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
);
553 scm_t_ptob_descriptor
*ptob
;
558 do_free (void *body_data
)
560 struct do_free_data
*data
= body_data
;
562 /* `close' is for explicit `close-port' by user. `free' is for this
563 purpose: ports collected by the GC. */
564 data
->ptob
->free (data
->port
);
569 /* Finalize the object (a port) pointed to by PTR. */
571 finalize_port (GC_PTR ptr
, GC_PTR data
)
573 SCM port
= SCM_PACK_POINTER (ptr
);
575 if (!SCM_PORTP (port
))
578 if (SCM_OPENP (port
))
580 if (SCM_REVEALED (port
) > 0)
581 /* Keep "revealed" ports alive and re-register a finalizer. */
582 register_finalizer_for_port (port
);
585 struct do_free_data data
;
587 SCM_CLR_PORT_OPEN_FLAG (port
);
589 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
592 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
593 scm_handle_by_message_noexit
, NULL
);
595 scm_gc_ports_collected
++;
604 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
605 const char *encoding
,
606 scm_t_string_failed_conversion_handler handler
,
611 scm_t_ptob_descriptor
*ptob
;
613 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
614 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
616 ret
= scm_words (tag
| mode_bits
, 3);
617 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
618 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
620 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
621 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
623 entry
->file_name
= SCM_BOOL_F
;
624 entry
->rw_active
= SCM_PORT_NEITHER
;
626 entry
->stream
= stream
;
627 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
628 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
629 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
631 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
632 entry
->ilseq_handler
= handler
;
633 entry
->iconv_descriptors
= NULL
;
635 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
636 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
638 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
639 register_finalizer_for_port (ret
);
645 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
647 return scm_c_make_port_with_encoding (tag
, mode_bits
,
648 scm_i_default_port_encoding (),
649 scm_i_get_conversion_strategy (SCM_BOOL_F
),
654 scm_new_port_table_entry (scm_t_bits tag
)
656 return scm_c_make_port (tag
, 0, 0);
663 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
665 "Return a boolean indicating whether @var{x} is a port.\n"
666 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
668 #define FUNC_NAME s_scm_port_p
670 return scm_from_bool (SCM_PORTP (x
));
674 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
676 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
677 "@code{#f}. Any object satisfying this predicate also satisfies\n"
679 #define FUNC_NAME s_scm_input_port_p
681 return scm_from_bool (SCM_INPUT_PORT_P (x
));
685 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
687 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
688 "@code{#f}. Any object satisfying this predicate also satisfies\n"
690 #define FUNC_NAME s_scm_output_port_p
692 x
= SCM_COERCE_OUTPORT (x
);
693 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
697 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
699 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
701 #define FUNC_NAME s_scm_port_closed_p
703 SCM_VALIDATE_PORT (1, port
);
704 return scm_from_bool (!SCM_OPPORTP (port
));
708 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
710 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
712 #define FUNC_NAME s_scm_eof_object_p
714 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
723 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
726 * Call the close operation on a port object.
727 * see also scm_close.
729 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
731 "Close the specified port object. Return @code{#t} if it\n"
732 "successfully closes a port or @code{#f} if it was already\n"
733 "closed. An exception may be raised if an error occurs, for\n"
734 "example when flushing buffered output. See also @ref{Ports and\n"
735 "File Descriptors, close}, for a procedure which can close file\n"
737 #define FUNC_NAME s_scm_close_port
742 port
= SCM_COERCE_OUTPORT (port
);
744 SCM_VALIDATE_PORT (1, port
);
745 if (SCM_CLOSEDP (port
))
748 p
= SCM_PTAB_ENTRY (port
);
749 SCM_CLR_PORT_OPEN_FLAG (port
);
751 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
752 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
754 if (SCM_PORT_DESCRIPTOR (port
)->close
)
755 /* Note! This may throw an exception. Anything after this point
756 should be resilient to non-local exits. */
757 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
761 if (p
->iconv_descriptors
)
763 /* If we don't get here, the iconv_descriptors finalizer will
765 close_iconv_descriptors (p
->iconv_descriptors
);
766 p
->iconv_descriptors
= NULL
;
769 return scm_from_bool (rv
>= 0);
773 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
775 "Close the specified input port object. The routine has no effect if\n"
776 "the file has already been closed. An exception may be raised if an\n"
777 "error occurs. The value returned is unspecified.\n\n"
778 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
779 "which can close file descriptors.")
780 #define FUNC_NAME s_scm_close_input_port
782 SCM_VALIDATE_INPUT_PORT (1, port
);
783 scm_close_port (port
);
784 return SCM_UNSPECIFIED
;
788 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
790 "Close the specified output port object. The routine has no effect if\n"
791 "the file has already been closed. An exception may be raised if an\n"
792 "error occurs. The value returned is unspecified.\n\n"
793 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
794 "which can close file descriptors.")
795 #define FUNC_NAME s_scm_close_output_port
797 port
= SCM_COERCE_OUTPORT (port
);
798 SCM_VALIDATE_OUTPUT_PORT (1, port
);
799 scm_close_port (port
);
800 return SCM_UNSPECIFIED
;
807 /* Encoding characters to byte streams, and decoding byte streams to
810 /* A fluid specifying the default encoding for newly created ports. If it is
811 a string, that is the encoding. If it is #f, it is in the "native"
812 (Latin-1) encoding. */
813 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
815 static int scm_port_encoding_init
= 0;
817 /* Use ENCODING as the default encoding for future ports. */
819 scm_i_set_default_port_encoding (const char *encoding
)
821 if (!scm_port_encoding_init
822 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
823 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
827 || !strcmp (encoding
, "ASCII")
828 || !strcmp (encoding
, "ANSI_X3.4-1968")
829 || !strcmp (encoding
, "ISO-8859-1"))
830 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
832 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
833 scm_from_locale_string (encoding
));
836 /* Return the name of the default encoding for newly created ports; a
837 return value of NULL means "ISO-8859-1". */
839 scm_i_default_port_encoding (void)
841 if (!scm_port_encoding_init
)
843 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
849 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
850 if (!scm_is_string (encoding
))
853 return scm_i_string_chars (encoding
);
858 finalize_iconv_descriptors (GC_PTR ptr
, GC_PTR data
)
860 close_iconv_descriptors (ptr
);
863 static scm_t_iconv_descriptors
*
864 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
866 scm_t_iconv_descriptors
*id
;
867 iconv_t input_cd
, output_cd
;
869 input_cd
= (iconv_t
) -1;
870 output_cd
= (iconv_t
) -1;
874 /* Open an input iconv conversion descriptor, from ENCODING
875 to UTF-8. We choose UTF-8, not UTF-32, because iconv
876 implementations can typically convert from anything to
877 UTF-8, but not to UTF-32 (see
878 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
880 /* Assume opening an iconv descriptor causes about 16 KB of
882 scm_gc_register_allocation (16 * 1024);
884 input_cd
= iconv_open ("UTF-8", encoding
);
885 if (input_cd
== (iconv_t
) -1)
886 goto invalid_encoding
;
891 /* Assume opening an iconv descriptor causes about 16 KB of
893 scm_gc_register_allocation (16 * 1024);
895 output_cd
= iconv_open (encoding
, "UTF-8");
896 if (output_cd
== (iconv_t
) -1)
898 if (input_cd
!= (iconv_t
) -1)
899 iconv_close (input_cd
);
900 goto invalid_encoding
;
904 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
905 id
->input_cd
= input_cd
;
906 id
->output_cd
= output_cd
;
909 GC_finalization_proc prev_finalizer
;
910 GC_PTR prev_finalization_data
;
912 /* Register a finalizer to close the descriptors. */
913 GC_REGISTER_FINALIZER_NO_ORDER (id
, finalize_iconv_descriptors
, 0,
914 &prev_finalizer
, &prev_finalization_data
);
922 err
= scm_from_locale_string (encoding
);
923 scm_misc_error ("open_iconv_descriptors",
924 "invalid or unknown character encoding ~s",
930 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
932 if (id
->input_cd
!= (iconv_t
) -1)
933 iconv_close (id
->input_cd
);
934 if (id
->output_cd
!= (iconv_t
) -1)
935 iconv_close (id
->output_cd
);
936 id
->input_cd
= (void *) -1;
937 id
->output_cd
= (void *) -1;
940 scm_t_iconv_descriptors
*
941 scm_i_port_iconv_descriptors (SCM port
)
945 pt
= SCM_PTAB_ENTRY (port
);
947 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
949 if (!pt
->iconv_descriptors
)
952 pt
->encoding
= "ISO-8859-1";
953 pt
->iconv_descriptors
=
954 open_iconv_descriptors (pt
->encoding
,
955 SCM_INPUT_PORT_P (port
),
956 SCM_OUTPUT_PORT_P (port
));
959 return pt
->iconv_descriptors
;
963 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
966 scm_t_iconv_descriptors
*prev
;
968 /* Set the character encoding for this port. */
969 pt
= SCM_PTAB_ENTRY (port
);
970 prev
= pt
->iconv_descriptors
;
972 if (encoding
== NULL
)
973 encoding
= "ISO-8859-1";
975 if (strcmp (encoding
, "UTF-8") == 0)
977 pt
->encoding
= "UTF-8";
978 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
979 pt
->iconv_descriptors
= NULL
;
983 /* Open descriptors before mutating the port. */
984 pt
->iconv_descriptors
=
985 open_iconv_descriptors (encoding
,
986 SCM_INPUT_PORT_P (port
),
987 SCM_OUTPUT_PORT_P (port
));
988 pt
->encoding
= scm_gc_strdup (encoding
, "port");
989 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
993 close_iconv_descriptors (prev
);
996 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
998 "Returns, as a string, the character encoding that @var{port}\n"
999 "uses to interpret its input and output.\n")
1000 #define FUNC_NAME s_scm_port_encoding
1005 SCM_VALIDATE_PORT (1, port
);
1007 pt
= SCM_PTAB_ENTRY (port
);
1010 return scm_from_locale_string (pt
->encoding
);
1016 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1017 (SCM port
, SCM enc
),
1018 "Sets the character encoding that will be used to interpret all\n"
1019 "port I/O. New ports are created with the encoding\n"
1020 "appropriate for the current locale if @code{setlocale} has \n"
1021 "been called or ISO-8859-1 otherwise\n"
1022 "and this procedure can be used to modify that encoding.\n")
1023 #define FUNC_NAME s_scm_set_port_encoding_x
1027 SCM_VALIDATE_PORT (1, port
);
1028 SCM_VALIDATE_STRING (2, enc
);
1030 enc_str
= scm_to_locale_string (enc
);
1031 scm_i_set_port_encoding_x (port
, enc_str
);
1034 return SCM_UNSPECIFIED
;
1039 /* This determines how conversions handle unconvertible characters. */
1040 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
1041 static int scm_conversion_strategy_init
= 0;
1043 scm_t_string_failed_conversion_handler
1044 scm_i_get_conversion_strategy (SCM port
)
1048 if (scm_is_false (port
))
1050 if (!scm_conversion_strategy_init
1051 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1052 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1055 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
1056 if (scm_is_false (encoding
))
1057 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
1059 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
1065 pt
= SCM_PTAB_ENTRY (port
);
1066 return pt
->ilseq_handler
;
1072 scm_i_set_conversion_strategy_x (SCM port
,
1073 scm_t_string_failed_conversion_handler handler
)
1078 strategy
= scm_from_int ((int) handler
);
1080 if (scm_is_false (port
))
1082 /* Set the default encoding for future ports. */
1083 if (!scm_conversion_strategy_init
1084 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1085 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1087 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1091 /* Set the character encoding for this port. */
1092 pt
= SCM_PTAB_ENTRY (port
);
1093 pt
->ilseq_handler
= handler
;
1097 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1098 1, 0, 0, (SCM port
),
1099 "Returns the behavior of the port when handling a character that\n"
1100 "is not representable in the port's current encoding.\n"
1101 "It returns the symbol @code{error} if unrepresentable characters\n"
1102 "should cause exceptions, @code{substitute} if the port should\n"
1103 "try to replace unrepresentable characters with question marks or\n"
1104 "approximate characters, or @code{escape} if unrepresentable\n"
1105 "characters should be converted to string escapes.\n"
1107 "If @var{port} is @code{#f}, then the current default behavior\n"
1108 "will be returned. New ports will have this default behavior\n"
1109 "when they are created.\n")
1110 #define FUNC_NAME s_scm_port_conversion_strategy
1112 scm_t_string_failed_conversion_handler h
;
1114 SCM_VALIDATE_OPPORT (1, port
);
1116 if (!scm_is_false (port
))
1118 SCM_VALIDATE_OPPORT (1, port
);
1121 h
= scm_i_get_conversion_strategy (port
);
1122 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1123 return scm_from_latin1_symbol ("error");
1124 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1125 return scm_from_latin1_symbol ("substitute");
1126 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1127 return scm_from_latin1_symbol ("escape");
1131 /* Never gets here. */
1132 return SCM_UNDEFINED
;
1136 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1138 (SCM port
, SCM sym
),
1139 "Sets the behavior of the interpreter when outputting a character\n"
1140 "that is not representable in the port's current encoding.\n"
1141 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1142 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1143 "when an unconvertible character is encountered. If it is\n"
1144 "@code{'substitute}, then unconvertible characters will \n"
1145 "be replaced with approximate characters, or with question marks\n"
1146 "if no approximately correct character is available.\n"
1147 "If it is @code{'escape},\n"
1148 "it will appear as a hex escape when output.\n"
1150 "If @var{port} is an open port, the conversion error behavior\n"
1151 "is set for that port. If it is @code{#f}, it is set as the\n"
1152 "default behavior for any future ports that get created in\n"
1154 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1160 if (!scm_is_false (port
))
1162 SCM_VALIDATE_OPPORT (1, port
);
1165 err
= scm_from_latin1_symbol ("error");
1166 if (scm_is_true (scm_eqv_p (sym
, err
)))
1168 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1169 return SCM_UNSPECIFIED
;
1172 qm
= scm_from_latin1_symbol ("substitute");
1173 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1175 scm_i_set_conversion_strategy_x (port
,
1176 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1177 return SCM_UNSPECIFIED
;
1180 esc
= scm_from_latin1_symbol ("escape");
1181 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1183 scm_i_set_conversion_strategy_x (port
,
1184 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1185 return SCM_UNSPECIFIED
;
1188 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1190 return SCM_UNSPECIFIED
;
1197 /* The port lock. */
1200 lock_port (void *mutex
)
1202 scm_i_pthread_mutex_lock (mutex
);
1206 unlock_port (void *mutex
)
1208 scm_i_pthread_mutex_unlock (mutex
);
1212 scm_dynwind_lock_port (SCM port
)
1213 #define FUNC_NAME "dynwind-lock-port"
1215 scm_i_pthread_mutex_t
*lock
;
1216 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1217 scm_c_lock_port (port
, &lock
);
1220 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1221 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1229 /* Revealed counts --- an oddity inherited from SCSH. */
1231 /* Find a port in the table and return its revealed count.
1232 Also used by the garbage collector.
1235 scm_revealed_count (SCM port
)
1237 scm_i_pthread_mutex_t
*lock
;
1240 scm_c_lock_port (port
, &lock
);
1241 ret
= SCM_REVEALED (port
);
1243 scm_i_pthread_mutex_unlock (lock
);
1248 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
1250 "Return the revealed count for @var{port}.")
1251 #define FUNC_NAME s_scm_port_revealed
1253 port
= SCM_COERCE_OUTPORT (port
);
1254 SCM_VALIDATE_OPENPORT (1, port
);
1255 return scm_from_int (scm_revealed_count (port
));
1259 /* Set the revealed count for a port. */
1260 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
1261 (SCM port
, SCM rcount
),
1262 "Sets the revealed count for a port to a given value.\n"
1263 "The return value is unspecified.")
1264 #define FUNC_NAME s_scm_set_port_revealed_x
1267 scm_i_pthread_mutex_t
*lock
;
1269 /* FIXME: It doesn't make sense to manipulate revealed counts on ports
1270 without a free function. */
1272 port
= SCM_COERCE_OUTPORT (port
);
1273 SCM_VALIDATE_OPENPORT (1, port
);
1274 r
= scm_to_int (rcount
);
1275 scm_c_lock_port (port
, &lock
);
1276 SCM_REVEALED (port
) = r
;
1278 scm_i_pthread_mutex_unlock (lock
);
1279 return SCM_UNSPECIFIED
;
1283 /* Set the revealed count for a port. */
1284 SCM_DEFINE (scm_adjust_port_revealed_x
, "adjust-port-revealed!", 2, 0, 0,
1285 (SCM port
, SCM addend
),
1286 "Add @var{addend} to the revealed count of @var{port}.\n"
1287 "The return value is unspecified.")
1288 #define FUNC_NAME s_scm_adjust_port_revealed_x
1290 scm_i_pthread_mutex_t
*lock
;
1292 port
= SCM_COERCE_OUTPORT (port
);
1293 SCM_VALIDATE_OPENPORT (1, port
);
1294 a
= scm_to_int (addend
);
1295 scm_c_lock_port (port
, &lock
);
1296 SCM_REVEALED (port
) += a
;
1298 scm_i_pthread_mutex_unlock (lock
);
1299 return SCM_UNSPECIFIED
;
1309 scm_get_byte_or_eof (SCM port
)
1311 scm_i_pthread_mutex_t
*lock
;
1314 scm_c_lock_port (port
, &lock
);
1315 ret
= scm_get_byte_or_eof_unlocked (port
);
1317 scm_i_pthread_mutex_unlock (lock
);
1323 scm_peek_byte_or_eof (SCM port
)
1325 scm_i_pthread_mutex_t
*lock
;
1328 scm_c_lock_port (port
, &lock
);
1329 ret
= scm_peek_byte_or_eof_unlocked (port
);
1331 scm_i_pthread_mutex_unlock (lock
);
1338 * Used by an application to read arbitrary number of bytes from an
1339 * SCM port. Same semantics as libc read, except that scm_c_read only
1340 * returns less than SIZE bytes if at end-of-file.
1342 * Warning: Doesn't update port line and column counts! */
1344 /* This structure, and the following swap_buffer function, are used
1345 for temporarily swapping a port's own read buffer, and the buffer
1346 that the caller of scm_c_read provides. */
1347 struct port_and_swap_buffer
1350 unsigned char *buffer
;
1355 swap_buffer (void *data
)
1357 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1358 unsigned char *old_buf
= psb
->pt
->read_buf
;
1359 size_t old_size
= psb
->pt
->read_buf_size
;
1361 /* Make the port use (buffer, size) from the struct. */
1362 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1363 psb
->pt
->read_buf_size
= psb
->size
;
1365 /* Save the port's old (buffer, size) in the struct. */
1366 psb
->buffer
= old_buf
;
1367 psb
->size
= old_size
;
1371 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1372 #define FUNC_NAME "scm_c_read"
1375 size_t n_read
= 0, n_available
;
1376 struct port_and_swap_buffer psb
;
1378 SCM_VALIDATE_OPINPORT (1, port
);
1380 pt
= SCM_PTAB_ENTRY (port
);
1381 if (pt
->rw_active
== SCM_PORT_WRITE
)
1382 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1385 pt
->rw_active
= SCM_PORT_READ
;
1387 /* Take bytes first from the port's read buffer. */
1388 if (pt
->read_pos
< pt
->read_end
)
1390 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1391 memcpy (buffer
, pt
->read_pos
, n_available
);
1392 buffer
= (char *) buffer
+ n_available
;
1393 pt
->read_pos
+= n_available
;
1394 n_read
+= n_available
;
1395 size
-= n_available
;
1398 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1402 /* Now we will call scm_fill_input repeatedly until we have read the
1403 requested number of bytes. (Note that a single scm_fill_input
1404 call does not guarantee to fill the whole of the port's read
1406 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1408 /* The port that we are reading from is unbuffered - i.e. does
1409 not have its own persistent buffer - but we have a buffer,
1410 provided by our caller, that is the right size for the data
1411 that is wanted. For the following scm_fill_input calls,
1412 therefore, we use the buffer in hand as the port's read
1415 We need to make sure that the port's normal (1 byte) buffer
1416 is reinstated in case one of the scm_fill_input () calls
1417 throws an exception; we use the scm_dynwind_* API to achieve
1420 A consequence of this optimization is that the fill_input
1421 functions can't unget characters. That'll push data to the
1422 pushback buffer instead of this psb buffer. */
1424 unsigned char *pback
= pt
->putback_buf
;
1427 psb
.buffer
= buffer
;
1429 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1430 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1431 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1433 /* Call scm_fill_input until we have all the bytes that we need,
1435 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1437 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1438 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1441 if (pback
!= pt
->putback_buf
1442 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1443 scm_misc_error (FUNC_NAME
,
1444 "scm_c_read must not call a fill function that pushes "
1445 "back characters onto an unbuffered port", SCM_EOL
);
1447 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1449 /* Reinstate the port's normal buffer. */
1454 /* The port has its own buffer. It is important that we use it,
1455 even if it happens to be smaller than our caller's buffer, so
1456 that a custom port implementation's entry points (in
1457 particular, fill_input) can rely on the buffer always being
1458 the same as they first set up. */
1459 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1461 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1462 memcpy (buffer
, pt
->read_pos
, n_available
);
1463 buffer
= (char *) buffer
+ n_available
;
1464 pt
->read_pos
+= n_available
;
1465 n_read
+= n_available
;
1466 size
-= n_available
;
1475 scm_c_read (SCM port
, void *buffer
, size_t size
)
1477 scm_i_pthread_mutex_t
*lock
;
1480 scm_c_lock_port (port
, &lock
);
1481 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1483 scm_i_pthread_mutex_unlock (lock
);
1489 /* Update the line and column number of PORT after consumption of C. */
1491 update_port_lf (scm_t_wchar c
, SCM port
)
1516 #define SCM_MBCHAR_BUF_SIZE (4)
1518 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1519 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1521 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1523 scm_t_wchar codepoint
;
1525 if (utf8_buf
[0] <= 0x7f)
1528 codepoint
= utf8_buf
[0];
1530 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1533 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1534 | (utf8_buf
[1] & 0x3f);
1536 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1539 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1540 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1541 | (utf8_buf
[2] & 0x3f);
1546 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1547 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1548 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1549 | (utf8_buf
[3] & 0x3f);
1555 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1556 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1557 representation, and set *LEN to the length in bytes. Return
1558 `EILSEQ' on error. */
1560 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1561 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1563 #define ASSERT_NOT_EOF(b) \
1564 if (SCM_UNLIKELY ((b) == EOF)) \
1566 #define CONSUME_PEEKED_BYTE() \
1573 pt
= SCM_PTAB_ENTRY (port
);
1575 byte
= scm_get_byte_or_eof_unlocked (port
);
1582 buf
[0] = (scm_t_uint8
) byte
;
1587 *codepoint
= buf
[0];
1588 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1591 byte
= scm_peek_byte_or_eof_unlocked (port
);
1592 ASSERT_NOT_EOF (byte
);
1594 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1597 CONSUME_PEEKED_BYTE ();
1598 buf
[1] = (scm_t_uint8
) byte
;
1601 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1604 else if ((buf
[0] & 0xf0) == 0xe0)
1607 byte
= scm_peek_byte_or_eof_unlocked (port
);
1608 ASSERT_NOT_EOF (byte
);
1610 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1611 || (buf
[0] == 0xe0 && byte
< 0xa0)
1612 || (buf
[0] == 0xed && byte
> 0x9f)))
1615 CONSUME_PEEKED_BYTE ();
1616 buf
[1] = (scm_t_uint8
) byte
;
1619 byte
= scm_peek_byte_or_eof_unlocked (port
);
1620 ASSERT_NOT_EOF (byte
);
1622 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1625 CONSUME_PEEKED_BYTE ();
1626 buf
[2] = (scm_t_uint8
) byte
;
1629 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1630 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1633 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1636 byte
= scm_peek_byte_or_eof_unlocked (port
);
1637 ASSERT_NOT_EOF (byte
);
1639 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1640 || (buf
[0] == 0xf0 && byte
< 0x90)
1641 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1644 CONSUME_PEEKED_BYTE ();
1645 buf
[1] = (scm_t_uint8
) byte
;
1648 byte
= scm_peek_byte_or_eof_unlocked (port
);
1649 ASSERT_NOT_EOF (byte
);
1651 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1654 CONSUME_PEEKED_BYTE ();
1655 buf
[2] = (scm_t_uint8
) byte
;
1658 byte
= scm_peek_byte_or_eof_unlocked (port
);
1659 ASSERT_NOT_EOF (byte
);
1661 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1664 CONSUME_PEEKED_BYTE ();
1665 buf
[3] = (scm_t_uint8
) byte
;
1668 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1669 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1670 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1679 /* Here we could choose the consume the faulty byte when it's not a
1680 valid starting byte, but it's not a requirement. What Section 3.9
1681 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1682 would otherwise be a valid starting byte. */
1686 #undef CONSUME_PEEKED_BYTE
1687 #undef ASSERT_NOT_EOF
1690 /* Likewise, read a byte sequence from PORT, passing it through its
1691 input conversion descriptor. */
1693 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1694 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1696 scm_t_iconv_descriptors
*id
;
1698 size_t bytes_consumed
, output_size
;
1700 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1702 id
= scm_i_port_iconv_descriptors (port
);
1704 for (output_size
= 0, output
= (char *) utf8_buf
,
1705 bytes_consumed
= 0, err
= 0;
1706 err
== 0 && output_size
== 0
1707 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1711 size_t input_left
, output_left
, done
;
1713 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1714 if (byte_read
== EOF
)
1716 if (bytes_consumed
== 0)
1718 *codepoint
= (scm_t_wchar
) EOF
;
1726 buf
[bytes_consumed
] = byte_read
;
1729 input_left
= bytes_consumed
+ 1;
1730 output_left
= sizeof (utf8_buf
);
1732 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1733 if (done
== (size_t) -1)
1737 /* Missing input: keep trying. */
1741 output_size
= sizeof (utf8_buf
) - output_left
;
1744 if (SCM_UNLIKELY (output_size
== 0))
1745 /* An unterminated sequence. */
1747 else if (SCM_LIKELY (err
== 0))
1749 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1750 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1751 *len
= bytes_consumed
;
1757 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1758 with the byte representation of the codepoint in PORT's encoding, and
1759 set *LEN to the length in bytes of that representation. Return 0 on
1760 success and an errno value on error. */
1762 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1763 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1766 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1768 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1769 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1771 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1773 if (SCM_LIKELY (err
== 0))
1774 update_port_lf (*codepoint
, port
);
1775 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1779 update_port_lf (*codepoint
, port
);
1785 /* Read a codepoint from PORT and return it. */
1787 scm_getc_unlocked (SCM port
)
1788 #define FUNC_NAME "scm_getc"
1792 scm_t_wchar codepoint
;
1793 char buf
[SCM_MBCHAR_BUF_SIZE
];
1795 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1796 if (SCM_UNLIKELY (err
!= 0))
1797 /* At this point PORT should point past the invalid encoding, as per
1798 R6RS-lib Section 8.2.4. */
1799 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1808 scm_i_pthread_mutex_t
*lock
;
1811 scm_c_lock_port (port
, &lock
);
1812 ret
= scm_getc_unlocked (port
);
1814 scm_i_pthread_mutex_unlock (lock
);
1820 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1822 "Return the next character available from @var{port}, updating\n"
1823 "@var{port} to point to the following character. If no more\n"
1824 "characters are available, the end-of-file object is returned.\n"
1826 "When @var{port}'s data cannot be decoded according to its\n"
1827 "character encoding, a @code{decoding-error} is raised and\n"
1828 "@var{port} points past the erroneous byte sequence.\n")
1829 #define FUNC_NAME s_scm_read_char
1832 if (SCM_UNBNDP (port
))
1833 port
= scm_current_input_port ();
1834 SCM_VALIDATE_OPINPORT (1, port
);
1835 c
= scm_getc_unlocked (port
);
1838 return SCM_MAKE_CHAR (c
);
1848 scm_unget_byte_unlocked (int c
, SCM port
)
1849 #define FUNC_NAME "scm_unget_byte"
1851 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1853 if (pt
->read_buf
== pt
->putback_buf
)
1854 /* already using the put-back buffer. */
1856 /* enlarge putback_buf if necessary. */
1857 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1858 && pt
->read_buf
== pt
->read_pos
)
1860 size_t new_size
= pt
->read_buf_size
* 2;
1861 unsigned char *tmp
= (unsigned char *)
1862 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1865 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1866 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1867 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1870 /* shift any existing bytes to buffer + 1. */
1871 if (pt
->read_pos
== pt
->read_end
)
1872 pt
->read_end
= pt
->read_buf
+ 1;
1873 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1875 int count
= pt
->read_end
- pt
->read_pos
;
1877 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1878 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1881 pt
->read_pos
= pt
->read_buf
;
1884 /* switch to the put-back buffer. */
1886 if (pt
->putback_buf
== NULL
)
1889 = (unsigned char *) scm_gc_malloc_pointerless
1890 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1891 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1894 pt
->saved_read_buf
= pt
->read_buf
;
1895 pt
->saved_read_pos
= pt
->read_pos
;
1896 pt
->saved_read_end
= pt
->read_end
;
1897 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1899 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1900 pt
->read_end
= pt
->read_buf
+ 1;
1901 pt
->read_buf_size
= pt
->putback_buf_size
;
1907 pt
->rw_active
= SCM_PORT_READ
;
1912 scm_unget_byte (int c
, SCM port
)
1914 scm_i_pthread_mutex_t
*lock
;
1915 scm_c_lock_port (port
, &lock
);
1916 scm_unget_byte_unlocked (c
, port
);
1918 scm_i_pthread_mutex_unlock (lock
);
1923 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1924 #define FUNC_NAME "scm_ungetc"
1926 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1928 char result_buf
[10];
1929 const char *encoding
;
1933 if (pt
->encoding
!= NULL
)
1934 encoding
= pt
->encoding
;
1936 encoding
= "ISO-8859-1";
1938 len
= sizeof (result_buf
);
1939 result
= u32_conv_to_encoding (encoding
,
1940 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1941 (uint32_t *) &c
, 1, NULL
,
1944 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1945 scm_encoding_error (FUNC_NAME
, errno
,
1946 "conversion to port encoding failed",
1947 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1949 for (i
= len
- 1; i
>= 0; i
--)
1950 scm_unget_byte_unlocked (result
[i
], port
);
1952 if (SCM_UNLIKELY (result
!= result_buf
))
1957 /* What should col be in this case?
1958 * We'll leave it at -1.
1960 SCM_LINUM (port
) -= 1;
1968 scm_ungetc (scm_t_wchar c
, SCM port
)
1970 scm_i_pthread_mutex_t
*lock
;
1971 scm_c_lock_port (port
, &lock
);
1972 scm_ungetc_unlocked (c
, port
);
1974 scm_i_pthread_mutex_unlock (lock
);
1979 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1981 /* This is simple minded and inefficient, but unreading strings is
1982 * probably not a common operation, and remember that line and
1983 * column numbers have to be handled...
1985 * Please feel free to write an optimized version!
1988 scm_ungetc_unlocked (s
[n
], port
);
1992 scm_ungets (const char *s
, int n
, SCM port
)
1994 scm_i_pthread_mutex_t
*lock
;
1995 scm_c_lock_port (port
, &lock
);
1996 scm_ungets_unlocked (s
, n
, port
);
1998 scm_i_pthread_mutex_unlock (lock
);
2002 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
2004 "Return the next character available from @var{port},\n"
2005 "@emph{without} updating @var{port} to point to the following\n"
2006 "character. If no more characters are available, the\n"
2007 "end-of-file object is returned.\n"
2009 "The value returned by\n"
2010 "a call to @code{peek-char} is the same as the value that would\n"
2011 "have been returned by a call to @code{read-char} on the same\n"
2012 "port. The only difference is that the very next call to\n"
2013 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
2014 "return the value returned by the preceding call to\n"
2015 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
2016 "an interactive port will hang waiting for input whenever a call\n"
2017 "to @code{read-char} would have hung.\n"
2019 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
2020 "if such a situation occurs. However, unlike with @code{read-char},\n"
2021 "@var{port} still points at the beginning of the erroneous byte\n"
2022 "sequence when the error is raised.\n")
2023 #define FUNC_NAME s_scm_peek_char
2028 char bytes
[SCM_MBCHAR_BUF_SIZE
];
2029 long column
, line
, i
;
2032 if (SCM_UNBNDP (port
))
2033 port
= scm_current_input_port ();
2034 SCM_VALIDATE_OPINPORT (1, port
);
2036 column
= SCM_COL (port
);
2037 line
= SCM_LINUM (port
);
2039 err
= get_codepoint (port
, &c
, bytes
, &len
);
2041 for (i
= len
- 1; i
>= 0; i
--)
2042 scm_unget_byte_unlocked (bytes
[i
], port
);
2044 SCM_COL (port
) = column
;
2045 SCM_LINUM (port
) = line
;
2047 if (SCM_UNLIKELY (err
!= 0))
2049 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
2051 /* Shouldn't happen since `catch' always aborts to prompt. */
2052 result
= SCM_BOOL_F
;
2055 result
= SCM_EOF_VAL
;
2057 result
= SCM_MAKE_CHAR (c
);
2063 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2064 (SCM cobj
, SCM port
),
2065 "Place character @var{cobj} in @var{port} so that it will be\n"
2066 "read by the next read operation. If called multiple times, the\n"
2067 "unread characters will be read again in last-in first-out\n"
2068 "order. If @var{port} is not supplied, the current input port\n"
2070 #define FUNC_NAME s_scm_unread_char
2074 SCM_VALIDATE_CHAR (1, cobj
);
2075 if (SCM_UNBNDP (port
))
2076 port
= scm_current_input_port ();
2077 SCM_VALIDATE_OPINPORT (2, port
);
2079 c
= SCM_CHAR (cobj
);
2081 scm_ungetc_unlocked (c
, port
);
2086 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2087 (SCM str
, SCM port
),
2088 "Place the string @var{str} in @var{port} so that its characters will be\n"
2089 "read in subsequent read operations. If called multiple times, the\n"
2090 "unread characters will be read again in last-in first-out order. If\n"
2091 "@var{port} is not supplied, the current-input-port is used.")
2092 #define FUNC_NAME s_scm_unread_string
2095 SCM_VALIDATE_STRING (1, str
);
2096 if (SCM_UNBNDP (port
))
2097 port
= scm_current_input_port ();
2098 SCM_VALIDATE_OPINPORT (2, port
);
2100 n
= scm_i_string_length (str
);
2103 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2112 /* Manipulating the buffers. */
2114 /* This routine does not take any locks, as it is usually called as part
2115 of a port implementation. */
2117 scm_port_non_buffer (scm_t_port
*pt
)
2119 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2120 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2121 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2122 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2125 /* this should only be called when the read buffer is empty. it
2126 tries to refill the read buffer. it returns the first char from
2127 the port, which is either EOF or *(pt->read_pos). */
2129 scm_fill_input_unlocked (SCM port
)
2131 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2133 assert (pt
->read_pos
== pt
->read_end
);
2135 if (pt
->read_buf
== pt
->putback_buf
)
2137 /* finished reading put-back chars. */
2138 pt
->read_buf
= pt
->saved_read_buf
;
2139 pt
->read_pos
= pt
->saved_read_pos
;
2140 pt
->read_end
= pt
->saved_read_end
;
2141 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2142 if (pt
->read_pos
< pt
->read_end
)
2143 return *(pt
->read_pos
);
2145 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2149 scm_fill_input (SCM port
)
2151 scm_i_pthread_mutex_t
*lock
;
2154 scm_c_lock_port (port
, &lock
);
2155 ret
= scm_fill_input_unlocked (port
);
2157 scm_i_pthread_mutex_unlock (lock
);
2163 /* move up to read_len chars from port's putback and/or read buffers
2164 into memory starting at dest. returns the number of chars moved. */
2166 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2168 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2169 size_t chars_read
= 0;
2170 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2174 memcpy (dest
, pt
->read_pos
, from_buf
);
2175 pt
->read_pos
+= from_buf
;
2176 chars_read
+= from_buf
;
2177 read_len
-= from_buf
;
2181 /* if putback was active, try the real input buffer too. */
2182 if (pt
->read_buf
== pt
->putback_buf
)
2184 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2187 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2188 pt
->saved_read_pos
+= from_buf
;
2189 chars_read
+= from_buf
;
2195 /* Clear a port's read buffers, returning the contents. */
2196 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2198 "This procedure clears a port's input buffers, similar\n"
2199 "to the way that force-output clears the output buffer. The\n"
2200 "contents of the buffers are returned as a single string, e.g.,\n"
2203 "(define p (open-input-file ...))\n"
2204 "(drain-input p) => empty string, nothing buffered yet.\n"
2205 "(unread-char (read-char p) p)\n"
2206 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2208 "Draining the buffers may be useful for cleanly finishing\n"
2209 "buffered I/O so that the file descriptor can be used directly\n"
2210 "for further input.")
2211 #define FUNC_NAME s_scm_drain_input
2218 SCM_VALIDATE_OPINPORT (1, port
);
2219 pt
= SCM_PTAB_ENTRY (port
);
2221 count
= pt
->read_end
- pt
->read_pos
;
2222 if (pt
->read_buf
== pt
->putback_buf
)
2223 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2227 result
= scm_i_make_string (count
, &data
, 0);
2228 scm_take_from_input_buffers (port
, data
, count
);
2231 result
= scm_nullstr
;
2238 scm_end_input_unlocked (SCM port
)
2241 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2243 if (pt
->read_buf
== pt
->putback_buf
)
2245 offset
= pt
->read_end
- pt
->read_pos
;
2246 pt
->read_buf
= pt
->saved_read_buf
;
2247 pt
->read_pos
= pt
->saved_read_pos
;
2248 pt
->read_end
= pt
->saved_read_end
;
2249 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2254 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2258 scm_end_input (SCM port
)
2260 scm_i_pthread_mutex_t
*lock
;
2261 scm_c_lock_port (port
, &lock
);
2262 scm_end_input_unlocked (port
);
2264 scm_i_pthread_mutex_unlock (lock
);
2268 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2270 "Flush the specified output port, or the current output port if @var{port}\n"
2271 "is omitted. The current output buffer contents are passed to the\n"
2272 "underlying port implementation (e.g., in the case of fports, the\n"
2273 "data will be written to the file and the output buffer will be cleared.)\n"
2274 "It has no effect on an unbuffered port.\n\n"
2275 "The return value is unspecified.")
2276 #define FUNC_NAME s_scm_force_output
2278 if (SCM_UNBNDP (port
))
2279 port
= scm_current_output_port ();
2282 port
= SCM_COERCE_OUTPORT (port
);
2283 SCM_VALIDATE_OPOUTPORT (1, port
);
2285 scm_flush_unlocked (port
);
2286 return SCM_UNSPECIFIED
;
2291 scm_flush_unlocked (SCM port
)
2293 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2297 scm_flush (SCM port
)
2299 scm_i_pthread_mutex_t
*lock
;
2300 scm_c_lock_port (port
, &lock
);
2301 scm_flush_unlocked (port
);
2303 scm_i_pthread_mutex_unlock (lock
);
2313 scm_putc (char c
, SCM port
)
2315 scm_i_pthread_mutex_t
*lock
;
2316 scm_c_lock_port (port
, &lock
);
2317 scm_putc_unlocked (c
, port
);
2319 scm_i_pthread_mutex_unlock (lock
);
2324 scm_puts (const char *s
, SCM port
)
2326 scm_i_pthread_mutex_t
*lock
;
2327 scm_c_lock_port (port
, &lock
);
2328 scm_puts_unlocked (s
, port
);
2330 scm_i_pthread_mutex_unlock (lock
);
2336 * Used by an application to write arbitrary number of bytes to an SCM
2337 * port. Similar semantics as libc write. However, unlike libc
2338 * write, scm_c_write writes the requested number of bytes and has no
2341 * Warning: Doesn't update port line and column counts!
2344 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2345 #define FUNC_NAME "scm_c_write"
2348 scm_t_ptob_descriptor
*ptob
;
2350 SCM_VALIDATE_OPOUTPORT (1, port
);
2352 pt
= SCM_PTAB_ENTRY (port
);
2353 ptob
= SCM_PORT_DESCRIPTOR (port
);
2355 if (pt
->rw_active
== SCM_PORT_READ
)
2356 scm_end_input_unlocked (port
);
2358 ptob
->write (port
, ptr
, size
);
2361 pt
->rw_active
= SCM_PORT_WRITE
;
2366 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2368 scm_i_pthread_mutex_t
*lock
;
2369 scm_c_lock_port (port
, &lock
);
2370 scm_c_write_unlocked (port
, ptr
, size
);
2372 scm_i_pthread_mutex_unlock (lock
);
2378 * This function differs from scm_c_write; it updates port line and
2381 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2383 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2384 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2386 if (pt
->rw_active
== SCM_PORT_READ
)
2387 scm_end_input_unlocked (port
);
2389 ptob
->write (port
, ptr
, size
);
2391 for (; size
; ptr
++, size
--)
2392 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2395 pt
->rw_active
= SCM_PORT_WRITE
;
2399 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2401 scm_i_pthread_mutex_t
*lock
;
2402 scm_c_lock_port (port
, &lock
);
2403 scm_lfwrite_unlocked (ptr
, size
, port
);
2405 scm_i_pthread_mutex_unlock (lock
);
2409 /* Write STR to PORT from START inclusive to END exclusive. */
2411 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2413 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2415 if (pt
->rw_active
== SCM_PORT_READ
)
2416 scm_end_input_unlocked (port
);
2418 if (end
== (size_t) -1)
2419 end
= scm_i_string_length (str
);
2421 scm_display (scm_c_substring (str
, start
, end
), port
);
2424 pt
->rw_active
= SCM_PORT_WRITE
;
2430 /* Querying and setting positions, and character availability. */
2432 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2434 "Return @code{#t} if a character is ready on input @var{port}\n"
2435 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2436 "@code{#t} then the next @code{read-char} operation on\n"
2437 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2438 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2440 "@code{char-ready?} exists to make it possible for a\n"
2441 "program to accept characters from interactive ports without\n"
2442 "getting stuck waiting for input. Any input editors associated\n"
2443 "with such ports must make sure that characters whose existence\n"
2444 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2445 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2446 "a port at end of file would be indistinguishable from an\n"
2447 "interactive port that has no ready characters.")
2448 #define FUNC_NAME s_scm_char_ready_p
2452 if (SCM_UNBNDP (port
))
2453 port
= scm_current_input_port ();
2454 /* It's possible to close the current input port, so validate even in
2456 SCM_VALIDATE_OPINPORT (1, port
);
2458 pt
= SCM_PTAB_ENTRY (port
);
2460 /* if the current read buffer is filled, or the
2461 last pushed-back char has been read and the saved buffer is
2462 filled, result is true. */
2463 if (pt
->read_pos
< pt
->read_end
2464 || (pt
->read_buf
== pt
->putback_buf
2465 && pt
->saved_read_pos
< pt
->saved_read_end
))
2469 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2471 if (ptob
->input_waiting
)
2472 return scm_from_bool(ptob
->input_waiting (port
));
2479 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2480 (SCM fd_port
, SCM offset
, SCM whence
),
2481 "Sets the current position of @var{fd_port} to the integer\n"
2482 "@var{offset}, which is interpreted according to the value of\n"
2485 "One of the following variables should be supplied for\n"
2487 "@defvar SEEK_SET\n"
2488 "Seek from the beginning of the file.\n"
2490 "@defvar SEEK_CUR\n"
2491 "Seek from the current position.\n"
2493 "@defvar SEEK_END\n"
2494 "Seek from the end of the file.\n"
2496 "If @var{fd_port} is a file descriptor, the underlying system\n"
2497 "call is @code{lseek}. @var{port} may be a string port.\n"
2499 "The value returned is the new position in the file. This means\n"
2500 "that the current position of a port can be obtained using:\n"
2502 "(seek port 0 SEEK_CUR)\n"
2504 #define FUNC_NAME s_scm_seek
2508 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2510 how
= scm_to_int (whence
);
2511 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2512 SCM_OUT_OF_RANGE (3, whence
);
2514 if (SCM_OPPORTP (fd_port
))
2516 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2517 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2518 off_t_or_off64_t rv
;
2521 SCM_MISC_ERROR ("port is not seekable",
2522 scm_cons (fd_port
, SCM_EOL
));
2524 rv
= ptob
->seek (fd_port
, off
, how
);
2525 return scm_from_off_t_or_off64_t (rv
);
2527 else /* file descriptor?. */
2529 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2530 off_t_or_off64_t rv
;
2531 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2534 return scm_from_off_t_or_off64_t (rv
);
2543 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2544 doesn't have the filename version truncate(), hence this code. */
2545 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2547 truncate (const char *file
, off_t length
)
2551 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2555 ret
= ftruncate (fdes
, length
);
2558 int save_errno
= errno
;
2564 return close (fdes
);
2566 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2568 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2569 (SCM object
, SCM length
),
2570 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2571 "can be a filename string, a port object, or an integer file\n"
2573 "The return value is unspecified.\n"
2575 "For a port or file descriptor @var{length} can be omitted, in\n"
2576 "which case the file is truncated at the current position (per\n"
2577 "@code{ftell} above).\n"
2579 "On most systems a file can be extended by giving a length\n"
2580 "greater than the current size, but this is not mandatory in the\n"
2582 #define FUNC_NAME s_scm_truncate_file
2586 /* "object" can be a port, fdes or filename.
2588 Negative "length" makes no sense, but it's left to truncate() or
2589 ftruncate() to give back an error for that (normally EINVAL).
2592 if (SCM_UNBNDP (length
))
2594 /* must supply length if object is a filename. */
2595 if (scm_is_string (object
))
2596 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2598 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2601 object
= SCM_COERCE_OUTPORT (object
);
2602 if (scm_is_integer (object
))
2604 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2605 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2608 else if (SCM_OPOUTPORTP (object
))
2610 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2611 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2612 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2614 if (!ptob
->truncate
)
2615 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2616 if (pt
->rw_active
== SCM_PORT_READ
)
2617 scm_end_input_unlocked (object
);
2618 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2619 ptob
->flush (object
);
2621 ptob
->truncate (object
, c_length
);
2626 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2627 char *str
= scm_to_locale_string (object
);
2629 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2636 return SCM_UNSPECIFIED
;
2640 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2642 "Return the current line number for @var{port}.\n"
2644 "The first line of a file is 0. But you might want to add 1\n"
2645 "when printing line numbers, since starting from 1 is\n"
2646 "traditional in error messages, and likely to be more natural to\n"
2648 #define FUNC_NAME s_scm_port_line
2650 port
= SCM_COERCE_OUTPORT (port
);
2651 SCM_VALIDATE_OPENPORT (1, port
);
2652 return scm_from_long (SCM_LINUM (port
));
2656 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2657 (SCM port
, SCM line
),
2658 "Set the current line number for @var{port} to @var{line}. The\n"
2659 "first line of a file is 0.")
2660 #define FUNC_NAME s_scm_set_port_line_x
2662 port
= SCM_COERCE_OUTPORT (port
);
2663 SCM_VALIDATE_OPENPORT (1, port
);
2664 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2665 return SCM_UNSPECIFIED
;
2669 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2671 "Return the current column number of @var{port}.\n"
2672 "If the number is\n"
2673 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2674 "- i.e. the first character of the first line is line 0, column 0.\n"
2675 "(However, when you display a file position, for example in an error\n"
2676 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2677 "because lines and column numbers traditionally start with 1, and that is\n"
2678 "what non-programmers will find most natural.)")
2679 #define FUNC_NAME s_scm_port_column
2681 port
= SCM_COERCE_OUTPORT (port
);
2682 SCM_VALIDATE_OPENPORT (1, port
);
2683 return scm_from_int (SCM_COL (port
));
2687 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2688 (SCM port
, SCM column
),
2689 "Set the current column of @var{port}. Before reading the first\n"
2690 "character on a line the column should be 0.")
2691 #define FUNC_NAME s_scm_set_port_column_x
2693 port
= SCM_COERCE_OUTPORT (port
);
2694 SCM_VALIDATE_OPENPORT (1, port
);
2695 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2696 return SCM_UNSPECIFIED
;
2700 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2702 "Return the filename associated with @var{port}, or @code{#f}\n"
2703 "if no filename is associated with the port.")
2704 #define FUNC_NAME s_scm_port_filename
2706 port
= SCM_COERCE_OUTPORT (port
);
2707 SCM_VALIDATE_OPENPORT (1, port
);
2708 return SCM_FILENAME (port
);
2712 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2713 (SCM port
, SCM filename
),
2714 "Change the filename associated with @var{port}, using the current input\n"
2715 "port if none is specified. Note that this does not change the port's\n"
2716 "source of data, but only the value that is returned by\n"
2717 "@code{port-filename} and reported in diagnostic output.")
2718 #define FUNC_NAME s_scm_set_port_filename_x
2720 port
= SCM_COERCE_OUTPORT (port
);
2721 SCM_VALIDATE_OPENPORT (1, port
);
2722 /* We allow the user to set the filename to whatever he likes. */
2723 SCM_SET_FILENAME (port
, filename
);
2724 return SCM_UNSPECIFIED
;
2731 /* Implementation helpers for port printing functions. */
2734 scm_print_port_mode (SCM exp
, SCM port
)
2736 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2738 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2739 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2742 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2749 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2751 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2754 scm_puts_unlocked ("#<", port
);
2755 scm_print_port_mode (exp
, port
);
2756 scm_puts_unlocked (type
, port
);
2757 scm_putc_unlocked (' ', port
);
2758 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2759 scm_putc_unlocked ('>', port
);
2766 /* Iterating over all ports. */
2768 struct for_each_data
2770 void (*proc
) (void *data
, SCM p
);
2775 for_each_trampoline (void *data
, SCM port
, SCM result
)
2777 struct for_each_data
*d
= data
;
2779 d
->proc (d
->data
, port
);
2785 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2787 struct for_each_data d
;
2792 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2793 scm_i_port_weak_set
);
2797 scm_for_each_trampoline (void *data
, SCM port
)
2799 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2802 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2804 "Apply @var{proc} to each port in the Guile port table\n"
2805 "in turn. The return value is unspecified. More specifically,\n"
2806 "@var{proc} is applied exactly once to every port that exists\n"
2807 "in the system at the time @code{port-for-each} is invoked.\n"
2808 "Changes to the port table while @code{port-for-each} is running\n"
2809 "have no effect as far as @code{port-for-each} is concerned.")
2810 #define FUNC_NAME s_scm_port_for_each
2812 SCM_VALIDATE_PROC (1, proc
);
2814 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2816 return SCM_UNSPECIFIED
;
2821 flush_output_port (void *closure
, SCM port
)
2823 if (SCM_OPOUTPORTP (port
))
2824 scm_flush_unlocked (port
);
2827 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2829 "Equivalent to calling @code{force-output} on\n"
2830 "all open output ports. The return value is unspecified.")
2831 #define FUNC_NAME s_scm_flush_all_ports
2833 scm_c_port_for_each (&flush_output_port
, NULL
);
2834 return SCM_UNSPECIFIED
;
2843 scm_t_bits scm_tc16_void_port
= 0;
2845 static int fill_input_void_port (SCM port SCM_UNUSED
)
2851 write_void_port (SCM port SCM_UNUSED
,
2852 const void *data SCM_UNUSED
,
2853 size_t size SCM_UNUSED
)
2858 scm_i_void_port (long mode_bits
)
2862 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2864 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2870 scm_void_port (char *mode_str
)
2872 return scm_i_void_port (scm_mode_bits (mode_str
));
2875 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2877 "Create and return a new void port. A void port acts like\n"
2878 "@file{/dev/null}. The @var{mode} argument\n"
2879 "specifies the input/output modes for this port: see the\n"
2880 "documentation for @code{open-file} in @ref{File Ports}.")
2881 #define FUNC_NAME s_scm_sys_make_void_port
2883 return scm_i_void_port (scm_i_mode_bits (mode
));
2890 /* Initialization. */
2895 /* lseek() symbols. */
2896 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2897 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2898 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2900 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2903 cur_inport_fluid
= scm_make_fluid ();
2904 cur_outport_fluid
= scm_make_fluid ();
2905 cur_errport_fluid
= scm_make_fluid ();
2906 cur_loadport_fluid
= scm_make_fluid ();
2908 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2910 #include "libguile/ports.x"
2912 /* Use Latin-1 as the default port encoding. */
2913 SCM_VARIABLE_SET (default_port_encoding_var
,
2914 scm_make_fluid_with_default (SCM_BOOL_F
));
2915 scm_port_encoding_init
= 1;
2917 SCM_VARIABLE_SET (scm_conversion_strategy
,
2918 scm_make_fluid_with_default
2919 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2920 scm_conversion_strategy_init
= 1;
2922 /* These bindings are used when boot-9 turns `current-input-port' et
2923 al into parameters. They are then removed from the guile module. */
2924 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2925 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2926 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);