[ViewVC] Diff of: cvs/JSON-XS/XS.xs

Comparing JSON-XS/XS.xs (file contents):
Revision 1.11 by root, Sat Mar 24 19:42:14 2007 UTC vs.
Revision 1.27 by root, Mon Apr 9 05:09:57 2007 UTC

…		…
3	#include "XSUB.h"	3	#include "XSUB.h"
4		4
5	#include "assert.h"	5	#include "assert.h"
6	#include "string.h"	6	#include "string.h"
7	#include "stdlib.h"	7	#include "stdlib.h"
		8	#include "stdio.h"
8		9
		10	#if defined(__BORLANDC__) \|\| defined(_MSC_VER)
		11	# define snprintf _snprintf // C compilers have this in stdio.h
		12	#endif
		13
9	#define F_ASCII 0x00000001	14	#define F_ASCII 0x00000001UL
10	#define F_UTF8 0x00000002	15	#define F_UTF8 0x00000002UL
11	#define F_INDENT 0x00000004	16	#define F_INDENT 0x00000004UL
12	#define F_CANONICAL 0x00000008	17	#define F_CANONICAL 0x00000008UL
13	#define F_SPACE_BEFORE 0x00000010	18	#define F_SPACE_BEFORE 0x00000010UL
14	#define F_SPACE_AFTER 0x00000020	19	#define F_SPACE_AFTER 0x00000020UL
15	#define F_ALLOW_NONREF 0x00000080	20	#define F_ALLOW_NONREF 0x00000080UL
16	#define F_SHRINK 0x00000100	21	#define F_SHRINK 0x00000100UL
		22	#define F_MAXDEPTH 0xf8000000UL
		23	#define S_MAXDEPTH 27
		24
		25	#define DEC_DEPTH(flags) (1UL << ((flags & F_MAXDEPTH) >> S_MAXDEPTH))
		26
		27	// F_SELFCONVERT? <=> to_json/toJson
		28	// F_BLESSED? <=> { $__class__$ => }
17		29
18	#define F_PRETTY F_INDENT \| F_SPACE_BEFORE \| F_SPACE_AFTER	30	#define F_PRETTY F_INDENT \| F_SPACE_BEFORE \| F_SPACE_AFTER
19	#define F_DEFAULT 0	31	#define F_DEFAULT (9UL << S_MAXDEPTH)
20		32
21	#define INIT_SIZE 32 // initial scalar size to be allocated	33	#define INIT_SIZE 32 // initial scalar size to be allocated
		34	#define INDENT_STEP 3 // spaces per indentation level
		35
		36	#define UTF8_MAX_LEN 11 // for perls UTF-X: max. number of octets per character
		37	#define SHORT_STRING_LEN 512 // special-case strings of up to this size
22		38
23	#define SB do {	39	#define SB do {
24	#define SE } while (0)	40	#define SE } while (0)
25		41
26	static HV *json_stash;	42	static HV *json_stash; // JSON::XS::
		43
		44	/////////////////////////////////////////////////////////////////////////////
		45	// utility functions
		46
		47	static UV *
		48	SvJSON (SV *sv)
		49	{
		50	if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
		51	croak ("object is not of type JSON::XS");
		52
		53	return &SvUVX (SvRV (sv));
		54	}
		55
		56	static void
		57	shrink (SV *sv)
		58	{
		59	sv_utf8_downgrade (sv, 1);
		60	if (SvLEN (sv) > SvCUR (sv) + 1)
		61	{
		62	#ifdef SvPV_shrink_to_cur
		63	SvPV_shrink_to_cur (sv);
		64	#elif defined (SvPV_renew)
		65	SvPV_renew (sv, SvCUR (sv) + 1);
		66	#endif
		67	}
		68	}
		69
		70	// decode an utf-8 character and return it, or (UV)-1 in
		71	// case of an error.
		72	// we special-case "safe" characters from U+80 .. U+7FF,
		73	// but use the very good perl function to parse anything else.
		74	// note that we never call this function for a ascii codepoints
		75	static UV
		76	decode_utf8 (unsigned char s, STRLEN len, STRLEN clen)
		77	{
		78	if (s[0] > 0xdf \|\| s[0] < 0xc2)
		79	return utf8n_to_uvuni (s, len, clen, UTF8_CHECK_ONLY);
		80	else if (len > 1 && s[1] >= 0x80 && s[1] <= 0xbf)
		81	{
		82	*clen = 2;
		83	return ((s[0] & 0x1f) << 6) \| (s[1] & 0x3f);
		84	}
		85	else
		86	{
		87	*clen = (STRLEN)-1;
		88	return (UV)-1;
		89	}
		90	}
		91
		92	/////////////////////////////////////////////////////////////////////////////
		93	// encoder
27		94
28	// structure used for encoding JSON	95	// structure used for encoding JSON
29	typedef struct	96	typedef struct
30	{	97	{
31	char *cur;	98	char *cur; // SvPVX (sv) + current output position
32	STRLEN len; // SvLEN (sv)
33	char *end; // SvEND (sv)	99	char *end; // SvEND (sv)
34	SV *sv;	100	SV *sv; // result scalar
35	UV flags;	101	U32 flags; // F_*
36	int max_recurse;	102	U32 indent; // indentation level
37	int indent;	103	U32 maxdepth; // max. indentation/recursion level
38	} enc_t;	104	} enc_t;
39
40	// structure used for decoding JSON
41	typedef struct
42	{
43	char *cur;
44	char *end;
45	const char *err;
46	UV flags;
47	} dec_t;
48
49	static UV *
50	SvJSON (SV *sv)
51	{
52	if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
53	croak ("object is not of type JSON::XS");
54
55	return &SvUVX (SvRV (sv));
56	}
57
58	static void
59	shrink (SV *sv)
60	{
61	sv_utf8_downgrade (sv, 1);
62	#ifdef SvPV_shrink_to_cur
63	SvPV_shrink_to_cur (sv);
64	#endif
65	}
66
67	/////////////////////////////////////////////////////////////////////////////
68		105
69	static void	106	static void
70	need (enc_t *enc, STRLEN len)	107	need (enc_t *enc, STRLEN len)
71	{	108	{
72	if (enc->cur + len >= enc->end)	109	if (enc->cur + len >= enc->end)
73	{	110	{
74	STRLEN cur = enc->cur - SvPVX (enc->sv);	111	STRLEN cur = enc->cur - SvPVX (enc->sv);
75	SvGROW (enc->sv, cur + len + 1);	112	SvGROW (enc->sv, cur + len + 1);
76	enc->cur = SvPVX (enc->sv) + cur;	113	enc->cur = SvPVX (enc->sv) + cur;
77	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv);	114	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
78	}	115	}
79	}	116	}
80		117
81	static void	118	static void
82	encode_ch (enc_t *enc, char ch)	119	encode_ch (enc_t *enc, char ch)
…		…
130	STRLEN clen;	167	STRLEN clen;
131	UV uch;	168	UV uch;
132		169
133	if (is_utf8)	170	if (is_utf8)
134	{	171	{
135	uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY);	172	//uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY);
		173	uch = decode_utf8 (str, end - str, &clen);
136	if (clen == (STRLEN)-1)	174	if (clen == (STRLEN)-1)
137	croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);	175	croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
138	}	176	}
139	else	177	else
140	{	178	{
…		…
178	}	216	}
179	while (--clen);	217	while (--clen);
180	}	218	}
181	else	219	else
182	{	220	{
183	need (enc, len += 10); // never more than 11 bytes needed	221	need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed
184	enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);	222	enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
185	++str;	223	++str;
186	}	224	}
187	}	225	}
188	}	226	}
…		…
190		228
191	--len;	229	--len;
192	}	230	}
193	}	231	}
194		232
195	#define INDENT SB \	233	static void
		234	encode_indent (enc_t *enc)
		235	{
196	if (enc->flags & F_INDENT) \	236	if (enc->flags & F_INDENT)
197	{ \	237	{
198	int i_; \	238	int spaces = enc->indent * INDENT_STEP;
199	need (enc, enc->indent); \
200	for (i_ = enc->indent * 3; i_--; )\
201	encode_ch (enc, ' '); \
202	} \
203	SE
204		239
205	#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE	240	need (enc, spaces);
206	#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE	241	memset (enc->cur, ' ', spaces);
207	#define COMMA SB \	242	enc->cur += spaces;
		243	}
		244	}
		245
		246	static void
		247	encode_space (enc_t *enc)
		248	{
		249	need (enc, 1);
208	encode_ch (enc, ','); \	250	encode_ch (enc, ' ');
		251	}
		252
		253	static void
		254	encode_nl (enc_t *enc)
		255	{
209	if (enc->flags & F_INDENT) \	256	if (enc->flags & F_INDENT)
210	NL; \	257	{
		258	need (enc, 1);
		259	encode_ch (enc, '\n');
		260	}
		261	}
		262
		263	static void
		264	encode_comma (enc_t *enc)
		265	{
		266	encode_ch (enc, ',');
		267
		268	if (enc->flags & F_INDENT)
		269	encode_nl (enc);
211	else if (enc->flags & F_SPACE_AFTER) \	270	else if (enc->flags & F_SPACE_AFTER)
212	SPACE; \	271	encode_space (enc);
213	SE	272	}
214		273
215	static void encode_sv (enc_t enc, SV sv);	274	static void encode_sv (enc_t enc, SV sv);
216		275
217	static void	276	static void
218	encode_av (enc_t enc, AV av)	277	encode_av (enc_t enc, AV av)
219	{	278	{
220	int i, len = av_len (av);	279	int i, len = av_len (av);
221		280
222	encode_ch (enc, '['); NL;	281	if (enc->indent >= enc->maxdepth)
		282	croak ("data structure too deep (hit recursion limit)");
		283
		284	encode_ch (enc, '['); encode_nl (enc);
223	++enc->indent;	285	++enc->indent;
224		286
225	for (i = 0; i <= len; ++i)	287	for (i = 0; i <= len; ++i)
226	{	288	{
227	INDENT;	289	encode_indent (enc);
228	encode_sv (enc, *av_fetch (av, i, 0));	290	encode_sv (enc, *av_fetch (av, i, 0));
229		291
230	if (i < len)	292	if (i < len)
231	COMMA;	293	encode_comma (enc);
232	}	294	}
233		295
234	NL;	296	encode_nl (enc);
235		297
236	--enc->indent;	298	--enc->indent;
237	INDENT; encode_ch (enc, ']');	299	encode_indent (enc); encode_ch (enc, ']');
238	}	300	}
239		301
240	static void	302	static void
241	encode_he (enc_t enc, HE he)	303	encode_he (enc_t enc, HE he)
242	{	304	{
…		…
256	else	318	else
257	encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));	319	encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
258		320
259	encode_ch (enc, '"');	321	encode_ch (enc, '"');
260		322
261	if (enc->flags & F_SPACE_BEFORE) SPACE;	323	if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
262	encode_ch (enc, ':');	324	encode_ch (enc, ':');
263	if (enc->flags & F_SPACE_AFTER ) SPACE;	325	if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
264	encode_sv (enc, HeVAL (he));	326	encode_sv (enc, HeVAL (he));
265	}	327	}
266		328
267	// compare hash entries, used when all keys are bytestrings	329	// compare hash entries, used when all keys are bytestrings
268	static int	330	static int
…		…
292	static void	354	static void
293	encode_hv (enc_t enc, HV hv)	355	encode_hv (enc_t enc, HV hv)
294	{	356	{
295	int count, i;	357	int count, i;
296		358
		359	if (enc->indent >= enc->maxdepth)
		360	croak ("data structure too deep (hit recursion limit)");
		361
297	encode_ch (enc, '{'); NL; ++enc->indent;	362	encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
298		363
299	if ((count = hv_iterinit (hv)))	364	if ((count = hv_iterinit (hv)))
300	{	365	{
301	// for canonical output we have to sort by keys first	366	// for canonical output we have to sort by keys first
302	// actually, this is mostly due to the stupid so-called	367	// actually, this is mostly due to the stupid so-called
303	// security workaround added somewhere in 5.8.x.	368	// security workaround added somewhere in 5.8.x.
304	// that randomises hash orderings	369	// that randomises hash orderings
305	if (enc->flags & F_CANONICAL)	370	if (enc->flags & F_CANONICAL)
306	{	371	{
307	HE he, hes [count];	372	HE he, hes [count]; // if your compiler dies here, you need to enable C99 mode
308	int fast = 1;	373	int fast = 1;
309		374
310	i = 0;	375	i = 0;
311	while ((he = hv_iternext (hv)))	376	while ((he = hv_iternext (hv)))
312	{	377	{
…		…
337	LEAVE;	402	LEAVE;
338	}	403	}
339		404
340	for (i = 0; i < count; ++i)	405	for (i = 0; i < count; ++i)
341	{	406	{
342	INDENT;	407	encode_indent (enc);
343	encode_he (enc, hes [i]);	408	encode_he (enc, hes [i]);
344		409
345	if (i < count - 1)	410	if (i < count - 1)
346	COMMA;	411	encode_comma (enc);
347	}	412	}
348		413
349	NL;	414	encode_nl (enc);
350	}	415	}
351	else	416	else
352	{	417	{
353	SV *sv;
354	HE *he = hv_iternext (hv);	418	HE *he = hv_iternext (hv);
355		419
356	for (;;)	420	for (;;)
357	{	421	{
358	INDENT;	422	encode_indent (enc);
359	encode_he (enc, he);	423	encode_he (enc, he);
360		424
361	if (!(he = hv_iternext (hv)))	425	if (!(he = hv_iternext (hv)))
362	break;	426	break;
363		427
364	COMMA;	428	encode_comma (enc);
365	}	429	}
366		430
367	NL;	431	encode_nl (enc);
368	}	432	}
369	}	433	}
370		434
371	--enc->indent; INDENT; encode_ch (enc, '}');	435	--enc->indent; encode_indent (enc); encode_ch (enc, '}');
		436	}
		437
		438	// encode objects, arrays and special \0=false and \1=true values.
		439	static void
		440	encode_rv (enc_t enc, SV sv)
		441	{
		442	svtype svt;
		443
		444	SvGETMAGIC (sv);
		445	svt = SvTYPE (sv);
		446
		447	if (svt == SVt_PVHV)
		448	encode_hv (enc, (HV *)sv);
		449	else if (svt == SVt_PVAV)
		450	encode_av (enc, (AV *)sv);
		451	else if (svt < SVt_PVAV)
		452	{
		453	if (SvNIOK (sv) && SvIV (sv) == 0)
		454	encode_str (enc, "false", 5, 0);
		455	else if (SvNIOK (sv) && SvIV (sv) == 1)
		456	encode_str (enc, "true", 4, 0);
		457	else
		458	croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
		459	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
		460	}
		461	else
		462	croak ("encountered %s, but JSON can only represent references to arrays or hashes",
		463	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
372	}	464	}
373		465
374	static void	466	static void
375	encode_sv (enc_t enc, SV sv)	467	encode_sv (enc_t enc, SV sv)
376	{	468	{
…		…
397	SvIsUV(sv)	489	SvIsUV(sv)
398	? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))	490	? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))
399	: snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));	491	: snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));
400	}	492	}
401	else if (SvROK (sv))	493	else if (SvROK (sv))
402	{	494	encode_rv (enc, SvRV (sv));
403	SV *rv = SvRV (sv);
404
405	if (!--enc->max_recurse)
406	croak ("data structure too deep (hit recursion limit)");
407
408	switch (SvTYPE (rv))
409	{
410	case SVt_PVAV: encode_av (enc, (AV *)rv); break;
411	case SVt_PVHV: encode_hv (enc, (HV *)rv); break;
412
413	default:
414	croak ("encountered %s, but JSON can only represent references to arrays or hashes",
415	SvPV_nolen (sv));
416	}
417	}
418	else if (!SvOK (sv))	495	else if (!SvOK (sv))
419	encode_str (enc, "null", 4, 0);	496	encode_str (enc, "null", 4, 0);
420	else	497	else
421	croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this",	498	croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this",
422	SvPV_nolen (sv), SvFLAGS (sv));	499	SvPV_nolen (sv), SvFLAGS (sv));
423	}	500	}
424		501
425	static SV *	502	static SV *
426	encode_json (SV *scalar, UV flags)	503	encode_json (SV *scalar, U32 flags)
427	{	504	{
		505	enc_t enc;
		506
428	if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))	507	if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
429	croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)");	508	croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)");
430		509
431	enc_t enc;
432	enc.flags = flags;	510	enc.flags = flags;
433	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	511	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
434	enc.cur = SvPVX (enc.sv);	512	enc.cur = SvPVX (enc.sv);
435	enc.end = SvEND (enc.sv);	513	enc.end = SvEND (enc.sv);
436	enc.max_recurse = 0;
437	enc.indent = 0;	514	enc.indent = 0;
		515	enc.maxdepth = DEC_DEPTH (flags);
438		516
439	SvPOK_only (enc.sv);	517	SvPOK_only (enc.sv);
440	encode_sv (&enc, scalar);	518	encode_sv (&enc, scalar);
441		519
442	if (!(flags & (F_ASCII \| F_UTF8)))	520	if (!(flags & (F_ASCII \| F_UTF8)))
443	SvUTF8_on (enc.sv);	521	SvUTF8_on (enc.sv);
444		522
445	SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));	523	SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
		524	*SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
446		525
447	if (enc.flags & F_SHRINK)	526	if (enc.flags & F_SHRINK)
448	shrink (enc.sv);	527	shrink (enc.sv);
449		528
450	return enc.sv;	529	return enc.sv;
451	}	530	}
452		531
453	/////////////////////////////////////////////////////////////////////////////	532	/////////////////////////////////////////////////////////////////////////////
		533	// decoder
454		534
455	#define WS \	535	// structure used for decoding JSON
		536	typedef struct
		537	{
		538	char *cur; // current parser pointer
		539	char *end; // end of input string
		540	const char *err; // parse error, if != 0
		541	U32 flags; // F_*
		542	U32 depth; // recursion depth
		543	U32 maxdepth; // recursion depth limit
		544	} dec_t;
		545
		546	static void
		547	decode_ws (dec_t *dec)
		548	{
456	for (;;) \	549	for (;;)
457	{ \	550	{
458	char ch = *dec->cur; \	551	char ch = *dec->cur;
		552
459	if (ch > 0x20 \	553	if (ch > 0x20
460	\|\| (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \	554	\|\| (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
461	break; \	555	break;
		556
462	++dec->cur; \	557	++dec->cur;
463	}	558	}
		559	}
464		560
465	#define ERR(reason) SB dec->err = reason; goto fail; SE	561	#define ERR(reason) SB dec->err = reason; goto fail; SE
		562
466	#define EXPECT_CH(ch) SB \	563	#define EXPECT_CH(ch) SB \
467	if (*dec->cur != ch) \	564	if (*dec->cur != ch) \
468	ERR (# ch " expected"); \	565	ERR (# ch " expected"); \
469	++dec->cur; \	566	++dec->cur; \
470	SE	567	SE
471		568
		569	#define DEC_INC_DEPTH if (++dec->depth > dec->maxdepth) ERR ("json datastructure exceeds maximum nesting level (set a higher max_depth)")
		570	#define DEC_DEC_DEPTH --dec->depth
		571
472	static SV decode_sv (dec_t dec);	572	static SV decode_sv (dec_t dec);
473		573
474	static signed char decode_hexdigit[256];	574	static signed char decode_hexdigit[256];
475		575
476	static UV	576	static UV
477	decode_4hex (dec_t *dec)	577	decode_4hex (dec_t *dec)
478	{	578	{
479	signed char d1, d2, d3, d4;	579	signed char d1, d2, d3, d4;
		580	unsigned char cur = (unsigned char )dec->cur;
480		581
481	d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
482	if (d1 < 0) ERR ("four hexadecimal digits expected");	582	d1 = decode_hexdigit [cur [0]]; if (d1 < 0) ERR ("four hexadecimal digits expected");
483	d2 = decode_hexdigit [((unsigned char *)dec->cur) [1]];
484	if (d2 < 0) ERR ("four hexadecimal digits expected");	583	d2 = decode_hexdigit [cur [1]]; if (d2 < 0) ERR ("four hexadecimal digits expected");
485	d3 = decode_hexdigit [((unsigned char *)dec->cur) [2]];
486	if (d3 < 0) ERR ("four hexadecimal digits expected");	584	d3 = decode_hexdigit [cur [2]]; if (d3 < 0) ERR ("four hexadecimal digits expected");
487	d4 = decode_hexdigit [((unsigned char *)dec->cur) [3]];
488	if (d4 < 0) ERR ("four hexadecimal digits expected");	585	d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
489		586
490	dec->cur += 4;	587	dec->cur += 4;
491		588
492	return ((UV)d1) << 12	589	return ((UV)d1) << 12
493	\| ((UV)d2) << 8	590	\| ((UV)d2) << 8
…		…
496		593
497	fail:	594	fail:
498	return (UV)-1;	595	return (UV)-1;
499	}	596	}
500		597
501	#define APPEND_GROW(n) SB \
502	if (cur + (n) >= end) \
503	{ \
504	STRLEN ofs = cur - SvPVX (sv); \
505	SvGROW (sv, ofs + (n) + 1); \
506	cur = SvPVX (sv) + ofs; \
507	end = SvEND (sv); \
508	} \
509	SE
510
511	#define APPEND_CH(ch) SB \
512	APPEND_GROW (1); \
513	*cur++ = (ch); \
514	SE
515
516	static SV *	598	static SV *
517	decode_str (dec_t *dec)	599	decode_str (dec_t *dec)
518	{	600	{
519	SV *sv = NEWSV (0,2);	601	SV *sv = 0;
520	int utf8 = 0;	602	int utf8 = 0;
521	char *cur = SvPVX (sv);
522	char *end = SvEND (sv);
523		603
524	for (;;)	604	do
525	{	605	{
526	unsigned char ch = (unsigned char )dec->cur;	606	char buf [SHORT_STRING_LEN + UTF8_MAX_LEN];
		607	char *cur = buf;
527		608
528	if (ch == '"')	609	do
529	break;
530	else if (ch == '\\')
531	{	610	{
532	switch (*++dec->cur)	611	unsigned char ch = (unsigned char )dec->cur++;
		612
		613	if (ch == '"')
533	{	614	{
534	case '\\':	615	--dec->cur;
535	case '/':	616	break;
536	case '"': APPEND_CH (*dec->cur++); break;	617	}
537		618	else if (ch == '\\')
538	case 'b': APPEND_CH ('\010'); ++dec->cur; break;	619	{
539	case 't': APPEND_CH ('\011'); ++dec->cur; break;	620	switch (*dec->cur)
540	case 'n': APPEND_CH ('\012'); ++dec->cur; break;
541	case 'f': APPEND_CH ('\014'); ++dec->cur; break;
542	case 'r': APPEND_CH ('\015'); ++dec->cur; break;
543
544	case 'u':
545	{	621	{
546	UV lo, hi;	622	case '\\':
547	++dec->cur;	623	case '/':
		624	case '"': cur++ = dec->cur++; break;
548		625
549	hi = decode_4hex (dec);	626	case 'b': ++dec->cur; *cur++ = '\010'; break;
550	if (hi == (UV)-1)	627	case 't': ++dec->cur; *cur++ = '\011'; break;
551	goto fail;	628	case 'n': ++dec->cur; *cur++ = '\012'; break;
		629	case 'f': ++dec->cur; *cur++ = '\014'; break;
		630	case 'r': ++dec->cur; *cur++ = '\015'; break;
552		631
553	// possibly a surrogate pair	632	case 'u':
554	if (hi >= 0xd800 && hi < 0xdc00)
555	{	633	{
556	if (dec->cur [0] != '\\' \|\| dec->cur [1] != 'u')	634	UV lo, hi;
557	ERR ("missing low surrogate character in surrogate pair");
558
559	dec->cur += 2;	635	++dec->cur;
560		636
561	lo = decode_4hex (dec);	637	hi = decode_4hex (dec);
562	if (lo == (UV)-1)	638	if (hi == (UV)-1)
563	goto fail;	639	goto fail;
564		640
		641	// possibly a surrogate pair
		642	if (hi >= 0xd800)
		643	if (hi < 0xdc00)
		644	{
		645	if (dec->cur [0] != '\\' \|\| dec->cur [1] != 'u')
		646	ERR ("missing low surrogate character in surrogate pair");
		647
		648	dec->cur += 2;
		649
		650	lo = decode_4hex (dec);
		651	if (lo == (UV)-1)
		652	goto fail;
		653
565	if (lo < 0xdc00 \|\| lo >= 0xe000)	654	if (lo < 0xdc00 \|\| lo >= 0xe000)
566	ERR ("surrogate pair expected");	655	ERR ("surrogate pair expected");
567		656
568	hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;	657	hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;
		658	}
		659	else if (hi < 0xe000)
		660	ERR ("missing high surrogate character in surrogate pair");
		661
		662	if (hi >= 0x80)
		663	{
		664	utf8 = 1;
		665
		666	cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
		667	}
		668	else
		669	*cur++ = hi;
569	}	670	}
570	else if (hi >= 0xdc00 && hi < 0xe000)
571	ERR ("missing high surrogate character in surrogate pair");
572
573	if (hi >= 0x80)
574	{	671	break;
575	utf8 = 1;
576		672
577	APPEND_GROW (4); // at most 4 bytes for 21 bits
578	cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
579	}
580	else	673	default:
581	APPEND_CH (hi);	674	--dec->cur;
		675	ERR ("illegal backslash escape sequence in string");
582	}	676	}
583	break;	677	}
		678	else if (ch >= 0x20 && ch <= 0x7f)
		679	*cur++ = ch;
		680	else if (ch >= 0x80)
		681	{
		682	STRLEN clen;
		683	UV uch;
584		684
585	default:
586	--dec->cur;	685	--dec->cur;
587	ERR ("illegal backslash escape sequence in string");	686
		687	uch = decode_utf8 (dec->cur, dec->end - dec->cur, &clen);
		688	if (clen == (STRLEN)-1)
		689	ERR ("malformed UTF-8 character in JSON string");
		690
		691	do
		692	cur++ = dec->cur++;
		693	while (--clen);
		694
		695	utf8 = 1;
		696	}
		697	else
		698	{
		699	--dec->cur;
		700
		701	if (!ch)
		702	ERR ("unexpected end of string while parsing JSON string");
		703	else
		704	ERR ("invalid character encountered while parsing JSON string");
588	}	705	}
589	}	706	}
590	else if (ch >= 0x20 && ch <= 0x7f)	707	while (cur < buf + SHORT_STRING_LEN);
591	APPEND_CH (*dec->cur++);	708
592	else if (ch >= 0x80)
593	{	709	{
594	STRLEN clen;	710	STRLEN len = cur - buf;
595	UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY);
596	if (clen == (STRLEN)-1)
597	ERR ("malformed UTF-8 character in JSON string");
598		711
599	APPEND_GROW (clen);	712	if (sv)
600	do
601	{	713	{
602	cur++ = dec->cur++;	714	SvGROW (sv, SvCUR (sv) + len + 1);
		715	memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
		716	SvCUR_set (sv, SvCUR (sv) + len);
603	}	717	}
604	while (--clen);
605
606	utf8 = 1;
607	}
608	else if (dec->cur == dec->end)
609	ERR ("unexpected end of string while parsing json string");
610	else	718	else
611	ERR ("invalid character encountered");	719	sv = newSVpvn (buf, len);
612	}	720	}
		721	}
		722	while (*dec->cur != '"');
613		723
614	++dec->cur;	724	++dec->cur;
615		725
616	SvCUR_set (sv, cur - SvPVX (sv));	726	if (sv)
617		727	{
618	SvPOK_only (sv);	728	SvPOK_only (sv);
619	*SvEND (sv) = 0;	729	*SvEND (sv) = 0;
620		730
621	if (utf8)	731	if (utf8)
622	SvUTF8_on (sv);	732	SvUTF8_on (sv);
623		733	}
624	if (dec->flags & F_SHRINK)	734	else
625	shrink (sv);	735	sv = newSVpvn ("", 0);
626		736
627	return sv;	737	return sv;
628		738
629	fail:	739	fail:
630	SvREFCNT_dec (sv);
631	return 0;	740	return 0;
632	}	741	}
633		742
634	static SV *	743	static SV *
635	decode_num (dec_t *dec)	744	decode_num (dec_t *dec)
…		…
716	static SV *	825	static SV *
717	decode_av (dec_t *dec)	826	decode_av (dec_t *dec)
718	{	827	{
719	AV *av = newAV ();	828	AV *av = newAV ();
720		829
721	WS;	830	DEC_INC_DEPTH;
		831	decode_ws (dec);
		832
722	if (*dec->cur == ']')	833	if (*dec->cur == ']')
723	++dec->cur;	834	++dec->cur;
724	else	835	else
725	for (;;)	836	for (;;)
726	{	837	{
…		…
730	if (!value)	841	if (!value)
731	goto fail;	842	goto fail;
732		843
733	av_push (av, value);	844	av_push (av, value);
734		845
735	WS;	846	decode_ws (dec);
736		847
737	if (*dec->cur == ']')	848	if (*dec->cur == ']')
738	{	849	{
739	++dec->cur;	850	++dec->cur;
740	break;	851	break;
…		…
744	ERR (", or ] expected while parsing array");	855	ERR (", or ] expected while parsing array");
745		856
746	++dec->cur;	857	++dec->cur;
747	}	858	}
748		859
		860	DEC_DEC_DEPTH;
749	return newRV_noinc ((SV *)av);	861	return newRV_noinc ((SV *)av);
750		862
751	fail:	863	fail:
752	SvREFCNT_dec (av);	864	SvREFCNT_dec (av);
		865	DEC_DEC_DEPTH;
753	return 0;	866	return 0;
754	}	867	}
755		868
756	static SV *	869	static SV *
757	decode_hv (dec_t *dec)	870	decode_hv (dec_t *dec)
758	{	871	{
759	HV *hv = newHV ();	872	HV *hv = newHV ();
760		873
761	WS;	874	DEC_INC_DEPTH;
		875	decode_ws (dec);
		876
762	if (*dec->cur == '}')	877	if (*dec->cur == '}')
763	++dec->cur;	878	++dec->cur;
764	else	879	else
765	for (;;)	880	for (;;)
766	{	881	{
767	SV key, value;	882	SV key, value;
768		883
769	WS; EXPECT_CH ('"');	884	decode_ws (dec); EXPECT_CH ('"');
770		885
771	key = decode_str (dec);	886	key = decode_str (dec);
772	if (!key)	887	if (!key)
773	goto fail;	888	goto fail;
774		889
775	WS; EXPECT_CH (':');	890	decode_ws (dec); EXPECT_CH (':');
776		891
777	value = decode_sv (dec);	892	value = decode_sv (dec);
778	if (!value)	893	if (!value)
779	{	894	{
780	SvREFCNT_dec (key);	895	SvREFCNT_dec (key);
781	goto fail;	896	goto fail;
782	}	897	}
783		898
784	//TODO: optimise
785	hv_store_ent (hv, key, value, 0);	899	hv_store_ent (hv, key, value, 0);
		900	SvREFCNT_dec (key);
786		901
787	WS;	902	decode_ws (dec);
788		903
789	if (*dec->cur == '}')	904	if (*dec->cur == '}')
790	{	905	{
791	++dec->cur;	906	++dec->cur;
792	break;	907	break;
…		…
796	ERR (", or } expected while parsing object/hash");	911	ERR (", or } expected while parsing object/hash");
797		912
798	++dec->cur;	913	++dec->cur;
799	}	914	}
800		915
		916	DEC_DEC_DEPTH;
801	return newRV_noinc ((SV *)hv);	917	return newRV_noinc ((SV *)hv);
802		918
803	fail:	919	fail:
804	SvREFCNT_dec (hv);	920	SvREFCNT_dec (hv);
		921	DEC_DEC_DEPTH;
805	return 0;	922	return 0;
806	}	923	}
807		924
808	static SV *	925	static SV *
809	decode_sv (dec_t *dec)	926	decode_sv (dec_t *dec)
810	{	927	{
811	WS;	928	decode_ws (dec);
812	switch (*dec->cur)	929	switch (*dec->cur)
813	{	930	{
814	case '"': ++dec->cur; return decode_str (dec);	931	case '"': ++dec->cur; return decode_str (dec);
815	case '[': ++dec->cur; return decode_av (dec);	932	case '[': ++dec->cur; return decode_av (dec);
816	case '{': ++dec->cur; return decode_hv (dec);	933	case '{': ++dec->cur; return decode_hv (dec);
…		…
852	ERR ("'null' expected");	969	ERR ("'null' expected");
853		970
854	break;	971	break;
855		972
856	default:	973	default:
857	ERR ("malformed json string, neither array, object, number, string or atom");	974	ERR ("malformed JSON string, neither array, object, number, string or atom");
858	break;	975	break;
859	}	976	}
860		977
861	fail:	978	fail:
862	return 0;	979	return 0;
863	}	980	}
864		981
865	static SV *	982	static SV *
866	decode_json (SV *string, UV flags)	983	decode_json (SV *string, U32 flags)
867	{	984	{
		985	dec_t dec;
868	SV *sv;	986	SV *sv;
		987
		988	SvUPGRADE (string, SVt_PV);
869		989
870	if (flags & F_UTF8)	990	if (flags & F_UTF8)
871	sv_utf8_downgrade (string, 0);	991	sv_utf8_downgrade (string, 0);
872	else	992	else
873	sv_utf8_upgrade (string);	993	sv_utf8_upgrade (string);
874		994
875	SvGROW (string, SvCUR (string) + 1); // should basically be a NOP	995	SvGROW (string, SvCUR (string) + 1); // should basically be a NOP
876		996
877	dec_t dec;
878	dec.flags = flags;	997	dec.flags = flags;
879	dec.cur = SvPVX (string);	998	dec.cur = SvPVX (string);
880	dec.end = SvEND (string);	999	dec.end = SvEND (string);
881	dec.err = 0;	1000	dec.err = 0;
		1001	dec.depth = 0;
		1002	dec.maxdepth = DEC_DEPTH (dec.flags);
882		1003
		1004	*dec.end = 0; // this should basically be a nop, too, but make sure its there
883	sv = decode_sv (&dec);	1005	sv = decode_sv (&dec);
884		1006
885	if (!sv)	1007	if (!sv)
886	{	1008	{
887	IV offset = dec.flags & F_UTF8	1009	IV offset = dec.flags & F_UTF8
…		…
896	SAVEVPTR (PL_curcop);	1018	SAVEVPTR (PL_curcop);
897	PL_curcop = &cop;	1019	PL_curcop = &cop;
898	pv_uni_display (uni, dec.cur, dec.end - dec.cur, 20, UNI_DISPLAY_QQ);	1020	pv_uni_display (uni, dec.cur, dec.end - dec.cur, 20, UNI_DISPLAY_QQ);
899	LEAVE;	1021	LEAVE;
900		1022
901	croak ("%s, at character offset %d (%s)",	1023	croak ("%s, at character offset %d [\"%s\"]",
902	dec.err,	1024	dec.err,
903	(int)offset,	1025	(int)offset,
904	dec.cur != dec.end ? SvPV_nolen (uni) : "(end of string)");	1026	dec.cur != dec.end ? SvPV_nolen (uni) : "(end of string)");
905	}	1027	}
906		1028
…		…
910	croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");	1032	croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");
911		1033
912	return sv;	1034	return sv;
913	}	1035	}
914		1036
		1037	/////////////////////////////////////////////////////////////////////////////
		1038	// XS interface functions
		1039
915	MODULE = JSON::XS PACKAGE = JSON::XS	1040	MODULE = JSON::XS PACKAGE = JSON::XS
916		1041
917	BOOT:	1042	BOOT:
918	{	1043	{
919	int i;	1044	int i;
920		1045
921	memset (decode_hexdigit, 0xff, 256);	1046	memset (decode_hexdigit, 0xff, 256);
		1047
922	for (i = 10; i--; )	1048	for (i = 0; i < 256; ++i)
923	decode_hexdigit ['0' + i] = i;	1049	decode_hexdigit [i] =
924		1050	i >= '0' && i <= '9' ? i - '0'
925	for (i = 7; i--; )	1051	: i >= 'a' && i <= 'f' ? i - 'a' + 10
926	{	1052	: i >= 'A' && i <= 'F' ? i - 'A' + 10
927	decode_hexdigit ['a' + i] = 10 + i;	1053	: -1;
928	decode_hexdigit ['A' + i] = 10 + i;
929	}
930		1054
931	json_stash = gv_stashpv ("JSON::XS", 1);	1055	json_stash = gv_stashpv ("JSON::XS", 1);
932	}	1056	}
933		1057
934	PROTOTYPES: DISABLE	1058	PROTOTYPES: DISABLE
…		…
961	RETVAL = newSVsv (self);	1085	RETVAL = newSVsv (self);
962	}	1086	}
963	OUTPUT:	1087	OUTPUT:
964	RETVAL	1088	RETVAL
965		1089
		1090	SV max_depth (SV self, UV max_depth = 0x80000000UL)
		1091	CODE:
		1092	{
		1093	UV *uv = SvJSON (self);
		1094	UV log2 = 0;
		1095
		1096	if (max_depth > 0x80000000UL) max_depth = 0x80000000UL;
		1097
		1098	while ((1UL << log2) < max_depth)
		1099	++log2;
		1100
		1101	uv = uv & ~F_MAXDEPTH \| (log2 << S_MAXDEPTH);
		1102
		1103	RETVAL = newSVsv (self);
		1104	}
		1105	OUTPUT:
		1106	RETVAL
		1107
966	void encode (SV self, SV scalar)	1108	void encode (SV self, SV scalar)
967	PPCODE:	1109	PPCODE:
968	XPUSHs (encode_json (scalar, *SvJSON (self)));	1110	XPUSHs (encode_json (scalar, *SvJSON (self)));
969		1111
970	void decode (SV self, SV jsonstr)	1112	void decode (SV self, SV jsonstr)
…		…
972	XPUSHs (decode_json (jsonstr, *SvJSON (self)));	1114	XPUSHs (decode_json (jsonstr, *SvJSON (self)));
973		1115
974	PROTOTYPES: ENABLE	1116	PROTOTYPES: ENABLE
975		1117
976	void to_json (SV *scalar)	1118	void to_json (SV *scalar)
		1119	ALIAS:
		1120	objToJson = 0
977	PPCODE:	1121	PPCODE:
978	XPUSHs (encode_json (scalar, F_UTF8));	1122	XPUSHs (encode_json (scalar, F_DEFAULT \| F_UTF8));
979		1123
980	void from_json (SV *jsonstr)	1124	void from_json (SV *jsonstr)
		1125	ALIAS:
		1126	jsonToObj = 0
981	PPCODE:	1127	PPCODE:
982	XPUSHs (decode_json (jsonstr, F_UTF8));	1128	XPUSHs (decode_json (jsonstr, F_DEFAULT \| F_UTF8));
983		1129

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing JSON-XS/XS.xs (file contents): Revision 1.11 by root, Sat Mar 24 19:42:14 2007 UTC vs. Revision 1.27 by root, Mon Apr 9 05:09:57 2007 UTC

Diff Legend

Comparing JSON-XS/XS.xs (file contents):
Revision 1.11 by root, Sat Mar 24 19:42:14 2007 UTC vs.
Revision 1.27 by root, Mon Apr 9 05:09:57 2007 UTC