ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.34 by root, Sat Nov 30 16:19:59 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
24	#ifndef SvREFCNT_dec_NN	29	#ifndef SvREFCNT_dec_NN
25	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)	30	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26	#endif	31	#endif
27		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	};
		69
28	// known tags	70	// known tags
29	enum cbor_tag	71	enum cbor_tag
30	{	72	{
31	// extensions	73	// extensions
32	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref	74	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
33	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object	75	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
34	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object	76	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
35	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing	77	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
36	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing	78	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
37	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref	79	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
38	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	80	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
39		81
40	// rfc7049	82	// rfc7049
41	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	83	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
42	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	84	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
43	CBOR_TAG_POS_BIGNUM = 2, // byte string	85	CBOR_TAG_POS_BIGNUM = 2, // byte string
44	CBOR_TAG_NEG_BIGNUM = 3, // byte string	86	CBOR_TAG_NEG_BIGNUM = 3, // byte string
45	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	87	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
46	CBOR_TAG_BIGFLOAT = 5, // array	88	CBOR_TAG_BIGFLOAT = 5, // array
47		89
48	CBOR_TAG_CONV_B64U = 21, // base64url, any	90	CBOR_TAG_CONV_B64U = 21, // base64url, any
49	CBOR_TAG_CONV_B64 = 22, // base64, any	91	CBOR_TAG_CONV_B64 = 22, // base64, any
50	CBOR_TAG_CONV_HEX = 23, // base16, any	92	CBOR_TAG_CONV_HEX = 23, // base16, any
51	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	93	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
52		94
53	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	95	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
54	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	96	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
55	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	97	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
56	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	98	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
57	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	99	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
58		100
59	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
60	};	114	};
61		115
62	#define F_SHRINK 0x00000001UL	116	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	117	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL	118	#define F_ALLOW_SHARING 0x00000004UL
		119	#define F_ALLOW_CYCLES 0x00000008UL
		120	#define F_FORBID_OBJECTS 0x00000010UL
65	#define F_PACK_STRINGS 0x00000008UL	121	#define F_PACK_STRINGS 0x00000020UL
		122	#define F_TEXT_KEYS 0x00000040UL
		123	#define F_TEXT_STRINGS 0x00000080UL
		124	#define F_VALIDATE_UTF8 0x00000100UL
66		125
67	#define INIT_SIZE 32 // initial scalar size to be allocated	126	#define INIT_SIZE 32 // initial scalar size to be allocated
68		127
69	#define SB do {	128	#define SB do {
70	#define SE } while (0)	129	#define SE } while (0)
…		…
89	typedef struct {	148	typedef struct {
90	U32 flags;	149	U32 flags;
91	U32 max_depth;	150	U32 max_depth;
92	STRLEN max_size;	151	STRLEN max_size;
93	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.
94	} CBOR;	158	} CBOR;
95		159
96	ecb_inline void	160	ecb_inline void
97	cbor_init (CBOR *cbor)	161	cbor_init (CBOR *cbor)
98	{	162	{
…		…
102		166
103	ecb_inline void	167	ecb_inline void
104	cbor_free (CBOR *cbor)	168	cbor_free (CBOR *cbor)
105	{	169	{
106	SvREFCNT_dec (cbor->filter);	170	SvREFCNT_dec (cbor->filter);
		171	SvREFCNT_dec (cbor->incr_count);
107	}	172	}
108		173
109	/////////////////////////////////////////////////////////////////////////////	174	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	175	// utility functions
111		176
…		…
137		202
138	// minimum length of a string to be registered for stringref	203	// minimum length of a string to be registered for stringref
139	ecb_inline int	204	ecb_inline int
140	minimum_string_length (UV idx)	205	minimum_string_length (UV idx)
141	{	206	{
142	return idx > 23	207	return idx <= 23 ? 3
143	? idx > 0xffU	208	: idx <= 0xffU ? 4
144	? idx > 0xffffU	209	: idx <= 0xffffU ? 5
145	? idx > 0xffffffffU	210	: idx <= 0xffffffffU ? 7
146	? 11	211	: 11;
147	: 7
148	: 5
149	: 4
150	: 3;
151	}	212	}
152		213
153	/////////////////////////////////////////////////////////////////////////////	214	/////////////////////////////////////////////////////////////////////////////
154	// encoder	215	// encoder
155		216
…		…
168	} enc_t;	229	} enc_t;
169		230
170	ecb_inline void	231	ecb_inline void
171	need (enc_t *enc, STRLEN len)	232	need (enc_t *enc, STRLEN len)
172	{	233	{
173	if (ecb_expect_false (enc->cur + len >= enc->end))	234	if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
174	{	235	{
175	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);	236	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
176	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);	237	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
177	enc->cur = SvPVX (enc->sv) + cur;	238	enc->cur = SvPVX (enc->sv) + cur;
178	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;	239	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
…		…
184	{	245	{
185	need (enc, 1);	246	need (enc, 1);
186	*enc->cur++ = ch;	247	*enc->cur++ = ch;
187	}	248	}
188		249
		250	// used for tags, intregers, element counts and so on
189	static void	251	static void
190	encode_uint (enc_t *enc, int major, UV len)	252	encode_uint (enc_t *enc, int major, UV len)
191	{	253	{
192	need (enc, 9);	254	need (enc, 9);
193		255
194	if (ecb_expect_true (len < 24))	256	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major | len;	257	*enc->cur++ = major | len;
196	else if (ecb_expect_true (len <= 0xff))	258	else if (ecb_expect_true (len <= 0xffU))
197	{	259	{
198	*enc->cur++ = major | 24;	260	*enc->cur++ = major | LENGTH_EXT1;
199	*enc->cur++ = len;	261	*enc->cur++ = len;
200	}	262	}
201	else if (len <= 0xffff)	263	else if (len <= 0xffffU)
202	{	264	{
203	*enc->cur++ = major | 25;	265	*enc->cur++ = major | LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	266	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	267	*enc->cur++ = len;
206	}	268	}
207	else if (len <= 0xffffffff)	269	else if (len <= 0xffffffffU)
208	{	270	{
209	*enc->cur++ = major | 26;	271	*enc->cur++ = major | LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	272	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	273	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	274	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	275	*enc->cur++ = len;
214	}	276	}
215	else	277	else
216	{	278	{
217	*enc->cur++ = major | 27;	279	*enc->cur++ = major | LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	280	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	281	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	282	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	283	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	284	*enc->cur++ = len >> 24;
…		…
224	*enc->cur++ = len >> 8;	286	*enc->cur++ = len >> 8;
225	*enc->cur++ = len;	287	*enc->cur++ = len;
226	}	288	}
227	}	289	}
228		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
229	ecb_inline void	303	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	304	encode_tag (enc_t *enc, UV tag)
231	{	305	{
232	encode_uint (enc, 0xc0, tag);	306	encode_uint (enc, MAJOR_TAG, tag);
		307	}
		308
		309	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		310	ecb_noinline static void
		311	encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
		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	}
233	}	330	}
234		331
235	ecb_inline void	332	ecb_inline void
236	encode_str (enc_t *enc, int utf8, char *str, STRLEN len)	333	encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
237	{	334	{
238	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	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);
239	need (enc, len);	343	need (enc, len);
240	memcpy (enc->cur, str, len);	344	memcpy (enc->cur, str, len);
241	enc->cur += len;	345	enc->cur += len;
242	}	346	}
243		347
244	static void	348	ecb_inline void
245	encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)	349	encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
246	{	350	{
247	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))	351	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
248	{	352	{
249	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	353	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
250		354
251	if (SvOK (*svp))	355	if (SvOK (*svp))
252	{	356	{
253	// already registered, use stringref	357	// already registered, use stringref
254	encode_tag (enc, CBOR_TAG_STRINGREF);	358	encode_tag (enc, CBOR_TAG_STRINGREF);
255	encode_uint (enc, 0x00, SvUV (*svp));	359	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
256	return;	360	return;
257	}	361	}
258	else if (len >= minimum_string_length (enc->stringref_idx))	362	else if (len >= minimum_string_length (enc->stringref_idx))
259	{	363	{
260	// register only	364	// register only
261	sv_setuv (*svp, enc->stringref_idx);	365	sv_setuv (*svp, enc->stringref_idx);
262	++enc->stringref_idx;	366	++enc->stringref_idx;
263	}	367	}
264	}	368	}
265		369
266	encode_str (enc, utf8, str, 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	{
		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
		385	memcpy (enc->cur, &fp, 2);
		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	}
267	}	460	}
268		461
269	static void encode_sv (enc_t *enc, SV *sv);	462	static void encode_sv (enc_t *enc, SV *sv);
270		463
271	static void	464	static void
…		…
276	if (enc->depth >= enc->cbor.max_depth)	469	if (enc->depth >= enc->cbor.max_depth)
277	croak (ERR_NESTING_EXCEEDED);	470	croak (ERR_NESTING_EXCEEDED);
278		471
279	++enc->depth;	472	++enc->depth;
280		473
281	encode_uint (enc, 0x80, len + 1);	474	encode_uint (enc, MAJOR_ARRAY, len + 1);
282		475
		476	if (ecb_expect_false (SvMAGICAL (av)))
283	for (i = 0; i <= len; ++i)	477	for (i = 0; i <= len; ++i)
284	{	478	{
285	SV **svp = av_fetch (av, i, 0);	479	SV **svp = av_fetch (av, i, 0);
286	encode_sv (enc, svp ? *svp : &PL_sv_undef);	480	encode_sv (enc, svp ? *svp : &PL_sv_undef);
287	}	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	}
288		488
289	--enc->depth;	489	--enc->depth;
290	}	490	}
291		491
292	static void	492	static void
…		…
300	++enc->depth;	500	++enc->depth;
301		501
302	int pairs = hv_iterinit (hv);	502	int pairs = hv_iterinit (hv);
303	int mg = SvMAGICAL (hv);	503	int mg = SvMAGICAL (hv);
304		504
305	if (mg)	505	if (ecb_expect_false (mg))
306	encode_ch (enc, 0xa0 | 31);	506	encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
307	else	507	else
308	encode_uint (enc, 0xa0, pairs);	508	encode_uint (enc, MAJOR_MAP, pairs);
309		509
310	while ((he = hv_iternext (hv)))	510	while ((he = hv_iternext (hv)))
311	{	511	{
312	if (HeKLEN (he) == HEf_SVKEY)	512	if (HeKLEN (he) == HEf_SVKEY)
313	encode_sv (enc, HeSVKEY (he));	513	encode_sv (enc, HeSVKEY (he));
314	else	514	else
315	encode_strref (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));
316		516
317	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));
318	}	518	}
319		519
320	if (mg)	520	if (ecb_expect_false (mg))
321	encode_ch (enc, 0xe0 | 31);	521	encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
322		522
323	--enc->depth;	523	--enc->depth;
324	}	524	}
325		525
326	// encode objects, arrays and special \0=false and \1=true values.	526	// encode objects, arrays and special \0=false and \1=true values.
…		…
345		545
346	HV *stash = SvSTASH (sv);	546	HV *stash = SvSTASH (sv);
347		547
348	if (stash == boolean_stash)	548	if (stash == boolean_stash)
349	{	549	{
350	encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);	550	encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
351	return;	551	return;
352	}	552	}
353	else if (stash == error_stash)	553	else if (stash == error_stash)
354	{	554	{
355	encode_ch (enc, 0xe0 | 23);	555	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
356	return;	556	return;
357	}	557	}
358	else if (stash == tagged_stash)	558	else if (stash == tagged_stash)
359	{	559	{
360	if (svt != SVt_PVAV)	560	if (svt != SVt_PVAV)
361	croak ("encountered CBOR::XS::Tagged object that isn't an array");	561	croak ("encountered CBOR::XS::Tagged object that isn't an array");
362		562
		563	switch (av_len ((AV *)sv))
		564	{
		565	case 2-1:
		566	// actually a tagged value
363	encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));	567	encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
364	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	}
365		579
366	return;	580	return;
367	}	581	}
368	}	582	}
369		583
370	if (ecb_expect_false (SvREFCNT (sv) > 1)	584	if (ecb_expect_false (SvREFCNT (sv) > 1)
371	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))	585	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
372	{	586	{
373	if (!enc->shareable)	587	if (ecb_expect_false (!enc->shareable))
374	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());	588	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
375		589
376	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);	590	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
377		591
378	if (SvOK (*svp))	592	if (SvOK (*svp))
379	{	593	{
380	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);	594	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381	encode_uint (enc, 0x00, SvUV (*svp));	595	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
382	return;	596	return;
383	}	597	}
384	else	598	else
385	{	599	{
386	sv_setuv (*svp, enc->shareable_idx);	600	sv_setuv (*svp, enc->shareable_idx);
…		…
392	if (ecb_expect_false (SvOBJECT (sv)))	606	if (ecb_expect_false (SvOBJECT (sv)))
393	{	607	{
394	HV *stash = SvSTASH (sv);	608	HV *stash = SvSTASH (sv);
395	GV *method;	609	GV *method;
396		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))));
397	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	614	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
398	{	615	{
399	dSP;	616	dSP;
400		617
401	ENTER; SAVETMPS; PUSHMARK (SP);	618	ENTER; SAVETMPS;
		619	PUSHMARK (SP);
402	// 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
403	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	621	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
404		622
405	PUTBACK;	623	PUTBACK;
406	// G_SCALAR ensures that return value is 1	624	// G_SCALAR ensures that return value is 1
…		…
419	}	637	}
420	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	638	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
421	{	639	{
422	dSP;	640	dSP;
423		641
424	ENTER; SAVETMPS; PUSHMARK (SP);	642	ENTER; SAVETMPS;
		643	PUSHMARK (SP);
425	EXTEND (SP, 2);	644	EXTEND (SP, 2);
426	// 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
427	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	646	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
428	PUSHs (sv_cbor);	647	PUSHs (sv_cbor);
429		648
…		…
434	// catch this surprisingly common error	653	// catch this surprisingly common error
435	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	654	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
436	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));
437		656
438	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	657	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
439	encode_uint (enc, 0x80, count + 1);	658	encode_uint (enc, MAJOR_ARRAY, count + 1);
440	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	659	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
441		660
442	while (count)	661	{
		662	int i;
		663
		664	for (i = 0; i < count; ++i)
443	encode_sv (enc, SP[1 - count--]);	665	encode_sv (enc, SP[i + 1 - count]);
		666
		667	SP -= count;
		668	}
444		669
445	PUTBACK;	670	PUTBACK;
446		671
447	FREETMPS; LEAVE;	672	FREETMPS; LEAVE;
448	}	673	}
…		…
466	{	691	{
467	double nv = SvNVX (sv);	692	double nv = SvNVX (sv);
468		693
469	need (enc, 9);	694	need (enc, 9);
470		695
471	if (ecb_expect_false (nv == (U32)nv))	696	if (ecb_expect_false (nv == (NV)(U32)nv))
472	encode_uint (enc, 0x00, (U32)nv);	697	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
473	//TODO: maybe I32?	698	//TODO: maybe I32?
474	else if (ecb_expect_false (nv == (float)nv))	699	else if (ecb_expect_false (nv == (float)nv))
475	{	700	encode_float32 (enc, nv);
476	uint32_t fp = ecb_float_to_binary32 (nv);
477
478	*enc->cur++ = 0xe0 | 26;
479
480	if (!ecb_big_endian ())
481	fp = ecb_bswap32 (fp);
482
483	memcpy (enc->cur, &fp, 4);
484	enc->cur += 4;
485	}
486	else	701	else
487	{	702	encode_float64 (enc, nv);
488	uint64_t fp = ecb_double_to_binary64 (nv);
489
490	*enc->cur++ = 0xe0 | 27;
491
492	if (!ecb_big_endian ())
493	fp = ecb_bswap64 (fp);
494
495	memcpy (enc->cur, &fp, 8);
496	enc->cur += 8;
497	}
498	}	703	}
499		704
500	static void	705	static void
501	encode_sv (enc_t *enc, SV *sv)	706	encode_sv (enc_t *enc, SV *sv)
502	{	707	{
…		…
504		709
505	if (SvPOKp (sv))	710	if (SvPOKp (sv))
506	{	711	{
507	STRLEN len;	712	STRLEN len;
508	char *str = SvPV (sv, len);	713	char *str = SvPV (sv, len);
509	encode_strref (enc, SvUTF8 (sv), str, len);	714	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
510	}	715	}
511	else if (SvNOKp (sv))	716	else if (SvNOKp (sv))
512	encode_nv (enc, sv);	717	encode_nv (enc, sv);
513	else if (SvIOKp (sv))	718	else if (SvIOKp (sv))
514	{	719	encode_int (e,v sv);
515	if (SvIsUV (sv))
516	encode_uint (enc, 0x00, SvUVX (sv));
517	else if (SvIVX (sv) >= 0)
518	encode_uint (enc, 0x00, SvIVX (sv));
519	else
520	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
521	}
522	else if (SvROK (sv))	720	else if (SvROK (sv))
523	encode_rv (enc, SvRV (sv));	721	encode_rv (enc, SvRV (sv));
524	else if (!SvOK (sv))	722	else if (!SvOK (sv))
525	encode_ch (enc, 0xe0 | 22);	723	encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
526	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	724	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
527	encode_ch (enc, 0xe0 | 23);	725	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
528	else	726	else
529	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",
530	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	728	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
531	}	729	}
532		730
533	static SV *	731	static SV *
534	encode_cbor (SV *scalar, CBOR *cbor)	732	encode_cbor (SV *scalar, CBOR *cbor)
535	{	733	{
536	enc_t enc = { };	734	enc_t enc = { 0 };
537		735
538	enc.cbor = *cbor;	736	enc.cbor = *cbor;
539	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	737	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
540	enc.cur = SvPVX (enc.sv);	738	enc.cur = SvPVX (enc.sv);
541	enc.end = SvEND (enc.sv);	739	enc.end = SvEND (enc.sv);
542		740
543	SvPOK_only (enc.sv);	741	SvPOK_only (enc.sv);
544		742
545	if (cbor->flags & F_PACK_STRINGS)	743	if (cbor->flags & F_PACK_STRINGS)
546	{	744	{
…		…
573	U32 depth; // recursion depth	771	U32 depth; // recursion depth
574	U32 maxdepth; // recursion depth limit	772	U32 maxdepth; // recursion depth limit
575	AV *shareable;	773	AV *shareable;
576	AV *stringref;	774	AV *stringref;
577	SV *decode_tagged;	775	SV *decode_tagged;
		776	SV *err_sv; // optional sv for error, needs to be freed
578	} dec_t;	777	} dec_t;
579		778
580	#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);
581		786
582	#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;
583		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
584	#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)
585	#define DEC_DEC_DEPTH --dec->depth	816	#define DEC_DEC_DEPTH --dec->depth
586		817
587	static UV	818	static UV
588	decode_uint (dec_t *dec)	819	decode_uint (dec_t *dec)
589	{	820	{
590	switch (*dec->cur & 31)	821	U8 m = *dec->cur & MINOR_MASK;
591	{	822	++dec->cur;
592	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
593	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
594	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
595	return *dec->cur++ & 31;
596		823
597	case 24:	824	if (ecb_expect_true (m < LENGTH_EXT1))
		825	return m;
		826	else if (ecb_expect_true (m == LENGTH_EXT1))
		827	{
598	WANT (2);	828	WANT (1);
599	dec->cur += 2;	829	dec->cur += 1;
600	return dec->cur[-1];	830	return dec->cur[-1];
601		831	}
602	case 25:	832	else if (ecb_expect_true (m == LENGTH_EXT2))
		833	{
603	WANT (3);	834	WANT (2);
604	dec->cur += 3;	835	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	836	return (((UV)dec->cur[-2]) << 8)
606	| ((UV)dec->cur[-1]);	837	| ((UV)dec->cur[-1]);
607		838	}
608	case 26:	839	else if (ecb_expect_true (m == LENGTH_EXT4))
		840	{
609	WANT (5);	841	WANT (4);
610	dec->cur += 5;	842	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	843	return (((UV)dec->cur[-4]) << 24)
612	| (((UV)dec->cur[-3]) << 16)	844	| (((UV)dec->cur[-3]) << 16)
613	| (((UV)dec->cur[-2]) << 8)	845	| (((UV)dec->cur[-2]) << 8)
614	| ((UV)dec->cur[-1]);	846	| ((UV)dec->cur[-1]);
615		847	}
616	case 27:	848	else if (ecb_expect_true (m == LENGTH_EXT8))
		849	{
617	WANT (9);	850	WANT (8);
618	dec->cur += 9;	851	dec->cur += 8;
619		852
620	return	853	return
621	#if UVSIZE < 8	854	#if UVSIZE < 8
622	0	855	0
623	#else	856	#else
624	(((UV)dec->cur[-8]) << 56)	857	(((UV)dec->cur[-8]) << 56)
625	| (((UV)dec->cur[-7]) << 48)	858	| (((UV)dec->cur[-7]) << 48)
626	| (((UV)dec->cur[-6]) << 40)	859	| (((UV)dec->cur[-6]) << 40)
627	| (((UV)dec->cur[-5]) << 32)	860	| (((UV)dec->cur[-5]) << 32)
628	#endif	861	#endif
629	| (((UV)dec->cur[-4]) << 24)	862	| (((UV)dec->cur[-4]) << 24)
630	| (((UV)dec->cur[-3]) << 16)	863	| (((UV)dec->cur[-3]) << 16)
631	| (((UV)dec->cur[-2]) << 8)	864	| (((UV)dec->cur[-2]) << 8)
632	| ((UV)dec->cur[-1]);	865	| ((UV)dec->cur[-1]);
633		866	}
634	default:	867	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	868	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		869
638	fail:	870	fail:
639	return 0;	871	return 0;
640	}	872	}
641		873
…		…
646	{	878	{
647	AV *av = newAV ();	879	AV *av = newAV ();
648		880
649	DEC_INC_DEPTH;	881	DEC_INC_DEPTH;
650		882
651	if ((*dec->cur & 31) == 31)	883	if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
652	{	884	{
653	++dec->cur;	885	++dec->cur;
654		886
655	for (;;)	887	for (;;)
656	{	888	{
657	WANT (1);	889	WANT (1);
658		890
659	if (*dec->cur == (0xe0 | 31))	891	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
660	{	892	{
661	++dec->cur;	893	++dec->cur;
662	break;	894	break;
663	}	895	}
664		896
665	av_push (av, decode_sv (dec));	897	av_push (av, decode_sv (dec));
666	}	898	}
667	}	899	}
668	else	900	else
669	{	901	{
670	int i, len = decode_uint (dec);	902	UV i, len = decode_uint (dec);
671		903
		904	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
672	av_fill (av, len - 1);	905	av_fill (av, len - 1);
673		906
674	for (i = 0; i < len; ++i)	907	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	908	AvARRAY (av)[i] = decode_sv (dec);
676	}	909	}
677		910
678	DEC_DEC_DEPTH;	911	DEC_DEC_DEPTH;
679	return newRV_noinc ((SV *)av);	912	return newRV_noinc ((SV *)av);
680		913
681	fail:	914	fail:
682	SvREFCNT_dec (av);	915	SvREFCNT_dec_NN (av);
683	DEC_DEC_DEPTH;	916	DEC_DEC_DEPTH;
684	return &PL_sv_undef;	917	return &PL_sv_undef;
685	}	918	}
686		919
687	static void	920	static void
…		…
689	{	922	{
690	// for speed reasons, we specialcase single-string	923	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	924	// byte or utf-8 strings as keys, but only when !stringref
692		925
693	if (ecb_expect_true (!dec->stringref))	926	if (ecb_expect_true (!dec->stringref))
694	if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)	927	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	928	{
696	I32 len = decode_uint (dec);	929	STRLEN len = decode_uint (dec);
697	char *key = (char *)dec->cur;	930	char *key = (char *)dec->cur;
698		931
		932	WANT (len);
699	dec->cur += len;	933	dec->cur += len;
700		934
701	if (ecb_expect_false (dec->stringref))
702	av_push (dec->stringref, newSVpvn (key, len));
703
704	hv_store (hv, key, len, decode_sv (dec), 0);	935	hv_store (hv, key, len, decode_sv (dec), 0);
705		936
706	return;	937	return;
707	}	938	}
708	else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)	939	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	940	{
710	I32 len = decode_uint (dec);	941	STRLEN len = decode_uint (dec);
711	char *key = (char *)dec->cur;	942	char *key = (char *)dec->cur;
712		943
		944	WANT (len);
713	dec->cur += len;	945	dec->cur += len;
714		946
715	if (ecb_expect_false (dec->stringref))	947	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	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)");
717		950
718	hv_store (hv, key, -len, decode_sv (dec), 0);	951	hv_store (hv, key, -len, decode_sv (dec), 0);
719		952
720	return;	953	return;
721	}	954	}
722		955
723	SV *k = decode_sv (dec);	956	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	957	SV *v = decode_sv (dec);
725		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
726	hv_store_ent (hv, k, v, 0);	987	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	988	SvREFCNT_dec_NN (k);
		989
		990	fail:
		991	;
728	}	992	}
729		993
730	static SV *	994	static SV *
731	decode_hv (dec_t *dec)	995	decode_hv (dec_t *dec)
732	{	996	{
733	HV *hv = newHV ();	997	HV *hv = newHV ();
734		998
735	DEC_INC_DEPTH;	999	DEC_INC_DEPTH;
736		1000
737	if ((*dec->cur & 31) == 31)	1001	if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
738	{	1002	{
739	++dec->cur;	1003	++dec->cur;
740		1004
741	for (;;)	1005	for (;;)
742	{	1006	{
743	WANT (1);	1007	WANT (1);
744		1008
745	if (*dec->cur == (0xe0 | 31))	1009	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
746	{	1010	{
747	++dec->cur;	1011	++dec->cur;
748	break;	1012	break;
749	}	1013	}
750		1014
751	decode_he (dec, hv);	1015	decode_he (dec, hv);
752	}	1016	}
753	}	1017	}
754	else	1018	else
755	{	1019	{
756	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
757		1023
758	while (pairs--)	1024	while (pairs--)
759	decode_he (dec, hv);	1025	decode_he (dec, hv);
760	}	1026	}
761		1027
762	DEC_DEC_DEPTH;	1028	DEC_DEC_DEPTH;
763	return newRV_noinc ((SV *)hv);	1029	return newRV_noinc ((SV *)hv);
764		1030
765	fail:	1031	fail:
766	SvREFCNT_dec (hv);	1032	SvREFCNT_dec_NN (hv);
767	DEC_DEC_DEPTH;	1033	DEC_DEC_DEPTH;
768	return &PL_sv_undef;	1034	return &PL_sv_undef;
769	}	1035	}
770		1036
771	static SV *	1037	static SV *
772	decode_str (dec_t *dec, int utf8)	1038	decode_str (dec_t *dec, int utf8)
773	{	1039	{
774	SV *sv = 0;	1040	SV *sv = 0;
775		1041
776	if ((*dec->cur & 31) == 31)	1042	if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
777	{	1043	{
778	// indefinite length strings	1044	// indefinite length strings
779	++dec->cur;	1045	++dec->cur;
780		1046
781	unsigned char major = *dec->cur & 0xe0;	1047	U8 major = *dec->cur & MAJOR_MISC;
782		1048
783	sv = newSVpvn ("", 0);	1049	sv = newSVpvn ("", 0);
784		1050
785	for (;;)	1051	for (;;)
786	{	1052	{
787	WANT (1);	1053	WANT (1);
788		1054
789	if ((*dec->cur ^ major) >= 31)	1055	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 | 31))	1056	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
791	{	1057	{
792	++dec->cur;	1058	++dec->cur;
793	break;	1059	break;
794	}	1060	}
795	else	1061	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	1080	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	1081	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	1082	}
817		1083
818	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
819	SvUTF8_on (sv);	1090	SvUTF8_on (sv);
		1091	}
820		1092
821	return sv;	1093	return sv;
822		1094
823	fail:	1095	fail:
824	SvREFCNT_dec (sv);	1096	SvREFCNT_dec (sv);
…		…
843	sv = newRV_noinc (decode_sv (dec));	1115	sv = newRV_noinc (decode_sv (dec));
844	break;	1116	break;
845		1117
846	case CBOR_TAG_STRINGREF_NAMESPACE:	1118	case CBOR_TAG_STRINGREF_NAMESPACE:
847	{	1119	{
		1120	// do not use SAVETMPS/FREETMPS, as these will
		1121	// erase mortalised caches, e.g. "shareable"
848	ENTER; SAVETMPS;	1122	ENTER;
849		1123
850	SAVESPTR (dec->stringref);	1124	SAVESPTR (dec->stringref);
851	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());	1125	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
852		1126
853	sv = decode_sv (dec);	1127	sv = decode_sv (dec);
854		1128
855	FREETMPS; LEAVE;	1129	LEAVE;
856	}	1130	}
857	break;	1131	break;
858		1132
859	case CBOR_TAG_STRINGREF:	1133	case CBOR_TAG_STRINGREF:
860	{	1134	{
861	if ((*dec->cur >> 5) != 0)	1135	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1136	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		1137
864	UV idx = decode_uint (dec);	1138	UV idx = decode_uint (dec);
865		1139
866	if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))	1140	if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
867	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)");
868		1142
869	sv = newSVsv (AvARRAY (dec->stringref)[idx]);	1143	sv = newSVsv (AvARRAY (dec->stringref)[idx]);
870	}	1144	}
871	break;	1145	break;
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	1147	case CBOR_TAG_VALUE_SHAREABLE:
874	{	1148	{
875	if (ecb_expect_false (!dec->shareable))	1149	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());	1150	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
877		1151
		1152	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1153	{
878	sv = newSV (0);	1154	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1155	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		1156
881	SV *osv = decode_sv (dec);	1157	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	1158	sv_setsv (sv, osv);
883	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	}
884	}	1168	}
885	break;	1169	break;
886		1170
887	case CBOR_TAG_VALUE_SHAREDREF:	1171	case CBOR_TAG_VALUE_SHAREDREF:
888	{	1172	{
889	if ((*dec->cur >> 5) != 0)	1173	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1174	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		1175
892	UV idx = decode_uint (dec);	1176	UV idx = decode_uint (dec);
893		1177
894	if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))	1178	if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1179	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		1180
897	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");
898	}	1185	}
899	break;	1186	break;
900		1187
901	case CBOR_TAG_PERL_OBJECT:	1188	case CBOR_TAG_PERL_OBJECT:
902	{	1189	{
		1190	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1191	goto filter;
		1192
903	sv = decode_sv (dec);	1193	sv = decode_sv (dec);
904		1194
905	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)	1195	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
906	ERR ("corrupted CBOR data (non-array perl object)");	1196	ERR ("corrupted CBOR data (non-array perl object)");
907		1197
…		…
917	if (!method)	1207	if (!method)
918	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)");
919		1209
920	dSP;	1210	dSP;
921		1211
922	ENTER; SAVETMPS; PUSHMARK (SP);	1212	ENTER; SAVETMPS;
		1213	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1214	EXTEND (SP, len + 1);
924	// 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
925	PUSHs (*av_fetch (av, 0, 1));	1216	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1217	PUSHs (sv_cbor);
927		1218
…		…
935	SPAGAIN;	1226	SPAGAIN;
936		1227
937	if (SvTRUE (ERRSV))	1228	if (SvTRUE (ERRSV))
938	{	1229	{
939	FREETMPS; LEAVE;	1230	FREETMPS; LEAVE;
940	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1231	ERR_ERRSV;
941	}	1232	}
942		1233
943	SvREFCNT_dec (sv);	1234	SvREFCNT_dec_NN (sv);
944	sv = SvREFCNT_inc (POPs);	1235	sv = SvREFCNT_inc (POPs);
945		1236
946	PUTBACK;	1237	PUTBACK;
947		1238
948	FREETMPS; LEAVE;	1239	FREETMPS; LEAVE;
949	}	1240	}
950	break;	1241	break;
951		1242
952	default:	1243	default:
		1244	filter:
953	{	1245	{
		1246	SV *tag_sv = newSVuv (tag);
		1247
954	sv = decode_sv (dec);	1248	sv = decode_sv (dec);
955		1249
956	dSP;	1250	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1251	ENTER; SAVETMPS;
		1252	PUSHMARK (SP);
958	EXTEND (SP, 2);	1253	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1254	PUSHs (tag_sv);
960	PUSHs (sv);	1255	PUSHs (sv);
961		1256
962	PUTBACK;	1257	PUTBACK;
963	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);
964	SPAGAIN;	1259	SPAGAIN;
965		1260
966	if (SvTRUE (ERRSV))	1261	if (SvTRUE (ERRSV))
967	{	1262	{
		1263	SvREFCNT_dec_NN (tag_sv);
968	FREETMPS; LEAVE;	1264	FREETMPS; LEAVE;
969	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1265	ERR_ERRSV;
970	}	1266	}
971		1267
972	if (count)	1268	if (count)
973	{	1269	{
		1270	SvREFCNT_dec_NN (tag_sv);
974	SvREFCNT_dec (sv);	1271	SvREFCNT_dec_NN (sv);
975	sv = SvREFCNT_inc (POPs);	1272	sv = SvREFCNT_inc_NN (TOPs);
		1273	SP -= count;
976	}	1274	}
977	else	1275	else
978	{	1276	{
979	AV *av = newAV ();	1277	AV *av = newAV ();
980	av_push (av, newSVuv (tag));	1278	av_push (av, tag_sv);
981	av_push (av, sv);	1279	av_push (av, sv);
982		1280
983	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	1281	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
984	? cbor_tagged_stash	1282	? cbor_tagged_stash
985	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1283	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
1003	static SV *	1301	static SV *
1004	decode_sv (dec_t *dec)	1302	decode_sv (dec_t *dec)
1005	{	1303	{
1006	WANT (1);	1304	WANT (1);
1007		1305
1008	switch (*dec->cur >> 5)	1306	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1307	{
1010	case 0: // unsigned int	1308	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1309	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1310	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1311	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1312	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1313	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1314	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1315
1018	case 4: // array	1316	case MAJOR_MISC >> MAJOR_SHIFT:
1019	return decode_av (dec);
1020	case 5: // map
1021	return decode_hv (dec);
1022	case 6: // tag
1023	return decode_tagged (dec);
1024	case 7: // misc
1025	switch (*dec->cur++ & 31)	1317	switch (*dec->cur++ & MINOR_MASK)
1026	{	1318	{
1027	case 20:	1319	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1320	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1321	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1322	#endif
1031	return newSVsv (types_false);	1323	return newSVsv (types_false);
1032	case 21:	1324	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1325	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1326	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1327	#endif
1036	return newSVsv (types_true);	1328	return newSVsv (types_true);
1037	case 22:	1329	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1330	return newSVsv (&PL_sv_undef);
1039	case 23:	1331	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1332	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1333	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1334	#endif
1043	return newSVsv (types_error);	1335	return newSVsv (types_error);
1044		1336
1045	case 25:	1337	case MISC_FLOAT16:
1046	{	1338	{
1047	WANT (2);	1339	WANT (2);
1048		1340
1049	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];	1341	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1050	dec->cur += 2;	1342	dec->cur += 2;
1051		1343
1052	return newSVnv (ecb_binary16_to_float (fp));	1344	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1345	}
1054		1346
1055	case 26:	1347	case MISC_FLOAT32:
1056	{	1348	{
1057	uint32_t fp;	1349	uint32_t fp;
1058	WANT (4);	1350	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1351	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1352	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1355	fp = ecb_bswap32 (fp);
1064		1356
1065	return newSVnv (ecb_binary32_to_float (fp));	1357	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1358	}
1067		1359
1068	case 27:	1360	case MISC_FLOAT64:
1069	{	1361	{
1070	uint64_t fp;	1362	uint64_t fp;
1071	WANT (8);	1363	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1364	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1365	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1368	fp = ecb_bswap64 (fp);
1077		1369
1078	return newSVnv (ecb_binary64_to_double (fp));	1370	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1371	}
1080		1372
1081	// 0..19 unassigned	1373	// 0..19 unassigned simple
1082	// 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
1083	default:	1377	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1378	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1379	}
1086		1380
1087	break;	1381	break;
1088	}	1382	}
1089		1383
…		…
1092	}	1386	}
1093		1387
1094	static SV *	1388	static SV *
1095	decode_cbor (SV *string, CBOR *cbor, char **offset_return)	1389	decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1096	{	1390	{
1097	dec_t dec = { };	1391	dec_t dec = { 0 };
1098	SV *sv;	1392	SV *sv;
1099	STRLEN len;	1393	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1394	char *data = SvPVbyte (string, len);
1101		1395
1102	if (len > cbor->max_size && cbor->max_size)	1396	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1410	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1411	dec.err = "garbage after CBOR object";
1118		1412
1119	if (dec.err)	1413	if (dec.err)
1120	{	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
1121	SvREFCNT_dec (sv);	1426	SvREFCNT_dec_NN (sv);
		1427
		1428	if (dec.err_sv)
		1429	sv_2mortal (dec.err_sv);
		1430
1122	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);
1123	}	1432	}
1124		1433
1125	sv = sv_2mortal (sv);	1434	sv = sv_2mortal (sv);
1126		1435
1127	return sv;	1436	return sv;
1128	}	1437	}
1129		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
1130	/////////////////////////////////////////////////////////////////////////////	1563	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1564	// XS interface functions
1132		1565
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1566	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1567
…		…
1146		1579
1147	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1580	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1148		1581
1149	sv_cbor = newSVpv ("CBOR", 0);	1582	sv_cbor = newSVpv ("CBOR", 0);
1150	SvREADONLY_on (sv_cbor);	1583	SvREADONLY_on (sv_cbor);
		1584
		1585	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1151	}	1586	}
1152		1587
1153	PROTOTYPES: DISABLE	1588	PROTOTYPES: DISABLE
1154		1589
1155	void CLONE (...)	1590	void CLONE (...)
…		…
1174	void shrink (CBOR *self, int enable = 1)	1609	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1610	ALIAS:
1176	shrink = F_SHRINK	1611	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1612	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1613	allow_sharing = F_ALLOW_SHARING
		1614	allow_cycles = F_ALLOW_CYCLES
		1615	forbid_objects = F_FORBID_OBJECTS
1179	pack_strings = F_PACK_STRINGS	1616	pack_strings = F_PACK_STRINGS
		1617	text_keys = F_TEXT_KEYS
		1618	text_strings = F_TEXT_STRINGS
		1619	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1620	PPCODE:
1181	{	1621	{
1182	if (enable)	1622	if (enable)
1183	self->flags |= ix;	1623	self->flags |= ix;
1184	else	1624	else
…		…
1190	void get_shrink (CBOR *self)	1630	void get_shrink (CBOR *self)
1191	ALIAS:	1631	ALIAS:
1192	get_shrink = F_SHRINK	1632	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1633	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1634	get_allow_sharing = F_ALLOW_SHARING
		1635	get_allow_cycles = F_ALLOW_CYCLES
		1636	get_forbid_objects = F_FORBID_OBJECTS
1195	get_pack_strings = F_PACK_STRINGS	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
1196	PPCODE:	1641	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1642	XPUSHs (boolSV (self->flags & ix));
1198		1643
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1644	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1645	PPCODE:
…		…
1249	EXTEND (SP, 2);	1694	EXTEND (SP, 2);
1250	PUSHs (sv);	1695	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1696	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1697	}
1253		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
1254	void DESTROY (CBOR *self)	1751	void DESTROY (CBOR *self)
1255	PPCODE:	1752	PPCODE:
1256	cbor_free (self);	1753	cbor_free (self);
1257		1754
1258	PROTOTYPES: ENABLE	1755	PROTOTYPES: ENABLE
1259		1756
1260	void encode_cbor (SV *scalar)	1757	void encode_cbor (SV *scalar)
		1758	ALIAS:
		1759	encode_cbor = 0
		1760	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1761	PPCODE:
1262	{	1762	{
1263	CBOR cbor;	1763	CBOR cbor;
1264	cbor_init (&cbor);	1764	cbor_init (&cbor);
		1765	cbor.flags |= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1766	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1767	XPUSHs (scalar);
1267	}	1768	}
1268		1769
1269	void decode_cbor (SV *cborstr)	1770	void decode_cbor (SV *cborstr)
…		…
1273	cbor_init (&cbor);	1774	cbor_init (&cbor);
1274	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;	1775	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1275	XPUSHs (cborstr);	1776	XPUSHs (cborstr);
1276	}	1777	}
1277		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