1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 3 of
7 * the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
32 #include <fcntl.h> /* for chsize on mingw */
37 #include <striconveh.h>
41 #include "libguile/_scm.h"
42 #include "libguile/async.h"
43 #include "libguile/deprecation.h"
44 #include "libguile/eval.h"
45 #include "libguile/fports.h" /* direct access for seek and truncate */
46 #include "libguile/goops.h"
47 #include "libguile/smob.h"
48 #include "libguile/chars.h"
49 #include "libguile/dynwind.h"
51 #include "libguile/keywords.h"
52 #include "libguile/hashtab.h"
53 #include "libguile/root.h"
54 #include "libguile/strings.h"
55 #include "libguile/mallocs.h"
56 #include "libguile/validate.h"
57 #include "libguile/ports.h"
58 #include "libguile/vectors.h"
59 #include "libguile/weak-set.h"
60 #include "libguile/fluids.h"
61 #include "libguile/eq.h"
75 #ifdef HAVE_SYS_IOCTL_H
76 #include <sys/ioctl.h>
79 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
80 already, but have this code here in case that wasn't so in past versions,
81 or perhaps to help other minimal DOS environments.
83 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
84 might be possibilities if we've got other systems without ftruncate. */
86 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
87 #define ftruncate(fd, size) chsize (fd, size)
89 #define HAVE_FTRUNCATE 1
93 /* The port kind table --- a dynamically resized array of port types. */
96 /* scm_ptobs scm_numptob
97 * implement a dynamically resized array of ptob records.
98 * Indexes into this table are used when generating type
99 * tags for smobjects (if you know a tag you can get an index and conversely).
101 static scm_t_ptob_descriptor
**scm_ptobs
= NULL
;
102 static long scm_numptob
= 0; /* Number of port types. */
103 static long scm_ptobs_size
= 0; /* Number of slots in the port type
105 static scm_i_pthread_mutex_t scm_ptobs_lock
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
108 scm_c_num_port_types (void)
112 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
114 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
119 scm_t_ptob_descriptor
*
120 scm_c_port_type_ref (long ptobnum
)
122 scm_t_ptob_descriptor
*ret
= NULL
;
124 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
126 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
127 ret
= scm_ptobs
[ptobnum
];
129 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
132 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
138 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
142 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
144 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
146 if (scm_numptob
== scm_ptobs_size
)
148 unsigned long old_size
= scm_ptobs_size
;
149 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
151 /* Currently there are only 9 predefined port types, so one
152 resize will cover it. */
153 scm_ptobs_size
= old_size
+ 10;
155 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
156 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
158 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
161 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
165 scm_ptobs
[ret
] = desc
;
168 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
171 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
177 * We choose to use an interface similar to the smob interface with
178 * fill_input and write as standard fields, passed to the port
179 * type constructor, and optional fields set by setters.
183 flush_port_default (SCM port SCM_UNUSED
)
188 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
193 scm_make_port_type (char *name
,
194 int (*fill_input
) (SCM port
),
195 void (*write
) (SCM port
, const void *data
, size_t size
))
197 scm_t_ptob_descriptor
*desc
;
200 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
201 memset (desc
, 0, sizeof (*desc
));
204 desc
->print
= scm_port_print
;
206 desc
->flush
= flush_port_default
;
207 desc
->end_input
= end_input_default
;
208 desc
->fill_input
= fill_input
;
210 ptobnum
= scm_c_port_type_add_x (desc
);
212 /* Make a class object if GOOPS is present. */
213 if (SCM_UNPACK (scm_port_class
[0]) != 0)
214 scm_make_port_classes (ptobnum
, name
);
216 return scm_tc7_port
+ ptobnum
* 256;
220 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
222 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
226 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
228 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
232 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
233 scm_print_state
*pstate
))
235 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
239 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
241 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
245 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
247 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
251 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
253 scm_t_ptob_descriptor
*ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
));
255 ptob
->flags
|= SCM_PORT_TYPE_HAS_FLUSH
;
259 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
261 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
265 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
267 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
271 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
273 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
277 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
279 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
284 /* Standard ports --- current input, output, error, and more(!). */
286 static SCM cur_inport_fluid
= SCM_BOOL_F
;
287 static SCM cur_outport_fluid
= SCM_BOOL_F
;
288 static SCM cur_errport_fluid
= SCM_BOOL_F
;
289 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
291 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
293 "Return the current input port. This is the default port used\n"
294 "by many input procedures. Initially, @code{current-input-port}\n"
295 "returns the @dfn{standard input} in Unix and C terminology.")
296 #define FUNC_NAME s_scm_current_input_port
298 if (scm_is_true (cur_inport_fluid
))
299 return scm_fluid_ref (cur_inport_fluid
);
305 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
307 "Return the current output port. This is the default port used\n"
308 "by many output procedures. Initially,\n"
309 "@code{current-output-port} returns the @dfn{standard output} in\n"
310 "Unix and C terminology.")
311 #define FUNC_NAME s_scm_current_output_port
313 if (scm_is_true (cur_outport_fluid
))
314 return scm_fluid_ref (cur_outport_fluid
);
320 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
322 "Return the port to which errors and warnings should be sent (the\n"
323 "@dfn{standard error} in Unix and C terminology).")
324 #define FUNC_NAME s_scm_current_error_port
326 if (scm_is_true (cur_errport_fluid
))
327 return scm_fluid_ref (cur_errport_fluid
);
334 scm_current_warning_port (void)
336 static SCM cwp_var
= SCM_BOOL_F
;
338 if (scm_is_false (cwp_var
))
339 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
341 return scm_call_0 (scm_variable_ref (cwp_var
));
344 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
346 "Return the current-load-port.\n"
347 "The load port is used internally by @code{primitive-load}.")
348 #define FUNC_NAME s_scm_current_load_port
350 return scm_fluid_ref (cur_loadport_fluid
);
354 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
356 "@deffnx {Scheme Procedure} set-current-output-port port\n"
357 "@deffnx {Scheme Procedure} set-current-error-port port\n"
358 "Change the ports returned by @code{current-input-port},\n"
359 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
360 "so that they use the supplied @var{port} for input or output.")
361 #define FUNC_NAME s_scm_set_current_input_port
363 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
364 SCM_VALIDATE_OPINPORT (1, port
);
365 scm_fluid_set_x (cur_inport_fluid
, port
);
371 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
373 "Set the current default output port to @var{port}.")
374 #define FUNC_NAME s_scm_set_current_output_port
376 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
377 port
= SCM_COERCE_OUTPORT (port
);
378 SCM_VALIDATE_OPOUTPORT (1, port
);
379 scm_fluid_set_x (cur_outport_fluid
, port
);
385 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
387 "Set the current default error port to @var{port}.")
388 #define FUNC_NAME s_scm_set_current_error_port
390 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
391 port
= SCM_COERCE_OUTPORT (port
);
392 SCM_VALIDATE_OPOUTPORT (1, port
);
393 scm_fluid_set_x (cur_errport_fluid
, port
);
400 scm_set_current_warning_port (SCM port
)
402 static SCM cwp_var
= SCM_BOOL_F
;
404 if (scm_is_false (cwp_var
))
405 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
407 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
412 scm_dynwind_current_input_port (SCM port
)
413 #define FUNC_NAME NULL
415 SCM_VALIDATE_OPINPORT (1, port
);
416 scm_dynwind_fluid (cur_inport_fluid
, port
);
421 scm_dynwind_current_output_port (SCM port
)
422 #define FUNC_NAME NULL
424 port
= SCM_COERCE_OUTPORT (port
);
425 SCM_VALIDATE_OPOUTPORT (1, port
);
426 scm_dynwind_fluid (cur_outport_fluid
, port
);
431 scm_dynwind_current_error_port (SCM port
)
432 #define FUNC_NAME NULL
434 port
= SCM_COERCE_OUTPORT (port
);
435 SCM_VALIDATE_OPOUTPORT (1, port
);
436 scm_dynwind_fluid (cur_errport_fluid
, port
);
441 scm_i_dynwind_current_load_port (SCM port
)
443 scm_dynwind_fluid (cur_loadport_fluid
, port
);
449 /* Retrieving a port's mode. */
451 /* Return the flags that characterize a port based on the mode
452 * string used to open a file for that port.
454 * See PORT FLAGS in scm.h
458 scm_i_mode_bits_n (SCM modes
)
461 | (scm_i_string_contains_char (modes
, 'r')
462 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
463 | (scm_i_string_contains_char (modes
, 'w')
464 || scm_i_string_contains_char (modes
, 'a')
465 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
466 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
467 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
471 scm_mode_bits (char *modes
)
473 return scm_i_mode_bits (scm_from_locale_string (modes
));
477 scm_i_mode_bits (SCM modes
)
481 if (!scm_is_string (modes
))
482 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
484 bits
= scm_i_mode_bits_n (modes
);
485 scm_remember_upto_here_1 (modes
);
489 /* Return the mode flags from an open port.
490 * Some modes such as "append" are only used when opening
491 * a file and are not returned here. */
493 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
495 "Return the port modes associated with the open port @var{port}.\n"
496 "These will not necessarily be identical to the modes used when\n"
497 "the port was opened, since modes such as \"append\" which are\n"
498 "used only during port creation are not retained.")
499 #define FUNC_NAME s_scm_port_mode
504 port
= SCM_COERCE_OUTPORT (port
);
505 SCM_VALIDATE_OPPORT (1, port
);
506 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
507 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
508 strcpy (modes
, "r+");
512 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
514 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
517 return scm_from_latin1_string (modes
);
524 /* The port table --- a weak set of all ports.
526 We need a global registry of ports to flush them all at exit, and to
527 get all the ports matching a file descriptor. */
528 SCM scm_i_port_weak_set
;
533 /* Port finalization. */
535 static void finalize_port (GC_PTR
, GC_PTR
);
537 /* Register a finalizer for PORT. */
538 static SCM_C_INLINE_KEYWORD
void
539 register_finalizer_for_port (SCM port
)
541 GC_finalization_proc prev_finalizer
;
542 GC_PTR prev_finalization_data
;
544 /* Register a finalizer for PORT so that its iconv CDs get freed and
545 optionally its type's `free' function gets called. */
546 GC_REGISTER_FINALIZER_NO_ORDER (SCM2PTR (port
), finalize_port
, 0,
548 &prev_finalization_data
);
551 /* Finalize the object (a port) pointed to by PTR. */
553 finalize_port (GC_PTR ptr
, GC_PTR data
)
555 SCM port
= SCM_PACK_POINTER (ptr
);
557 if (!SCM_PORTP (port
))
560 if (SCM_OPENP (port
))
562 if (SCM_REVEALED (port
) > 0)
563 /* Keep "revealed" ports alive and re-register a finalizer. */
564 register_finalizer_for_port (port
);
567 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
570 /* Yes, I really do mean `free' rather than `close'. `close'
571 is for explicit `close-port' by user. */
574 SCM_SETSTREAM (port
, 0);
575 SCM_CLR_PORT_OPEN_FLAG (port
);
577 scm_gc_ports_collected
++;
586 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
587 const char *encoding
,
588 scm_t_string_failed_conversion_handler handler
,
593 scm_t_ptob_descriptor
*ptob
;
595 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
596 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
598 ret
= scm_words (tag
| mode_bits
, 3);
599 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
600 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
602 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
603 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
605 entry
->file_name
= SCM_BOOL_F
;
606 entry
->rw_active
= SCM_PORT_NEITHER
;
608 entry
->stream
= stream
;
609 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
610 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
611 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
613 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
614 entry
->ilseq_handler
= handler
;
615 entry
->iconv_descriptors
= NULL
;
617 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
618 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
620 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
621 register_finalizer_for_port (ret
);
627 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
629 return scm_c_make_port_with_encoding (tag
, mode_bits
,
630 scm_i_default_port_encoding (),
631 scm_i_get_conversion_strategy (SCM_BOOL_F
),
636 scm_new_port_table_entry (scm_t_bits tag
)
638 return scm_c_make_port (tag
, 0, 0);
645 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
647 "Return a boolean indicating whether @var{x} is a port.\n"
648 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
650 #define FUNC_NAME s_scm_port_p
652 return scm_from_bool (SCM_PORTP (x
));
656 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
658 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
659 "@code{#f}. Any object satisfying this predicate also satisfies\n"
661 #define FUNC_NAME s_scm_input_port_p
663 return scm_from_bool (SCM_INPUT_PORT_P (x
));
667 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
669 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
670 "@code{#f}. Any object satisfying this predicate also satisfies\n"
672 #define FUNC_NAME s_scm_output_port_p
674 x
= SCM_COERCE_OUTPORT (x
);
675 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
679 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
681 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
683 #define FUNC_NAME s_scm_port_closed_p
685 SCM_VALIDATE_PORT (1, port
);
686 return scm_from_bool (!SCM_OPPORTP (port
));
690 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
692 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
694 #define FUNC_NAME s_scm_eof_object_p
696 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
705 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
708 * Call the close operation on a port object.
709 * see also scm_close.
711 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
713 "Close the specified port object. Return @code{#t} if it\n"
714 "successfully closes a port or @code{#f} if it was already\n"
715 "closed. An exception may be raised if an error occurs, for\n"
716 "example when flushing buffered output. See also @ref{Ports and\n"
717 "File Descriptors, close}, for a procedure which can close file\n"
719 #define FUNC_NAME s_scm_close_port
724 port
= SCM_COERCE_OUTPORT (port
);
726 SCM_VALIDATE_PORT (1, port
);
727 if (SCM_CLOSEDP (port
))
729 if (SCM_PORT_DESCRIPTOR (port
)->close
)
730 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
734 p
= SCM_PTAB_ENTRY (port
);
736 scm_port_non_buffer (p
);
737 SCM_SETPTAB_ENTRY (port
, 0);
739 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
740 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
742 p
->putback_buf
= NULL
;
743 p
->putback_buf_size
= 0;
745 if (p
->iconv_descriptors
)
747 close_iconv_descriptors (p
->iconv_descriptors
);
748 p
->iconv_descriptors
= NULL
;
751 SCM_CLR_PORT_OPEN_FLAG (port
);
753 return scm_from_bool (rv
>= 0);
757 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
759 "Close the specified input port object. The routine has no effect if\n"
760 "the file has already been closed. An exception may be raised if an\n"
761 "error occurs. The value returned is unspecified.\n\n"
762 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
763 "which can close file descriptors.")
764 #define FUNC_NAME s_scm_close_input_port
766 SCM_VALIDATE_INPUT_PORT (1, port
);
767 scm_close_port (port
);
768 return SCM_UNSPECIFIED
;
772 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
774 "Close the specified output port object. The routine has no effect if\n"
775 "the file has already been closed. An exception may be raised if an\n"
776 "error occurs. The value returned is unspecified.\n\n"
777 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
778 "which can close file descriptors.")
779 #define FUNC_NAME s_scm_close_output_port
781 port
= SCM_COERCE_OUTPORT (port
);
782 SCM_VALIDATE_OUTPUT_PORT (1, port
);
783 scm_close_port (port
);
784 return SCM_UNSPECIFIED
;
791 /* Encoding characters to byte streams, and decoding byte streams to
794 /* A fluid specifying the default encoding for newly created ports. If it is
795 a string, that is the encoding. If it is #f, it is in the "native"
796 (Latin-1) encoding. */
797 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
799 static int scm_port_encoding_init
= 0;
801 /* Use ENCODING as the default encoding for future ports. */
803 scm_i_set_default_port_encoding (const char *encoding
)
805 if (!scm_port_encoding_init
806 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
807 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
811 || !strcmp (encoding
, "ASCII")
812 || !strcmp (encoding
, "ANSI_X3.4-1968")
813 || !strcmp (encoding
, "ISO-8859-1"))
814 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
816 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
817 scm_from_locale_string (encoding
));
820 /* Return the name of the default encoding for newly created ports; a
821 return value of NULL means "ISO-8859-1". */
823 scm_i_default_port_encoding (void)
825 if (!scm_port_encoding_init
)
827 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
833 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
834 if (!scm_is_string (encoding
))
837 return scm_i_string_chars (encoding
);
842 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
844 close_iconv_descriptors (ptr
);
847 static scm_t_iconv_descriptors
*
848 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
850 scm_t_iconv_descriptors
*id
;
851 iconv_t input_cd
, output_cd
;
853 input_cd
= (iconv_t
) -1;
854 output_cd
= (iconv_t
) -1;
858 /* Open an input iconv conversion descriptor, from ENCODING
859 to UTF-8. We choose UTF-8, not UTF-32, because iconv
860 implementations can typically convert from anything to
861 UTF-8, but not to UTF-32 (see
862 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
864 /* Assume opening an iconv descriptor causes about 16 KB of
866 scm_gc_register_allocation (16 * 1024);
868 input_cd
= iconv_open ("UTF-8", encoding
);
869 if (input_cd
== (iconv_t
) -1)
870 goto invalid_encoding
;
875 /* Assume opening an iconv descriptor causes about 16 KB of
877 scm_gc_register_allocation (16 * 1024);
879 output_cd
= iconv_open (encoding
, "UTF-8");
880 if (output_cd
== (iconv_t
) -1)
882 if (input_cd
!= (iconv_t
) -1)
883 iconv_close (input_cd
);
884 goto invalid_encoding
;
888 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
889 id
->input_cd
= input_cd
;
890 id
->output_cd
= output_cd
;
893 GC_finalization_proc prev_finalizer
;
894 GC_PTR prev_finalization_data
;
896 /* Register a finalizer to close the descriptors. */
897 GC_REGISTER_FINALIZER_NO_ORDER (id
, finalize_iconv_descriptors
, 0,
898 &prev_finalizer
, &prev_finalization_data
);
906 err
= scm_from_locale_string (encoding
);
907 scm_misc_error ("open_iconv_descriptors",
908 "invalid or unknown character encoding ~s",
914 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
916 if (id
->input_cd
!= (iconv_t
) -1)
917 iconv_close (id
->input_cd
);
918 if (id
->output_cd
!= (iconv_t
) -1)
919 iconv_close (id
->output_cd
);
920 id
->input_cd
= (void *) -1;
921 id
->output_cd
= (void *) -1;
924 scm_t_iconv_descriptors
*
925 scm_i_port_iconv_descriptors (SCM port
)
929 pt
= SCM_PTAB_ENTRY (port
);
931 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
933 if (!pt
->iconv_descriptors
)
936 pt
->encoding
= "ISO-8859-1";
937 pt
->iconv_descriptors
=
938 open_iconv_descriptors (pt
->encoding
,
939 SCM_INPUT_PORT_P (port
),
940 SCM_OUTPUT_PORT_P (port
));
943 return pt
->iconv_descriptors
;
947 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
950 scm_t_iconv_descriptors
*prev
;
952 /* Set the character encoding for this port. */
953 pt
= SCM_PTAB_ENTRY (port
);
954 prev
= pt
->iconv_descriptors
;
956 if (encoding
== NULL
)
957 encoding
= "ISO-8859-1";
959 if (strcmp (encoding
, "UTF-8") == 0)
961 pt
->encoding
= "UTF-8";
962 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
963 pt
->iconv_descriptors
= NULL
;
967 /* Open descriptors before mutating the port. */
968 pt
->iconv_descriptors
=
969 open_iconv_descriptors (encoding
,
970 SCM_INPUT_PORT_P (port
),
971 SCM_OUTPUT_PORT_P (port
));
972 pt
->encoding
= scm_gc_strdup (encoding
, "port");
973 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
977 close_iconv_descriptors (prev
);
980 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
982 "Returns, as a string, the character encoding that @var{port}\n"
983 "uses to interpret its input and output.\n")
984 #define FUNC_NAME s_scm_port_encoding
989 SCM_VALIDATE_PORT (1, port
);
991 pt
= SCM_PTAB_ENTRY (port
);
994 return scm_from_locale_string (pt
->encoding
);
1000 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1001 (SCM port
, SCM enc
),
1002 "Sets the character encoding that will be used to interpret all\n"
1003 "port I/O. New ports are created with the encoding\n"
1004 "appropriate for the current locale if @code{setlocale} has \n"
1005 "been called or ISO-8859-1 otherwise\n"
1006 "and this procedure can be used to modify that encoding.\n")
1007 #define FUNC_NAME s_scm_set_port_encoding_x
1011 SCM_VALIDATE_PORT (1, port
);
1012 SCM_VALIDATE_STRING (2, enc
);
1014 enc_str
= scm_to_locale_string (enc
);
1015 scm_i_set_port_encoding_x (port
, enc_str
);
1018 return SCM_UNSPECIFIED
;
1023 /* This determines how conversions handle unconvertible characters. */
1024 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
1025 static int scm_conversion_strategy_init
= 0;
1027 scm_t_string_failed_conversion_handler
1028 scm_i_get_conversion_strategy (SCM port
)
1032 if (scm_is_false (port
))
1034 if (!scm_conversion_strategy_init
1035 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1036 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1039 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
1040 if (scm_is_false (encoding
))
1041 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1043 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
1049 pt
= SCM_PTAB_ENTRY (port
);
1050 return pt
->ilseq_handler
;
1056 scm_i_set_conversion_strategy_x (SCM port
,
1057 scm_t_string_failed_conversion_handler handler
)
1062 strategy
= scm_from_int ((int) handler
);
1064 if (scm_is_false (port
))
1066 /* Set the default encoding for future ports. */
1067 if (!scm_conversion_strategy_init
1068 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1069 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1071 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1075 /* Set the character encoding for this port. */
1076 pt
= SCM_PTAB_ENTRY (port
);
1077 pt
->ilseq_handler
= handler
;
1081 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1082 1, 0, 0, (SCM port
),
1083 "Returns the behavior of the port when handling a character that\n"
1084 "is not representable in the port's current encoding.\n"
1085 "It returns the symbol @code{error} if unrepresentable characters\n"
1086 "should cause exceptions, @code{substitute} if the port should\n"
1087 "try to replace unrepresentable characters with question marks or\n"
1088 "approximate characters, or @code{escape} if unrepresentable\n"
1089 "characters should be converted to string escapes.\n"
1091 "If @var{port} is @code{#f}, then the current default behavior\n"
1092 "will be returned. New ports will have this default behavior\n"
1093 "when they are created.\n")
1094 #define FUNC_NAME s_scm_port_conversion_strategy
1096 scm_t_string_failed_conversion_handler h
;
1098 SCM_VALIDATE_OPPORT (1, port
);
1100 if (!scm_is_false (port
))
1102 SCM_VALIDATE_OPPORT (1, port
);
1105 h
= scm_i_get_conversion_strategy (port
);
1106 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1107 return scm_from_latin1_symbol ("error");
1108 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1109 return scm_from_latin1_symbol ("substitute");
1110 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1111 return scm_from_latin1_symbol ("escape");
1115 /* Never gets here. */
1116 return SCM_UNDEFINED
;
1120 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1122 (SCM port
, SCM sym
),
1123 "Sets the behavior of the interpreter when outputting a character\n"
1124 "that is not representable in the port's current encoding.\n"
1125 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1126 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1127 "when an unconvertible character is encountered. If it is\n"
1128 "@code{'substitute}, then unconvertible characters will \n"
1129 "be replaced with approximate characters, or with question marks\n"
1130 "if no approximately correct character is available.\n"
1131 "If it is @code{'escape},\n"
1132 "it will appear as a hex escape when output.\n"
1134 "If @var{port} is an open port, the conversion error behavior\n"
1135 "is set for that port. If it is @code{#f}, it is set as the\n"
1136 "default behavior for any future ports that get created in\n"
1138 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1144 if (!scm_is_false (port
))
1146 SCM_VALIDATE_OPPORT (1, port
);
1149 err
= scm_from_latin1_symbol ("error");
1150 if (scm_is_true (scm_eqv_p (sym
, err
)))
1152 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1153 return SCM_UNSPECIFIED
;
1156 qm
= scm_from_latin1_symbol ("substitute");
1157 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1159 scm_i_set_conversion_strategy_x (port
,
1160 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1161 return SCM_UNSPECIFIED
;
1164 esc
= scm_from_latin1_symbol ("escape");
1165 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1167 scm_i_set_conversion_strategy_x (port
,
1168 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1169 return SCM_UNSPECIFIED
;
1172 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1174 return SCM_UNSPECIFIED
;
1181 /* The port lock. */
1184 lock_port (void *mutex
)
1186 scm_i_pthread_mutex_lock (mutex
);
1190 unlock_port (void *mutex
)
1192 scm_i_pthread_mutex_unlock (mutex
);
1196 scm_dynwind_lock_port (SCM port
)
1197 #define FUNC_NAME "dynwind-lock-port"
1199 scm_i_pthread_mutex_t
*lock
;
1200 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1201 scm_c_lock_port (port
, &lock
);
1204 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1205 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1213 /* Revealed counts --- an oddity inherited from SCSH. */
1215 /* Find a port in the table and return its revealed count.
1216 Also used by the garbage collector.
1219 scm_revealed_count (SCM port
)
1221 scm_i_pthread_mutex_t
*lock
;
1224 scm_c_lock_port (port
, &lock
);
1225 ret
= SCM_REVEALED (port
);
1227 scm_i_pthread_mutex_unlock (lock
);
1232 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
1234 "Return the revealed count for @var{port}.")
1235 #define FUNC_NAME s_scm_port_revealed
1237 port
= SCM_COERCE_OUTPORT (port
);
1238 SCM_VALIDATE_OPENPORT (1, port
);
1239 return scm_from_int (scm_revealed_count (port
));
1243 /* Set the revealed count for a port. */
1244 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
1245 (SCM port
, SCM rcount
),
1246 "Sets the revealed count for a port to a given value.\n"
1247 "The return value is unspecified.")
1248 #define FUNC_NAME s_scm_set_port_revealed_x
1251 scm_i_pthread_mutex_t
*lock
;
1253 /* FIXME: It doesn't make sense to manipulate revealed counts on ports
1254 without a free function. */
1256 port
= SCM_COERCE_OUTPORT (port
);
1257 SCM_VALIDATE_OPENPORT (1, port
);
1258 r
= scm_to_int (rcount
);
1259 scm_c_lock_port (port
, &lock
);
1260 SCM_REVEALED (port
) = r
;
1262 scm_i_pthread_mutex_unlock (lock
);
1263 return SCM_UNSPECIFIED
;
1267 /* Set the revealed count for a port. */
1268 SCM_DEFINE (scm_adjust_port_revealed_x
, "adjust-port-revealed!", 2, 0, 0,
1269 (SCM port
, SCM addend
),
1270 "Add @var{addend} to the revealed count of @var{port}.\n"
1271 "The return value is unspecified.")
1272 #define FUNC_NAME s_scm_adjust_port_revealed_x
1274 scm_i_pthread_mutex_t
*lock
;
1276 port
= SCM_COERCE_OUTPORT (port
);
1277 SCM_VALIDATE_OPENPORT (1, port
);
1278 a
= scm_to_int (addend
);
1279 scm_c_lock_port (port
, &lock
);
1280 SCM_REVEALED (port
) += a
;
1282 scm_i_pthread_mutex_unlock (lock
);
1283 return SCM_UNSPECIFIED
;
1293 scm_get_byte_or_eof (SCM port
)
1295 scm_i_pthread_mutex_t
*lock
;
1298 scm_c_lock_port (port
, &lock
);
1299 ret
= scm_get_byte_or_eof_unlocked (port
);
1301 scm_i_pthread_mutex_unlock (lock
);
1307 scm_peek_byte_or_eof (SCM port
)
1309 scm_i_pthread_mutex_t
*lock
;
1312 scm_c_lock_port (port
, &lock
);
1313 ret
= scm_peek_byte_or_eof_unlocked (port
);
1315 scm_i_pthread_mutex_unlock (lock
);
1322 * Used by an application to read arbitrary number of bytes from an
1323 * SCM port. Same semantics as libc read, except that scm_c_read only
1324 * returns less than SIZE bytes if at end-of-file.
1326 * Warning: Doesn't update port line and column counts! */
1328 /* This structure, and the following swap_buffer function, are used
1329 for temporarily swapping a port's own read buffer, and the buffer
1330 that the caller of scm_c_read provides. */
1331 struct port_and_swap_buffer
1334 unsigned char *buffer
;
1339 swap_buffer (void *data
)
1341 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1342 unsigned char *old_buf
= psb
->pt
->read_buf
;
1343 size_t old_size
= psb
->pt
->read_buf_size
;
1345 /* Make the port use (buffer, size) from the struct. */
1346 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1347 psb
->pt
->read_buf_size
= psb
->size
;
1349 /* Save the port's old (buffer, size) in the struct. */
1350 psb
->buffer
= old_buf
;
1351 psb
->size
= old_size
;
1355 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1356 #define FUNC_NAME "scm_c_read"
1359 size_t n_read
= 0, n_available
;
1360 struct port_and_swap_buffer psb
;
1362 SCM_VALIDATE_OPINPORT (1, port
);
1364 pt
= SCM_PTAB_ENTRY (port
);
1365 if (pt
->rw_active
== SCM_PORT_WRITE
)
1366 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1369 pt
->rw_active
= SCM_PORT_READ
;
1371 /* Take bytes first from the port's read buffer. */
1372 if (pt
->read_pos
< pt
->read_end
)
1374 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1375 memcpy (buffer
, pt
->read_pos
, n_available
);
1376 buffer
= (char *) buffer
+ n_available
;
1377 pt
->read_pos
+= n_available
;
1378 n_read
+= n_available
;
1379 size
-= n_available
;
1382 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1386 /* Now we will call scm_fill_input repeatedly until we have read the
1387 requested number of bytes. (Note that a single scm_fill_input
1388 call does not guarantee to fill the whole of the port's read
1390 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1392 /* The port that we are reading from is unbuffered - i.e. does
1393 not have its own persistent buffer - but we have a buffer,
1394 provided by our caller, that is the right size for the data
1395 that is wanted. For the following scm_fill_input calls,
1396 therefore, we use the buffer in hand as the port's read
1399 We need to make sure that the port's normal (1 byte) buffer
1400 is reinstated in case one of the scm_fill_input () calls
1401 throws an exception; we use the scm_dynwind_* API to achieve
1404 A consequence of this optimization is that the fill_input
1405 functions can't unget characters. That'll push data to the
1406 pushback buffer instead of this psb buffer. */
1408 unsigned char *pback
= pt
->putback_buf
;
1411 psb
.buffer
= buffer
;
1413 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1414 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1415 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1417 /* Call scm_fill_input until we have all the bytes that we need,
1419 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1421 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1422 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1425 if (pback
!= pt
->putback_buf
1426 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1427 scm_misc_error (FUNC_NAME
,
1428 "scm_c_read must not call a fill function that pushes "
1429 "back characters onto an unbuffered port", SCM_EOL
);
1431 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1433 /* Reinstate the port's normal buffer. */
1438 /* The port has its own buffer. It is important that we use it,
1439 even if it happens to be smaller than our caller's buffer, so
1440 that a custom port implementation's entry points (in
1441 particular, fill_input) can rely on the buffer always being
1442 the same as they first set up. */
1443 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1445 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1446 memcpy (buffer
, pt
->read_pos
, n_available
);
1447 buffer
= (char *) buffer
+ n_available
;
1448 pt
->read_pos
+= n_available
;
1449 n_read
+= n_available
;
1450 size
-= n_available
;
1459 scm_c_read (SCM port
, void *buffer
, size_t size
)
1461 scm_i_pthread_mutex_t
*lock
;
1464 scm_c_lock_port (port
, &lock
);
1465 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1467 scm_i_pthread_mutex_unlock (lock
);
1473 /* Update the line and column number of PORT after consumption of C. */
1475 update_port_lf (scm_t_wchar c
, SCM port
)
1500 #define SCM_MBCHAR_BUF_SIZE (4)
1502 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1503 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1505 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1507 scm_t_wchar codepoint
;
1509 if (utf8_buf
[0] <= 0x7f)
1512 codepoint
= utf8_buf
[0];
1514 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1517 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1518 | (utf8_buf
[1] & 0x3f);
1520 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1523 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1524 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1525 | (utf8_buf
[2] & 0x3f);
1530 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1531 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1532 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1533 | (utf8_buf
[3] & 0x3f);
1539 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1540 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1541 representation, and set *LEN to the length in bytes. Return
1542 `EILSEQ' on error. */
1544 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1545 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1547 #define ASSERT_NOT_EOF(b) \
1548 if (SCM_UNLIKELY ((b) == EOF)) \
1550 #define CONSUME_PEEKED_BYTE() \
1557 pt
= SCM_PTAB_ENTRY (port
);
1559 byte
= scm_get_byte_or_eof_unlocked (port
);
1566 buf
[0] = (scm_t_uint8
) byte
;
1571 *codepoint
= buf
[0];
1572 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1575 byte
= scm_peek_byte_or_eof_unlocked (port
);
1576 ASSERT_NOT_EOF (byte
);
1578 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1581 CONSUME_PEEKED_BYTE ();
1582 buf
[1] = (scm_t_uint8
) byte
;
1585 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1588 else if ((buf
[0] & 0xf0) == 0xe0)
1591 byte
= scm_peek_byte_or_eof_unlocked (port
);
1592 ASSERT_NOT_EOF (byte
);
1594 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1595 || (buf
[0] == 0xe0 && byte
< 0xa0)
1596 || (buf
[0] == 0xed && byte
> 0x9f)))
1599 CONSUME_PEEKED_BYTE ();
1600 buf
[1] = (scm_t_uint8
) byte
;
1603 byte
= scm_peek_byte_or_eof_unlocked (port
);
1604 ASSERT_NOT_EOF (byte
);
1606 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1609 CONSUME_PEEKED_BYTE ();
1610 buf
[2] = (scm_t_uint8
) byte
;
1613 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1614 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1617 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1620 byte
= scm_peek_byte_or_eof_unlocked (port
);
1621 ASSERT_NOT_EOF (byte
);
1623 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1624 || (buf
[0] == 0xf0 && byte
< 0x90)
1625 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1628 CONSUME_PEEKED_BYTE ();
1629 buf
[1] = (scm_t_uint8
) byte
;
1632 byte
= scm_peek_byte_or_eof_unlocked (port
);
1633 ASSERT_NOT_EOF (byte
);
1635 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1638 CONSUME_PEEKED_BYTE ();
1639 buf
[2] = (scm_t_uint8
) byte
;
1642 byte
= scm_peek_byte_or_eof_unlocked (port
);
1643 ASSERT_NOT_EOF (byte
);
1645 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1648 CONSUME_PEEKED_BYTE ();
1649 buf
[3] = (scm_t_uint8
) byte
;
1652 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1653 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1654 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1663 /* Here we could choose the consume the faulty byte when it's not a
1664 valid starting byte, but it's not a requirement. What Section 3.9
1665 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1666 would otherwise be a valid starting byte. */
1670 #undef CONSUME_PEEKED_BYTE
1671 #undef ASSERT_NOT_EOF
1674 /* Likewise, read a byte sequence from PORT, passing it through its
1675 input conversion descriptor. */
1677 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1678 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1680 scm_t_iconv_descriptors
*id
;
1682 size_t bytes_consumed
, output_size
;
1684 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1686 id
= scm_i_port_iconv_descriptors (port
);
1688 for (output_size
= 0, output
= (char *) utf8_buf
,
1689 bytes_consumed
= 0, err
= 0;
1690 err
== 0 && output_size
== 0
1691 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1695 size_t input_left
, output_left
, done
;
1697 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1698 if (byte_read
== EOF
)
1700 if (bytes_consumed
== 0)
1702 *codepoint
= (scm_t_wchar
) EOF
;
1710 buf
[bytes_consumed
] = byte_read
;
1713 input_left
= bytes_consumed
+ 1;
1714 output_left
= sizeof (utf8_buf
);
1716 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1717 if (done
== (size_t) -1)
1721 /* Missing input: keep trying. */
1725 output_size
= sizeof (utf8_buf
) - output_left
;
1728 if (SCM_UNLIKELY (output_size
== 0))
1729 /* An unterminated sequence. */
1731 else if (SCM_LIKELY (err
== 0))
1733 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1734 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1735 *len
= bytes_consumed
;
1741 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1742 with the byte representation of the codepoint in PORT's encoding, and
1743 set *LEN to the length in bytes of that representation. Return 0 on
1744 success and an errno value on error. */
1746 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1747 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1750 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1752 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1753 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1755 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1757 if (SCM_LIKELY (err
== 0))
1758 update_port_lf (*codepoint
, port
);
1759 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1763 update_port_lf (*codepoint
, port
);
1769 /* Read a codepoint from PORT and return it. */
1771 scm_getc_unlocked (SCM port
)
1772 #define FUNC_NAME "scm_getc"
1776 scm_t_wchar codepoint
;
1777 char buf
[SCM_MBCHAR_BUF_SIZE
];
1779 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1780 if (SCM_UNLIKELY (err
!= 0))
1781 /* At this point PORT should point past the invalid encoding, as per
1782 R6RS-lib Section 8.2.4. */
1783 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1792 scm_i_pthread_mutex_t
*lock
;
1795 scm_c_lock_port (port
, &lock
);
1796 ret
= scm_getc_unlocked (port
);
1798 scm_i_pthread_mutex_unlock (lock
);
1804 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1806 "Return the next character available from @var{port}, updating\n"
1807 "@var{port} to point to the following character. If no more\n"
1808 "characters are available, the end-of-file object is returned.\n"
1810 "When @var{port}'s data cannot be decoded according to its\n"
1811 "character encoding, a @code{decoding-error} is raised and\n"
1812 "@var{port} points past the erroneous byte sequence.\n")
1813 #define FUNC_NAME s_scm_read_char
1816 if (SCM_UNBNDP (port
))
1817 port
= scm_current_input_port ();
1818 SCM_VALIDATE_OPINPORT (1, port
);
1819 c
= scm_getc_unlocked (port
);
1822 return SCM_MAKE_CHAR (c
);
1832 scm_unget_byte_unlocked (int c
, SCM port
)
1833 #define FUNC_NAME "scm_unget_byte"
1835 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1837 if (pt
->read_buf
== pt
->putback_buf
)
1838 /* already using the put-back buffer. */
1840 /* enlarge putback_buf if necessary. */
1841 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1842 && pt
->read_buf
== pt
->read_pos
)
1844 size_t new_size
= pt
->read_buf_size
* 2;
1845 unsigned char *tmp
= (unsigned char *)
1846 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1849 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1850 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1851 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1854 /* shift any existing bytes to buffer + 1. */
1855 if (pt
->read_pos
== pt
->read_end
)
1856 pt
->read_end
= pt
->read_buf
+ 1;
1857 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1859 int count
= pt
->read_end
- pt
->read_pos
;
1861 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1862 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1865 pt
->read_pos
= pt
->read_buf
;
1868 /* switch to the put-back buffer. */
1870 if (pt
->putback_buf
== NULL
)
1873 = (unsigned char *) scm_gc_malloc_pointerless
1874 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1875 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1878 pt
->saved_read_buf
= pt
->read_buf
;
1879 pt
->saved_read_pos
= pt
->read_pos
;
1880 pt
->saved_read_end
= pt
->read_end
;
1881 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1883 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1884 pt
->read_end
= pt
->read_buf
+ 1;
1885 pt
->read_buf_size
= pt
->putback_buf_size
;
1891 pt
->rw_active
= SCM_PORT_READ
;
1896 scm_unget_byte (int c
, SCM port
)
1898 scm_i_pthread_mutex_t
*lock
;
1899 scm_c_lock_port (port
, &lock
);
1900 scm_unget_byte_unlocked (c
, port
);
1902 scm_i_pthread_mutex_unlock (lock
);
1907 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1908 #define FUNC_NAME "scm_ungetc"
1910 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1912 char result_buf
[10];
1913 const char *encoding
;
1917 if (pt
->encoding
!= NULL
)
1918 encoding
= pt
->encoding
;
1920 encoding
= "ISO-8859-1";
1922 len
= sizeof (result_buf
);
1923 result
= u32_conv_to_encoding (encoding
,
1924 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1925 (uint32_t *) &c
, 1, NULL
,
1928 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1929 scm_encoding_error (FUNC_NAME
, errno
,
1930 "conversion to port encoding failed",
1931 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1933 for (i
= len
- 1; i
>= 0; i
--)
1934 scm_unget_byte_unlocked (result
[i
], port
);
1936 if (SCM_UNLIKELY (result
!= result_buf
))
1941 /* What should col be in this case?
1942 * We'll leave it at -1.
1944 SCM_LINUM (port
) -= 1;
1952 scm_ungetc (scm_t_wchar c
, SCM port
)
1954 scm_i_pthread_mutex_t
*lock
;
1955 scm_c_lock_port (port
, &lock
);
1956 scm_ungetc_unlocked (c
, port
);
1958 scm_i_pthread_mutex_unlock (lock
);
1963 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1965 /* This is simple minded and inefficient, but unreading strings is
1966 * probably not a common operation, and remember that line and
1967 * column numbers have to be handled...
1969 * Please feel free to write an optimized version!
1972 scm_ungetc_unlocked (s
[n
], port
);
1976 scm_ungets (const char *s
, int n
, SCM port
)
1978 scm_i_pthread_mutex_t
*lock
;
1979 scm_c_lock_port (port
, &lock
);
1980 scm_ungets_unlocked (s
, n
, port
);
1982 scm_i_pthread_mutex_unlock (lock
);
1986 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1988 "Return the next character available from @var{port},\n"
1989 "@emph{without} updating @var{port} to point to the following\n"
1990 "character. If no more characters are available, the\n"
1991 "end-of-file object is returned.\n"
1993 "The value returned by\n"
1994 "a call to @code{peek-char} is the same as the value that would\n"
1995 "have been returned by a call to @code{read-char} on the same\n"
1996 "port. The only difference is that the very next call to\n"
1997 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1998 "return the value returned by the preceding call to\n"
1999 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
2000 "an interactive port will hang waiting for input whenever a call\n"
2001 "to @code{read-char} would have hung.\n"
2003 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
2004 "if such a situation occurs. However, unlike with @code{read-char},\n"
2005 "@var{port} still points at the beginning of the erroneous byte\n"
2006 "sequence when the error is raised.\n")
2007 #define FUNC_NAME s_scm_peek_char
2012 char bytes
[SCM_MBCHAR_BUF_SIZE
];
2013 long column
, line
, i
;
2016 if (SCM_UNBNDP (port
))
2017 port
= scm_current_input_port ();
2018 SCM_VALIDATE_OPINPORT (1, port
);
2020 column
= SCM_COL (port
);
2021 line
= SCM_LINUM (port
);
2023 err
= get_codepoint (port
, &c
, bytes
, &len
);
2025 for (i
= len
- 1; i
>= 0; i
--)
2026 scm_unget_byte_unlocked (bytes
[i
], port
);
2028 SCM_COL (port
) = column
;
2029 SCM_LINUM (port
) = line
;
2031 if (SCM_UNLIKELY (err
!= 0))
2033 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
2035 /* Shouldn't happen since `catch' always aborts to prompt. */
2036 result
= SCM_BOOL_F
;
2039 result
= SCM_EOF_VAL
;
2041 result
= SCM_MAKE_CHAR (c
);
2047 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2048 (SCM cobj
, SCM port
),
2049 "Place character @var{cobj} in @var{port} so that it will be\n"
2050 "read by the next read operation. If called multiple times, the\n"
2051 "unread characters will be read again in last-in first-out\n"
2052 "order. If @var{port} is not supplied, the current input port\n"
2054 #define FUNC_NAME s_scm_unread_char
2058 SCM_VALIDATE_CHAR (1, cobj
);
2059 if (SCM_UNBNDP (port
))
2060 port
= scm_current_input_port ();
2061 SCM_VALIDATE_OPINPORT (2, port
);
2063 c
= SCM_CHAR (cobj
);
2065 scm_ungetc_unlocked (c
, port
);
2070 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2071 (SCM str
, SCM port
),
2072 "Place the string @var{str} in @var{port} so that its characters will be\n"
2073 "read in subsequent read operations. If called multiple times, the\n"
2074 "unread characters will be read again in last-in first-out order. If\n"
2075 "@var{port} is not supplied, the current-input-port is used.")
2076 #define FUNC_NAME s_scm_unread_string
2079 SCM_VALIDATE_STRING (1, str
);
2080 if (SCM_UNBNDP (port
))
2081 port
= scm_current_input_port ();
2082 SCM_VALIDATE_OPINPORT (2, port
);
2084 n
= scm_i_string_length (str
);
2087 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2096 /* Manipulating the buffers. */
2098 /* This routine does not take any locks, as it is usually called as part
2099 of a port implementation. */
2101 scm_port_non_buffer (scm_t_port
*pt
)
2103 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2104 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2105 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2106 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2109 /* this should only be called when the read buffer is empty. it
2110 tries to refill the read buffer. it returns the first char from
2111 the port, which is either EOF or *(pt->read_pos). */
2113 scm_fill_input_unlocked (SCM port
)
2115 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2117 assert (pt
->read_pos
== pt
->read_end
);
2119 if (pt
->read_buf
== pt
->putback_buf
)
2121 /* finished reading put-back chars. */
2122 pt
->read_buf
= pt
->saved_read_buf
;
2123 pt
->read_pos
= pt
->saved_read_pos
;
2124 pt
->read_end
= pt
->saved_read_end
;
2125 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2126 if (pt
->read_pos
< pt
->read_end
)
2127 return *(pt
->read_pos
);
2129 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2133 scm_fill_input (SCM port
)
2135 scm_i_pthread_mutex_t
*lock
;
2138 scm_c_lock_port (port
, &lock
);
2139 ret
= scm_fill_input_unlocked (port
);
2141 scm_i_pthread_mutex_unlock (lock
);
2147 /* move up to read_len chars from port's putback and/or read buffers
2148 into memory starting at dest. returns the number of chars moved. */
2150 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2152 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2153 size_t chars_read
= 0;
2154 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2158 memcpy (dest
, pt
->read_pos
, from_buf
);
2159 pt
->read_pos
+= from_buf
;
2160 chars_read
+= from_buf
;
2161 read_len
-= from_buf
;
2165 /* if putback was active, try the real input buffer too. */
2166 if (pt
->read_buf
== pt
->putback_buf
)
2168 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2171 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2172 pt
->saved_read_pos
+= from_buf
;
2173 chars_read
+= from_buf
;
2179 /* Clear a port's read buffers, returning the contents. */
2180 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2182 "This procedure clears a port's input buffers, similar\n"
2183 "to the way that force-output clears the output buffer. The\n"
2184 "contents of the buffers are returned as a single string, e.g.,\n"
2187 "(define p (open-input-file ...))\n"
2188 "(drain-input p) => empty string, nothing buffered yet.\n"
2189 "(unread-char (read-char p) p)\n"
2190 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2192 "Draining the buffers may be useful for cleanly finishing\n"
2193 "buffered I/O so that the file descriptor can be used directly\n"
2194 "for further input.")
2195 #define FUNC_NAME s_scm_drain_input
2202 SCM_VALIDATE_OPINPORT (1, port
);
2203 pt
= SCM_PTAB_ENTRY (port
);
2205 count
= pt
->read_end
- pt
->read_pos
;
2206 if (pt
->read_buf
== pt
->putback_buf
)
2207 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2211 result
= scm_i_make_string (count
, &data
, 0);
2212 scm_take_from_input_buffers (port
, data
, count
);
2215 result
= scm_nullstr
;
2222 scm_end_input_unlocked (SCM port
)
2225 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2227 if (pt
->read_buf
== pt
->putback_buf
)
2229 offset
= pt
->read_end
- pt
->read_pos
;
2230 pt
->read_buf
= pt
->saved_read_buf
;
2231 pt
->read_pos
= pt
->saved_read_pos
;
2232 pt
->read_end
= pt
->saved_read_end
;
2233 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2238 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2242 scm_end_input (SCM port
)
2244 scm_i_pthread_mutex_t
*lock
;
2245 scm_c_lock_port (port
, &lock
);
2246 scm_end_input_unlocked (port
);
2248 scm_i_pthread_mutex_unlock (lock
);
2252 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2254 "Flush the specified output port, or the current output port if @var{port}\n"
2255 "is omitted. The current output buffer contents are passed to the\n"
2256 "underlying port implementation (e.g., in the case of fports, the\n"
2257 "data will be written to the file and the output buffer will be cleared.)\n"
2258 "It has no effect on an unbuffered port.\n\n"
2259 "The return value is unspecified.")
2260 #define FUNC_NAME s_scm_force_output
2262 if (SCM_UNBNDP (port
))
2263 port
= scm_current_output_port ();
2266 port
= SCM_COERCE_OUTPORT (port
);
2267 SCM_VALIDATE_OPOUTPORT (1, port
);
2269 scm_flush_unlocked (port
);
2270 return SCM_UNSPECIFIED
;
2275 scm_flush_unlocked (SCM port
)
2277 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2281 scm_flush (SCM port
)
2283 scm_i_pthread_mutex_t
*lock
;
2284 scm_c_lock_port (port
, &lock
);
2285 scm_flush_unlocked (port
);
2287 scm_i_pthread_mutex_unlock (lock
);
2297 scm_putc (char c
, SCM port
)
2299 scm_i_pthread_mutex_t
*lock
;
2300 scm_c_lock_port (port
, &lock
);
2301 scm_putc_unlocked (c
, port
);
2303 scm_i_pthread_mutex_unlock (lock
);
2308 scm_puts (const char *s
, SCM port
)
2310 scm_i_pthread_mutex_t
*lock
;
2311 scm_c_lock_port (port
, &lock
);
2312 scm_puts_unlocked (s
, port
);
2314 scm_i_pthread_mutex_unlock (lock
);
2320 * Used by an application to write arbitrary number of bytes to an SCM
2321 * port. Similar semantics as libc write. However, unlike libc
2322 * write, scm_c_write writes the requested number of bytes and has no
2325 * Warning: Doesn't update port line and column counts!
2328 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2329 #define FUNC_NAME "scm_c_write"
2332 scm_t_ptob_descriptor
*ptob
;
2334 SCM_VALIDATE_OPOUTPORT (1, port
);
2336 pt
= SCM_PTAB_ENTRY (port
);
2337 ptob
= SCM_PORT_DESCRIPTOR (port
);
2339 if (pt
->rw_active
== SCM_PORT_READ
)
2340 scm_end_input_unlocked (port
);
2342 ptob
->write (port
, ptr
, size
);
2345 pt
->rw_active
= SCM_PORT_WRITE
;
2350 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2352 scm_i_pthread_mutex_t
*lock
;
2353 scm_c_lock_port (port
, &lock
);
2354 scm_c_write_unlocked (port
, ptr
, size
);
2356 scm_i_pthread_mutex_unlock (lock
);
2362 * This function differs from scm_c_write; it updates port line and
2365 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2367 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2368 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2370 if (pt
->rw_active
== SCM_PORT_READ
)
2371 scm_end_input_unlocked (port
);
2373 ptob
->write (port
, ptr
, size
);
2375 for (; size
; ptr
++, size
--)
2376 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2379 pt
->rw_active
= SCM_PORT_WRITE
;
2383 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2385 scm_i_pthread_mutex_t
*lock
;
2386 scm_c_lock_port (port
, &lock
);
2387 scm_lfwrite_unlocked (ptr
, size
, port
);
2389 scm_i_pthread_mutex_unlock (lock
);
2393 /* Write STR to PORT from START inclusive to END exclusive. */
2395 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2397 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2399 if (pt
->rw_active
== SCM_PORT_READ
)
2400 scm_end_input_unlocked (port
);
2402 if (end
== (size_t) -1)
2403 end
= scm_i_string_length (str
);
2405 scm_display (scm_c_substring (str
, start
, end
), port
);
2408 pt
->rw_active
= SCM_PORT_WRITE
;
2414 /* Querying and setting positions, and character availability. */
2416 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2418 "Return @code{#t} if a character is ready on input @var{port}\n"
2419 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2420 "@code{#t} then the next @code{read-char} operation on\n"
2421 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2422 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2424 "@code{char-ready?} exists to make it possible for a\n"
2425 "program to accept characters from interactive ports without\n"
2426 "getting stuck waiting for input. Any input editors associated\n"
2427 "with such ports must make sure that characters whose existence\n"
2428 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2429 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2430 "a port at end of file would be indistinguishable from an\n"
2431 "interactive port that has no ready characters.")
2432 #define FUNC_NAME s_scm_char_ready_p
2436 if (SCM_UNBNDP (port
))
2437 port
= scm_current_input_port ();
2438 /* It's possible to close the current input port, so validate even in
2440 SCM_VALIDATE_OPINPORT (1, port
);
2442 pt
= SCM_PTAB_ENTRY (port
);
2444 /* if the current read buffer is filled, or the
2445 last pushed-back char has been read and the saved buffer is
2446 filled, result is true. */
2447 if (pt
->read_pos
< pt
->read_end
2448 || (pt
->read_buf
== pt
->putback_buf
2449 && pt
->saved_read_pos
< pt
->saved_read_end
))
2453 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2455 if (ptob
->input_waiting
)
2456 return scm_from_bool(ptob
->input_waiting (port
));
2463 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2464 (SCM fd_port
, SCM offset
, SCM whence
),
2465 "Sets the current position of @var{fd_port} to the integer\n"
2466 "@var{offset}, which is interpreted according to the value of\n"
2469 "One of the following variables should be supplied for\n"
2471 "@defvar SEEK_SET\n"
2472 "Seek from the beginning of the file.\n"
2474 "@defvar SEEK_CUR\n"
2475 "Seek from the current position.\n"
2477 "@defvar SEEK_END\n"
2478 "Seek from the end of the file.\n"
2480 "If @var{fd_port} is a file descriptor, the underlying system\n"
2481 "call is @code{lseek}. @var{port} may be a string port.\n"
2483 "The value returned is the new position in the file. This means\n"
2484 "that the current position of a port can be obtained using:\n"
2486 "(seek port 0 SEEK_CUR)\n"
2488 #define FUNC_NAME s_scm_seek
2492 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2494 how
= scm_to_int (whence
);
2495 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2496 SCM_OUT_OF_RANGE (3, whence
);
2498 if (SCM_OPPORTP (fd_port
))
2500 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2501 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2502 off_t_or_off64_t rv
;
2505 SCM_MISC_ERROR ("port is not seekable",
2506 scm_cons (fd_port
, SCM_EOL
));
2508 rv
= ptob
->seek (fd_port
, off
, how
);
2509 return scm_from_off_t_or_off64_t (rv
);
2511 else /* file descriptor?. */
2513 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2514 off_t_or_off64_t rv
;
2515 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2518 return scm_from_off_t_or_off64_t (rv
);
2527 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2528 doesn't have the filename version truncate(), hence this code. */
2529 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2531 truncate (const char *file
, off_t length
)
2535 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2539 ret
= ftruncate (fdes
, length
);
2542 int save_errno
= errno
;
2548 return close (fdes
);
2550 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2552 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2553 (SCM object
, SCM length
),
2554 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2555 "can be a filename string, a port object, or an integer file\n"
2557 "The return value is unspecified.\n"
2559 "For a port or file descriptor @var{length} can be omitted, in\n"
2560 "which case the file is truncated at the current position (per\n"
2561 "@code{ftell} above).\n"
2563 "On most systems a file can be extended by giving a length\n"
2564 "greater than the current size, but this is not mandatory in the\n"
2566 #define FUNC_NAME s_scm_truncate_file
2570 /* "object" can be a port, fdes or filename.
2572 Negative "length" makes no sense, but it's left to truncate() or
2573 ftruncate() to give back an error for that (normally EINVAL).
2576 if (SCM_UNBNDP (length
))
2578 /* must supply length if object is a filename. */
2579 if (scm_is_string (object
))
2580 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2582 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2585 object
= SCM_COERCE_OUTPORT (object
);
2586 if (scm_is_integer (object
))
2588 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2589 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2592 else if (SCM_OPOUTPORTP (object
))
2594 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2595 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2596 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2598 if (!ptob
->truncate
)
2599 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2600 if (pt
->rw_active
== SCM_PORT_READ
)
2601 scm_end_input_unlocked (object
);
2602 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2603 ptob
->flush (object
);
2605 ptob
->truncate (object
, c_length
);
2610 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2611 char *str
= scm_to_locale_string (object
);
2613 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2620 return SCM_UNSPECIFIED
;
2624 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2626 "Return the current line number for @var{port}.\n"
2628 "The first line of a file is 0. But you might want to add 1\n"
2629 "when printing line numbers, since starting from 1 is\n"
2630 "traditional in error messages, and likely to be more natural to\n"
2632 #define FUNC_NAME s_scm_port_line
2634 port
= SCM_COERCE_OUTPORT (port
);
2635 SCM_VALIDATE_OPENPORT (1, port
);
2636 return scm_from_long (SCM_LINUM (port
));
2640 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2641 (SCM port
, SCM line
),
2642 "Set the current line number for @var{port} to @var{line}. The\n"
2643 "first line of a file is 0.")
2644 #define FUNC_NAME s_scm_set_port_line_x
2646 port
= SCM_COERCE_OUTPORT (port
);
2647 SCM_VALIDATE_OPENPORT (1, port
);
2648 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2649 return SCM_UNSPECIFIED
;
2653 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2655 "Return the current column number of @var{port}.\n"
2656 "If the number is\n"
2657 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2658 "- i.e. the first character of the first line is line 0, column 0.\n"
2659 "(However, when you display a file position, for example in an error\n"
2660 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2661 "because lines and column numbers traditionally start with 1, and that is\n"
2662 "what non-programmers will find most natural.)")
2663 #define FUNC_NAME s_scm_port_column
2665 port
= SCM_COERCE_OUTPORT (port
);
2666 SCM_VALIDATE_OPENPORT (1, port
);
2667 return scm_from_int (SCM_COL (port
));
2671 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2672 (SCM port
, SCM column
),
2673 "Set the current column of @var{port}. Before reading the first\n"
2674 "character on a line the column should be 0.")
2675 #define FUNC_NAME s_scm_set_port_column_x
2677 port
= SCM_COERCE_OUTPORT (port
);
2678 SCM_VALIDATE_OPENPORT (1, port
);
2679 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2680 return SCM_UNSPECIFIED
;
2684 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2686 "Return the filename associated with @var{port}, or @code{#f}\n"
2687 "if no filename is associated with the port.")
2688 #define FUNC_NAME s_scm_port_filename
2690 port
= SCM_COERCE_OUTPORT (port
);
2691 SCM_VALIDATE_OPENPORT (1, port
);
2692 return SCM_FILENAME (port
);
2696 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2697 (SCM port
, SCM filename
),
2698 "Change the filename associated with @var{port}, using the current input\n"
2699 "port if none is specified. Note that this does not change the port's\n"
2700 "source of data, but only the value that is returned by\n"
2701 "@code{port-filename} and reported in diagnostic output.")
2702 #define FUNC_NAME s_scm_set_port_filename_x
2704 port
= SCM_COERCE_OUTPORT (port
);
2705 SCM_VALIDATE_OPENPORT (1, port
);
2706 /* We allow the user to set the filename to whatever he likes. */
2707 SCM_SET_FILENAME (port
, filename
);
2708 return SCM_UNSPECIFIED
;
2715 /* Implementation helpers for port printing functions. */
2718 scm_print_port_mode (SCM exp
, SCM port
)
2720 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2722 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2723 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2726 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2733 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2735 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2738 scm_puts_unlocked ("#<", port
);
2739 scm_print_port_mode (exp
, port
);
2740 scm_puts_unlocked (type
, port
);
2741 scm_putc_unlocked (' ', port
);
2742 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2743 scm_putc_unlocked ('>', port
);
2750 /* Iterating over all ports. */
2752 struct for_each_data
2754 void (*proc
) (void *data
, SCM p
);
2759 for_each_trampoline (void *data
, SCM port
, SCM result
)
2761 struct for_each_data
*d
= data
;
2763 d
->proc (d
->data
, port
);
2769 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2771 struct for_each_data d
;
2776 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2777 scm_i_port_weak_set
);
2781 scm_for_each_trampoline (void *data
, SCM port
)
2783 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2786 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2788 "Apply @var{proc} to each port in the Guile port table\n"
2789 "in turn. The return value is unspecified. More specifically,\n"
2790 "@var{proc} is applied exactly once to every port that exists\n"
2791 "in the system at the time @code{port-for-each} is invoked.\n"
2792 "Changes to the port table while @code{port-for-each} is running\n"
2793 "have no effect as far as @code{port-for-each} is concerned.")
2794 #define FUNC_NAME s_scm_port_for_each
2796 SCM_VALIDATE_PROC (1, proc
);
2798 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2800 return SCM_UNSPECIFIED
;
2805 flush_output_port (void *closure
, SCM port
)
2807 if (SCM_OPOUTPORTP (port
))
2808 scm_flush_unlocked (port
);
2811 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2813 "Equivalent to calling @code{force-output} on\n"
2814 "all open output ports. The return value is unspecified.")
2815 #define FUNC_NAME s_scm_flush_all_ports
2817 scm_c_port_for_each (&flush_output_port
, NULL
);
2818 return SCM_UNSPECIFIED
;
2827 scm_t_bits scm_tc16_void_port
= 0;
2829 static int fill_input_void_port (SCM port SCM_UNUSED
)
2835 write_void_port (SCM port SCM_UNUSED
,
2836 const void *data SCM_UNUSED
,
2837 size_t size SCM_UNUSED
)
2842 scm_i_void_port (long mode_bits
)
2846 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2848 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2854 scm_void_port (char *mode_str
)
2856 return scm_i_void_port (scm_mode_bits (mode_str
));
2859 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2861 "Create and return a new void port. A void port acts like\n"
2862 "@file{/dev/null}. The @var{mode} argument\n"
2863 "specifies the input/output modes for this port: see the\n"
2864 "documentation for @code{open-file} in @ref{File Ports}.")
2865 #define FUNC_NAME s_scm_sys_make_void_port
2867 return scm_i_void_port (scm_i_mode_bits (mode
));
2874 /* Initialization. */
2879 /* lseek() symbols. */
2880 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2881 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2882 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2884 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2887 cur_inport_fluid
= scm_make_fluid ();
2888 cur_outport_fluid
= scm_make_fluid ();
2889 cur_errport_fluid
= scm_make_fluid ();
2890 cur_loadport_fluid
= scm_make_fluid ();
2892 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2894 #include "libguile/ports.x"
2896 /* Use Latin-1 as the default port encoding. */
2897 SCM_VARIABLE_SET (default_port_encoding_var
,
2898 scm_make_fluid_with_default (SCM_BOOL_F
));
2899 scm_port_encoding_init
= 1;
2901 SCM_VARIABLE_SET (scm_conversion_strategy
,
2902 scm_make_fluid_with_default
2903 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2904 scm_conversion_strategy_init
= 1;
2906 /* These bindings are used when boot-9 turns `current-input-port' et
2907 al into parameters. They are then removed from the guile module. */
2908 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2909 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2910 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);