1 /* Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2003, 2004,
2 * 2006, 2007, 2008, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 3 of
7 * the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
24 #define _LARGEFILE64_SOURCE /* ask for stat64 etc */
32 #include <fcntl.h> /* for chsize on mingw */
37 #include <striconveh.h>
41 #include "libguile/_scm.h"
42 #include "libguile/async.h"
43 #include "libguile/deprecation.h"
44 #include "libguile/eval.h"
45 #include "libguile/fports.h" /* direct access for seek and truncate */
46 #include "libguile/goops.h"
47 #include "libguile/smob.h"
48 #include "libguile/chars.h"
49 #include "libguile/dynwind.h"
51 #include "libguile/keywords.h"
52 #include "libguile/hashtab.h"
53 #include "libguile/root.h"
54 #include "libguile/strings.h"
55 #include "libguile/mallocs.h"
56 #include "libguile/validate.h"
57 #include "libguile/ports.h"
58 #include "libguile/vectors.h"
59 #include "libguile/weak-set.h"
60 #include "libguile/fluids.h"
61 #include "libguile/eq.h"
75 #ifdef HAVE_SYS_IOCTL_H
76 #include <sys/ioctl.h>
79 /* Mingw (version 3.4.5, circa 2006) has ftruncate as an alias for chsize
80 already, but have this code here in case that wasn't so in past versions,
81 or perhaps to help other minimal DOS environments.
83 gnulib ftruncate.c has code using fcntl F_CHSIZE and F_FREESP, which
84 might be possibilities if we've got other systems without ftruncate. */
86 #if defined HAVE_CHSIZE && ! defined HAVE_FTRUNCATE
87 #define ftruncate(fd, size) chsize (fd, size)
89 #define HAVE_FTRUNCATE 1
93 /* The port kind table --- a dynamically resized array of port types. */
96 /* scm_ptobs scm_numptob
97 * implement a dynamically resized array of ptob records.
98 * Indexes into this table are used when generating type
99 * tags for smobjects (if you know a tag you can get an index and conversely).
101 static scm_t_ptob_descriptor
**scm_ptobs
= NULL
;
102 static long scm_numptob
= 0; /* Number of port types. */
103 static long scm_ptobs_size
= 0; /* Number of slots in the port type
105 static scm_i_pthread_mutex_t scm_ptobs_lock
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
108 scm_c_num_port_types (void)
112 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
114 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
119 scm_t_ptob_descriptor
*
120 scm_c_port_type_ref (long ptobnum
)
122 scm_t_ptob_descriptor
*ret
= NULL
;
124 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
126 if (0 <= ptobnum
&& ptobnum
< scm_numptob
)
127 ret
= scm_ptobs
[ptobnum
];
129 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
132 scm_out_of_range ("scm_c_port_type_ref", scm_from_long (ptobnum
));
138 scm_c_port_type_add_x (scm_t_ptob_descriptor
*desc
)
142 scm_i_pthread_mutex_lock (&scm_ptobs_lock
);
144 if (scm_numptob
+ 1 < SCM_I_MAX_PORT_TYPE_COUNT
)
146 if (scm_numptob
== scm_ptobs_size
)
148 unsigned long old_size
= scm_ptobs_size
;
149 scm_t_ptob_descriptor
**old_ptobs
= scm_ptobs
;
151 /* Currently there are only 9 predefined port types, so one
152 resize will cover it. */
153 scm_ptobs_size
= old_size
+ 10;
155 if (scm_ptobs_size
>= SCM_I_MAX_PORT_TYPE_COUNT
)
156 scm_ptobs_size
= SCM_I_MAX_PORT_TYPE_COUNT
;
158 scm_ptobs
= scm_gc_malloc (sizeof (*scm_ptobs
) * scm_ptobs_size
,
161 memcpy (scm_ptobs
, old_ptobs
, sizeof (*scm_ptobs
) * scm_numptob
);
165 scm_ptobs
[ret
] = desc
;
168 scm_i_pthread_mutex_unlock (&scm_ptobs_lock
);
171 scm_out_of_range ("scm_c_port_type_add_x", scm_from_long (scm_numptob
));
177 * We choose to use an interface similar to the smob interface with
178 * fill_input and write as standard fields, passed to the port
179 * type constructor, and optional fields set by setters.
183 flush_port_default (SCM port SCM_UNUSED
)
188 end_input_default (SCM port SCM_UNUSED
, int offset SCM_UNUSED
)
193 scm_make_port_type (char *name
,
194 int (*fill_input
) (SCM port
),
195 void (*write
) (SCM port
, const void *data
, size_t size
))
197 scm_t_ptob_descriptor
*desc
;
200 desc
= scm_gc_malloc_pointerless (sizeof (*desc
), "port-type");
201 memset (desc
, 0, sizeof (*desc
));
204 desc
->print
= scm_port_print
;
206 desc
->flush
= flush_port_default
;
207 desc
->end_input
= end_input_default
;
208 desc
->fill_input
= fill_input
;
210 ptobnum
= scm_c_port_type_add_x (desc
);
212 /* Make a class object if GOOPS is present. */
213 if (SCM_UNPACK (scm_port_class
[0]) != 0)
214 scm_make_port_classes (ptobnum
, name
);
216 return scm_tc7_port
+ ptobnum
* 256;
220 scm_set_port_mark (scm_t_bits tc
, SCM (*mark
) (SCM
))
222 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->mark
= mark
;
226 scm_set_port_free (scm_t_bits tc
, size_t (*free
) (SCM
))
228 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->free
= free
;
232 scm_set_port_print (scm_t_bits tc
, int (*print
) (SCM exp
, SCM port
,
233 scm_print_state
*pstate
))
235 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->print
= print
;
239 scm_set_port_equalp (scm_t_bits tc
, SCM (*equalp
) (SCM
, SCM
))
241 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->equalp
= equalp
;
245 scm_set_port_close (scm_t_bits tc
, int (*close
) (SCM
))
247 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->close
= close
;
251 scm_set_port_flush (scm_t_bits tc
, void (*flush
) (SCM port
))
253 scm_t_ptob_descriptor
*ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
));
255 ptob
->flags
|= SCM_PORT_TYPE_HAS_FLUSH
;
259 scm_set_port_end_input (scm_t_bits tc
, void (*end_input
) (SCM port
, int offset
))
261 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->end_input
= end_input
;
265 scm_set_port_seek (scm_t_bits tc
, scm_t_off (*seek
) (SCM
, scm_t_off
, int))
267 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->seek
= seek
;
271 scm_set_port_truncate (scm_t_bits tc
, void (*truncate
) (SCM
, scm_t_off
))
273 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->truncate
= truncate
;
277 scm_set_port_input_waiting (scm_t_bits tc
, int (*input_waiting
) (SCM
))
279 scm_c_port_type_ref (SCM_TC2PTOBNUM (tc
))->input_waiting
= input_waiting
;
284 /* Standard ports --- current input, output, error, and more(!). */
286 static SCM cur_inport_fluid
= SCM_BOOL_F
;
287 static SCM cur_outport_fluid
= SCM_BOOL_F
;
288 static SCM cur_errport_fluid
= SCM_BOOL_F
;
289 static SCM cur_loadport_fluid
= SCM_BOOL_F
;
291 SCM_DEFINE (scm_current_input_port
, "current-input-port", 0, 0, 0,
293 "Return the current input port. This is the default port used\n"
294 "by many input procedures. Initially, @code{current-input-port}\n"
295 "returns the @dfn{standard input} in Unix and C terminology.")
296 #define FUNC_NAME s_scm_current_input_port
298 if (scm_is_true (cur_inport_fluid
))
299 return scm_fluid_ref (cur_inport_fluid
);
305 SCM_DEFINE (scm_current_output_port
, "current-output-port", 0, 0, 0,
307 "Return the current output port. This is the default port used\n"
308 "by many output procedures. Initially,\n"
309 "@code{current-output-port} returns the @dfn{standard output} in\n"
310 "Unix and C terminology.")
311 #define FUNC_NAME s_scm_current_output_port
313 if (scm_is_true (cur_outport_fluid
))
314 return scm_fluid_ref (cur_outport_fluid
);
320 SCM_DEFINE (scm_current_error_port
, "current-error-port", 0, 0, 0,
322 "Return the port to which errors and warnings should be sent (the\n"
323 "@dfn{standard error} in Unix and C terminology).")
324 #define FUNC_NAME s_scm_current_error_port
326 if (scm_is_true (cur_errport_fluid
))
327 return scm_fluid_ref (cur_errport_fluid
);
334 scm_current_warning_port (void)
336 static SCM cwp_var
= SCM_BOOL_F
;
338 if (scm_is_false (cwp_var
))
339 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
341 return scm_call_0 (scm_variable_ref (cwp_var
));
344 SCM_DEFINE (scm_current_load_port
, "current-load-port", 0, 0, 0,
346 "Return the current-load-port.\n"
347 "The load port is used internally by @code{primitive-load}.")
348 #define FUNC_NAME s_scm_current_load_port
350 return scm_fluid_ref (cur_loadport_fluid
);
354 SCM_DEFINE (scm_set_current_input_port
, "set-current-input-port", 1, 0, 0,
356 "@deffnx {Scheme Procedure} set-current-output-port port\n"
357 "@deffnx {Scheme Procedure} set-current-error-port port\n"
358 "Change the ports returned by @code{current-input-port},\n"
359 "@code{current-output-port} and @code{current-error-port}, respectively,\n"
360 "so that they use the supplied @var{port} for input or output.")
361 #define FUNC_NAME s_scm_set_current_input_port
363 SCM oinp
= scm_fluid_ref (cur_inport_fluid
);
364 SCM_VALIDATE_OPINPORT (1, port
);
365 scm_fluid_set_x (cur_inport_fluid
, port
);
371 SCM_DEFINE (scm_set_current_output_port
, "set-current-output-port", 1, 0, 0,
373 "Set the current default output port to @var{port}.")
374 #define FUNC_NAME s_scm_set_current_output_port
376 SCM ooutp
= scm_fluid_ref (cur_outport_fluid
);
377 port
= SCM_COERCE_OUTPORT (port
);
378 SCM_VALIDATE_OPOUTPORT (1, port
);
379 scm_fluid_set_x (cur_outport_fluid
, port
);
385 SCM_DEFINE (scm_set_current_error_port
, "set-current-error-port", 1, 0, 0,
387 "Set the current default error port to @var{port}.")
388 #define FUNC_NAME s_scm_set_current_error_port
390 SCM oerrp
= scm_fluid_ref (cur_errport_fluid
);
391 port
= SCM_COERCE_OUTPORT (port
);
392 SCM_VALIDATE_OPOUTPORT (1, port
);
393 scm_fluid_set_x (cur_errport_fluid
, port
);
400 scm_set_current_warning_port (SCM port
)
402 static SCM cwp_var
= SCM_BOOL_F
;
404 if (scm_is_false (cwp_var
))
405 cwp_var
= scm_c_private_lookup ("guile", "current-warning-port");
407 return scm_call_1 (scm_variable_ref (cwp_var
), port
);
412 scm_dynwind_current_input_port (SCM port
)
413 #define FUNC_NAME NULL
415 SCM_VALIDATE_OPINPORT (1, port
);
416 scm_dynwind_fluid (cur_inport_fluid
, port
);
421 scm_dynwind_current_output_port (SCM port
)
422 #define FUNC_NAME NULL
424 port
= SCM_COERCE_OUTPORT (port
);
425 SCM_VALIDATE_OPOUTPORT (1, port
);
426 scm_dynwind_fluid (cur_outport_fluid
, port
);
431 scm_dynwind_current_error_port (SCM port
)
432 #define FUNC_NAME NULL
434 port
= SCM_COERCE_OUTPORT (port
);
435 SCM_VALIDATE_OPOUTPORT (1, port
);
436 scm_dynwind_fluid (cur_errport_fluid
, port
);
441 scm_i_dynwind_current_load_port (SCM port
)
443 scm_dynwind_fluid (cur_loadport_fluid
, port
);
449 /* Retrieving a port's mode. */
451 /* Return the flags that characterize a port based on the mode
452 * string used to open a file for that port.
454 * See PORT FLAGS in scm.h
458 scm_i_mode_bits_n (SCM modes
)
461 | (scm_i_string_contains_char (modes
, 'r')
462 || scm_i_string_contains_char (modes
, '+') ? SCM_RDNG
: 0)
463 | (scm_i_string_contains_char (modes
, 'w')
464 || scm_i_string_contains_char (modes
, 'a')
465 || scm_i_string_contains_char (modes
, '+') ? SCM_WRTNG
: 0)
466 | (scm_i_string_contains_char (modes
, '0') ? SCM_BUF0
: 0)
467 | (scm_i_string_contains_char (modes
, 'l') ? SCM_BUFLINE
: 0));
471 scm_mode_bits (char *modes
)
473 /* Valid characters are rw+a0l. So, use latin1. */
474 return scm_i_mode_bits (scm_from_latin1_string (modes
));
478 scm_i_mode_bits (SCM modes
)
482 if (!scm_is_string (modes
))
483 scm_wrong_type_arg_msg (NULL
, 0, modes
, "string");
485 bits
= scm_i_mode_bits_n (modes
);
486 scm_remember_upto_here_1 (modes
);
490 /* Return the mode flags from an open port.
491 * Some modes such as "append" are only used when opening
492 * a file and are not returned here. */
494 SCM_DEFINE (scm_port_mode
, "port-mode", 1, 0, 0,
496 "Return the port modes associated with the open port @var{port}.\n"
497 "These will not necessarily be identical to the modes used when\n"
498 "the port was opened, since modes such as \"append\" which are\n"
499 "used only during port creation are not retained.")
500 #define FUNC_NAME s_scm_port_mode
505 port
= SCM_COERCE_OUTPORT (port
);
506 SCM_VALIDATE_OPPORT (1, port
);
507 if (SCM_CELL_WORD_0 (port
) & SCM_RDNG
) {
508 if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
509 strcpy (modes
, "r+");
513 else if (SCM_CELL_WORD_0 (port
) & SCM_WRTNG
)
515 if (SCM_CELL_WORD_0 (port
) & SCM_BUF0
)
518 return scm_from_latin1_string (modes
);
525 /* The port table --- a weak set of all ports.
527 We need a global registry of ports to flush them all at exit, and to
528 get all the ports matching a file descriptor. */
529 SCM scm_i_port_weak_set
;
534 /* Port finalization. */
538 scm_t_ptob_descriptor
*ptob
;
543 do_free (void *body_data
)
545 struct do_free_data
*data
= body_data
;
547 /* `close' is for explicit `close-port' by user. `free' is for this
548 purpose: ports collected by the GC. */
549 data
->ptob
->free (data
->port
);
554 /* Finalize the object (a port) pointed to by PTR. */
556 finalize_port (void *ptr
, void *data
)
558 SCM port
= SCM_PACK_POINTER (ptr
);
560 if (!SCM_PORTP (port
))
563 if (SCM_OPENP (port
))
565 struct do_free_data data
;
567 SCM_CLR_PORT_OPEN_FLAG (port
);
569 data
.ptob
= SCM_PORT_DESCRIPTOR (port
);
572 scm_internal_catch (SCM_BOOL_T
, do_free
, &data
,
573 scm_handle_by_message_noexit
, NULL
);
575 scm_gc_ports_collected
++;
583 scm_c_make_port_with_encoding (scm_t_bits tag
, unsigned long mode_bits
,
584 const char *encoding
,
585 scm_t_string_failed_conversion_handler handler
,
590 scm_t_ptob_descriptor
*ptob
;
592 entry
= (scm_t_port
*) scm_gc_calloc (sizeof (scm_t_port
), "port");
593 ptob
= scm_c_port_type_ref (SCM_TC2PTOBNUM (tag
));
595 ret
= scm_words (tag
| mode_bits
, 3);
596 SCM_SET_CELL_WORD_1 (ret
, (scm_t_bits
) entry
);
597 SCM_SET_CELL_WORD_2 (ret
, (scm_t_bits
) ptob
);
599 entry
->lock
= scm_gc_malloc_pointerless (sizeof (*entry
->lock
), "port lock");
600 scm_i_pthread_mutex_init (entry
->lock
, scm_i_pthread_mutexattr_recursive
);
602 entry
->file_name
= SCM_BOOL_F
;
603 entry
->rw_active
= SCM_PORT_NEITHER
;
605 entry
->stream
= stream
;
606 entry
->encoding
= encoding
? scm_gc_strdup (encoding
, "port") : NULL
;
607 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
608 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
609 else if (!encoding
|| strcmp (encoding
, "ISO-8859-1") == 0)
610 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
612 entry
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
613 entry
->ilseq_handler
= handler
;
614 entry
->iconv_descriptors
= NULL
;
616 entry
->alist
= SCM_EOL
;
618 if (SCM_PORT_DESCRIPTOR (ret
)->free
)
619 scm_i_set_finalizer (SCM2PTR (ret
), finalize_port
, NULL
);
621 if (SCM_PORT_DESCRIPTOR (ret
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
622 scm_weak_set_add_x (scm_i_port_weak_set
, ret
);
628 scm_c_make_port (scm_t_bits tag
, unsigned long mode_bits
, scm_t_bits stream
)
630 return scm_c_make_port_with_encoding (tag
, mode_bits
,
631 scm_i_default_port_encoding (),
632 scm_i_default_port_conversion_handler (),
637 scm_new_port_table_entry (scm_t_bits tag
)
639 return scm_c_make_port (tag
, 0, 0);
646 SCM_DEFINE (scm_port_p
, "port?", 1, 0, 0,
648 "Return a boolean indicating whether @var{x} is a port.\n"
649 "Equivalent to @code{(or (input-port? @var{x}) (output-port?\n"
651 #define FUNC_NAME s_scm_port_p
653 return scm_from_bool (SCM_PORTP (x
));
657 SCM_DEFINE (scm_input_port_p
, "input-port?", 1, 0, 0,
659 "Return @code{#t} if @var{x} is an input port, otherwise return\n"
660 "@code{#f}. Any object satisfying this predicate also satisfies\n"
662 #define FUNC_NAME s_scm_input_port_p
664 return scm_from_bool (SCM_INPUT_PORT_P (x
));
668 SCM_DEFINE (scm_output_port_p
, "output-port?", 1, 0, 0,
670 "Return @code{#t} if @var{x} is an output port, otherwise return\n"
671 "@code{#f}. Any object satisfying this predicate also satisfies\n"
673 #define FUNC_NAME s_scm_output_port_p
675 x
= SCM_COERCE_OUTPORT (x
);
676 return scm_from_bool (SCM_OUTPUT_PORT_P (x
));
680 SCM_DEFINE (scm_port_closed_p
, "port-closed?", 1, 0, 0,
682 "Return @code{#t} if @var{port} is closed or @code{#f} if it is\n"
684 #define FUNC_NAME s_scm_port_closed_p
686 SCM_VALIDATE_PORT (1, port
);
687 return scm_from_bool (!SCM_OPPORTP (port
));
691 SCM_DEFINE (scm_eof_object_p
, "eof-object?", 1, 0, 0,
693 "Return @code{#t} if @var{x} is an end-of-file object; otherwise\n"
695 #define FUNC_NAME s_scm_eof_object_p
697 return scm_from_bool (SCM_EOF_OBJECT_P (x
));
706 static void close_iconv_descriptors (scm_t_iconv_descriptors
*id
);
709 * Call the close operation on a port object.
710 * see also scm_close.
712 SCM_DEFINE (scm_close_port
, "close-port", 1, 0, 0,
714 "Close the specified port object. Return @code{#t} if it\n"
715 "successfully closes a port or @code{#f} if it was already\n"
716 "closed. An exception may be raised if an error occurs, for\n"
717 "example when flushing buffered output. See also @ref{Ports and\n"
718 "File Descriptors, close}, for a procedure which can close file\n"
720 #define FUNC_NAME s_scm_close_port
725 port
= SCM_COERCE_OUTPORT (port
);
727 SCM_VALIDATE_PORT (1, port
);
728 if (SCM_CLOSEDP (port
))
731 p
= SCM_PTAB_ENTRY (port
);
732 SCM_CLR_PORT_OPEN_FLAG (port
);
734 if (SCM_PORT_DESCRIPTOR (port
)->flags
& SCM_PORT_TYPE_HAS_FLUSH
)
735 scm_weak_set_remove_x (scm_i_port_weak_set
, port
);
737 if (SCM_PORT_DESCRIPTOR (port
)->close
)
738 /* Note! This may throw an exception. Anything after this point
739 should be resilient to non-local exits. */
740 rv
= SCM_PORT_DESCRIPTOR (port
)->close (port
);
744 if (p
->iconv_descriptors
)
746 /* If we don't get here, the iconv_descriptors finalizer will
748 close_iconv_descriptors (p
->iconv_descriptors
);
749 p
->iconv_descriptors
= NULL
;
752 return scm_from_bool (rv
>= 0);
756 SCM_DEFINE (scm_close_input_port
, "close-input-port", 1, 0, 0,
758 "Close the specified input port object. The routine has no effect if\n"
759 "the file has already been closed. An exception may be raised if an\n"
760 "error occurs. The value returned is unspecified.\n\n"
761 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
762 "which can close file descriptors.")
763 #define FUNC_NAME s_scm_close_input_port
765 SCM_VALIDATE_INPUT_PORT (1, port
);
766 scm_close_port (port
);
767 return SCM_UNSPECIFIED
;
771 SCM_DEFINE (scm_close_output_port
, "close-output-port", 1, 0, 0,
773 "Close the specified output port object. The routine has no effect if\n"
774 "the file has already been closed. An exception may be raised if an\n"
775 "error occurs. The value returned is unspecified.\n\n"
776 "See also @ref{Ports and File Descriptors, close}, for a procedure\n"
777 "which can close file descriptors.")
778 #define FUNC_NAME s_scm_close_output_port
780 port
= SCM_COERCE_OUTPORT (port
);
781 SCM_VALIDATE_OUTPUT_PORT (1, port
);
782 scm_close_port (port
);
783 return SCM_UNSPECIFIED
;
790 /* Encoding characters to byte streams, and decoding byte streams to
793 /* A fluid specifying the default encoding for newly created ports. If it is
794 a string, that is the encoding. If it is #f, it is in the "native"
795 (Latin-1) encoding. */
796 SCM_VARIABLE (default_port_encoding_var
, "%default-port-encoding");
798 static int scm_port_encoding_init
= 0;
800 /* Use ENCODING as the default encoding for future ports. */
802 scm_i_set_default_port_encoding (const char *encoding
)
804 if (!scm_port_encoding_init
805 || !scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
806 scm_misc_error (NULL
, "tried to set port encoding fluid before it is initialized",
810 || !strcmp (encoding
, "ASCII")
811 || !strcmp (encoding
, "ANSI_X3.4-1968")
812 || !strcmp (encoding
, "ISO-8859-1"))
813 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), SCM_BOOL_F
);
819 str
= scm_from_latin1_string (encoding
);
821 /* Restrict to ASCII. */
822 for (i
= 0; encoding
[i
]; i
++)
823 if (encoding
[i
] > 127)
824 scm_misc_error ("scm_i_set_default_port_encoding",
825 "invalid character encoding ~s", scm_list_1 (str
));
827 scm_fluid_set_x (SCM_VARIABLE_REF (default_port_encoding_var
), str
);
831 /* Return the name of the default encoding for newly created ports; a
832 return value of NULL means "ISO-8859-1". */
834 scm_i_default_port_encoding (void)
836 if (!scm_port_encoding_init
)
838 else if (!scm_is_fluid (SCM_VARIABLE_REF (default_port_encoding_var
)))
844 encoding
= scm_fluid_ref (SCM_VARIABLE_REF (default_port_encoding_var
));
845 if (!scm_is_string (encoding
))
848 return scm_i_string_chars (encoding
);
852 /* A fluid specifying the default conversion handler for newly created
853 ports. Its value should be one of the symbols below. */
854 SCM_VARIABLE (default_conversion_strategy_var
,
855 "%default-port-conversion-strategy");
857 /* Whether the above fluid is initialized. */
858 static int scm_conversion_strategy_init
= 0;
860 /* The possible conversion strategies. */
861 SCM_SYMBOL (sym_error
, "error");
862 SCM_SYMBOL (sym_substitute
, "substitute");
863 SCM_SYMBOL (sym_escape
, "escape");
865 /* Return the default failed encoding conversion policy for new created
867 scm_t_string_failed_conversion_handler
868 scm_i_default_port_conversion_handler (void)
870 scm_t_string_failed_conversion_handler handler
;
872 if (!scm_conversion_strategy_init
873 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
874 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
879 fluid
= SCM_VARIABLE_REF (default_conversion_strategy_var
);
880 value
= scm_fluid_ref (fluid
);
882 if (scm_is_eq (sym_substitute
, value
))
883 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
884 else if (scm_is_eq (sym_escape
, value
))
885 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
887 /* Default to 'error also when the fluid's value is not one of
888 the valid symbols. */
889 handler
= SCM_FAILED_CONVERSION_ERROR
;
895 /* Use HANDLER as the default conversion strategy for future ports. */
897 scm_i_set_default_port_conversion_handler (scm_t_string_failed_conversion_handler
902 if (!scm_conversion_strategy_init
903 || !scm_is_fluid (SCM_VARIABLE_REF (default_conversion_strategy_var
)))
904 scm_misc_error (NULL
, "tried to set conversion strategy fluid before it is initialized",
909 case SCM_FAILED_CONVERSION_ERROR
:
910 strategy
= sym_error
;
913 case SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
:
914 strategy
= sym_escape
;
917 case SCM_FAILED_CONVERSION_QUESTION_MARK
:
918 strategy
= sym_substitute
;
925 scm_fluid_set_x (SCM_VARIABLE_REF (default_conversion_strategy_var
),
930 finalize_iconv_descriptors (void *ptr
, void *data
)
932 close_iconv_descriptors (ptr
);
935 static scm_t_iconv_descriptors
*
936 open_iconv_descriptors (const char *encoding
, int reading
, int writing
)
938 scm_t_iconv_descriptors
*id
;
939 iconv_t input_cd
, output_cd
;
942 input_cd
= (iconv_t
) -1;
943 output_cd
= (iconv_t
) -1;
945 for (i
= 0; encoding
[i
]; i
++)
946 if (encoding
[i
] > 127)
947 goto invalid_encoding
;
951 /* Open an input iconv conversion descriptor, from ENCODING
952 to UTF-8. We choose UTF-8, not UTF-32, because iconv
953 implementations can typically convert from anything to
954 UTF-8, but not to UTF-32 (see
955 <http://lists.gnu.org/archive/html/bug-libunistring/2010-09/msg00007.html>). */
957 /* Assume opening an iconv descriptor causes about 16 KB of
959 scm_gc_register_allocation (16 * 1024);
961 input_cd
= iconv_open ("UTF-8", encoding
);
962 if (input_cd
== (iconv_t
) -1)
963 goto invalid_encoding
;
968 /* Assume opening an iconv descriptor causes about 16 KB of
970 scm_gc_register_allocation (16 * 1024);
972 output_cd
= iconv_open (encoding
, "UTF-8");
973 if (output_cd
== (iconv_t
) -1)
975 if (input_cd
!= (iconv_t
) -1)
976 iconv_close (input_cd
);
977 goto invalid_encoding
;
981 id
= scm_gc_malloc_pointerless (sizeof (*id
), "iconv descriptors");
982 id
->input_cd
= input_cd
;
983 id
->output_cd
= output_cd
;
985 /* Register a finalizer to close the descriptors. */
986 scm_i_set_finalizer (id
, finalize_iconv_descriptors
, NULL
);
993 err
= scm_from_latin1_string (encoding
);
994 scm_misc_error ("open_iconv_descriptors",
995 "invalid or unknown character encoding ~s",
1001 close_iconv_descriptors (scm_t_iconv_descriptors
*id
)
1003 if (id
->input_cd
!= (iconv_t
) -1)
1004 iconv_close (id
->input_cd
);
1005 if (id
->output_cd
!= (iconv_t
) -1)
1006 iconv_close (id
->output_cd
);
1007 id
->input_cd
= (void *) -1;
1008 id
->output_cd
= (void *) -1;
1011 scm_t_iconv_descriptors
*
1012 scm_i_port_iconv_descriptors (SCM port
)
1016 pt
= SCM_PTAB_ENTRY (port
);
1018 assert (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_ICONV
);
1020 if (!pt
->iconv_descriptors
)
1023 pt
->encoding
= "ISO-8859-1";
1024 pt
->iconv_descriptors
=
1025 open_iconv_descriptors (pt
->encoding
,
1026 SCM_INPUT_PORT_P (port
),
1027 SCM_OUTPUT_PORT_P (port
));
1030 return pt
->iconv_descriptors
;
1033 /* The name of the encoding is itself encoded in ASCII. */
1035 scm_i_set_port_encoding_x (SCM port
, const char *encoding
)
1038 scm_t_iconv_descriptors
*prev
;
1040 /* Set the character encoding for this port. */
1041 pt
= SCM_PTAB_ENTRY (port
);
1042 prev
= pt
->iconv_descriptors
;
1044 if (encoding
&& strcmp (encoding
, "UTF-8") == 0)
1046 pt
->encoding
= "UTF-8";
1047 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_UTF8
;
1048 pt
->iconv_descriptors
= NULL
;
1050 else if (!encoding
|| strcmp (encoding
, "ISO-8859-1") == 0)
1052 pt
->encoding
= "ISO-8859-1";
1053 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_LATIN1
;
1054 pt
->iconv_descriptors
= NULL
;
1058 /* Open descriptors before mutating the port. */
1059 pt
->iconv_descriptors
=
1060 open_iconv_descriptors (encoding
,
1061 SCM_INPUT_PORT_P (port
),
1062 SCM_OUTPUT_PORT_P (port
));
1063 pt
->encoding
= scm_gc_strdup (encoding
, "port");
1064 pt
->encoding_mode
= SCM_PORT_ENCODING_MODE_ICONV
;
1068 close_iconv_descriptors (prev
);
1071 SCM_DEFINE (scm_port_encoding
, "port-encoding", 1, 0, 0,
1073 "Returns, as a string, the character encoding that @var{port}\n"
1074 "uses to interpret its input and output.\n")
1075 #define FUNC_NAME s_scm_port_encoding
1080 SCM_VALIDATE_PORT (1, port
);
1082 pt
= SCM_PTAB_ENTRY (port
);
1085 return scm_from_latin1_string (pt
->encoding
);
1091 SCM_DEFINE (scm_set_port_encoding_x
, "set-port-encoding!", 2, 0, 0,
1092 (SCM port
, SCM enc
),
1093 "Sets the character encoding that will be used to interpret all\n"
1094 "port I/O. New ports are created with the encoding\n"
1095 "appropriate for the current locale if @code{setlocale} has \n"
1096 "been called or ISO-8859-1 otherwise\n"
1097 "and this procedure can be used to modify that encoding.\n")
1098 #define FUNC_NAME s_scm_set_port_encoding_x
1102 SCM_VALIDATE_PORT (1, port
);
1103 SCM_VALIDATE_STRING (2, enc
);
1105 enc_str
= scm_to_latin1_string (enc
);
1106 scm_i_set_port_encoding_x (port
, enc_str
);
1109 return SCM_UNSPECIFIED
;
1113 SCM_DEFINE (scm_port_conversion_strategy
, "port-conversion-strategy",
1114 1, 0, 0, (SCM port
),
1115 "Returns the behavior of the port when handling a character that\n"
1116 "is not representable in the port's current encoding.\n"
1117 "It returns the symbol @code{error} if unrepresentable characters\n"
1118 "should cause exceptions, @code{substitute} if the port should\n"
1119 "try to replace unrepresentable characters with question marks or\n"
1120 "approximate characters, or @code{escape} if unrepresentable\n"
1121 "characters should be converted to string escapes.\n"
1123 "If @var{port} is @code{#f}, then the current default behavior\n"
1124 "will be returned. New ports will have this default behavior\n"
1125 "when they are created.\n")
1126 #define FUNC_NAME s_scm_port_conversion_strategy
1128 scm_t_string_failed_conversion_handler h
;
1130 SCM_VALIDATE_OPPORT (1, port
);
1132 if (scm_is_false (port
))
1133 h
= scm_i_default_port_conversion_handler ();
1138 SCM_VALIDATE_OPPORT (1, port
);
1139 pt
= SCM_PTAB_ENTRY (port
);
1141 h
= pt
->ilseq_handler
;
1144 if (h
== SCM_FAILED_CONVERSION_ERROR
)
1145 return scm_from_latin1_symbol ("error");
1146 else if (h
== SCM_FAILED_CONVERSION_QUESTION_MARK
)
1147 return scm_from_latin1_symbol ("substitute");
1148 else if (h
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
1149 return scm_from_latin1_symbol ("escape");
1153 /* Never gets here. */
1154 return SCM_UNDEFINED
;
1158 SCM_DEFINE (scm_set_port_conversion_strategy_x
, "set-port-conversion-strategy!",
1160 (SCM port
, SCM sym
),
1161 "Sets the behavior of the interpreter when outputting a character\n"
1162 "that is not representable in the port's current encoding.\n"
1163 "@var{sym} can be either @code{'error}, @code{'substitute}, or\n"
1164 "@code{'escape}. If it is @code{'error}, an error will be thrown\n"
1165 "when an unconvertible character is encountered. If it is\n"
1166 "@code{'substitute}, then unconvertible characters will \n"
1167 "be replaced with approximate characters, or with question marks\n"
1168 "if no approximately correct character is available.\n"
1169 "If it is @code{'escape},\n"
1170 "it will appear as a hex escape when output.\n"
1172 "If @var{port} is an open port, the conversion error behavior\n"
1173 "is set for that port. If it is @code{#f}, it is set as the\n"
1174 "default behavior for any future ports that get created in\n"
1176 #define FUNC_NAME s_scm_set_port_conversion_strategy_x
1178 scm_t_string_failed_conversion_handler handler
;
1180 if (scm_is_eq (sym
, sym_error
))
1181 handler
= SCM_FAILED_CONVERSION_ERROR
;
1182 else if (scm_is_eq (sym
, sym_substitute
))
1183 handler
= SCM_FAILED_CONVERSION_QUESTION_MARK
;
1184 else if (scm_is_eq (sym
, sym_escape
))
1185 handler
= SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
;
1187 SCM_MISC_ERROR ("unknown conversion strategy ~s", scm_list_1 (sym
));
1189 if (scm_is_false (port
))
1190 scm_i_set_default_port_conversion_handler (handler
);
1193 SCM_VALIDATE_OPPORT (1, port
);
1194 SCM_PTAB_ENTRY (port
)->ilseq_handler
= handler
;
1197 return SCM_UNSPECIFIED
;
1204 /* The port lock. */
1207 lock_port (void *mutex
)
1209 scm_i_pthread_mutex_lock (mutex
);
1213 unlock_port (void *mutex
)
1215 scm_i_pthread_mutex_unlock (mutex
);
1219 scm_dynwind_lock_port (SCM port
)
1220 #define FUNC_NAME "dynwind-lock-port"
1222 scm_i_pthread_mutex_t
*lock
;
1223 SCM_VALIDATE_OPPORT (SCM_ARG1
, port
);
1224 scm_c_lock_port (port
, &lock
);
1227 scm_dynwind_unwind_handler (unlock_port
, lock
, SCM_F_WIND_EXPLICITLY
);
1228 scm_dynwind_rewind_handler (lock_port
, lock
, 0);
1239 scm_get_byte_or_eof (SCM port
)
1241 scm_i_pthread_mutex_t
*lock
;
1244 scm_c_lock_port (port
, &lock
);
1245 ret
= scm_get_byte_or_eof_unlocked (port
);
1247 scm_i_pthread_mutex_unlock (lock
);
1253 scm_peek_byte_or_eof (SCM port
)
1255 scm_i_pthread_mutex_t
*lock
;
1258 scm_c_lock_port (port
, &lock
);
1259 ret
= scm_peek_byte_or_eof_unlocked (port
);
1261 scm_i_pthread_mutex_unlock (lock
);
1268 * Used by an application to read arbitrary number of bytes from an
1269 * SCM port. Same semantics as libc read, except that scm_c_read only
1270 * returns less than SIZE bytes if at end-of-file.
1272 * Warning: Doesn't update port line and column counts! */
1274 /* This structure, and the following swap_buffer function, are used
1275 for temporarily swapping a port's own read buffer, and the buffer
1276 that the caller of scm_c_read provides. */
1277 struct port_and_swap_buffer
1280 unsigned char *buffer
;
1285 swap_buffer (void *data
)
1287 struct port_and_swap_buffer
*psb
= (struct port_and_swap_buffer
*) data
;
1288 unsigned char *old_buf
= psb
->pt
->read_buf
;
1289 size_t old_size
= psb
->pt
->read_buf_size
;
1291 /* Make the port use (buffer, size) from the struct. */
1292 psb
->pt
->read_pos
= psb
->pt
->read_buf
= psb
->pt
->read_end
= psb
->buffer
;
1293 psb
->pt
->read_buf_size
= psb
->size
;
1295 /* Save the port's old (buffer, size) in the struct. */
1296 psb
->buffer
= old_buf
;
1297 psb
->size
= old_size
;
1301 scm_c_read_unlocked (SCM port
, void *buffer
, size_t size
)
1302 #define FUNC_NAME "scm_c_read"
1305 size_t n_read
= 0, n_available
;
1306 struct port_and_swap_buffer psb
;
1308 SCM_VALIDATE_OPINPORT (1, port
);
1310 pt
= SCM_PTAB_ENTRY (port
);
1311 if (pt
->rw_active
== SCM_PORT_WRITE
)
1312 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
1315 pt
->rw_active
= SCM_PORT_READ
;
1317 /* Take bytes first from the port's read buffer. */
1318 if (pt
->read_pos
< pt
->read_end
)
1320 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1321 memcpy (buffer
, pt
->read_pos
, n_available
);
1322 buffer
= (char *) buffer
+ n_available
;
1323 pt
->read_pos
+= n_available
;
1324 n_read
+= n_available
;
1325 size
-= n_available
;
1328 /* Avoid the scm_dynwind_* costs if we now have enough data. */
1332 /* Now we will call scm_fill_input repeatedly until we have read the
1333 requested number of bytes. (Note that a single scm_fill_input
1334 call does not guarantee to fill the whole of the port's read
1336 if (pt
->read_buf_size
<= 1 && pt
->encoding
== NULL
)
1338 /* The port that we are reading from is unbuffered - i.e. does
1339 not have its own persistent buffer - but we have a buffer,
1340 provided by our caller, that is the right size for the data
1341 that is wanted. For the following scm_fill_input calls,
1342 therefore, we use the buffer in hand as the port's read
1345 We need to make sure that the port's normal (1 byte) buffer
1346 is reinstated in case one of the scm_fill_input () calls
1347 throws an exception; we use the scm_dynwind_* API to achieve
1350 A consequence of this optimization is that the fill_input
1351 functions can't unget characters. That'll push data to the
1352 pushback buffer instead of this psb buffer. */
1354 unsigned char *pback
= pt
->putback_buf
;
1357 psb
.buffer
= buffer
;
1359 scm_dynwind_begin (SCM_F_DYNWIND_REWINDABLE
);
1360 scm_dynwind_rewind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1361 scm_dynwind_unwind_handler (swap_buffer
, &psb
, SCM_F_WIND_EXPLICITLY
);
1363 /* Call scm_fill_input until we have all the bytes that we need,
1365 while (pt
->read_buf_size
&& (scm_fill_input_unlocked (port
) != EOF
))
1367 pt
->read_buf_size
-= (pt
->read_end
- pt
->read_pos
);
1368 pt
->read_pos
= pt
->read_buf
= pt
->read_end
;
1371 if (pback
!= pt
->putback_buf
1372 || pt
->read_buf
- (unsigned char *) buffer
< 0)
1373 scm_misc_error (FUNC_NAME
,
1374 "scm_c_read must not call a fill function that pushes "
1375 "back characters onto an unbuffered port", SCM_EOL
);
1377 n_read
+= pt
->read_buf
- (unsigned char *) buffer
;
1379 /* Reinstate the port's normal buffer. */
1384 /* The port has its own buffer. It is important that we use it,
1385 even if it happens to be smaller than our caller's buffer, so
1386 that a custom port implementation's entry points (in
1387 particular, fill_input) can rely on the buffer always being
1388 the same as they first set up. */
1389 while (size
&& (scm_fill_input_unlocked (port
) != EOF
))
1391 n_available
= min (size
, pt
->read_end
- pt
->read_pos
);
1392 memcpy (buffer
, pt
->read_pos
, n_available
);
1393 buffer
= (char *) buffer
+ n_available
;
1394 pt
->read_pos
+= n_available
;
1395 n_read
+= n_available
;
1396 size
-= n_available
;
1405 scm_c_read (SCM port
, void *buffer
, size_t size
)
1407 scm_i_pthread_mutex_t
*lock
;
1410 scm_c_lock_port (port
, &lock
);
1411 ret
= scm_c_read_unlocked (port
, buffer
, size
);
1413 scm_i_pthread_mutex_unlock (lock
);
1419 /* Update the line and column number of PORT after consumption of C. */
1421 update_port_lf (scm_t_wchar c
, SCM port
)
1446 #define SCM_MBCHAR_BUF_SIZE (4)
1448 /* Convert the SIZE-byte UTF-8 sequence in UTF8_BUF to a codepoint.
1449 UTF8_BUF is assumed to contain a valid UTF-8 sequence. */
1451 utf8_to_codepoint (const scm_t_uint8
*utf8_buf
, size_t size
)
1453 scm_t_wchar codepoint
;
1455 if (utf8_buf
[0] <= 0x7f)
1458 codepoint
= utf8_buf
[0];
1460 else if ((utf8_buf
[0] & 0xe0) == 0xc0)
1463 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x1f) << 6UL
1464 | (utf8_buf
[1] & 0x3f);
1466 else if ((utf8_buf
[0] & 0xf0) == 0xe0)
1469 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x0f) << 12UL
1470 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 6UL
1471 | (utf8_buf
[2] & 0x3f);
1476 codepoint
= ((scm_t_wchar
) utf8_buf
[0] & 0x07) << 18UL
1477 | ((scm_t_wchar
) utf8_buf
[1] & 0x3f) << 12UL
1478 | ((scm_t_wchar
) utf8_buf
[2] & 0x3f) << 6UL
1479 | (utf8_buf
[3] & 0x3f);
1485 /* Read a UTF-8 sequence from PORT. On success, return 0 and set
1486 *CODEPOINT to the codepoint that was read, fill BUF with its UTF-8
1487 representation, and set *LEN to the length in bytes. Return
1488 `EILSEQ' on error. */
1490 get_utf8_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1491 scm_t_uint8 buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1493 #define ASSERT_NOT_EOF(b) \
1494 if (SCM_UNLIKELY ((b) == EOF)) \
1496 #define CONSUME_PEEKED_BYTE() \
1503 pt
= SCM_PTAB_ENTRY (port
);
1505 byte
= scm_get_byte_or_eof_unlocked (port
);
1512 buf
[0] = (scm_t_uint8
) byte
;
1517 *codepoint
= buf
[0];
1518 else if (buf
[0] >= 0xc2 && buf
[0] <= 0xdf)
1521 byte
= scm_peek_byte_or_eof_unlocked (port
);
1522 ASSERT_NOT_EOF (byte
);
1524 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1527 CONSUME_PEEKED_BYTE ();
1528 buf
[1] = (scm_t_uint8
) byte
;
1531 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x1f) << 6UL
1534 else if ((buf
[0] & 0xf0) == 0xe0)
1537 byte
= scm_peek_byte_or_eof_unlocked (port
);
1538 ASSERT_NOT_EOF (byte
);
1540 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80
1541 || (buf
[0] == 0xe0 && byte
< 0xa0)
1542 || (buf
[0] == 0xed && byte
> 0x9f)))
1545 CONSUME_PEEKED_BYTE ();
1546 buf
[1] = (scm_t_uint8
) byte
;
1549 byte
= scm_peek_byte_or_eof_unlocked (port
);
1550 ASSERT_NOT_EOF (byte
);
1552 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1555 CONSUME_PEEKED_BYTE ();
1556 buf
[2] = (scm_t_uint8
) byte
;
1559 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x0f) << 12UL
1560 | ((scm_t_wchar
) buf
[1] & 0x3f) << 6UL
1563 else if (buf
[0] >= 0xf0 && buf
[0] <= 0xf4)
1566 byte
= scm_peek_byte_or_eof_unlocked (port
);
1567 ASSERT_NOT_EOF (byte
);
1569 if (SCM_UNLIKELY (((byte
& 0xc0) != 0x80)
1570 || (buf
[0] == 0xf0 && byte
< 0x90)
1571 || (buf
[0] == 0xf4 && byte
> 0x8f)))
1574 CONSUME_PEEKED_BYTE ();
1575 buf
[1] = (scm_t_uint8
) byte
;
1578 byte
= scm_peek_byte_or_eof_unlocked (port
);
1579 ASSERT_NOT_EOF (byte
);
1581 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1584 CONSUME_PEEKED_BYTE ();
1585 buf
[2] = (scm_t_uint8
) byte
;
1588 byte
= scm_peek_byte_or_eof_unlocked (port
);
1589 ASSERT_NOT_EOF (byte
);
1591 if (SCM_UNLIKELY ((byte
& 0xc0) != 0x80))
1594 CONSUME_PEEKED_BYTE ();
1595 buf
[3] = (scm_t_uint8
) byte
;
1598 *codepoint
= ((scm_t_wchar
) buf
[0] & 0x07) << 18UL
1599 | ((scm_t_wchar
) buf
[1] & 0x3f) << 12UL
1600 | ((scm_t_wchar
) buf
[2] & 0x3f) << 6UL
1609 /* Here we could choose the consume the faulty byte when it's not a
1610 valid starting byte, but it's not a requirement. What Section 3.9
1611 of Unicode 6.0.0 mandates, though, is to not consume a byte that
1612 would otherwise be a valid starting byte. */
1616 #undef CONSUME_PEEKED_BYTE
1617 #undef ASSERT_NOT_EOF
1620 /* Read an ISO-8859-1 codepoint (a byte) from PORT. On success, return
1621 0 and set *CODEPOINT to the codepoint that was read, fill BUF with
1622 its UTF-8 representation, and set *LEN to the length in bytes.
1623 Return `EILSEQ' on error. */
1625 get_latin1_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1626 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1628 *codepoint
= scm_get_byte_or_eof_unlocked (port
);
1630 if (*codepoint
== EOF
)
1635 buf
[0] = *codepoint
;
1640 /* Likewise, read a byte sequence from PORT, passing it through its
1641 input conversion descriptor. */
1643 get_iconv_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1644 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1646 scm_t_iconv_descriptors
*id
;
1648 size_t bytes_consumed
, output_size
;
1650 scm_t_uint8 utf8_buf
[SCM_MBCHAR_BUF_SIZE
];
1652 id
= scm_i_port_iconv_descriptors (port
);
1654 for (output_size
= 0, output
= (char *) utf8_buf
,
1655 bytes_consumed
= 0, err
= 0;
1656 err
== 0 && output_size
== 0
1657 && (bytes_consumed
== 0 || byte_read
!= EOF
);
1661 size_t input_left
, output_left
, done
;
1663 byte_read
= scm_get_byte_or_eof_unlocked (port
);
1664 if (byte_read
== EOF
)
1666 if (bytes_consumed
== 0)
1668 *codepoint
= (scm_t_wchar
) EOF
;
1676 buf
[bytes_consumed
] = byte_read
;
1679 input_left
= bytes_consumed
+ 1;
1680 output_left
= sizeof (utf8_buf
);
1682 done
= iconv (id
->input_cd
, &input
, &input_left
, &output
, &output_left
);
1683 if (done
== (size_t) -1)
1687 /* Missing input: keep trying. */
1691 output_size
= sizeof (utf8_buf
) - output_left
;
1694 if (SCM_UNLIKELY (output_size
== 0))
1695 /* An unterminated sequence. */
1697 else if (SCM_LIKELY (err
== 0))
1699 /* Convert the UTF8_BUF sequence to a Unicode code point. */
1700 *codepoint
= utf8_to_codepoint (utf8_buf
, output_size
);
1701 *len
= bytes_consumed
;
1707 /* Read a codepoint from PORT and return it in *CODEPOINT. Fill BUF
1708 with the byte representation of the codepoint in PORT's encoding, and
1709 set *LEN to the length in bytes of that representation. Return 0 on
1710 success and an errno value on error. */
1711 static SCM_C_INLINE
int
1712 get_codepoint (SCM port
, scm_t_wchar
*codepoint
,
1713 char buf
[SCM_MBCHAR_BUF_SIZE
], size_t *len
)
1716 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1718 if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_UTF8
)
1719 err
= get_utf8_codepoint (port
, codepoint
, (scm_t_uint8
*) buf
, len
);
1720 else if (pt
->encoding_mode
== SCM_PORT_ENCODING_MODE_LATIN1
)
1721 err
= get_latin1_codepoint (port
, codepoint
, buf
, len
);
1723 err
= get_iconv_codepoint (port
, codepoint
, buf
, len
);
1725 if (SCM_LIKELY (err
== 0))
1726 update_port_lf (*codepoint
, port
);
1727 else if (pt
->ilseq_handler
== SCM_ICONVEH_QUESTION_MARK
)
1731 update_port_lf (*codepoint
, port
);
1737 /* Read a codepoint from PORT and return it. */
1739 scm_getc_unlocked (SCM port
)
1740 #define FUNC_NAME "scm_getc"
1744 scm_t_wchar codepoint
;
1745 char buf
[SCM_MBCHAR_BUF_SIZE
];
1747 err
= get_codepoint (port
, &codepoint
, buf
, &len
);
1748 if (SCM_UNLIKELY (err
!= 0))
1749 /* At this point PORT should point past the invalid encoding, as per
1750 R6RS-lib Section 8.2.4. */
1751 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
1760 scm_i_pthread_mutex_t
*lock
;
1763 scm_c_lock_port (port
, &lock
);
1764 ret
= scm_getc_unlocked (port
);
1766 scm_i_pthread_mutex_unlock (lock
);
1772 SCM_DEFINE (scm_read_char
, "read-char", 0, 1, 0,
1774 "Return the next character available from @var{port}, updating\n"
1775 "@var{port} to point to the following character. If no more\n"
1776 "characters are available, the end-of-file object is returned.\n"
1778 "When @var{port}'s data cannot be decoded according to its\n"
1779 "character encoding, a @code{decoding-error} is raised and\n"
1780 "@var{port} points past the erroneous byte sequence.\n")
1781 #define FUNC_NAME s_scm_read_char
1784 if (SCM_UNBNDP (port
))
1785 port
= scm_current_input_port ();
1786 SCM_VALIDATE_OPINPORT (1, port
);
1787 c
= scm_getc_unlocked (port
);
1790 return SCM_MAKE_CHAR (c
);
1800 scm_unget_byte_unlocked (int c
, SCM port
)
1801 #define FUNC_NAME "scm_unget_byte"
1803 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1805 if (pt
->read_buf
== pt
->putback_buf
)
1806 /* already using the put-back buffer. */
1808 /* enlarge putback_buf if necessary. */
1809 if (pt
->read_end
== pt
->read_buf
+ pt
->read_buf_size
1810 && pt
->read_buf
== pt
->read_pos
)
1812 size_t new_size
= pt
->read_buf_size
* 2;
1813 unsigned char *tmp
= (unsigned char *)
1814 scm_gc_realloc (pt
->putback_buf
, pt
->read_buf_size
, new_size
,
1817 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
= tmp
;
1818 pt
->read_end
= pt
->read_buf
+ pt
->read_buf_size
;
1819 pt
->read_buf_size
= pt
->putback_buf_size
= new_size
;
1822 /* shift any existing bytes to buffer + 1. */
1823 if (pt
->read_pos
== pt
->read_end
)
1824 pt
->read_end
= pt
->read_buf
+ 1;
1825 else if (pt
->read_pos
!= pt
->read_buf
+ 1)
1827 int count
= pt
->read_end
- pt
->read_pos
;
1829 memmove (pt
->read_buf
+ 1, pt
->read_pos
, count
);
1830 pt
->read_end
= pt
->read_buf
+ 1 + count
;
1833 pt
->read_pos
= pt
->read_buf
;
1836 /* switch to the put-back buffer. */
1838 if (pt
->putback_buf
== NULL
)
1841 = (unsigned char *) scm_gc_malloc_pointerless
1842 (SCM_INITIAL_PUTBACK_BUF_SIZE
, "putback buffer");
1843 pt
->putback_buf_size
= SCM_INITIAL_PUTBACK_BUF_SIZE
;
1846 pt
->saved_read_buf
= pt
->read_buf
;
1847 pt
->saved_read_pos
= pt
->read_pos
;
1848 pt
->saved_read_end
= pt
->read_end
;
1849 pt
->saved_read_buf_size
= pt
->read_buf_size
;
1851 pt
->read_pos
= pt
->read_buf
= pt
->putback_buf
;
1852 pt
->read_end
= pt
->read_buf
+ 1;
1853 pt
->read_buf_size
= pt
->putback_buf_size
;
1859 pt
->rw_active
= SCM_PORT_READ
;
1864 scm_unget_byte (int c
, SCM port
)
1866 scm_i_pthread_mutex_t
*lock
;
1867 scm_c_lock_port (port
, &lock
);
1868 scm_unget_byte_unlocked (c
, port
);
1870 scm_i_pthread_mutex_unlock (lock
);
1875 scm_ungetc_unlocked (scm_t_wchar c
, SCM port
)
1876 #define FUNC_NAME "scm_ungetc"
1878 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
1880 char result_buf
[10];
1881 const char *encoding
;
1885 if (pt
->encoding
!= NULL
)
1886 encoding
= pt
->encoding
;
1888 encoding
= "ISO-8859-1";
1890 len
= sizeof (result_buf
);
1891 result
= u32_conv_to_encoding (encoding
,
1892 (enum iconv_ilseq_handler
) pt
->ilseq_handler
,
1893 (uint32_t *) &c
, 1, NULL
,
1896 if (SCM_UNLIKELY (result
== NULL
|| len
== 0))
1897 scm_encoding_error (FUNC_NAME
, errno
,
1898 "conversion to port encoding failed",
1899 SCM_BOOL_F
, SCM_MAKE_CHAR (c
));
1901 for (i
= len
- 1; i
>= 0; i
--)
1902 scm_unget_byte_unlocked (result
[i
], port
);
1904 if (SCM_UNLIKELY (result
!= result_buf
))
1909 /* What should col be in this case?
1910 * We'll leave it at -1.
1912 SCM_LINUM (port
) -= 1;
1920 scm_ungetc (scm_t_wchar c
, SCM port
)
1922 scm_i_pthread_mutex_t
*lock
;
1923 scm_c_lock_port (port
, &lock
);
1924 scm_ungetc_unlocked (c
, port
);
1926 scm_i_pthread_mutex_unlock (lock
);
1931 scm_ungets_unlocked (const char *s
, int n
, SCM port
)
1933 /* This is simple minded and inefficient, but unreading strings is
1934 * probably not a common operation, and remember that line and
1935 * column numbers have to be handled...
1937 * Please feel free to write an optimized version!
1940 scm_ungetc_unlocked (s
[n
], port
);
1944 scm_ungets (const char *s
, int n
, SCM port
)
1946 scm_i_pthread_mutex_t
*lock
;
1947 scm_c_lock_port (port
, &lock
);
1948 scm_ungets_unlocked (s
, n
, port
);
1950 scm_i_pthread_mutex_unlock (lock
);
1954 SCM_DEFINE (scm_peek_char
, "peek-char", 0, 1, 0,
1956 "Return the next character available from @var{port},\n"
1957 "@emph{without} updating @var{port} to point to the following\n"
1958 "character. If no more characters are available, the\n"
1959 "end-of-file object is returned.\n"
1961 "The value returned by\n"
1962 "a call to @code{peek-char} is the same as the value that would\n"
1963 "have been returned by a call to @code{read-char} on the same\n"
1964 "port. The only difference is that the very next call to\n"
1965 "@code{read-char} or @code{peek-char} on that @var{port} will\n"
1966 "return the value returned by the preceding call to\n"
1967 "@code{peek-char}. In particular, a call to @code{peek-char} on\n"
1968 "an interactive port will hang waiting for input whenever a call\n"
1969 "to @code{read-char} would have hung.\n"
1971 "As for @code{read-char}, a @code{decoding-error} may be raised\n"
1972 "if such a situation occurs. However, unlike with @code{read-char},\n"
1973 "@var{port} still points at the beginning of the erroneous byte\n"
1974 "sequence when the error is raised.\n")
1975 #define FUNC_NAME s_scm_peek_char
1980 char bytes
[SCM_MBCHAR_BUF_SIZE
];
1981 long column
, line
, i
;
1984 if (SCM_UNBNDP (port
))
1985 port
= scm_current_input_port ();
1986 SCM_VALIDATE_OPINPORT (1, port
);
1988 column
= SCM_COL (port
);
1989 line
= SCM_LINUM (port
);
1991 err
= get_codepoint (port
, &c
, bytes
, &len
);
1993 for (i
= len
- 1; i
>= 0; i
--)
1994 scm_unget_byte_unlocked (bytes
[i
], port
);
1996 SCM_COL (port
) = column
;
1997 SCM_LINUM (port
) = line
;
1999 if (SCM_UNLIKELY (err
!= 0))
2001 scm_decoding_error (FUNC_NAME
, err
, "input decoding error", port
);
2003 /* Shouldn't happen since `catch' always aborts to prompt. */
2004 result
= SCM_BOOL_F
;
2007 result
= SCM_EOF_VAL
;
2009 result
= SCM_MAKE_CHAR (c
);
2015 SCM_DEFINE (scm_unread_char
, "unread-char", 1, 1, 0,
2016 (SCM cobj
, SCM port
),
2017 "Place character @var{cobj} in @var{port} so that it will be\n"
2018 "read by the next read operation. If called multiple times, the\n"
2019 "unread characters will be read again in last-in first-out\n"
2020 "order. If @var{port} is not supplied, the current input port\n"
2022 #define FUNC_NAME s_scm_unread_char
2026 SCM_VALIDATE_CHAR (1, cobj
);
2027 if (SCM_UNBNDP (port
))
2028 port
= scm_current_input_port ();
2029 SCM_VALIDATE_OPINPORT (2, port
);
2031 c
= SCM_CHAR (cobj
);
2033 scm_ungetc_unlocked (c
, port
);
2038 SCM_DEFINE (scm_unread_string
, "unread-string", 2, 0, 0,
2039 (SCM str
, SCM port
),
2040 "Place the string @var{str} in @var{port} so that its characters will be\n"
2041 "read in subsequent read operations. If called multiple times, the\n"
2042 "unread characters will be read again in last-in first-out order. If\n"
2043 "@var{port} is not supplied, the current-input-port is used.")
2044 #define FUNC_NAME s_scm_unread_string
2047 SCM_VALIDATE_STRING (1, str
);
2048 if (SCM_UNBNDP (port
))
2049 port
= scm_current_input_port ();
2050 SCM_VALIDATE_OPINPORT (2, port
);
2052 n
= scm_i_string_length (str
);
2055 scm_ungetc_unlocked (scm_i_string_ref (str
, n
), port
);
2064 /* Manipulating the buffers. */
2066 /* This routine does not take any locks, as it is usually called as part
2067 of a port implementation. */
2069 scm_port_non_buffer (scm_t_port
*pt
)
2071 pt
->read_pos
= pt
->read_buf
= pt
->read_end
= &pt
->shortbuf
;
2072 pt
->write_buf
= pt
->write_pos
= &pt
->shortbuf
;
2073 pt
->read_buf_size
= pt
->write_buf_size
= 1;
2074 pt
->write_end
= pt
->write_buf
+ pt
->write_buf_size
;
2077 /* this should only be called when the read buffer is empty. it
2078 tries to refill the read buffer. it returns the first char from
2079 the port, which is either EOF or *(pt->read_pos). */
2081 scm_fill_input_unlocked (SCM port
)
2083 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2085 assert (pt
->read_pos
== pt
->read_end
);
2087 if (pt
->read_buf
== pt
->putback_buf
)
2089 /* finished reading put-back chars. */
2090 pt
->read_buf
= pt
->saved_read_buf
;
2091 pt
->read_pos
= pt
->saved_read_pos
;
2092 pt
->read_end
= pt
->saved_read_end
;
2093 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2094 if (pt
->read_pos
< pt
->read_end
)
2095 return *(pt
->read_pos
);
2097 return SCM_PORT_DESCRIPTOR (port
)->fill_input (port
);
2101 scm_fill_input (SCM port
)
2103 scm_i_pthread_mutex_t
*lock
;
2106 scm_c_lock_port (port
, &lock
);
2107 ret
= scm_fill_input_unlocked (port
);
2109 scm_i_pthread_mutex_unlock (lock
);
2115 /* Move up to READ_LEN bytes from PORT's putback and/or read buffers
2116 into memory starting at DEST. Return the number of bytes moved.
2117 PORT's line/column numbers are left unchanged. */
2119 scm_take_from_input_buffers (SCM port
, char *dest
, size_t read_len
)
2121 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2122 size_t bytes_read
= 0;
2123 size_t from_buf
= min (pt
->read_end
- pt
->read_pos
, read_len
);
2127 memcpy (dest
, pt
->read_pos
, from_buf
);
2128 pt
->read_pos
+= from_buf
;
2129 bytes_read
+= from_buf
;
2130 read_len
-= from_buf
;
2134 /* if putback was active, try the real input buffer too. */
2135 if (pt
->read_buf
== pt
->putback_buf
)
2137 from_buf
= min (pt
->saved_read_end
- pt
->saved_read_pos
, read_len
);
2140 memcpy (dest
, pt
->saved_read_pos
, from_buf
);
2141 pt
->saved_read_pos
+= from_buf
;
2142 bytes_read
+= from_buf
;
2149 /* Clear a port's read buffers, returning the contents. */
2150 SCM_DEFINE (scm_drain_input
, "drain-input", 1, 0, 0,
2152 "This procedure clears a port's input buffers, similar\n"
2153 "to the way that force-output clears the output buffer. The\n"
2154 "contents of the buffers are returned as a single string, e.g.,\n"
2157 "(define p (open-input-file ...))\n"
2158 "(drain-input p) => empty string, nothing buffered yet.\n"
2159 "(unread-char (read-char p) p)\n"
2160 "(drain-input p) => initial chars from p, up to the buffer size.\n"
2162 "Draining the buffers may be useful for cleanly finishing\n"
2163 "buffered I/O so that the file descriptor can be used directly\n"
2164 "for further input.")
2165 #define FUNC_NAME s_scm_drain_input
2172 SCM_VALIDATE_OPINPORT (1, port
);
2173 pt
= SCM_PTAB_ENTRY (port
);
2175 count
= pt
->read_end
- pt
->read_pos
;
2176 if (pt
->read_buf
== pt
->putback_buf
)
2177 count
+= pt
->saved_read_end
- pt
->saved_read_pos
;
2181 result
= scm_i_make_string (count
, &data
, 0);
2182 scm_take_from_input_buffers (port
, data
, count
);
2185 result
= scm_nullstr
;
2192 scm_end_input_unlocked (SCM port
)
2195 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2197 if (pt
->read_buf
== pt
->putback_buf
)
2199 offset
= pt
->read_end
- pt
->read_pos
;
2200 pt
->read_buf
= pt
->saved_read_buf
;
2201 pt
->read_pos
= pt
->saved_read_pos
;
2202 pt
->read_end
= pt
->saved_read_end
;
2203 pt
->read_buf_size
= pt
->saved_read_buf_size
;
2208 SCM_PORT_DESCRIPTOR (port
)->end_input (port
, offset
);
2212 scm_end_input (SCM port
)
2214 scm_i_pthread_mutex_t
*lock
;
2215 scm_c_lock_port (port
, &lock
);
2216 scm_end_input_unlocked (port
);
2218 scm_i_pthread_mutex_unlock (lock
);
2222 SCM_DEFINE (scm_force_output
, "force-output", 0, 1, 0,
2224 "Flush the specified output port, or the current output port if @var{port}\n"
2225 "is omitted. The current output buffer contents are passed to the\n"
2226 "underlying port implementation (e.g., in the case of fports, the\n"
2227 "data will be written to the file and the output buffer will be cleared.)\n"
2228 "It has no effect on an unbuffered port.\n\n"
2229 "The return value is unspecified.")
2230 #define FUNC_NAME s_scm_force_output
2232 if (SCM_UNBNDP (port
))
2233 port
= scm_current_output_port ();
2236 port
= SCM_COERCE_OUTPORT (port
);
2237 SCM_VALIDATE_OPOUTPORT (1, port
);
2239 scm_flush_unlocked (port
);
2240 return SCM_UNSPECIFIED
;
2245 scm_flush_unlocked (SCM port
)
2247 SCM_PORT_DESCRIPTOR (port
)->flush (port
);
2251 scm_flush (SCM port
)
2253 scm_i_pthread_mutex_t
*lock
;
2254 scm_c_lock_port (port
, &lock
);
2255 scm_flush_unlocked (port
);
2257 scm_i_pthread_mutex_unlock (lock
);
2267 scm_putc (char c
, SCM port
)
2269 scm_i_pthread_mutex_t
*lock
;
2270 scm_c_lock_port (port
, &lock
);
2271 scm_putc_unlocked (c
, port
);
2273 scm_i_pthread_mutex_unlock (lock
);
2278 scm_puts (const char *s
, SCM port
)
2280 scm_i_pthread_mutex_t
*lock
;
2281 scm_c_lock_port (port
, &lock
);
2282 scm_puts_unlocked (s
, port
);
2284 scm_i_pthread_mutex_unlock (lock
);
2290 * Used by an application to write arbitrary number of bytes to an SCM
2291 * port. Similar semantics as libc write. However, unlike libc
2292 * write, scm_c_write writes the requested number of bytes and has no
2295 * Warning: Doesn't update port line and column counts!
2298 scm_c_write_unlocked (SCM port
, const void *ptr
, size_t size
)
2299 #define FUNC_NAME "scm_c_write"
2302 scm_t_ptob_descriptor
*ptob
;
2304 SCM_VALIDATE_OPOUTPORT (1, port
);
2306 pt
= SCM_PTAB_ENTRY (port
);
2307 ptob
= SCM_PORT_DESCRIPTOR (port
);
2309 if (pt
->rw_active
== SCM_PORT_READ
)
2310 scm_end_input_unlocked (port
);
2312 ptob
->write (port
, ptr
, size
);
2315 pt
->rw_active
= SCM_PORT_WRITE
;
2320 scm_c_write (SCM port
, const void *ptr
, size_t size
)
2322 scm_i_pthread_mutex_t
*lock
;
2323 scm_c_lock_port (port
, &lock
);
2324 scm_c_write_unlocked (port
, ptr
, size
);
2326 scm_i_pthread_mutex_unlock (lock
);
2332 * This function differs from scm_c_write; it updates port line and
2335 scm_lfwrite_unlocked (const char *ptr
, size_t size
, SCM port
)
2337 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2338 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2340 if (pt
->rw_active
== SCM_PORT_READ
)
2341 scm_end_input_unlocked (port
);
2343 ptob
->write (port
, ptr
, size
);
2345 for (; size
; ptr
++, size
--)
2346 update_port_lf ((scm_t_wchar
) (unsigned char) *ptr
, port
);
2349 pt
->rw_active
= SCM_PORT_WRITE
;
2353 scm_lfwrite (const char *ptr
, size_t size
, SCM port
)
2355 scm_i_pthread_mutex_t
*lock
;
2356 scm_c_lock_port (port
, &lock
);
2357 scm_lfwrite_unlocked (ptr
, size
, port
);
2359 scm_i_pthread_mutex_unlock (lock
);
2363 /* Write STR to PORT from START inclusive to END exclusive. */
2365 scm_lfwrite_substr (SCM str
, size_t start
, size_t end
, SCM port
)
2367 scm_t_port
*pt
= SCM_PTAB_ENTRY (port
);
2369 if (pt
->rw_active
== SCM_PORT_READ
)
2370 scm_end_input_unlocked (port
);
2372 if (end
== (size_t) -1)
2373 end
= scm_i_string_length (str
);
2375 scm_i_display_substring (str
, start
, end
, port
);
2378 pt
->rw_active
= SCM_PORT_WRITE
;
2384 /* Querying and setting positions, and character availability. */
2386 SCM_DEFINE (scm_char_ready_p
, "char-ready?", 0, 1, 0,
2388 "Return @code{#t} if a character is ready on input @var{port}\n"
2389 "and return @code{#f} otherwise. If @code{char-ready?} returns\n"
2390 "@code{#t} then the next @code{read-char} operation on\n"
2391 "@var{port} is guaranteed not to hang. If @var{port} is a file\n"
2392 "port at end of file then @code{char-ready?} returns @code{#t}.\n"
2394 "@code{char-ready?} exists to make it possible for a\n"
2395 "program to accept characters from interactive ports without\n"
2396 "getting stuck waiting for input. Any input editors associated\n"
2397 "with such ports must make sure that characters whose existence\n"
2398 "has been asserted by @code{char-ready?} cannot be rubbed out.\n"
2399 "If @code{char-ready?} were to return @code{#f} at end of file,\n"
2400 "a port at end of file would be indistinguishable from an\n"
2401 "interactive port that has no ready characters.")
2402 #define FUNC_NAME s_scm_char_ready_p
2406 if (SCM_UNBNDP (port
))
2407 port
= scm_current_input_port ();
2408 /* It's possible to close the current input port, so validate even in
2410 SCM_VALIDATE_OPINPORT (1, port
);
2412 pt
= SCM_PTAB_ENTRY (port
);
2414 /* if the current read buffer is filled, or the
2415 last pushed-back char has been read and the saved buffer is
2416 filled, result is true. */
2417 if (pt
->read_pos
< pt
->read_end
2418 || (pt
->read_buf
== pt
->putback_buf
2419 && pt
->saved_read_pos
< pt
->saved_read_end
))
2423 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (port
);
2425 if (ptob
->input_waiting
)
2426 return scm_from_bool(ptob
->input_waiting (port
));
2433 SCM_DEFINE (scm_seek
, "seek", 3, 0, 0,
2434 (SCM fd_port
, SCM offset
, SCM whence
),
2435 "Sets the current position of @var{fd_port} to the integer\n"
2436 "@var{offset}, which is interpreted according to the value of\n"
2439 "One of the following variables should be supplied for\n"
2441 "@defvar SEEK_SET\n"
2442 "Seek from the beginning of the file.\n"
2444 "@defvar SEEK_CUR\n"
2445 "Seek from the current position.\n"
2447 "@defvar SEEK_END\n"
2448 "Seek from the end of the file.\n"
2450 "If @var{fd_port} is a file descriptor, the underlying system\n"
2451 "call is @code{lseek}. @var{port} may be a string port.\n"
2453 "The value returned is the new position in the file. This means\n"
2454 "that the current position of a port can be obtained using:\n"
2456 "(seek port 0 SEEK_CUR)\n"
2458 #define FUNC_NAME s_scm_seek
2462 fd_port
= SCM_COERCE_OUTPORT (fd_port
);
2464 how
= scm_to_int (whence
);
2465 if (how
!= SEEK_SET
&& how
!= SEEK_CUR
&& how
!= SEEK_END
)
2466 SCM_OUT_OF_RANGE (3, whence
);
2468 if (SCM_OPPORTP (fd_port
))
2470 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (fd_port
);
2471 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2472 off_t_or_off64_t rv
;
2475 SCM_MISC_ERROR ("port is not seekable",
2476 scm_cons (fd_port
, SCM_EOL
));
2478 rv
= ptob
->seek (fd_port
, off
, how
);
2479 return scm_from_off_t_or_off64_t (rv
);
2481 else /* file descriptor?. */
2483 off_t_or_off64_t off
= scm_to_off_t_or_off64_t (offset
);
2484 off_t_or_off64_t rv
;
2485 rv
= lseek_or_lseek64 (scm_to_int (fd_port
), off
, how
);
2488 return scm_from_off_t_or_off64_t (rv
);
2497 /* Mingw has ftruncate(), perhaps implemented above using chsize, but
2498 doesn't have the filename version truncate(), hence this code. */
2499 #if HAVE_FTRUNCATE && ! HAVE_TRUNCATE
2501 truncate (const char *file
, off_t length
)
2505 fdes
= open (file
, O_BINARY
| O_WRONLY
);
2509 ret
= ftruncate (fdes
, length
);
2512 int save_errno
= errno
;
2518 return close (fdes
);
2520 #endif /* HAVE_FTRUNCATE && ! HAVE_TRUNCATE */
2522 SCM_DEFINE (scm_truncate_file
, "truncate-file", 1, 1, 0,
2523 (SCM object
, SCM length
),
2524 "Truncate file @var{object} to @var{length} bytes. @var{object}\n"
2525 "can be a filename string, a port object, or an integer file\n"
2527 "The return value is unspecified.\n"
2529 "For a port or file descriptor @var{length} can be omitted, in\n"
2530 "which case the file is truncated at the current position (per\n"
2531 "@code{ftell} above).\n"
2533 "On most systems a file can be extended by giving a length\n"
2534 "greater than the current size, but this is not mandatory in the\n"
2536 #define FUNC_NAME s_scm_truncate_file
2540 /* "object" can be a port, fdes or filename.
2542 Negative "length" makes no sense, but it's left to truncate() or
2543 ftruncate() to give back an error for that (normally EINVAL).
2546 if (SCM_UNBNDP (length
))
2548 /* must supply length if object is a filename. */
2549 if (scm_is_string (object
))
2550 SCM_MISC_ERROR("must supply length if OBJECT is a filename", SCM_EOL
);
2552 length
= scm_seek (object
, SCM_INUM0
, scm_from_int (SEEK_CUR
));
2555 object
= SCM_COERCE_OUTPORT (object
);
2556 if (scm_is_integer (object
))
2558 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2559 SCM_SYSCALL (rv
= ftruncate_or_ftruncate64 (scm_to_int (object
),
2562 else if (SCM_OPOUTPORTP (object
))
2564 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2565 scm_t_port
*pt
= SCM_PTAB_ENTRY (object
);
2566 scm_t_ptob_descriptor
*ptob
= SCM_PORT_DESCRIPTOR (object
);
2568 if (!ptob
->truncate
)
2569 SCM_MISC_ERROR ("port is not truncatable", SCM_EOL
);
2570 if (pt
->rw_active
== SCM_PORT_READ
)
2571 scm_end_input_unlocked (object
);
2572 else if (pt
->rw_active
== SCM_PORT_WRITE
)
2573 ptob
->flush (object
);
2575 ptob
->truncate (object
, c_length
);
2580 off_t_or_off64_t c_length
= scm_to_off_t_or_off64_t (length
);
2581 char *str
= scm_to_locale_string (object
);
2583 SCM_SYSCALL (rv
= truncate_or_truncate64 (str
, c_length
));
2590 return SCM_UNSPECIFIED
;
2594 SCM_DEFINE (scm_port_line
, "port-line", 1, 0, 0,
2596 "Return the current line number for @var{port}.\n"
2598 "The first line of a file is 0. But you might want to add 1\n"
2599 "when printing line numbers, since starting from 1 is\n"
2600 "traditional in error messages, and likely to be more natural to\n"
2602 #define FUNC_NAME s_scm_port_line
2604 port
= SCM_COERCE_OUTPORT (port
);
2605 SCM_VALIDATE_OPENPORT (1, port
);
2606 return scm_from_long (SCM_LINUM (port
));
2610 SCM_DEFINE (scm_set_port_line_x
, "set-port-line!", 2, 0, 0,
2611 (SCM port
, SCM line
),
2612 "Set the current line number for @var{port} to @var{line}. The\n"
2613 "first line of a file is 0.")
2614 #define FUNC_NAME s_scm_set_port_line_x
2616 port
= SCM_COERCE_OUTPORT (port
);
2617 SCM_VALIDATE_OPENPORT (1, port
);
2618 SCM_PTAB_ENTRY (port
)->line_number
= scm_to_long (line
);
2619 return SCM_UNSPECIFIED
;
2623 SCM_DEFINE (scm_port_column
, "port-column", 1, 0, 0,
2625 "Return the current column number of @var{port}.\n"
2626 "If the number is\n"
2627 "unknown, the result is #f. Otherwise, the result is a 0-origin integer\n"
2628 "- i.e. the first character of the first line is line 0, column 0.\n"
2629 "(However, when you display a file position, for example in an error\n"
2630 "message, we recommend you add 1 to get 1-origin integers. This is\n"
2631 "because lines and column numbers traditionally start with 1, and that is\n"
2632 "what non-programmers will find most natural.)")
2633 #define FUNC_NAME s_scm_port_column
2635 port
= SCM_COERCE_OUTPORT (port
);
2636 SCM_VALIDATE_OPENPORT (1, port
);
2637 return scm_from_int (SCM_COL (port
));
2641 SCM_DEFINE (scm_set_port_column_x
, "set-port-column!", 2, 0, 0,
2642 (SCM port
, SCM column
),
2643 "Set the current column of @var{port}. Before reading the first\n"
2644 "character on a line the column should be 0.")
2645 #define FUNC_NAME s_scm_set_port_column_x
2647 port
= SCM_COERCE_OUTPORT (port
);
2648 SCM_VALIDATE_OPENPORT (1, port
);
2649 SCM_PTAB_ENTRY (port
)->column_number
= scm_to_int (column
);
2650 return SCM_UNSPECIFIED
;
2654 SCM_DEFINE (scm_port_filename
, "port-filename", 1, 0, 0,
2656 "Return the filename associated with @var{port}, or @code{#f}\n"
2657 "if no filename is associated with the port.")
2658 #define FUNC_NAME s_scm_port_filename
2660 port
= SCM_COERCE_OUTPORT (port
);
2661 SCM_VALIDATE_OPENPORT (1, port
);
2662 return SCM_FILENAME (port
);
2666 SCM_DEFINE (scm_set_port_filename_x
, "set-port-filename!", 2, 0, 0,
2667 (SCM port
, SCM filename
),
2668 "Change the filename associated with @var{port}, using the current input\n"
2669 "port if none is specified. Note that this does not change the port's\n"
2670 "source of data, but only the value that is returned by\n"
2671 "@code{port-filename} and reported in diagnostic output.")
2672 #define FUNC_NAME s_scm_set_port_filename_x
2674 port
= SCM_COERCE_OUTPORT (port
);
2675 SCM_VALIDATE_OPENPORT (1, port
);
2676 /* We allow the user to set the filename to whatever he likes. */
2677 SCM_SET_FILENAME (port
, filename
);
2678 return SCM_UNSPECIFIED
;
2685 /* Implementation helpers for port printing functions. */
2688 scm_print_port_mode (SCM exp
, SCM port
)
2690 scm_puts_unlocked (SCM_CLOSEDP (exp
)
2692 : (SCM_RDNG
& SCM_CELL_WORD_0 (exp
)
2693 ? (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2696 : (SCM_WRTNG
& SCM_CELL_WORD_0 (exp
)
2703 scm_port_print (SCM exp
, SCM port
, scm_print_state
*pstate SCM_UNUSED
)
2705 char *type
= SCM_PTOBNAME (SCM_PTOBNUM (exp
));
2708 scm_puts_unlocked ("#<", port
);
2709 scm_print_port_mode (exp
, port
);
2710 scm_puts_unlocked (type
, port
);
2711 scm_putc_unlocked (' ', port
);
2712 scm_uintprint (SCM_CELL_WORD_1 (exp
), 16, port
);
2713 scm_putc_unlocked ('>', port
);
2720 /* Iterating over all ports. */
2722 struct for_each_data
2724 void (*proc
) (void *data
, SCM p
);
2729 for_each_trampoline (void *data
, SCM port
, SCM result
)
2731 struct for_each_data
*d
= data
;
2733 d
->proc (d
->data
, port
);
2739 scm_c_port_for_each (void (*proc
)(void *data
, SCM p
), void *data
)
2741 struct for_each_data d
;
2746 scm_c_weak_set_fold (for_each_trampoline
, &d
, SCM_EOL
,
2747 scm_i_port_weak_set
);
2751 scm_for_each_trampoline (void *data
, SCM port
)
2753 scm_call_1 (SCM_PACK_POINTER (data
), port
);
2756 SCM_DEFINE (scm_port_for_each
, "port-for-each", 1, 0, 0,
2758 "Apply @var{proc} to each port in the Guile port table\n"
2759 "in turn. The return value is unspecified. More specifically,\n"
2760 "@var{proc} is applied exactly once to every port that exists\n"
2761 "in the system at the time @code{port-for-each} is invoked.\n"
2762 "Changes to the port table while @code{port-for-each} is running\n"
2763 "have no effect as far as @code{port-for-each} is concerned.")
2764 #define FUNC_NAME s_scm_port_for_each
2766 SCM_VALIDATE_PROC (1, proc
);
2768 scm_c_port_for_each (scm_for_each_trampoline
, SCM_UNPACK_POINTER (proc
));
2770 return SCM_UNSPECIFIED
;
2775 flush_output_port (void *closure
, SCM port
)
2777 if (SCM_OPOUTPORTP (port
))
2778 scm_flush_unlocked (port
);
2781 SCM_DEFINE (scm_flush_all_ports
, "flush-all-ports", 0, 0, 0,
2783 "Equivalent to calling @code{force-output} on\n"
2784 "all open output ports. The return value is unspecified.")
2785 #define FUNC_NAME s_scm_flush_all_ports
2787 scm_c_port_for_each (&flush_output_port
, NULL
);
2788 return SCM_UNSPECIFIED
;
2797 scm_t_bits scm_tc16_void_port
= 0;
2799 static int fill_input_void_port (SCM port SCM_UNUSED
)
2805 write_void_port (SCM port SCM_UNUSED
,
2806 const void *data SCM_UNUSED
,
2807 size_t size SCM_UNUSED
)
2812 scm_i_void_port (long mode_bits
)
2816 ret
= scm_c_make_port (scm_tc16_void_port
, mode_bits
, 0);
2818 scm_port_non_buffer (SCM_PTAB_ENTRY (ret
));
2824 scm_void_port (char *mode_str
)
2826 return scm_i_void_port (scm_mode_bits (mode_str
));
2829 SCM_DEFINE (scm_sys_make_void_port
, "%make-void-port", 1, 0, 0,
2831 "Create and return a new void port. A void port acts like\n"
2832 "@file{/dev/null}. The @var{mode} argument\n"
2833 "specifies the input/output modes for this port: see the\n"
2834 "documentation for @code{open-file} in @ref{File Ports}.")
2835 #define FUNC_NAME s_scm_sys_make_void_port
2837 return scm_i_void_port (scm_i_mode_bits (mode
));
2844 /* Initialization. */
2849 /* lseek() symbols. */
2850 scm_c_define ("SEEK_SET", scm_from_int (SEEK_SET
));
2851 scm_c_define ("SEEK_CUR", scm_from_int (SEEK_CUR
));
2852 scm_c_define ("SEEK_END", scm_from_int (SEEK_END
));
2854 scm_tc16_void_port
= scm_make_port_type ("void", fill_input_void_port
,
2857 cur_inport_fluid
= scm_make_fluid ();
2858 cur_outport_fluid
= scm_make_fluid ();
2859 cur_errport_fluid
= scm_make_fluid ();
2860 cur_loadport_fluid
= scm_make_fluid ();
2862 scm_i_port_weak_set
= scm_c_make_weak_set (31);
2864 #include "libguile/ports.x"
2866 /* Use Latin-1 as the default port encoding. */
2867 SCM_VARIABLE_SET (default_port_encoding_var
,
2868 scm_make_fluid_with_default (SCM_BOOL_F
));
2869 scm_port_encoding_init
= 1;
2871 SCM_VARIABLE_SET (default_conversion_strategy_var
,
2872 scm_make_fluid_with_default (sym_substitute
));
2873 scm_conversion_strategy_init
= 1;
2875 /* These bindings are used when boot-9 turns `current-input-port' et
2876 al into parameters. They are then removed from the guile module. */
2877 scm_c_define ("%current-input-port-fluid", cur_inport_fluid
);
2878 scm_c_define ("%current-output-port-fluid", cur_outport_fluid
);
2879 scm_c_define ("%current-error-port-fluid", cur_errport_fluid
);