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

Comparing JSON-XS/XS.xs (file contents):
Revision 1.8 by root, Fri Mar 23 16:13:59 2007 UTC vs.
Revision 1.31 by root, Wed May 9 16:33:53 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
		15	#define F_LATIN1 0x00000002UL
10	#define F_UTF8 0x00000002	16	#define F_UTF8 0x00000004UL
11	#define F_INDENT 0x00000004	17	#define F_INDENT 0x00000008UL
12	#define F_CANONICAL 0x00000008	18	#define F_CANONICAL 0x00000010UL
13	#define F_SPACE_BEFORE 0x00000010	19	#define F_SPACE_BEFORE 0x00000020UL
14	#define F_SPACE_AFTER 0x00000020	20	#define F_SPACE_AFTER 0x00000040UL
15	#define F_JSON_RPC 0x00000040
16	#define F_ALLOW_NONREF 0x00000080	21	#define F_ALLOW_NONREF 0x00000100UL
17	#define F_SHRINK 0x00000100	22	#define F_SHRINK 0x00000200UL
		23	#define F_MAXDEPTH 0xf8000000UL
		24	#define S_MAXDEPTH 27
		25
		26	#define DEC_DEPTH(flags) (1UL << ((flags & F_MAXDEPTH) >> S_MAXDEPTH))
		27
		28	// F_SELFCONVERT? <=> to_json/toJson
		29	// F_BLESSED? <=> { $__class__$ => }
18		30
19	#define F_PRETTY F_INDENT \| F_SPACE_BEFORE \| F_SPACE_AFTER	31	#define F_PRETTY F_INDENT \| F_SPACE_BEFORE \| F_SPACE_AFTER
20	#define F_DEFAULT 0	32	#define F_DEFAULT (9UL << S_MAXDEPTH)
21		33
22	#define INIT_SIZE 32 // initial scalar size to be allocated	34	#define INIT_SIZE 32 // initial scalar size to be allocated
		35	#define INDENT_STEP 3 // spaces per indentation level
		36
		37	#define SHORT_STRING_LEN 512 // special-case strings of up to this size
23		38
24	#define SB do {	39	#define SB do {
25	#define SE } while (0)	40	#define SE } while (0)
26		41
27	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
28		94
29	// structure used for encoding JSON	95	// structure used for encoding JSON
30	typedef struct	96	typedef struct
31	{	97	{
32	char *cur;	98	char *cur; // SvPVX (sv) + current output position
33	STRLEN len; // SvLEN (sv)
34	char *end; // SvEND (sv)	99	char *end; // SvEND (sv)
35	SV *sv;	100	SV *sv; // result scalar
36	UV flags;	101	U32 flags; // F_*
37	int max_recurse;	102	U32 indent; // indentation level
38	int indent;	103	U32 maxdepth; // max. indentation/recursion level
39	} enc_t;	104	} enc_t;
40
41	// structure used for decoding JSON
42	typedef struct
43	{
44	char *cur;
45	char *end;
46	const char *err;
47	UV flags;
48	} dec_t;
49
50	static UV *
51	SvJSON (SV *sv)
52	{
53	if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
54	croak ("object is not of type JSON::XS");
55
56	return &SvUVX (SvRV (sv));
57	}
58
59	static void
60	shrink (SV *sv)
61	{
62	sv_utf8_downgrade (sv, 1);
63	#ifdef SvPV_shrink_to_cur
64	SvPV_shrink_to_cur (sv);
65	#endif
66	}
67
68	/////////////////////////////////////////////////////////////////////////////
69		105
70	static void	106	static void
71	need (enc_t *enc, STRLEN len)	107	need (enc_t *enc, STRLEN len)
72	{	108	{
73	if (enc->cur + len >= enc->end)	109	if (enc->cur + len >= enc->end)
74	{	110	{
75	STRLEN cur = enc->cur - SvPVX (enc->sv);	111	STRLEN cur = enc->cur - SvPVX (enc->sv);
76	SvGROW (enc->sv, cur + len + 1);	112	SvGROW (enc->sv, cur + len + 1);
77	enc->cur = SvPVX (enc->sv) + cur;	113	enc->cur = SvPVX (enc->sv) + cur;
78	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv);	114	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
79	}	115	}
80	}	116	}
81		117
82	static void	118	static void
83	encode_ch (enc_t *enc, char ch)	119	encode_ch (enc_t *enc, char ch)
…		…
131	STRLEN clen;	167	STRLEN clen;
132	UV uch;	168	UV uch;
133		169
134	if (is_utf8)	170	if (is_utf8)
135	{	171	{
136	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);
137	if (clen == (STRLEN)-1)	174	if (clen == (STRLEN)-1)
138	croak ("malformed UTF-8 character in string, cannot convert to JSON");	175	croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
139	}	176	}
140	else	177	else
141	{	178	{
142	uch = ch;	179	uch = ch;
143	clen = 1;	180	clen = 1;
144	}	181	}
145		182
146	if (uch < 0x80 \|\| enc->flags & F_ASCII)	183	if (uch > 0x10FFFFUL)
		184	croak ("out of range codepoint (0x%lx) encountered, unrepresentable in JSON", (unsigned long)uch);
		185
		186	if (uch < 0x80 \|\| enc->flags & F_ASCII \|\| (enc->flags & F_LATIN1 && uch > 0xFF))
147	{	187	{
148	if (uch > 0xFFFFUL)	188	if (uch > 0xFFFFUL)
149	{	189	{
150	need (enc, len += 11);	190	need (enc, len += 11);
151	sprintf (enc->cur, "\\u%04x\\u%04x",	191	sprintf (enc->cur, "\\u%04x\\u%04x",
152	(uch - 0x10000) / 0x400 + 0xD800,	192	(int)((uch - 0x10000) / 0x400 + 0xD800),
153	(uch - 0x10000) % 0x400 + 0xDC00);	193	(int)((uch - 0x10000) % 0x400 + 0xDC00));
154	enc->cur += 12;	194	enc->cur += 12;
155	}	195	}
156	else	196	else
157	{	197	{
158	static char hexdigit [16] = "0123456789abcdef";	198	static char hexdigit [16] = "0123456789abcdef";
…		…
165	*enc->cur++ = hexdigit [(uch >> 0) & 15];	205	*enc->cur++ = hexdigit [(uch >> 0) & 15];
166	}	206	}
167		207
168	str += clen;	208	str += clen;
169	}	209	}
		210	else if (enc->flags & F_LATIN1)
		211	{
		212	*enc->cur++ = uch;
		213	str += clen;
		214	}
170	else if (is_utf8)	215	else if (is_utf8)
171	{	216	{
172	need (enc, len += clen);	217	need (enc, len += clen);
173	do	218	do
174	{	219	{
…		…
176	}	221	}
177	while (--clen);	222	while (--clen);
178	}	223	}
179	else	224	else
180	{	225	{
181	need (enc, len += 10); // never more than 11 bytes needed	226	need (enc, len += UTF8_MAXBYTES - 1); // never more than 11 bytes needed
182	enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);	227	enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
183	++str;	228	++str;
184	}	229	}
185	}	230	}
186	}	231	}
…		…
188		233
189	--len;	234	--len;
190	}	235	}
191	}	236	}
192		237
193	#define INDENT SB \	238	static void
		239	encode_indent (enc_t *enc)
		240	{
194	if (enc->flags & F_INDENT) \	241	if (enc->flags & F_INDENT)
195	{ \	242	{
196	int i_; \	243	int spaces = enc->indent * INDENT_STEP;
197	need (enc, enc->indent); \
198	for (i_ = enc->indent * 3; i_--; )\
199	encode_ch (enc, ' '); \
200	} \
201	SE
202		244
203	#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE	245	need (enc, spaces);
204	#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE	246	memset (enc->cur, ' ', spaces);
205	#define COMMA SB \	247	enc->cur += spaces;
		248	}
		249	}
		250
		251	static void
		252	encode_space (enc_t *enc)
		253	{
		254	need (enc, 1);
206	encode_ch (enc, ','); \	255	encode_ch (enc, ' ');
		256	}
		257
		258	static void
		259	encode_nl (enc_t *enc)
		260	{
207	if (enc->flags & F_INDENT) \	261	if (enc->flags & F_INDENT)
208	NL; \	262	{
		263	need (enc, 1);
		264	encode_ch (enc, '\n');
		265	}
		266	}
		267
		268	static void
		269	encode_comma (enc_t *enc)
		270	{
		271	encode_ch (enc, ',');
		272
		273	if (enc->flags & F_INDENT)
		274	encode_nl (enc);
209	else if (enc->flags & F_SPACE_AFTER) \	275	else if (enc->flags & F_SPACE_AFTER)
210	SPACE; \	276	encode_space (enc);
211	SE	277	}
212		278
213	static void encode_sv (enc_t enc, SV sv);	279	static void encode_sv (enc_t enc, SV sv);
214		280
215	static void	281	static void
216	encode_av (enc_t enc, AV av)	282	encode_av (enc_t enc, AV av)
217	{	283	{
218	int i, len = av_len (av);	284	int i, len = av_len (av);
219		285
220	encode_ch (enc, '['); NL;	286	if (enc->indent >= enc->maxdepth)
		287	croak ("data structure too deep (hit recursion limit)");
		288
		289	encode_ch (enc, '['); encode_nl (enc);
221	++enc->indent;	290	++enc->indent;
222		291
223	for (i = 0; i <= len; ++i)	292	for (i = 0; i <= len; ++i)
224	{	293	{
225	INDENT;	294	encode_indent (enc);
226	encode_sv (enc, *av_fetch (av, i, 0));	295	encode_sv (enc, *av_fetch (av, i, 0));
227		296
228	if (i < len)	297	if (i < len)
229	COMMA;	298	encode_comma (enc);
230	}	299	}
231		300
232	NL;	301	encode_nl (enc);
233		302
234	--enc->indent;	303	--enc->indent;
235	INDENT; encode_ch (enc, ']');	304	encode_indent (enc); encode_ch (enc, ']');
236	}	305	}
237		306
238	static void	307	static void
239	encode_he (enc_t enc, HE he)	308	encode_he (enc_t enc, HE he)
240	{	309	{
…		…
254	else	323	else
255	encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));	324	encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
256		325
257	encode_ch (enc, '"');	326	encode_ch (enc, '"');
258		327
259	if (enc->flags & F_SPACE_BEFORE) SPACE;	328	if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
260	encode_ch (enc, ':');	329	encode_ch (enc, ':');
261	if (enc->flags & F_SPACE_AFTER ) SPACE;	330	if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
262	encode_sv (enc, HeVAL (he));	331	encode_sv (enc, HeVAL (he));
263	}	332	}
264		333
265	// compare hash entries, used when all keys are bytestrings	334	// compare hash entries, used when all keys are bytestrings
266	static int	335	static int
…		…
272	HE b = (HE **)b_;	341	HE b = (HE **)b_;
273		342
274	STRLEN la = HeKLEN (a);	343	STRLEN la = HeKLEN (a);
275	STRLEN lb = HeKLEN (b);	344	STRLEN lb = HeKLEN (b);
276		345
277	if (!(cmp == memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))	346	if (!(cmp = memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))
278	cmp = la < lb ? -1 : la == lb ? 0 : 1;	347	cmp = la - lb;
279		348
280	return cmp;	349	return cmp;
281	}	350	}
282		351
283	// compare hash entries, used when some keys are sv's or utf-x	352	// compare hash entries, used when some keys are sv's or utf-x
…		…
290	static void	359	static void
291	encode_hv (enc_t enc, HV hv)	360	encode_hv (enc_t enc, HV hv)
292	{	361	{
293	int count, i;	362	int count, i;
294		363
		364	if (enc->indent >= enc->maxdepth)
		365	croak ("data structure too deep (hit recursion limit)");
		366
295	encode_ch (enc, '{'); NL; ++enc->indent;	367	encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
296		368
297	if ((count = hv_iterinit (hv)))	369	if ((count = hv_iterinit (hv)))
298	{	370	{
299	// for canonical output we have to sort by keys first	371	// for canonical output we have to sort by keys first
300	// actually, this is mostly due to the stupid so-called	372	// actually, this is mostly due to the stupid so-called
301	// security workaround added somewhere in 5.8.x.	373	// security workaround added somewhere in 5.8.x.
302	// that randomises hash orderings	374	// that randomises hash orderings
303	if (enc->flags & F_CANONICAL)	375	if (enc->flags & F_CANONICAL)
304	{	376	{
305	HE he, hes [count];	377	HE he, hes [count]; // if your compiler dies here, you need to enable C99 mode
306	int fast = 1;	378	int fast = 1;
307		379
308	i = 0;	380	i = 0;
309	while ((he = hv_iternext (hv)))	381	while ((he = hv_iternext (hv)))
310	{	382	{
…		…
335	LEAVE;	407	LEAVE;
336	}	408	}
337		409
338	for (i = 0; i < count; ++i)	410	for (i = 0; i < count; ++i)
339	{	411	{
340	INDENT;	412	encode_indent (enc);
341	encode_he (enc, hes [i]);	413	encode_he (enc, hes [i]);
342		414
343	if (i < count - 1)	415	if (i < count - 1)
344	COMMA;	416	encode_comma (enc);
345	}	417	}
346		418
347	NL;	419	encode_nl (enc);
348	}	420	}
349	else	421	else
350	{	422	{
351	SV *sv;
352	HE *he = hv_iternext (hv);	423	HE *he = hv_iternext (hv);
353		424
354	for (;;)	425	for (;;)
355	{	426	{
356	INDENT;	427	encode_indent (enc);
357	encode_he (enc, he);	428	encode_he (enc, he);
358		429
359	if (!(he = hv_iternext (hv)))	430	if (!(he = hv_iternext (hv)))
360	break;	431	break;
361		432
362	COMMA;	433	encode_comma (enc);
363	}	434	}
364		435
365	NL;	436	encode_nl (enc);
366	}	437	}
367	}	438	}
368		439
369	--enc->indent; INDENT; encode_ch (enc, '}');	440	--enc->indent; encode_indent (enc); encode_ch (enc, '}');
		441	}
		442
		443	// encode objects, arrays and special \0=false and \1=true values.
		444	static void
		445	encode_rv (enc_t enc, SV sv)
		446	{
		447	svtype svt;
		448
		449	SvGETMAGIC (sv);
		450	svt = SvTYPE (sv);
		451
		452	if (svt == SVt_PVHV)
		453	encode_hv (enc, (HV *)sv);
		454	else if (svt == SVt_PVAV)
		455	encode_av (enc, (AV *)sv);
		456	else if (svt < SVt_PVAV)
		457	{
		458	if (SvNIOK (sv) && SvIV (sv) == 0)
		459	encode_str (enc, "false", 5, 0);
		460	else if (SvNIOK (sv) && SvIV (sv) == 1)
		461	encode_str (enc, "true", 4, 0);
		462	else
		463	croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
		464	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
		465	}
		466	else
		467	croak ("encountered %s, but JSON can only represent references to arrays or hashes",
		468	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
370	}	469	}
371		470
372	static void	471	static void
373	encode_sv (enc_t enc, SV sv)	472	encode_sv (enc_t enc, SV sv)
374	{	473	{
…		…
395	SvIsUV(sv)	494	SvIsUV(sv)
396	? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))	495	? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))
397	: snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));	496	: snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));
398	}	497	}
399	else if (SvROK (sv))	498	else if (SvROK (sv))
400	{	499	encode_rv (enc, SvRV (sv));
401	if (!--enc->max_recurse)
402	croak ("data structure too deep (hit recursion limit)");
403
404	sv = SvRV (sv);
405
406	switch (SvTYPE (sv))
407	{
408	case SVt_PVAV: encode_av (enc, (AV *)sv); break;
409	case SVt_PVHV: encode_hv (enc, (HV *)sv); break;
410
411	default:
412	croak ("JSON can only represent references to arrays or hashes");
413	}
414	}
415	else if (!SvOK (sv))	500	else if (!SvOK (sv))
416	encode_str (enc, "null", 4, 0);	501	encode_str (enc, "null", 4, 0);
417	else	502	else
418	croak ("encountered perl type that JSON cannot handle");	503	croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this",
		504	SvPV_nolen (sv), SvFLAGS (sv));
419	}	505	}
420		506
421	static SV *	507	static SV *
422	encode_json (SV *scalar, UV flags)	508	encode_json (SV *scalar, U32 flags)
423	{	509	{
		510	enc_t enc;
		511
424	if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))	512	if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
425	croak ("hash- or arraref required (not a simple scalar, use allow_nonref to allow this)");	513	croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)");
426		514
427	enc_t enc;
428	enc.flags = flags;	515	enc.flags = flags;
429	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	516	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
430	enc.cur = SvPVX (enc.sv);	517	enc.cur = SvPVX (enc.sv);
431	enc.end = SvEND (enc.sv);	518	enc.end = SvEND (enc.sv);
432	enc.max_recurse = 0;
433	enc.indent = 0;	519	enc.indent = 0;
		520	enc.maxdepth = DEC_DEPTH (flags);
434		521
435	SvPOK_only (enc.sv);	522	SvPOK_only (enc.sv);
436	encode_sv (&enc, scalar);	523	encode_sv (&enc, scalar);
437		524
		525	SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
		526	*SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
		527
438	if (!(flags & (F_ASCII \| F_UTF8)))	528	if (!(flags & (F_ASCII \| F_LATIN1 \| F_UTF8)))
439	SvUTF8_on (enc.sv);	529	SvUTF8_on (enc.sv);
440
441	SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
442		530
443	if (enc.flags & F_SHRINK)	531	if (enc.flags & F_SHRINK)
444	shrink (enc.sv);	532	shrink (enc.sv);
445		533
446	return enc.sv;	534	return enc.sv;
447	}	535	}
448		536
449	/////////////////////////////////////////////////////////////////////////////	537	/////////////////////////////////////////////////////////////////////////////
		538	// decoder
450		539
451	#define WS \	540	// structure used for decoding JSON
		541	typedef struct
		542	{
		543	char *cur; // current parser pointer
		544	char *end; // end of input string
		545	const char *err; // parse error, if != 0
		546	U32 flags; // F_*
		547	U32 depth; // recursion depth
		548	U32 maxdepth; // recursion depth limit
		549	} dec_t;
		550
		551	static void
		552	decode_ws (dec_t *dec)
		553	{
452	for (;;) \	554	for (;;)
453	{ \	555	{
454	char ch = *dec->cur; \	556	char ch = *dec->cur;
		557
455	if (ch > 0x20 \	558	if (ch > 0x20
456	\|\| (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \	559	\|\| (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
457	break; \	560	break;
		561
458	++dec->cur; \	562	++dec->cur;
459	}	563	}
		564	}
460		565
461	#define ERR(reason) SB dec->err = reason; goto fail; SE	566	#define ERR(reason) SB dec->err = reason; goto fail; SE
		567
462	#define EXPECT_CH(ch) SB \	568	#define EXPECT_CH(ch) SB \
463	if (*dec->cur != ch) \	569	if (*dec->cur != ch) \
464	ERR (# ch " expected"); \	570	ERR (# ch " expected"); \
465	++dec->cur; \	571	++dec->cur; \
466	SE	572	SE
467		573
		574	#define DEC_INC_DEPTH if (++dec->depth > dec->maxdepth) ERR ("json datastructure exceeds maximum nesting level (set a higher max_depth)")
		575	#define DEC_DEC_DEPTH --dec->depth
		576
468	static SV decode_sv (dec_t dec);	577	static SV decode_sv (dec_t dec);
469		578
470	static signed char decode_hexdigit[256];	579	static signed char decode_hexdigit[256];
471		580
472	static UV	581	static UV
473	decode_4hex (dec_t *dec)	582	decode_4hex (dec_t *dec)
474	{	583	{
475	signed char d1, d2, d3, d4;	584	signed char d1, d2, d3, d4;
		585	unsigned char cur = (unsigned char )dec->cur;
476		586
477	d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
478	if (d1 < 0) ERR ("four hexadecimal digits expected");	587	d1 = decode_hexdigit [cur [0]]; if (d1 < 0) ERR ("four hexadecimal digits expected");
479	d2 = decode_hexdigit [((unsigned char *)dec->cur) [1]];
480	if (d2 < 0) ERR ("four hexadecimal digits expected");	588	d2 = decode_hexdigit [cur [1]]; if (d2 < 0) ERR ("four hexadecimal digits expected");
481	d3 = decode_hexdigit [((unsigned char *)dec->cur) [2]];
482	if (d3 < 0) ERR ("four hexadecimal digits expected");	589	d3 = decode_hexdigit [cur [2]]; if (d3 < 0) ERR ("four hexadecimal digits expected");
483	d4 = decode_hexdigit [((unsigned char *)dec->cur) [3]];
484	if (d4 < 0) ERR ("four hexadecimal digits expected");	590	d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
485		591
486	dec->cur += 4;	592	dec->cur += 4;
487		593
488	return ((UV)d1) << 12	594	return ((UV)d1) << 12
489	\| ((UV)d2) << 8	595	\| ((UV)d2) << 8
…		…
492		598
493	fail:	599	fail:
494	return (UV)-1;	600	return (UV)-1;
495	}	601	}
496		602
497	#define APPEND_GROW(n) SB \
498	if (cur + (n) >= end) \
499	{ \
500	STRLEN ofs = cur - SvPVX (sv); \
501	SvGROW (sv, ofs + (n) + 1); \
502	cur = SvPVX (sv) + ofs; \
503	end = SvEND (sv); \
504	} \
505	SE
506
507	#define APPEND_CH(ch) SB \
508	APPEND_GROW (1); \
509	*cur++ = (ch); \
510	SE
511
512	static SV *	603	static SV *
513	decode_str (dec_t *dec)	604	decode_str (dec_t *dec)
514	{	605	{
515	SV *sv = NEWSV (0,2);	606	SV *sv = 0;
516	int utf8 = 0;	607	int utf8 = 0;
517	char *cur = SvPVX (sv);
518	char *end = SvEND (sv);
519		608
520	for (;;)	609	do
521	{	610	{
522	unsigned char ch = (unsigned char )dec->cur;	611	char buf [SHORT_STRING_LEN + UTF8_MAXBYTES];
		612	char *cur = buf;
523		613
524	if (ch == '"')	614	do
525	break;
526	else if (ch == '\\')
527	{	615	{
528	switch (*++dec->cur)	616	unsigned char ch = (unsigned char )dec->cur++;
		617
		618	if (ch == '"')
529	{	619	{
530	case '\\':	620	--dec->cur;
531	case '/':	621	break;
532	case '"': APPEND_CH (*dec->cur++); break;	622	}
533		623	else if (ch == '\\')
534	case 'b': APPEND_CH ('\010'); ++dec->cur; break;	624	{
535	case 't': APPEND_CH ('\011'); ++dec->cur; break;	625	switch (*dec->cur)
536	case 'n': APPEND_CH ('\012'); ++dec->cur; break;
537	case 'f': APPEND_CH ('\014'); ++dec->cur; break;
538	case 'r': APPEND_CH ('\015'); ++dec->cur; break;
539
540	case 'u':
541	{	626	{
542	UV lo, hi;	627	case '\\':
543	++dec->cur;	628	case '/':
		629	case '"': cur++ = dec->cur++; break;
544		630
545	hi = decode_4hex (dec);	631	case 'b': ++dec->cur; *cur++ = '\010'; break;
546	if (hi == (UV)-1)	632	case 't': ++dec->cur; *cur++ = '\011'; break;
547	goto fail;	633	case 'n': ++dec->cur; *cur++ = '\012'; break;
		634	case 'f': ++dec->cur; *cur++ = '\014'; break;
		635	case 'r': ++dec->cur; *cur++ = '\015'; break;
548		636
549	// possibly a surrogate pair	637	case 'u':
550	if (hi >= 0xd800 && hi < 0xdc00)
551	{	638	{
552	if (dec->cur [0] != '\\' \|\| dec->cur [1] != 'u')	639	UV lo, hi;
553	ERR ("missing low surrogate character in surrogate pair");
554
555	dec->cur += 2;	640	++dec->cur;
556		641
557	lo = decode_4hex (dec);	642	hi = decode_4hex (dec);
558	if (lo == (UV)-1)	643	if (hi == (UV)-1)
559	goto fail;	644	goto fail;
560		645
		646	// possibly a surrogate pair
		647	if (hi >= 0xd800)
		648	if (hi < 0xdc00)
		649	{
		650	if (dec->cur [0] != '\\' \|\| dec->cur [1] != 'u')
		651	ERR ("missing low surrogate character in surrogate pair");
		652
		653	dec->cur += 2;
		654
		655	lo = decode_4hex (dec);
		656	if (lo == (UV)-1)
		657	goto fail;
		658
561	if (lo < 0xdc00 \|\| lo >= 0xe000)	659	if (lo < 0xdc00 \|\| lo >= 0xe000)
562	ERR ("surrogate pair expected");	660	ERR ("surrogate pair expected");
563		661
564	hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;	662	hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;
		663	}
		664	else if (hi < 0xe000)
		665	ERR ("missing high surrogate character in surrogate pair");
		666
		667	if (hi >= 0x80)
		668	{
		669	utf8 = 1;
		670
		671	cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
		672	}
		673	else
		674	*cur++ = hi;
565	}	675	}
566	else if (hi >= 0xdc00 && hi < 0xe000)
567	ERR ("missing high surrogate character in surrogate pair");
568
569	if (hi >= 0x80)
570	{	676	break;
571	utf8 = 1;
572		677
573	APPEND_GROW (4); // at most 4 bytes for 21 bits
574	cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
575	}
576	else	678	default:
577	APPEND_CH (hi);	679	--dec->cur;
		680	ERR ("illegal backslash escape sequence in string");
578	}	681	}
579	break;	682	}
		683	else if (ch >= 0x20 && ch <= 0x7f)
		684	*cur++ = ch;
		685	else if (ch >= 0x80)
		686	{
		687	STRLEN clen;
		688	UV uch;
580		689
581	default:
582	--dec->cur;	690	--dec->cur;
583	ERR ("illegal backslash escape sequence in string");	691
		692	uch = decode_utf8 (dec->cur, dec->end - dec->cur, &clen);
		693	if (clen == (STRLEN)-1)
		694	ERR ("malformed UTF-8 character in JSON string");
		695
		696	do
		697	cur++ = dec->cur++;
		698	while (--clen);
		699
		700	utf8 = 1;
		701	}
		702	else
		703	{
		704	--dec->cur;
		705
		706	if (!ch)
		707	ERR ("unexpected end of string while parsing JSON string");
		708	else
		709	ERR ("invalid character encountered while parsing JSON string");
584	}	710	}
585	}	711	}
586	else if (ch >= 0x20 && ch <= 0x7f)	712	while (cur < buf + SHORT_STRING_LEN);
587	APPEND_CH (*dec->cur++);	713
588	else if (ch >= 0x80)
589	{	714	{
590	STRLEN clen;	715	STRLEN len = cur - buf;
591	UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY);
592	if (clen == (STRLEN)-1)
593	ERR ("malformed UTF-8 character in JSON string");
594		716
595	APPEND_GROW (clen);	717	if (sv)
596	do
597	{	718	{
598	cur++ = dec->cur++;	719	SvGROW (sv, SvCUR (sv) + len + 1);
		720	memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
		721	SvCUR_set (sv, SvCUR (sv) + len);
599	}	722	}
600	while (--clen);
601
602	utf8 = 1;
603	}
604	else if (dec->cur == dec->end)
605	ERR ("unexpected end of string while parsing json string");
606	else	723	else
607	ERR ("invalid character encountered");	724	sv = newSVpvn (buf, len);
608	}	725	}
		726	}
		727	while (*dec->cur != '"');
609		728
610	++dec->cur;	729	++dec->cur;
611		730
612	SvCUR_set (sv, cur - SvPVX (sv));	731	if (sv)
613		732	{
614	SvPOK_only (sv);	733	SvPOK_only (sv);
615	*SvEND (sv) = 0;	734	*SvEND (sv) = 0;
616		735
617	if (utf8)	736	if (utf8)
618	SvUTF8_on (sv);	737	SvUTF8_on (sv);
619		738	}
620	if (dec->flags & F_SHRINK)	739	else
621	shrink (sv);	740	sv = newSVpvn ("", 0);
622		741
623	return sv;	742	return sv;
624		743
625	fail:	744	fail:
626	SvREFCNT_dec (sv);
627	return 0;	745	return 0;
628	}	746	}
629		747
630	static SV *	748	static SV *
631	decode_num (dec_t *dec)	749	decode_num (dec_t *dec)
…		…
712	static SV *	830	static SV *
713	decode_av (dec_t *dec)	831	decode_av (dec_t *dec)
714	{	832	{
715	AV *av = newAV ();	833	AV *av = newAV ();
716		834
717	WS;	835	DEC_INC_DEPTH;
		836	decode_ws (dec);
		837
718	if (*dec->cur == ']')	838	if (*dec->cur == ']')
719	++dec->cur;	839	++dec->cur;
720	else	840	else
721	for (;;)	841	for (;;)
722	{	842	{
…		…
726	if (!value)	846	if (!value)
727	goto fail;	847	goto fail;
728		848
729	av_push (av, value);	849	av_push (av, value);
730		850
731	WS;	851	decode_ws (dec);
732		852
733	if (*dec->cur == ']')	853	if (*dec->cur == ']')
734	{	854	{
735	++dec->cur;	855	++dec->cur;
736	break;	856	break;
…		…
740	ERR (", or ] expected while parsing array");	860	ERR (", or ] expected while parsing array");
741		861
742	++dec->cur;	862	++dec->cur;
743	}	863	}
744		864
		865	DEC_DEC_DEPTH;
745	return newRV_noinc ((SV *)av);	866	return newRV_noinc ((SV *)av);
746		867
747	fail:	868	fail:
748	SvREFCNT_dec (av);	869	SvREFCNT_dec (av);
		870	DEC_DEC_DEPTH;
749	return 0;	871	return 0;
750	}	872	}
751		873
752	static SV *	874	static SV *
753	decode_hv (dec_t *dec)	875	decode_hv (dec_t *dec)
754	{	876	{
755	HV *hv = newHV ();	877	HV *hv = newHV ();
756		878
757	WS;	879	DEC_INC_DEPTH;
		880	decode_ws (dec);
		881
758	if (*dec->cur == '}')	882	if (*dec->cur == '}')
759	++dec->cur;	883	++dec->cur;
760	else	884	else
761	for (;;)	885	for (;;)
762	{	886	{
763	SV key, value;	887	SV key, value;
764		888
765	WS; EXPECT_CH ('"');	889	decode_ws (dec); EXPECT_CH ('"');
766		890
767	key = decode_str (dec);	891	key = decode_str (dec);
768	if (!key)	892	if (!key)
769	goto fail;	893	goto fail;
770		894
771	WS; EXPECT_CH (':');	895	decode_ws (dec); EXPECT_CH (':');
772		896
773	value = decode_sv (dec);	897	value = decode_sv (dec);
774	if (!value)	898	if (!value)
775	{	899	{
776	SvREFCNT_dec (key);	900	SvREFCNT_dec (key);
777	goto fail;	901	goto fail;
778	}	902	}
779		903
780	//TODO: optimise
781	hv_store_ent (hv, key, value, 0);	904	hv_store_ent (hv, key, value, 0);
		905	SvREFCNT_dec (key);
782		906
783	WS;	907	decode_ws (dec);
784		908
785	if (*dec->cur == '}')	909	if (*dec->cur == '}')
786	{	910	{
787	++dec->cur;	911	++dec->cur;
788	break;	912	break;
…		…
792	ERR (", or } expected while parsing object/hash");	916	ERR (", or } expected while parsing object/hash");
793		917
794	++dec->cur;	918	++dec->cur;
795	}	919	}
796		920
		921	DEC_DEC_DEPTH;
797	return newRV_noinc ((SV *)hv);	922	return newRV_noinc ((SV *)hv);
798		923
799	fail:	924	fail:
800	SvREFCNT_dec (hv);	925	SvREFCNT_dec (hv);
		926	DEC_DEC_DEPTH;
801	return 0;	927	return 0;
802	}	928	}
803		929
804	static SV *	930	static SV *
805	decode_sv (dec_t *dec)	931	decode_sv (dec_t *dec)
806	{	932	{
807	WS;	933	decode_ws (dec);
808	switch (*dec->cur)	934	switch (*dec->cur)
809	{	935	{
810	case '"': ++dec->cur; return decode_str (dec);	936	case '"': ++dec->cur; return decode_str (dec);
811	case '[': ++dec->cur; return decode_av (dec);	937	case '[': ++dec->cur; return decode_av (dec);
812	case '{': ++dec->cur; return decode_hv (dec);	938	case '{': ++dec->cur; return decode_hv (dec);
…		…
848	ERR ("'null' expected");	974	ERR ("'null' expected");
849		975
850	break;	976	break;
851		977
852	default:	978	default:
853	ERR ("malformed json string, neither array, object, number, string or atom");	979	ERR ("malformed JSON string, neither array, object, number, string or atom");
854	break;	980	break;
855	}	981	}
856		982
857	fail:	983	fail:
858	return 0;	984	return 0;
859	}	985	}
860		986
861	static SV *	987	static SV *
862	decode_json (SV *string, UV flags)	988	decode_json (SV string, U32 flags, UV offset_return)
863	{	989	{
		990	dec_t dec;
		991	UV offset;
864	SV *sv;	992	SV *sv;
		993
		994	SvGETMAGIC (string);
		995	SvUPGRADE (string, SVt_PV);
865		996
866	if (flags & F_UTF8)	997	if (flags & F_UTF8)
867	sv_utf8_downgrade (string, 0);	998	sv_utf8_downgrade (string, 0);
868	else	999	else
869	sv_utf8_upgrade (string);	1000	sv_utf8_upgrade (string);
870		1001
871	SvGROW (string, SvCUR (string) + 1); // should basically be a NOP	1002	SvGROW (string, SvCUR (string) + 1); // should basically be a NOP
872		1003
873	dec_t dec;
874	dec.flags = flags;	1004	dec.flags = flags;
875	dec.cur = SvPVX (string);	1005	dec.cur = SvPVX (string);
876	dec.end = SvEND (string);	1006	dec.end = SvEND (string);
877	dec.err = 0;	1007	dec.err = 0;
		1008	dec.depth = 0;
		1009	dec.maxdepth = DEC_DEPTH (dec.flags);
878		1010
		1011	*dec.end = 0; // this should basically be a nop, too, but make sure it's there
879	sv = decode_sv (&dec);	1012	sv = decode_sv (&dec);
880		1013
		1014	if (offset_return \|\| !sv)
		1015	{
		1016	offset = dec.flags & F_UTF8
		1017	? dec.cur - SvPVX (string)
		1018	: utf8_distance (dec.cur, SvPVX (string));
		1019
		1020	if (offset_return)
		1021	*offset_return = offset;
		1022	}
		1023	else
		1024	{
		1025	// check for trailing garbage
		1026	decode_ws (&dec);
		1027
		1028	if (*dec.cur)
		1029	{
		1030	dec.err = "garbage after JSON object";
		1031	SvREFCNT_dec (sv);
		1032	sv = 0;
		1033	}
		1034	}
		1035
881	if (!sv)	1036	if (!sv)
882	{	1037	{
883	IV offset = dec.flags & F_UTF8
884	? dec.cur - SvPVX (string)
885	: utf8_distance (dec.cur, SvPVX (string));
886	SV *uni = sv_newmortal ();	1038	SV *uni = sv_newmortal ();
887		1039
888	// horrible hack to silence warning inside pv_uni_display	1040	// horrible hack to silence warning inside pv_uni_display
889	COP cop = *PL_curcop;	1041	COP cop = *PL_curcop;
890	cop.cop_warnings = pWARN_NONE;	1042	cop.cop_warnings = pWARN_NONE;
…		…
892	SAVEVPTR (PL_curcop);	1044	SAVEVPTR (PL_curcop);
893	PL_curcop = &cop;	1045	PL_curcop = &cop;
894	pv_uni_display (uni, dec.cur, dec.end - dec.cur, 20, UNI_DISPLAY_QQ);	1046	pv_uni_display (uni, dec.cur, dec.end - dec.cur, 20, UNI_DISPLAY_QQ);
895	LEAVE;	1047	LEAVE;
896		1048
897	croak ("%s, at character offset %d (%s)",	1049	croak ("%s, at character offset %d [\"%s\"]",
898	dec.err,	1050	dec.err,
899	(int)offset,	1051	(int)offset,
900	dec.cur != dec.end ? SvPV_nolen (uni) : "(end of string)");	1052	dec.cur != dec.end ? SvPV_nolen (uni) : "(end of string)");
901	}	1053	}
902		1054
903	sv = sv_2mortal (sv);	1055	sv = sv_2mortal (sv);
904		1056
905	if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv))	1057	if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv))
906	croak ("JSON object or array expected (but number, string, true, false or null found, use allow_nonref to allow this)");	1058	croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");
907		1059
908	return sv;	1060	return sv;
909	}	1061	}
910		1062
		1063	/////////////////////////////////////////////////////////////////////////////
		1064	// XS interface functions
		1065
911	MODULE = JSON::XS PACKAGE = JSON::XS	1066	MODULE = JSON::XS PACKAGE = JSON::XS
912		1067
913	BOOT:	1068	BOOT:
914	{	1069	{
915	int i;	1070	int i;
916		1071
917	memset (decode_hexdigit, 0xff, 256);	1072	memset (decode_hexdigit, 0xff, 256);
		1073
918	for (i = 10; i--; )	1074	for (i = 0; i < 256; ++i)
919	decode_hexdigit ['0' + i] = i;	1075	decode_hexdigit [i] =
920		1076	i >= '0' && i <= '9' ? i - '0'
921	for (i = 7; i--; )	1077	: i >= 'a' && i <= 'f' ? i - 'a' + 10
922	{	1078	: i >= 'A' && i <= 'F' ? i - 'A' + 10
923	decode_hexdigit ['a' + i] = 10 + i;	1079	: -1;
924	decode_hexdigit ['A' + i] = 10 + i;
925	}
926		1080
927	json_stash = gv_stashpv ("JSON::XS", 1);	1081	json_stash = gv_stashpv ("JSON::XS", 1);
928	}	1082	}
929		1083
930	PROTOTYPES: DISABLE	1084	PROTOTYPES: DISABLE
…		…
936	RETVAL	1090	RETVAL
937		1091
938	SV ascii (SV self, int enable = 1)	1092	SV ascii (SV self, int enable = 1)
939	ALIAS:	1093	ALIAS:
940	ascii = F_ASCII	1094	ascii = F_ASCII
		1095	latin1 = F_LATIN1
941	utf8 = F_UTF8	1096	utf8 = F_UTF8
942	indent = F_INDENT	1097	indent = F_INDENT
943	canonical = F_CANONICAL	1098	canonical = F_CANONICAL
944	space_before = F_SPACE_BEFORE	1099	space_before = F_SPACE_BEFORE
945	space_after = F_SPACE_AFTER	1100	space_after = F_SPACE_AFTER
946	json_rpc = F_JSON_RPC
947	pretty = F_PRETTY	1101	pretty = F_PRETTY
948	allow_nonref = F_ALLOW_NONREF	1102	allow_nonref = F_ALLOW_NONREF
949	shrink = F_SHRINK	1103	shrink = F_SHRINK
950	CODE:	1104	CODE:
951	{	1105	{
…		…
958	RETVAL = newSVsv (self);	1112	RETVAL = newSVsv (self);
959	}	1113	}
960	OUTPUT:	1114	OUTPUT:
961	RETVAL	1115	RETVAL
962		1116
		1117	SV max_depth (SV self, UV max_depth = 0x80000000UL)
		1118	CODE:
		1119	{
		1120	UV *uv = SvJSON (self);
		1121	UV log2 = 0;
		1122
		1123	if (max_depth > 0x80000000UL) max_depth = 0x80000000UL;
		1124
		1125	while ((1UL << log2) < max_depth)
		1126	++log2;
		1127
		1128	uv = uv & ~F_MAXDEPTH \| (log2 << S_MAXDEPTH);
		1129
		1130	RETVAL = newSVsv (self);
		1131	}
		1132	OUTPUT:
		1133	RETVAL
		1134
963	void encode (SV self, SV scalar)	1135	void encode (SV self, SV scalar)
964	PPCODE:	1136	PPCODE:
965	XPUSHs (encode_json (scalar, *SvJSON (self)));	1137	XPUSHs (encode_json (scalar, *SvJSON (self)));
966		1138
967	void decode (SV self, SV jsonstr)	1139	void decode (SV self, SV jsonstr)
968	PPCODE:	1140	PPCODE:
969	XPUSHs (decode_json (jsonstr, *SvJSON (self)));	1141	XPUSHs (decode_json (jsonstr, *SvJSON (self), 0));
		1142
		1143	void decode_prefix (SV self, SV jsonstr)
		1144	PPCODE:
		1145	{
		1146	UV offset;
		1147	EXTEND (SP, 2);
		1148	PUSHs (decode_json (jsonstr, *SvJSON (self), &offset));
		1149	PUSHs (sv_2mortal (newSVuv (offset)));
		1150	}
970		1151
971	PROTOTYPES: ENABLE	1152	PROTOTYPES: ENABLE
972		1153
973	void to_json (SV *scalar)	1154	void to_json (SV *scalar)
		1155	ALIAS:
		1156	objToJson = 0
974	PPCODE:	1157	PPCODE:
975	XPUSHs (encode_json (scalar, F_UTF8));	1158	XPUSHs (encode_json (scalar, F_DEFAULT \| F_UTF8));
976		1159
977	void from_json (SV *jsonstr)	1160	void from_json (SV *jsonstr)
		1161	ALIAS:
		1162	jsonToObj = 0
978	PPCODE:	1163	PPCODE:
979	XPUSHs (decode_json (jsonstr, F_UTF8));	1164	XPUSHs (decode_json (jsonstr, F_DEFAULT \| F_UTF8, 0));
980		1165

Diff Legend

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

Comparing JSON-XS/XS.xs (file contents): Revision 1.8 by root, Fri Mar 23 16:13:59 2007 UTC vs. Revision 1.31 by root, Wed May 9 16:33:53 2007 UTC

Diff Legend

Comparing JSON-XS/XS.xs (file contents):
Revision 1.8 by root, Fri Mar 23 16:13:59 2007 UTC vs.
Revision 1.31 by root, Wed May 9 16:33:53 2007 UTC