1 /* Copyright (C) 1995-1999, 2000, 2001, 2002, 2003, 2004, 2006, 2008,
2 * 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
34 #include "libguile/_scm.h"
35 #include "libguile/chars.h"
36 #include "libguile/continuations.h"
37 #include "libguile/smob.h"
38 #include "libguile/control.h"
39 #include "libguile/eval.h"
40 #include "libguile/macros.h"
41 #include "libguile/procprop.h"
42 #include "libguile/read.h"
43 #include "libguile/weaks.h"
44 #include "libguile/programs.h"
45 #include "libguile/alist.h"
46 #include "libguile/struct.h"
47 #include "libguile/ports.h"
48 #include "libguile/root.h"
49 #include "libguile/strings.h"
50 #include "libguile/strports.h"
51 #include "libguile/vectors.h"
52 #include "libguile/numbers.h"
53 #include "libguile/vm.h"
55 #include "libguile/validate.h"
56 #include "libguile/print.h"
58 #include "libguile/private-options.h"
62 /* Character printers. */
64 static size_t display_string (const void *, int, size_t, SCM
,
65 scm_t_string_failed_conversion_handler
);
67 static int display_character (scm_t_wchar
, SCM
,
68 scm_t_string_failed_conversion_handler
);
70 static void write_character (scm_t_wchar
, SCM
, int);
72 static void write_character_escaped (scm_t_wchar
, int, SCM
);
76 /* {Names of immediate symbols}
78 * This table must agree with the declarations in scm.h: {Immediate Symbols}.
81 /* This table must agree with the list of flags in tags.h. */
82 static const char *iflagnames
[] =
85 "#nil", /* Elisp nil value. Should print from elisp as symbol `nil'. */
86 "#<XXX UNUSED LISP FALSE -- DO NOT USE -- SHOULD NEVER BE SEEN XXX>",
89 "#<XXX UNUSED BOOLEAN 0 -- DO NOT USE -- SHOULD NEVER BE SEEN XXX>",
90 "#<XXX UNUSED BOOLEAN 1 -- DO NOT USE -- SHOULD NEVER BE SEEN XXX>",
91 "#<XXX UNUSED BOOLEAN 2 -- DO NOT USE -- SHOULD NEVER BE SEEN XXX>",
96 /* Unbound slot marker for GOOPS. For internal use in GOOPS only. */
100 SCM_SYMBOL (sym_reader
, "reader");
102 scm_t_option scm_print_opts
[] = {
103 { SCM_OPTION_SCM
, "highlight-prefix", (scm_t_bits
)SCM_BOOL_F
,
104 "The string to print before highlighted values." },
105 { SCM_OPTION_SCM
, "highlight-suffix", (scm_t_bits
)SCM_BOOL_F
,
106 "The string to print after highlighted values." },
107 { SCM_OPTION_SCM
, "quote-keywordish-symbols", (scm_t_bits
)SCM_BOOL_F
,
108 "How to print symbols that have a colon as their first or last character. "
109 "The value '#f' does not quote the colons; '#t' quotes them; "
110 "'reader' quotes them when the reader option 'keywords' is not '#f'."
115 SCM_DEFINE (scm_print_options
, "print-options-interface", 0, 1, 0,
117 "Option interface for the print options. Instead of using\n"
118 "this procedure directly, use the procedures\n"
119 "@code{print-enable}, @code{print-disable}, @code{print-set!}\n"
120 "and @code{print-options}.")
121 #define FUNC_NAME s_scm_print_options
123 SCM ans
= scm_options (setting
,
131 /* {Printing of Scheme Objects}
134 /* Detection of circular references.
136 * Due to other constraints in the implementation, this code has bad
137 * time complexity (O (depth * N)), The printer code can be
138 * rewritten to be O(N).
140 #define PUSH_REF(pstate, obj) \
143 PSTATE_STACK_SET (pstate, pstate->top, obj); \
145 if (pstate->top == pstate->ceiling) \
146 grow_ref_stack (pstate); \
149 #define ENTER_NESTED_DATA(pstate, obj, label) \
152 register unsigned long i; \
153 for (i = 0; i < pstate->top; ++i) \
154 if (scm_is_eq (PSTATE_STACK_REF (pstate, i), (obj))) \
156 if (pstate->fancyp) \
158 if (pstate->top - pstate->list_offset >= pstate->level) \
160 scm_putc ('#', port); \
164 PUSH_REF(pstate, obj); \
167 #define EXIT_NESTED_DATA(pstate) \
171 PSTATE_STACK_SET (pstate, pstate->top, SCM_UNDEFINED); \
175 SCM scm_print_state_vtable
= SCM_BOOL_F
;
176 static SCM print_state_pool
= SCM_EOL
;
177 scm_i_pthread_mutex_t print_state_mutex
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
179 #ifdef GUILE_DEBUG /* Used for debugging purposes */
181 SCM_DEFINE (scm_current_pstate
, "current-pstate", 0, 0, 0,
183 "Return the current-pstate -- the car of the\n"
184 "@code{print_state_pool}. @code{current-pstate} is only\n"
185 "included in @code{--enable-guile-debug} builds.")
186 #define FUNC_NAME s_scm_current_pstate
188 if (!scm_is_null (print_state_pool
))
189 return SCM_CAR (print_state_pool
);
197 #define PSTATE_SIZE 50L
200 make_print_state (void)
203 = scm_make_struct (scm_print_state_vtable
, SCM_INUM0
, SCM_EOL
);
204 scm_print_state
*pstate
= SCM_PRINT_STATE (print_state
);
205 pstate
->ref_vect
= scm_c_make_vector (PSTATE_SIZE
, SCM_UNDEFINED
);
206 pstate
->ceiling
= SCM_SIMPLE_VECTOR_LENGTH (pstate
->ref_vect
);
207 pstate
->highlight_objects
= SCM_EOL
;
212 scm_make_print_state ()
214 SCM answer
= SCM_BOOL_F
;
216 /* First try to allocate a print state from the pool */
217 scm_i_pthread_mutex_lock (&print_state_mutex
);
218 if (!scm_is_null (print_state_pool
))
220 answer
= SCM_CAR (print_state_pool
);
221 print_state_pool
= SCM_CDR (print_state_pool
);
223 scm_i_pthread_mutex_unlock (&print_state_mutex
);
225 return scm_is_false (answer
) ? make_print_state () : answer
;
229 scm_free_print_state (SCM print_state
)
232 scm_print_state
*pstate
= SCM_PRINT_STATE (print_state
);
233 /* Cleanup before returning print state to pool.
234 * It is better to do it here. Doing it in scm_prin1
235 * would cost more since that function is called much more
239 pstate
->revealed
= 0;
240 pstate
->highlight_objects
= SCM_EOL
;
241 scm_i_pthread_mutex_lock (&print_state_mutex
);
242 handle
= scm_cons (print_state
, print_state_pool
);
243 print_state_pool
= handle
;
244 scm_i_pthread_mutex_unlock (&print_state_mutex
);
248 scm_i_port_with_print_state (SCM port
, SCM print_state
)
250 if (SCM_UNBNDP (print_state
))
252 if (SCM_PORT_WITH_PS_P (port
))
255 print_state
= scm_make_print_state ();
256 /* port does not need to be coerced since it doesn't have ps */
259 port
= SCM_COERCE_OUTPORT (port
);
260 SCM_RETURN_NEWSMOB (scm_tc16_port_with_ps
,
261 SCM_UNPACK (scm_cons (port
, print_state
)));
265 grow_ref_stack (scm_print_state
*pstate
)
267 SCM old_vect
= pstate
->ref_vect
;
268 size_t old_size
= SCM_SIMPLE_VECTOR_LENGTH (old_vect
);
269 size_t new_size
= 2 * pstate
->ceiling
;
270 SCM new_vect
= scm_c_make_vector (new_size
, SCM_UNDEFINED
);
273 for (i
= 0; i
!= old_size
; ++i
)
274 SCM_SIMPLE_VECTOR_SET (new_vect
, i
, SCM_SIMPLE_VECTOR_REF (old_vect
, i
));
276 pstate
->ref_vect
= new_vect
;
277 pstate
->ceiling
= new_size
;
280 #define PSTATE_STACK_REF(p,i) SCM_SIMPLE_VECTOR_REF((p)->ref_vect, (i))
281 #define PSTATE_STACK_SET(p,i,v) SCM_SIMPLE_VECTOR_SET((p)->ref_vect, (i), (v))
284 print_circref (SCM port
, scm_print_state
*pstate
, SCM ref
)
287 long self
= pstate
->top
- 1;
289 if (scm_is_pair (PSTATE_STACK_REF (pstate
, i
)))
293 if (!scm_is_pair (PSTATE_STACK_REF (pstate
, i
-1))
294 || !scm_is_eq (SCM_CDR (PSTATE_STACK_REF (pstate
, i
-1)),
295 SCM_CDR (PSTATE_STACK_REF (pstate
, i
))))
301 for (i
= pstate
->top
- 1; 1; --i
)
302 if (scm_is_eq (PSTATE_STACK_REF(pstate
, i
), ref
))
304 scm_putc ('#', port
);
305 scm_intprint (i
- self
, 10, port
);
306 scm_putc ('#', port
);
309 /* Print the name of a symbol. */
312 quote_keywordish_symbol (SCM symbol
)
316 if (scm_i_symbol_ref (symbol
, 0) != ':'
317 && scm_i_symbol_ref (symbol
, scm_i_symbol_length (symbol
) - 1) != ':')
320 option
= SCM_PRINT_KEYWORD_STYLE
;
321 if (scm_is_false (option
))
323 if (scm_is_eq (option
, sym_reader
))
324 return scm_is_true (SCM_PACK (SCM_KEYWORD_STYLE
));
329 scm_i_print_symbol_name (SCM str
, SCM port
)
331 /* This points to the first character that has not yet been written to the
334 /* This points to the character we're currently looking at. */
336 /* If the name contains weird characters, we'll escape them with
337 * backslashes and set this flag; it indicates that we should surround the
338 * name with "#{" and "}#". */
340 /* Backslashes are not sufficient to make a name weird, but if a name is
341 * weird because of other characters, backslahes need to be escaped too.
342 * The first time we see a backslash, we set maybe_weird, and mw_pos points
343 * to the backslash. Then if the name turns out to be weird, we re-process
344 * everything starting from mw_pos.
345 * We could instead make backslashes always weird. This is not necessary
346 * to ensure that the output is (read)-able, but it would make this code
347 * simpler and faster. */
350 size_t len
= scm_i_symbol_length (str
);
351 scm_t_wchar str0
= scm_i_symbol_ref (str
, 0);
353 if (len
== 0 || str0
== '\'' || str0
== '`' || str0
== ','
354 || quote_keywordish_symbol (str
)
355 || (str0
== '.' && len
== 1)
356 || scm_is_true (scm_i_string_to_number (scm_symbol_to_string (str
), 10)))
358 scm_lfwrite ("#{", 2, port
);
362 for (end
= pos
; end
< len
; ++end
)
363 switch (scm_i_symbol_ref (str
, end
))
365 #ifdef BRACKETS_AS_PARENS
374 case SCM_WHITE_SPACES
:
375 case SCM_LINE_INCREMENTORS
:
384 scm_lfwrite ("#{", 2, port
);
388 scm_lfwrite_substr (scm_symbol_to_string (str
), pos
, end
, port
);
392 buf
[1] = (char) (unsigned char) scm_i_symbol_ref (str
, end
);
393 scm_lfwrite (buf
, 2, port
);
410 scm_lfwrite_substr (scm_symbol_to_string (str
), pos
, end
, port
);
412 scm_lfwrite ("}#", 2, port
);
416 scm_print_symbol_name (const char *str
, size_t len
, SCM port
)
418 SCM symbol
= scm_from_locale_symboln (str
, len
);
419 scm_i_print_symbol_name (symbol
, port
);
422 /* Print generally. Handles both write and display according to PSTATE.
424 SCM_GPROC(s_write
, "write", 1, 1, 0, scm_write
, g_write
);
425 SCM_GPROC(s_display
, "display", 1, 1, 0, scm_display
, g_display
);
427 static void iprin1 (SCM exp
, SCM port
, scm_print_state
*pstate
);
430 /* Print a character as an octal or hex escape. */
431 #define PRINT_CHAR_ESCAPE(i, port) \
434 if (!SCM_R6RS_ESCAPES_P) \
435 scm_intprint (i, 8, port); \
438 scm_puts ("x", port); \
439 scm_intprint (i, 16, port); \
446 scm_iprin1 (SCM exp
, SCM port
, scm_print_state
*pstate
)
449 && scm_is_true (scm_memq (exp
, pstate
->highlight_objects
)))
451 scm_display (SCM_PRINT_HIGHLIGHT_PREFIX
, port
);
452 iprin1 (exp
, port
, pstate
);
453 scm_display (SCM_PRINT_HIGHLIGHT_SUFFIX
, port
);
456 iprin1 (exp
, port
, pstate
);
460 iprin1 (SCM exp
, SCM port
, scm_print_state
*pstate
)
462 switch (SCM_ITAG3 (exp
))
466 /* These tc3 tags should never occur in an immediate value. They are
467 * only used in cell types of non-immediates, i. e. the value returned
468 * by SCM_CELL_TYPE (exp) can use these tags.
470 scm_ipruk ("immediate", exp
, port
);
474 scm_intprint (SCM_I_INUM (exp
), 10, port
);
479 if (SCM_WRITINGP (pstate
))
480 write_character (SCM_CHAR (exp
), port
, 0);
483 if (!display_character (SCM_CHAR (exp
), port
,
484 scm_i_get_conversion_strategy (port
)))
485 scm_encoding_error (__func__
, errno
,
486 "cannot convert to output locale",
490 else if (SCM_IFLAGP (exp
)
491 && ((size_t) SCM_IFLAGNUM (exp
) < (sizeof iflagnames
/ sizeof (char *))))
493 scm_puts (iflagnames
[SCM_IFLAGNUM (exp
)], port
);
497 /* unknown immediate value */
498 scm_ipruk ("immediate", exp
, port
);
502 switch (SCM_TYP7 (exp
))
506 ENTER_NESTED_DATA (pstate
, exp
, circref
);
507 if (SCM_OBJ_CLASS_FLAGS (exp
) & SCM_CLASSF_GOOPS
)
509 SCM pwps
, print
= pstate
->writingp
? g_write
: g_display
;
512 pwps
= scm_i_port_with_print_state (port
, pstate
->handle
);
513 pstate
->revealed
= 1;
514 scm_call_generic_2 (print
, exp
, pwps
);
519 scm_print_struct (exp
, port
, pstate
);
521 EXIT_NESTED_DATA (pstate
);
524 case scm_tcs_cons_imcar
:
525 case scm_tcs_cons_nimcar
:
526 ENTER_NESTED_DATA (pstate
, exp
, circref
);
527 scm_iprlist ("(", exp
, ')', port
, pstate
);
528 EXIT_NESTED_DATA (pstate
);
531 print_circref (port
, pstate
, exp
);
534 switch SCM_TYP16 (exp
) {
536 scm_bigprint (exp
, port
, pstate
);
539 scm_print_real (exp
, port
, pstate
);
541 case scm_tc16_complex
:
542 scm_print_complex (exp
, port
, pstate
);
544 case scm_tc16_fraction
:
545 scm_i_print_fraction (exp
, port
, pstate
);
550 if (SCM_WRITINGP (pstate
))
554 display_character ('"', port
, iconveh_question_mark
);
555 len
= scm_i_string_length (exp
);
556 for (i
= 0; i
< len
; ++i
)
557 write_character (scm_i_string_ref (exp
, i
), port
, 1);
559 display_character ('"', port
, iconveh_question_mark
);
560 scm_remember_upto_here_1 (exp
);
566 len
= scm_i_string_length (exp
);
567 printed
= display_string (scm_i_string_data (exp
),
568 scm_i_is_narrow_string (exp
),
570 scm_i_get_conversion_strategy (port
));
571 if (SCM_UNLIKELY (printed
< len
))
572 scm_encoding_error (__func__
, errno
,
573 "cannot convert to output locale",
574 port
, scm_c_string_ref (exp
, printed
));
577 scm_remember_upto_here_1 (exp
);
580 if (scm_i_symbol_is_interned (exp
))
582 scm_i_print_symbol_name (exp
, port
);
583 scm_remember_upto_here_1 (exp
);
587 scm_puts ("#<uninterned-symbol ", port
);
588 scm_i_print_symbol_name (exp
, port
);
589 scm_putc (' ', port
);
590 scm_uintprint (SCM_UNPACK (exp
), 16, port
);
591 scm_putc ('>', port
);
594 case scm_tc7_variable
:
595 scm_i_variable_print (exp
, port
, pstate
);
597 case scm_tc7_program
:
598 scm_i_program_print (exp
, port
, pstate
);
600 case scm_tc7_pointer
:
601 scm_i_pointer_print (exp
, port
, pstate
);
603 case scm_tc7_hashtable
:
604 scm_i_hashtable_print (exp
, port
, pstate
);
607 scm_i_fluid_print (exp
, port
, pstate
);
609 case scm_tc7_dynamic_state
:
610 scm_i_dynamic_state_print (exp
, port
, pstate
);
613 scm_i_frame_print (exp
, port
, pstate
);
615 case scm_tc7_objcode
:
616 scm_i_objcode_print (exp
, port
, pstate
);
619 scm_i_vm_print (exp
, port
, pstate
);
621 case scm_tc7_vm_cont
:
622 scm_i_vm_cont_print (exp
, port
, pstate
);
625 scm_i_prompt_print (exp
, port
, pstate
);
627 case scm_tc7_with_fluids
:
628 scm_i_with_fluids_print (exp
, port
, pstate
);
631 ENTER_NESTED_DATA (pstate
, exp
, circref
);
632 if (SCM_IS_WHVEC (exp
))
633 scm_puts ("#wh(", port
);
635 scm_puts ("#w(", port
);
636 goto common_vector_printer
;
638 case scm_tc7_bytevector
:
639 scm_i_print_bytevector (exp
, port
, pstate
);
642 ENTER_NESTED_DATA (pstate
, exp
, circref
);
643 scm_puts ("#(", port
);
644 common_vector_printer
:
647 long last
= SCM_SIMPLE_VECTOR_LENGTH (exp
) - 1;
650 && SCM_SIMPLE_VECTOR_LENGTH (exp
) > pstate
->length
)
652 last
= pstate
->length
- 1;
655 if (SCM_I_WVECTP (exp
))
657 /* Elements of weak vectors may not be accessed via the
658 `SIMPLE_VECTOR_REF ()' macro. */
659 for (i
= 0; i
< last
; ++i
)
661 scm_iprin1 (scm_c_vector_ref (exp
, i
),
663 scm_putc (' ', port
);
668 for (i
= 0; i
< last
; ++i
)
670 scm_iprin1 (SCM_SIMPLE_VECTOR_REF (exp
, i
), port
, pstate
);
671 scm_putc (' ', port
);
678 scm_iprin1 (scm_c_vector_ref (exp
, i
), port
, pstate
);
681 scm_puts (" ...", port
);
682 scm_putc (')', port
);
684 EXIT_NESTED_DATA (pstate
);
688 register long i
= SCM_PTOBNUM (exp
);
690 && scm_ptobs
[i
].print
691 && (scm_ptobs
[i
].print
) (exp
, port
, pstate
))
696 ENTER_NESTED_DATA (pstate
, exp
, circref
);
697 SCM_SMOB_DESCRIPTOR (exp
).print (exp
, port
, pstate
);
698 EXIT_NESTED_DATA (pstate
);
701 /* case scm_tcs_closures: */
703 scm_ipruk ("type", exp
, port
);
708 /* Print states are necessary for circular reference safe printing.
709 * They are also expensive to allocate. Therefore print states are
710 * kept in a pool so that they can be reused.
713 /* The PORT argument can also be a print-state/port pair, which will
714 * then be used instead of allocating a new print state. This is
715 * useful for continuing a chain of print calls from Scheme. */
718 scm_prin1 (SCM exp
, SCM port
, int writingp
)
720 SCM handle
= SCM_BOOL_F
; /* Will GC protect the handle whilst unlinked */
722 scm_print_state
*pstate
;
725 /* If PORT is a print-state/port pair, use that. Else create a new
728 if (SCM_PORT_WITH_PS_P (port
))
730 pstate_scm
= SCM_PORT_WITH_PS_PS (port
);
731 port
= SCM_PORT_WITH_PS_PORT (port
);
735 /* First try to allocate a print state from the pool */
736 scm_i_pthread_mutex_lock (&print_state_mutex
);
737 if (!scm_is_null (print_state_pool
))
739 handle
= print_state_pool
;
740 print_state_pool
= SCM_CDR (print_state_pool
);
742 scm_i_pthread_mutex_unlock (&print_state_mutex
);
743 if (scm_is_false (handle
))
744 handle
= scm_list_1 (make_print_state ());
745 pstate_scm
= SCM_CAR (handle
);
748 pstate
= SCM_PRINT_STATE (pstate_scm
);
749 old_writingp
= pstate
->writingp
;
750 pstate
->writingp
= writingp
;
751 scm_iprin1 (exp
, port
, pstate
);
752 pstate
->writingp
= old_writingp
;
754 /* Return print state to pool if it has been created above and
755 hasn't escaped to Scheme. */
757 if (scm_is_true (handle
) && !pstate
->revealed
)
759 scm_i_pthread_mutex_lock (&print_state_mutex
);
760 SCM_SETCDR (handle
, print_state_pool
);
761 print_state_pool
= handle
;
762 scm_i_pthread_mutex_unlock (&print_state_mutex
);
766 /* Convert codepoint CH to UTF-8 and store the result in UTF8. Return
767 the number of bytes of the UTF-8-encoded string. */
769 codepoint_to_utf8 (scm_t_wchar ch
, scm_t_uint8 utf8
[4])
772 scm_t_uint32 codepoint
;
774 codepoint
= (scm_t_uint32
) ch
;
776 if (codepoint
<= 0x7f)
779 utf8
[0] = (scm_t_uint8
) codepoint
;
781 else if (codepoint
<= 0x7ffUL
)
784 utf8
[0] = 0xc0 | (codepoint
>> 6);
785 utf8
[1] = 0x80 | (codepoint
& 0x3f);
787 else if (codepoint
<= 0xffffUL
)
790 utf8
[0] = 0xe0 | (codepoint
>> 12);
791 utf8
[1] = 0x80 | ((codepoint
>> 6) & 0x3f);
792 utf8
[2] = 0x80 | (codepoint
& 0x3f);
797 utf8
[0] = 0xf0 | (codepoint
>> 18);
798 utf8
[1] = 0x80 | ((codepoint
>> 12) & 0x3f);
799 utf8
[2] = 0x80 | ((codepoint
>> 6) & 0x3f);
800 utf8
[3] = 0x80 | (codepoint
& 0x3f);
806 /* Display the LEN codepoints in STR to PORT according to STRATEGY;
807 return the number of codepoints successfully displayed. If NARROW_P,
808 then STR is interpreted as a sequence of `char', denoting a Latin-1
809 string; otherwise it's interpreted as a sequence of
812 display_string (const void *str
, int narrow_p
,
813 size_t len
, SCM port
,
814 scm_t_string_failed_conversion_handler strategy
)
817 #define STR_REF(s, x) \
819 ? (scm_t_wchar) ((unsigned char *) (s))[x] \
820 : ((scm_t_wchar *) (s))[x])
825 pt
= SCM_PTAB_ENTRY (port
);
827 if (SCM_UNLIKELY (pt
->output_cd
== (iconv_t
) -1))
828 /* Initialize the conversion descriptors. */
829 scm_i_set_port_encoding_x (port
, pt
->encoding
);
833 while (len
> printed
)
835 size_t done
, utf8_len
, input_left
, output_left
, i
;
836 size_t codepoints_read
, output_len
;
837 char *input
, *output
;
838 char utf8_buf
[256], encoded_output
[256];
841 /* Convert STR to UTF-8. */
842 for (i
= printed
, utf8_len
= 0, input
= utf8_buf
;
843 i
< len
&& utf8_len
+ 4 < sizeof (utf8_buf
);
846 offsets
[utf8_len
] = i
;
847 utf8_len
+= codepoint_to_utf8 (STR_REF (str
, i
),
848 (scm_t_uint8
*) input
);
849 input
= utf8_buf
+ utf8_len
;
853 input_left
= utf8_len
;
855 output
= encoded_output
;
856 output_left
= sizeof (encoded_output
);
858 done
= iconv (pt
->output_cd
, &input
, &input_left
,
859 &output
, &output_left
);
861 output_len
= sizeof (encoded_output
) - output_left
;
863 if (SCM_UNLIKELY (done
== (size_t) -1))
865 int errno_save
= errno
;
867 /* Reset the `iconv' state. */
868 iconv (pt
->output_cd
, NULL
, NULL
, NULL
, NULL
);
870 /* Print the OUTPUT_LEN bytes successfully converted. */
871 scm_lfwrite (encoded_output
, output_len
, port
);
873 /* See how many input codepoints these OUTPUT_LEN bytes
875 codepoints_read
= offsets
[input
- utf8_buf
] - printed
;
876 printed
+= codepoints_read
;
878 if (errno_save
== EILSEQ
&&
879 strategy
!= SCM_FAILED_CONVERSION_ERROR
)
881 /* Conversion failed somewhere in INPUT and we want to
882 escape or substitute the offending input character. */
884 if (strategy
== SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE
)
888 /* Find CH, the offending codepoint, and escape it. */
889 ch
= STR_REF (str
, offsets
[input
- utf8_buf
]);
890 write_character_escaped (ch
, 1, port
);
893 /* STRATEGY is `SCM_FAILED_CONVERSION_QUESTION_MARK'. */
894 display_string ("?", 1, 1, port
, strategy
);
899 /* Something bad happened that we can't handle: bail out. */
904 /* INPUT was successfully converted, entirely; print the
906 scm_lfwrite (encoded_output
, output_len
, port
);
907 codepoints_read
= i
- printed
;
908 printed
+= codepoints_read
;
916 /* Attempt to display CH to PORT according to STRATEGY. Return non-zero
917 if CH was successfully displayed, zero otherwise (e.g., if it was not
918 representable in PORT's encoding.) */
920 display_character (scm_t_wchar ch
, SCM port
,
921 scm_t_string_failed_conversion_handler strategy
)
923 return display_string (&ch
, 0, 1, port
, strategy
) == 1;
926 /* Attempt to pretty-print CH, a combining character, to PORT. Return
927 zero upon failure, non-zero otherwise. The idea is to print CH above
928 a dotted circle to make it more visible. */
930 write_combining_character (scm_t_wchar ch
, SCM port
)
934 str
[0] = SCM_CODEPOINT_DOTTED_CIRCLE
;
937 return display_string (str
, 0, 2, port
, iconveh_error
) == 2;
940 /* Write CH to PORT in its escaped form, using the string escape syntax
941 if STRING_ESCAPES_P is non-zero. */
943 write_character_escaped (scm_t_wchar ch
, int string_escapes_p
, SCM port
)
945 if (string_escapes_p
)
947 /* Represent CH using the in-string escape syntax. */
949 static const char hex
[] = "0123456789abcdef";
950 static const char escapes
[7] = "abtnvfr";
953 if (ch
>= 0x07 && ch
<= 0x0D && ch
!= 0x0A)
955 /* Use special escapes for some C0 controls. */
957 buf
[1] = escapes
[ch
- 0x07];
958 scm_lfwrite (buf
, 2, port
);
960 else if (!SCM_R6RS_ESCAPES_P
)
966 buf
[2] = hex
[ch
/ 16];
967 buf
[3] = hex
[ch
% 16];
968 scm_lfwrite (buf
, 4, port
);
970 else if (ch
<= 0xFFFF)
974 buf
[2] = hex
[(ch
& 0xF000) >> 12];
975 buf
[3] = hex
[(ch
& 0xF00) >> 8];
976 buf
[4] = hex
[(ch
& 0xF0) >> 4];
977 buf
[5] = hex
[(ch
& 0xF)];
978 scm_lfwrite (buf
, 6, port
);
980 else if (ch
> 0xFFFF)
984 buf
[2] = hex
[(ch
& 0xF00000) >> 20];
985 buf
[3] = hex
[(ch
& 0xF0000) >> 16];
986 buf
[4] = hex
[(ch
& 0xF000) >> 12];
987 buf
[5] = hex
[(ch
& 0xF00) >> 8];
988 buf
[6] = hex
[(ch
& 0xF0) >> 4];
989 buf
[7] = hex
[(ch
& 0xF)];
990 scm_lfwrite (buf
, 8, port
);
995 /* Print an R6RS variable-length hex escape: "\xNNNN;". */
996 scm_t_wchar ch2
= ch
;
1006 buf
[i
] = hex
[ch2
& 0xF];
1013 scm_lfwrite (buf
+ i
, 9 - i
, port
);
1018 /* Represent CH using the character escape syntax. */
1021 name
= scm_i_charname (SCM_MAKE_CHAR (ch
));
1023 scm_puts (name
, port
);
1025 PRINT_CHAR_ESCAPE (ch
, port
);
1029 /* Write CH to PORT, escaping it if it's non-graphic or not
1030 representable in PORT's encoding. If STRING_ESCAPES_P is true and CH
1031 needs to be escaped, it is escaped using the in-string escape syntax;
1032 otherwise the character escape syntax is used. */
1034 write_character (scm_t_wchar ch
, SCM port
, int string_escapes_p
)
1037 scm_t_string_failed_conversion_handler strategy
;
1039 strategy
= scm_i_get_conversion_strategy (port
);
1041 if (string_escapes_p
)
1043 /* Check if CH deserves special treatment. */
1044 if (ch
== '"' || ch
== '\\')
1046 display_character ('\\', port
, iconveh_question_mark
);
1047 display_character (ch
, port
, strategy
);
1050 else if (ch
== ' ' || ch
== '\n')
1052 display_character (ch
, port
, strategy
);
1058 display_string ("#\\", 1, 2, port
, iconveh_question_mark
);
1060 if (uc_combining_class (ch
) != UC_CCC_NR
)
1061 /* Character is a combining character, so attempt to
1063 printed
= write_combining_character (ch
, port
);
1067 && uc_is_general_category_withtable (ch
,
1068 UC_CATEGORY_MASK_L
|
1069 UC_CATEGORY_MASK_M
|
1070 UC_CATEGORY_MASK_N
|
1071 UC_CATEGORY_MASK_P
|
1072 UC_CATEGORY_MASK_S
))
1073 /* CH is graphic; attempt to display it. */
1074 printed
= display_character (ch
, port
, iconveh_error
);
1077 /* CH isn't graphic or cannot be represented in PORT's encoding. */
1078 write_character_escaped (ch
, string_escapes_p
, port
);
1081 /* Print an integer.
1085 scm_intprint (scm_t_intmax n
, int radix
, SCM port
)
1087 char num_buf
[SCM_INTBUFLEN
];
1088 scm_lfwrite (num_buf
, scm_iint2str (n
, radix
, num_buf
), port
);
1092 scm_uintprint (scm_t_uintmax n
, int radix
, SCM port
)
1094 char num_buf
[SCM_INTBUFLEN
];
1095 scm_lfwrite (num_buf
, scm_iuint2str (n
, radix
, num_buf
), port
);
1098 /* Print an object of unrecognized type.
1102 scm_ipruk (char *hdr
, SCM ptr
, SCM port
)
1104 scm_puts ("#<unknown-", port
);
1105 scm_puts (hdr
, port
);
1106 if (1) /* (scm_in_heap_p (ptr)) */ /* FIXME */
1108 scm_puts (" (0x", port
);
1109 scm_uintprint (SCM_CELL_WORD_0 (ptr
), 16, port
);
1110 scm_puts (" . 0x", port
);
1111 scm_uintprint (SCM_CELL_WORD_1 (ptr
), 16, port
);
1112 scm_puts (") @", port
);
1114 scm_puts (" 0x", port
);
1115 scm_uintprint (SCM_UNPACK (ptr
), 16, port
);
1116 scm_putc ('>', port
);
1123 scm_iprlist (char *hdr
, SCM exp
, int tlr
, SCM port
, scm_print_state
*pstate
)
1125 register SCM hare
, tortoise
;
1126 long floor
= pstate
->top
- 2;
1127 scm_puts (hdr
, port
);
1130 goto fancy_printing
;
1132 /* Run a hare and tortoise so that total time complexity will be
1133 O(depth * N) instead of O(N^2). */
1134 hare
= SCM_CDR (exp
);
1136 while (scm_is_pair (hare
))
1138 if (scm_is_eq (hare
, tortoise
))
1139 goto fancy_printing
;
1140 hare
= SCM_CDR (hare
);
1141 if (!scm_is_pair (hare
))
1143 hare
= SCM_CDR (hare
);
1144 tortoise
= SCM_CDR (tortoise
);
1147 /* No cdr cycles intrinsic to this list */
1148 scm_iprin1 (SCM_CAR (exp
), port
, pstate
);
1149 for (exp
= SCM_CDR (exp
); scm_is_pair (exp
); exp
= SCM_CDR (exp
))
1153 for (i
= floor
; i
>= 0; --i
)
1154 if (scm_is_eq (PSTATE_STACK_REF(pstate
, i
), exp
))
1156 PUSH_REF (pstate
, exp
);
1157 scm_putc (' ', port
);
1159 scm_iprin1 (SCM_CAR (exp
), port
, pstate
);
1161 if (!SCM_NULL_OR_NIL_P (exp
))
1163 scm_puts (" . ", port
);
1164 scm_iprin1 (exp
, port
, pstate
);
1168 scm_putc (tlr
, port
);
1169 pstate
->top
= floor
+ 2;
1174 long n
= pstate
->length
;
1176 scm_iprin1 (SCM_CAR (exp
), port
, pstate
);
1177 exp
= SCM_CDR (exp
); --n
;
1178 for (; scm_is_pair (exp
); exp
= SCM_CDR (exp
))
1180 register unsigned long i
;
1182 for (i
= 0; i
< pstate
->top
; ++i
)
1183 if (scm_is_eq (PSTATE_STACK_REF(pstate
, i
), exp
))
1189 scm_puts (" ...", port
);
1195 PUSH_REF(pstate
, exp
);
1196 ++pstate
->list_offset
;
1197 scm_putc (' ', port
);
1199 scm_iprin1 (SCM_CAR (exp
), port
, pstate
);
1202 if (!SCM_NULL_OR_NIL_P (exp
))
1204 scm_puts (" . ", port
);
1205 scm_iprin1 (exp
, port
, pstate
);
1208 pstate
->list_offset
-= pstate
->top
- floor
- 2;
1212 pstate
->list_offset
-= pstate
->top
- floor
- 2;
1215 scm_puts (" . ", port
);
1216 print_circref (port
, pstate
, exp
);
1223 scm_valid_oport_value_p (SCM val
)
1225 return (SCM_OPOUTPORTP (val
)
1226 || (SCM_PORT_WITH_PS_P (val
)
1227 && SCM_OPOUTPORTP (SCM_PORT_WITH_PS_PORT (val
))));
1230 /* SCM_GPROC(s_write, "write", 1, 1, 0, scm_write, g_write); */
1233 scm_write (SCM obj
, SCM port
)
1235 if (SCM_UNBNDP (port
))
1236 port
= scm_current_output_port ();
1238 SCM_ASSERT (scm_valid_oport_value_p (port
), port
, SCM_ARG2
, s_write
);
1240 scm_prin1 (obj
, port
, 1);
1241 return SCM_UNSPECIFIED
;
1245 /* SCM_GPROC(s_display, "display", 1, 1, 0, scm_display, g_display); */
1248 scm_display (SCM obj
, SCM port
)
1250 if (SCM_UNBNDP (port
))
1251 port
= scm_current_output_port ();
1253 SCM_ASSERT (scm_valid_oport_value_p (port
), port
, SCM_ARG2
, s_display
);
1255 scm_prin1 (obj
, port
, 0);
1256 return SCM_UNSPECIFIED
;
1260 SCM_DEFINE (scm_simple_format
, "simple-format", 2, 0, 1,
1261 (SCM destination
, SCM message
, SCM args
),
1262 "Write @var{message} to @var{destination}, defaulting to\n"
1263 "the current output port.\n"
1264 "@var{message} can contain @code{~A} (was @code{%s}) and\n"
1265 "@code{~S} (was @code{%S}) escapes. When printed,\n"
1266 "the escapes are replaced with corresponding members of\n"
1268 "@code{~A} formats using @code{display} and @code{~S} formats\n"
1269 "using @code{write}.\n"
1270 "If @var{destination} is @code{#t}, then use the current output\n"
1271 "port, if @var{destination} is @code{#f}, then return a string\n"
1272 "containing the formatted text. Does not add a trailing newline.")
1273 #define FUNC_NAME s_scm_simple_format
1275 SCM port
, answer
= SCM_UNSPECIFIED
;
1276 int fReturnString
= 0;
1278 size_t start
, p
, end
;
1280 if (scm_is_eq (destination
, SCM_BOOL_T
))
1282 destination
= port
= scm_current_output_port ();
1284 else if (scm_is_false (destination
))
1287 port
= scm_mkstrport (SCM_INUM0
, SCM_BOOL_F
,
1288 SCM_OPN
| SCM_WRTNG
,
1294 SCM_VALIDATE_OPORT_VALUE (1, destination
);
1295 port
= SCM_COERCE_OUTPORT (destination
);
1297 SCM_VALIDATE_STRING (2, message
);
1298 SCM_VALIDATE_REST_ARGUMENT (args
);
1302 end
= scm_i_string_length (message
);
1303 for (p
= start
; p
!= end
; ++p
)
1304 if (scm_i_string_ref (message
, p
) == '~')
1309 switch (scm_i_string_ref (message
, p
))
1318 scm_lfwrite_substr (message
, start
, p
, port
);
1322 scm_lfwrite_substr (message
, start
, p
- 1, port
);
1327 SCM_MISC_ERROR ("FORMAT: Unsupported format option ~~~A - use (ice-9 format) instead",
1328 scm_list_1 (SCM_MAKE_CHAR (scm_i_string_ref (message
, p
))));
1333 if (!scm_is_pair (args
))
1334 SCM_MISC_ERROR ("FORMAT: Missing argument for ~~~A",
1335 scm_list_1 (SCM_MAKE_CHAR (scm_i_string_ref (message
, p
))));
1337 scm_lfwrite_substr (message
, start
, p
- 1, port
);
1338 /* we pass destination here */
1339 scm_prin1 (SCM_CAR (args
), destination
, writingp
);
1340 args
= SCM_CDR (args
);
1344 scm_lfwrite_substr (message
, start
, p
, port
);
1345 if (!scm_is_eq (args
, SCM_EOL
))
1346 SCM_MISC_ERROR ("FORMAT: ~A superfluous arguments",
1347 scm_list_1 (scm_length (args
)));
1350 answer
= scm_strport_to_string (destination
);
1352 return scm_return_first (answer
, message
);
1357 SCM_DEFINE (scm_newline
, "newline", 0, 1, 0,
1359 "Send a newline to @var{port}.\n"
1360 "If @var{port} is omitted, send to the current output port.")
1361 #define FUNC_NAME s_scm_newline
1363 if (SCM_UNBNDP (port
))
1364 port
= scm_current_output_port ();
1366 SCM_VALIDATE_OPORT_VALUE (1, port
);
1368 scm_putc ('\n', SCM_COERCE_OUTPORT (port
));
1369 return SCM_UNSPECIFIED
;
1373 SCM_DEFINE (scm_write_char
, "write-char", 1, 1, 0,
1374 (SCM chr
, SCM port
),
1375 "Send character @var{chr} to @var{port}.")
1376 #define FUNC_NAME s_scm_write_char
1378 if (SCM_UNBNDP (port
))
1379 port
= scm_current_output_port ();
1381 SCM_VALIDATE_CHAR (1, chr
);
1382 SCM_VALIDATE_OPORT_VALUE (2, port
);
1384 port
= SCM_COERCE_OUTPORT (port
);
1385 if (!display_character (SCM_CHAR (chr
), port
,
1386 scm_i_get_conversion_strategy (port
)))
1387 scm_encoding_error (__func__
, errno
,
1388 "cannot convert to output locale",
1391 return SCM_UNSPECIFIED
;
1397 /* Call back to Scheme code to do the printing of special objects
1398 * (like structs). SCM_PRINTER_APPLY applies PROC to EXP and a smob
1399 * containing PORT and PSTATE. This object can be used as the port for
1400 * display/write etc to continue the current print chain. The REVEALED
1401 * field of PSTATE is set to true to indicate that the print state has
1402 * escaped to Scheme and thus has to be freed by the GC.
1405 scm_t_bits scm_tc16_port_with_ps
;
1407 /* Print exactly as the port itself would */
1410 port_with_ps_print (SCM obj
, SCM port
, scm_print_state
*pstate
)
1412 obj
= SCM_PORT_WITH_PS_PORT (obj
);
1413 return scm_ptobs
[SCM_PTOBNUM (obj
)].print (obj
, port
, pstate
);
1417 scm_printer_apply (SCM proc
, SCM exp
, SCM port
, scm_print_state
*pstate
)
1419 pstate
->revealed
= 1;
1420 return scm_call_2 (proc
, exp
,
1421 scm_i_port_with_print_state (port
, pstate
->handle
));
1424 SCM_DEFINE (scm_port_with_print_state
, "port-with-print-state", 1, 1, 0,
1425 (SCM port
, SCM pstate
),
1426 "Create a new port which behaves like @var{port}, but with an\n"
1427 "included print state @var{pstate}. @var{pstate} is optional.\n"
1428 "If @var{pstate} isn't supplied and @var{port} already has\n"
1429 "a print state, the old print state is reused.")
1430 #define FUNC_NAME s_scm_port_with_print_state
1432 SCM_VALIDATE_OPORT_VALUE (1, port
);
1433 if (!SCM_UNBNDP (pstate
))
1434 SCM_VALIDATE_PRINTSTATE (2, pstate
);
1435 return scm_i_port_with_print_state (port
, pstate
);
1439 SCM_DEFINE (scm_get_print_state
, "get-print-state", 1, 0, 0,
1441 "Return the print state of the port @var{port}. If @var{port}\n"
1442 "has no associated print state, @code{#f} is returned.")
1443 #define FUNC_NAME s_scm_get_print_state
1445 if (SCM_PORT_WITH_PS_P (port
))
1446 return SCM_PORT_WITH_PS_PS (port
);
1447 if (SCM_OUTPUT_PORT_P (port
))
1449 SCM_WRONG_TYPE_ARG (1, port
);
1458 SCM vtable
, layout
, type
;
1460 scm_init_opts (scm_print_options
, scm_print_opts
);
1462 scm_print_options (scm_list_4 (scm_from_latin1_symbol ("highlight-prefix"),
1463 scm_from_locale_string ("{"),
1464 scm_from_latin1_symbol ("highlight-suffix"),
1465 scm_from_locale_string ("}")));
1467 scm_gc_register_root (&print_state_pool
);
1468 scm_gc_register_root (&scm_print_state_vtable
);
1469 vtable
= scm_make_vtable_vtable (scm_nullstr
, SCM_INUM0
, SCM_EOL
);
1471 scm_make_struct_layout (scm_from_locale_string (SCM_PRINT_STATE_LAYOUT
));
1472 type
= scm_make_struct (vtable
, SCM_INUM0
, scm_list_1 (layout
));
1473 scm_set_struct_vtable_name_x (type
, scm_from_latin1_symbol ("print-state"));
1474 scm_print_state_vtable
= type
;
1476 /* Don't want to bind a wrapper class in GOOPS, so pass 0 as arg1. */
1477 scm_tc16_port_with_ps
= scm_make_smob_type (0, 0);
1478 scm_set_smob_print (scm_tc16_port_with_ps
, port_with_ps_print
);
1480 #include "libguile/print.x"
1482 scm_print_opts
[SCM_PRINT_KEYWORD_STYLE_I
].val
= SCM_UNPACK (sym_reader
);