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
);
331 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
333 "Return the current-load-port.\n"
334 "The load port is used internally by @code{primitive-load}.")
335 #define FUNC_NAME s_scm_current_load_port
337 return scm_fluid_ref (cur_loadport_fluid
);
341 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
343 "@deffnx {Scheme Procedure} set-current-output-port port\n"
344 "@deffnx {Scheme Procedure} set-current-error-port port\n"
345 "Change the ports returned by @code{current-input-port},\n"
346 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
347 "so that they use the supplied @var{port} for input or output.")
348 #define FUNC_NAME s_scm_set_current_input_port
350 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
351 SCM_VALIDATE_OPINPORT (1, port
);
352 scm_fluid_set_x (cur_inport_fluid
, port
);
358 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
360 "Set the current default output port to @var{port}.")
361 #define FUNC_NAME s_scm_set_current_output_port
363 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
364 port
= SCM_COERCE_OUTPORT (port
);
365 SCM_VALIDATE_OPOUTPORT (1, port
);
366 scm_fluid_set_x (cur_outport_fluid
, port
);
372 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
374 "Set the current default error port to @var{port}.")
375 #define FUNC_NAME s_scm_set_current_error_port
377 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
378 port
= SCM_COERCE_OUTPORT (port
);
379 SCM_VALIDATE_OPOUTPORT (1, port
);
380 scm_fluid_set_x (cur_errport_fluid
, port
);
386 scm_dynwind_current_input_port (SCM port
)
387 #define FUNC_NAME NULL
389 SCM_VALIDATE_OPINPORT (1, port
);
390 scm_dynwind_fluid (cur_inport_fluid
, port
);
395 scm_dynwind_current_output_port (SCM port
)
396 #define FUNC_NAME NULL
398 port
= SCM_COERCE_OUTPORT (port
);
399 SCM_VALIDATE_OPOUTPORT (1, port
);
400 scm_dynwind_fluid (cur_outport_fluid
, port
);
405 scm_dynwind_current_error_port (SCM port
)
406 #define FUNC_NAME NULL
408 port
= SCM_COERCE_OUTPORT (port
);
409 SCM_VALIDATE_OPOUTPORT (1, port
);
410 scm_dynwind_fluid (cur_errport_fluid
, port
);
415 scm_i_dynwind_current_load_port (SCM port
)
417 scm_dynwind_fluid (cur_loadport_fluid
, port
);
423 /* Retrieving a port's mode. */
425 /* Return the flags that characterize a port based on the mode
426 * string used to open a file for that port.
428 * See PORT FLAGS in scm.h
432 scm_i_mode_bits_n (SCM modes
)
435 | (scm_i_string_contains_char (modes
, 'r')
436 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
437 | (scm_i_string_contains_char (modes
, 'w')
438 || scm_i_string_contains_char (modes
, 'a')
439 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
440 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
441 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
445 scm_mode_bits (char *modes
)
447 return scm_i_mode_bits (scm_from_locale_string (modes
));
451 scm_i_mode_bits (SCM modes
)
455 if (!scm_is_string (modes
))
456 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
458 bits
= scm_i_mode_bits_n (modes
);
459 scm_remember_upto_here_1 (modes
);
463 /* Return the mode flags from an open port.
464 * Some modes such as "append" are only used when opening
465 * a file and are not returned here. */
467 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
469 "Return the port modes associated with the open port @var{port}.\n"
470 "These will not necessarily be identical to the modes used when\n"
471 "the port was opened, since modes such as \"append\" which are\n"
472 "used only during port creation are not retained.")
473 #define FUNC_NAME s_scm_port_mode
478 port
= SCM_COERCE_OUTPORT (port
);
479 SCM_VALIDATE_OPPORT (1, port
);
480 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
481 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
482 strcpy (modes
, "r+");
486 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
488 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
491 return scm_from_latin1_string (modes
);
498 /* The port table --- a weak set of all ports.
500 We need a global registry of ports to flush them all at exit, and to
501 get all the ports matching a file descriptor. */
502 SCM scm_i_port_weak_set
;
507 /* Port finalization. */
509 static void finalize_port (GC_PTR
, GC_PTR
);
511 /* Register a finalizer for PORT. */
512 static SCM_C_INLINE_KEYWORD
void
513 register_finalizer_for_port (SCM port
)
515 GC_finalization_proc prev_finalizer
;
516 GC_PTR prev_finalization_data
;
518 /* Register a finalizer for PORT so that its iconv CDs get freed and
519 optionally its type's `free' function gets called. */
520 GC_REGISTER_FINALIZER_NO_ORDER (SCM_HEAP_OBJECT_BASE (port
),
523 &prev_finalization_data
);
526 /* Finalize the object (a port) pointed to by PTR. */
528 finalize_port (GC_PTR ptr
, GC_PTR data
)
530 SCM port
= SCM_PACK_POINTER (ptr
);
532 if (!SCM_PORTP (port
))
535 if (SCM_OPENP (port
))
537 if (SCM_REVEALED (port
) > 0)
538 /* Keep "revealed" ports alive and re-register a finalizer. */
539 register_finalizer_for_port (port
);
542 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
546 /* Yes, I really do mean `free' rather than `close'. `close'
547 is for explicit `close-port' by user. */
550 entry
= SCM_PTAB_ENTRY (port
);
552 if (entry
->input_cd
!= (iconv_t
) -1)
553 iconv_close (entry
->input_cd
);
554 if (entry
->output_cd
!= (iconv_t
) -1)
555 iconv_close (entry
->output_cd
);
557 SCM_SETSTREAM (port
, 0);
558 SCM_CLR_PORT_OPEN_FLAG (port
);
560 scm_gc_ports_collected
++;
569 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
570 const char *encoding
,
571 scm_t_string_failed_conversion_handler handler
,
576 scm_t_ptob_descriptor
*ptob
;
578 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
579 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
581 ret
= scm_words (tag
| mode_bits
, 3);
582 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
583 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
585 #if SCM_USE_PTHREAD_THREADS
586 scm_i_pthread_mutex_init (&entry
->lock
, scm_i_pthread_mutexattr_recursive
);
589 entry
->file_name
= SCM_BOOL_F
;
590 entry
->rw_active
= SCM_PORT_NEITHER
;
592 entry
->stream
= stream
;
593 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
594 /* The conversion descriptors will be opened lazily. */
595 entry
->input_cd
= (iconv_t
) -1;
596 entry
->output_cd
= (iconv_t
) -1;
597 entry
->ilseq_handler
= handler
;
599 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
601 /* For each new port, register a finalizer so that it port type's free
602 function can be invoked eventually. */
603 register_finalizer_for_port (ret
);
609 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
611 return scm_c_make_port_with_encoding (tag
, mode_bits
,
612 scm_i_default_port_encoding (),
613 scm_i_get_conversion_strategy (SCM_BOOL_F
),
618 scm_new_port_table_entry (scm_t_bits tag
)
620 return scm_c_make_port (tag
, 0, 0);
623 /* Remove a port from the table and destroy it. */
626 scm_i_remove_port (SCM port
)
627 #define FUNC_NAME "scm_remove_port"
631 p
= SCM_PTAB_ENTRY (port
);
632 scm_port_non_buffer (p
);
633 SCM_SETPTAB_ENTRY (port
, 0);
634 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
636 p
->putback_buf
= NULL
;
637 p
->putback_buf_size
= 0;
639 if (p
->input_cd
!= (iconv_t
) -1)
641 iconv_close (p
->input_cd
);
642 p
->input_cd
= (iconv_t
) -1;
645 if (p
->output_cd
!= (iconv_t
) -1)
647 iconv_close (p
->output_cd
);
648 p
->output_cd
= (iconv_t
) -1;
658 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
660 "Return a boolean indicating whether @var{x} is a port.\n"
661 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
663 #define FUNC_NAME s_scm_port_p
665 return scm_from_bool (SCM_PORTP (x
));
669 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
671 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
672 "@code{#f}. Any object satisfying this predicate also satisfies\n"
674 #define FUNC_NAME s_scm_input_port_p
676 return scm_from_bool (SCM_INPUT_PORT_P (x
));
680 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
682 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
683 "@code{#f}. Any object satisfying this predicate also satisfies\n"
685 #define FUNC_NAME s_scm_output_port_p
687 x
= SCM_COERCE_OUTPORT (x
);
688 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
692 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
694 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
696 #define FUNC_NAME s_scm_port_closed_p
698 SCM_VALIDATE_PORT (1, port
);
699 return scm_from_bool (!SCM_OPPORTP (port
));
703 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
705 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
707 #define FUNC_NAME s_scm_eof_object_p
709 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
719 * Call the close operation on a port object.
720 * see also scm_close.
722 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
724 "Close the specified port object. Return @code{#t} if it\n"
725 "successfully closes a port or @code{#f} if it was already\n"
726 "closed. An exception may be raised if an error occurs, for\n"
727 "example when flushing buffered output. See also @ref{Ports and\n"
728 "File Descriptors, close}, for a procedure which can close file\n"
730 #define FUNC_NAME s_scm_close_port
734 port
= SCM_COERCE_OUTPORT (port
);
736 SCM_VALIDATE_PORT (1, port
);
737 if (SCM_CLOSEDP (port
))
739 if (SCM_PORT_DESCRIPTOR (port
)->close
)
740 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
743 scm_i_remove_port (port
);
744 SCM_CLR_PORT_OPEN_FLAG (port
);
745 return scm_from_bool (rv
>= 0);
749 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
751 "Close the specified input port object. The routine has no effect if\n"
752 "the file has already been closed. An exception may be raised if an\n"
753 "error occurs. The value returned is unspecified.\n\n"
754 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
755 "which can close file descriptors.")
756 #define FUNC_NAME s_scm_close_input_port
758 SCM_VALIDATE_INPUT_PORT (1, port
);
759 scm_close_port (port
);
760 return SCM_UNSPECIFIED
;
764 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
766 "Close the specified output port object. The routine has no effect if\n"
767 "the file has already been closed. An exception may be raised if an\n"
768 "error occurs. The value returned is unspecified.\n\n"
769 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
770 "which can close file descriptors.")
771 #define FUNC_NAME s_scm_close_output_port
773 port
= SCM_COERCE_OUTPORT (port
);
774 SCM_VALIDATE_OUTPUT_PORT (1, port
);
775 scm_close_port (port
);
776 return SCM_UNSPECIFIED
;
783 /* Encoding characters to byte streams, and decoding byte streams to
786 /* A fluid specifying the default encoding for newly created ports. If it is
787 a string, that is the encoding. If it is #f, it is in the "native"
788 (Latin-1) encoding. */
789 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
791 static int scm_port_encoding_init
= 0;
793 /* Use ENCODING as the default encoding for future ports. */
795 scm_i_set_default_port_encoding (const char *encoding
)
797 if (!scm_port_encoding_init
798 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
799 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
803 || !strcmp (encoding
, "ASCII")
804 || !strcmp (encoding
, "ANSI_X3.4-1968")
805 || !strcmp (encoding
, "ISO-8859-1"))
806 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
808 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
),
809 scm_from_locale_string (encoding
));
812 /* Return the name of the default encoding for newly created ports; a
813 return value of NULL means "ISO-8859-1". */
815 scm_i_default_port_encoding (void)
817 if (!scm_port_encoding_init
)
819 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
825 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
826 if (!scm_is_string (encoding
))
829 return scm_i_string_chars (encoding
);
834 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
837 iconv_t new_input_cd
, new_output_cd
;
839 new_input_cd
= (iconv_t
) -1;
840 new_output_cd
= (iconv_t
) -1;
842 /* Set the character encoding for this port. */
843 pt
= SCM_PTAB_ENTRY (port
);
845 if (encoding
== NULL
)
846 encoding
= "ISO-8859-1";
848 if (pt
->encoding
!= encoding
)
849 pt
->encoding
= scm_gc_strdup (encoding
, "port");
851 /* If ENCODING is UTF-8, then no conversion descriptor is opened
852 because we do I/O ourselves. This saves 100+ KiB for each
854 if (strcmp (encoding
, "UTF-8"))
856 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
)
858 /* Open an input iconv conversion descriptor, from ENCODING
859 to UTF-8. We choose UTF-8, not UTF-32, because iconv
860 implementations can typically convert from anything to
861 UTF-8, but not to UTF-32 (see
862 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
863 new_input_cd
= iconv_open ("UTF-8", encoding
);
864 if (new_input_cd
== (iconv_t
) -1)
865 goto invalid_encoding
;
868 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
870 new_output_cd
= iconv_open (encoding
, "UTF-8");
871 if (new_output_cd
== (iconv_t
) -1)
873 if (new_input_cd
!= (iconv_t
) -1)
874 iconv_close (new_input_cd
);
875 goto invalid_encoding
;
880 if (pt
->input_cd
!= (iconv_t
) -1)
881 iconv_close (pt
->input_cd
);
882 if (pt
->output_cd
!= (iconv_t
) -1)
883 iconv_close (pt
->output_cd
);
885 pt
->input_cd
= new_input_cd
;
886 pt
->output_cd
= new_output_cd
;
893 err
= scm_from_locale_string (encoding
);
894 scm_misc_error ("scm_i_set_port_encoding_x",
895 "invalid or unknown character encoding ~s",
900 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
902 "Returns, as a string, the character encoding that @var{port}\n"
903 "uses to interpret its input and output.\n")
904 #define FUNC_NAME s_scm_port_encoding
909 SCM_VALIDATE_PORT (1, port
);
911 pt
= SCM_PTAB_ENTRY (port
);
914 return scm_from_locale_string (pt
->encoding
);
920 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
922 "Sets the character encoding that will be used to interpret all\n"
923 "port I/O. New ports are created with the encoding\n"
924 "appropriate for the current locale if @code{setlocale} has \n"
925 "been called or ISO-8859-1 otherwise\n"
926 "and this procedure can be used to modify that encoding.\n")
927 #define FUNC_NAME s_scm_set_port_encoding_x
931 SCM_VALIDATE_PORT (1, port
);
932 SCM_VALIDATE_STRING (2, enc
);
934 enc_str
= scm_to_locale_string (enc
);
935 scm_i_set_port_encoding_x (port
, enc_str
);
938 return SCM_UNSPECIFIED
;
943 /* This determines how conversions handle unconvertible characters. */
944 SCM_GLOBAL_VARIABLE (scm_conversion_strategy
, "%port-conversion-strategy");
945 static int scm_conversion_strategy_init
= 0;
947 scm_t_string_failed_conversion_handler
948 scm_i_get_conversion_strategy (SCM port
)
952 if (scm_is_false (port
))
954 if (!scm_conversion_strategy_init
955 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
956 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
959 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (scm_conversion_strategy
));
960 if (scm_is_false (encoding
))
961 return SCM_FAILED_CONVERSION_QUESTION_MARK
;
963 return (scm_t_string_failed_conversion_handler
) scm_to_int (encoding
);
969 pt
= SCM_PTAB_ENTRY (port
);
970 return pt
->ilseq_handler
;
976 scm_i_set_conversion_strategy_x (SCM port
,
977 scm_t_string_failed_conversion_handler handler
)
982 strategy
= scm_from_int ((int) handler
);
984 if (scm_is_false (port
))
986 /* Set the default encoding for future ports. */
987 if (!scm_conversion_strategy_init
988 || !scm_is_fluid (SCM_VARIABLE_REF (scm_conversion_strategy
)))
989 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
991 scm_fluid_set_x (SCM_VARIABLE_REF (scm_conversion_strategy
), strategy
);
995 /* Set the character encoding for this port. */
996 pt
= SCM_PTAB_ENTRY (port
);
997 pt
->ilseq_handler
= handler
;
1001 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1002 1, 0, 0, (SCM port
),
1003 "Returns the behavior of the port when handling a character that\n"
1004 "is not representable in the port's current encoding.\n"
1005 "It returns the symbol @code{error} if unrepresentable characters\n"
1006 "should cause exceptions, @code{substitute} if the port should\n"
1007 "try to replace unrepresentable characters with question marks or\n"
1008 "approximate characters, or @code{escape} if unrepresentable\n"
1009 "characters should be converted to string escapes.\n"
1011 "If @var{port} is @code{#f}, then the current default behavior\n"
1012 "will be returned. New ports will have this default behavior\n"
1013 "when they are created.\n")
1014 #define FUNC_NAME s_scm_port_conversion_strategy
1016 scm_t_string_failed_conversion_handler h
;
1018 SCM_VALIDATE_OPPORT (1, port
);
1020 if (!scm_is_false (port
))
1022 SCM_VALIDATE_OPPORT (1, port
);
1025 h
= scm_i_get_conversion_strategy (port
);
1026 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1027 return scm_from_latin1_symbol ("error");
1028 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1029 return scm_from_latin1_symbol ("substitute");
1030 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1031 return scm_from_latin1_symbol ("escape");
1035 /* Never gets here. */
1036 return SCM_UNDEFINED
;
1040 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1042 (SCM port
, SCM sym
),
1043 "Sets the behavior of the interpreter when outputting a character\n"
1044 "that is not representable in the port's current encoding.\n"
1045 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1046 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1047 "when an unconvertible character is encountered. If it is\n"
1048 "@code{'substitute}, then unconvertible characters will \n"
1049 "be replaced with approximate characters, or with question marks\n"
1050 "if no approximately correct character is available.\n"
1051 "If it is @code{'escape},\n"
1052 "it will appear as a hex escape when output.\n"
1054 "If @var{port} is an open port, the conversion error behavior\n"
1055 "is set for that port. If it is @code{#f}, it is set as the\n"
1056 "default behavior for any future ports that get created in\n"
1058 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1064 if (!scm_is_false (port
))
1066 SCM_VALIDATE_OPPORT (1, port
);
1069 err
= scm_from_latin1_symbol ("error");
1070 if (scm_is_true (scm_eqv_p (sym
, err
)))
1072 scm_i_set_conversion_strategy_x (port
, SCM_FAILED_CONVERSION_ERROR
);
1073 return SCM_UNSPECIFIED
;
1076 qm
= scm_from_latin1_symbol ("substitute");
1077 if (scm_is_true (scm_eqv_p (sym
, qm
)))
1079 scm_i_set_conversion_strategy_x (port
,
1080 SCM_FAILED_CONVERSION_QUESTION_MARK
);
1081 return SCM_UNSPECIFIED
;
1084 esc
= scm_from_latin1_symbol ("escape");
1085 if (scm_is_true (scm_eqv_p (sym
, esc
)))
1087 scm_i_set_conversion_strategy_x (port
,
1088 SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
);
1089 return SCM_UNSPECIFIED
;
1092 SCM_MISC_ERROR ("unknown conversion behavior ~s", scm_list_1 (sym
));
1094 return SCM_UNSPECIFIED
;
1101 /* The port lock. */
1104 lock_port (SCM port
)
1106 scm_c_lock_port (port
);
1110 unlock_port (SCM port
)
1112 scm_c_unlock_port (port
);
1116 scm_dynwind_lock_port (SCM port
)
1118 scm_dynwind_unwind_handler_with_scm (unlock_port
, port
,
1119 SCM_F_WIND_EXPLICITLY
);
1120 scm_dynwind_rewind_handler_with_scm (lock_port
, port
,
1121 SCM_F_WIND_EXPLICITLY
);
1127 /* Revealed counts --- an oddity inherited from SCSH. */
1129 /* Find a port in the table and return its revealed count.
1130 Also used by the garbage collector.
1133 scm_revealed_count (SCM port
)
1137 scm_c_lock_port (port
);
1138 ret
= SCM_REVEALED (port
);
1139 scm_c_unlock_port (port
);
1144 SCM_DEFINE (scm_port_revealed
, "port-revealed", 1, 0, 0,
1146 "Return the revealed count for @var{port}.")
1147 #define FUNC_NAME s_scm_port_revealed
1149 port
= SCM_COERCE_OUTPORT (port
);
1150 SCM_VALIDATE_OPENPORT (1, port
);
1151 return scm_from_int (scm_revealed_count (port
));
1155 /* Set the revealed count for a port. */
1156 SCM_DEFINE (scm_set_port_revealed_x
, "set-port-revealed!", 2, 0, 0,
1157 (SCM port
, SCM rcount
),
1158 "Sets the revealed count for a port to a given value.\n"
1159 "The return value is unspecified.")
1160 #define FUNC_NAME s_scm_set_port_revealed_x
1163 port
= SCM_COERCE_OUTPORT (port
);
1164 SCM_VALIDATE_OPENPORT (1, port
);
1165 r
= scm_to_int (rcount
);
1166 scm_c_lock_port (port
);
1167 SCM_REVEALED (port
) = r
;
1168 scm_c_unlock_port (port
);
1169 return SCM_UNSPECIFIED
;
1173 /* Set the revealed count for a port. */
1174 SCM_DEFINE (scm_adjust_port_revealed_x
, "adjust-port-revealed!", 2, 0, 0,
1175 (SCM port
, SCM addend
),
1176 "Add @var{addend} to the revealed count of @var{port}.\n"
1177 "The return value is unspecified.")
1178 #define FUNC_NAME s_scm_set_port_revealed_x
1181 port
= SCM_COERCE_OUTPORT (port
);
1182 SCM_VALIDATE_OPENPORT (1, port
);
1183 a
= scm_to_int (addend
);
1184 scm_c_lock_port (port
);
1185 SCM_REVEALED (port
) += a
;
1186 scm_c_unlock_port (port
);
1187 return SCM_UNSPECIFIED
;
1197 scm_get_byte_or_eof (SCM port
)
1201 scm_c_lock_port (port
);
1202 ret
= scm_get_byte_or_eof_unlocked (port
);
1203 scm_c_unlock_port (port
);
1209 scm_peek_byte_or_eof (SCM port
)
1213 scm_c_lock_port (port
);
1214 ret
= scm_peek_byte_or_eof_unlocked (port
);
1215 scm_c_unlock_port (port
);
1222 * Used by an application to read arbitrary number of bytes from an
1223 * SCM port. Same semantics as libc read, except that scm_c_read only
1224 * returns less than SIZE bytes if at end-of-file.
1226 * Warning: Doesn't update port line and column counts! */
1228 /* This structure, and the following swap_buffer function, are used
1229 for temporarily swapping a port's own read buffer, and the buffer
1230 that the caller of scm_c_read provides. */
1231 struct port_and_swap_buffer
1234 unsigned char *buffer
;
1239 swap_buffer (void *data
)
1241 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1242 unsigned char *old_buf
= psb
->pt
->read_buf
;
1243 size_t old_size
= psb
->pt
->read_buf_size
;
1245 /* Make the port use (buffer, size) from the struct. */
1246 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1247 psb
->pt
->read_buf_size
= psb
->size
;
1249 /* Save the port's old (buffer, size) in the struct. */
1250 psb
->buffer
= old_buf
;
1251 psb
->size
= old_size
;
1255 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1256 #define FUNC_NAME "scm_c_read"
1259 size_t n_read
= 0, n_available
;
1260 struct port_and_swap_buffer psb
;
1262 SCM_VALIDATE_OPINPORT (1, port
);
1264 pt
= SCM_PTAB_ENTRY (port
);
1265 if (pt
->rw_active
== SCM_PORT_WRITE
)
1266 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1269 pt
->rw_active
= SCM_PORT_READ
;
1271 /* Take bytes first from the port's read buffer. */
1272 if (pt
->read_pos
< pt
->read_end
)
1274 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1275 memcpy (buffer
, pt
->read_pos
, n_available
);
1276 buffer
= (char *) buffer
+ n_available
;
1277 pt
->read_pos
+= n_available
;
1278 n_read
+= n_available
;
1279 size
-= n_available
;
1282 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1286 /* Now we will call scm_fill_input repeatedly until we have read the
1287 requested number of bytes. (Note that a single scm_fill_input
1288 call does not guarantee to fill the whole of the port's read
1290 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1292 /* The port that we are reading from is unbuffered - i.e. does
1293 not have its own persistent buffer - but we have a buffer,
1294 provided by our caller, that is the right size for the data
1295 that is wanted. For the following scm_fill_input calls,
1296 therefore, we use the buffer in hand as the port's read
1299 We need to make sure that the port's normal (1 byte) buffer
1300 is reinstated in case one of the scm_fill_input () calls
1301 throws an exception; we use the scm_dynwind_* API to achieve
1304 A consequence of this optimization is that the fill_input
1305 functions can't unget characters. That'll push data to the
1306 pushback buffer instead of this psb buffer. */
1308 unsigned char *pback
= pt
->putback_buf
;
1311 psb
.buffer
= buffer
;
1313 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1314 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1315 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1317 /* Call scm_fill_input until we have all the bytes that we need,
1319 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1321 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1322 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1325 if (pback
!= pt
->putback_buf
1326 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1327 scm_misc_error (FUNC_NAME
,
1328 "scm_c_read must not call a fill function that pushes "
1329 "back characters onto an unbuffered port", SCM_EOL
);
1331 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1333 /* Reinstate the port's normal buffer. */
1338 /* The port has its own buffer. It is important that we use it,
1339 even if it happens to be smaller than our caller's buffer, so
1340 that a custom port implementation's entry points (in
1341 particular, fill_input) can rely on the buffer always being
1342 the same as they first set up. */
1343 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1345 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1346 memcpy (buffer
, pt
->read_pos
, n_available
);
1347 buffer
= (char *) buffer
+ n_available
;
1348 pt
->read_pos
+= n_available
;
1349 n_read
+= n_available
;
1350 size
-= n_available
;
1359 scm_c_read (SCM port
, void *buffer
, size_t size
)
1363 scm_c_lock_port (port
);
1364 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1365 scm_c_unlock_port (port
);
1370 /* Update the line and column number of PORT after consumption of C. */
1372 update_port_lf (scm_t_wchar c
, SCM port
)
1397 #define SCM_MBCHAR_BUF_SIZE (4)
1399 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1400 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1402 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1404 scm_t_wchar codepoint
;
1406 if (utf8_buf
[0] <= 0x7f)
1409 codepoint
= utf8_buf
[0];
1411 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1414 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1415 | (utf8_buf
[1] & 0x3f);
1417 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1420 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1421 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1422 | (utf8_buf
[2] & 0x3f);
1427 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1428 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1429 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1430 | (utf8_buf
[3] & 0x3f);
1436 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1437 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1438 representation, and set *LEN to the length in bytes. Return
1439 `EILSEQ' on error. */
1441 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1442 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1444 #define ASSERT_NOT_EOF(b) \
1445 if (SCM_UNLIKELY ((b) == EOF)) \
1447 #define CONSUME_PEEKED_BYTE() \
1454 pt
= SCM_PTAB_ENTRY (port
);
1456 byte
= scm_get_byte_or_eof_unlocked (port
);
1463 buf
[0] = (scm_t_uint8
) byte
;
1468 *codepoint
= buf
[0];
1469 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1472 byte
= scm_peek_byte_or_eof_unlocked (port
);
1473 ASSERT_NOT_EOF (byte
);
1475 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1478 CONSUME_PEEKED_BYTE ();
1479 buf
[1] = (scm_t_uint8
) byte
;
1482 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1485 else if ((buf
[0] & 0xf0) == 0xe0)
1488 byte
= scm_peek_byte_or_eof_unlocked (port
);
1489 ASSERT_NOT_EOF (byte
);
1491 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1492 || (buf
[0] == 0xe0 && byte
< 0xa0)
1493 || (buf
[0] == 0xed && byte
> 0x9f)))
1496 CONSUME_PEEKED_BYTE ();
1497 buf
[1] = (scm_t_uint8
) byte
;
1500 byte
= scm_peek_byte_or_eof_unlocked (port
);
1501 ASSERT_NOT_EOF (byte
);
1503 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1506 CONSUME_PEEKED_BYTE ();
1507 buf
[2] = (scm_t_uint8
) byte
;
1510 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1511 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1514 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1517 byte
= scm_peek_byte_or_eof_unlocked (port
);
1518 ASSERT_NOT_EOF (byte
);
1520 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1521 || (buf
[0] == 0xf0 && byte
< 0x90)
1522 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1525 CONSUME_PEEKED_BYTE ();
1526 buf
[1] = (scm_t_uint8
) byte
;
1529 byte
= scm_peek_byte_or_eof_unlocked (port
);
1530 ASSERT_NOT_EOF (byte
);
1532 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1535 CONSUME_PEEKED_BYTE ();
1536 buf
[2] = (scm_t_uint8
) byte
;
1539 byte
= scm_peek_byte_or_eof_unlocked (port
);
1540 ASSERT_NOT_EOF (byte
);
1542 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1545 CONSUME_PEEKED_BYTE ();
1546 buf
[3] = (scm_t_uint8
) byte
;
1549 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1550 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1551 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1560 /* Here we could choose the consume the faulty byte when it's not a
1561 valid starting byte, but it's not a requirement. What Section 3.9
1562 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1563 would otherwise be a valid starting byte. */
1567 #undef CONSUME_PEEKED_BYTE
1568 #undef ASSERT_NOT_EOF
1571 /* Likewise, read a byte sequence from PORT, passing it through its
1572 input conversion descriptor. */
1574 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1575 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1579 size_t bytes_consumed
, output_size
;
1581 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1583 pt
= SCM_PTAB_ENTRY (port
);
1585 for (output_size
= 0, output
= (char *) utf8_buf
,
1586 bytes_consumed
= 0, err
= 0;
1587 err
== 0 && output_size
== 0
1588 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1592 size_t input_left
, output_left
, done
;
1594 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1595 if (byte_read
== EOF
)
1597 if (bytes_consumed
== 0)
1599 *codepoint
= (scm_t_wchar
) EOF
;
1607 buf
[bytes_consumed
] = byte_read
;
1610 input_left
= bytes_consumed
+ 1;
1611 output_left
= sizeof (utf8_buf
);
1613 done
= iconv (pt
->input_cd
, &input
, &input_left
,
1614 &output
, &output_left
);
1615 if (done
== (size_t) -1)
1619 /* Missing input: keep trying. */
1623 output_size
= sizeof (utf8_buf
) - output_left
;
1626 if (SCM_UNLIKELY (output_size
== 0))
1627 /* An unterminated sequence. */
1629 else if (SCM_LIKELY (err
== 0))
1631 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1632 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1633 *len
= bytes_consumed
;
1639 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1640 with the byte representation of the codepoint in PORT's encoding, and
1641 set *LEN to the length in bytes of that representation. Return 0 on
1642 success and an errno value on error. */
1644 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1645 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1648 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1650 if (pt
->input_cd
== (iconv_t
) -1)
1651 /* Initialize the conversion descriptors, if needed. */
1652 scm_i_set_port_encoding_x (port
, pt
->encoding
);
1654 /* FIXME: In 2.1, add a flag to determine whether a port is UTF-8. */
1655 if (pt
->input_cd
== (iconv_t
) -1)
1656 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1658 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1660 if (SCM_LIKELY (err
== 0))
1661 update_port_lf (*codepoint
, port
);
1662 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1666 update_port_lf (*codepoint
, port
);
1672 /* Read a codepoint from PORT and return it. */
1674 scm_getc_unlocked (SCM port
)
1675 #define FUNC_NAME "scm_getc"
1679 scm_t_wchar codepoint
;
1680 char buf
[SCM_MBCHAR_BUF_SIZE
];
1682 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1683 if (SCM_UNLIKELY (err
!= 0))
1684 /* At this point PORT should point past the invalid encoding, as per
1685 R6RS-lib Section 8.2.4. */
1686 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1697 scm_c_lock_port (port
);
1698 ret
= scm_getc_unlocked (port
);
1699 scm_c_unlock_port (port
);
1704 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1706 "Return the next character available from @var{port}, updating\n"
1707 "@var{port} to point to the following character. If no more\n"
1708 "characters are available, the end-of-file object is returned.\n"
1710 "When @var{port}'s data cannot be decoded according to its\n"
1711 "character encoding, a @code{decoding-error} is raised and\n"
1712 "@var{port} points past the erroneous byte sequence.\n")
1713 #define FUNC_NAME s_scm_read_char
1716 if (SCM_UNBNDP (port
))
1717 port
= scm_current_input_port ();
1718 SCM_VALIDATE_OPINPORT (1, port
);
1719 c
= scm_getc_unlocked (port
);
1722 return SCM_MAKE_CHAR (c
);
1732 scm_unget_byte_unlocked (int c
, SCM port
)
1733 #define FUNC_NAME "scm_unget_byte"
1735 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1737 if (pt
->read_buf
== pt
->putback_buf
)
1738 /* already using the put-back buffer. */
1740 /* enlarge putback_buf if necessary. */
1741 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1742 && pt
->read_buf
== pt
->read_pos
)
1744 size_t new_size
= pt
->read_buf_size
* 2;
1745 unsigned char *tmp
= (unsigned char *)
1746 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1749 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1750 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1751 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1754 /* shift any existing bytes to buffer + 1. */
1755 if (pt
->read_pos
== pt
->read_end
)
1756 pt
->read_end
= pt
->read_buf
+ 1;
1757 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1759 int count
= pt
->read_end
- pt
->read_pos
;
1761 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1762 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1765 pt
->read_pos
= pt
->read_buf
;
1768 /* switch to the put-back buffer. */
1770 if (pt
->putback_buf
== NULL
)
1773 = (unsigned char *) scm_gc_malloc_pointerless
1774 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1775 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1778 pt
->saved_read_buf
= pt
->read_buf
;
1779 pt
->saved_read_pos
= pt
->read_pos
;
1780 pt
->saved_read_end
= pt
->read_end
;
1781 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1783 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1784 pt
->read_end
= pt
->read_buf
+ 1;
1785 pt
->read_buf_size
= pt
->putback_buf_size
;
1791 pt
->rw_active
= SCM_PORT_READ
;
1796 scm_unget_byte (int c
, SCM port
)
1798 scm_c_lock_port (port
);
1799 scm_unget_byte_unlocked (c
, port
);
1800 scm_c_unlock_port (port
);
1804 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1805 #define FUNC_NAME "scm_ungetc"
1807 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1809 char result_buf
[10];
1810 const char *encoding
;
1814 if (pt
->encoding
!= NULL
)
1815 encoding
= pt
->encoding
;
1817 encoding
= "ISO-8859-1";
1819 len
= sizeof (result_buf
);
1820 result
= u32_conv_to_encoding (encoding
,
1821 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1822 (uint32_t *) &c
, 1, NULL
,
1825 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1826 scm_encoding_error (FUNC_NAME
, errno
,
1827 "conversion to port encoding failed",
1828 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1830 for (i
= len
- 1; i
>= 0; i
--)
1831 scm_unget_byte_unlocked (result
[i
], port
);
1833 if (SCM_UNLIKELY (result
!= result_buf
))
1838 /* What should col be in this case?
1839 * We'll leave it at -1.
1841 SCM_LINUM (port
) -= 1;
1849 scm_ungetc (scm_t_wchar c
, SCM port
)
1851 scm_c_lock_port (port
);
1852 scm_ungetc_unlocked (c
, port
);
1853 scm_c_unlock_port (port
);
1857 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1859 /* This is simple minded and inefficient, but unreading strings is
1860 * probably not a common operation, and remember that line and
1861 * column numbers have to be handled...
1863 * Please feel free to write an optimized version!
1866 scm_ungetc_unlocked (s
[n
], port
);
1870 scm_ungets (const char *s
, int n
, SCM port
)
1872 scm_c_lock_port (port
);
1873 scm_ungets_unlocked (s
, n
, port
);
1874 scm_c_unlock_port (port
);
1877 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1879 "Return the next character available from @var{port},\n"
1880 "@emph{without} updating @var{port} to point to the following\n"
1881 "character. If no more characters are available, the\n"
1882 "end-of-file object is returned.\n"
1884 "The value returned by\n"
1885 "a call to @code{peek-char} is the same as the value that would\n"
1886 "have been returned by a call to @code{read-char} on the same\n"
1887 "port. The only difference is that the very next call to\n"
1888 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1889 "return the value returned by the preceding call to\n"
1890 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1891 "an interactive port will hang waiting for input whenever a call\n"
1892 "to @code{read-char} would have hung.\n"
1894 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1895 "if such a situation occurs. However, unlike with @code{read-char},\n"
1896 "@var{port} still points at the beginning of the erroneous byte\n"
1897 "sequence when the error is raised.\n")
1898 #define FUNC_NAME s_scm_peek_char
1903 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1904 long column
, line
, i
;
1907 if (SCM_UNBNDP (port
))
1908 port
= scm_current_input_port ();
1909 SCM_VALIDATE_OPINPORT (1, port
);
1911 column
= SCM_COL (port
);
1912 line
= SCM_LINUM (port
);
1914 err
= get_codepoint (port
, &c
, bytes
, &len
);
1916 for (i
= len
- 1; i
>= 0; i
--)
1917 scm_unget_byte_unlocked (bytes
[i
], port
);
1919 SCM_COL (port
) = column
;
1920 SCM_LINUM (port
) = line
;
1922 if (SCM_UNLIKELY (err
!= 0))
1924 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1926 /* Shouldn't happen since `catch' always aborts to prompt. */
1927 result
= SCM_BOOL_F
;
1930 result
= SCM_EOF_VAL
;
1932 result
= SCM_MAKE_CHAR (c
);
1938 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
1939 (SCM cobj
, SCM port
),
1940 "Place @var{char} in @var{port} so that it will be read by the\n"
1941 "next read operation. If called multiple times, the unread characters\n"
1942 "will be read again in last-in first-out order. If @var{port} is\n"
1943 "not supplied, the current input port is used.")
1944 #define FUNC_NAME s_scm_unread_char
1948 SCM_VALIDATE_CHAR (1, cobj
);
1949 if (SCM_UNBNDP (port
))
1950 port
= scm_current_input_port ();
1951 SCM_VALIDATE_OPINPORT (2, port
);
1953 c
= SCM_CHAR (cobj
);
1955 scm_ungetc_unlocked (c
, port
);
1960 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
1961 (SCM str
, SCM port
),
1962 "Place the string @var{str} in @var{port} so that its characters will be\n"
1963 "read in subsequent read operations. If called multiple times, the\n"
1964 "unread characters will be read again in last-in first-out order. If\n"
1965 "@var{port} is not supplied, the current-input-port is used.")
1966 #define FUNC_NAME s_scm_unread_string
1969 SCM_VALIDATE_STRING (1, str
);
1970 if (SCM_UNBNDP (port
))
1971 port
= scm_current_input_port ();
1972 SCM_VALIDATE_OPINPORT (2, port
);
1974 n
= scm_i_string_length (str
);
1977 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
1986 /* Manipulating the buffers. */
1988 /* This routine does not take any locks, as it is usually called as part
1989 of a port implementation. */
1991 scm_port_non_buffer (scm_t_port
*pt
)
1993 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
1994 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
1995 pt
->read_buf_size
= pt
->write_buf_size
= 1;
1996 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
1999 /* this should only be called when the read buffer is empty. it
2000 tries to refill the read buffer. it returns the first char from
2001 the port, which is either EOF or *(pt->read_pos). */
2003 scm_fill_input_unlocked (SCM port
)
2005 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2007 assert (pt
->read_pos
== pt
->read_end
);
2009 if (pt
->read_buf
== pt
->putback_buf
)
2011 /* finished reading put-back chars. */
2012 pt
->read_buf
= pt
->saved_read_buf
;
2013 pt
->read_pos
= pt
->saved_read_pos
;
2014 pt
->read_end
= pt
->saved_read_end
;
2015 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2016 if (pt
->read_pos
< pt
->read_end
)
2017 return *(pt
->read_pos
);
2019 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2023 scm_fill_input (SCM port
)
2027 scm_c_lock_port (port
);
2028 ret
= scm_fill_input_unlocked (port
);
2029 scm_c_unlock_port (port
);
2034 /* move up to read_len chars from port's putback and/or read buffers
2035 into memory starting at dest. returns the number of chars moved. */
2037 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2039 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2040 size_t chars_read
= 0;
2041 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2045 memcpy (dest
, pt
->read_pos
, from_buf
);
2046 pt
->read_pos
+= from_buf
;
2047 chars_read
+= from_buf
;
2048 read_len
-= from_buf
;
2052 /* if putback was active, try the real input buffer too. */
2053 if (pt
->read_buf
== pt
->putback_buf
)
2055 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2058 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2059 pt
->saved_read_pos
+= from_buf
;
2060 chars_read
+= from_buf
;
2066 /* Clear a port's read buffers, returning the contents. */
2067 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2069 "This procedure clears a port's input buffers, similar\n"
2070 "to the way that force-output clears the output buffer. The\n"
2071 "contents of the buffers are returned as a single string, e.g.,\n"
2074 "(define p (open-input-file ...))\n"
2075 "(drain-input p) => empty string, nothing buffered yet.\n"
2076 "(unread-char (read-char p) p)\n"
2077 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2079 "Draining the buffers may be useful for cleanly finishing\n"
2080 "buffered I/O so that the file descriptor can be used directly\n"
2081 "for further input.")
2082 #define FUNC_NAME s_scm_drain_input
2089 SCM_VALIDATE_OPINPORT (1, port
);
2090 pt
= SCM_PTAB_ENTRY (port
);
2092 count
= pt
->read_end
- pt
->read_pos
;
2093 if (pt
->read_buf
== pt
->putback_buf
)
2094 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2098 result
= scm_i_make_string (count
, &data
, 0);
2099 scm_take_from_input_buffers (port
, data
, count
);
2102 result
= scm_nullstr
;
2109 scm_end_input_unlocked (SCM port
)
2112 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2114 if (pt
->read_buf
== pt
->putback_buf
)
2116 offset
= pt
->read_end
- pt
->read_pos
;
2117 pt
->read_buf
= pt
->saved_read_buf
;
2118 pt
->read_pos
= pt
->saved_read_pos
;
2119 pt
->read_end
= pt
->saved_read_end
;
2120 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2125 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2129 scm_end_input (SCM port
)
2131 scm_c_lock_port (port
);
2132 scm_end_input_unlocked (port
);
2133 scm_c_unlock_port (port
);
2136 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2138 "Flush the specified output port, or the current output port if @var{port}\n"
2139 "is omitted. The current output buffer contents are passed to the\n"
2140 "underlying port implementation (e.g., in the case of fports, the\n"
2141 "data will be written to the file and the output buffer will be cleared.)\n"
2142 "It has no effect on an unbuffered port.\n\n"
2143 "The return value is unspecified.")
2144 #define FUNC_NAME s_scm_force_output
2146 if (SCM_UNBNDP (port
))
2147 port
= scm_current_output_port ();
2150 port
= SCM_COERCE_OUTPORT (port
);
2151 SCM_VALIDATE_OPOUTPORT (1, port
);
2153 scm_flush_unlocked (port
);
2154 return SCM_UNSPECIFIED
;
2159 scm_flush_unlocked (SCM port
)
2161 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2165 scm_flush (SCM port
)
2167 scm_c_lock_port (port
);
2168 scm_flush_unlocked (port
);
2169 scm_c_unlock_port (port
);
2178 scm_putc (char c
, SCM port
)
2180 scm_c_lock_port (port
);
2181 scm_putc_unlocked (c
, port
);
2182 scm_c_unlock_port (port
);
2186 scm_puts (const char *s
, SCM port
)
2188 scm_c_lock_port (port
);
2189 scm_puts_unlocked (s
, port
);
2190 scm_c_unlock_port (port
);
2195 * Used by an application to write arbitrary number of bytes to an SCM
2196 * port. Similar semantics as libc write. However, unlike libc
2197 * write, scm_c_write writes the requested number of bytes and has no
2200 * Warning: Doesn't update port line and column counts!
2203 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2204 #define FUNC_NAME "scm_c_write"
2207 scm_t_ptob_descriptor
*ptob
;
2209 SCM_VALIDATE_OPOUTPORT (1, port
);
2211 pt
= SCM_PTAB_ENTRY (port
);
2212 ptob
= SCM_PORT_DESCRIPTOR (port
);
2214 if (pt
->rw_active
== SCM_PORT_READ
)
2215 scm_end_input_unlocked (port
);
2217 ptob
->write (port
, ptr
, size
);
2220 pt
->rw_active
= SCM_PORT_WRITE
;
2225 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2227 scm_c_lock_port (port
);
2228 scm_c_write_unlocked (port
, ptr
, size
);
2229 scm_c_unlock_port (port
);
2234 * This function differs from scm_c_write; it updates port line and
2237 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2239 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2240 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2242 if (pt
->rw_active
== SCM_PORT_READ
)
2243 scm_end_input_unlocked (port
);
2245 ptob
->write (port
, ptr
, size
);
2247 for (; size
; ptr
++, size
--)
2248 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2251 pt
->rw_active
= SCM_PORT_WRITE
;
2255 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2257 scm_c_lock_port (port
);
2258 scm_lfwrite_unlocked (ptr
, size
, port
);
2259 scm_c_unlock_port (port
);
2262 /* Write STR to PORT from START inclusive to END exclusive. */
2264 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2266 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2268 if (pt
->rw_active
== SCM_PORT_READ
)
2269 scm_end_input_unlocked (port
);
2271 if (end
== (size_t) -1)
2272 end
= scm_i_string_length (str
);
2274 scm_display (scm_c_substring (str
, start
, end
), port
);
2277 pt
->rw_active
= SCM_PORT_WRITE
;
2283 /* Querying and setting positions, and character availability. */
2285 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2287 "Return @code{#t} if a character is ready on input @var{port}\n"
2288 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2289 "@code{#t} then the next @code{read-char} operation on\n"
2290 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2291 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2293 "@code{char-ready?} exists to make it possible for a\n"
2294 "program to accept characters from interactive ports without\n"
2295 "getting stuck waiting for input. Any input editors associated\n"
2296 "with such ports must make sure that characters whose existence\n"
2297 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2298 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2299 "a port at end of file would be indistinguishable from an\n"
2300 "interactive port that has no ready characters.")
2301 #define FUNC_NAME s_scm_char_ready_p
2305 if (SCM_UNBNDP (port
))
2306 port
= scm_current_input_port ();
2307 /* It's possible to close the current input port, so validate even in
2309 SCM_VALIDATE_OPINPORT (1, port
);
2311 pt
= SCM_PTAB_ENTRY (port
);
2313 /* if the current read buffer is filled, or the
2314 last pushed-back char has been read and the saved buffer is
2315 filled, result is true. */
2316 if (pt
->read_pos
< pt
->read_end
2317 || (pt
->read_buf
== pt
->putback_buf
2318 && pt
->saved_read_pos
< pt
->saved_read_end
))
2322 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2324 if (ptob
->input_waiting
)
2325 return scm_from_bool(ptob
->input_waiting (port
));
2332 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2333 (SCM fd_port
, SCM offset
, SCM whence
),
2334 "Sets the current position of @var{fd/port} to the integer\n"
2335 "@var{offset}, which is interpreted according to the value of\n"
2338 "One of the following variables should be supplied for\n"
2340 "@defvar SEEK_SET\n"
2341 "Seek from the beginning of the file.\n"
2343 "@defvar SEEK_CUR\n"
2344 "Seek from the current position.\n"
2346 "@defvar SEEK_END\n"
2347 "Seek from the end of the file.\n"
2349 "If @var{fd/port} is a file descriptor, the underlying system\n"
2350 "call is @code{lseek}. @var{port} may be a string port.\n"
2352 "The value returned is the new position in the file. This means\n"
2353 "that the current position of a port can be obtained using:\n"
2355 "(seek port 0 SEEK_CUR)\n"
2357 #define FUNC_NAME s_scm_seek
2361 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2363 how
= scm_to_int (whence
);
2364 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2365 SCM_OUT_OF_RANGE (3, whence
);
2367 if (SCM_OPPORTP (fd_port
))
2369 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2370 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2371 off_t_or_off64_t rv
;
2374 SCM_MISC_ERROR ("port is not seekable",
2375 scm_cons (fd_port
, SCM_EOL
));
2377 rv
= ptob
->seek (fd_port
, off
, how
);
2378 return scm_from_off_t_or_off64_t (rv
);
2380 else /* file descriptor?. */
2382 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2383 off_t_or_off64_t rv
;
2384 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2387 return scm_from_off_t_or_off64_t (rv
);
2396 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2397 doesn't have the filename version truncate(), hence this code. */
2398 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2400 truncate (const char *file
, off_t length
)
2404 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2408 ret
= ftruncate (fdes
, length
);
2411 int save_errno
= errno
;
2417 return close (fdes
);
2419 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2421 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2422 (SCM object
, SCM length
),
2423 "Truncate @var{file} to @var{length} bytes. @var{file} can be a\n"
2424 "filename string, a port object, or an integer file descriptor.\n"
2425 "The return value is unspecified.\n"
2427 "For a port or file descriptor @var{length} can be omitted, in\n"
2428 "which case the file is truncated at the current position (per\n"
2429 "@code{ftell} above).\n"
2431 "On most systems a file can be extended by giving a length\n"
2432 "greater than the current size, but this is not mandatory in the\n"
2434 #define FUNC_NAME s_scm_truncate_file
2438 /* "object" can be a port, fdes or filename.
2440 Negative "length" makes no sense, but it's left to truncate() or
2441 ftruncate() to give back an error for that (normally EINVAL).
2444 if (SCM_UNBNDP (length
))
2446 /* must supply length if object is a filename. */
2447 if (scm_is_string (object
))
2448 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2450 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2453 object
= SCM_COERCE_OUTPORT (object
);
2454 if (scm_is_integer (object
))
2456 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2457 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2460 else if (SCM_OPOUTPORTP (object
))
2462 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2463 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2464 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2466 if (!ptob
->truncate
)
2467 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2468 if (pt
->rw_active
== SCM_PORT_READ
)
2469 scm_end_input_unlocked (object
);
2470 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2471 ptob
->flush (object
);
2473 ptob
->truncate (object
, c_length
);
2478 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2479 char *str
= scm_to_locale_string (object
);
2481 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2488 return SCM_UNSPECIFIED
;
2492 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2494 "Return the current line number for @var{port}.\n"
2496 "The first line of a file is 0. But you might want to add 1\n"
2497 "when printing line numbers, since starting from 1 is\n"
2498 "traditional in error messages, and likely to be more natural to\n"
2500 #define FUNC_NAME s_scm_port_line
2502 port
= SCM_COERCE_OUTPORT (port
);
2503 SCM_VALIDATE_OPENPORT (1, port
);
2504 return scm_from_long (SCM_LINUM (port
));
2508 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2509 (SCM port
, SCM line
),
2510 "Set the current line number for @var{port} to @var{line}. The\n"
2511 "first line of a file is 0.")
2512 #define FUNC_NAME s_scm_set_port_line_x
2514 port
= SCM_COERCE_OUTPORT (port
);
2515 SCM_VALIDATE_OPENPORT (1, port
);
2516 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2517 return SCM_UNSPECIFIED
;
2521 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2523 "Return the current column number of @var{port}.\n"
2524 "If the number is\n"
2525 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2526 "- i.e. the first character of the first line is line 0, column 0.\n"
2527 "(However, when you display a file position, for example in an error\n"
2528 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2529 "because lines and column numbers traditionally start with 1, and that is\n"
2530 "what non-programmers will find most natural.)")
2531 #define FUNC_NAME s_scm_port_column
2533 port
= SCM_COERCE_OUTPORT (port
);
2534 SCM_VALIDATE_OPENPORT (1, port
);
2535 return scm_from_int (SCM_COL (port
));
2539 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2540 (SCM port
, SCM column
),
2541 "Set the current column of @var{port}. Before reading the first\n"
2542 "character on a line the column should be 0.")
2543 #define FUNC_NAME s_scm_set_port_column_x
2545 port
= SCM_COERCE_OUTPORT (port
);
2546 SCM_VALIDATE_OPENPORT (1, port
);
2547 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2548 return SCM_UNSPECIFIED
;
2552 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2554 "Return the filename associated with @var{port}, or @code{#f}\n"
2555 "if no filename is associated with the port.")
2556 #define FUNC_NAME s_scm_port_filename
2558 port
= SCM_COERCE_OUTPORT (port
);
2559 SCM_VALIDATE_OPENPORT (1, port
);
2560 return SCM_FILENAME (port
);
2564 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2565 (SCM port
, SCM filename
),
2566 "Change the filename associated with @var{port}, using the current input\n"
2567 "port if none is specified. Note that this does not change the port's\n"
2568 "source of data, but only the value that is returned by\n"
2569 "@code{port-filename} and reported in diagnostic output.")
2570 #define FUNC_NAME s_scm_set_port_filename_x
2572 port
= SCM_COERCE_OUTPORT (port
);
2573 SCM_VALIDATE_OPENPORT (1, port
);
2574 /* We allow the user to set the filename to whatever he likes. */
2575 SCM_SET_FILENAME (port
, filename
);
2576 return SCM_UNSPECIFIED
;
2583 /* Implementation helpers for port printing functions. */
2586 scm_print_port_mode (SCM exp
, SCM port
)
2588 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2590 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2591 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2594 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2601 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2603 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2606 scm_puts_unlocked ("#<", port
);
2607 scm_print_port_mode (exp
, port
);
2608 scm_puts_unlocked (type
, port
);
2609 scm_putc_unlocked (' ', port
);
2610 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2611 scm_putc_unlocked ('>', port
);
2618 /* Iterating over all ports. */
2620 struct for_each_data
2622 void (*proc
) (void *data
, SCM p
);
2627 for_each_trampoline (void *data
, SCM port
, SCM result
)
2629 struct for_each_data
*d
= data
;
2631 d
->proc (d
->data
, port
);
2637 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2639 struct for_each_data d
;
2644 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2645 scm_i_port_weak_set
);
2649 scm_for_each_trampoline (void *data
, SCM port
)
2651 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2654 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2656 "Apply @var{proc} to each port in the Guile port table\n"
2657 "in turn. The return value is unspecified. More specifically,\n"
2658 "@var{proc} is applied exactly once to every port that exists\n"
2659 "in the system at the time @var{port-for-each} is invoked.\n"
2660 "Changes to the port table while @var{port-for-each} is running\n"
2661 "have no effect as far as @var{port-for-each} is concerned.")
2662 #define FUNC_NAME s_scm_port_for_each
2664 SCM_VALIDATE_PROC (1, proc
);
2666 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2668 return SCM_UNSPECIFIED
;
2673 flush_output_port (void *closure
, SCM port
)
2675 if (SCM_OPOUTPORTP (port
))
2676 scm_flush_unlocked (port
);
2679 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2681 "Equivalent to calling @code{force-output} on\n"
2682 "all open output ports. The return value is unspecified.")
2683 #define FUNC_NAME s_scm_flush_all_ports
2685 scm_c_port_for_each (&flush_output_port
, NULL
);
2686 return SCM_UNSPECIFIED
;
2695 scm_t_bits scm_tc16_void_port
= 0;
2697 static int fill_input_void_port (SCM port SCM_UNUSED
)
2703 write_void_port (SCM port SCM_UNUSED
,
2704 const void *data SCM_UNUSED
,
2705 size_t size SCM_UNUSED
)
2710 scm_i_void_port (long mode_bits
)
2714 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2716 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2722 scm_void_port (char *mode_str
)
2724 return scm_i_void_port (scm_mode_bits (mode_str
));
2727 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2729 "Create and return a new void port. A void port acts like\n"
2730 "@file{/dev/null}. The @var{mode} argument\n"
2731 "specifies the input/output modes for this port: see the\n"
2732 "documentation for @code{open-file} in @ref{File Ports}.")
2733 #define FUNC_NAME s_scm_sys_make_void_port
2735 return scm_i_void_port (scm_i_mode_bits (mode
));
2742 /* Initialization. */
2747 /* lseek() symbols. */
2748 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2749 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2750 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2752 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2755 cur_inport_fluid
= scm_make_fluid ();
2756 cur_outport_fluid
= scm_make_fluid ();
2757 cur_errport_fluid
= scm_make_fluid ();
2758 cur_loadport_fluid
= scm_make_fluid ();
2760 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2762 #include "libguile/ports.x"
2764 /* Use Latin-1 as the default port encoding. */
2765 SCM_VARIABLE_SET (default_port_encoding_var
,
2766 scm_make_fluid_with_default (SCM_BOOL_F
));
2767 scm_port_encoding_init
= 1;
2769 SCM_VARIABLE_SET (scm_conversion_strategy
,
2770 scm_make_fluid_with_default
2771 (scm_from_int ((int) SCM_FAILED_CONVERSION_QUESTION_MARK
)));
2772 scm_conversion_strategy_init
= 1;