1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 3 of
7 * the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
32 #include <fcntl.h> /* for chsize on mingw */
37 #include <striconveh.h>
41 #include "libguile/_scm.h"
42 #include "libguile/async.h"
43 #include "libguile/deprecation.h"
44 #include "libguile/eval.h"
45 #include "libguile/fports.h" /* direct access for seek and truncate */
46 #include "libguile/goops.h"
47 #include "libguile/smob.h"
48 #include "libguile/chars.h"
49 #include "libguile/dynwind.h"
51 #include "libguile/keywords.h"
52 #include "libguile/hashtab.h"
53 #include "libguile/root.h"
54 #include "libguile/strings.h"
55 #include "libguile/mallocs.h"
56 #include "libguile/validate.h"
57 #include "libguile/ports.h"
58 #include "libguile/vectors.h"
59 #include "libguile/weak-set.h"
60 #include "libguile/fluids.h"
61 #include "libguile/eq.h"
75 #ifdef HAVE_SYS_IOCTL_H
76 #include <sys/ioctl.h>
79 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
80 already, but have this code here in case that wasn't so in past versions,
81 or perhaps to help other minimal DOS environments.
83 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
84 might be possibilities if we've got other systems without ftruncate. */
86 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
87 #define ftruncate(fd, size) chsize (fd, size)
89 #define HAVE_FTRUNCATE 1
93 /* The port kind table --- a dynamically resized array of port types. */
96 /* scm_ptobs scm_numptob
97 * implement a dynamically resized array of ptob records.
98 * Indexes into this table are used when generating type
99 * tags for smobjects (if you know a tag you can get an index and conversely).
101 static scm_t_ptob_descriptor
**scm_ptobs
= NULL
;
102 static long scm_numptob
= 0; /* Number of port types. */
103 static long scm_ptobs_size
= 0; /* Number of slots in the port type
105 static scm_i_pthread_mutex_t scm_ptobs_lock
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
108 scm_c_num_port_types (void)
112 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
114 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
119 scm_t_ptob_descriptor
*
120 scm_c_port_type_ref (long ptobnum
)
122 scm_t_ptob_descriptor
*ret
= NULL
;
124 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
126 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
127 ret
= scm_ptobs
[ptobnum
];
129 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
132 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
138 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
142 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
144 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
146 if (scm_numptob
== scm_ptobs_size
)
148 unsigned long old_size
= scm_ptobs_size
;
149 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
151 /* Currently there are only 9 predefined port types, so one
152 resize will cover it. */
153 scm_ptobs_size
= old_size
+ 10;
155 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
156 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
158 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
161 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
165 scm_ptobs
[ret
] = desc
;
168 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
171 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
177 * We choose to use an interface similar to the smob interface with
178 * fill_input and write as standard fields, passed to the port
179 * type constructor, and optional fields set by setters.
183 flush_port_default (SCM port SCM_UNUSED
)
188 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
193 scm_make_port_type (char *name
,
194 int (*fill_input
) (SCM port
),
195 void (*write
) (SCM port
, const void *data
, size_t size
))
197 scm_t_ptob_descriptor
*desc
;
200 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
201 memset (desc
, 0, sizeof (*desc
));
204 desc
->print
= scm_port_print
;
206 desc
->flush
= flush_port_default
;
207 desc
->end_input
= end_input_default
;
208 desc
->fill_input
= fill_input
;
210 ptobnum
= scm_c_port_type_add_x (desc
);
212 /* Make a class object if GOOPS is present. */
213 if (SCM_UNPACK (scm_port_class
[0]) != 0)
214 scm_make_port_classes (ptobnum
, name
);
216 return scm_tc7_port
+ ptobnum
* 256;
220 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
222 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
226 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
228 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
232 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
233 scm_print_state
*pstate
))
235 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
239 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
241 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
245 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
247 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
251 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
253 scm_t_ptob_descriptor
*ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
));
255 ptob
->flags
|= SCM_PORT_TYPE_HAS_FLUSH
;
259 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
261 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
265 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
267 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
271 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
273 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
277 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
279 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
284 /* Standard ports --- current input, output, error, and more(!). */
286 static SCM cur_inport_fluid
= SCM_BOOL_F
;
287 static SCM cur_outport_fluid
= SCM_BOOL_F
;
288 static SCM cur_errport_fluid
= SCM_BOOL_F
;
289 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
291 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
293 "Return the current input port. This is the default port used\n"
294 "by many input procedures. Initially, @code{current-input-port}\n"
295 "returns the @dfn{standard input} in Unix and C terminology.")
296 #define FUNC_NAME s_scm_current_input_port
298 if (scm_is_true (cur_inport_fluid
))
299 return scm_fluid_ref (cur_inport_fluid
);
305 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
307 "Return the current output port. This is the default port used\n"
308 "by many output procedures. Initially,\n"
309 "@code{current-output-port} returns the @dfn{standard output} in\n"
310 "Unix and C terminology.")
311 #define FUNC_NAME s_scm_current_output_port
313 if (scm_is_true (cur_outport_fluid
))
314 return scm_fluid_ref (cur_outport_fluid
);
320 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
322 "Return the port to which errors and warnings should be sent (the\n"
323 "@dfn{standard error} in Unix and C terminology).")
324 #define FUNC_NAME s_scm_current_error_port
326 if (scm_is_true (cur_errport_fluid
))
327 return scm_fluid_ref (cur_errport_fluid
);
334 scm_current_warning_port (void)
336 static SCM cwp_var
= SCM_BOOL_F
;
338 if (scm_is_false (cwp_var
))
339 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
341 return scm_call_0 (scm_variable_ref (cwp_var
));
344 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
346 "Return the current-load-port.\n"
347 "The load port is used internally by @code{primitive-load}.")
348 #define FUNC_NAME s_scm_current_load_port
350 return scm_fluid_ref (cur_loadport_fluid
);
354 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
356 "@deffnx {Scheme Procedure} set-current-output-port port\n"
357 "@deffnx {Scheme Procedure} set-current-error-port port\n"
358 "Change the ports returned by @code{current-input-port},\n"
359 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
360 "so that they use the supplied @var{port} for input or output.")
361 #define FUNC_NAME s_scm_set_current_input_port
363 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
364 SCM_VALIDATE_OPINPORT (1, port
);
365 scm_fluid_set_x (cur_inport_fluid
, port
);
371 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
373 "Set the current default output port to @var{port}.")
374 #define FUNC_NAME s_scm_set_current_output_port
376 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
377 port
= SCM_COERCE_OUTPORT (port
);
378 SCM_VALIDATE_OPOUTPORT (1, port
);
379 scm_fluid_set_x (cur_outport_fluid
, port
);
385 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
387 "Set the current default error port to @var{port}.")
388 #define FUNC_NAME s_scm_set_current_error_port
390 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
391 port
= SCM_COERCE_OUTPORT (port
);
392 SCM_VALIDATE_OPOUTPORT (1, port
);
393 scm_fluid_set_x (cur_errport_fluid
, port
);
400 scm_set_current_warning_port (SCM port
)
402 static SCM cwp_var
= SCM_BOOL_F
;
404 if (scm_is_false (cwp_var
))
405 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
407 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
412 scm_dynwind_current_input_port (SCM port
)
413 #define FUNC_NAME NULL
415 SCM_VALIDATE_OPINPORT (1, port
);
416 scm_dynwind_fluid (cur_inport_fluid
, port
);
421 scm_dynwind_current_output_port (SCM port
)
422 #define FUNC_NAME NULL
424 port
= SCM_COERCE_OUTPORT (port
);
425 SCM_VALIDATE_OPOUTPORT (1, port
);
426 scm_dynwind_fluid (cur_outport_fluid
, port
);
431 scm_dynwind_current_error_port (SCM port
)
432 #define FUNC_NAME NULL
434 port
= SCM_COERCE_OUTPORT (port
);
435 SCM_VALIDATE_OPOUTPORT (1, port
);
436 scm_dynwind_fluid (cur_errport_fluid
, port
);
441 scm_i_dynwind_current_load_port (SCM port
)
443 scm_dynwind_fluid (cur_loadport_fluid
, port
);
449 /* Retrieving a port's mode. */
451 /* Return the flags that characterize a port based on the mode
452 * string used to open a file for that port.
454 * See PORT FLAGS in scm.h
458 scm_i_mode_bits_n (SCM modes
)
461 | (scm_i_string_contains_char (modes
, 'r')
462 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
463 | (scm_i_string_contains_char (modes
, 'w')
464 || scm_i_string_contains_char (modes
, 'a')
465 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
466 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
467 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
471 scm_mode_bits (char *modes
)
473 /* Valid characters are rw+a0l. So, use latin1. */
474 return scm_i_mode_bits (scm_from_latin1_string (modes
));
478 scm_i_mode_bits (SCM modes
)
482 if (!scm_is_string (modes
))
483 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
485 bits
= scm_i_mode_bits_n (modes
);
486 scm_remember_upto_here_1 (modes
);
490 /* Return the mode flags from an open port.
491 * Some modes such as "append" are only used when opening
492 * a file and are not returned here. */
494 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
496 "Return the port modes associated with the open port @var{port}.\n"
497 "These will not necessarily be identical to the modes used when\n"
498 "the port was opened, since modes such as \"append\" which are\n"
499 "used only during port creation are not retained.")
500 #define FUNC_NAME s_scm_port_mode
505 port
= SCM_COERCE_OUTPORT (port
);
506 SCM_VALIDATE_OPPORT (1, port
);
507 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
508 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
509 strcpy (modes
, "r+");
513 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
515 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
518 return scm_from_latin1_string (modes
);
525 /* The port table --- a weak set of all ports.
527 We need a global registry of ports to flush them all at exit, and to
528 get all the ports matching a file descriptor. */
529 SCM scm_i_port_weak_set
;
534 /* Port finalization. */
538 scm_t_ptob_descriptor
*ptob
;
543 do_free (void *body_data
)
545 struct do_free_data
*data
= body_data
;
547 /* `close' is for explicit `close-port' by user. `free' is for this
548 purpose: ports collected by the GC. */
549 data
->ptob
->free (data
->port
);
554 /* Finalize the object (a port) pointed to by PTR. */
556 finalize_port (GC_PTR ptr
, GC_PTR data
)
558 SCM port
= SCM_PACK_POINTER (ptr
);
560 if (!SCM_PORTP (port
))
563 if (SCM_OPENP (port
))
565 struct do_free_data data
;
567 SCM_CLR_PORT_OPEN_FLAG (port
);
569 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
572 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
573 scm_handle_by_message_noexit
, NULL
);
575 scm_gc_ports_collected
++;
583 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
584 const char *encoding
,
585 scm_t_string_failed_conversion_handler handler
,
590 scm_t_ptob_descriptor
*ptob
;
592 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
593 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
595 ret
= scm_words (tag
| mode_bits
, 3);
596 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
597 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
599 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
600 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
602 entry
->file_name
= SCM_BOOL_F
;
603 entry
->rw_active
= SCM_PORT_NEITHER
;
605 entry
->stream
= stream
;
606 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
607 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
608 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
609 else if (!encoding
|| strcmp (encoding
, "ISO-8859-1") == 0)
610 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
612 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
613 entry
->ilseq_handler
= handler
;
614 entry
->iconv_descriptors
= NULL
;
616 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
617 scm_i_set_finalizer (SCM2PTR (ret
), finalize_port
, NULL
);
619 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
620 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
626 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
628 return scm_c_make_port_with_encoding (tag
, mode_bits
,
629 scm_i_default_port_encoding (),
630 scm_i_get_conversion_strategy (SCM_BOOL_F
),
635 scm_new_port_table_entry (scm_t_bits tag
)
637 return scm_c_make_port (tag
, 0, 0);
644 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
646 "Return a boolean indicating whether @var{x} is a port.\n"
647 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
649 #define FUNC_NAME s_scm_port_p
651 return scm_from_bool (SCM_PORTP (x
));
655 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
657 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
658 "@code{#f}. Any object satisfying this predicate also satisfies\n"
660 #define FUNC_NAME s_scm_input_port_p
662 return scm_from_bool (SCM_INPUT_PORT_P (x
));
666 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
668 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
669 "@code{#f}. Any object satisfying this predicate also satisfies\n"
671 #define FUNC_NAME s_scm_output_port_p
673 x
= SCM_COERCE_OUTPORT (x
);
674 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
678 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
680 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
682 #define FUNC_NAME s_scm_port_closed_p
684 SCM_VALIDATE_PORT (1, port
);
685 return scm_from_bool (!SCM_OPPORTP (port
));
689 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
691 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
693 #define FUNC_NAME s_scm_eof_object_p
695 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
704 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
707 * Call the close operation on a port object.
708 * see also scm_close.
710 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
712 "Close the specified port object. Return @code{#t} if it\n"
713 "successfully closes a port or @code{#f} if it was already\n"
714 "closed. An exception may be raised if an error occurs, for\n"
715 "example when flushing buffered output. See also @ref{Ports and\n"
716 "File Descriptors, close}, for a procedure which can close file\n"
718 #define FUNC_NAME s_scm_close_port
723 port
= SCM_COERCE_OUTPORT (port
);
725 SCM_VALIDATE_PORT (1, port
);
726 if (SCM_CLOSEDP (port
))
729 p
= SCM_PTAB_ENTRY (port
);
730 SCM_CLR_PORT_OPEN_FLAG (port
);
732 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
733 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
735 if (SCM_PORT_DESCRIPTOR (port
)->close
)
736 /* Note! This may throw an exception. Anything after this point
737 should be resilient to non-local exits. */
738 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
742 if (p
->iconv_descriptors
)
744 /* If we don't get here, the iconv_descriptors finalizer will
746 close_iconv_descriptors (p
->iconv_descriptors
);
747 p
->iconv_descriptors
= NULL
;
750 return scm_from_bool (rv
>= 0);
754 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
756 "Close the specified input port object. The routine has no effect if\n"
757 "the file has already been closed. An exception may be raised if an\n"
758 "error occurs. The value returned is unspecified.\n\n"
759 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
760 "which can close file descriptors.")
761 #define FUNC_NAME s_scm_close_input_port
763 SCM_VALIDATE_INPUT_PORT (1, port
);
764 scm_close_port (port
);
765 return SCM_UNSPECIFIED
;
769 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
771 "Close the specified output port object. The routine has no effect if\n"
772 "the file has already been closed. An exception may be raised if an\n"
773 "error occurs. The value returned is unspecified.\n\n"
774 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
775 "which can close file descriptors.")
776 #define FUNC_NAME s_scm_close_output_port
778 port
= SCM_COERCE_OUTPORT (port
);
779 SCM_VALIDATE_OUTPUT_PORT (1, port
);
780 scm_close_port (port
);
781 return SCM_UNSPECIFIED
;
788 /* Encoding characters to byte streams, and decoding byte streams to
791 /* A fluid specifying the default encoding for newly created ports. If it is
792 a string, that is the encoding. If it is #f, it is in the "native"
793 (Latin-1) encoding. */
794 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
796 static int scm_port_encoding_init
= 0;
798 /* Use ENCODING as the default encoding for future ports. */
800 scm_i_set_default_port_encoding (const char *encoding
)
802 if (!scm_port_encoding_init
803 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
804 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
808 || !strcmp (encoding
, "ASCII")
809 || !strcmp (encoding
, "ANSI_X3.4-1968")
810 || !strcmp (encoding
, "ISO-8859-1"))
811 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
817 str
= scm_from_latin1_string (encoding
);
819 /* Restrict to ASCII. */
820 for (i
= 0; encoding
[i
]; i
++)
821 if (encoding
[i
] > 127)
822 scm_misc_error ("scm_i_set_default_port_encoding",
823 "invalid character encoding ~s", scm_list_1 (str
));
825 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), str
);
829 /* Return the name of the default encoding for newly created ports; a
830 return value of NULL means "ISO-8859-1". */
832 scm_i_default_port_encoding (void)
834 if (!scm_port_encoding_init
)
836 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
842 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
843 if (!scm_is_string (encoding
))
846 return scm_i_string_chars (encoding
);
851 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
853 close_iconv_descriptors (ptr
);
856 static scm_t_iconv_descriptors
*
857 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
859 scm_t_iconv_descriptors
*id
;
860 iconv_t input_cd
, output_cd
;
863 input_cd
= (iconv_t
) -1;
864 output_cd
= (iconv_t
) -1;
866 for (i
= 0; encoding
[i
]; i
++)
867 if (encoding
[i
] > 127)
868 goto invalid_encoding
;
872 /* Open an input iconv conversion descriptor, from ENCODING
873 to UTF-8. We choose UTF-8, not UTF-32, because iconv
874 implementations can typically convert from anything to
875 UTF-8, but not to UTF-32 (see
876 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
878 /* Assume opening an iconv descriptor causes about 16 KB of
880 scm_gc_register_allocation (16 * 1024);
882 input_cd
= iconv_open ("UTF-8", encoding
);
883 if (input_cd
== (iconv_t
) -1)
884 goto invalid_encoding
;
889 /* Assume opening an iconv descriptor causes about 16 KB of
891 scm_gc_register_allocation (16 * 1024);
893 output_cd
= iconv_open (encoding
, "UTF-8");
894 if (output_cd
== (iconv_t
) -1)
896 if (input_cd
!= (iconv_t
) -1)
897 iconv_close (input_cd
);
898 goto invalid_encoding
;
902 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
903 id
->input_cd
= input_cd
;
904 id
->output_cd
= output_cd
;
906 /* Register a finalizer to close the descriptors. */
907 scm_i_set_finalizer (id
, finalize_iconv_descriptors
, NULL
);
914 err
= scm_from_latin1_string (encoding
);
915 scm_misc_error ("open_iconv_descriptors",
916 "invalid or unknown character encoding ~s",
922 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
924 if (id
->input_cd
!= (iconv_t
) -1)
925 iconv_close (id
->input_cd
);
926 if (id
->output_cd
!= (iconv_t
) -1)
927 iconv_close (id
->output_cd
);
928 id
->input_cd
= (void *) -1;
929 id
->output_cd
= (void *) -1;
932 scm_t_iconv_descriptors
*
933 scm_i_port_iconv_descriptors (SCM port
)
937 pt
= SCM_PTAB_ENTRY (port
);
939 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
941 if (!pt
->iconv_descriptors
)
944 pt
->encoding
= "ISO-8859-1";
945 pt
->iconv_descriptors
=
946 open_iconv_descriptors (pt
->encoding
,
947 SCM_INPUT_PORT_P (port
),
948 SCM_OUTPUT_PORT_P (port
));
951 return pt
->iconv_descriptors
;
954 /* The name of the encoding is itself encoded in ASCII. */
956 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
959 scm_t_iconv_descriptors
*prev
;
961 /* Set the character encoding for this port. */
962 pt
= SCM_PTAB_ENTRY (port
);
963 prev
= pt
->iconv_descriptors
;
965 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
967 pt
->encoding
= "UTF-8";
968 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
969 pt
->iconv_descriptors
= NULL
;
971 else if (!encoding
|| strcmp (encoding
, "ISO-8859-1") == 0)
973 pt
->encoding
= "ISO-8859-1";
974 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
975 pt
->iconv_descriptors
= NULL
;
979 /* Open descriptors before mutating the port. */
980 pt
->iconv_descriptors
=
981 open_iconv_descriptors (encoding
,
982 SCM_INPUT_PORT_P (port
),
983 SCM_OUTPUT_PORT_P (port
));
984 pt
->encoding
= scm_gc_strdup (encoding
, "port");
985 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
989 close_iconv_descriptors (prev
);
992 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
994 "Returns, as a string, the character encoding that @var{port}\n"
995 "uses to interpret its input and output.\n")
996 #define FUNC_NAME s_scm_port_encoding
1001 SCM_VALIDATE_PORT (1, port
);
1003 pt
= SCM_PTAB_ENTRY (port
);
1006 return scm_from_latin1_string (pt
->encoding
);
1012 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1013 (SCM port
, SCM enc
),
1014 "Sets the character encoding that will be used to interpret all\n"
1015 "port I/O. New ports are created with the encoding\n"
1016 "appropriate for the current locale if @code{setlocale} has \n"
1017 "been called or ISO-8859-1 otherwise\n"
1018 "and this procedure can be used to modify that encoding.\n")
1019 #define FUNC_NAME s_scm_set_port_encoding_x
1023 SCM_VALIDATE_PORT (1, port
);
1024 SCM_VALIDATE_STRING (2, enc
);
1026 enc_str
= scm_to_latin1_string (enc
);
1027 scm_i_set_port_encoding_x (port
, enc_str
);
1030 return SCM_UNSPECIFIED
;
1035 /* This determines how conversions handle unconvertible characters. */
1036 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
1037 static int scm_conversion_strategy_init
= 0;
1039 scm_t_string_failed_conversion_handler
1040 scm_i_get_conversion_strategy (SCM port
)
1044 if (scm_is_false (port
))
1046 if (!scm_conversion_strategy_init
1047 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1048 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1051 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
1052 if (scm_is_false (encoding
))
1053 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1055 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
1061 pt
= SCM_PTAB_ENTRY (port
);
1062 return pt
->ilseq_handler
;
1068 scm_i_set_conversion_strategy_x (SCM port
,
1069 scm_t_string_failed_conversion_handler handler
)
1074 strategy
= scm_from_int ((int) handler
);
1076 if (scm_is_false (port
))
1078 /* Set the default encoding for future ports. */
1079 if (!scm_conversion_strategy_init
1080 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1081 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1083 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1087 /* Set the character encoding for this port. */
1088 pt
= SCM_PTAB_ENTRY (port
);
1089 pt
->ilseq_handler
= handler
;
1093 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1094 1, 0, 0, (SCM port
),
1095 "Returns the behavior of the port when handling a character that\n"
1096 "is not representable in the port's current encoding.\n"
1097 "It returns the symbol @code{error} if unrepresentable characters\n"
1098 "should cause exceptions, @code{substitute} if the port should\n"
1099 "try to replace unrepresentable characters with question marks or\n"
1100 "approximate characters, or @code{escape} if unrepresentable\n"
1101 "characters should be converted to string escapes.\n"
1103 "If @var{port} is @code{#f}, then the current default behavior\n"
1104 "will be returned. New ports will have this default behavior\n"
1105 "when they are created.\n")
1106 #define FUNC_NAME s_scm_port_conversion_strategy
1108 scm_t_string_failed_conversion_handler h
;
1110 SCM_VALIDATE_OPPORT (1, port
);
1112 if (!scm_is_false (port
))
1114 SCM_VALIDATE_OPPORT (1, port
);
1117 h
= scm_i_get_conversion_strategy (port
);
1118 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1119 return scm_from_latin1_symbol ("error");
1120 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1121 return scm_from_latin1_symbol ("substitute");
1122 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1123 return scm_from_latin1_symbol ("escape");
1127 /* Never gets here. */
1128 return SCM_UNDEFINED
;
1132 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1134 (SCM port
, SCM sym
),
1135 "Sets the behavior of the interpreter when outputting a character\n"
1136 "that is not representable in the port's current encoding.\n"
1137 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1138 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1139 "when an unconvertible character is encountered. If it is\n"
1140 "@code{'substitute}, then unconvertible characters will \n"
1141 "be replaced with approximate characters, or with question marks\n"
1142 "if no approximately correct character is available.\n"
1143 "If it is @code{'escape},\n"
1144 "it will appear as a hex escape when output.\n"
1146 "If @var{port} is an open port, the conversion error behavior\n"
1147 "is set for that port. If it is @code{#f}, it is set as the\n"
1148 "default behavior for any future ports that get created in\n"
1150 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1156 if (!scm_is_false (port
))
1158 SCM_VALIDATE_OPPORT (1, port
);
1161 err
= scm_from_latin1_symbol ("error");
1162 if (scm_is_true (scm_eqv_p (sym
, err
)))
1164 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1165 return SCM_UNSPECIFIED
;
1168 qm
= scm_from_latin1_symbol ("substitute");
1169 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1171 scm_i_set_conversion_strategy_x (port
,
1172 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1173 return SCM_UNSPECIFIED
;
1176 esc
= scm_from_latin1_symbol ("escape");
1177 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1179 scm_i_set_conversion_strategy_x (port
,
1180 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1181 return SCM_UNSPECIFIED
;
1184 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1186 return SCM_UNSPECIFIED
;
1193 /* The port lock. */
1196 lock_port (void *mutex
)
1198 scm_i_pthread_mutex_lock (mutex
);
1202 unlock_port (void *mutex
)
1204 scm_i_pthread_mutex_unlock (mutex
);
1208 scm_dynwind_lock_port (SCM port
)
1209 #define FUNC_NAME "dynwind-lock-port"
1211 scm_i_pthread_mutex_t
*lock
;
1212 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1213 scm_c_lock_port (port
, &lock
);
1216 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1217 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1228 scm_get_byte_or_eof (SCM port
)
1230 scm_i_pthread_mutex_t
*lock
;
1233 scm_c_lock_port (port
, &lock
);
1234 ret
= scm_get_byte_or_eof_unlocked (port
);
1236 scm_i_pthread_mutex_unlock (lock
);
1242 scm_peek_byte_or_eof (SCM port
)
1244 scm_i_pthread_mutex_t
*lock
;
1247 scm_c_lock_port (port
, &lock
);
1248 ret
= scm_peek_byte_or_eof_unlocked (port
);
1250 scm_i_pthread_mutex_unlock (lock
);
1257 * Used by an application to read arbitrary number of bytes from an
1258 * SCM port. Same semantics as libc read, except that scm_c_read only
1259 * returns less than SIZE bytes if at end-of-file.
1261 * Warning: Doesn't update port line and column counts! */
1263 /* This structure, and the following swap_buffer function, are used
1264 for temporarily swapping a port's own read buffer, and the buffer
1265 that the caller of scm_c_read provides. */
1266 struct port_and_swap_buffer
1269 unsigned char *buffer
;
1274 swap_buffer (void *data
)
1276 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1277 unsigned char *old_buf
= psb
->pt
->read_buf
;
1278 size_t old_size
= psb
->pt
->read_buf_size
;
1280 /* Make the port use (buffer, size) from the struct. */
1281 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1282 psb
->pt
->read_buf_size
= psb
->size
;
1284 /* Save the port's old (buffer, size) in the struct. */
1285 psb
->buffer
= old_buf
;
1286 psb
->size
= old_size
;
1290 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1291 #define FUNC_NAME "scm_c_read"
1294 size_t n_read
= 0, n_available
;
1295 struct port_and_swap_buffer psb
;
1297 SCM_VALIDATE_OPINPORT (1, port
);
1299 pt
= SCM_PTAB_ENTRY (port
);
1300 if (pt
->rw_active
== SCM_PORT_WRITE
)
1301 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1304 pt
->rw_active
= SCM_PORT_READ
;
1306 /* Take bytes first from the port's read buffer. */
1307 if (pt
->read_pos
< pt
->read_end
)
1309 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1310 memcpy (buffer
, pt
->read_pos
, n_available
);
1311 buffer
= (char *) buffer
+ n_available
;
1312 pt
->read_pos
+= n_available
;
1313 n_read
+= n_available
;
1314 size
-= n_available
;
1317 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1321 /* Now we will call scm_fill_input repeatedly until we have read the
1322 requested number of bytes. (Note that a single scm_fill_input
1323 call does not guarantee to fill the whole of the port's read
1325 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1327 /* The port that we are reading from is unbuffered - i.e. does
1328 not have its own persistent buffer - but we have a buffer,
1329 provided by our caller, that is the right size for the data
1330 that is wanted. For the following scm_fill_input calls,
1331 therefore, we use the buffer in hand as the port's read
1334 We need to make sure that the port's normal (1 byte) buffer
1335 is reinstated in case one of the scm_fill_input () calls
1336 throws an exception; we use the scm_dynwind_* API to achieve
1339 A consequence of this optimization is that the fill_input
1340 functions can't unget characters. That'll push data to the
1341 pushback buffer instead of this psb buffer. */
1343 unsigned char *pback
= pt
->putback_buf
;
1346 psb
.buffer
= buffer
;
1348 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1349 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1350 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1352 /* Call scm_fill_input until we have all the bytes that we need,
1354 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1356 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1357 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1360 if (pback
!= pt
->putback_buf
1361 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1362 scm_misc_error (FUNC_NAME
,
1363 "scm_c_read must not call a fill function that pushes "
1364 "back characters onto an unbuffered port", SCM_EOL
);
1366 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1368 /* Reinstate the port's normal buffer. */
1373 /* The port has its own buffer. It is important that we use it,
1374 even if it happens to be smaller than our caller's buffer, so
1375 that a custom port implementation's entry points (in
1376 particular, fill_input) can rely on the buffer always being
1377 the same as they first set up. */
1378 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1380 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1381 memcpy (buffer
, pt
->read_pos
, n_available
);
1382 buffer
= (char *) buffer
+ n_available
;
1383 pt
->read_pos
+= n_available
;
1384 n_read
+= n_available
;
1385 size
-= n_available
;
1394 scm_c_read (SCM port
, void *buffer
, size_t size
)
1396 scm_i_pthread_mutex_t
*lock
;
1399 scm_c_lock_port (port
, &lock
);
1400 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1402 scm_i_pthread_mutex_unlock (lock
);
1408 /* Update the line and column number of PORT after consumption of C. */
1410 update_port_lf (scm_t_wchar c
, SCM port
)
1435 #define SCM_MBCHAR_BUF_SIZE (4)
1437 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1438 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1440 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1442 scm_t_wchar codepoint
;
1444 if (utf8_buf
[0] <= 0x7f)
1447 codepoint
= utf8_buf
[0];
1449 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1452 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1453 | (utf8_buf
[1] & 0x3f);
1455 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1458 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1459 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1460 | (utf8_buf
[2] & 0x3f);
1465 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1466 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1467 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1468 | (utf8_buf
[3] & 0x3f);
1474 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1475 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1476 representation, and set *LEN to the length in bytes. Return
1477 `EILSEQ' on error. */
1479 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1480 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1482 #define ASSERT_NOT_EOF(b) \
1483 if (SCM_UNLIKELY ((b) == EOF)) \
1485 #define CONSUME_PEEKED_BYTE() \
1492 pt
= SCM_PTAB_ENTRY (port
);
1494 byte
= scm_get_byte_or_eof_unlocked (port
);
1501 buf
[0] = (scm_t_uint8
) byte
;
1506 *codepoint
= buf
[0];
1507 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1510 byte
= scm_peek_byte_or_eof_unlocked (port
);
1511 ASSERT_NOT_EOF (byte
);
1513 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1516 CONSUME_PEEKED_BYTE ();
1517 buf
[1] = (scm_t_uint8
) byte
;
1520 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1523 else if ((buf
[0] & 0xf0) == 0xe0)
1526 byte
= scm_peek_byte_or_eof_unlocked (port
);
1527 ASSERT_NOT_EOF (byte
);
1529 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1530 || (buf
[0] == 0xe0 && byte
< 0xa0)
1531 || (buf
[0] == 0xed && byte
> 0x9f)))
1534 CONSUME_PEEKED_BYTE ();
1535 buf
[1] = (scm_t_uint8
) byte
;
1538 byte
= scm_peek_byte_or_eof_unlocked (port
);
1539 ASSERT_NOT_EOF (byte
);
1541 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1544 CONSUME_PEEKED_BYTE ();
1545 buf
[2] = (scm_t_uint8
) byte
;
1548 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1549 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1552 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1555 byte
= scm_peek_byte_or_eof_unlocked (port
);
1556 ASSERT_NOT_EOF (byte
);
1558 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1559 || (buf
[0] == 0xf0 && byte
< 0x90)
1560 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1563 CONSUME_PEEKED_BYTE ();
1564 buf
[1] = (scm_t_uint8
) byte
;
1567 byte
= scm_peek_byte_or_eof_unlocked (port
);
1568 ASSERT_NOT_EOF (byte
);
1570 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1573 CONSUME_PEEKED_BYTE ();
1574 buf
[2] = (scm_t_uint8
) byte
;
1577 byte
= scm_peek_byte_or_eof_unlocked (port
);
1578 ASSERT_NOT_EOF (byte
);
1580 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1583 CONSUME_PEEKED_BYTE ();
1584 buf
[3] = (scm_t_uint8
) byte
;
1587 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1588 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1589 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1598 /* Here we could choose the consume the faulty byte when it's not a
1599 valid starting byte, but it's not a requirement. What Section 3.9
1600 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1601 would otherwise be a valid starting byte. */
1605 #undef CONSUME_PEEKED_BYTE
1606 #undef ASSERT_NOT_EOF
1609 /* Read an ISO-8859-1 codepoint (a byte) from PORT. On success, return
1610 0 and set *CODEPOINT to the codepoint that was read, fill BUF with
1611 its UTF-8 representation, and set *LEN to the length in bytes.
1612 Return `EILSEQ' on error. */
1614 get_latin1_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1615 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1617 *codepoint
= scm_get_byte_or_eof_unlocked (port
);
1619 if (*codepoint
== EOF
)
1624 buf
[0] = *codepoint
;
1629 /* Likewise, read a byte sequence from PORT, passing it through its
1630 input conversion descriptor. */
1632 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1633 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1635 scm_t_iconv_descriptors
*id
;
1637 size_t bytes_consumed
, output_size
;
1639 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1641 id
= scm_i_port_iconv_descriptors (port
);
1643 for (output_size
= 0, output
= (char *) utf8_buf
,
1644 bytes_consumed
= 0, err
= 0;
1645 err
== 0 && output_size
== 0
1646 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1650 size_t input_left
, output_left
, done
;
1652 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1653 if (byte_read
== EOF
)
1655 if (bytes_consumed
== 0)
1657 *codepoint
= (scm_t_wchar
) EOF
;
1665 buf
[bytes_consumed
] = byte_read
;
1668 input_left
= bytes_consumed
+ 1;
1669 output_left
= sizeof (utf8_buf
);
1671 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1672 if (done
== (size_t) -1)
1676 /* Missing input: keep trying. */
1680 output_size
= sizeof (utf8_buf
) - output_left
;
1683 if (SCM_UNLIKELY (output_size
== 0))
1684 /* An unterminated sequence. */
1686 else if (SCM_LIKELY (err
== 0))
1688 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1689 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1690 *len
= bytes_consumed
;
1696 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1697 with the byte representation of the codepoint in PORT's encoding, and
1698 set *LEN to the length in bytes of that representation. Return 0 on
1699 success and an errno value on error. */
1700 static SCM_C_INLINE
int
1701 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1702 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1705 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1707 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1708 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1709 else if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_LATIN1
)
1710 err
= get_latin1_codepoint (port
, codepoint
, buf
, len
);
1712 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1714 if (SCM_LIKELY (err
== 0))
1715 update_port_lf (*codepoint
, port
);
1716 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1720 update_port_lf (*codepoint
, port
);
1726 /* Read a codepoint from PORT and return it. */
1728 scm_getc_unlocked (SCM port
)
1729 #define FUNC_NAME "scm_getc"
1733 scm_t_wchar codepoint
;
1734 char buf
[SCM_MBCHAR_BUF_SIZE
];
1736 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1737 if (SCM_UNLIKELY (err
!= 0))
1738 /* At this point PORT should point past the invalid encoding, as per
1739 R6RS-lib Section 8.2.4. */
1740 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1749 scm_i_pthread_mutex_t
*lock
;
1752 scm_c_lock_port (port
, &lock
);
1753 ret
= scm_getc_unlocked (port
);
1755 scm_i_pthread_mutex_unlock (lock
);
1761 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1763 "Return the next character available from @var{port}, updating\n"
1764 "@var{port} to point to the following character. If no more\n"
1765 "characters are available, the end-of-file object is returned.\n"
1767 "When @var{port}'s data cannot be decoded according to its\n"
1768 "character encoding, a @code{decoding-error} is raised and\n"
1769 "@var{port} points past the erroneous byte sequence.\n")
1770 #define FUNC_NAME s_scm_read_char
1773 if (SCM_UNBNDP (port
))
1774 port
= scm_current_input_port ();
1775 SCM_VALIDATE_OPINPORT (1, port
);
1776 c
= scm_getc_unlocked (port
);
1779 return SCM_MAKE_CHAR (c
);
1789 scm_unget_byte_unlocked (int c
, SCM port
)
1790 #define FUNC_NAME "scm_unget_byte"
1792 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1794 if (pt
->read_buf
== pt
->putback_buf
)
1795 /* already using the put-back buffer. */
1797 /* enlarge putback_buf if necessary. */
1798 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1799 && pt
->read_buf
== pt
->read_pos
)
1801 size_t new_size
= pt
->read_buf_size
* 2;
1802 unsigned char *tmp
= (unsigned char *)
1803 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1806 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1807 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1808 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1811 /* shift any existing bytes to buffer + 1. */
1812 if (pt
->read_pos
== pt
->read_end
)
1813 pt
->read_end
= pt
->read_buf
+ 1;
1814 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1816 int count
= pt
->read_end
- pt
->read_pos
;
1818 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1819 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1822 pt
->read_pos
= pt
->read_buf
;
1825 /* switch to the put-back buffer. */
1827 if (pt
->putback_buf
== NULL
)
1830 = (unsigned char *) scm_gc_malloc_pointerless
1831 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1832 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1835 pt
->saved_read_buf
= pt
->read_buf
;
1836 pt
->saved_read_pos
= pt
->read_pos
;
1837 pt
->saved_read_end
= pt
->read_end
;
1838 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1840 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1841 pt
->read_end
= pt
->read_buf
+ 1;
1842 pt
->read_buf_size
= pt
->putback_buf_size
;
1848 pt
->rw_active
= SCM_PORT_READ
;
1853 scm_unget_byte (int c
, SCM port
)
1855 scm_i_pthread_mutex_t
*lock
;
1856 scm_c_lock_port (port
, &lock
);
1857 scm_unget_byte_unlocked (c
, port
);
1859 scm_i_pthread_mutex_unlock (lock
);
1864 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1865 #define FUNC_NAME "scm_ungetc"
1867 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1869 char result_buf
[10];
1870 const char *encoding
;
1874 if (pt
->encoding
!= NULL
)
1875 encoding
= pt
->encoding
;
1877 encoding
= "ISO-8859-1";
1879 len
= sizeof (result_buf
);
1880 result
= u32_conv_to_encoding (encoding
,
1881 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1882 (uint32_t *) &c
, 1, NULL
,
1885 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1886 scm_encoding_error (FUNC_NAME
, errno
,
1887 "conversion to port encoding failed",
1888 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1890 for (i
= len
- 1; i
>= 0; i
--)
1891 scm_unget_byte_unlocked (result
[i
], port
);
1893 if (SCM_UNLIKELY (result
!= result_buf
))
1898 /* What should col be in this case?
1899 * We'll leave it at -1.
1901 SCM_LINUM (port
) -= 1;
1909 scm_ungetc (scm_t_wchar c
, SCM port
)
1911 scm_i_pthread_mutex_t
*lock
;
1912 scm_c_lock_port (port
, &lock
);
1913 scm_ungetc_unlocked (c
, port
);
1915 scm_i_pthread_mutex_unlock (lock
);
1920 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1922 /* This is simple minded and inefficient, but unreading strings is
1923 * probably not a common operation, and remember that line and
1924 * column numbers have to be handled...
1926 * Please feel free to write an optimized version!
1929 scm_ungetc_unlocked (s
[n
], port
);
1933 scm_ungets (const char *s
, int n
, SCM port
)
1935 scm_i_pthread_mutex_t
*lock
;
1936 scm_c_lock_port (port
, &lock
);
1937 scm_ungets_unlocked (s
, n
, port
);
1939 scm_i_pthread_mutex_unlock (lock
);
1943 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1945 "Return the next character available from @var{port},\n"
1946 "@emph{without} updating @var{port} to point to the following\n"
1947 "character. If no more characters are available, the\n"
1948 "end-of-file object is returned.\n"
1950 "The value returned by\n"
1951 "a call to @code{peek-char} is the same as the value that would\n"
1952 "have been returned by a call to @code{read-char} on the same\n"
1953 "port. The only difference is that the very next call to\n"
1954 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1955 "return the value returned by the preceding call to\n"
1956 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1957 "an interactive port will hang waiting for input whenever a call\n"
1958 "to @code{read-char} would have hung.\n"
1960 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1961 "if such a situation occurs. However, unlike with @code{read-char},\n"
1962 "@var{port} still points at the beginning of the erroneous byte\n"
1963 "sequence when the error is raised.\n")
1964 #define FUNC_NAME s_scm_peek_char
1969 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1970 long column
, line
, i
;
1973 if (SCM_UNBNDP (port
))
1974 port
= scm_current_input_port ();
1975 SCM_VALIDATE_OPINPORT (1, port
);
1977 column
= SCM_COL (port
);
1978 line
= SCM_LINUM (port
);
1980 err
= get_codepoint (port
, &c
, bytes
, &len
);
1982 for (i
= len
- 1; i
>= 0; i
--)
1983 scm_unget_byte_unlocked (bytes
[i
], port
);
1985 SCM_COL (port
) = column
;
1986 SCM_LINUM (port
) = line
;
1988 if (SCM_UNLIKELY (err
!= 0))
1990 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1992 /* Shouldn't happen since `catch' always aborts to prompt. */
1993 result
= SCM_BOOL_F
;
1996 result
= SCM_EOF_VAL
;
1998 result
= SCM_MAKE_CHAR (c
);
2004 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2005 (SCM cobj
, SCM port
),
2006 "Place character @var{cobj} in @var{port} so that it will be\n"
2007 "read by the next read operation. If called multiple times, the\n"
2008 "unread characters will be read again in last-in first-out\n"
2009 "order. If @var{port} is not supplied, the current input port\n"
2011 #define FUNC_NAME s_scm_unread_char
2015 SCM_VALIDATE_CHAR (1, cobj
);
2016 if (SCM_UNBNDP (port
))
2017 port
= scm_current_input_port ();
2018 SCM_VALIDATE_OPINPORT (2, port
);
2020 c
= SCM_CHAR (cobj
);
2022 scm_ungetc_unlocked (c
, port
);
2027 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2028 (SCM str
, SCM port
),
2029 "Place the string @var{str} in @var{port} so that its characters will be\n"
2030 "read in subsequent read operations. If called multiple times, the\n"
2031 "unread characters will be read again in last-in first-out order. If\n"
2032 "@var{port} is not supplied, the current-input-port is used.")
2033 #define FUNC_NAME s_scm_unread_string
2036 SCM_VALIDATE_STRING (1, str
);
2037 if (SCM_UNBNDP (port
))
2038 port
= scm_current_input_port ();
2039 SCM_VALIDATE_OPINPORT (2, port
);
2041 n
= scm_i_string_length (str
);
2044 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2053 /* Manipulating the buffers. */
2055 /* This routine does not take any locks, as it is usually called as part
2056 of a port implementation. */
2058 scm_port_non_buffer (scm_t_port
*pt
)
2060 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2061 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2062 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2063 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2066 /* this should only be called when the read buffer is empty. it
2067 tries to refill the read buffer. it returns the first char from
2068 the port, which is either EOF or *(pt->read_pos). */
2070 scm_fill_input_unlocked (SCM port
)
2072 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2074 assert (pt
->read_pos
== pt
->read_end
);
2076 if (pt
->read_buf
== pt
->putback_buf
)
2078 /* finished reading put-back chars. */
2079 pt
->read_buf
= pt
->saved_read_buf
;
2080 pt
->read_pos
= pt
->saved_read_pos
;
2081 pt
->read_end
= pt
->saved_read_end
;
2082 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2083 if (pt
->read_pos
< pt
->read_end
)
2084 return *(pt
->read_pos
);
2086 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2090 scm_fill_input (SCM port
)
2092 scm_i_pthread_mutex_t
*lock
;
2095 scm_c_lock_port (port
, &lock
);
2096 ret
= scm_fill_input_unlocked (port
);
2098 scm_i_pthread_mutex_unlock (lock
);
2104 /* move up to read_len chars from port's putback and/or read buffers
2105 into memory starting at dest. returns the number of chars moved. */
2107 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2109 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2110 size_t chars_read
= 0;
2111 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2115 memcpy (dest
, pt
->read_pos
, from_buf
);
2116 pt
->read_pos
+= from_buf
;
2117 chars_read
+= from_buf
;
2118 read_len
-= from_buf
;
2122 /* if putback was active, try the real input buffer too. */
2123 if (pt
->read_buf
== pt
->putback_buf
)
2125 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2128 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2129 pt
->saved_read_pos
+= from_buf
;
2130 chars_read
+= from_buf
;
2136 /* Clear a port's read buffers, returning the contents. */
2137 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2139 "This procedure clears a port's input buffers, similar\n"
2140 "to the way that force-output clears the output buffer. The\n"
2141 "contents of the buffers are returned as a single string, e.g.,\n"
2144 "(define p (open-input-file ...))\n"
2145 "(drain-input p) => empty string, nothing buffered yet.\n"
2146 "(unread-char (read-char p) p)\n"
2147 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2149 "Draining the buffers may be useful for cleanly finishing\n"
2150 "buffered I/O so that the file descriptor can be used directly\n"
2151 "for further input.")
2152 #define FUNC_NAME s_scm_drain_input
2159 SCM_VALIDATE_OPINPORT (1, port
);
2160 pt
= SCM_PTAB_ENTRY (port
);
2162 count
= pt
->read_end
- pt
->read_pos
;
2163 if (pt
->read_buf
== pt
->putback_buf
)
2164 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2168 result
= scm_i_make_string (count
, &data
, 0);
2169 scm_take_from_input_buffers (port
, data
, count
);
2172 result
= scm_nullstr
;
2179 scm_end_input_unlocked (SCM port
)
2182 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2184 if (pt
->read_buf
== pt
->putback_buf
)
2186 offset
= pt
->read_end
- pt
->read_pos
;
2187 pt
->read_buf
= pt
->saved_read_buf
;
2188 pt
->read_pos
= pt
->saved_read_pos
;
2189 pt
->read_end
= pt
->saved_read_end
;
2190 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2195 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2199 scm_end_input (SCM port
)
2201 scm_i_pthread_mutex_t
*lock
;
2202 scm_c_lock_port (port
, &lock
);
2203 scm_end_input_unlocked (port
);
2205 scm_i_pthread_mutex_unlock (lock
);
2209 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2211 "Flush the specified output port, or the current output port if @var{port}\n"
2212 "is omitted. The current output buffer contents are passed to the\n"
2213 "underlying port implementation (e.g., in the case of fports, the\n"
2214 "data will be written to the file and the output buffer will be cleared.)\n"
2215 "It has no effect on an unbuffered port.\n\n"
2216 "The return value is unspecified.")
2217 #define FUNC_NAME s_scm_force_output
2219 if (SCM_UNBNDP (port
))
2220 port
= scm_current_output_port ();
2223 port
= SCM_COERCE_OUTPORT (port
);
2224 SCM_VALIDATE_OPOUTPORT (1, port
);
2226 scm_flush_unlocked (port
);
2227 return SCM_UNSPECIFIED
;
2232 scm_flush_unlocked (SCM port
)
2234 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2238 scm_flush (SCM port
)
2240 scm_i_pthread_mutex_t
*lock
;
2241 scm_c_lock_port (port
, &lock
);
2242 scm_flush_unlocked (port
);
2244 scm_i_pthread_mutex_unlock (lock
);
2254 scm_putc (char c
, SCM port
)
2256 scm_i_pthread_mutex_t
*lock
;
2257 scm_c_lock_port (port
, &lock
);
2258 scm_putc_unlocked (c
, port
);
2260 scm_i_pthread_mutex_unlock (lock
);
2265 scm_puts (const char *s
, SCM port
)
2267 scm_i_pthread_mutex_t
*lock
;
2268 scm_c_lock_port (port
, &lock
);
2269 scm_puts_unlocked (s
, port
);
2271 scm_i_pthread_mutex_unlock (lock
);
2277 * Used by an application to write arbitrary number of bytes to an SCM
2278 * port. Similar semantics as libc write. However, unlike libc
2279 * write, scm_c_write writes the requested number of bytes and has no
2282 * Warning: Doesn't update port line and column counts!
2285 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2286 #define FUNC_NAME "scm_c_write"
2289 scm_t_ptob_descriptor
*ptob
;
2291 SCM_VALIDATE_OPOUTPORT (1, port
);
2293 pt
= SCM_PTAB_ENTRY (port
);
2294 ptob
= SCM_PORT_DESCRIPTOR (port
);
2296 if (pt
->rw_active
== SCM_PORT_READ
)
2297 scm_end_input_unlocked (port
);
2299 ptob
->write (port
, ptr
, size
);
2302 pt
->rw_active
= SCM_PORT_WRITE
;
2307 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2309 scm_i_pthread_mutex_t
*lock
;
2310 scm_c_lock_port (port
, &lock
);
2311 scm_c_write_unlocked (port
, ptr
, size
);
2313 scm_i_pthread_mutex_unlock (lock
);
2319 * This function differs from scm_c_write; it updates port line and
2322 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2324 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2325 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2327 if (pt
->rw_active
== SCM_PORT_READ
)
2328 scm_end_input_unlocked (port
);
2330 ptob
->write (port
, ptr
, size
);
2332 for (; size
; ptr
++, size
--)
2333 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2336 pt
->rw_active
= SCM_PORT_WRITE
;
2340 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2342 scm_i_pthread_mutex_t
*lock
;
2343 scm_c_lock_port (port
, &lock
);
2344 scm_lfwrite_unlocked (ptr
, size
, port
);
2346 scm_i_pthread_mutex_unlock (lock
);
2350 /* Write STR to PORT from START inclusive to END exclusive. */
2352 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2354 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2356 if (pt
->rw_active
== SCM_PORT_READ
)
2357 scm_end_input_unlocked (port
);
2359 if (end
== (size_t) -1)
2360 end
= scm_i_string_length (str
);
2362 scm_display (scm_c_substring (str
, start
, end
), port
);
2365 pt
->rw_active
= SCM_PORT_WRITE
;
2371 /* Querying and setting positions, and character availability. */
2373 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2375 "Return @code{#t} if a character is ready on input @var{port}\n"
2376 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2377 "@code{#t} then the next @code{read-char} operation on\n"
2378 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2379 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2381 "@code{char-ready?} exists to make it possible for a\n"
2382 "program to accept characters from interactive ports without\n"
2383 "getting stuck waiting for input. Any input editors associated\n"
2384 "with such ports must make sure that characters whose existence\n"
2385 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2386 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2387 "a port at end of file would be indistinguishable from an\n"
2388 "interactive port that has no ready characters.")
2389 #define FUNC_NAME s_scm_char_ready_p
2393 if (SCM_UNBNDP (port
))
2394 port
= scm_current_input_port ();
2395 /* It's possible to close the current input port, so validate even in
2397 SCM_VALIDATE_OPINPORT (1, port
);
2399 pt
= SCM_PTAB_ENTRY (port
);
2401 /* if the current read buffer is filled, or the
2402 last pushed-back char has been read and the saved buffer is
2403 filled, result is true. */
2404 if (pt
->read_pos
< pt
->read_end
2405 || (pt
->read_buf
== pt
->putback_buf
2406 && pt
->saved_read_pos
< pt
->saved_read_end
))
2410 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2412 if (ptob
->input_waiting
)
2413 return scm_from_bool(ptob
->input_waiting (port
));
2420 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2421 (SCM fd_port
, SCM offset
, SCM whence
),
2422 "Sets the current position of @var{fd_port} to the integer\n"
2423 "@var{offset}, which is interpreted according to the value of\n"
2426 "One of the following variables should be supplied for\n"
2428 "@defvar SEEK_SET\n"
2429 "Seek from the beginning of the file.\n"
2431 "@defvar SEEK_CUR\n"
2432 "Seek from the current position.\n"
2434 "@defvar SEEK_END\n"
2435 "Seek from the end of the file.\n"
2437 "If @var{fd_port} is a file descriptor, the underlying system\n"
2438 "call is @code{lseek}. @var{port} may be a string port.\n"
2440 "The value returned is the new position in the file. This means\n"
2441 "that the current position of a port can be obtained using:\n"
2443 "(seek port 0 SEEK_CUR)\n"
2445 #define FUNC_NAME s_scm_seek
2449 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2451 how
= scm_to_int (whence
);
2452 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2453 SCM_OUT_OF_RANGE (3, whence
);
2455 if (SCM_OPPORTP (fd_port
))
2457 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2458 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2459 off_t_or_off64_t rv
;
2462 SCM_MISC_ERROR ("port is not seekable",
2463 scm_cons (fd_port
, SCM_EOL
));
2465 rv
= ptob
->seek (fd_port
, off
, how
);
2466 return scm_from_off_t_or_off64_t (rv
);
2468 else /* file descriptor?. */
2470 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2471 off_t_or_off64_t rv
;
2472 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2475 return scm_from_off_t_or_off64_t (rv
);
2484 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2485 doesn't have the filename version truncate(), hence this code. */
2486 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2488 truncate (const char *file
, off_t length
)
2492 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2496 ret
= ftruncate (fdes
, length
);
2499 int save_errno
= errno
;
2505 return close (fdes
);
2507 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2509 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2510 (SCM object
, SCM length
),
2511 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2512 "can be a filename string, a port object, or an integer file\n"
2514 "The return value is unspecified.\n"
2516 "For a port or file descriptor @var{length} can be omitted, in\n"
2517 "which case the file is truncated at the current position (per\n"
2518 "@code{ftell} above).\n"
2520 "On most systems a file can be extended by giving a length\n"
2521 "greater than the current size, but this is not mandatory in the\n"
2523 #define FUNC_NAME s_scm_truncate_file
2527 /* "object" can be a port, fdes or filename.
2529 Negative "length" makes no sense, but it's left to truncate() or
2530 ftruncate() to give back an error for that (normally EINVAL).
2533 if (SCM_UNBNDP (length
))
2535 /* must supply length if object is a filename. */
2536 if (scm_is_string (object
))
2537 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2539 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2542 object
= SCM_COERCE_OUTPORT (object
);
2543 if (scm_is_integer (object
))
2545 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2546 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2549 else if (SCM_OPOUTPORTP (object
))
2551 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2552 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2553 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2555 if (!ptob
->truncate
)
2556 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2557 if (pt
->rw_active
== SCM_PORT_READ
)
2558 scm_end_input_unlocked (object
);
2559 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2560 ptob
->flush (object
);
2562 ptob
->truncate (object
, c_length
);
2567 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2568 char *str
= scm_to_locale_string (object
);
2570 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2577 return SCM_UNSPECIFIED
;
2581 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2583 "Return the current line number for @var{port}.\n"
2585 "The first line of a file is 0. But you might want to add 1\n"
2586 "when printing line numbers, since starting from 1 is\n"
2587 "traditional in error messages, and likely to be more natural to\n"
2589 #define FUNC_NAME s_scm_port_line
2591 port
= SCM_COERCE_OUTPORT (port
);
2592 SCM_VALIDATE_OPENPORT (1, port
);
2593 return scm_from_long (SCM_LINUM (port
));
2597 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2598 (SCM port
, SCM line
),
2599 "Set the current line number for @var{port} to @var{line}. The\n"
2600 "first line of a file is 0.")
2601 #define FUNC_NAME s_scm_set_port_line_x
2603 port
= SCM_COERCE_OUTPORT (port
);
2604 SCM_VALIDATE_OPENPORT (1, port
);
2605 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2606 return SCM_UNSPECIFIED
;
2610 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2612 "Return the current column number of @var{port}.\n"
2613 "If the number is\n"
2614 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2615 "- i.e. the first character of the first line is line 0, column 0.\n"
2616 "(However, when you display a file position, for example in an error\n"
2617 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2618 "because lines and column numbers traditionally start with 1, and that is\n"
2619 "what non-programmers will find most natural.)")
2620 #define FUNC_NAME s_scm_port_column
2622 port
= SCM_COERCE_OUTPORT (port
);
2623 SCM_VALIDATE_OPENPORT (1, port
);
2624 return scm_from_int (SCM_COL (port
));
2628 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2629 (SCM port
, SCM column
),
2630 "Set the current column of @var{port}. Before reading the first\n"
2631 "character on a line the column should be 0.")
2632 #define FUNC_NAME s_scm_set_port_column_x
2634 port
= SCM_COERCE_OUTPORT (port
);
2635 SCM_VALIDATE_OPENPORT (1, port
);
2636 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2637 return SCM_UNSPECIFIED
;
2641 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2643 "Return the filename associated with @var{port}, or @code{#f}\n"
2644 "if no filename is associated with the port.")
2645 #define FUNC_NAME s_scm_port_filename
2647 port
= SCM_COERCE_OUTPORT (port
);
2648 SCM_VALIDATE_OPENPORT (1, port
);
2649 return SCM_FILENAME (port
);
2653 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2654 (SCM port
, SCM filename
),
2655 "Change the filename associated with @var{port}, using the current input\n"
2656 "port if none is specified. Note that this does not change the port's\n"
2657 "source of data, but only the value that is returned by\n"
2658 "@code{port-filename} and reported in diagnostic output.")
2659 #define FUNC_NAME s_scm_set_port_filename_x
2661 port
= SCM_COERCE_OUTPORT (port
);
2662 SCM_VALIDATE_OPENPORT (1, port
);
2663 /* We allow the user to set the filename to whatever he likes. */
2664 SCM_SET_FILENAME (port
, filename
);
2665 return SCM_UNSPECIFIED
;
2672 /* Implementation helpers for port printing functions. */
2675 scm_print_port_mode (SCM exp
, SCM port
)
2677 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2679 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2680 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2683 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2690 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2692 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2695 scm_puts_unlocked ("#<", port
);
2696 scm_print_port_mode (exp
, port
);
2697 scm_puts_unlocked (type
, port
);
2698 scm_putc_unlocked (' ', port
);
2699 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2700 scm_putc_unlocked ('>', port
);
2707 /* Iterating over all ports. */
2709 struct for_each_data
2711 void (*proc
) (void *data
, SCM p
);
2716 for_each_trampoline (void *data
, SCM port
, SCM result
)
2718 struct for_each_data
*d
= data
;
2720 d
->proc (d
->data
, port
);
2726 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2728 struct for_each_data d
;
2733 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2734 scm_i_port_weak_set
);
2738 scm_for_each_trampoline (void *data
, SCM port
)
2740 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2743 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2745 "Apply @var{proc} to each port in the Guile port table\n"
2746 "in turn. The return value is unspecified. More specifically,\n"
2747 "@var{proc} is applied exactly once to every port that exists\n"
2748 "in the system at the time @code{port-for-each} is invoked.\n"
2749 "Changes to the port table while @code{port-for-each} is running\n"
2750 "have no effect as far as @code{port-for-each} is concerned.")
2751 #define FUNC_NAME s_scm_port_for_each
2753 SCM_VALIDATE_PROC (1, proc
);
2755 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2757 return SCM_UNSPECIFIED
;
2762 flush_output_port (void *closure
, SCM port
)
2764 if (SCM_OPOUTPORTP (port
))
2765 scm_flush_unlocked (port
);
2768 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2770 "Equivalent to calling @code{force-output} on\n"
2771 "all open output ports. The return value is unspecified.")
2772 #define FUNC_NAME s_scm_flush_all_ports
2774 scm_c_port_for_each (&flush_output_port
, NULL
);
2775 return SCM_UNSPECIFIED
;
2784 scm_t_bits scm_tc16_void_port
= 0;
2786 static int fill_input_void_port (SCM port SCM_UNUSED
)
2792 write_void_port (SCM port SCM_UNUSED
,
2793 const void *data SCM_UNUSED
,
2794 size_t size SCM_UNUSED
)
2799 scm_i_void_port (long mode_bits
)
2803 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2805 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2811 scm_void_port (char *mode_str
)
2813 return scm_i_void_port (scm_mode_bits (mode_str
));
2816 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2818 "Create and return a new void port. A void port acts like\n"
2819 "@file{/dev/null}. The @var{mode} argument\n"
2820 "specifies the input/output modes for this port: see the\n"
2821 "documentation for @code{open-file} in @ref{File Ports}.")
2822 #define FUNC_NAME s_scm_sys_make_void_port
2824 return scm_i_void_port (scm_i_mode_bits (mode
));
2831 /* Initialization. */
2836 /* lseek() symbols. */
2837 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2838 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2839 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2841 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2844 cur_inport_fluid
= scm_make_fluid ();
2845 cur_outport_fluid
= scm_make_fluid ();
2846 cur_errport_fluid
= scm_make_fluid ();
2847 cur_loadport_fluid
= scm_make_fluid ();
2849 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2851 #include "libguile/ports.x"
2853 /* Use Latin-1 as the default port encoding. */
2854 SCM_VARIABLE_SET (default_port_encoding_var
,
2855 scm_make_fluid_with_default (SCM_BOOL_F
));
2856 scm_port_encoding_init
= 1;
2858 SCM_VARIABLE_SET (scm_conversion_strategy
,
2859 scm_make_fluid_with_default
2860 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2861 scm_conversion_strategy_init
= 1;
2863 /* These bindings are used when boot-9 turns `current-input-port' et
2864 al into parameters. They are then removed from the guile module. */
2865 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2866 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2867 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);