input calls,
1315 therefore, we use the buffer in hand as the port's read
1318 We need to make sure that the port's normal (1 byte) buffer
1319 is reinstated in case one of the scm_fill_input () calls
1320 throws an exception; we use the scm_dynwind_* API to achieve
1323 A consequence of this optimization is that the fill_input
1324 functions can't unget characters. That'll push data to the
1325 pushback buffer instead of this psb buffer. */
1327 unsigned char *pback
= pt
->putback_buf
;
1330 psb
.buffer
= buffer
;
1332 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1333 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1334 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1336 /* Call scm_fill_input until we have all the bytes that we need,
1338 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1340 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1341 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1344 if (pback
!= pt
->putback_buf
1345 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1346 scm_misc_error (FUNC_NAME
,
1347 "scm_c_read must not call a fill function that pushes "
1348 "back characters onto an unbuffered port", SCM_EOL
);
1350 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1352 /* Reinstate the port's normal buffer. */
1357 /* The port has its own buffer. It is important that we use it,
1358 even if it happens to be smaller than our caller's buffer, so
1359 that a custom port implementation's entry points (in
1360 particular, fill_input) can rely on the buffer always being
1361 the same as they first set up. */
1362 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1364 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1365 memcpy (buffer
, pt
->read_pos
, n_available
);
1366 buffer
= (char *) buffer
+ n_available
;
1367 pt
->read_pos
+= n_available
;
1368 n_read
+= n_available
;
1369 size
-= n_available
;
1378 scm_c_read (SCM port
, void *buffer
, size_t size
)
1380 scm_i_pthread_mutex_t
*lock
;
1383 scm_c_lock_port (port
, &lock
);
1384 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1386 scm_i_pthread_mutex_unlock (lock
);
1392 /* Update the line and column number of PORT after consumption of C. */
1394 update_port_lf (scm_t_wchar c
, SCM port
)
1419 #define SCM_MBCHAR_BUF_SIZE (4)
1421 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1422 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1424 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1426 scm_t_wchar codepoint
;
1428 if (utf8_buf
[0] <= 0x7f)
1431 codepoint
= utf8_buf
[0];
1433 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1436 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1437 | (utf8_buf
[1] & 0x3f);
1439 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1442 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1443 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1444 | (utf8_buf
[2] & 0x3f);
1449 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1450 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1451 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1452 | (utf8_buf
[3] & 0x3f);
1458 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1459 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1460 representation, and set *LEN to the length in bytes. Return
1461 `EILSEQ' on error. */
1463 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1464 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1466 #define ASSERT_NOT_EOF(b) \
1467 if (SCM_UNLIKELY ((b) == EOF)) \
1469 #define CONSUME_PEEKED_BYTE() \
1476 pt
= SCM_PTAB_ENTRY (port
);
1478 byte
= scm_get_byte_or_eof_unlocked (port
);
1485 buf
[0] = (scm_t_uint8
) byte
;
1490 *codepoint
= buf
[0];
1491 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1494 byte
= scm_peek_byte_or_eof_unlocked (port
);
1495 ASSERT_NOT_EOF (byte
);
1497 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1500 CONSUME_PEEKED_BYTE ();
1501 buf
[1] = (scm_t_uint8
) byte
;
1504 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1507 else if ((buf
[0] & 0xf0) == 0xe0)
1510 byte
= scm_peek_byte_or_eof_unlocked (port
);
1511 ASSERT_NOT_EOF (byte
);
1513 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1514 || (buf
[0] == 0xe0 && byte
< 0xa0)
1515 || (buf
[0] == 0xed && byte
> 0x9f)))
1518 CONSUME_PEEKED_BYTE ();
1519 buf
[1] = (scm_t_uint8
) byte
;
1522 byte
= scm_peek_byte_or_eof_unlocked (port
);
1523 ASSERT_NOT_EOF (byte
);
1525 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1528 CONSUME_PEEKED_BYTE ();
1529 buf
[2] = (scm_t_uint8
) byte
;
1532 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1533 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1536 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1539 byte
= scm_peek_byte_or_eof_unlocked (port
);
1540 ASSERT_NOT_EOF (byte
);
1542 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1543 || (buf
[0] == 0xf0 && byte
< 0x90)
1544 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1547 CONSUME_PEEKED_BYTE ();
1548 buf
[1] = (scm_t_uint8
) byte
;
1551 byte
= scm_peek_byte_or_eof_unlocked (port
);
1552 ASSERT_NOT_EOF (byte
);
1554 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1557 CONSUME_PEEKED_BYTE ();
1558 buf
[2] = (scm_t_uint8
) byte
;
1561 byte
= scm_peek_byte_or_eof_unlocked (port
);
1562 ASSERT_NOT_EOF (byte
);
1564 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1567 CONSUME_PEEKED_BYTE ();
1568 buf
[3] = (scm_t_uint8
) byte
;
1571 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1572 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1573 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1582 /* Here we could choose the consume the faulty byte when it's not a
1583 valid starting byte, but it's not a requirement. What Section 3.9
1584 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1585 would otherwise be a valid starting byte. */
1589 #undef CONSUME_PEEKED_BYTE
1590 #undef ASSERT_NOT_EOF
1593 /* Likewise, read a byte sequence from PORT, passing it through its
1594 input conversion descriptor. */
1596 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1597 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1601 size_t bytes_consumed
, output_size
;
1603 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1605 pt
= SCM_PTAB_ENTRY (port
);
1607 for (output_size
= 0, output
= (char *) utf8_buf
,
1608 bytes_consumed
= 0, err
= 0;
1609 err
== 0 && output_size
== 0
1610 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1614 size_t input_left
, output_left
, done
;
1616 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1617 if (byte_read
== EOF
)
1619 if (bytes_consumed
== 0)
1621 *codepoint
= (scm_t_wchar
) EOF
;
1629 buf
[bytes_consumed
] = byte_read
;
1632 input_left
= bytes_consumed
+ 1;
1633 output_left
= sizeof (utf8_buf
);
1635 done
= iconv (pt
->input_cd
, &input
, &input_left
,
1636 &output
, &output_left
);
1637 if (done
== (size_t) -1)
1641 /* Missing input: keep trying. */
1645 output_size
= sizeof (utf8_buf
) - output_left
;
1648 if (SCM_UNLIKELY (output_size
== 0))
1649 /* An unterminated sequence. */
1651 else if (SCM_LIKELY (err
== 0))
1653 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1654 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1655 *len
= bytes_consumed
;
1661 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1662 with the byte representation of the codepoint in PORT's encoding, and
1663 set *LEN to the length in bytes of that representation. Return 0 on
1664 success and an errno value on error. */
1666 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1667 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1670 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1672 if (pt
->input_cd
== (iconv_t
) -1)
1673 /* Initialize the conversion descriptors, if needed. */
1674 scm_i_set_port_encoding_x (port
, pt
->encoding
);
1676 /* FIXME: In 2.1, add a flag to determine whether a port is UTF-8. */
1677 if (pt
->input_cd
== (iconv_t
) -1)
1678 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1680 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1682 if (SCM_LIKELY (err
== 0))
1683 update_port_lf (*codepoint
, port
);
1684 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1688 update_port_lf (*codepoint
, port
);
1694 /* Read a codepoint from PORT and return it. */
1696 scm_getc_unlocked (SCM port
)
1697 #define FUNC_NAME "scm_getc"
1701 scm_t_wchar codepoint
;
1702 char buf
[SCM_MBCHAR_BUF_SIZE
];
1704 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1705 if (SCM_UNLIKELY (err
!= 0))
1706 /* At this point PORT should point past the invalid encoding, as per
1707 R6RS-lib Section 8.2.4. */
1708 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1717 scm_i_pthread_mutex_t
*lock
;
1720 scm_c_lock_port (port
, &lock
);
1721 ret
= scm_getc_unlocked (port
);
1723 scm_i_pthread_mutex_unlock (lock
);
1729 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1731 "Return the next character available from @var{port}, updating\n"
1732 "@var{port} to point to the following character. If no more\n"
1733 "characters are available, the end-of-file object is returned.\n"
1735 "When @var{port}'s data cannot be decoded according to its\n"
1736 "character encoding, a @code{decoding-error} is raised and\n"
1737 "@var{port} points past the erroneous byte sequence.\n")
1738 #define FUNC_NAME s_scm_read_char
1741 if (SCM_UNBNDP (port
))
1742 port
= scm_current_input_port ();
1743 SCM_VALIDATE_OPINPORT (1, port
);
1744 c
= scm_getc_unlocked (port
);
1747 return SCM_MAKE_CHAR (c
);
1757 scm_unget_byte_unlocked (int c
, SCM port
)
1758 #define FUNC_NAME "scm_unget_byte"
1760 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1762 if (pt
->read_buf
== pt
->putback_buf
)
1763 /* already using the put-back buffer. */
1765 /* enlarge putback_buf if necessary. */
1766 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1767 && pt
->read_buf
== pt
->read_pos
)
1769 size_t new_size
= pt
->read_buf_size
* 2;
1770 unsigned char *tmp
= (unsigned char *)
1771 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1774 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1775 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1776 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1779 /* shift any existing bytes to buffer + 1. */
1780 if (pt
->read_pos
== pt
->read_end
)
1781 pt
->read_end
= pt
->read_buf
+ 1;
1782 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1784 int count
= pt
->read_end
- pt
->read_pos
;
1786 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1787 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1790 pt
->read_pos
= pt
->read_buf
;
1793 /* switch to the put-back buffer. */
1795 if (pt
->putback_buf
== NULL
)
1798 = (unsigned char *) scm_gc_malloc_pointerless
1799 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1800 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1803 pt
->saved_read_buf
= pt
->read_buf
;
1804 pt
->saved_read_pos
= pt
->read_pos
;
1805 pt
->saved_read_end
= pt
->read_end
;
1806 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1808 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1809 pt
->read_end
= pt
->read_buf
+ 1;
1810 pt
->read_buf_size
= pt
->putback_buf_size
;
1816 pt
->rw_active
= SCM_PORT_READ
;
1821 scm_unget_byte (int c
, SCM port
)
1823 scm_i_pthread_mutex_t
*lock
;
1824 scm_c_lock_port (port
, &lock
);
1825 scm_unget_byte_unlocked (c
, port
);
1827 scm_i_pthread_mutex_unlock (lock
);
1832 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1833 #define FUNC_NAME "scm_ungetc"
1835 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1837 char result_buf
[10];
1838 const char *encoding
;
1842 if (pt
->encoding
!= NULL
)
1843 encoding
= pt
->encoding
;
1845 encoding
= "ISO-8859-1";
1847 len
= sizeof (result_buf
);
1848 result
= u32_conv_to_encoding (encoding
,
1849 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1850 (uint32_t *) &c
, 1, NULL
,
1853 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1854 scm_encoding_error (FUNC_NAME
, errno
,
1855 "conversion to port encoding failed",
1856 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1858 for (i
= len
- 1; i
>= 0; i
--)
1859 scm_unget_byte_unlocked (result
[i
], port
);
1861 if (SCM_UNLIKELY (result
!= result_buf
))
1866 /* What should col be in this case?
1867 * We'll leave it at -1.
1869 SCM_LINUM (port
) -= 1;
1877 scm_ungetc (scm_t_wchar c
, SCM port
)
1879 scm_i_pthread_mutex_t
*lock
;
1880 scm_c_lock_port (port
, &lock
);
1881 scm_ungetc_unlocked (c
, port
);
1883 scm_i_pthread_mutex_unlock (lock
);
1888 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1890 /* This is simple minded and inefficient, but unreading strings is
1891 * probably not a common operation, and remember that line and
1892 * column numbers have to be handled...
1894 * Please feel free to write an optimized version!
1897 scm_ungetc_unlocked (s
[n
], port
);
1901 scm_ungets (const char *s
, int n
, SCM port
)
1903 scm_i_pthread_mutex_t
*lock
;
1904 scm_c_lock_port (port
, &lock
);
1905 scm_ungets_unlocked (s
, n
, port
);
1907 scm_i_pthread_mutex_unlock (lock
);
1911 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1913 "Return the next character available from @var{port},\n"
1914 "@emph{without} updating @var{port} to point to the following\n"
1915 "character. If no more characters are available, the\n"
1916 "end-of-file object is returned.\n"
1918 "The value returned by\n"
1919 "a call to @code{peek-char} is the same as the value that would\n"
1920 "have been returned by a call to @code{read-char} on the same\n"
1921 "port. The only difference is that the very next call to\n"
1922 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1923 "return the value returned by the preceding call to\n"
1924 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1925 "an interactive port will hang waiting for input whenever a call\n"
1926 "to @code{read-char} would have hung.\n"
1928 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1929 "if such a situation occurs. However, unlike with @code{read-char},\n"
1930 "@var{port} still points at the beginning of the erroneous byte\n"
1931 "sequence when the error is raised.\n")
1932 #define FUNC_NAME s_scm_peek_char
1937 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1938 long column
, line
, i
;
1941 if (SCM_UNBNDP (port
))
1942 port
= scm_current_input_port ();
1943 SCM_VALIDATE_OPINPORT (1, port
);
1945 column
= SCM_COL (port
);
1946 line
= SCM_LINUM (port
);
1948 err
= get_codepoint (port
, &c
, bytes
, &len
);
1950 for (i
= len
- 1; i
>= 0; i
--)
1951 scm_unget_byte_unlocked (bytes
[i
], port
);
1953 SCM_COL (port
) = column
;
1954 SCM_LINUM (port
) = line
;
1956 if (SCM_UNLIKELY (err
!= 0))
1958 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1960 /* Shouldn't happen since `catch' always aborts to prompt. */
1961 result
= SCM_BOOL_F
;
1964 result
= SCM_EOF_VAL
;
1966 result
= SCM_MAKE_CHAR (c
);
1972 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
1973 (SCM cobj
, SCM port
),
1974 "Place @var{char} in @var{port} so that it will be read by the\n"
1975 "next read operation. If called multiple times, the unread characters\n"
1976 "will be read again in last-in first-out order. If @var{port} is\n"
1977 "not supplied, the current input port is used.")
1978 #define FUNC_NAME s_scm_unread_char
1982 SCM_VALIDATE_CHAR (1, cobj
);
1983 if (SCM_UNBNDP (port
))
1984 port
= scm_current_input_port ();
1985 SCM_VALIDATE_OPINPORT (2, port
);
1987 c
= SCM_CHAR (cobj
);
1989 scm_ungetc_unlocked (c
, port
);
1994 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
1995 (SCM str
, SCM port
),
1996 "Place the string @var{str} in @var{port} so that its characters will be\n"
1997 "read in subsequent read operations. If called multiple times, the\n"
1998 "unread characters will be read again in last-in first-out order. If\n"
1999 "@var{port} is not supplied, the current-input-port is used.")
2000 #define FUNC_NAME s_scm_unread_string
2003 SCM_VALIDATE_STRING (1, str
);
2004 if (SCM_UNBNDP (port
))
2005 port
= scm_current_input_port ();
2006 SCM_VALIDATE_OPINPORT (2, port
);
2008 n
= scm_i_string_length (str
);
2011 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2020 /* Manipulating the buffers. */
2022 /* This routine does not take any locks, as it is usually called as part
2023 of a port implementation. */
2025 scm_port_non_buffer (scm_t_port
*pt
)
2027 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2028 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2029 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2030 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2033 /* this should only be called when the read buffer is empty. it
2034 tries to refill the read buffer. it returns the first char from
2035 the port, which is either EOF or *(pt->read_pos). */
2037 scm_fill_input_unlocked (SCM port
)
2039 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2041 assert (pt
->read_pos
== pt
->read_end
);
2043 if (pt
->read_buf
== pt
->putback_buf
)
2045 /* finished reading put-back chars. */
2046 pt
->read_buf
= pt
->saved_read_buf
;
2047 pt
->read_pos
= pt
->saved_read_pos
;
2048 pt
->read_end
= pt
->saved_read_end
;
2049 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2050 if (pt
->read_pos
< pt
->read_end
)
2051 return *(pt
->read_pos
);
2053 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2057 scm_fill_input (SCM port
)
2059 scm_i_pthread_mutex_t
*lock
;
2062 scm_c_lock_port (port
, &lock
);
2063 ret
= scm_fill_input_unlocked (port
);
2065 scm_i_pthread_mutex_unlock (lock
);
2071 /* move up to read_len chars from port's putback and/or read buffers
2072 into memory starting at dest. returns the number of chars moved. */
2074 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2076 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2077 size_t chars_read
= 0;
2078 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2082 memcpy (dest
, pt
->read_pos
, from_buf
);
2083 pt
->read_pos
+= from_buf
;
2084 chars_read
+= from_buf
;
2085 read_len
-= from_buf
;
2089 /* if putback was active, try the real input buffer too. */
2090 if (pt
->read_buf
== pt
->putback_buf
)
2092 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2095 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2096 pt
->saved_read_pos
+= from_buf
;
2097 chars_read
+= from_buf
;
2103 /* Clear a port's read buffers, returning the contents. */
2104 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2106 "This procedure clears a port's input buffers, similar\n"
2107 "to the way that force-output clears the output buffer. The\n"
2108 "contents of the buffers are returned as a single string, e.g.,\n"
2111 "(define p (open-input-file ...))\n"
2112 "(drain-input p) => empty string, nothing buffered yet.\n"
2113 "(unread-char (read-char p) p)\n"
2114 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2116 "Draining the buffers may be useful for cleanly finishing\n"
2117 "buffered I/O so that the file descriptor can be used directly\n"
2118 "for further input.")
2119 #define FUNC_NAME s_scm_drain_input
2126 SCM_VALIDATE_OPINPORT (1, port
);
2127 pt
= SCM_PTAB_ENTRY (port
);
2129 count
= pt
->read_end
- pt
->read_pos
;
2130 if (pt
->read_buf
== pt
->putback_buf
)
2131 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2135 result
= scm_i_make_string (count
, &data
, 0);
2136 scm_take_from_input_buffers (port
, data
, count
);
2139 result
= scm_nullstr
;
2146 scm_end_input_unlocked (SCM port
)
2149 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2151 if (pt
->read_buf
== pt
->putback_buf
)
2153 offset
= pt
->read_end
- pt
->read_pos
;
2154 pt
->read_buf
= pt
->saved_read_buf
;
2155 pt
->read_pos
= pt
->saved_read_pos
;
2156 pt
->read_end
= pt
->saved_read_end
;
2157 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2162 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2166 scm_end_input (SCM port
)
2168 scm_i_pthread_mutex_t
*lock
;
2169 scm_c_lock_port (port
, &lock
);
2170 scm_end_input_unlocked (port
);
2172 scm_i_pthread_mutex_unlock (lock
);
2176 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2178 "Flush the specified output port, or the current output port if @var{port}\n"
2179 "is omitted. The current output buffer contents are passed to the\n"
2180 "underlying port implementation (e.g., in the case of fports, the\n"
2181 "data will be written to the file and the output buffer will be cleared.)\n"
2182 "It has no effect on an unbuffered port.\n\n"
2183 "The return value is unspecified.")
2184 #define FUNC_NAME s_scm_force_output
2186 if (SCM_UNBNDP (port
))
2187 port
= scm_current_output_port ();
2190 port
= SCM_COERCE_OUTPORT (port
);
2191 SCM_VALIDATE_OPOUTPORT (1, port
);
2193 scm_flush_unlocked (port
);
2194 return SCM_UNSPECIFIED
;
2199 scm_flush_unlocked (SCM port
)
2201 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2205 scm_flush (SCM port
)
2207 scm_i_pthread_mutex_t
*lock
;
2208 scm_c_lock_port (port
, &lock
);
2209 scm_flush_unlocked (port
);
2211 scm_i_pthread_mutex_unlock (lock
);
2221 scm_putc (char c
, SCM port
)
2223 scm_i_pthread_mutex_t
*lock
;
2224 scm_c_lock_port (port
, &lock
);
2225 scm_putc_unlocked (c
, port
);
2227 scm_i_pthread_mutex_unlock (lock
);
2232 scm_puts (const char *s
, SCM port
)
2234 scm_i_pthread_mutex_t
*lock
;
2235 scm_c_lock_port (port
, &lock
);
2236 scm_puts_unlocked (s
, port
);
2238 scm_i_pthread_mutex_unlock (lock
);
2244 * Used by an application to write arbitrary number of bytes to an SCM
2245 * port. Similar semantics as libc write. However, unlike libc
2246 * write, scm_c_write writes the requested number of bytes and has no
2249 * Warning: Doesn't update port line and column counts!
2252 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2253 #define FUNC_NAME "scm_c_write"
2256 scm_t_ptob_descriptor
*ptob
;
2258 SCM_VALIDATE_OPOUTPORT (1, port
);
2260 pt
= SCM_PTAB_ENTRY (port
);
2261 ptob
= SCM_PORT_DESCRIPTOR (port
);
2263 if (pt
->rw_active
== SCM_PORT_READ
)
2264 scm_end_input_unlocked (port
);
2266 ptob
->write (port
, ptr
, size
);
2269 pt
->rw_active
= SCM_PORT_WRITE
;
2274 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2276 scm_i_pthread_mutex_t
*lock
;
2277 scm_c_lock_port (port
, &lock
);
2278 scm_c_write_unlocked (port
, ptr
, size
);
2280 scm_i_pthread_mutex_unlock (lock
);
2286 * This function differs from scm_c_write; it updates port line and
2289 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2291 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2292 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2294 if (pt
->rw_active
== SCM_PORT_READ
)
2295 scm_end_input_unlocked (port
);
2297 ptob
->write (port
, ptr
, size
);
2299 for (; size
; ptr
++, size
--)
2300 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2303 pt
->rw_active
= SCM_PORT_WRITE
;
2307 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2309 scm_i_pthread_mutex_t
*lock
;
2310 scm_c_lock_port (port
, &lock
);
2311 scm_lfwrite_unlocked (ptr
, size
, port
);
2313 scm_i_pthread_mutex_unlock (lock
);
2317 /* Write STR to PORT from START inclusive to END exclusive. */
2319 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2321 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2323 if (pt
->rw_active
== SCM_PORT_READ
)
2324 scm_end_input_unlocked (port
);
2326 if (end
== (size_t) -1)
2327 end
= scm_i_string_length (str
);
2329 scm_display (scm_c_substring (str
, start
, end
), port
);
2332 pt
->rw_active
= SCM_PORT_WRITE
;
2338 /* Querying and setting positions, and character availability. */
2340 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2342 "Return @code{#t} if a character is ready on input @var{port}\n"
2343 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2344 "@code{#t} then the next @code{read-char} operation on\n"
2345 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2346 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2348 "@code{char-ready?} exists to make it possible for a\n"
2349 "program to accept characters from interactive ports without\n"
2350 "getting stuck waiting for input. Any input editors associated\n"
2351 "with such ports must make sure that characters whose existence\n"
2352 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2353 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2354 "a port at end of file would be indistinguishable from an\n"
2355 "interactive port that has no ready characters.")
2356 #define FUNC_NAME s_scm_char_ready_p
2360 if (SCM_UNBNDP (port
))
2361 port
= scm_current_input_port ();
2362 /* It's possible to close the current input port, so validate even in
2364 SCM_VALIDATE_OPINPORT (1, port
);
2366 pt
= SCM_PTAB_ENTRY (port
);
2368 /* if the current read buffer is filled, or the
2369 last pushed-back char has been read and the saved buffer is
2370 filled, result is true. */
2371 if (pt
->read_pos
< pt
->read_end
2372 || (pt
->read_buf
== pt
->putback_buf
2373 && pt
->saved_read_pos
< pt
->saved_read_end
))
2377 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2379 if (ptob
->input_waiting
)
2380 return scm_from_bool(ptob
->input_waiting (port
));
2387 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2388 (SCM fd_port
, SCM offset
, SCM whence
),
2389 "Sets the current position of @var{fd/port} to the integer\n"
2390 "@var{offset}, which is interpreted according to the value of\n"
2393 "One of the following variables should be supplied for\n"
2395 "@defvar SEEK_SET\n"
2396 "Seek from the beginning of the file.\n"
2398 "@defvar SEEK_CUR\n"
2399 "Seek from the current position.\n"
2401 "@defvar SEEK_END\n"
2402 "Seek from the end of the file.\n"
2404 "If @var{fd/port} is a file descriptor, the underlying system\n"
2405 "call is @code{lseek}. @var{port} may be a string port.\n"
2407 "The value returned is the new position in the file. This means\n"
2408 "that the current position of a port can be obtained using:\n"
2410 "(seek port 0 SEEK_CUR)\n"
2412 #define FUNC_NAME s_scm_seek
2416 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2418 how
= scm_to_int (whence
);
2419 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2420 SCM_OUT_OF_RANGE (3, whence
);
2422 if (SCM_OPPORTP (fd_port
))
2424 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2425 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2426 off_t_or_off64_t rv
;
2429 SCM_MISC_ERROR ("port is not seekable",
2430 scm_cons (fd_port
, SCM_EOL
));
2432 rv
= ptob
->seek (fd_port
, off
, how
);
2433 return scm_from_off_t_or_off64_t (rv
);
2435 else /* file descriptor?. */
2437 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2438 off_t_or_off64_t rv
;
2439 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2442 return scm_from_off_t_or_off64_t (rv
);
2451 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2452 doesn't have the filename version truncate(), hence this code. */
2453 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2455 truncate (const char *file
, off_t length
)
2459 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2463 ret
= ftruncate (fdes
, length
);
2466 int save_errno
= errno
;
2472 return close (fdes
);
2474 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2476 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2477 (SCM object
, SCM length
),
2478 "Truncate @var{file} to @var{length} bytes. @var{file} can be a\n"
2479 "filename string, a port object, or an integer file descriptor.\n"
2480 "The return value is unspecified.\n"
2482 "For a port or file descriptor @var{length} can be omitted, in\n"
2483 "which case the file is truncated at the current position (per\n"
2484 "@code{ftell} above).\n"
2486 "On most systems a file can be extended by giving a length\n"
2487 "greater than the current size, but this is not mandatory in the\n"
2489 #define FUNC_NAME s_scm_truncate_file
2493 /* "object" can be a port, fdes or filename.
2495 Negative "length" makes no sense, but it's left to truncate() or
2496 ftruncate() to give back an error for that (normally EINVAL).
2499 if (SCM_UNBNDP (length
))
2501 /* must supply length if object is a filename. */
2502 if (scm_is_string (object
))
2503 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2505 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2508 object
= SCM_COERCE_OUTPORT (object
);
2509 if (scm_is_integer (object
))
2511 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2512 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2515 else if (SCM_OPOUTPORTP (object
))
2517 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2518 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2519 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2521 if (!ptob
->truncate
)
2522 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2523 if (pt
->rw_active
== SCM_PORT_READ
)
2524 scm_end_input_unlocked (object
);
2525 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2526 ptob
->flush (object
);
2528 ptob
->truncate (object
, c_length
);
2533 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2534 char *str
= scm_to_locale_string (object
);
2536 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2543 return SCM_UNSPECIFIED
;
2547 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2549 "Return the current line number for @var{port}.\n"
2551 "The first line of a file is 0. But you might want to add 1\n"
2552 "when printing line numbers, since starting from 1 is\n"
2553 "traditional in error messages, and likely to be more natural to\n"
2555 #define FUNC_NAME s_scm_port_line
2557 port
= SCM_COERCE_OUTPORT (port
);
2558 SCM_VALIDATE_OPENPORT (1, port
);
2559 return scm_from_long (SCM_LINUM (port
));
2563 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2564 (SCM port
, SCM line
),
2565 "Set the current line number for @var{port} to @var{line}. The\n"
2566 "first line of a file is 0.")
2567 #define FUNC_NAME s_scm_set_port_line_x
2569 port
= SCM_COERCE_OUTPORT (port
);
2570 SCM_VALIDATE_OPENPORT (1, port
);
2571 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2572 return SCM_UNSPECIFIED
;
2576 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2578 "Return the current column number of @var{port}.\n"
2579 "If the number is\n"
2580 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2581 "- i.e. the first character of the first line is line 0, column 0.\n"
2582 "(However, when you display a file position, for example in an error\n"
2583 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2584 "because lines and column numbers traditionally start with 1, and that is\n"
2585 "what non-programmers will find most natural.)")
2586 #define FUNC_NAME s_scm_port_column
2588 port
= SCM_COERCE_OUTPORT (port
);
2589 SCM_VALIDATE_OPENPORT (1, port
);
2590 return scm_from_int (SCM_COL (port
));
2594 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2595 (SCM port
, SCM column
),
2596 "Set the current column of @var{port}. Before reading the first\n"
2597 "character on a line the column should be 0.")
2598 #define FUNC_NAME s_scm_set_port_column_x
2600 port
= SCM_COERCE_OUTPORT (port
);
2601 SCM_VALIDATE_OPENPORT (1, port
);
2602 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2603 return SCM_UNSPECIFIED
;
2607 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2609 "Return the filename associated with @var{port}, or @code{#f}\n"
2610 "if no filename is associated with the port.")
2611 #define FUNC_NAME s_scm_port_filename
2613 port
= SCM_COERCE_OUTPORT (port
);
2614 SCM_VALIDATE_OPENPORT (1, port
);
2615 return SCM_FILENAME (port
);
2619 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2620 (SCM port
, SCM filename
),
2621 "Change the filename associated with @var{port}, using the current input\n"
2622 "port if none is specified. Note that this does not change the port's\n"
2623 "source of data, but only the value that is returned by\n"
2624 "@code{port-filename} and reported in diagnostic output.")
2625 #define FUNC_NAME s_scm_set_port_filename_x
2627 port
= SCM_COERCE_OUTPORT (port
);
2628 SCM_VALIDATE_OPENPORT (1, port
);
2629 /* We allow the user to set the filename to whatever he likes. */
2630 SCM_SET_FILENAME (port
, filename
);
2631 return SCM_UNSPECIFIED
;
2638 /* Implementation helpers for port printing functions. */
2641 scm_print_port_mode (SCM exp
, SCM port
)
2643 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2645 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2646 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2649 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2656 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2658 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2661 scm_puts_unlocked ("#<", port
);
2662 scm_print_port_mode (exp
, port
);
2663 scm_puts_unlocked (type
, port
);
2664 scm_putc_unlocked (' ', port
);
2665 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2666 scm_putc_unlocked ('>', port
);
2673 /* Iterating over all ports. */
2675 struct for_each_data
2677 void (*proc
) (void *data
, SCM p
);
2682 for_each_trampoline (void *data
, SCM port
, SCM result
)
2684 struct for_each_data
*d
= data
;
2686 d
->proc (d
->data
, port
);
2692 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2694 struct for_each_data d
;
2699 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2700 scm_i_port_weak_set
);
2704 scm_for_each_trampoline (void *data
, SCM port
)
2706 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2709 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2711 "Apply @var{proc} to each port in the Guile port table\n"
2712 "in turn. The return value is unspecified. More specifically,\n"
2713 "@var{proc} is applied exactly once to every port that exists\n"
2714 "in the system at the time @var{port-for-each} is invoked.\n"
2715 "Changes to the port table while @var{port-for-each} is running\n"
2716 "have no effect as far as @var{port-for-each} is concerned.")
2717 #define FUNC_NAME s_scm_port_for_each
2719 SCM_VALIDATE_PROC (1, proc
);
2721 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2723 return SCM_UNSPECIFIED
;
2728 flush_output_port (void *closure
, SCM port
)
2730 if (SCM_OPOUTPORTP (port
))
2731 scm_flush_unlocked (port
);
2734 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2736 "Equivalent to calling @code{force-output} on\n"
2737 "all open output ports. The return value is unspecified.")
2738 #define FUNC_NAME s_scm_flush_all_ports
2740 scm_c_port_for_each (&flush_output_port
, NULL
);
2741 return SCM_UNSPECIFIED
;
2750 scm_t_bits scm_tc16_void_port
= 0;
2752 static int fill_input_void_port (SCM port SCM_UNUSED
)
2758 write_void_port (SCM port SCM_UNUSED
,
2759 const void *data SCM_UNUSED
,
2760 size_t size SCM_UNUSED
)
2765 scm_i_void_port (long mode_bits
)
2769 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2771 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2777 scm_void_port (char *mode_str
)
2779 return scm_i_void_port (scm_mode_bits (mode_str
));
2782 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2784 "Create and return a new void port. A void port acts like\n"
2785 "@file{/dev/null}. The @var{mode} argument\n"
2786 "specifies the input/output modes for this port: see the\n"
2787 "documentation for @code{open-file} in @ref{File Ports}.")
2788 #define FUNC_NAME s_scm_sys_make_void_port
2790 return scm_i_void_port (scm_i_mode_bits (mode
));
2797 /* Initialization. */
2802 /* lseek() symbols. */
2803 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2804 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2805 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2807 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2810 cur_inport_fluid
= scm_make_fluid ();
2811 cur_outport_fluid
= scm_make_fluid ();
2812 cur_errport_fluid
= scm_make_fluid ();
2813 cur_loadport_fluid
= scm_make_fluid ();
2815 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2817 #include "libguile/ports.x"
2819 /* Use Latin-1 as the default port encoding. */
2820 SCM_VARIABLE_SET (default_port_encoding_var
,
2821 scm_make_fluid_with_default (SCM_BOOL_F
));
2822 scm_port_encoding_init
= 1;
2824 SCM_VARIABLE_SET (scm_conversion_strategy
,
2825 scm_make_fluid_with_default
2826 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2827 scm_conversion_strategy_init
= 1;