1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013 Free Software Foundation, Inc.
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 3 of
7 * the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
32 #include <fcntl.h> /* for chsize on mingw */
37 #include <striconveh.h>
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 /* Port encodings are case-insensitive ASCII strings. */
95 ascii_toupper (char c
)
97 return (c
< 'a' || c
> 'z') ? c
: ('A' + (c
- 'a'));
100 /* It is only necessary to use this function on encodings that come from
101 the user and have not been canonicalized yet. Encodings that are set
102 on ports or in the default encoding fluid are in upper-case, and can
103 be compared with strcmp. */
105 encoding_matches (const char *enc
, const char *upper
)
111 if (ascii_toupper (*enc
++) != *upper
++)
118 canonicalize_encoding (const char *enc
)
126 ret
= scm_gc_strdup (enc
, "port");
128 for (i
= 0; ret
[i
]; i
++)
131 /* Restrict to ASCII. */
132 scm_misc_error (NULL
, "invalid character encoding ~s",
133 scm_list_1 (scm_from_latin1_string (enc
)));
135 ret
[i
] = ascii_toupper (ret
[i
]);
143 /* The port kind table --- a dynamically resized array of port types. */
146 /* scm_ptobs scm_numptob
147 * implement a dynamically resized array of ptob records.
148 * Indexes into this table are used when generating type
149 * tags for smobjects (if you know a tag you can get an index and conversely).
151 static scm_t_ptob_descriptor
**scm_ptobs
= NULL
;
152 static long scm_numptob
= 0; /* Number of port types. */
153 static long scm_ptobs_size
= 0; /* Number of slots in the port type
155 static scm_i_pthread_mutex_t scm_ptobs_lock
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
158 scm_c_num_port_types (void)
162 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
164 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
169 scm_t_ptob_descriptor
*
170 scm_c_port_type_ref (long ptobnum
)
172 scm_t_ptob_descriptor
*ret
= NULL
;
174 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
176 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
177 ret
= scm_ptobs
[ptobnum
];
179 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
182 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
188 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
192 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
194 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
196 if (scm_numptob
== scm_ptobs_size
)
198 unsigned long old_size
= scm_ptobs_size
;
199 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
201 /* Currently there are only 9 predefined port types, so one
202 resize will cover it. */
203 scm_ptobs_size
= old_size
+ 10;
205 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
206 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
208 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
211 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
215 scm_ptobs
[ret
] = desc
;
218 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
221 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
227 * We choose to use an interface similar to the smob interface with
228 * fill_input and write as standard fields, passed to the port
229 * type constructor, and optional fields set by setters.
233 flush_port_default (SCM port SCM_UNUSED
)
238 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
243 scm_make_port_type (char *name
,
244 int (*fill_input
) (SCM port
),
245 void (*write
) (SCM port
, const void *data
, size_t size
))
247 scm_t_ptob_descriptor
*desc
;
250 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
251 memset (desc
, 0, sizeof (*desc
));
254 desc
->print
= scm_port_print
;
256 desc
->flush
= flush_port_default
;
257 desc
->end_input
= end_input_default
;
258 desc
->fill_input
= fill_input
;
260 ptobnum
= scm_c_port_type_add_x (desc
);
262 /* Make a class object if GOOPS is present. */
263 if (SCM_UNPACK (scm_port_class
[0]) != 0)
264 scm_make_port_classes (ptobnum
, name
);
266 return scm_tc7_port
+ ptobnum
* 256;
270 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
272 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
276 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
278 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
282 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
283 scm_print_state
*pstate
))
285 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
289 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
291 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
295 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
297 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
301 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
303 scm_t_ptob_descriptor
*ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
));
305 ptob
->flags
|= SCM_PORT_TYPE_HAS_FLUSH
;
309 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
311 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
315 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
317 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
321 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
323 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
327 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
329 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
334 /* Standard ports --- current input, output, error, and more(!). */
336 static SCM cur_inport_fluid
= SCM_BOOL_F
;
337 static SCM cur_outport_fluid
= SCM_BOOL_F
;
338 static SCM cur_errport_fluid
= SCM_BOOL_F
;
339 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
341 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
343 "Return the current input port. This is the default port used\n"
344 "by many input procedures. Initially, @code{current-input-port}\n"
345 "returns the @dfn{standard input} in Unix and C terminology.")
346 #define FUNC_NAME s_scm_current_input_port
348 if (scm_is_true (cur_inport_fluid
))
349 return scm_fluid_ref (cur_inport_fluid
);
355 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
357 "Return the current output port. This is the default port used\n"
358 "by many output procedures. Initially,\n"
359 "@code{current-output-port} returns the @dfn{standard output} in\n"
360 "Unix and C terminology.")
361 #define FUNC_NAME s_scm_current_output_port
363 if (scm_is_true (cur_outport_fluid
))
364 return scm_fluid_ref (cur_outport_fluid
);
370 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
372 "Return the port to which errors and warnings should be sent (the\n"
373 "@dfn{standard error} in Unix and C terminology).")
374 #define FUNC_NAME s_scm_current_error_port
376 if (scm_is_true (cur_errport_fluid
))
377 return scm_fluid_ref (cur_errport_fluid
);
384 scm_current_warning_port (void)
386 static SCM cwp_var
= SCM_BOOL_F
;
388 if (scm_is_false (cwp_var
))
389 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
391 return scm_call_0 (scm_variable_ref (cwp_var
));
394 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
396 "Return the current-load-port.\n"
397 "The load port is used internally by @code{primitive-load}.")
398 #define FUNC_NAME s_scm_current_load_port
400 return scm_fluid_ref (cur_loadport_fluid
);
404 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
406 "@deffnx {Scheme Procedure} set-current-output-port port\n"
407 "@deffnx {Scheme Procedure} set-current-error-port port\n"
408 "Change the ports returned by @code{current-input-port},\n"
409 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
410 "so that they use the supplied @var{port} for input or output.")
411 #define FUNC_NAME s_scm_set_current_input_port
413 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
414 SCM_VALIDATE_OPINPORT (1, port
);
415 scm_fluid_set_x (cur_inport_fluid
, port
);
421 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
423 "Set the current default output port to @var{port}.")
424 #define FUNC_NAME s_scm_set_current_output_port
426 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
427 port
= SCM_COERCE_OUTPORT (port
);
428 SCM_VALIDATE_OPOUTPORT (1, port
);
429 scm_fluid_set_x (cur_outport_fluid
, port
);
435 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
437 "Set the current default error port to @var{port}.")
438 #define FUNC_NAME s_scm_set_current_error_port
440 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
441 port
= SCM_COERCE_OUTPORT (port
);
442 SCM_VALIDATE_OPOUTPORT (1, port
);
443 scm_fluid_set_x (cur_errport_fluid
, port
);
450 scm_set_current_warning_port (SCM port
)
452 static SCM cwp_var
= SCM_BOOL_F
;
454 if (scm_is_false (cwp_var
))
455 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
457 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
462 scm_dynwind_current_input_port (SCM port
)
463 #define FUNC_NAME NULL
465 SCM_VALIDATE_OPINPORT (1, port
);
466 scm_dynwind_fluid (cur_inport_fluid
, port
);
471 scm_dynwind_current_output_port (SCM port
)
472 #define FUNC_NAME NULL
474 port
= SCM_COERCE_OUTPORT (port
);
475 SCM_VALIDATE_OPOUTPORT (1, port
);
476 scm_dynwind_fluid (cur_outport_fluid
, port
);
481 scm_dynwind_current_error_port (SCM port
)
482 #define FUNC_NAME NULL
484 port
= SCM_COERCE_OUTPORT (port
);
485 SCM_VALIDATE_OPOUTPORT (1, port
);
486 scm_dynwind_fluid (cur_errport_fluid
, port
);
491 scm_i_dynwind_current_load_port (SCM port
)
493 scm_dynwind_fluid (cur_loadport_fluid
, port
);
499 /* Retrieving a port's mode. */
501 /* Return the flags that characterize a port based on the mode
502 * string used to open a file for that port.
504 * See PORT FLAGS in scm.h
508 scm_i_mode_bits_n (SCM modes
)
511 | (scm_i_string_contains_char (modes
, 'r')
512 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
513 | (scm_i_string_contains_char (modes
, 'w')
514 || scm_i_string_contains_char (modes
, 'a')
515 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
516 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
517 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
521 scm_mode_bits (char *modes
)
523 /* Valid characters are rw+a0l. So, use latin1. */
524 return scm_i_mode_bits (scm_from_latin1_string (modes
));
528 scm_i_mode_bits (SCM modes
)
532 if (!scm_is_string (modes
))
533 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
535 bits
= scm_i_mode_bits_n (modes
);
536 scm_remember_upto_here_1 (modes
);
540 /* Return the mode flags from an open port.
541 * Some modes such as "append" are only used when opening
542 * a file and are not returned here. */
544 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
546 "Return the port modes associated with the open port @var{port}.\n"
547 "These will not necessarily be identical to the modes used when\n"
548 "the port was opened, since modes such as \"append\" which are\n"
549 "used only during port creation are not retained.")
550 #define FUNC_NAME s_scm_port_mode
555 port
= SCM_COERCE_OUTPORT (port
);
556 SCM_VALIDATE_OPPORT (1, port
);
557 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
558 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
559 strcpy (modes
, "r+");
563 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
565 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
568 return scm_from_latin1_string (modes
);
575 /* The port table --- a weak set of all ports.
577 We need a global registry of ports to flush them all at exit, and to
578 get all the ports matching a file descriptor. */
579 SCM scm_i_port_weak_set
;
584 /* Port finalization. */
588 scm_t_ptob_descriptor
*ptob
;
593 do_free (void *body_data
)
595 struct do_free_data
*data
= body_data
;
597 /* `close' is for explicit `close-port' by user. `free' is for this
598 purpose: ports collected by the GC. */
599 data
->ptob
->free (data
->port
);
604 /* Finalize the object (a port) pointed to by PTR. */
606 finalize_port (void *ptr
, void *data
)
608 SCM port
= SCM_PACK_POINTER (ptr
);
610 if (!SCM_PORTP (port
))
613 if (SCM_OPENP (port
))
615 struct do_free_data data
;
617 SCM_CLR_PORT_OPEN_FLAG (port
);
619 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
622 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
623 scm_handle_by_message_noexit
, NULL
);
625 scm_gc_ports_collected
++;
633 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
634 const char *encoding
,
635 scm_t_string_failed_conversion_handler handler
,
640 scm_t_ptob_descriptor
*ptob
;
642 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
643 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
645 ret
= scm_words (tag
| mode_bits
, 3);
646 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
647 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
649 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
650 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
652 entry
->file_name
= SCM_BOOL_F
;
653 entry
->rw_active
= SCM_PORT_NEITHER
;
655 entry
->stream
= stream
;
657 if (encoding_matches (encoding
, "UTF-8"))
659 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
660 entry
->encoding
= "UTF-8";
662 else if (encoding_matches (encoding
, "ISO-8859-1"))
664 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
665 entry
->encoding
= "ISO-8859-1";
669 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
670 entry
->encoding
= canonicalize_encoding (encoding
);
673 entry
->ilseq_handler
= handler
;
674 entry
->iconv_descriptors
= NULL
;
676 entry
->alist
= SCM_EOL
;
678 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
679 scm_i_set_finalizer (SCM2PTR (ret
), finalize_port
, NULL
);
681 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
682 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
688 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
690 return scm_c_make_port_with_encoding (tag
, mode_bits
,
691 scm_i_default_port_encoding (),
692 scm_i_default_port_conversion_handler (),
697 scm_new_port_table_entry (scm_t_bits tag
)
699 return scm_c_make_port (tag
, 0, 0);
706 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
708 "Return a boolean indicating whether @var{x} is a port.\n"
709 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
711 #define FUNC_NAME s_scm_port_p
713 return scm_from_bool (SCM_PORTP (x
));
717 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
719 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
720 "@code{#f}. Any object satisfying this predicate also satisfies\n"
722 #define FUNC_NAME s_scm_input_port_p
724 return scm_from_bool (SCM_INPUT_PORT_P (x
));
728 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
730 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
731 "@code{#f}. Any object satisfying this predicate also satisfies\n"
733 #define FUNC_NAME s_scm_output_port_p
735 x
= SCM_COERCE_OUTPORT (x
);
736 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
740 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
742 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
744 #define FUNC_NAME s_scm_port_closed_p
746 SCM_VALIDATE_PORT (1, port
);
747 return scm_from_bool (!SCM_OPPORTP (port
));
751 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
753 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
755 #define FUNC_NAME s_scm_eof_object_p
757 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
766 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
769 * Call the close operation on a port object.
770 * see also scm_close.
772 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
774 "Close the specified port object. Return @code{#t} if it\n"
775 "successfully closes a port or @code{#f} if it was already\n"
776 "closed. An exception may be raised if an error occurs, for\n"
777 "example when flushing buffered output. See also @ref{Ports and\n"
778 "File Descriptors, close}, for a procedure which can close file\n"
780 #define FUNC_NAME s_scm_close_port
785 port
= SCM_COERCE_OUTPORT (port
);
787 SCM_VALIDATE_PORT (1, port
);
788 if (SCM_CLOSEDP (port
))
791 p
= SCM_PTAB_ENTRY (port
);
792 SCM_CLR_PORT_OPEN_FLAG (port
);
794 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
795 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
797 if (SCM_PORT_DESCRIPTOR (port
)->close
)
798 /* Note! This may throw an exception. Anything after this point
799 should be resilient to non-local exits. */
800 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
804 if (p
->iconv_descriptors
)
806 /* If we don't get here, the iconv_descriptors finalizer will
808 close_iconv_descriptors (p
->iconv_descriptors
);
809 p
->iconv_descriptors
= NULL
;
812 return scm_from_bool (rv
>= 0);
816 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
818 "Close the specified input port object. The routine has no effect if\n"
819 "the file has already been closed. An exception may be raised if an\n"
820 "error occurs. The value returned is unspecified.\n\n"
821 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
822 "which can close file descriptors.")
823 #define FUNC_NAME s_scm_close_input_port
825 SCM_VALIDATE_INPUT_PORT (1, port
);
826 scm_close_port (port
);
827 return SCM_UNSPECIFIED
;
831 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
833 "Close the specified output port object. The routine has no effect if\n"
834 "the file has already been closed. An exception may be raised if an\n"
835 "error occurs. The value returned is unspecified.\n\n"
836 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
837 "which can close file descriptors.")
838 #define FUNC_NAME s_scm_close_output_port
840 port
= SCM_COERCE_OUTPORT (port
);
841 SCM_VALIDATE_OUTPUT_PORT (1, port
);
842 scm_close_port (port
);
843 return SCM_UNSPECIFIED
;
850 /* Encoding characters to byte streams, and decoding byte streams to
853 /* A fluid specifying the default encoding for newly created ports. If it is
854 a string, that is the encoding. If it is #f, it is in the "native"
855 (Latin-1) encoding. */
856 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
858 static int scm_port_encoding_init
= 0;
860 /* Use ENCODING as the default encoding for future ports. */
862 scm_i_set_default_port_encoding (const char *encoding
)
864 if (!scm_port_encoding_init
865 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
866 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
869 if (encoding_matches (encoding
, "ASCII")
870 || encoding_matches (encoding
, "ANSI_X3.4-1968")
871 || encoding_matches (encoding
, "ISO-8859-1"))
872 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
874 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
875 scm_from_latin1_string (canonicalize_encoding (encoding
)));
878 /* Return the name of the default encoding for newly created ports. */
880 scm_i_default_port_encoding (void)
882 if (!scm_port_encoding_init
)
884 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
890 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
891 if (!scm_is_string (encoding
))
894 return scm_i_string_chars (encoding
);
898 /* A fluid specifying the default conversion handler for newly created
899 ports. Its value should be one of the symbols below. */
900 SCM_VARIABLE (default_conversion_strategy_var
,
901 "%default-port-conversion-strategy");
903 /* Whether the above fluid is initialized. */
904 static int scm_conversion_strategy_init
= 0;
906 /* The possible conversion strategies. */
907 SCM_SYMBOL (sym_error
, "error");
908 SCM_SYMBOL (sym_substitute
, "substitute");
909 SCM_SYMBOL (sym_escape
, "escape");
911 /* Return the default failed encoding conversion policy for new created
913 scm_t_string_failed_conversion_handler
914 scm_i_default_port_conversion_handler (void)
916 scm_t_string_failed_conversion_handler handler
;
918 if (!scm_conversion_strategy_init
919 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
920 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
925 fluid
= SCM_VARIABLE_REF (default_conversion_strategy_var
);
926 value
= scm_fluid_ref (fluid
);
928 if (scm_is_eq (sym_substitute
, value
))
929 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
930 else if (scm_is_eq (sym_escape
, value
))
931 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
933 /* Default to 'error also when the fluid's value is not one of
934 the valid symbols. */
935 handler
= SCM_FAILED_CONVERSION_ERROR
;
941 /* Use HANDLER as the default conversion strategy for future ports. */
943 scm_i_set_default_port_conversion_handler (scm_t_string_failed_conversion_handler
948 if (!scm_conversion_strategy_init
949 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
950 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
955 case SCM_FAILED_CONVERSION_ERROR
:
956 strategy
= sym_error
;
959 case SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
:
960 strategy
= sym_escape
;
963 case SCM_FAILED_CONVERSION_QUESTION_MARK
:
964 strategy
= sym_substitute
;
971 scm_fluid_set_x (SCM_VARIABLE_REF (default_conversion_strategy_var
),
976 finalize_iconv_descriptors (void *ptr
, void *data
)
978 close_iconv_descriptors (ptr
);
981 static scm_t_iconv_descriptors
*
982 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
984 scm_t_iconv_descriptors
*id
;
985 iconv_t input_cd
, output_cd
;
988 input_cd
= (iconv_t
) -1;
989 output_cd
= (iconv_t
) -1;
991 for (i
= 0; encoding
[i
]; i
++)
992 if (encoding
[i
] > 127)
993 goto invalid_encoding
;
997 /* Open an input iconv conversion descriptor, from ENCODING
998 to UTF-8. We choose UTF-8, not UTF-32, because iconv
999 implementations can typically convert from anything to
1000 UTF-8, but not to UTF-32 (see
1001 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
1003 /* Assume opening an iconv descriptor causes about 16 KB of
1005 scm_gc_register_allocation (16 * 1024);
1007 input_cd
= iconv_open ("UTF-8", encoding
);
1008 if (input_cd
== (iconv_t
) -1)
1009 goto invalid_encoding
;
1014 /* Assume opening an iconv descriptor causes about 16 KB of
1016 scm_gc_register_allocation (16 * 1024);
1018 output_cd
= iconv_open (encoding
, "UTF-8");
1019 if (output_cd
== (iconv_t
) -1)
1021 if (input_cd
!= (iconv_t
) -1)
1022 iconv_close (input_cd
);
1023 goto invalid_encoding
;
1027 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
1028 id
->input_cd
= input_cd
;
1029 id
->output_cd
= output_cd
;
1031 /* Register a finalizer to close the descriptors. */
1032 scm_i_set_finalizer (id
, finalize_iconv_descriptors
, NULL
);
1039 err
= scm_from_latin1_string (encoding
);
1040 scm_misc_error ("open_iconv_descriptors",
1041 "invalid or unknown character encoding ~s",
1047 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
1049 if (id
->input_cd
!= (iconv_t
) -1)
1050 iconv_close (id
->input_cd
);
1051 if (id
->output_cd
!= (iconv_t
) -1)
1052 iconv_close (id
->output_cd
);
1053 id
->input_cd
= (void *) -1;
1054 id
->output_cd
= (void *) -1;
1057 scm_t_iconv_descriptors
*
1058 scm_i_port_iconv_descriptors (SCM port
)
1062 pt
= SCM_PTAB_ENTRY (port
);
1064 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
1066 if (!pt
->iconv_descriptors
)
1069 pt
->encoding
= "ISO-8859-1";
1070 pt
->iconv_descriptors
=
1071 open_iconv_descriptors (pt
->encoding
,
1072 SCM_INPUT_PORT_P (port
),
1073 SCM_OUTPUT_PORT_P (port
));
1076 return pt
->iconv_descriptors
;
1079 /* The name of the encoding is itself encoded in ASCII. */
1081 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
1084 scm_t_iconv_descriptors
*prev
;
1086 /* Set the character encoding for this port. */
1087 pt
= SCM_PTAB_ENTRY (port
);
1088 prev
= pt
->iconv_descriptors
;
1090 if (encoding_matches (encoding
, "UTF-8"))
1092 pt
->encoding
= "UTF-8";
1093 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
1094 pt
->iconv_descriptors
= NULL
;
1096 else if (encoding_matches (encoding
, "ISO-8859-1"))
1098 pt
->encoding
= "ISO-8859-1";
1099 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
1100 pt
->iconv_descriptors
= NULL
;
1104 /* Open descriptors before mutating the port. */
1105 char *gc_encoding
= canonicalize_encoding (encoding
);
1106 pt
->iconv_descriptors
=
1107 open_iconv_descriptors (gc_encoding
,
1108 SCM_INPUT_PORT_P (port
),
1109 SCM_OUTPUT_PORT_P (port
));
1110 pt
->encoding
= gc_encoding
;
1111 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
1115 close_iconv_descriptors (prev
);
1118 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
1120 "Returns, as a string, the character encoding that @var{port}\n"
1121 "uses to interpret its input and output.\n")
1122 #define FUNC_NAME s_scm_port_encoding
1124 SCM_VALIDATE_PORT (1, port
);
1126 return scm_from_latin1_string (SCM_PTAB_ENTRY (port
)->encoding
);
1130 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1131 (SCM port
, SCM enc
),
1132 "Sets the character encoding that will be used to interpret all\n"
1133 "port I/O. New ports are created with the encoding\n"
1134 "appropriate for the current locale if @code{setlocale} has \n"
1135 "been called or ISO-8859-1 otherwise\n"
1136 "and this procedure can be used to modify that encoding.\n")
1137 #define FUNC_NAME s_scm_set_port_encoding_x
1141 SCM_VALIDATE_PORT (1, port
);
1142 SCM_VALIDATE_STRING (2, enc
);
1144 enc_str
= scm_to_latin1_string (enc
);
1145 scm_i_set_port_encoding_x (port
, enc_str
);
1148 return SCM_UNSPECIFIED
;
1152 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1153 1, 0, 0, (SCM port
),
1154 "Returns the behavior of the port when handling a character that\n"
1155 "is not representable in the port's current encoding.\n"
1156 "It returns the symbol @code{error} if unrepresentable characters\n"
1157 "should cause exceptions, @code{substitute} if the port should\n"
1158 "try to replace unrepresentable characters with question marks or\n"
1159 "approximate characters, or @code{escape} if unrepresentable\n"
1160 "characters should be converted to string escapes.\n"
1162 "If @var{port} is @code{#f}, then the current default behavior\n"
1163 "will be returned. New ports will have this default behavior\n"
1164 "when they are created.\n")
1165 #define FUNC_NAME s_scm_port_conversion_strategy
1167 scm_t_string_failed_conversion_handler h
;
1169 SCM_VALIDATE_OPPORT (1, port
);
1171 if (scm_is_false (port
))
1172 h
= scm_i_default_port_conversion_handler ();
1177 SCM_VALIDATE_OPPORT (1, port
);
1178 pt
= SCM_PTAB_ENTRY (port
);
1180 h
= pt
->ilseq_handler
;
1183 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1184 return scm_from_latin1_symbol ("error");
1185 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1186 return scm_from_latin1_symbol ("substitute");
1187 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1188 return scm_from_latin1_symbol ("escape");
1192 /* Never gets here. */
1193 return SCM_UNDEFINED
;
1197 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1199 (SCM port
, SCM sym
),
1200 "Sets the behavior of the interpreter when outputting a character\n"
1201 "that is not representable in the port's current encoding.\n"
1202 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1203 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1204 "when an unconvertible character is encountered. If it is\n"
1205 "@code{'substitute}, then unconvertible characters will \n"
1206 "be replaced with approximate characters, or with question marks\n"
1207 "if no approximately correct character is available.\n"
1208 "If it is @code{'escape},\n"
1209 "it will appear as a hex escape when output.\n"
1211 "If @var{port} is an open port, the conversion error behavior\n"
1212 "is set for that port. If it is @code{#f}, it is set as the\n"
1213 "default behavior for any future ports that get created in\n"
1215 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1217 scm_t_string_failed_conversion_handler handler
;
1219 if (scm_is_eq (sym
, sym_error
))
1220 handler
= SCM_FAILED_CONVERSION_ERROR
;
1221 else if (scm_is_eq (sym
, sym_substitute
))
1222 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
1223 else if (scm_is_eq (sym
, sym_escape
))
1224 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
1226 SCM_MISC_ERROR ("unknown conversion strategy ~s", scm_list_1 (sym
));
1228 if (scm_is_false (port
))
1229 scm_i_set_default_port_conversion_handler (handler
);
1232 SCM_VALIDATE_OPPORT (1, port
);
1233 SCM_PTAB_ENTRY (port
)->ilseq_handler
= handler
;
1236 return SCM_UNSPECIFIED
;
1243 /* The port lock. */
1246 lock_port (void *mutex
)
1248 scm_i_pthread_mutex_lock (mutex
);
1252 unlock_port (void *mutex
)
1254 scm_i_pthread_mutex_unlock (mutex
);
1258 scm_dynwind_lock_port (SCM port
)
1259 #define FUNC_NAME "dynwind-lock-port"
1261 scm_i_pthread_mutex_t
*lock
;
1262 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1263 scm_c_lock_port (port
, &lock
);
1266 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1267 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1278 scm_get_byte_or_eof (SCM port
)
1280 scm_i_pthread_mutex_t
*lock
;
1283 scm_c_lock_port (port
, &lock
);
1284 ret
= scm_get_byte_or_eof_unlocked (port
);
1286 scm_i_pthread_mutex_unlock (lock
);
1292 scm_peek_byte_or_eof (SCM port
)
1294 scm_i_pthread_mutex_t
*lock
;
1297 scm_c_lock_port (port
, &lock
);
1298 ret
= scm_peek_byte_or_eof_unlocked (port
);
1300 scm_i_pthread_mutex_unlock (lock
);
1307 * Used by an application to read arbitrary number of bytes from an
1308 * SCM port. Same semantics as libc read, except that scm_c_read only
1309 * returns less than SIZE bytes if at end-of-file.
1311 * Warning: Doesn't update port line and column counts! */
1313 /* This structure, and the following swap_buffer function, are used
1314 for temporarily swapping a port's own read buffer, and the buffer
1315 that the caller of scm_c_read provides. */
1316 struct port_and_swap_buffer
1319 unsigned char *buffer
;
1324 swap_buffer (void *data
)
1326 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1327 unsigned char *old_buf
= psb
->pt
->read_buf
;
1328 size_t old_size
= psb
->pt
->read_buf_size
;
1330 /* Make the port use (buffer, size) from the struct. */
1331 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1332 psb
->pt
->read_buf_size
= psb
->size
;
1334 /* Save the port's old (buffer, size) in the struct. */
1335 psb
->buffer
= old_buf
;
1336 psb
->size
= old_size
;
1340 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1341 #define FUNC_NAME "scm_c_read"
1344 size_t n_read
= 0, n_available
;
1345 struct port_and_swap_buffer psb
;
1347 SCM_VALIDATE_OPINPORT (1, port
);
1349 pt
= SCM_PTAB_ENTRY (port
);
1350 if (pt
->rw_active
== SCM_PORT_WRITE
)
1351 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1354 pt
->rw_active
= SCM_PORT_READ
;
1356 /* Take bytes first from the port's read buffer. */
1357 if (pt
->read_pos
< pt
->read_end
)
1359 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1360 memcpy (buffer
, pt
->read_pos
, n_available
);
1361 buffer
= (char *) buffer
+ n_available
;
1362 pt
->read_pos
+= n_available
;
1363 n_read
+= n_available
;
1364 size
-= n_available
;
1367 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1371 /* Now we will call scm_fill_input repeatedly until we have read the
1372 requested number of bytes. (Note that a single scm_fill_input
1373 call does not guarantee to fill the whole of the port's read
1375 if (pt
->read_buf_size
<= 1
1376 && pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_LATIN1
)
1378 /* The port that we are reading from is unbuffered - i.e. does
1379 not have its own persistent buffer - but we have a buffer,
1380 provided by our caller, that is the right size for the data
1381 that is wanted. For the following scm_fill_input calls,
1382 therefore, we use the buffer in hand as the port's read
1385 We need to make sure that the port's normal (1 byte) buffer
1386 is reinstated in case one of the scm_fill_input () calls
1387 throws an exception; we use the scm_dynwind_* API to achieve
1390 A consequence of this optimization is that the fill_input
1391 functions can't unget characters. That'll push data to the
1392 pushback buffer instead of this psb buffer. */
1394 unsigned char *pback
= pt
->putback_buf
;
1397 psb
.buffer
= buffer
;
1399 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1400 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1401 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1403 /* Call scm_fill_input until we have all the bytes that we need,
1405 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1407 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1408 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1411 if (pback
!= pt
->putback_buf
1412 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1413 scm_misc_error (FUNC_NAME
,
1414 "scm_c_read must not call a fill function that pushes "
1415 "back characters onto an unbuffered port", SCM_EOL
);
1417 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1419 /* Reinstate the port's normal buffer. */
1424 /* The port has its own buffer. It is important that we use it,
1425 even if it happens to be smaller than our caller's buffer, so
1426 that a custom port implementation's entry points (in
1427 particular, fill_input) can rely on the buffer always being
1428 the same as they first set up. */
1429 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1431 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1432 memcpy (buffer
, pt
->read_pos
, n_available
);
1433 buffer
= (char *) buffer
+ n_available
;
1434 pt
->read_pos
+= n_available
;
1435 n_read
+= n_available
;
1436 size
-= n_available
;
1445 scm_c_read (SCM port
, void *buffer
, size_t size
)
1447 scm_i_pthread_mutex_t
*lock
;
1450 scm_c_lock_port (port
, &lock
);
1451 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1453 scm_i_pthread_mutex_unlock (lock
);
1459 /* Update the line and column number of PORT after consumption of C. */
1461 update_port_lf (scm_t_wchar c
, SCM port
)
1486 #define SCM_MBCHAR_BUF_SIZE (4)
1488 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1489 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1491 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1493 scm_t_wchar codepoint
;
1495 if (utf8_buf
[0] <= 0x7f)
1498 codepoint
= utf8_buf
[0];
1500 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1503 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1504 | (utf8_buf
[1] & 0x3f);
1506 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1509 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1510 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1511 | (utf8_buf
[2] & 0x3f);
1516 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1517 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1518 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1519 | (utf8_buf
[3] & 0x3f);
1525 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1526 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1527 representation, and set *LEN to the length in bytes. Return
1528 `EILSEQ' on error. */
1530 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1531 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1533 #define ASSERT_NOT_EOF(b) \
1534 if (SCM_UNLIKELY ((b) == EOF)) \
1536 #define CONSUME_PEEKED_BYTE() \
1543 pt
= SCM_PTAB_ENTRY (port
);
1545 byte
= scm_get_byte_or_eof_unlocked (port
);
1552 buf
[0] = (scm_t_uint8
) byte
;
1557 *codepoint
= buf
[0];
1558 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
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
[1] = (scm_t_uint8
) byte
;
1571 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1574 else if ((buf
[0] & 0xf0) == 0xe0)
1577 byte
= scm_peek_byte_or_eof_unlocked (port
);
1578 ASSERT_NOT_EOF (byte
);
1580 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1581 || (buf
[0] == 0xe0 && byte
< 0xa0)
1582 || (buf
[0] == 0xed && byte
> 0x9f)))
1585 CONSUME_PEEKED_BYTE ();
1586 buf
[1] = (scm_t_uint8
) byte
;
1589 byte
= scm_peek_byte_or_eof_unlocked (port
);
1590 ASSERT_NOT_EOF (byte
);
1592 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1595 CONSUME_PEEKED_BYTE ();
1596 buf
[2] = (scm_t_uint8
) byte
;
1599 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1600 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1603 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1606 byte
= scm_peek_byte_or_eof_unlocked (port
);
1607 ASSERT_NOT_EOF (byte
);
1609 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1610 || (buf
[0] == 0xf0 && byte
< 0x90)
1611 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1614 CONSUME_PEEKED_BYTE ();
1615 buf
[1] = (scm_t_uint8
) byte
;
1618 byte
= scm_peek_byte_or_eof_unlocked (port
);
1619 ASSERT_NOT_EOF (byte
);
1621 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1624 CONSUME_PEEKED_BYTE ();
1625 buf
[2] = (scm_t_uint8
) byte
;
1628 byte
= scm_peek_byte_or_eof_unlocked (port
);
1629 ASSERT_NOT_EOF (byte
);
1631 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1634 CONSUME_PEEKED_BYTE ();
1635 buf
[3] = (scm_t_uint8
) byte
;
1638 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1639 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1640 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1649 /* Here we could choose the consume the faulty byte when it's not a
1650 valid starting byte, but it's not a requirement. What Section 3.9
1651 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1652 would otherwise be a valid starting byte. */
1656 #undef CONSUME_PEEKED_BYTE
1657 #undef ASSERT_NOT_EOF
1660 /* Read an ISO-8859-1 codepoint (a byte) from PORT. On success, return
1661 0 and set *CODEPOINT to the codepoint that was read, fill BUF with
1662 its UTF-8 representation, and set *LEN to the length in bytes.
1663 Return `EILSEQ' on error. */
1665 get_latin1_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1666 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1668 *codepoint
= scm_get_byte_or_eof_unlocked (port
);
1670 if (*codepoint
== EOF
)
1675 buf
[0] = *codepoint
;
1680 /* Likewise, read a byte sequence from PORT, passing it through its
1681 input conversion descriptor. */
1683 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1684 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1686 scm_t_iconv_descriptors
*id
;
1688 size_t bytes_consumed
, output_size
;
1690 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1692 id
= scm_i_port_iconv_descriptors (port
);
1694 for (output_size
= 0, output
= (char *) utf8_buf
,
1695 bytes_consumed
= 0, err
= 0;
1696 err
== 0 && output_size
== 0
1697 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1701 size_t input_left
, output_left
, done
;
1703 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1704 if (byte_read
== EOF
)
1706 if (bytes_consumed
== 0)
1708 *codepoint
= (scm_t_wchar
) EOF
;
1716 buf
[bytes_consumed
] = byte_read
;
1719 input_left
= bytes_consumed
+ 1;
1720 output_left
= sizeof (utf8_buf
);
1722 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1723 if (done
== (size_t) -1)
1727 /* Missing input: keep trying. */
1731 output_size
= sizeof (utf8_buf
) - output_left
;
1734 if (SCM_UNLIKELY (output_size
== 0))
1735 /* An unterminated sequence. */
1737 else if (SCM_LIKELY (err
== 0))
1739 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1740 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1741 *len
= bytes_consumed
;
1747 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1748 with the byte representation of the codepoint in PORT's encoding, and
1749 set *LEN to the length in bytes of that representation. Return 0 on
1750 success and an errno value on error. */
1751 static SCM_C_INLINE
int
1752 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1753 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1756 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1758 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1759 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1760 else if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_LATIN1
)
1761 err
= get_latin1_codepoint (port
, codepoint
, buf
, len
);
1763 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1765 if (SCM_LIKELY (err
== 0))
1766 update_port_lf (*codepoint
, port
);
1767 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1771 update_port_lf (*codepoint
, port
);
1777 /* Read a codepoint from PORT and return it. */
1779 scm_getc_unlocked (SCM port
)
1780 #define FUNC_NAME "scm_getc"
1784 scm_t_wchar codepoint
;
1785 char buf
[SCM_MBCHAR_BUF_SIZE
];
1787 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1788 if (SCM_UNLIKELY (err
!= 0))
1789 /* At this point PORT should point past the invalid encoding, as per
1790 R6RS-lib Section 8.2.4. */
1791 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1800 scm_i_pthread_mutex_t
*lock
;
1803 scm_c_lock_port (port
, &lock
);
1804 ret
= scm_getc_unlocked (port
);
1806 scm_i_pthread_mutex_unlock (lock
);
1812 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1814 "Return the next character available from @var{port}, updating\n"
1815 "@var{port} to point to the following character. If no more\n"
1816 "characters are available, the end-of-file object is returned.\n"
1818 "When @var{port}'s data cannot be decoded according to its\n"
1819 "character encoding, a @code{decoding-error} is raised and\n"
1820 "@var{port} points past the erroneous byte sequence.\n")
1821 #define FUNC_NAME s_scm_read_char
1824 if (SCM_UNBNDP (port
))
1825 port
= scm_current_input_port ();
1826 SCM_VALIDATE_OPINPORT (1, port
);
1827 c
= scm_getc_unlocked (port
);
1830 return SCM_MAKE_CHAR (c
);
1840 scm_unget_byte_unlocked (int c
, SCM port
)
1841 #define FUNC_NAME "scm_unget_byte"
1843 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1845 if (pt
->read_buf
== pt
->putback_buf
)
1846 /* already using the put-back buffer. */
1848 /* enlarge putback_buf if necessary. */
1849 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1850 && pt
->read_buf
== pt
->read_pos
)
1852 size_t new_size
= pt
->read_buf_size
* 2;
1853 unsigned char *tmp
= (unsigned char *)
1854 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1857 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1858 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1859 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1862 /* shift any existing bytes to buffer + 1. */
1863 if (pt
->read_pos
== pt
->read_end
)
1864 pt
->read_end
= pt
->read_buf
+ 1;
1865 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1867 int count
= pt
->read_end
- pt
->read_pos
;
1869 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1870 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1873 pt
->read_pos
= pt
->read_buf
;
1876 /* switch to the put-back buffer. */
1878 if (pt
->putback_buf
== NULL
)
1881 = (unsigned char *) scm_gc_malloc_pointerless
1882 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1883 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1886 pt
->saved_read_buf
= pt
->read_buf
;
1887 pt
->saved_read_pos
= pt
->read_pos
;
1888 pt
->saved_read_end
= pt
->read_end
;
1889 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1891 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1892 pt
->read_end
= pt
->read_buf
+ 1;
1893 pt
->read_buf_size
= pt
->putback_buf_size
;
1899 pt
->rw_active
= SCM_PORT_READ
;
1904 scm_unget_byte (int c
, SCM port
)
1906 scm_i_pthread_mutex_t
*lock
;
1907 scm_c_lock_port (port
, &lock
);
1908 scm_unget_byte_unlocked (c
, port
);
1910 scm_i_pthread_mutex_unlock (lock
);
1915 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1916 #define FUNC_NAME "scm_ungetc"
1918 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1920 char result_buf
[10];
1924 len
= sizeof (result_buf
);
1925 result
= u32_conv_to_encoding (pt
->encoding
,
1926 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1927 (uint32_t *) &c
, 1, NULL
,
1930 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1931 scm_encoding_error (FUNC_NAME
, errno
,
1932 "conversion to port encoding failed",
1933 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1935 for (i
= len
- 1; i
>= 0; i
--)
1936 scm_unget_byte_unlocked (result
[i
], port
);
1938 if (SCM_UNLIKELY (result
!= result_buf
))
1943 /* What should col be in this case?
1944 * We'll leave it at -1.
1946 SCM_LINUM (port
) -= 1;
1954 scm_ungetc (scm_t_wchar c
, SCM port
)
1956 scm_i_pthread_mutex_t
*lock
;
1957 scm_c_lock_port (port
, &lock
);
1958 scm_ungetc_unlocked (c
, port
);
1960 scm_i_pthread_mutex_unlock (lock
);
1965 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1967 /* This is simple minded and inefficient, but unreading strings is
1968 * probably not a common operation, and remember that line and
1969 * column numbers have to be handled...
1971 * Please feel free to write an optimized version!
1974 scm_ungetc_unlocked (s
[n
], port
);
1978 scm_ungets (const char *s
, int n
, SCM port
)
1980 scm_i_pthread_mutex_t
*lock
;
1981 scm_c_lock_port (port
, &lock
);
1982 scm_ungets_unlocked (s
, n
, port
);
1984 scm_i_pthread_mutex_unlock (lock
);
1988 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1990 "Return the next character available from @var{port},\n"
1991 "@emph{without} updating @var{port} to point to the following\n"
1992 "character. If no more characters are available, the\n"
1993 "end-of-file object is returned.\n"
1995 "The value returned by\n"
1996 "a call to @code{peek-char} is the same as the value that would\n"
1997 "have been returned by a call to @code{read-char} on the same\n"
1998 "port. The only difference is that the very next call to\n"
1999 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
2000 "return the value returned by the preceding call to\n"
2001 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
2002 "an interactive port will hang waiting for input whenever a call\n"
2003 "to @code{read-char} would have hung.\n"
2005 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
2006 "if such a situation occurs. However, unlike with @code{read-char},\n"
2007 "@var{port} still points at the beginning of the erroneous byte\n"
2008 "sequence when the error is raised.\n")
2009 #define FUNC_NAME s_scm_peek_char
2014 char bytes
[SCM_MBCHAR_BUF_SIZE
];
2015 long column
, line
, i
;
2018 if (SCM_UNBNDP (port
))
2019 port
= scm_current_input_port ();
2020 SCM_VALIDATE_OPINPORT (1, port
);
2022 column
= SCM_COL (port
);
2023 line
= SCM_LINUM (port
);
2025 err
= get_codepoint (port
, &c
, bytes
, &len
);
2027 for (i
= len
- 1; i
>= 0; i
--)
2028 scm_unget_byte_unlocked (bytes
[i
], port
);
2030 SCM_COL (port
) = column
;
2031 SCM_LINUM (port
) = line
;
2033 if (SCM_UNLIKELY (err
!= 0))
2035 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
2037 /* Shouldn't happen since `catch' always aborts to prompt. */
2038 result
= SCM_BOOL_F
;
2041 result
= SCM_EOF_VAL
;
2043 result
= SCM_MAKE_CHAR (c
);
2049 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2050 (SCM cobj
, SCM port
),
2051 "Place character @var{cobj} in @var{port} so that it will be\n"
2052 "read by the next read operation. If called multiple times, the\n"
2053 "unread characters will be read again in last-in first-out\n"
2054 "order. If @var{port} is not supplied, the current input port\n"
2056 #define FUNC_NAME s_scm_unread_char
2060 SCM_VALIDATE_CHAR (1, cobj
);
2061 if (SCM_UNBNDP (port
))
2062 port
= scm_current_input_port ();
2063 SCM_VALIDATE_OPINPORT (2, port
);
2065 c
= SCM_CHAR (cobj
);
2067 scm_ungetc_unlocked (c
, port
);
2072 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2073 (SCM str
, SCM port
),
2074 "Place the string @var{str} in @var{port} so that its characters will be\n"
2075 "read in subsequent read operations. If called multiple times, the\n"
2076 "unread characters will be read again in last-in first-out order. If\n"
2077 "@var{port} is not supplied, the current-input-port is used.")
2078 #define FUNC_NAME s_scm_unread_string
2081 SCM_VALIDATE_STRING (1, str
);
2082 if (SCM_UNBNDP (port
))
2083 port
= scm_current_input_port ();
2084 SCM_VALIDATE_OPINPORT (2, port
);
2086 n
= scm_i_string_length (str
);
2089 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2098 /* Manipulating the buffers. */
2100 /* This routine does not take any locks, as it is usually called as part
2101 of a port implementation. */
2103 scm_port_non_buffer (scm_t_port
*pt
)
2105 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2106 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2107 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2108 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2111 /* this should only be called when the read buffer is empty. it
2112 tries to refill the read buffer. it returns the first char from
2113 the port, which is either EOF or *(pt->read_pos). */
2115 scm_fill_input_unlocked (SCM port
)
2117 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2119 assert (pt
->read_pos
== pt
->read_end
);
2121 if (pt
->read_buf
== pt
->putback_buf
)
2123 /* finished reading put-back chars. */
2124 pt
->read_buf
= pt
->saved_read_buf
;
2125 pt
->read_pos
= pt
->saved_read_pos
;
2126 pt
->read_end
= pt
->saved_read_end
;
2127 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2128 if (pt
->read_pos
< pt
->read_end
)
2129 return *(pt
->read_pos
);
2131 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2135 scm_fill_input (SCM port
)
2137 scm_i_pthread_mutex_t
*lock
;
2140 scm_c_lock_port (port
, &lock
);
2141 ret
= scm_fill_input_unlocked (port
);
2143 scm_i_pthread_mutex_unlock (lock
);
2149 /* Move up to READ_LEN bytes from PORT's putback and/or read buffers
2150 into memory starting at DEST. Return the number of bytes moved.
2151 PORT's line/column numbers are left unchanged. */
2153 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2155 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2156 size_t bytes_read
= 0;
2157 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2161 memcpy (dest
, pt
->read_pos
, from_buf
);
2162 pt
->read_pos
+= from_buf
;
2163 bytes_read
+= from_buf
;
2164 read_len
-= from_buf
;
2168 /* if putback was active, try the real input buffer too. */
2169 if (pt
->read_buf
== pt
->putback_buf
)
2171 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2174 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2175 pt
->saved_read_pos
+= from_buf
;
2176 bytes_read
+= from_buf
;
2183 /* Clear a port's read buffers, returning the contents. */
2184 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2186 "This procedure clears a port's input buffers, similar\n"
2187 "to the way that force-output clears the output buffer. The\n"
2188 "contents of the buffers are returned as a single string, e.g.,\n"
2191 "(define p (open-input-file ...))\n"
2192 "(drain-input p) => empty string, nothing buffered yet.\n"
2193 "(unread-char (read-char p) p)\n"
2194 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2196 "Draining the buffers may be useful for cleanly finishing\n"
2197 "buffered I/O so that the file descriptor can be used directly\n"
2198 "for further input.")
2199 #define FUNC_NAME s_scm_drain_input
2206 SCM_VALIDATE_OPINPORT (1, port
);
2207 pt
= SCM_PTAB_ENTRY (port
);
2209 count
= pt
->read_end
- pt
->read_pos
;
2210 if (pt
->read_buf
== pt
->putback_buf
)
2211 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2215 result
= scm_i_make_string (count
, &data
, 0);
2216 scm_take_from_input_buffers (port
, data
, count
);
2219 result
= scm_nullstr
;
2226 scm_end_input_unlocked (SCM port
)
2229 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2231 if (pt
->read_buf
== pt
->putback_buf
)
2233 offset
= pt
->read_end
- pt
->read_pos
;
2234 pt
->read_buf
= pt
->saved_read_buf
;
2235 pt
->read_pos
= pt
->saved_read_pos
;
2236 pt
->read_end
= pt
->saved_read_end
;
2237 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2242 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2246 scm_end_input (SCM port
)
2248 scm_i_pthread_mutex_t
*lock
;
2249 scm_c_lock_port (port
, &lock
);
2250 scm_end_input_unlocked (port
);
2252 scm_i_pthread_mutex_unlock (lock
);
2256 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2258 "Flush the specified output port, or the current output port if @var{port}\n"
2259 "is omitted. The current output buffer contents are passed to the\n"
2260 "underlying port implementation (e.g., in the case of fports, the\n"
2261 "data will be written to the file and the output buffer will be cleared.)\n"
2262 "It has no effect on an unbuffered port.\n\n"
2263 "The return value is unspecified.")
2264 #define FUNC_NAME s_scm_force_output
2266 if (SCM_UNBNDP (port
))
2267 port
= scm_current_output_port ();
2270 port
= SCM_COERCE_OUTPORT (port
);
2271 SCM_VALIDATE_OPOUTPORT (1, port
);
2273 scm_flush_unlocked (port
);
2274 return SCM_UNSPECIFIED
;
2279 scm_flush_unlocked (SCM port
)
2281 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2285 scm_flush (SCM port
)
2287 scm_i_pthread_mutex_t
*lock
;
2288 scm_c_lock_port (port
, &lock
);
2289 scm_flush_unlocked (port
);
2291 scm_i_pthread_mutex_unlock (lock
);
2301 scm_putc (char c
, SCM port
)
2303 scm_i_pthread_mutex_t
*lock
;
2304 scm_c_lock_port (port
, &lock
);
2305 scm_putc_unlocked (c
, port
);
2307 scm_i_pthread_mutex_unlock (lock
);
2312 scm_puts (const char *s
, SCM port
)
2314 scm_i_pthread_mutex_t
*lock
;
2315 scm_c_lock_port (port
, &lock
);
2316 scm_puts_unlocked (s
, port
);
2318 scm_i_pthread_mutex_unlock (lock
);
2324 * Used by an application to write arbitrary number of bytes to an SCM
2325 * port. Similar semantics as libc write. However, unlike libc
2326 * write, scm_c_write writes the requested number of bytes and has no
2329 * Warning: Doesn't update port line and column counts!
2332 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2333 #define FUNC_NAME "scm_c_write"
2336 scm_t_ptob_descriptor
*ptob
;
2338 SCM_VALIDATE_OPOUTPORT (1, port
);
2340 pt
= SCM_PTAB_ENTRY (port
);
2341 ptob
= SCM_PORT_DESCRIPTOR (port
);
2343 if (pt
->rw_active
== SCM_PORT_READ
)
2344 scm_end_input_unlocked (port
);
2346 ptob
->write (port
, ptr
, size
);
2349 pt
->rw_active
= SCM_PORT_WRITE
;
2354 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2356 scm_i_pthread_mutex_t
*lock
;
2357 scm_c_lock_port (port
, &lock
);
2358 scm_c_write_unlocked (port
, ptr
, size
);
2360 scm_i_pthread_mutex_unlock (lock
);
2366 * This function differs from scm_c_write; it updates port line and
2369 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2371 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2372 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2374 if (pt
->rw_active
== SCM_PORT_READ
)
2375 scm_end_input_unlocked (port
);
2377 ptob
->write (port
, ptr
, size
);
2379 for (; size
; ptr
++, size
--)
2380 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2383 pt
->rw_active
= SCM_PORT_WRITE
;
2387 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2389 scm_i_pthread_mutex_t
*lock
;
2390 scm_c_lock_port (port
, &lock
);
2391 scm_lfwrite_unlocked (ptr
, size
, port
);
2393 scm_i_pthread_mutex_unlock (lock
);
2397 /* Write STR to PORT from START inclusive to END exclusive. */
2399 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2401 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2403 if (pt
->rw_active
== SCM_PORT_READ
)
2404 scm_end_input_unlocked (port
);
2406 if (end
== (size_t) -1)
2407 end
= scm_i_string_length (str
);
2409 scm_i_display_substring (str
, start
, end
, port
);
2412 pt
->rw_active
= SCM_PORT_WRITE
;
2418 /* Querying and setting positions, and character availability. */
2420 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2422 "Return @code{#t} if a character is ready on input @var{port}\n"
2423 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2424 "@code{#t} then the next @code{read-char} operation on\n"
2425 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2426 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2428 "@code{char-ready?} exists to make it possible for a\n"
2429 "program to accept characters from interactive ports without\n"
2430 "getting stuck waiting for input. Any input editors associated\n"
2431 "with such ports must make sure that characters whose existence\n"
2432 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2433 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2434 "a port at end of file would be indistinguishable from an\n"
2435 "interactive port that has no ready characters.")
2436 #define FUNC_NAME s_scm_char_ready_p
2440 if (SCM_UNBNDP (port
))
2441 port
= scm_current_input_port ();
2442 /* It's possible to close the current input port, so validate even in
2444 SCM_VALIDATE_OPINPORT (1, port
);
2446 pt
= SCM_PTAB_ENTRY (port
);
2448 /* if the current read buffer is filled, or the
2449 last pushed-back char has been read and the saved buffer is
2450 filled, result is true. */
2451 if (pt
->read_pos
< pt
->read_end
2452 || (pt
->read_buf
== pt
->putback_buf
2453 && pt
->saved_read_pos
< pt
->saved_read_end
))
2457 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2459 if (ptob
->input_waiting
)
2460 return scm_from_bool(ptob
->input_waiting (port
));
2467 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2468 (SCM fd_port
, SCM offset
, SCM whence
),
2469 "Sets the current position of @var{fd_port} to the integer\n"
2470 "@var{offset}, which is interpreted according to the value of\n"
2473 "One of the following variables should be supplied for\n"
2475 "@defvar SEEK_SET\n"
2476 "Seek from the beginning of the file.\n"
2478 "@defvar SEEK_CUR\n"
2479 "Seek from the current position.\n"
2481 "@defvar SEEK_END\n"
2482 "Seek from the end of the file.\n"
2484 "If @var{fd_port} is a file descriptor, the underlying system\n"
2485 "call is @code{lseek}. @var{port} may be a string port.\n"
2487 "The value returned is the new position in the file. This means\n"
2488 "that the current position of a port can be obtained using:\n"
2490 "(seek port 0 SEEK_CUR)\n"
2492 #define FUNC_NAME s_scm_seek
2496 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2498 how
= scm_to_int (whence
);
2499 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2500 SCM_OUT_OF_RANGE (3, whence
);
2502 if (SCM_OPPORTP (fd_port
))
2504 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2505 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2506 off_t_or_off64_t rv
;
2509 SCM_MISC_ERROR ("port is not seekable",
2510 scm_cons (fd_port
, SCM_EOL
));
2512 rv
= ptob
->seek (fd_port
, off
, how
);
2513 return scm_from_off_t_or_off64_t (rv
);
2515 else /* file descriptor?. */
2517 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2518 off_t_or_off64_t rv
;
2519 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2522 return scm_from_off_t_or_off64_t (rv
);
2531 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2532 doesn't have the filename version truncate(), hence this code. */
2533 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2535 truncate (const char *file
, off_t length
)
2539 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2543 ret
= ftruncate (fdes
, length
);
2546 int save_errno
= errno
;
2552 return close (fdes
);
2554 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2556 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2557 (SCM object
, SCM length
),
2558 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2559 "can be a filename string, a port object, or an integer file\n"
2561 "The return value is unspecified.\n"
2563 "For a port or file descriptor @var{length} can be omitted, in\n"
2564 "which case the file is truncated at the current position (per\n"
2565 "@code{ftell} above).\n"
2567 "On most systems a file can be extended by giving a length\n"
2568 "greater than the current size, but this is not mandatory in the\n"
2570 #define FUNC_NAME s_scm_truncate_file
2574 /* "object" can be a port, fdes or filename.
2576 Negative "length" makes no sense, but it's left to truncate() or
2577 ftruncate() to give back an error for that (normally EINVAL).
2580 if (SCM_UNBNDP (length
))
2582 /* must supply length if object is a filename. */
2583 if (scm_is_string (object
))
2584 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2586 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2589 object
= SCM_COERCE_OUTPORT (object
);
2590 if (scm_is_integer (object
))
2592 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2593 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2596 else if (SCM_OPOUTPORTP (object
))
2598 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2599 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2600 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2602 if (!ptob
->truncate
)
2603 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2604 if (pt
->rw_active
== SCM_PORT_READ
)
2605 scm_end_input_unlocked (object
);
2606 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2607 ptob
->flush (object
);
2609 ptob
->truncate (object
, c_length
);
2614 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2615 char *str
= scm_to_locale_string (object
);
2617 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2624 return SCM_UNSPECIFIED
;
2628 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2630 "Return the current line number for @var{port}.\n"
2632 "The first line of a file is 0. But you might want to add 1\n"
2633 "when printing line numbers, since starting from 1 is\n"
2634 "traditional in error messages, and likely to be more natural to\n"
2636 #define FUNC_NAME s_scm_port_line
2638 port
= SCM_COERCE_OUTPORT (port
);
2639 SCM_VALIDATE_OPENPORT (1, port
);
2640 return scm_from_long (SCM_LINUM (port
));
2644 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2645 (SCM port
, SCM line
),
2646 "Set the current line number for @var{port} to @var{line}. The\n"
2647 "first line of a file is 0.")
2648 #define FUNC_NAME s_scm_set_port_line_x
2650 port
= SCM_COERCE_OUTPORT (port
);
2651 SCM_VALIDATE_OPENPORT (1, port
);
2652 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2653 return SCM_UNSPECIFIED
;
2657 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2659 "Return the current column number of @var{port}.\n"
2660 "If the number is\n"
2661 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2662 "- i.e. the first character of the first line is line 0, column 0.\n"
2663 "(However, when you display a file position, for example in an error\n"
2664 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2665 "because lines and column numbers traditionally start with 1, and that is\n"
2666 "what non-programmers will find most natural.)")
2667 #define FUNC_NAME s_scm_port_column
2669 port
= SCM_COERCE_OUTPORT (port
);
2670 SCM_VALIDATE_OPENPORT (1, port
);
2671 return scm_from_int (SCM_COL (port
));
2675 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2676 (SCM port
, SCM column
),
2677 "Set the current column of @var{port}. Before reading the first\n"
2678 "character on a line the column should be 0.")
2679 #define FUNC_NAME s_scm_set_port_column_x
2681 port
= SCM_COERCE_OUTPORT (port
);
2682 SCM_VALIDATE_OPENPORT (1, port
);
2683 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2684 return SCM_UNSPECIFIED
;
2688 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2690 "Return the filename associated with @var{port}, or @code{#f}\n"
2691 "if no filename is associated with the port.")
2692 #define FUNC_NAME s_scm_port_filename
2694 port
= SCM_COERCE_OUTPORT (port
);
2695 SCM_VALIDATE_OPENPORT (1, port
);
2696 return SCM_FILENAME (port
);
2700 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2701 (SCM port
, SCM filename
),
2702 "Change the filename associated with @var{port}, using the current input\n"
2703 "port if none is specified. Note that this does not change the port's\n"
2704 "source of data, but only the value that is returned by\n"
2705 "@code{port-filename} and reported in diagnostic output.")
2706 #define FUNC_NAME s_scm_set_port_filename_x
2708 port
= SCM_COERCE_OUTPORT (port
);
2709 SCM_VALIDATE_OPENPORT (1, port
);
2710 /* We allow the user to set the filename to whatever he likes. */
2711 SCM_SET_FILENAME (port
, filename
);
2712 return SCM_UNSPECIFIED
;
2719 /* Implementation helpers for port printing functions. */
2722 scm_print_port_mode (SCM exp
, SCM port
)
2724 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2726 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2727 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2730 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2737 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2739 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2742 scm_puts_unlocked ("#<", port
);
2743 scm_print_port_mode (exp
, port
);
2744 scm_puts_unlocked (type
, port
);
2745 scm_putc_unlocked (' ', port
);
2746 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2747 scm_putc_unlocked ('>', port
);
2754 /* Iterating over all ports. */
2756 struct for_each_data
2758 void (*proc
) (void *data
, SCM p
);
2763 for_each_trampoline (void *data
, SCM port
, SCM result
)
2765 struct for_each_data
*d
= data
;
2767 d
->proc (d
->data
, port
);
2773 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2775 struct for_each_data d
;
2780 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2781 scm_i_port_weak_set
);
2785 scm_for_each_trampoline (void *data
, SCM port
)
2787 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2790 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2792 "Apply @var{proc} to each port in the Guile port table\n"
2793 "in turn. The return value is unspecified. More specifically,\n"
2794 "@var{proc} is applied exactly once to every port that exists\n"
2795 "in the system at the time @code{port-for-each} is invoked.\n"
2796 "Changes to the port table while @code{port-for-each} is running\n"
2797 "have no effect as far as @code{port-for-each} is concerned.")
2798 #define FUNC_NAME s_scm_port_for_each
2800 SCM_VALIDATE_PROC (1, proc
);
2802 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2804 return SCM_UNSPECIFIED
;
2809 flush_output_port (void *closure
, SCM port
)
2811 if (SCM_OPOUTPORTP (port
))
2812 scm_flush_unlocked (port
);
2815 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2817 "Equivalent to calling @code{force-output} on\n"
2818 "all open output ports. The return value is unspecified.")
2819 #define FUNC_NAME s_scm_flush_all_ports
2821 scm_c_port_for_each (&flush_output_port
, NULL
);
2822 return SCM_UNSPECIFIED
;
2831 scm_t_bits scm_tc16_void_port
= 0;
2833 static int fill_input_void_port (SCM port SCM_UNUSED
)
2839 write_void_port (SCM port SCM_UNUSED
,
2840 const void *data SCM_UNUSED
,
2841 size_t size SCM_UNUSED
)
2846 scm_i_void_port (long mode_bits
)
2850 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2852 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2858 scm_void_port (char *mode_str
)
2860 return scm_i_void_port (scm_mode_bits (mode_str
));
2863 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2865 "Create and return a new void port. A void port acts like\n"
2866 "@file{/dev/null}. The @var{mode} argument\n"
2867 "specifies the input/output modes for this port: see the\n"
2868 "documentation for @code{open-file} in @ref{File Ports}.")
2869 #define FUNC_NAME s_scm_sys_make_void_port
2871 return scm_i_void_port (scm_i_mode_bits (mode
));
2878 /* Initialization. */
2883 /* lseek() symbols. */
2884 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2885 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2886 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2888 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2891 cur_inport_fluid
= scm_make_fluid ();
2892 cur_outport_fluid
= scm_make_fluid ();
2893 cur_errport_fluid
= scm_make_fluid ();
2894 cur_loadport_fluid
= scm_make_fluid ();
2896 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2898 #include "libguile/ports.x"
2900 /* Use Latin-1 as the default port encoding. */
2901 SCM_VARIABLE_SET (default_port_encoding_var
,
2902 scm_make_fluid_with_default (SCM_BOOL_F
));
2903 scm_port_encoding_init
= 1;
2905 SCM_VARIABLE_SET (default_conversion_strategy_var
,
2906 scm_make_fluid_with_default (sym_substitute
));
2907 scm_conversion_strategy_init
= 1;
2909 /* These bindings are used when boot-9 turns `current-input-port' et
2910 al into parameters. They are then removed from the guile module. */
2911 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2912 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2913 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);