1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 3 of
7 * the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
32 #include <fcntl.h> /* for chsize on mingw */
37 #include <striconveh.h>
41 #include "libguile/_scm.h"
42 #include "libguile/async.h"
43 #include "libguile/deprecation.h"
44 #include "libguile/eval.h"
45 #include "libguile/fports.h" /* direct access for seek and truncate */
46 #include "libguile/goops.h"
47 #include "libguile/smob.h"
48 #include "libguile/chars.h"
49 #include "libguile/dynwind.h"
51 #include "libguile/keywords.h"
52 #include "libguile/hashtab.h"
53 #include "libguile/root.h"
54 #include "libguile/strings.h"
55 #include "libguile/mallocs.h"
56 #include "libguile/validate.h"
57 #include "libguile/ports.h"
58 #include "libguile/vectors.h"
59 #include "libguile/weak-set.h"
60 #include "libguile/fluids.h"
61 #include "libguile/eq.h"
75 #ifdef HAVE_SYS_IOCTL_H
76 #include <sys/ioctl.h>
79 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
80 already, but have this code here in case that wasn't so in past versions,
81 or perhaps to help other minimal DOS environments.
83 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
84 might be possibilities if we've got other systems without ftruncate. */
86 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
87 #define ftruncate(fd, size) chsize (fd, size)
89 #define HAVE_FTRUNCATE 1
93 /* The port kind table --- a dynamically resized array of port types. */
96 /* scm_ptobs scm_numptob
97 * implement a dynamically resized array of ptob records.
98 * Indexes into this table are used when generating type
99 * tags for smobjects (if you know a tag you can get an index and conversely).
101 static scm_t_ptob_descriptor
**scm_ptobs
= NULL
;
102 static long scm_numptob
= 0; /* Number of port types. */
103 static long scm_ptobs_size
= 0; /* Number of slots in the port type
105 static scm_i_pthread_mutex_t scm_ptobs_lock
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
108 scm_c_num_port_types (void)
112 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
114 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
119 scm_t_ptob_descriptor
*
120 scm_c_port_type_ref (long ptobnum
)
122 scm_t_ptob_descriptor
*ret
= NULL
;
124 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
126 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
127 ret
= scm_ptobs
[ptobnum
];
129 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
132 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
138 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
142 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
144 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
146 if (scm_numptob
== scm_ptobs_size
)
148 unsigned long old_size
= scm_ptobs_size
;
149 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
151 /* Currently there are only 9 predefined port types, so one
152 resize will cover it. */
153 scm_ptobs_size
= old_size
+ 10;
155 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
156 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
158 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
161 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
165 scm_ptobs
[ret
] = desc
;
168 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
171 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
177 * We choose to use an interface similar to the smob interface with
178 * fill_input and write as standard fields, passed to the port
179 * type constructor, and optional fields set by setters.
183 flush_port_default (SCM port SCM_UNUSED
)
188 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
193 scm_make_port_type (char *name
,
194 int (*fill_input
) (SCM port
),
195 void (*write
) (SCM port
, const void *data
, size_t size
))
197 scm_t_ptob_descriptor
*desc
;
200 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
201 memset (desc
, 0, sizeof (*desc
));
204 desc
->print
= scm_port_print
;
206 desc
->flush
= flush_port_default
;
207 desc
->end_input
= end_input_default
;
208 desc
->fill_input
= fill_input
;
210 ptobnum
= scm_c_port_type_add_x (desc
);
212 /* Make a class object if GOOPS is present. */
213 if (SCM_UNPACK (scm_port_class
[0]) != 0)
214 scm_make_port_classes (ptobnum
, name
);
216 return scm_tc7_port
+ ptobnum
* 256;
220 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
222 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
226 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
228 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
232 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
233 scm_print_state
*pstate
))
235 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
239 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
241 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
245 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
247 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
251 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
253 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->flush
= flush
;
257 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
259 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
263 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
265 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
269 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
271 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
275 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
277 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
282 /* Standard ports --- current input, output, error, and more(!). */
284 static SCM cur_inport_fluid
= SCM_BOOL_F
;
285 static SCM cur_outport_fluid
= SCM_BOOL_F
;
286 static SCM cur_errport_fluid
= SCM_BOOL_F
;
287 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
289 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
291 "Return the current input port. This is the default port used\n"
292 "by many input procedures. Initially, @code{current-input-port}\n"
293 "returns the @dfn{standard input} in Unix and C terminology.")
294 #define FUNC_NAME s_scm_current_input_port
296 if (scm_is_true (cur_inport_fluid
))
297 return scm_fluid_ref (cur_inport_fluid
);
303 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
305 "Return the current output port. This is the default port used\n"
306 "by many output procedures. Initially,\n"
307 "@code{current-output-port} returns the @dfn{standard output} in\n"
308 "Unix and C terminology.")
309 #define FUNC_NAME s_scm_current_output_port
311 if (scm_is_true (cur_outport_fluid
))
312 return scm_fluid_ref (cur_outport_fluid
);
318 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
320 "Return the port to which errors and warnings should be sent (the\n"
321 "@dfn{standard error} in Unix and C terminology).")
322 #define FUNC_NAME s_scm_current_error_port
324 if (scm_is_true (cur_errport_fluid
))
325 return scm_fluid_ref (cur_errport_fluid
);
332 scm_current_warning_port (void)
334 static SCM cwp_var
= SCM_BOOL_F
;
336 if (scm_is_false (cwp_var
))
337 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
339 return scm_call_0 (scm_variable_ref (cwp_var
));
342 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
344 "Return the current-load-port.\n"
345 "The load port is used internally by @code{primitive-load}.")
346 #define FUNC_NAME s_scm_current_load_port
348 return scm_fluid_ref (cur_loadport_fluid
);
352 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
354 "@deffnx {Scheme Procedure} set-current-output-port port\n"
355 "@deffnx {Scheme Procedure} set-current-error-port port\n"
356 "Change the ports returned by @code{current-input-port},\n"
357 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
358 "so that they use the supplied @var{port} for input or output.")
359 #define FUNC_NAME s_scm_set_current_input_port
361 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
362 SCM_VALIDATE_OPINPORT (1, port
);
363 scm_fluid_set_x (cur_inport_fluid
, port
);
369 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
371 "Set the current default output port to @var{port}.")
372 #define FUNC_NAME s_scm_set_current_output_port
374 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
375 port
= SCM_COERCE_OUTPORT (port
);
376 SCM_VALIDATE_OPOUTPORT (1, port
);
377 scm_fluid_set_x (cur_outport_fluid
, port
);
383 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
385 "Set the current default error port to @var{port}.")
386 #define FUNC_NAME s_scm_set_current_error_port
388 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
389 port
= SCM_COERCE_OUTPORT (port
);
390 SCM_VALIDATE_OPOUTPORT (1, port
);
391 scm_fluid_set_x (cur_errport_fluid
, port
);
398 scm_set_current_warning_port (SCM port
)
400 static SCM cwp_var
= SCM_BOOL_F
;
402 if (scm_is_false (cwp_var
))
403 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
405 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
410 scm_dynwind_current_input_port (SCM port
)
411 #define FUNC_NAME NULL
413 SCM_VALIDATE_OPINPORT (1, port
);
414 scm_dynwind_fluid (cur_inport_fluid
, port
);
419 scm_dynwind_current_output_port (SCM port
)
420 #define FUNC_NAME NULL
422 port
= SCM_COERCE_OUTPORT (port
);
423 SCM_VALIDATE_OPOUTPORT (1, port
);
424 scm_dynwind_fluid (cur_outport_fluid
, port
);
429 scm_dynwind_current_error_port (SCM port
)
430 #define FUNC_NAME NULL
432 port
= SCM_COERCE_OUTPORT (port
);
433 SCM_VALIDATE_OPOUTPORT (1, port
);
434 scm_dynwind_fluid (cur_errport_fluid
, port
);
439 scm_i_dynwind_current_load_port (SCM port
)
441 scm_dynwind_fluid (cur_loadport_fluid
, port
);
447 /* Retrieving a port's mode. */
449 /* Return the flags that characterize a port based on the mode
450 * string used to open a file for that port.
452 * See PORT FLAGS in scm.h
456 scm_i_mode_bits_n (SCM modes
)
459 | (scm_i_string_contains_char (modes
, 'r')
460 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
461 | (scm_i_string_contains_char (modes
, 'w')
462 || scm_i_string_contains_char (modes
, 'a')
463 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
464 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
465 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
469 scm_mode_bits (char *modes
)
471 return scm_i_mode_bits (scm_from_locale_string (modes
));
475 scm_i_mode_bits (SCM modes
)
479 if (!scm_is_string (modes
))
480 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
482 bits
= scm_i_mode_bits_n (modes
);
483 scm_remember_upto_here_1 (modes
);
487 /* Return the mode flags from an open port.
488 * Some modes such as "append" are only used when opening
489 * a file and are not returned here. */
491 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
493 "Return the port modes associated with the open port @var{port}.\n"
494 "These will not necessarily be identical to the modes used when\n"
495 "the port was opened, since modes such as \"append\" which are\n"
496 "used only during port creation are not retained.")
497 #define FUNC_NAME s_scm_port_mode
502 port
= SCM_COERCE_OUTPORT (port
);
503 SCM_VALIDATE_OPPORT (1, port
);
504 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
505 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
506 strcpy (modes
, "r+");
510 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
512 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
515 return scm_from_latin1_string (modes
);
522 /* The port table --- a weak set of all ports.
524 We need a global registry of ports to flush them all at exit, and to
525 get all the ports matching a file descriptor. */
526 SCM scm_i_port_weak_set
;
531 /* Port finalization. */
533 static void finalize_port (GC_PTR
, GC_PTR
);
535 /* Register a finalizer for PORT. */
536 static SCM_C_INLINE_KEYWORD
void
537 register_finalizer_for_port (SCM port
)
539 GC_finalization_proc prev_finalizer
;
540 GC_PTR prev_finalization_data
;
542 /* Register a finalizer for PORT so that its iconv CDs get freed and
543 optionally its type's `free' function gets called. */
544 GC_REGISTER_FINALIZER_NO_ORDER (SCM_HEAP_OBJECT_BASE (port
),
547 &prev_finalization_data
);
550 /* Finalize the object (a port) pointed to by PTR. */
552 finalize_port (GC_PTR ptr
, GC_PTR data
)
554 SCM port
= SCM_PACK_POINTER (ptr
);
556 if (!SCM_PORTP (port
))
559 if (SCM_OPENP (port
))
561 if (SCM_REVEALED (port
) > 0)
562 /* Keep "revealed" ports alive and re-register a finalizer. */
563 register_finalizer_for_port (port
);
566 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 entry
= SCM_PTAB_ENTRY (port
);
576 if (entry
->input_cd
!= (iconv_t
) -1)
577 iconv_close (entry
->input_cd
);
578 if (entry
->output_cd
!= (iconv_t
) -1)
579 iconv_close (entry
->output_cd
);
581 SCM_SETSTREAM (port
, 0);
582 SCM_CLR_PORT_OPEN_FLAG (port
);
584 scm_gc_ports_collected
++;
593 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
594 const char *encoding
,
595 scm_t_string_failed_conversion_handler handler
,
600 scm_t_ptob_descriptor
*ptob
;
602 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
603 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
605 ret
= scm_words (tag
| mode_bits
, 3);
606 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
607 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
609 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
610 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
612 entry
->file_name
= SCM_BOOL_F
;
613 entry
->rw_active
= SCM_PORT_NEITHER
;
615 entry
->stream
= stream
;
616 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
617 /* The conversion descriptors will be opened lazily. */
618 entry
->input_cd
= (iconv_t
) -1;
619 entry
->output_cd
= (iconv_t
) -1;
620 entry
->ilseq_handler
= handler
;
622 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
624 /* For each new port, register a finalizer so that it port type's free
625 function can be invoked eventually. */
626 register_finalizer_for_port (ret
);
632 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
634 return scm_c_make_port_with_encoding (tag
, mode_bits
,
635 scm_i_default_port_encoding (),
636 scm_i_get_conversion_strategy (SCM_BOOL_F
),
641 scm_new_port_table_entry (scm_t_bits tag
)
643 return scm_c_make_port (tag
, 0, 0);
646 /* Remove a port from the table and destroy it. */
649 scm_i_remove_port (SCM port
)
650 #define FUNC_NAME "scm_remove_port"
654 p
= SCM_PTAB_ENTRY (port
);
655 scm_port_non_buffer (p
);
656 SCM_SETPTAB_ENTRY (port
, 0);
657 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
659 p
->putback_buf
= NULL
;
660 p
->putback_buf_size
= 0;
662 if (p
->input_cd
!= (iconv_t
) -1)
664 iconv_close (p
->input_cd
);
665 p
->input_cd
= (iconv_t
) -1;
668 if (p
->output_cd
!= (iconv_t
) -1)
670 iconv_close (p
->output_cd
);
671 p
->output_cd
= (iconv_t
) -1;
681 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
683 "Return a boolean indicating whether @var{x} is a port.\n"
684 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
686 #define FUNC_NAME s_scm_port_p
688 return scm_from_bool (SCM_PORTP (x
));
692 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
694 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
695 "@code{#f}. Any object satisfying this predicate also satisfies\n"
697 #define FUNC_NAME s_scm_input_port_p
699 return scm_from_bool (SCM_INPUT_PORT_P (x
));
703 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
705 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
706 "@code{#f}. Any object satisfying this predicate also satisfies\n"
708 #define FUNC_NAME s_scm_output_port_p
710 x
= SCM_COERCE_OUTPORT (x
);
711 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
715 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
717 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
719 #define FUNC_NAME s_scm_port_closed_p
721 SCM_VALIDATE_PORT (1, port
);
722 return scm_from_bool (!SCM_OPPORTP (port
));
726 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
728 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
730 #define FUNC_NAME s_scm_eof_object_p
732 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
742 * Call the close operation on a port object.
743 * see also scm_close.
745 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
747 "Close the specified port object. Return @code{#t} if it\n"
748 "successfully closes a port or @code{#f} if it was already\n"
749 "closed. An exception may be raised if an error occurs, for\n"
750 "example when flushing buffered output. See also @ref{Ports and\n"
751 "File Descriptors, close}, for a procedure which can close file\n"
753 #define FUNC_NAME s_scm_close_port
757 port
= SCM_COERCE_OUTPORT (port
);
759 SCM_VALIDATE_PORT (1, port
);
760 if (SCM_CLOSEDP (port
))
762 if (SCM_PORT_DESCRIPTOR (port
)->close
)
763 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
766 scm_i_remove_port (port
);
767 SCM_CLR_PORT_OPEN_FLAG (port
);
768 return scm_from_bool (rv
>= 0);
772 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
774 "Close the specified input 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_input_port
781 SCM_VALIDATE_INPUT_PORT (1, port
);
782 scm_close_port (port
);
783 return SCM_UNSPECIFIED
;
787 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
789 "Close the specified output port object. The routine has no effect if\n"
790 "the file has already been closed. An exception may be raised if an\n"
791 "error occurs. The value returned is unspecified.\n\n"
792 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
793 "which can close file descriptors.")
794 #define FUNC_NAME s_scm_close_output_port
796 port
= SCM_COERCE_OUTPORT (port
);
797 SCM_VALIDATE_OUTPUT_PORT (1, port
);
798 scm_close_port (port
);
799 return SCM_UNSPECIFIED
;
806 /* Encoding characters to byte streams, and decoding byte streams to
809 /* A fluid specifying the default encoding for newly created ports. If it is
810 a string, that is the encoding. If it is #f, it is in the "native"
811 (Latin-1) encoding. */
812 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
814 static int scm_port_encoding_init
= 0;
816 /* Use ENCODING as the default encoding for future ports. */
818 scm_i_set_default_port_encoding (const char *encoding
)
820 if (!scm_port_encoding_init
821 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
822 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
826 || !strcmp (encoding
, "ASCII")
827 || !strcmp (encoding
, "ANSI_X3.4-1968")
828 || !strcmp (encoding
, "ISO-8859-1"))
829 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
831 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
832 scm_from_locale_string (encoding
));
835 /* Return the name of the default encoding for newly created ports; a
836 return value of NULL means "ISO-8859-1". */
838 scm_i_default_port_encoding (void)
840 if (!scm_port_encoding_init
)
842 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
848 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
849 if (!scm_is_string (encoding
))
852 return scm_i_string_chars (encoding
);
857 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
860 iconv_t new_input_cd
, new_output_cd
;
862 new_input_cd
= (iconv_t
) -1;
863 new_output_cd
= (iconv_t
) -1;
865 /* Set the character encoding for this port. */
866 pt
= SCM_PTAB_ENTRY (port
);
868 if (encoding
== NULL
)
869 encoding
= "ISO-8859-1";
871 if (pt
->encoding
!= encoding
)
872 pt
->encoding
= scm_gc_strdup (encoding
, "port");
874 /* If ENCODING is UTF-8, then no conversion descriptor is opened
875 because we do I/O ourselves. This saves 100+ KiB for each
877 if (strcmp (encoding
, "UTF-8"))
879 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
)
881 /* Open an input iconv conversion descriptor, from ENCODING
882 to UTF-8. We choose UTF-8, not UTF-32, because iconv
883 implementations can typically convert from anything to
884 UTF-8, but not to UTF-32 (see
885 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
886 new_input_cd
= iconv_open ("UTF-8", encoding
);
887 if (new_input_cd
== (iconv_t
) -1)
888 goto invalid_encoding
;
891 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
893 new_output_cd
= iconv_open (encoding
, "UTF-8");
894 if (new_output_cd
== (iconv_t
) -1)
896 if (new_input_cd
!= (iconv_t
) -1)
897 iconv_close (new_input_cd
);
898 goto invalid_encoding
;
903 if (pt
->input_cd
!= (iconv_t
) -1)
904 iconv_close (pt
->input_cd
);
905 if (pt
->output_cd
!= (iconv_t
) -1)
906 iconv_close (pt
->output_cd
);
908 pt
->input_cd
= new_input_cd
;
909 pt
->output_cd
= new_output_cd
;
916 err
= scm_from_locale_string (encoding
);
917 scm_misc_error ("scm_i_set_port_encoding_x",
918 "invalid or unknown character encoding ~s",
923 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
925 "Returns, as a string, the character encoding that @var{port}\n"
926 "uses to interpret its input and output.\n")
927 #define FUNC_NAME s_scm_port_encoding
932 SCM_VALIDATE_PORT (1, port
);
934 pt
= SCM_PTAB_ENTRY (port
);
937 return scm_from_locale_string (pt
->encoding
);
943 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
945 "Sets the character encoding that will be used to interpret all\n"
946 "port I/O. New ports are created with the encoding\n"
947 "appropriate for the current locale if @code{setlocale} has \n"
948 "been called or ISO-8859-1 otherwise\n"
949 "and this procedure can be used to modify that encoding.\n")
950 #define FUNC_NAME s_scm_set_port_encoding_x
954 SCM_VALIDATE_PORT (1, port
);
955 SCM_VALIDATE_STRING (2, enc
);
957 enc_str
= scm_to_locale_string (enc
);
958 scm_i_set_port_encoding_x (port
, enc_str
);
961 return SCM_UNSPECIFIED
;
966 /* This determines how conversions handle unconvertible characters. */
967 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
968 static int scm_conversion_strategy_init
= 0;
970 scm_t_string_failed_conversion_handler
971 scm_i_get_conversion_strategy (SCM port
)
975 if (scm_is_false (port
))
977 if (!scm_conversion_strategy_init
978 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
979 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
982 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
983 if (scm_is_false (encoding
))
984 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
986 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
992 pt
= SCM_PTAB_ENTRY (port
);
993 return pt
->ilseq_handler
;
999 scm_i_set_conversion_strategy_x (SCM port
,
1000 scm_t_string_failed_conversion_handler handler
)
1005 strategy
= scm_from_int ((int) handler
);
1007 if (scm_is_false (port
))
1009 /* Set the default encoding for future ports. */
1010 if (!scm_conversion_strategy_init
1011 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
1012 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
1014 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
1018 /* Set the character encoding for this port. */
1019 pt
= SCM_PTAB_ENTRY (port
);
1020 pt
->ilseq_handler
= handler
;
1024 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1025 1, 0, 0, (SCM port
),
1026 "Returns the behavior of the port when handling a character that\n"
1027 "is not representable in the port's current encoding.\n"
1028 "It returns the symbol @code{error} if unrepresentable characters\n"
1029 "should cause exceptions, @code{substitute} if the port should\n"
1030 "try to replace unrepresentable characters with question marks or\n"
1031 "approximate characters, or @code{escape} if unrepresentable\n"
1032 "characters should be converted to string escapes.\n"
1034 "If @var{port} is @code{#f}, then the current default behavior\n"
1035 "will be returned. New ports will have this default behavior\n"
1036 "when they are created.\n")
1037 #define FUNC_NAME s_scm_port_conversion_strategy
1039 scm_t_string_failed_conversion_handler h
;
1041 SCM_VALIDATE_OPPORT (1, port
);
1043 if (!scm_is_false (port
))
1045 SCM_VALIDATE_OPPORT (1, port
);
1048 h
= scm_i_get_conversion_strategy (port
);
1049 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1050 return scm_from_latin1_symbol ("error");
1051 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1052 return scm_from_latin1_symbol ("substitute");
1053 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1054 return scm_from_latin1_symbol ("escape");
1058 /* Never gets here. */
1059 return SCM_UNDEFINED
;
1063 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1065 (SCM port
, SCM sym
),
1066 "Sets the behavior of the interpreter when outputting a character\n"
1067 "that is not representable in the port's current encoding.\n"
1068 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1069 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1070 "when an unconvertible character is encountered. If it is\n"
1071 "@code{'substitute}, then unconvertible characters will \n"
1072 "be replaced with approximate characters, or with question marks\n"
1073 "if no approximately correct character is available.\n"
1074 "If it is @code{'escape},\n"
1075 "it will appear as a hex escape when output.\n"
1077 "If @var{port} is an open port, the conversion error behavior\n"
1078 "is set for that port. If it is @code{#f}, it is set as the\n"
1079 "default behavior for any future ports that get created in\n"
1081 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1087 if (!scm_is_false (port
))
1089 SCM_VALIDATE_OPPORT (1, port
);
1092 err
= scm_from_latin1_symbol ("error");
1093 if (scm_is_true (scm_eqv_p (sym
, err
)))
1095 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1096 return SCM_UNSPECIFIED
;
1099 qm
= scm_from_latin1_symbol ("substitute");
1100 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1102 scm_i_set_conversion_strategy_x (port
,
1103 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1104 return SCM_UNSPECIFIED
;
1107 esc
= scm_from_latin1_symbol ("escape");
1108 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1110 scm_i_set_conversion_strategy_x (port
,
1111 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1112 return SCM_UNSPECIFIED
;
1115 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1117 return SCM_UNSPECIFIED
;
1124 /* The port lock. */
1127 lock_port (void *mutex
)
1129 scm_i_pthread_mutex_lock (mutex
);
1133 unlock_port (void *mutex
)
1135 scm_i_pthread_mutex_unlock (mutex
);
1139 scm_dynwind_lock_port (SCM port
)
1140 #define FUNC_NAME "dynwind-lock-port"
1142 scm_i_pthread_mutex_t
*lock
;
1143 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1144 scm_c_lock_port (port
, &lock
);
1147 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1148 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1156 /* Revealed counts --- an oddity inherited from SCSH. */
1158 /* Find a port in the table and return its revealed count.
1159 Also used by the garbage collector.
1162 scm_revealed_count (SCM port
)
1164 scm_i_pthread_mutex_t
*lock
;
1167 scm_c_lock_port (port
, &lock
);
1168 ret
= SCM_REVEALED (port
);
1170 scm_i_pthread_mutex_unlock (lock
);
1175 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
1177 "Return the revealed count for @var{port}.")
1178 #define FUNC_NAME s_scm_port_revealed
1180 port
= SCM_COERCE_OUTPORT (port
);
1181 SCM_VALIDATE_OPENPORT (1, port
);
1182 return scm_from_int (scm_revealed_count (port
));
1186 /* Set the revealed count for a port. */
1187 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
1188 (SCM port
, SCM rcount
),
1189 "Sets the revealed count for a port to a given value.\n"
1190 "The return value is unspecified.")
1191 #define FUNC_NAME s_scm_set_port_revealed_x
1194 scm_i_pthread_mutex_t
*lock
;
1196 port
= SCM_COERCE_OUTPORT (port
);
1197 SCM_VALIDATE_OPENPORT (1, port
);
1198 r
= scm_to_int (rcount
);
1199 scm_c_lock_port (port
, &lock
);
1200 SCM_REVEALED (port
) = r
;
1202 scm_i_pthread_mutex_unlock (lock
);
1203 return SCM_UNSPECIFIED
;
1207 /* Set the revealed count for a port. */
1208 SCM_DEFINE (scm_adjust_port_revealed_x
, "adjust-port-revealed!", 2, 0, 0,
1209 (SCM port
, SCM addend
),
1210 "Add @var{addend} to the revealed count of @var{port}.\n"
1211 "The return value is unspecified.")
1212 #define FUNC_NAME s_scm_adjust_port_revealed_x
1214 scm_i_pthread_mutex_t
*lock
;
1216 port
= SCM_COERCE_OUTPORT (port
);
1217 SCM_VALIDATE_OPENPORT (1, port
);
1218 a
= scm_to_int (addend
);
1219 scm_c_lock_port (port
, &lock
);
1220 SCM_REVEALED (port
) += a
;
1222 scm_i_pthread_mutex_unlock (lock
);
1223 return SCM_UNSPECIFIED
;
1233 scm_get_byte_or_eof (SCM port
)
1235 scm_i_pthread_mutex_t
*lock
;
1238 scm_c_lock_port (port
, &lock
);
1239 ret
= scm_get_byte_or_eof_unlocked (port
);
1241 scm_i_pthread_mutex_unlock (lock
);
1247 scm_peek_byte_or_eof (SCM port
)
1249 scm_i_pthread_mutex_t
*lock
;
1252 scm_c_lock_port (port
, &lock
);
1253 ret
= scm_peek_byte_or_eof_unlocked (port
);
1255 scm_i_pthread_mutex_unlock (lock
);
1262 * Used by an application to read arbitrary number of bytes from an
1263 * SCM port. Same semantics as libc read, except that scm_c_read only
1264 * returns less than SIZE bytes if at end-of-file.
1266 * Warning: Doesn't update port line and column counts! */
1268 /* This structure, and the following swap_buffer function, are used
1269 for temporarily swapping a port's own read buffer, and the buffer
1270 that the caller of scm_c_read provides. */
1271 struct port_and_swap_buffer
1274 unsigned char *buffer
;
1279 swap_buffer (void *data
)
1281 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1282 unsigned char *old_buf
= psb
->pt
->read_buf
;
1283 size_t old_size
= psb
->pt
->read_buf_size
;
1285 /* Make the port use (buffer, size) from the struct. */
1286 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1287 psb
->pt
->read_buf_size
= psb
->size
;
1289 /* Save the port's old (buffer, size) in the struct. */
1290 psb
->buffer
= old_buf
;
1291 psb
->size
= old_size
;
1295 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1296 #define FUNC_NAME "scm_c_read"
1299 size_t n_read
= 0, n_available
;
1300 struct port_and_swap_buffer psb
;
1302 SCM_VALIDATE_OPINPORT (1, port
);
1304 pt
= SCM_PTAB_ENTRY (port
);
1305 if (pt
->rw_active
== SCM_PORT_WRITE
)
1306 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1309 pt
->rw_active
= SCM_PORT_READ
;
1311 /* Take bytes first from the port's read buffer. */
1312 if (pt
->read_pos
< pt
->read_end
)
1314 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1315 memcpy (buffer
, pt
->read_pos
, n_available
);
1316 buffer
= (char *) buffer
+ n_available
;
1317 pt
->read_pos
+= n_available
;
1318 n_read
+= n_available
;
1319 size
-= n_available
;
1322 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1326 /* Now we will call scm_fill_input repeatedly until we have read the
1327 requested number of bytes. (Note that a single scm_fill_input
1328 call does not guarantee to fill the whole of the port's read
1330 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1332 /* The port that we are reading from is unbuffered - i.e. does
1333 not have its own persistent buffer - but we have a buffer,
1334 provided by our caller, that is the right size for the data
1335 that is wanted. For the following scm_fill_input calls,
1336 therefore, we use the buffer in hand as the port's read
1339 We need to make sure that the port's normal (1 byte) buffer
1340 is reinstated in case one of the scm_fill_input () calls
1341 throws an exception; we use the scm_dynwind_* API to achieve
1344 A consequence of this optimization is that the fill_input
1345 functions can't unget characters. That'll push data to the
1346 pushback buffer instead of this psb buffer. */
1348 unsigned char *pback
= pt
->putback_buf
;
1351 psb
.buffer
= buffer
;
1353 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1354 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1355 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1357 /* Call scm_fill_input until we have all the bytes that we need,
1359 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1361 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1362 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1365 if (pback
!= pt
->putback_buf
1366 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1367 scm_misc_error (FUNC_NAME
,
1368 "scm_c_read must not call a fill function that pushes "
1369 "back characters onto an unbuffered port", SCM_EOL
);
1371 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1373 /* Reinstate the port's normal buffer. */
1378 /* The port has its own buffer. It is important that we use it,
1379 even if it happens to be smaller than our caller's buffer, so
1380 that a custom port implementation's entry points (in
1381 particular, fill_input) can rely on the buffer always being
1382 the same as they first set up. */
1383 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1385 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1386 memcpy (buffer
, pt
->read_pos
, n_available
);
1387 buffer
= (char *) buffer
+ n_available
;
1388 pt
->read_pos
+= n_available
;
1389 n_read
+= n_available
;
1390 size
-= n_available
;
1399 scm_c_read (SCM port
, void *buffer
, size_t size
)
1401 scm_i_pthread_mutex_t
*lock
;
1404 scm_c_lock_port (port
, &lock
);
1405 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1407 scm_i_pthread_mutex_unlock (lock
);
1413 /* Update the line and column number of PORT after consumption of C. */
1415 update_port_lf (scm_t_wchar c
, SCM port
)
1440 #define SCM_MBCHAR_BUF_SIZE (4)
1442 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1443 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1445 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1447 scm_t_wchar codepoint
;
1449 if (utf8_buf
[0] <= 0x7f)
1452 codepoint
= utf8_buf
[0];
1454 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1457 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1458 | (utf8_buf
[1] & 0x3f);
1460 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1463 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1464 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1465 | (utf8_buf
[2] & 0x3f);
1470 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1471 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1472 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1473 | (utf8_buf
[3] & 0x3f);
1479 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1480 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1481 representation, and set *LEN to the length in bytes. Return
1482 `EILSEQ' on error. */
1484 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1485 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1487 #define ASSERT_NOT_EOF(b) \
1488 if (SCM_UNLIKELY ((b) == EOF)) \
1490 #define CONSUME_PEEKED_BYTE() \
1497 pt
= SCM_PTAB_ENTRY (port
);
1499 byte
= scm_get_byte_or_eof_unlocked (port
);
1506 buf
[0] = (scm_t_uint8
) byte
;
1511 *codepoint
= buf
[0];
1512 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1515 byte
= scm_peek_byte_or_eof_unlocked (port
);
1516 ASSERT_NOT_EOF (byte
);
1518 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1521 CONSUME_PEEKED_BYTE ();
1522 buf
[1] = (scm_t_uint8
) byte
;
1525 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1528 else if ((buf
[0] & 0xf0) == 0xe0)
1531 byte
= scm_peek_byte_or_eof_unlocked (port
);
1532 ASSERT_NOT_EOF (byte
);
1534 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1535 || (buf
[0] == 0xe0 && byte
< 0xa0)
1536 || (buf
[0] == 0xed && byte
> 0x9f)))
1539 CONSUME_PEEKED_BYTE ();
1540 buf
[1] = (scm_t_uint8
) byte
;
1543 byte
= scm_peek_byte_or_eof_unlocked (port
);
1544 ASSERT_NOT_EOF (byte
);
1546 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1549 CONSUME_PEEKED_BYTE ();
1550 buf
[2] = (scm_t_uint8
) byte
;
1553 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1554 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1557 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1560 byte
= scm_peek_byte_or_eof_unlocked (port
);
1561 ASSERT_NOT_EOF (byte
);
1563 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1564 || (buf
[0] == 0xf0 && byte
< 0x90)
1565 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1568 CONSUME_PEEKED_BYTE ();
1569 buf
[1] = (scm_t_uint8
) byte
;
1572 byte
= scm_peek_byte_or_eof_unlocked (port
);
1573 ASSERT_NOT_EOF (byte
);
1575 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1578 CONSUME_PEEKED_BYTE ();
1579 buf
[2] = (scm_t_uint8
) byte
;
1582 byte
= scm_peek_byte_or_eof_unlocked (port
);
1583 ASSERT_NOT_EOF (byte
);
1585 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1588 CONSUME_PEEKED_BYTE ();
1589 buf
[3] = (scm_t_uint8
) byte
;
1592 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1593 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1594 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1603 /* Here we could choose the consume the faulty byte when it's not a
1604 valid starting byte, but it's not a requirement. What Section 3.9
1605 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1606 would otherwise be a valid starting byte. */
1610 #undef CONSUME_PEEKED_BYTE
1611 #undef ASSERT_NOT_EOF
1614 /* Likewise, read a byte sequence from PORT, passing it through its
1615 input conversion descriptor. */
1617 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1618 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1622 size_t bytes_consumed
, output_size
;
1624 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1626 pt
= SCM_PTAB_ENTRY (port
);
1628 for (output_size
= 0, output
= (char *) utf8_buf
,
1629 bytes_consumed
= 0, err
= 0;
1630 err
== 0 && output_size
== 0
1631 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1635 size_t input_left
, output_left
, done
;
1637 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1638 if (byte_read
== EOF
)
1640 if (bytes_consumed
== 0)
1642 *codepoint
= (scm_t_wchar
) EOF
;
1650 buf
[bytes_consumed
] = byte_read
;
1653 input_left
= bytes_consumed
+ 1;
1654 output_left
= sizeof (utf8_buf
);
1656 done
= iconv (pt
->input_cd
, &input
, &input_left
,
1657 &output
, &output_left
);
1658 if (done
== (size_t) -1)
1662 /* Missing input: keep trying. */
1666 output_size
= sizeof (utf8_buf
) - output_left
;
1669 if (SCM_UNLIKELY (output_size
== 0))
1670 /* An unterminated sequence. */
1672 else if (SCM_LIKELY (err
== 0))
1674 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1675 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1676 *len
= bytes_consumed
;
1682 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1683 with the byte representation of the codepoint in PORT's encoding, and
1684 set *LEN to the length in bytes of that representation. Return 0 on
1685 success and an errno value on error. */
1687 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1688 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1691 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1693 if (pt
->input_cd
== (iconv_t
) -1)
1694 /* Initialize the conversion descriptors, if needed. */
1695 scm_i_set_port_encoding_x (port
, pt
->encoding
);
1697 /* FIXME: In 2.1, add a flag to determine whether a port is UTF-8. */
1698 if (pt
->input_cd
== (iconv_t
) -1)
1699 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1701 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1703 if (SCM_LIKELY (err
== 0))
1704 update_port_lf (*codepoint
, port
);
1705 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1709 update_port_lf (*codepoint
, port
);
1715 /* Read a codepoint from PORT and return it. */
1717 scm_getc_unlocked (SCM port
)
1718 #define FUNC_NAME "scm_getc"
1722 scm_t_wchar codepoint
;
1723 char buf
[SCM_MBCHAR_BUF_SIZE
];
1725 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1726 if (SCM_UNLIKELY (err
!= 0))
1727 /* At this point PORT should point past the invalid encoding, as per
1728 R6RS-lib Section 8.2.4. */
1729 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1738 scm_i_pthread_mutex_t
*lock
;
1741 scm_c_lock_port (port
, &lock
);
1742 ret
= scm_getc_unlocked (port
);
1744 scm_i_pthread_mutex_unlock (lock
);
1750 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1752 "Return the next character available from @var{port}, updating\n"
1753 "@var{port} to point to the following character. If no more\n"
1754 "characters are available, the end-of-file object is returned.\n"
1756 "When @var{port}'s data cannot be decoded according to its\n"
1757 "character encoding, a @code{decoding-error} is raised and\n"
1758 "@var{port} points past the erroneous byte sequence.\n")
1759 #define FUNC_NAME s_scm_read_char
1762 if (SCM_UNBNDP (port
))
1763 port
= scm_current_input_port ();
1764 SCM_VALIDATE_OPINPORT (1, port
);
1765 c
= scm_getc_unlocked (port
);
1768 return SCM_MAKE_CHAR (c
);
1778 scm_unget_byte_unlocked (int c
, SCM port
)
1779 #define FUNC_NAME "scm_unget_byte"
1781 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1783 if (pt
->read_buf
== pt
->putback_buf
)
1784 /* already using the put-back buffer. */
1786 /* enlarge putback_buf if necessary. */
1787 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1788 && pt
->read_buf
== pt
->read_pos
)
1790 size_t new_size
= pt
->read_buf_size
* 2;
1791 unsigned char *tmp
= (unsigned char *)
1792 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1795 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1796 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1797 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1800 /* shift any existing bytes to buffer + 1. */
1801 if (pt
->read_pos
== pt
->read_end
)
1802 pt
->read_end
= pt
->read_buf
+ 1;
1803 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1805 int count
= pt
->read_end
- pt
->read_pos
;
1807 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1808 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1811 pt
->read_pos
= pt
->read_buf
;
1814 /* switch to the put-back buffer. */
1816 if (pt
->putback_buf
== NULL
)
1819 = (unsigned char *) scm_gc_malloc_pointerless
1820 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1821 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1824 pt
->saved_read_buf
= pt
->read_buf
;
1825 pt
->saved_read_pos
= pt
->read_pos
;
1826 pt
->saved_read_end
= pt
->read_end
;
1827 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1829 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1830 pt
->read_end
= pt
->read_buf
+ 1;
1831 pt
->read_buf_size
= pt
->putback_buf_size
;
1837 pt
->rw_active
= SCM_PORT_READ
;
1842 scm_unget_byte (int c
, SCM port
)
1844 scm_i_pthread_mutex_t
*lock
;
1845 scm_c_lock_port (port
, &lock
);
1846 scm_unget_byte_unlocked (c
, port
);
1848 scm_i_pthread_mutex_unlock (lock
);
1853 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1854 #define FUNC_NAME "scm_ungetc"
1856 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1858 char result_buf
[10];
1859 const char *encoding
;
1863 if (pt
->encoding
!= NULL
)
1864 encoding
= pt
->encoding
;
1866 encoding
= "ISO-8859-1";
1868 len
= sizeof (result_buf
);
1869 result
= u32_conv_to_encoding (encoding
,
1870 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1871 (uint32_t *) &c
, 1, NULL
,
1874 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1875 scm_encoding_error (FUNC_NAME
, errno
,
1876 "conversion to port encoding failed",
1877 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1879 for (i
= len
- 1; i
>= 0; i
--)
1880 scm_unget_byte_unlocked (result
[i
], port
);
1882 if (SCM_UNLIKELY (result
!= result_buf
))
1887 /* What should col be in this case?
1888 * We'll leave it at -1.
1890 SCM_LINUM (port
) -= 1;
1898 scm_ungetc (scm_t_wchar c
, SCM port
)
1900 scm_i_pthread_mutex_t
*lock
;
1901 scm_c_lock_port (port
, &lock
);
1902 scm_ungetc_unlocked (c
, port
);
1904 scm_i_pthread_mutex_unlock (lock
);
1909 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1911 /* This is simple minded and inefficient, but unreading strings is
1912 * probably not a common operation, and remember that line and
1913 * column numbers have to be handled...
1915 * Please feel free to write an optimized version!
1918 scm_ungetc_unlocked (s
[n
], port
);
1922 scm_ungets (const char *s
, int n
, SCM port
)
1924 scm_i_pthread_mutex_t
*lock
;
1925 scm_c_lock_port (port
, &lock
);
1926 scm_ungets_unlocked (s
, n
, port
);
1928 scm_i_pthread_mutex_unlock (lock
);
1932 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1934 "Return the next character available from @var{port},\n"
1935 "@emph{without} updating @var{port} to point to the following\n"
1936 "character. If no more characters are available, the\n"
1937 "end-of-file object is returned.\n"
1939 "The value returned by\n"
1940 "a call to @code{peek-char} is the same as the value that would\n"
1941 "have been returned by a call to @code{read-char} on the same\n"
1942 "port. The only difference is that the very next call to\n"
1943 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1944 "return the value returned by the preceding call to\n"
1945 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1946 "an interactive port will hang waiting for input whenever a call\n"
1947 "to @code{read-char} would have hung.\n"
1949 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1950 "if such a situation occurs. However, unlike with @code{read-char},\n"
1951 "@var{port} still points at the beginning of the erroneous byte\n"
1952 "sequence when the error is raised.\n")
1953 #define FUNC_NAME s_scm_peek_char
1958 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1959 long column
, line
, i
;
1962 if (SCM_UNBNDP (port
))
1963 port
= scm_current_input_port ();
1964 SCM_VALIDATE_OPINPORT (1, port
);
1966 column
= SCM_COL (port
);
1967 line
= SCM_LINUM (port
);
1969 err
= get_codepoint (port
, &c
, bytes
, &len
);
1971 for (i
= len
- 1; i
>= 0; i
--)
1972 scm_unget_byte_unlocked (bytes
[i
], port
);
1974 SCM_COL (port
) = column
;
1975 SCM_LINUM (port
) = line
;
1977 if (SCM_UNLIKELY (err
!= 0))
1979 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1981 /* Shouldn't happen since `catch' always aborts to prompt. */
1982 result
= SCM_BOOL_F
;
1985 result
= SCM_EOF_VAL
;
1987 result
= SCM_MAKE_CHAR (c
);
1993 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
1994 (SCM cobj
, SCM port
),
1995 "Place @var{char} in @var{port} so that it will be read by the\n"
1996 "next read operation. If called multiple times, the unread characters\n"
1997 "will be read again in last-in first-out order. If @var{port} is\n"
1998 "not supplied, the current input port is used.")
1999 #define FUNC_NAME s_scm_unread_char
2003 SCM_VALIDATE_CHAR (1, cobj
);
2004 if (SCM_UNBNDP (port
))
2005 port
= scm_current_input_port ();
2006 SCM_VALIDATE_OPINPORT (2, port
);
2008 c
= SCM_CHAR (cobj
);
2010 scm_ungetc_unlocked (c
, port
);
2015 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2016 (SCM str
, SCM port
),
2017 "Place the string @var{str} in @var{port} so that its characters will be\n"
2018 "read in subsequent read operations. If called multiple times, the\n"
2019 "unread characters will be read again in last-in first-out order. If\n"
2020 "@var{port} is not supplied, the current-input-port is used.")
2021 #define FUNC_NAME s_scm_unread_string
2024 SCM_VALIDATE_STRING (1, str
);
2025 if (SCM_UNBNDP (port
))
2026 port
= scm_current_input_port ();
2027 SCM_VALIDATE_OPINPORT (2, port
);
2029 n
= scm_i_string_length (str
);
2032 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2041 /* Manipulating the buffers. */
2043 /* This routine does not take any locks, as it is usually called as part
2044 of a port implementation. */
2046 scm_port_non_buffer (scm_t_port
*pt
)
2048 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2049 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2050 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2051 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2054 /* this should only be called when the read buffer is empty. it
2055 tries to refill the read buffer. it returns the first char from
2056 the port, which is either EOF or *(pt->read_pos). */
2058 scm_fill_input_unlocked (SCM port
)
2060 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2062 assert (pt
->read_pos
== pt
->read_end
);
2064 if (pt
->read_buf
== pt
->putback_buf
)
2066 /* finished reading put-back chars. */
2067 pt
->read_buf
= pt
->saved_read_buf
;
2068 pt
->read_pos
= pt
->saved_read_pos
;
2069 pt
->read_end
= pt
->saved_read_end
;
2070 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2071 if (pt
->read_pos
< pt
->read_end
)
2072 return *(pt
->read_pos
);
2074 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2078 scm_fill_input (SCM port
)
2080 scm_i_pthread_mutex_t
*lock
;
2083 scm_c_lock_port (port
, &lock
);
2084 ret
= scm_fill_input_unlocked (port
);
2086 scm_i_pthread_mutex_unlock (lock
);
2092 /* move up to read_len chars from port's putback and/or read buffers
2093 into memory starting at dest. returns the number of chars moved. */
2095 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2097 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2098 size_t chars_read
= 0;
2099 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2103 memcpy (dest
, pt
->read_pos
, from_buf
);
2104 pt
->read_pos
+= from_buf
;
2105 chars_read
+= from_buf
;
2106 read_len
-= from_buf
;
2110 /* if putback was active, try the real input buffer too. */
2111 if (pt
->read_buf
== pt
->putback_buf
)
2113 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2116 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2117 pt
->saved_read_pos
+= from_buf
;
2118 chars_read
+= from_buf
;
2124 /* Clear a port's read buffers, returning the contents. */
2125 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2127 "This procedure clears a port's input buffers, similar\n"
2128 "to the way that force-output clears the output buffer. The\n"
2129 "contents of the buffers are returned as a single string, e.g.,\n"
2132 "(define p (open-input-file ...))\n"
2133 "(drain-input p) => empty string, nothing buffered yet.\n"
2134 "(unread-char (read-char p) p)\n"
2135 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2137 "Draining the buffers may be useful for cleanly finishing\n"
2138 "buffered I/O so that the file descriptor can be used directly\n"
2139 "for further input.")
2140 #define FUNC_NAME s_scm_drain_input
2147 SCM_VALIDATE_OPINPORT (1, port
);
2148 pt
= SCM_PTAB_ENTRY (port
);
2150 count
= pt
->read_end
- pt
->read_pos
;
2151 if (pt
->read_buf
== pt
->putback_buf
)
2152 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2156 result
= scm_i_make_string (count
, &data
, 0);
2157 scm_take_from_input_buffers (port
, data
, count
);
2160 result
= scm_nullstr
;
2167 scm_end_input_unlocked (SCM port
)
2170 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2172 if (pt
->read_buf
== pt
->putback_buf
)
2174 offset
= pt
->read_end
- pt
->read_pos
;
2175 pt
->read_buf
= pt
->saved_read_buf
;
2176 pt
->read_pos
= pt
->saved_read_pos
;
2177 pt
->read_end
= pt
->saved_read_end
;
2178 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2183 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2187 scm_end_input (SCM port
)
2189 scm_i_pthread_mutex_t
*lock
;
2190 scm_c_lock_port (port
, &lock
);
2191 scm_end_input_unlocked (port
);
2193 scm_i_pthread_mutex_unlock (lock
);
2197 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2199 "Flush the specified output port, or the current output port if @var{port}\n"
2200 "is omitted. The current output buffer contents are passed to the\n"
2201 "underlying port implementation (e.g., in the case of fports, the\n"
2202 "data will be written to the file and the output buffer will be cleared.)\n"
2203 "It has no effect on an unbuffered port.\n\n"
2204 "The return value is unspecified.")
2205 #define FUNC_NAME s_scm_force_output
2207 if (SCM_UNBNDP (port
))
2208 port
= scm_current_output_port ();
2211 port
= SCM_COERCE_OUTPORT (port
);
2212 SCM_VALIDATE_OPOUTPORT (1, port
);
2214 scm_flush_unlocked (port
);
2215 return SCM_UNSPECIFIED
;
2220 scm_flush_unlocked (SCM port
)
2222 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2226 scm_flush (SCM port
)
2228 scm_i_pthread_mutex_t
*lock
;
2229 scm_c_lock_port (port
, &lock
);
2230 scm_flush_unlocked (port
);
2232 scm_i_pthread_mutex_unlock (lock
);
2242 scm_putc (char c
, SCM port
)
2244 scm_i_pthread_mutex_t
*lock
;
2245 scm_c_lock_port (port
, &lock
);
2246 scm_putc_unlocked (c
, port
);
2248 scm_i_pthread_mutex_unlock (lock
);
2253 scm_puts (const char *s
, SCM port
)
2255 scm_i_pthread_mutex_t
*lock
;
2256 scm_c_lock_port (port
, &lock
);
2257 scm_puts_unlocked (s
, port
);
2259 scm_i_pthread_mutex_unlock (lock
);
2265 * Used by an application to write arbitrary number of bytes to an SCM
2266 * port. Similar semantics as libc write. However, unlike libc
2267 * write, scm_c_write writes the requested number of bytes and has no
2270 * Warning: Doesn't update port line and column counts!
2273 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2274 #define FUNC_NAME "scm_c_write"
2277 scm_t_ptob_descriptor
*ptob
;
2279 SCM_VALIDATE_OPOUTPORT (1, port
);
2281 pt
= SCM_PTAB_ENTRY (port
);
2282 ptob
= SCM_PORT_DESCRIPTOR (port
);
2284 if (pt
->rw_active
== SCM_PORT_READ
)
2285 scm_end_input_unlocked (port
);
2287 ptob
->write (port
, ptr
, size
);
2290 pt
->rw_active
= SCM_PORT_WRITE
;
2295 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2297 scm_i_pthread_mutex_t
*lock
;
2298 scm_c_lock_port (port
, &lock
);
2299 scm_c_write_unlocked (port
, ptr
, size
);
2301 scm_i_pthread_mutex_unlock (lock
);
2307 * This function differs from scm_c_write; it updates port line and
2310 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2312 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2313 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2315 if (pt
->rw_active
== SCM_PORT_READ
)
2316 scm_end_input_unlocked (port
);
2318 ptob
->write (port
, ptr
, size
);
2320 for (; size
; ptr
++, size
--)
2321 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2324 pt
->rw_active
= SCM_PORT_WRITE
;
2328 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2330 scm_i_pthread_mutex_t
*lock
;
2331 scm_c_lock_port (port
, &lock
);
2332 scm_lfwrite_unlocked (ptr
, size
, port
);
2334 scm_i_pthread_mutex_unlock (lock
);
2338 /* Write STR to PORT from START inclusive to END exclusive. */
2340 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2342 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2344 if (pt
->rw_active
== SCM_PORT_READ
)
2345 scm_end_input_unlocked (port
);
2347 if (end
== (size_t) -1)
2348 end
= scm_i_string_length (str
);
2350 scm_display (scm_c_substring (str
, start
, end
), port
);
2353 pt
->rw_active
= SCM_PORT_WRITE
;
2359 /* Querying and setting positions, and character availability. */
2361 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2363 "Return @code{#t} if a character is ready on input @var{port}\n"
2364 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2365 "@code{#t} then the next @code{read-char} operation on\n"
2366 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2367 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2369 "@code{char-ready?} exists to make it possible for a\n"
2370 "program to accept characters from interactive ports without\n"
2371 "getting stuck waiting for input. Any input editors associated\n"
2372 "with such ports must make sure that characters whose existence\n"
2373 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2374 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2375 "a port at end of file would be indistinguishable from an\n"
2376 "interactive port that has no ready characters.")
2377 #define FUNC_NAME s_scm_char_ready_p
2381 if (SCM_UNBNDP (port
))
2382 port
= scm_current_input_port ();
2383 /* It's possible to close the current input port, so validate even in
2385 SCM_VALIDATE_OPINPORT (1, port
);
2387 pt
= SCM_PTAB_ENTRY (port
);
2389 /* if the current read buffer is filled, or the
2390 last pushed-back char has been read and the saved buffer is
2391 filled, result is true. */
2392 if (pt
->read_pos
< pt
->read_end
2393 || (pt
->read_buf
== pt
->putback_buf
2394 && pt
->saved_read_pos
< pt
->saved_read_end
))
2398 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2400 if (ptob
->input_waiting
)
2401 return scm_from_bool(ptob
->input_waiting (port
));
2408 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2409 (SCM fd_port
, SCM offset
, SCM whence
),
2410 "Sets the current position of @var{fd/port} to the integer\n"
2411 "@var{offset}, which is interpreted according to the value of\n"
2414 "One of the following variables should be supplied for\n"
2416 "@defvar SEEK_SET\n"
2417 "Seek from the beginning of the file.\n"
2419 "@defvar SEEK_CUR\n"
2420 "Seek from the current position.\n"
2422 "@defvar SEEK_END\n"
2423 "Seek from the end of the file.\n"
2425 "If @var{fd/port} is a file descriptor, the underlying system\n"
2426 "call is @code{lseek}. @var{port} may be a string port.\n"
2428 "The value returned is the new position in the file. This means\n"
2429 "that the current position of a port can be obtained using:\n"
2431 "(seek port 0 SEEK_CUR)\n"
2433 #define FUNC_NAME s_scm_seek
2437 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2439 how
= scm_to_int (whence
);
2440 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2441 SCM_OUT_OF_RANGE (3, whence
);
2443 if (SCM_OPPORTP (fd_port
))
2445 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2446 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2447 off_t_or_off64_t rv
;
2450 SCM_MISC_ERROR ("port is not seekable",
2451 scm_cons (fd_port
, SCM_EOL
));
2453 rv
= ptob
->seek (fd_port
, off
, how
);
2454 return scm_from_off_t_or_off64_t (rv
);
2456 else /* file descriptor?. */
2458 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2459 off_t_or_off64_t rv
;
2460 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2463 return scm_from_off_t_or_off64_t (rv
);
2472 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2473 doesn't have the filename version truncate(), hence this code. */
2474 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2476 truncate (const char *file
, off_t length
)
2480 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2484 ret
= ftruncate (fdes
, length
);
2487 int save_errno
= errno
;
2493 return close (fdes
);
2495 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2497 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2498 (SCM object
, SCM length
),
2499 "Truncate @var{file} to @var{length} bytes. @var{file} can be a\n"
2500 "filename string, a port object, or an integer file descriptor.\n"
2501 "The return value is unspecified.\n"
2503 "For a port or file descriptor @var{length} can be omitted, in\n"
2504 "which case the file is truncated at the current position (per\n"
2505 "@code{ftell} above).\n"
2507 "On most systems a file can be extended by giving a length\n"
2508 "greater than the current size, but this is not mandatory in the\n"
2510 #define FUNC_NAME s_scm_truncate_file
2514 /* "object" can be a port, fdes or filename.
2516 Negative "length" makes no sense, but it's left to truncate() or
2517 ftruncate() to give back an error for that (normally EINVAL).
2520 if (SCM_UNBNDP (length
))
2522 /* must supply length if object is a filename. */
2523 if (scm_is_string (object
))
2524 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2526 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2529 object
= SCM_COERCE_OUTPORT (object
);
2530 if (scm_is_integer (object
))
2532 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2533 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2536 else if (SCM_OPOUTPORTP (object
))
2538 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2539 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2540 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2542 if (!ptob
->truncate
)
2543 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2544 if (pt
->rw_active
== SCM_PORT_READ
)
2545 scm_end_input_unlocked (object
);
2546 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2547 ptob
->flush (object
);
2549 ptob
->truncate (object
, c_length
);
2554 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2555 char *str
= scm_to_locale_string (object
);
2557 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2564 return SCM_UNSPECIFIED
;
2568 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2570 "Return the current line number for @var{port}.\n"
2572 "The first line of a file is 0. But you might want to add 1\n"
2573 "when printing line numbers, since starting from 1 is\n"
2574 "traditional in error messages, and likely to be more natural to\n"
2576 #define FUNC_NAME s_scm_port_line
2578 port
= SCM_COERCE_OUTPORT (port
);
2579 SCM_VALIDATE_OPENPORT (1, port
);
2580 return scm_from_long (SCM_LINUM (port
));
2584 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2585 (SCM port
, SCM line
),
2586 "Set the current line number for @var{port} to @var{line}. The\n"
2587 "first line of a file is 0.")
2588 #define FUNC_NAME s_scm_set_port_line_x
2590 port
= SCM_COERCE_OUTPORT (port
);
2591 SCM_VALIDATE_OPENPORT (1, port
);
2592 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2593 return SCM_UNSPECIFIED
;
2597 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2599 "Return the current column number of @var{port}.\n"
2600 "If the number is\n"
2601 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2602 "- i.e. the first character of the first line is line 0, column 0.\n"
2603 "(However, when you display a file position, for example in an error\n"
2604 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2605 "because lines and column numbers traditionally start with 1, and that is\n"
2606 "what non-programmers will find most natural.)")
2607 #define FUNC_NAME s_scm_port_column
2609 port
= SCM_COERCE_OUTPORT (port
);
2610 SCM_VALIDATE_OPENPORT (1, port
);
2611 return scm_from_int (SCM_COL (port
));
2615 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2616 (SCM port
, SCM column
),
2617 "Set the current column of @var{port}. Before reading the first\n"
2618 "character on a line the column should be 0.")
2619 #define FUNC_NAME s_scm_set_port_column_x
2621 port
= SCM_COERCE_OUTPORT (port
);
2622 SCM_VALIDATE_OPENPORT (1, port
);
2623 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2624 return SCM_UNSPECIFIED
;
2628 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2630 "Return the filename associated with @var{port}, or @code{#f}\n"
2631 "if no filename is associated with the port.")
2632 #define FUNC_NAME s_scm_port_filename
2634 port
= SCM_COERCE_OUTPORT (port
);
2635 SCM_VALIDATE_OPENPORT (1, port
);
2636 return SCM_FILENAME (port
);
2640 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2641 (SCM port
, SCM filename
),
2642 "Change the filename associated with @var{port}, using the current input\n"
2643 "port if none is specified. Note that this does not change the port's\n"
2644 "source of data, but only the value that is returned by\n"
2645 "@code{port-filename} and reported in diagnostic output.")
2646 #define FUNC_NAME s_scm_set_port_filename_x
2648 port
= SCM_COERCE_OUTPORT (port
);
2649 SCM_VALIDATE_OPENPORT (1, port
);
2650 /* We allow the user to set the filename to whatever he likes. */
2651 SCM_SET_FILENAME (port
, filename
);
2652 return SCM_UNSPECIFIED
;
2659 /* Implementation helpers for port printing functions. */
2662 scm_print_port_mode (SCM exp
, SCM port
)
2664 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2666 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2667 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2670 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2677 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2679 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2682 scm_puts_unlocked ("#<", port
);
2683 scm_print_port_mode (exp
, port
);
2684 scm_puts_unlocked (type
, port
);
2685 scm_putc_unlocked (' ', port
);
2686 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2687 scm_putc_unlocked ('>', port
);
2694 /* Iterating over all ports. */
2696 struct for_each_data
2698 void (*proc
) (void *data
, SCM p
);
2703 for_each_trampoline (void *data
, SCM port
, SCM result
)
2705 struct for_each_data
*d
= data
;
2707 d
->proc (d
->data
, port
);
2713 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2715 struct for_each_data d
;
2720 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2721 scm_i_port_weak_set
);
2725 scm_for_each_trampoline (void *data
, SCM port
)
2727 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2730 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2732 "Apply @var{proc} to each port in the Guile port table\n"
2733 "in turn. The return value is unspecified. More specifically,\n"
2734 "@var{proc} is applied exactly once to every port that exists\n"
2735 "in the system at the time @var{port-for-each} is invoked.\n"
2736 "Changes to the port table while @var{port-for-each} is running\n"
2737 "have no effect as far as @var{port-for-each} is concerned.")
2738 #define FUNC_NAME s_scm_port_for_each
2740 SCM_VALIDATE_PROC (1, proc
);
2742 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2744 return SCM_UNSPECIFIED
;
2749 flush_output_port (void *closure
, SCM port
)
2751 if (SCM_OPOUTPORTP (port
))
2752 scm_flush_unlocked (port
);
2755 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2757 "Equivalent to calling @code{force-output} on\n"
2758 "all open output ports. The return value is unspecified.")
2759 #define FUNC_NAME s_scm_flush_all_ports
2761 scm_c_port_for_each (&flush_output_port
, NULL
);
2762 return SCM_UNSPECIFIED
;
2771 scm_t_bits scm_tc16_void_port
= 0;
2773 static int fill_input_void_port (SCM port SCM_UNUSED
)
2779 write_void_port (SCM port SCM_UNUSED
,
2780 const void *data SCM_UNUSED
,
2781 size_t size SCM_UNUSED
)
2786 scm_i_void_port (long mode_bits
)
2790 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2792 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2798 scm_void_port (char *mode_str
)
2800 return scm_i_void_port (scm_mode_bits (mode_str
));
2803 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2805 "Create and return a new void port. A void port acts like\n"
2806 "@file{/dev/null}. The @var{mode} argument\n"
2807 "specifies the input/output modes for this port: see the\n"
2808 "documentation for @code{open-file} in @ref{File Ports}.")
2809 #define FUNC_NAME s_scm_sys_make_void_port
2811 return scm_i_void_port (scm_i_mode_bits (mode
));
2818 /* Initialization. */
2823 /* lseek() symbols. */
2824 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2825 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2826 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2828 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2831 cur_inport_fluid
= scm_make_fluid ();
2832 cur_outport_fluid
= scm_make_fluid ();
2833 cur_errport_fluid
= scm_make_fluid ();
2834 cur_loadport_fluid
= scm_make_fluid ();
2836 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2838 #include "libguile/ports.x"
2840 /* Use Latin-1 as the default port encoding. */
2841 SCM_VARIABLE_SET (default_port_encoding_var
,
2842 scm_make_fluid_with_default (SCM_BOOL_F
));
2843 scm_port_encoding_init
= 1;
2845 SCM_VARIABLE_SET (scm_conversion_strategy
,
2846 scm_make_fluid_with_default
2847 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2848 scm_conversion_strategy_init
= 1;
2850 /* These bindings are used when boot-9 turns `current-input-port' et
2851 al into parameters. They are then removed from the guile module. */
2852 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2853 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2854 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);