ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs.
Revision 1.66 by root, Sun Nov 29 21:32:01 2020 UTC

…		…
6	#include <string.h>	6	#include <string.h>
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
		11	#include <inttypes.h>
11		12
		13	#define ECB_NO_THREADS 1
12	#include "ecb.h"	14	#include "ecb.h"
13		15
14	// compatibility with perl <5.18	16	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	17	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	18	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	21	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	22	#endif
21	#ifndef HvNAMEUTF8	23	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	24	# define HvNAMEUTF8(hv) 0
23	#endif	25	#endif
		26	#ifndef SvREFCNT_inc_NN
		27	# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
		28	#endif
		29	#ifndef SvREFCNT_dec_NN
		30	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		31	#endif
		32
		33	// known major and minor types
		34	enum cbor_type
		35	{
		36	MAJOR_SHIFT = 5,
		37	MINOR_MASK = 0x1f,
		38
		39	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		40	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		41	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		42	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		43	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		44	MAJOR_MAP = 5 << MAJOR_SHIFT,
		45	MAJOR_TAG = 6 << MAJOR_SHIFT,
		46	MAJOR_MISC = 7 << MAJOR_SHIFT,
		47
		48	// INT/STRING/ARRAY/MAP subtypes
		49	LENGTH_EXT1 = 24,
		50	LENGTH_EXT2 = 25,
		51	LENGTH_EXT4 = 26,
		52	LENGTH_EXT8 = 27,
		53
		54	// SIMPLE types (effectively MISC subtypes)
		55	SIMPLE_FALSE = 20,
		56	SIMPLE_TRUE = 21,
		57	SIMPLE_NULL = 22,
		58	SIMPLE_UNDEF = 23,
		59
		60	// MISC subtype (unused)
		61	MISC_EXT1 = 24,
		62	MISC_FLOAT16 = 25,
		63	MISC_FLOAT32 = 26,
		64	MISC_FLOAT64 = 27,
		65
		66	// BYTES/TEXT/ARRAY/MAP
		67	MINOR_INDEF = 31,
		68	};
24		69
25	// known tags	70	// known tags
26	enum cbor_tag	71	enum cbor_tag
27	{	72	{
28	// inofficial extensions (pending iana registration)	73	// extensions
		74	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	75	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	76	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	77	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	78	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	79	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	80	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		81
37	// rfc7049	82	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	83	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	84	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	85	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	86	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	87	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	88	CBOR_TAG_BIGFLOAT = 5, // array
44		89
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	90	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	91	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	92	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	93	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		94
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	95	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	96	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	97	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	98	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	99	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		100
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	101	CBOR_TAG_MAGIC = 55799, // self-describe cbor
		102	};
		103
		104	// known forced types, also hardcoded in CBOR.pm
		105	enum
		106	{
		107	AS_CBOR = 0,
		108	AS_BYTES = 1,
		109	AS_TEXT = 2,
		110	AS_FLOAT16 = 3,
		111	AS_FLOAT32 = 4,
		112	AS_FLOAT64 = 5,
		113	// possibly future enhancements: float, integer
57	};	114	};
58		115
59	#define F_SHRINK 0x00000001UL	116	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	117	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	118	#define F_ALLOW_SHARING 0x00000004UL
62	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	119	#define F_ALLOW_CYCLES 0x00000008UL
		120	#define F_FORBID_OBJECTS 0x00000010UL
		121	#define F_PACK_STRINGS 0x00000020UL
		122	#define F_TEXT_KEYS 0x00000040UL
		123	#define F_TEXT_STRINGS 0x00000080UL
		124	#define F_VALIDATE_UTF8 0x00000100UL
63		125
64	#define INIT_SIZE 32 // initial scalar size to be allocated	126	#define INIT_SIZE 32 // initial scalar size to be allocated
65		127
66	#define SB do {	128	#define SB do {
67	#define SE } while (0)	129	#define SE } while (0)
…		…
86	typedef struct {	148	typedef struct {
87	U32 flags;	149	U32 flags;
88	U32 max_depth;	150	U32 max_depth;
89	STRLEN max_size;	151	STRLEN max_size;
90	SV *filter;	152	SV *filter;
		153
		154	// for the incremental parser
		155	STRLEN incr_pos; // the current offset into the text
		156	STRLEN incr_need; // minimum bytes needed to decode
		157	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91	} CBOR;	158	} CBOR;
92		159
93	ecb_inline void	160	ecb_inline void
94	cbor_init (CBOR *cbor)	161	cbor_init (CBOR *cbor)
95	{	162	{
…		…
99		166
100	ecb_inline void	167	ecb_inline void
101	cbor_free (CBOR *cbor)	168	cbor_free (CBOR *cbor)
102	{	169	{
103	SvREFCNT_dec (cbor->filter);	170	SvREFCNT_dec (cbor->filter);
		171	SvREFCNT_dec (cbor->incr_count);
104	}	172	}
105		173
106	/////////////////////////////////////////////////////////////////////////////	174	/////////////////////////////////////////////////////////////////////////////
107	// utility functions	175	// utility functions
108		176
…		…
134		202
135	// minimum length of a string to be registered for stringref	203	// minimum length of a string to be registered for stringref
136	ecb_inline int	204	ecb_inline int
137	minimum_string_length (UV idx)	205	minimum_string_length (UV idx)
138	{	206	{
139	return idx > 23	207	return idx <= 23 ? 3
140	? idx > 0xffU	208	: idx <= 0xffU ? 4
141	? idx > 0xffffU	209	: idx <= 0xffffU ? 5
142	? idx > 0xffffffffU	210	: idx <= 0xffffffffU ? 7
143	? 7	211	: 11;
144	: 6
145	: 5
146	: 4
147	: 3;
148	}	212	}
149		213
150	/////////////////////////////////////////////////////////////////////////////	214	/////////////////////////////////////////////////////////////////////////////
151	// encoder	215	// encoder
152		216
…		…
165	} enc_t;	229	} enc_t;
166		230
167	ecb_inline void	231	ecb_inline void
168	need (enc_t *enc, STRLEN len)	232	need (enc_t *enc, STRLEN len)
169	{	233	{
170	if (ecb_expect_false (enc->cur + len >= enc->end))	234	if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
171	{	235	{
172	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);	236	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
173	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);	237	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
174	enc->cur = SvPVX (enc->sv) + cur;	238	enc->cur = SvPVX (enc->sv) + cur;
175	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;	239	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
…		…
181	{	245	{
182	need (enc, 1);	246	need (enc, 1);
183	*enc->cur++ = ch;	247	*enc->cur++ = ch;
184	}	248	}
185		249
		250	// used for tags, intregers, element counts and so on
186	static void	251	static void
187	encode_uint (enc_t *enc, int major, UV len)	252	encode_uint (enc_t *enc, int major, UV len)
188	{	253	{
189	need (enc, 9);	254	need (enc, 9);
190		255
191	if (len < 24)	256	if (ecb_expect_true (len < LENGTH_EXT1))
192	*enc->cur++ = major | len;	257	*enc->cur++ = major | len;
193	else if (len <= 0xff)	258	else if (ecb_expect_true (len <= 0xffU))
194	{	259	{
195	*enc->cur++ = major | 24;	260	*enc->cur++ = major | LENGTH_EXT1;
196	*enc->cur++ = len;	261	*enc->cur++ = len;
197	}	262	}
198	else if (len <= 0xffff)	263	else if (len <= 0xffffU)
199	{	264	{
200	*enc->cur++ = major | 25;	265	*enc->cur++ = major | LENGTH_EXT2;
201	*enc->cur++ = len >> 8;	266	*enc->cur++ = len >> 8;
202	*enc->cur++ = len;	267	*enc->cur++ = len;
203	}	268	}
204	else if (len <= 0xffffffff)	269	else if (len <= 0xffffffffU)
205	{	270	{
206	*enc->cur++ = major | 26;	271	*enc->cur++ = major | LENGTH_EXT4;
207	*enc->cur++ = len >> 24;	272	*enc->cur++ = len >> 24;
208	*enc->cur++ = len >> 16;	273	*enc->cur++ = len >> 16;
209	*enc->cur++ = len >> 8;	274	*enc->cur++ = len >> 8;
210	*enc->cur++ = len;	275	*enc->cur++ = len;
211	}	276	}
212	else	277	else
213	{	278	{
214	*enc->cur++ = major | 27;	279	*enc->cur++ = major | LENGTH_EXT8;
215	*enc->cur++ = len >> 56;	280	*enc->cur++ = len >> 56;
216	*enc->cur++ = len >> 48;	281	*enc->cur++ = len >> 48;
217	*enc->cur++ = len >> 40;	282	*enc->cur++ = len >> 40;
218	*enc->cur++ = len >> 32;	283	*enc->cur++ = len >> 32;
219	*enc->cur++ = len >> 24;	284	*enc->cur++ = len >> 24;
…		…
221	*enc->cur++ = len >> 8;	286	*enc->cur++ = len >> 8;
222	*enc->cur++ = len;	287	*enc->cur++ = len;
223	}	288	}
224	}	289	}
225		290
		291	// encodes a perl value into a CBOR integer
		292	ecb_inline
		293	encode_int (enc_t *enc, SV *sv)
		294	{
		295	if (SvIsUV (sv))
		296	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
		297	else if (SvIVX (sv) >= 0)
		298	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
		299	else
		300	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
		301	}
		302
226	ecb_inline void	303	ecb_inline void
227	encode_tag (enc_t *enc, UV tag)	304	encode_tag (enc_t *enc, UV tag)
228	{	305	{
229	encode_uint (enc, 0xc0, tag);	306	encode_uint (enc, MAJOR_TAG, tag);
230	}	307	}
231		308
232	static void	309	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		310	ecb_noinline static void
233	encode_str (enc_t *enc, int utf8, char *str, STRLEN len)	311	encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
234	{	312	{
		313	STRLEN ulen = len;
		314	U8 *p, *pend = (U8 *)str + len;
		315
		316	for (p = (U8 *)str; p < pend; ++p)
		317	ulen += *p >> 7; // count set high bits
		318
		319	encode_uint (enc, MAJOR_TEXT, ulen);
		320
		321	need (enc, ulen);
		322	for (p = (U8 *)str; p < pend; ++p)
		323	if (*p < 0x80)
		324	*enc->cur++ = *p;
		325	else
		326	{
		327	*enc->cur++ = 0xc0 + (*p >> 6);
		328	*enc->cur++ = 0x80 + (*p & 63);
		329	}
		330	}
		331
		332	ecb_inline void
		333	encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
		334	{
		335	if (ecb_expect_false (upgrade_utf8))
		336	if (!utf8)
		337	{
		338	encode_str_utf8 (enc, utf8, str, len);
		339	return;
		340	}
		341
		342	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		343	need (enc, len);
		344	memcpy (enc->cur, str, len);
		345	enc->cur += len;
		346	}
		347
		348	ecb_inline void
		349	encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
		350	{
235	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	351	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
236	{	352	{
237	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	353	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
238		354
239	if (SvOK (*svp))	355	if (SvOK (*svp))
240	{	356	{
241	// already registered, use stringref	357	// already registered, use stringref
242	encode_tag (enc, CBOR_TAG_STRINGREF);	358	encode_tag (enc, CBOR_TAG_STRINGREF);
243	encode_uint (enc, 0x00, SvUV (*svp));	359	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
244	return;	360	return;
245	}	361	}
246	else if (len >= minimum_string_length (enc->stringref_idx))	362	else if (len >= minimum_string_length (enc->stringref_idx))
247	{	363	{
248	// register only	364	// register only
249	sv_setuv (*svp, enc->stringref_idx);	365	sv_setuv (*svp, enc->stringref_idx);
250	++enc->stringref_idx;	366	++enc->stringref_idx;
251	}	367	}
252	}	368	}
253		369
254	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	370	encode_str (enc, upgrade_utf8, utf8, str, len);
		371	}
		372
		373	ecb_inline void
		374	encode_float16 (enc_t *enc, NV nv)
		375	{
255	need (enc, len);	376	need (enc, 1+2);
		377
		378	*enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
		379
		380	uint16_t fp = ecb_float_to_binary16 (nv);
		381
		382	if (!ecb_big_endian ())
		383	fp = ecb_bswap16 (fp);
		384
256	memcpy (enc->cur, str, len);	385	memcpy (enc->cur, &fp, 2);
257	enc->cur += len;	386	enc->cur += 2;
		387	}
		388
		389	ecb_inline void
		390	encode_float32 (enc_t *enc, NV nv)
		391	{
		392	need (enc, 1+4);
		393
		394	*enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
		395
		396	uint32_t fp = ecb_float_to_binary32 (nv);
		397
		398	if (!ecb_big_endian ())
		399	fp = ecb_bswap32 (fp);
		400
		401	memcpy (enc->cur, &fp, 4);
		402	enc->cur += 4;
		403	}
		404
		405	ecb_inline void
		406	encode_float64 (enc_t *enc, NV nv)
		407	{
		408	need (enc, 1+8);
		409
		410	*enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
		411
		412	uint64_t fp = ecb_double_to_binary64 (nv);
		413
		414	if (!ecb_big_endian ())
		415	fp = ecb_bswap64 (fp);
		416
		417	memcpy (enc->cur, &fp, 8);
		418	enc->cur += 8;
		419	}
		420
		421	ecb_inline void
		422	encode_forced (enc_t *enc, UV type, SV *sv)
		423	{
		424	switch (type)
		425	{
		426	case AS_CBOR:
		427	{
		428	STRLEN len;
		429	char *str = SvPVbyte (sv, len);
		430
		431	need (enc, len);
		432	memcpy (enc->cur, str, len);
		433	enc->cur += len;
		434	}
		435	break;
		436
		437	case AS_BYTES:
		438	{
		439	STRLEN len;
		440	char *str = SvPVbyte (sv, len);
		441	encode_strref (enc, 0, 0, str, len);
		442	}
		443	break;
		444
		445	case AS_TEXT:
		446	{
		447	STRLEN len;
		448	char *str = SvPVutf8 (sv, len);
		449	encode_strref (enc, 1, 1, str, len);
		450	}
		451	break;
		452
		453	case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
		454	case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
		455	case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
		456
		457	default:
		458	croak ("encountered malformed CBOR::XS::Tagged object");
		459	}
258	}	460	}
259		461
260	static void encode_sv (enc_t *enc, SV *sv);	462	static void encode_sv (enc_t *enc, SV *sv);
261		463
262	static void	464	static void
…		…
267	if (enc->depth >= enc->cbor.max_depth)	469	if (enc->depth >= enc->cbor.max_depth)
268	croak (ERR_NESTING_EXCEEDED);	470	croak (ERR_NESTING_EXCEEDED);
269		471
270	++enc->depth;	472	++enc->depth;
271		473
272	encode_uint (enc, 0x80, len + 1);	474	encode_uint (enc, MAJOR_ARRAY, len + 1);
273		475
		476	if (ecb_expect_false (SvMAGICAL (av)))
274	for (i = 0; i <= len; ++i)	477	for (i = 0; i <= len; ++i)
275	{	478	{
276	SV **svp = av_fetch (av, i, 0);	479	SV **svp = av_fetch (av, i, 0);
277	encode_sv (enc, svp ? *svp : &PL_sv_undef);	480	encode_sv (enc, svp ? *svp : &PL_sv_undef);
278	}	481	}
		482	else
		483	for (i = 0; i <= len; ++i)
		484	{
		485	SV *sv = AvARRAY (av)[i];
		486	encode_sv (enc, sv ? sv : &PL_sv_undef);
		487	}
279		488
280	--enc->depth;	489	--enc->depth;
281	}	490	}
282		491
283	static void	492	static void
…		…
291	++enc->depth;	500	++enc->depth;
292		501
293	int pairs = hv_iterinit (hv);	502	int pairs = hv_iterinit (hv);
294	int mg = SvMAGICAL (hv);	503	int mg = SvMAGICAL (hv);
295		504
296	if (mg)	505	if (ecb_expect_false (mg))
297	encode_ch (enc, 0xa0 | 31);	506	encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
298	else	507	else
299	encode_uint (enc, 0xa0, pairs);	508	encode_uint (enc, MAJOR_MAP, pairs);
300		509
301	while ((he = hv_iternext (hv)))	510	while ((he = hv_iternext (hv)))
302	{	511	{
303	if (HeKLEN (he) == HEf_SVKEY)	512	if (HeKLEN (he) == HEf_SVKEY)
304	encode_sv (enc, HeSVKEY (he));	513	encode_sv (enc, HeSVKEY (he));
305	else	514	else
306	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	515	encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
307		516
308	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	517	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
309	}	518	}
310		519
311	if (mg)	520	if (ecb_expect_false (mg))
312	encode_ch (enc, 0xe0 | 31);	521	encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
313		522
314	--enc->depth;	523	--enc->depth;
315	}	524	}
316		525
317	// encode objects, arrays and special \0=false and \1=true values.	526	// encode objects, arrays and special \0=false and \1=true values.
318	static void	527	static void
319	encode_rv (enc_t *enc, SV *sv)	528	encode_rv (enc_t *enc, SV *sv)
320	{	529	{
321	SvGETMAGIC (sv);	530	SvGETMAGIC (sv);
322
323	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
324	&& ecb_expect_false (SvREFCNT (sv) > 1))
325	{
326	if (!enc->shareable)
327	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
328
329	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
330
331	if (SvOK (*svp))
332	{
333	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
334	encode_uint (enc, 0x00, SvUV (*svp));
335	return;
336	}
337	else
338	{
339	sv_setuv (*svp, enc->shareable_idx);
340	++enc->shareable_idx;
341	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
342	}
343	}
344		531
345	svtype svt = SvTYPE (sv);	532	svtype svt = SvTYPE (sv);
346		533
347	if (ecb_expect_false (SvOBJECT (sv)))	534	if (ecb_expect_false (SvOBJECT (sv)))
348	{	535	{
…		…
355	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	542	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
356	? cbor_tagged_stash	543	? cbor_tagged_stash
357	: gv_stashpv ("CBOR::XS::Tagged" , 1);	544	: gv_stashpv ("CBOR::XS::Tagged" , 1);
358		545
359	HV *stash = SvSTASH (sv);	546	HV *stash = SvSTASH (sv);
360	GV *method;
361		547
362	if (stash == boolean_stash)	548	if (stash == boolean_stash)
363	encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);	549	{
		550	encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
		551	return;
		552	}
364	else if (stash == error_stash)	553	else if (stash == error_stash)
365	encode_ch (enc, 0xe0 | 23);	554	{
		555	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
		556	return;
		557	}
366	else if (stash == tagged_stash)	558	else if (stash == tagged_stash)
367	{	559	{
368	if (svt != SVt_PVAV)	560	if (svt != SVt_PVAV)
369	croak ("encountered CBOR::XS::Tagged object that isn't an array");	561	croak ("encountered CBOR::XS::Tagged object that isn't an array");
370		562
		563	switch (av_len ((AV *)sv))
		564	{
		565	case 2-1:
		566	// actually a tagged value
371	encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));	567	encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
372	encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));	568	encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
		569	break;
		570
		571	case 3-1:
		572	// a forced type [value, type, undef]
		573	encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
		574	break;
		575
		576	default:
		577	croak ("encountered malformed CBOR::XS::Tagged object");
		578	}
		579
		580	return;
		581	}
		582	}
		583
		584	if (ecb_expect_false (SvREFCNT (sv) > 1)
		585	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		586	{
		587	if (ecb_expect_false (!enc->shareable))
		588	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
		589
		590	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
		591
		592	if (SvOK (*svp))
373	}	593	{
		594	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		595	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		596	return;
		597	}
		598	else
		599	{
		600	sv_setuv (*svp, enc->shareable_idx);
		601	++enc->shareable_idx;
		602	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		603	}
		604	}
		605
		606	if (ecb_expect_false (SvOBJECT (sv)))
		607	{
		608	HV *stash = SvSTASH (sv);
		609	GV *method;
		610
		611	if (enc->cbor.flags & F_FORBID_OBJECTS)
		612	croak ("encountered object '%s', but forbid_objects is enabled",
		613	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
374	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	614	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
375	{	615	{
376	dSP;	616	dSP;
377		617
378	ENTER; SAVETMPS; PUSHMARK (SP);	618	ENTER; SAVETMPS;
		619	PUSHMARK (SP);
379	// we re-bless the reference to get overload and other niceties right	620	// we re-bless the reference to get overload and other niceties right
380	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	621	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
381		622
382	PUTBACK;	623	PUTBACK;
383	// G_SCALAR ensures that return value is 1	624	// G_SCALAR ensures that return value is 1
…		…
396	}	637	}
397	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	638	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
398	{	639	{
399	dSP;	640	dSP;
400		641
401	ENTER; SAVETMPS; PUSHMARK (SP);	642	ENTER; SAVETMPS;
		643	PUSHMARK (SP);
402	EXTEND (SP, 2);	644	EXTEND (SP, 2);
403	// we re-bless the reference to get overload and other niceties right	645	// we re-bless the reference to get overload and other niceties right
404	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	646	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
405	PUSHs (sv_cbor);	647	PUSHs (sv_cbor);
406		648
…		…
411	// catch this surprisingly common error	653	// catch this surprisingly common error
412	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	654	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
413	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	655	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
414		656
415	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	657	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
416	encode_uint (enc, 0x80, count + 1);	658	encode_uint (enc, MAJOR_ARRAY, count + 1);
417	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	659	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
418		660
419	while (count)	661	{
		662	int i;
		663
		664	for (i = 0; i < count; ++i)
420	encode_sv (enc, SP[1 - count--]);	665	encode_sv (enc, SP[i + 1 - count]);
		666
		667	SP -= count;
		668	}
421		669
422	PUTBACK;	670	PUTBACK;
423		671
424	FREETMPS; LEAVE;	672	FREETMPS; LEAVE;
425	}	673	}
…		…
443	{	691	{
444	double nv = SvNVX (sv);	692	double nv = SvNVX (sv);
445		693
446	need (enc, 9);	694	need (enc, 9);
447		695
448	if (ecb_expect_false (nv == (U32)nv))	696	if (ecb_expect_false (nv == (NV)(U32)nv))
449	encode_uint (enc, 0x00, (U32)nv);	697	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
450	//TODO: maybe I32?	698	//TODO: maybe I32?
451	else if (ecb_expect_false (nv == (float)nv))	699	else if (ecb_expect_false (nv == (float)nv))
452	{	700	encode_float32 (enc, nv);
453	uint32_t fp = ecb_float_to_binary32 (nv);
454
455	*enc->cur++ = 0xe0 | 26;
456
457	if (!ecb_big_endian ())
458	fp = ecb_bswap32 (fp);
459
460	memcpy (enc->cur, &fp, 4);
461	enc->cur += 4;
462	}
463	else	701	else
464	{	702	encode_float64 (enc, nv);
465	uint64_t fp = ecb_double_to_binary64 (nv);
466
467	*enc->cur++ = 0xe0 | 27;
468
469	if (!ecb_big_endian ())
470	fp = ecb_bswap64 (fp);
471
472	memcpy (enc->cur, &fp, 8);
473	enc->cur += 8;
474	}
475	}	703	}
476		704
477	static void	705	static void
478	encode_sv (enc_t *enc, SV *sv)	706	encode_sv (enc_t *enc, SV *sv)
479	{	707	{
…		…
481		709
482	if (SvPOKp (sv))	710	if (SvPOKp (sv))
483	{	711	{
484	STRLEN len;	712	STRLEN len;
485	char *str = SvPV (sv, len);	713	char *str = SvPV (sv, len);
486	encode_str (enc, SvUTF8 (sv), str, len);	714	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
487	}	715	}
488	else if (SvNOKp (sv))	716	else if (SvNOKp (sv))
489	encode_nv (enc, sv);	717	encode_nv (enc, sv);
490	else if (SvIOKp (sv))	718	else if (SvIOKp (sv))
491	{	719	encode_int (e,v sv);
492	if (SvIsUV (sv))
493	encode_uint (enc, 0x00, SvUVX (sv));
494	else if (SvIVX (sv) >= 0)
495	encode_uint (enc, 0x00, SvIVX (sv));
496	else
497	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
498	}
499	else if (SvROK (sv))	720	else if (SvROK (sv))
500	encode_rv (enc, SvRV (sv));	721	encode_rv (enc, SvRV (sv));
501	else if (!SvOK (sv))	722	else if (!SvOK (sv))
502	encode_ch (enc, 0xe0 | 22);	723	encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
503	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	724	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
504	encode_ch (enc, 0xe0 | 23);	725	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
505	else	726	else
506	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	727	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
507	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	728	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
508	}	729	}
509		730
510	static SV *	731	static SV *
511	encode_cbor (SV *scalar, CBOR *cbor)	732	encode_cbor (SV *scalar, CBOR *cbor)
512	{	733	{
513	enc_t enc = { };	734	enc_t enc = { 0 };
514		735
515	enc.cbor = *cbor;	736	enc.cbor = *cbor;
516	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	737	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
517	enc.cur = SvPVX (enc.sv);	738	enc.cur = SvPVX (enc.sv);
518	enc.end = SvEND (enc.sv);	739	enc.end = SvEND (enc.sv);
519		740
520	SvPOK_only (enc.sv);	741	SvPOK_only (enc.sv);
521		742
522	if (cbor->flags & F_ALLOW_STRINGREF)	743	if (cbor->flags & F_PACK_STRINGS)
523	{	744	{
524	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	745	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
525	enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());	746	enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
526	enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());	747	enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
527	}	748	}
…		…
550	U32 depth; // recursion depth	771	U32 depth; // recursion depth
551	U32 maxdepth; // recursion depth limit	772	U32 maxdepth; // recursion depth limit
552	AV *shareable;	773	AV *shareable;
553	AV *stringref;	774	AV *stringref;
554	SV *decode_tagged;	775	SV *decode_tagged;
		776	SV *err_sv; // optional sv for error, needs to be freed
555	} dec_t;	777	} dec_t;
556		778
557	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	779	// set dec->err to ERRSV
		780	ecb_cold static void
		781	err_errsv (dec_t *dec)
		782	{
		783	if (!dec->err)
		784	{
		785	dec->err_sv = newSVsv (ERRSV);
558		786
559	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	787	// chop off the trailing \n
		788	SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
		789	*SvEND (dec->err_sv) = 0;
560		790
		791	dec->err = SvPVutf8_nolen (dec->err_sv);
		792	}
		793	}
		794
		795	// the following functions are used to reduce code size and help the compiler to optimise
		796	ecb_cold static void
		797	err_set (dec_t *dec, const char *reason)
		798	{
		799	if (!dec->err)
		800	dec->err = reason;
		801	}
		802
		803	ecb_cold static void
		804	err_unexpected_end (dec_t *dec)
		805	{
		806	err_set (dec, "unexpected end of CBOR data");
		807	}
		808
		809	#define ERR_DO(do) SB do; goto fail; SE
		810	#define ERR(reason) ERR_DO (err_set (dec, reason))
		811	#define ERR_ERRSV ERR_DO (err_errsv (dec))
		812
		813	#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
		814
561	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	815	#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
562	#define DEC_DEC_DEPTH --dec->depth	816	#define DEC_DEC_DEPTH --dec->depth
563		817
564	static UV	818	static UV
565	decode_uint (dec_t *dec)	819	decode_uint (dec_t *dec)
566	{	820	{
567	switch (*dec->cur & 31)	821	U8 m = *dec->cur & MINOR_MASK;
568	{	822	++dec->cur;
569	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
570	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
571	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
572	return *dec->cur++ & 31;
573		823
574	case 24:	824	if (ecb_expect_true (m < LENGTH_EXT1))
		825	return m;
		826	else if (ecb_expect_true (m == LENGTH_EXT1))
		827	{
575	WANT (2);	828	WANT (1);
576	dec->cur += 2;	829	dec->cur += 1;
577	return dec->cur[-1];	830	return dec->cur[-1];
578		831	}
579	case 25:	832	else if (ecb_expect_true (m == LENGTH_EXT2))
		833	{
580	WANT (3);	834	WANT (2);
581	dec->cur += 3;	835	dec->cur += 2;
582	return (((UV)dec->cur[-2]) << 8)	836	return (((UV)dec->cur[-2]) << 8)
583	| ((UV)dec->cur[-1]);	837	| ((UV)dec->cur[-1]);
584		838	}
585	case 26:	839	else if (ecb_expect_true (m == LENGTH_EXT4))
		840	{
586	WANT (5);	841	WANT (4);
587	dec->cur += 5;	842	dec->cur += 4;
588	return (((UV)dec->cur[-4]) << 24)	843	return (((UV)dec->cur[-4]) << 24)
589	| (((UV)dec->cur[-3]) << 16)	844	| (((UV)dec->cur[-3]) << 16)
590	| (((UV)dec->cur[-2]) << 8)	845	| (((UV)dec->cur[-2]) << 8)
591	| ((UV)dec->cur[-1]);	846	| ((UV)dec->cur[-1]);
592		847	}
593	case 27:	848	else if (ecb_expect_true (m == LENGTH_EXT8))
		849	{
594	WANT (9);	850	WANT (8);
595	dec->cur += 9;	851	dec->cur += 8;
		852
		853	return
		854	#if UVSIZE < 8
		855	0
		856	#else
596	return (((UV)dec->cur[-8]) << 56)	857	(((UV)dec->cur[-8]) << 56)
597	| (((UV)dec->cur[-7]) << 48)	858	| (((UV)dec->cur[-7]) << 48)
598	| (((UV)dec->cur[-6]) << 40)	859	| (((UV)dec->cur[-6]) << 40)
599	| (((UV)dec->cur[-5]) << 32)	860	| (((UV)dec->cur[-5]) << 32)
		861	#endif
600	| (((UV)dec->cur[-4]) << 24)	862	| (((UV)dec->cur[-4]) << 24)
601	| (((UV)dec->cur[-3]) << 16)	863	| (((UV)dec->cur[-3]) << 16)
602	| (((UV)dec->cur[-2]) << 8)	864	| (((UV)dec->cur[-2]) << 8)
603	| ((UV)dec->cur[-1]);	865	| ((UV)dec->cur[-1]);
604		866	}
605	default:	867	else
606	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	868	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
607	}
608		869
609	fail:	870	fail:
610	return 0;	871	return 0;
611	}	872	}
612		873
…		…
617	{	878	{
618	AV *av = newAV ();	879	AV *av = newAV ();
619		880
620	DEC_INC_DEPTH;	881	DEC_INC_DEPTH;
621		882
622	if ((*dec->cur & 31) == 31)	883	if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
623	{	884	{
624	++dec->cur;	885	++dec->cur;
625		886
626	for (;;)	887	for (;;)
627	{	888	{
628	WANT (1);	889	WANT (1);
629		890
630	if (*dec->cur == (0xe0 | 31))	891	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
631	{	892	{
632	++dec->cur;	893	++dec->cur;
633	break;	894	break;
634	}	895	}
635		896
636	av_push (av, decode_sv (dec));	897	av_push (av, decode_sv (dec));
637	}	898	}
638	}	899	}
639	else	900	else
640	{	901	{
641	int i, len = decode_uint (dec);	902	UV i, len = decode_uint (dec);
642		903
		904	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
643	av_fill (av, len - 1);	905	av_fill (av, len - 1);
644		906
645	for (i = 0; i < len; ++i)	907	for (i = 0; i < len; ++i)
646	AvARRAY (av)[i] = decode_sv (dec);	908	AvARRAY (av)[i] = decode_sv (dec);
647	}	909	}
648		910
649	DEC_DEC_DEPTH;	911	DEC_DEC_DEPTH;
650	return newRV_noinc ((SV *)av);	912	return newRV_noinc ((SV *)av);
651		913
652	fail:	914	fail:
653	SvREFCNT_dec (av);	915	SvREFCNT_dec_NN (av);
654	DEC_DEC_DEPTH;	916	DEC_DEC_DEPTH;
655	return &PL_sv_undef;	917	return &PL_sv_undef;
656	}	918	}
657		919
658	static void	920	static void
…		…
660	{	922	{
661	// for speed reasons, we specialcase single-string	923	// for speed reasons, we specialcase single-string
662	// byte or utf-8 strings as keys, but only when !stringref	924	// byte or utf-8 strings as keys, but only when !stringref
663		925
664	if (ecb_expect_true (!dec->stringref))	926	if (ecb_expect_true (!dec->stringref))
665	if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)	927	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
666	{	928	{
667	I32 len = decode_uint (dec);	929	STRLEN len = decode_uint (dec);
668	char *key = (char *)dec->cur;	930	char *key = (char *)dec->cur;
669		931
		932	WANT (len);
670	dec->cur += len;	933	dec->cur += len;
671		934
672	if (ecb_expect_false (dec->stringref))
673	av_push (dec->stringref, newSVpvn (key, len));
674
675	hv_store (hv, key, len, decode_sv (dec), 0);	935	hv_store (hv, key, len, decode_sv (dec), 0);
676		936
677	return;	937	return;
678	}	938	}
679	else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)	939	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
680	{	940	{
681	I32 len = decode_uint (dec);	941	STRLEN len = decode_uint (dec);
682	char *key = (char *)dec->cur;	942	char *key = (char *)dec->cur;
683		943
		944	WANT (len);
684	dec->cur += len;	945	dec->cur += len;
685		946
686	if (ecb_expect_false (dec->stringref))	947	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
687	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	948	if (!is_utf8_string (key, len))
		949	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
688		950
689	hv_store (hv, key, -len, decode_sv (dec), 0);	951	hv_store (hv, key, -len, decode_sv (dec), 0);
690		952
691	return;	953	return;
692	}	954	}
693		955
694	SV *k = decode_sv (dec);	956	SV *k = decode_sv (dec);
695	SV *v = decode_sv (dec);	957	SV *v = decode_sv (dec);
696		958
		959	// we leak memory if uncaught exceptions are thrown by random magical
		960	// methods, and this is hopefully the only place where it can happen,
		961	// so if there is a chance of an exception, take the very slow path.
		962	// since catching exceptions is "undocumented/internal/forbidden" by
		963	// the new p5p powers, we need to call out to a perl function :/
		964	if (ecb_expect_false (SvAMAGIC (k)))
		965	{
		966	dSP;
		967
		968	ENTER; SAVETMPS;
		969	PUSHMARK (SP);
		970	EXTEND (SP, 3);
		971	PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
		972	PUSHs (sv_2mortal (k));
		973	PUSHs (sv_2mortal (v));
		974
		975	PUTBACK;
		976	call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
		977	SPAGAIN;
		978
		979	FREETMPS; LEAVE;
		980
		981	if (SvTRUE (ERRSV))
		982	ERR_ERRSV;
		983
		984	return;
		985	}
		986
697	hv_store_ent (hv, k, v, 0);	987	hv_store_ent (hv, k, v, 0);
698	SvREFCNT_dec (k);	988	SvREFCNT_dec_NN (k);
		989
		990	fail:
		991	;
699	}	992	}
700		993
701	static SV *	994	static SV *
702	decode_hv (dec_t *dec)	995	decode_hv (dec_t *dec)
703	{	996	{
704	HV *hv = newHV ();	997	HV *hv = newHV ();
705		998
706	DEC_INC_DEPTH;	999	DEC_INC_DEPTH;
707		1000
708	if ((*dec->cur & 31) == 31)	1001	if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
709	{	1002	{
710	++dec->cur;	1003	++dec->cur;
711		1004
712	for (;;)	1005	for (;;)
713	{	1006	{
714	WANT (1);	1007	WANT (1);
715		1008
716	if (*dec->cur == (0xe0 | 31))	1009	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
717	{	1010	{
718	++dec->cur;	1011	++dec->cur;
719	break;	1012	break;
720	}	1013	}
721		1014
722	decode_he (dec, hv);	1015	decode_he (dec, hv);
723	}	1016	}
724	}	1017	}
725	else	1018	else
726	{	1019	{
727	int pairs = decode_uint (dec);	1020	UV pairs = decode_uint (dec);
		1021
		1022	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
728		1023
729	while (pairs--)	1024	while (pairs--)
730	decode_he (dec, hv);	1025	decode_he (dec, hv);
731	}	1026	}
732		1027
733	DEC_DEC_DEPTH;	1028	DEC_DEC_DEPTH;
734	return newRV_noinc ((SV *)hv);	1029	return newRV_noinc ((SV *)hv);
735		1030
736	fail:	1031	fail:
737	SvREFCNT_dec (hv);	1032	SvREFCNT_dec_NN (hv);
738	DEC_DEC_DEPTH;	1033	DEC_DEC_DEPTH;
739	return &PL_sv_undef;	1034	return &PL_sv_undef;
740	}	1035	}
741		1036
742	static SV *	1037	static SV *
743	decode_str (dec_t *dec, int utf8)	1038	decode_str (dec_t *dec, int utf8)
744	{	1039	{
745	SV *sv = 0;	1040	SV *sv = 0;
746		1041
747	if ((*dec->cur & 31) == 31)	1042	if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
748	{	1043	{
		1044	// indefinite length strings
749	++dec->cur;	1045	++dec->cur;
750		1046
		1047	U8 major = *dec->cur & MAJOR_MISC;
		1048
751	sv = newSVpvn ("", 0);	1049	sv = newSVpvn ("", 0);
752		1050
753	// not very fast, and certainly not robust against illegal input
754	for (;;)	1051	for (;;)
755	{	1052	{
756	WANT (1);	1053	WANT (1);
757		1054
758	if (*dec->cur == (0xe0 | 31))	1055	if ((*dec->cur - major) > LENGTH_EXT8)
		1056	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
759	{	1057	{
760	++dec->cur;	1058	++dec->cur;
761	break;	1059	break;
762	}	1060	}
		1061	else
		1062	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
763		1063
764	sv_catsv (sv, decode_sv (dec));	1064	STRLEN len = decode_uint (dec);
		1065
		1066	WANT (len);
		1067	sv_catpvn (sv, dec->cur, len);
		1068	dec->cur += len;
765	}	1069	}
766	}	1070	}
767	else	1071	else
768	{	1072	{
769	STRLEN len = decode_uint (dec);	1073	STRLEN len = decode_uint (dec);
…		…
776	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	1080	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
777	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	1081	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
778	}	1082	}
779		1083
780	if (utf8)	1084	if (utf8)
		1085	{
		1086	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		1087	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		1088	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		1089
781	SvUTF8_on (sv);	1090	SvUTF8_on (sv);
		1091	}
782		1092
783	return sv;	1093	return sv;
784		1094
785	fail:	1095	fail:
786	SvREFCNT_dec (sv);	1096	SvREFCNT_dec (sv);
…		…
805	sv = newRV_noinc (decode_sv (dec));	1115	sv = newRV_noinc (decode_sv (dec));
806	break;	1116	break;
807		1117
808	case CBOR_TAG_STRINGREF_NAMESPACE:	1118	case CBOR_TAG_STRINGREF_NAMESPACE:
809	{	1119	{
		1120	// do not use SAVETMPS/FREETMPS, as these will
		1121	// erase mortalised caches, e.g. "shareable"
810	ENTER; SAVETMPS;	1122	ENTER;
811		1123
812	SAVESPTR (dec->stringref);	1124	SAVESPTR (dec->stringref);
813	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());	1125	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
814		1126
815	sv = decode_sv (dec);	1127	sv = decode_sv (dec);
816		1128
817	FREETMPS; LEAVE;	1129	LEAVE;
818	}	1130	}
819	break;	1131	break;
820		1132
821	case CBOR_TAG_STRINGREF:	1133	case CBOR_TAG_STRINGREF:
822	{	1134	{
823	if ((*dec->cur >> 5) != 0)	1135	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
824	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1136	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
825		1137
826	UV idx = decode_uint (dec);	1138	UV idx = decode_uint (dec);
827		1139
828	if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))	1140	if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
829	ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");	1141	ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
830		1142
831	sv = newSVsv (AvARRAY (dec->stringref)[idx]);	1143	sv = newSVsv (AvARRAY (dec->stringref)[idx]);
832	}	1144	}
833	break;	1145	break;
…		…
835	case CBOR_TAG_VALUE_SHAREABLE:	1147	case CBOR_TAG_VALUE_SHAREABLE:
836	{	1148	{
837	if (ecb_expect_false (!dec->shareable))	1149	if (ecb_expect_false (!dec->shareable))
838	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());	1150	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
839		1151
		1152	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1153	{
840	sv = newSV (0);	1154	sv = newSV (0);
841	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1155	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
842		1156
843	SV *osv = decode_sv (dec);	1157	SV *osv = decode_sv (dec);
844	sv_setsv (sv, osv);	1158	sv_setsv (sv, osv);
845	SvREFCNT_dec_NN (osv);	1159	SvREFCNT_dec_NN (osv);
		1160	}
		1161	else
		1162	{
		1163	av_push (dec->shareable, &PL_sv_undef);
		1164	int idx = AvFILLp (dec->shareable);
		1165	sv = decode_sv (dec);
		1166	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		1167	}
846	}	1168	}
847	break;	1169	break;
848		1170
849	case CBOR_TAG_VALUE_SHAREDREF:	1171	case CBOR_TAG_VALUE_SHAREDREF:
850	{	1172	{
851	if ((*dec->cur >> 5) != 0)	1173	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
852	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1174	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
853		1175
854	UV idx = decode_uint (dec);	1176	UV idx = decode_uint (dec);
855		1177
856	if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))	1178	if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
857	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1179	ERR ("corrupted CBOR data (sharedref index out of bounds)");
858		1180
859	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1181	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1182
		1183	if (sv == &PL_sv_undef)
		1184	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
860	}	1185	}
861	break;	1186	break;
862		1187
863	case CBOR_TAG_PERL_OBJECT:	1188	case CBOR_TAG_PERL_OBJECT:
864	{	1189	{
		1190	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1191	goto filter;
		1192
865	sv = decode_sv (dec);	1193	sv = decode_sv (dec);
866		1194
867	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)	1195	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
868	ERR ("corrupted CBOR data (non-array perl object)");	1196	ERR ("corrupted CBOR data (non-array perl object)");
869		1197
…		…
879	if (!method)	1207	if (!method)
880	ERR ("cannot decode perl-object (package does not have a THAW method)");	1208	ERR ("cannot decode perl-object (package does not have a THAW method)");
881		1209
882	dSP;	1210	dSP;
883		1211
884	ENTER; SAVETMPS; PUSHMARK (SP);	1212	ENTER; SAVETMPS;
		1213	PUSHMARK (SP);
885	EXTEND (SP, len + 1);	1214	EXTEND (SP, len + 1);
886	// we re-bless the reference to get overload and other niceties right	1215	// we re-bless the reference to get overload and other niceties right
887	PUSHs (*av_fetch (av, 0, 1));	1216	PUSHs (*av_fetch (av, 0, 1));
888	PUSHs (sv_cbor);	1217	PUSHs (sv_cbor);
889		1218
…		…
897	SPAGAIN;	1226	SPAGAIN;
898		1227
899	if (SvTRUE (ERRSV))	1228	if (SvTRUE (ERRSV))
900	{	1229	{
901	FREETMPS; LEAVE;	1230	FREETMPS; LEAVE;
902	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1231	ERR_ERRSV;
903	}	1232	}
904		1233
905	SvREFCNT_dec (sv);	1234	SvREFCNT_dec_NN (sv);
906	sv = SvREFCNT_inc (POPs);	1235	sv = SvREFCNT_inc (POPs);
907		1236
908	PUTBACK;	1237	PUTBACK;
909		1238
910	FREETMPS; LEAVE;	1239	FREETMPS; LEAVE;
911	}	1240	}
912	break;	1241	break;
913		1242
914	default:	1243	default:
		1244	filter:
915	{	1245	{
		1246	SV *tag_sv = newSVuv (tag);
		1247
916	sv = decode_sv (dec);	1248	sv = decode_sv (dec);
917		1249
918	dSP;	1250	dSP;
919	ENTER; SAVETMPS; PUSHMARK (SP);	1251	ENTER; SAVETMPS;
		1252	PUSHMARK (SP);
920	EXTEND (SP, 2);	1253	EXTEND (SP, 2);
921	PUSHs (newSVuv (tag));	1254	PUSHs (tag_sv);
922	PUSHs (sv);	1255	PUSHs (sv);
923		1256
924	PUTBACK;	1257	PUTBACK;
925	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);	1258	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
926	SPAGAIN;	1259	SPAGAIN;
927		1260
928	if (SvTRUE (ERRSV))	1261	if (SvTRUE (ERRSV))
929	{	1262	{
		1263	SvREFCNT_dec_NN (tag_sv);
930	FREETMPS; LEAVE;	1264	FREETMPS; LEAVE;
931	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1265	ERR_ERRSV;
932	}	1266	}
933		1267
934	if (count)	1268	if (count)
935	{	1269	{
		1270	SvREFCNT_dec_NN (tag_sv);
936	SvREFCNT_dec (sv);	1271	SvREFCNT_dec_NN (sv);
937	sv = SvREFCNT_inc (POPs);	1272	sv = SvREFCNT_inc_NN (TOPs);
		1273	SP -= count;
938	}	1274	}
939	else	1275	else
940	{	1276	{
941	AV *av = newAV ();	1277	AV *av = newAV ();
942	av_push (av, newSVuv (tag));	1278	av_push (av, tag_sv);
943	av_push (av, sv);	1279	av_push (av, sv);
944		1280
945	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	1281	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
946	? cbor_tagged_stash	1282	? cbor_tagged_stash
947	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1283	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
965	static SV *	1301	static SV *
966	decode_sv (dec_t *dec)	1302	decode_sv (dec_t *dec)
967	{	1303	{
968	WANT (1);	1304	WANT (1);
969		1305
970	switch (*dec->cur >> 5)	1306	switch (*dec->cur >> MAJOR_SHIFT)
971	{	1307	{
972	case 0: // unsigned int	1308	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
973	return newSVuv (decode_uint (dec));	1309	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
974	case 1: // negative int	1310	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
975	return newSViv (-1 - (IV)decode_uint (dec));	1311	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
976	case 2: // octet string	1312	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
977	return decode_str (dec, 0);	1313	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
978	case 3: // utf-8 string	1314	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
979	return decode_str (dec, 1);	1315
980	case 4: // array	1316	case MAJOR_MISC >> MAJOR_SHIFT:
981	return decode_av (dec);
982	case 5: // map
983	return decode_hv (dec);
984	case 6: // tag
985	return decode_tagged (dec);
986	case 7: // misc
987	switch (*dec->cur++ & 31)	1317	switch (*dec->cur++ & MINOR_MASK)
988	{	1318	{
989	case 20:	1319	case SIMPLE_FALSE:
990	#if CBOR_SLOW	1320	#if CBOR_SLOW
991	types_false = get_bool ("Types::Serialiser::false");	1321	types_false = get_bool ("Types::Serialiser::false");
992	#endif	1322	#endif
993	return newSVsv (types_false);	1323	return newSVsv (types_false);
994	case 21:	1324	case SIMPLE_TRUE:
995	#if CBOR_SLOW	1325	#if CBOR_SLOW
996	types_true = get_bool ("Types::Serialiser::true");	1326	types_true = get_bool ("Types::Serialiser::true");
997	#endif	1327	#endif
998	return newSVsv (types_true);	1328	return newSVsv (types_true);
999	case 22:	1329	case SIMPLE_NULL:
1000	return newSVsv (&PL_sv_undef);	1330	return newSVsv (&PL_sv_undef);
1001	case 23:	1331	case SIMPLE_UNDEF:
1002	#if CBOR_SLOW	1332	#if CBOR_SLOW
1003	types_error = get_bool ("Types::Serialiser::error");	1333	types_error = get_bool ("Types::Serialiser::error");
1004	#endif	1334	#endif
1005	return newSVsv (types_error);	1335	return newSVsv (types_error);
1006		1336
1007	case 25:	1337	case MISC_FLOAT16:
1008	{	1338	{
1009	WANT (2);	1339	WANT (2);
1010		1340
1011	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];	1341	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1012	dec->cur += 2;	1342	dec->cur += 2;
1013		1343
1014	return newSVnv (ecb_binary16_to_float (fp));	1344	return newSVnv (ecb_binary16_to_float (fp));
1015	}	1345	}
1016		1346
1017	case 26:	1347	case MISC_FLOAT32:
1018	{	1348	{
1019	uint32_t fp;	1349	uint32_t fp;
1020	WANT (4);	1350	WANT (4);
1021	memcpy (&fp, dec->cur, 4);	1351	memcpy (&fp, dec->cur, 4);
1022	dec->cur += 4;	1352	dec->cur += 4;
…		…
1025	fp = ecb_bswap32 (fp);	1355	fp = ecb_bswap32 (fp);
1026		1356
1027	return newSVnv (ecb_binary32_to_float (fp));	1357	return newSVnv (ecb_binary32_to_float (fp));
1028	}	1358	}
1029		1359
1030	case 27:	1360	case MISC_FLOAT64:
1031	{	1361	{
1032	uint64_t fp;	1362	uint64_t fp;
1033	WANT (8);	1363	WANT (8);
1034	memcpy (&fp, dec->cur, 8);	1364	memcpy (&fp, dec->cur, 8);
1035	dec->cur += 8;	1365	dec->cur += 8;
…		…
1038	fp = ecb_bswap64 (fp);	1368	fp = ecb_bswap64 (fp);
1039		1369
1040	return newSVnv (ecb_binary64_to_double (fp));	1370	return newSVnv (ecb_binary64_to_double (fp));
1041	}	1371	}
1042		1372
1043	// 0..19 unassigned	1373	// 0..19 unassigned simple
1044	// 24 reserved + unassigned (reserved values are not encodable)	1374	// 24 reserved + unassigned simple (reserved values are not encodable)
		1375	// 28-30 unassigned misc
		1376	// 31 break code
1045	default:	1377	default:
1046	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1378	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1047	}	1379	}
1048		1380
1049	break;	1381	break;
1050	}	1382	}
1051		1383
…		…
1054	}	1386	}
1055		1387
1056	static SV *	1388	static SV *
1057	decode_cbor (SV *string, CBOR *cbor, char **offset_return)	1389	decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1058	{	1390	{
1059	dec_t dec = { };	1391	dec_t dec = { 0 };
1060	SV *sv;	1392	SV *sv;
1061	STRLEN len;	1393	STRLEN len;
1062	char *data = SvPVbyte (string, len);	1394	char *data = SvPVbyte (string, len);
1063		1395
1064	if (len > cbor->max_size && cbor->max_size)	1396	if (len > cbor->max_size && cbor->max_size)
…		…
1078	if (dec.cur != dec.end && !dec.err)	1410	if (dec.cur != dec.end && !dec.err)
1079	dec.err = "garbage after CBOR object";	1411	dec.err = "garbage after CBOR object";
1080		1412
1081	if (dec.err)	1413	if (dec.err)
1082	{	1414	{
		1415	if (dec.shareable)
		1416	{
		1417	// need to break cyclic links, which would all be in shareable
		1418	int i;
		1419	SV **svp;
		1420
		1421	for (i = av_len (dec.shareable) + 1; i--; )
		1422	if ((svp = av_fetch (dec.shareable, i, 0)))
		1423	sv_setsv (*svp, &PL_sv_undef);
		1424	}
		1425
1083	SvREFCNT_dec (sv);	1426	SvREFCNT_dec_NN (sv);
		1427
		1428	if (dec.err_sv)
		1429	sv_2mortal (dec.err_sv);
		1430
1084	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);	1431	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1085	}	1432	}
1086		1433
1087	sv = sv_2mortal (sv);	1434	sv = sv_2mortal (sv);
1088		1435
1089	return sv;	1436	return sv;
1090	}	1437	}
1091		1438
		1439	/////////////////////////////////////////////////////////////////////////////
		1440	// incremental parser
		1441
		1442	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1443
		1444	// returns 0 for notyet, 1 for success or error
		1445	static int
		1446	incr_parse (CBOR *self, SV *cborstr)
		1447	{
		1448	STRLEN cur;
		1449	SvPV (cborstr, cur);
		1450
		1451	while (ecb_expect_true (self->incr_need <= cur))
		1452	{
		1453	// table of integer count bytes
		1454	static I8 incr_len[MINOR_MASK + 1] = {
		1455	0, 0, 0, 0, 0, 0, 0, 0,
		1456	0, 0, 0, 0, 0, 0, 0, 0,
		1457	0, 0, 0, 0, 0, 0, 0, 0,
		1458	1, 2, 4, 8,-1,-1,-1,-2
		1459	};
		1460
		1461	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1462	U8 m = *p & MINOR_MASK;
		1463	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1464	I8 ilen = incr_len[m];
		1465
		1466	self->incr_need = self->incr_pos + 1;
		1467
		1468	if (ecb_expect_false (ilen < 0))
		1469	{
		1470	if (m != MINOR_INDEF)
		1471	return 1; // error
		1472
		1473	if (*p == (MAJOR_MISC | MINOR_INDEF))
		1474	{
		1475	if (count >= 0)
		1476	return 1; // error
		1477
		1478	count = 1;
		1479	}
		1480	else
		1481	{
		1482	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1483	count = -1;
		1484	}
		1485	}
		1486	else
		1487	{
		1488	self->incr_need += ilen;
		1489	if (ecb_expect_false (self->incr_need > cur))
		1490	return 0;
		1491
		1492	int major = *p >> MAJOR_SHIFT;
		1493
		1494	switch (major)
		1495	{
		1496	case MAJOR_TAG >> MAJOR_SHIFT:
		1497	++count; // tags merely prefix another value
		1498	break;
		1499
		1500	case MAJOR_BYTES >> MAJOR_SHIFT:
		1501	case MAJOR_TEXT >> MAJOR_SHIFT:
		1502	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1503	case MAJOR_MAP >> MAJOR_SHIFT:
		1504	{
		1505	UV len;
		1506
		1507	if (ecb_expect_false (ilen))
		1508	{
		1509	len = 0;
		1510
		1511	do {
		1512	len = (len << 8) | *++p;
		1513	} while (--ilen);
		1514	}
		1515	else
		1516	len = m;
		1517
		1518	switch (major)
		1519	{
		1520	case MAJOR_BYTES >> MAJOR_SHIFT:
		1521	case MAJOR_TEXT >> MAJOR_SHIFT:
		1522	self->incr_need += len;
		1523	if (ecb_expect_false (self->incr_need > cur))
		1524	return 0;
		1525
		1526	break;
		1527
		1528	case MAJOR_MAP >> MAJOR_SHIFT:
		1529	len <<= 1;
		1530	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1531	if (len)
		1532	{
		1533	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1534	count = len + 1;
		1535	}
		1536	break;
		1537	}
		1538	}
		1539	}
		1540	}
		1541
		1542	self->incr_pos = self->incr_need;
		1543
		1544	if (count > 0)
		1545	{
		1546	while (!--count)
		1547	{
		1548	if (!AvFILLp (self->incr_count))
		1549	return 1; // done
		1550
		1551	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1552	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1553	}
		1554
		1555	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1556	}
		1557	}
		1558
		1559	return 0;
		1560	}
		1561
		1562
1092	/////////////////////////////////////////////////////////////////////////////	1563	/////////////////////////////////////////////////////////////////////////////
1093	// XS interface functions	1564	// XS interface functions
1094		1565
1095	MODULE = CBOR::XS PACKAGE = CBOR::XS	1566	MODULE = CBOR::XS PACKAGE = CBOR::XS
1096		1567
…		…
1108		1579
1109	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1580	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1110		1581
1111	sv_cbor = newSVpv ("CBOR", 0);	1582	sv_cbor = newSVpv ("CBOR", 0);
1112	SvREADONLY_on (sv_cbor);	1583	SvREADONLY_on (sv_cbor);
		1584
		1585	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1113	}	1586	}
1114		1587
1115	PROTOTYPES: DISABLE	1588	PROTOTYPES: DISABLE
1116		1589
1117	void CLONE (...)	1590	void CLONE (...)
…		…
1136	void shrink (CBOR *self, int enable = 1)	1609	void shrink (CBOR *self, int enable = 1)
1137	ALIAS:	1610	ALIAS:
1138	shrink = F_SHRINK	1611	shrink = F_SHRINK
1139	allow_unknown = F_ALLOW_UNKNOWN	1612	allow_unknown = F_ALLOW_UNKNOWN
1140	allow_sharing = F_ALLOW_SHARING	1613	allow_sharing = F_ALLOW_SHARING
1141	allow_stringref = F_ALLOW_STRINGREF	1614	allow_cycles = F_ALLOW_CYCLES
		1615	forbid_objects = F_FORBID_OBJECTS
		1616	pack_strings = F_PACK_STRINGS
		1617	text_keys = F_TEXT_KEYS
		1618	text_strings = F_TEXT_STRINGS
		1619	validate_utf8 = F_VALIDATE_UTF8
1142	PPCODE:	1620	PPCODE:
1143	{	1621	{
1144	if (enable)	1622	if (enable)
1145	self->flags |= ix;	1623	self->flags |= ix;
1146	else	1624	else
…		…
1152	void get_shrink (CBOR *self)	1630	void get_shrink (CBOR *self)
1153	ALIAS:	1631	ALIAS:
1154	get_shrink = F_SHRINK	1632	get_shrink = F_SHRINK
1155	get_allow_unknown = F_ALLOW_UNKNOWN	1633	get_allow_unknown = F_ALLOW_UNKNOWN
1156	get_allow_sharing = F_ALLOW_SHARING	1634	get_allow_sharing = F_ALLOW_SHARING
1157	get_allow_stringref = F_ALLOW_STRINGREF	1635	get_allow_cycles = F_ALLOW_CYCLES
		1636	get_forbid_objects = F_FORBID_OBJECTS
		1637	get_pack_strings = F_PACK_STRINGS
		1638	get_text_keys = F_TEXT_KEYS
		1639	get_text_strings = F_TEXT_STRINGS
		1640	get_validate_utf8 = F_VALIDATE_UTF8
1158	PPCODE:	1641	PPCODE:
1159	XPUSHs (boolSV (self->flags & ix));	1642	XPUSHs (boolSV (self->flags & ix));
1160		1643
1161	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1644	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1162	PPCODE:	1645	PPCODE:
…		…
1211	EXTEND (SP, 2);	1694	EXTEND (SP, 2);
1212	PUSHs (sv);	1695	PUSHs (sv);
1213	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1696	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1214	}	1697	}
1215		1698
		1699	void incr_parse (CBOR *self, SV *cborstr)
		1700	ALIAS:
		1701	incr_parse_multiple = 1
		1702	PPCODE:
		1703	{
		1704	if (SvUTF8 (cborstr))
		1705	sv_utf8_downgrade (cborstr, 0);
		1706
		1707	if (!self->incr_count)
		1708	{
		1709	self->incr_count = newAV ();
		1710	self->incr_pos = 0;
		1711	self->incr_need = 1;
		1712
		1713	av_push (self->incr_count, newSViv (1));
		1714	}
		1715
		1716	do
		1717	{
		1718	if (!incr_parse (self, cborstr))
		1719	{
		1720	if (self->incr_need > self->max_size && self->max_size)
		1721	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1722	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1723
		1724	break;
		1725	}
		1726
		1727	SV *sv;
		1728	char *offset;
		1729
		1730	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1731	XPUSHs (sv);
		1732
		1733	sv_chop (cborstr, offset);
		1734
		1735	av_clear (self->incr_count);
		1736	av_push (self->incr_count, newSViv (1));
		1737
		1738	self->incr_pos = 0;
		1739	self->incr_need = self->incr_pos + 1;
		1740	}
		1741	while (ix);
		1742	}
		1743
		1744	void incr_reset (CBOR *self)
		1745	CODE:
		1746	{
		1747	SvREFCNT_dec (self->incr_count);
		1748	self->incr_count = 0;
		1749	}
		1750
1216	void DESTROY (CBOR *self)	1751	void DESTROY (CBOR *self)
1217	PPCODE:	1752	PPCODE:
1218	cbor_free (self);	1753	cbor_free (self);
1219		1754
1220	PROTOTYPES: ENABLE	1755	PROTOTYPES: ENABLE
1221		1756
1222	void encode_cbor (SV *scalar)	1757	void encode_cbor (SV *scalar)
		1758	ALIAS:
		1759	encode_cbor = 0
		1760	encode_cbor_sharing = F_ALLOW_SHARING
1223	PPCODE:	1761	PPCODE:
1224	{	1762	{
1225	CBOR cbor;	1763	CBOR cbor;
1226	cbor_init (&cbor);	1764	cbor_init (&cbor);
		1765	cbor.flags |= ix;
1227	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1766	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1228	XPUSHs (scalar);	1767	XPUSHs (scalar);
1229	}	1768	}
1230		1769
1231	void decode_cbor (SV *cborstr)	1770	void decode_cbor (SV *cborstr)
…		…
1235	cbor_init (&cbor);	1774	cbor_init (&cbor);
1236	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;	1775	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1237	XPUSHs (cborstr);	1776	XPUSHs (cborstr);
1238	}	1777	}
1239		1778
		1779	#ifdef __AFL_COMPILER
		1780
		1781	void
		1782	afl_init ()
		1783	CODE:
		1784	__AFL_INIT ();
		1785
		1786	int
		1787	afl_loop (unsigned int count = 10000)
		1788	CODE:
		1789	RETVAL = __AFL_LOOP (count);
		1790	OUTPUT:
		1791	RETVAL
		1792
		1793	#endif
		1794

Diff Legend

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