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 1.70 by root, Fri Dec 4 02:57:14 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_INT = 1,
		109	AS_BYTES = 2,
		110	AS_TEXT = 3,
		111	AS_FLOAT16 = 4,
		112	AS_FLOAT32 = 5,
		113	AS_FLOAT64 = 6,
		114	AS_MAP = 7,
		115	// possibly future enhancements: (generic) float, (generic) string
60	};	116	};
61		117
62	#define F_SHRINK 0x00000001UL	118	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	119	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL	120	#define F_ALLOW_SHARING 0x00000004UL
		121	#define F_ALLOW_CYCLES 0x00000008UL
		122	#define F_FORBID_OBJECTS 0x00000010UL
65	#define F_PACK_STRINGS 0x00000008UL	123	#define F_PACK_STRINGS 0x00000020UL
		124	#define F_TEXT_KEYS 0x00000040UL
		125	#define F_TEXT_STRINGS 0x00000080UL
		126	#define F_VALIDATE_UTF8 0x00000100UL
66		127
67	#define INIT_SIZE 32 // initial scalar size to be allocated	128	#define INIT_SIZE 32 // initial scalar size to be allocated
68		129
69	#define SB do {	130	#define SB do {
70	#define SE } while (0)	131	#define SE } while (0)
…		…
89	typedef struct {	150	typedef struct {
90	U32 flags;	151	U32 flags;
91	U32 max_depth;	152	U32 max_depth;
92	STRLEN max_size;	153	STRLEN max_size;
93	SV *filter;	154	SV *filter;
		155
		156	// for the incremental parser
		157	STRLEN incr_pos; // the current offset into the text
		158	STRLEN incr_need; // minimum bytes needed to decode
		159	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
94	} CBOR;	160	} CBOR;
95		161
96	ecb_inline void	162	ecb_inline void
97	cbor_init (CBOR *cbor)	163	cbor_init (CBOR *cbor)
98	{	164	{
…		…
102		168
103	ecb_inline void	169	ecb_inline void
104	cbor_free (CBOR *cbor)	170	cbor_free (CBOR *cbor)
105	{	171	{
106	SvREFCNT_dec (cbor->filter);	172	SvREFCNT_dec (cbor->filter);
		173	SvREFCNT_dec (cbor->incr_count);
107	}	174	}
108		175
109	/////////////////////////////////////////////////////////////////////////////	176	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	177	// utility functions
111		178
…		…
137		204
138	// minimum length of a string to be registered for stringref	205	// minimum length of a string to be registered for stringref
139	ecb_inline int	206	ecb_inline int
140	minimum_string_length (UV idx)	207	minimum_string_length (UV idx)
141	{	208	{
142	return idx > 23	209	return idx <= 23 ? 3
143	? idx > 0xffU	210	: idx <= 0xffU ? 4
144	? idx > 0xffffU	211	: idx <= 0xffffU ? 5
145	? idx > 0xffffffffU	212	: idx <= 0xffffffffU ? 7
146	? 11	213	: 11;
147	: 7
148	: 5
149	: 4
150	: 3;
151	}	214	}
152		215
153	/////////////////////////////////////////////////////////////////////////////	216	/////////////////////////////////////////////////////////////////////////////
154	// encoder	217	// encoder
155		218
…		…
168	} enc_t;	231	} enc_t;
169		232
170	ecb_inline void	233	ecb_inline void
171	need (enc_t *enc, STRLEN len)	234	need (enc_t *enc, STRLEN len)
172	{	235	{
173	if (ecb_expect_false (enc->cur + len >= enc->end))	236	if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
174	{	237	{
175	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);	238	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
176	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);	239	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
177	enc->cur = SvPVX (enc->sv) + cur;	240	enc->cur = SvPVX (enc->sv) + cur;
178	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;	241	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
179	}	242	}
180	}	243	}
181		244
		245	static void encode_sv (enc_t *enc, SV *sv);
		246
182	ecb_inline void	247	ecb_inline void
183	encode_ch (enc_t *enc, char ch)	248	encode_ch (enc_t *enc, char ch)
184	{	249	{
185	need (enc, 1);	250	need (enc, 1);
186	*enc->cur++ = ch;	251	*enc->cur++ = ch;
187	}	252	}
188		253
		254	// used for tags, intregers, element counts and so on
189	static void	255	static void
190	encode_uint (enc_t *enc, int major, UV len)	256	encode_uint (enc_t *enc, int major, UV len)
191	{	257	{
192	need (enc, 9);	258	need (enc, 9);
193		259
194	if (ecb_expect_true (len < 24))	260	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major | len;	261	*enc->cur++ = major | len;
196	else if (ecb_expect_true (len <= 0xff))	262	else if (ecb_expect_true (len <= 0xffU))
197	{	263	{
198	*enc->cur++ = major | 24;	264	*enc->cur++ = major | LENGTH_EXT1;
199	*enc->cur++ = len;	265	*enc->cur++ = len;
200	}	266	}
201	else if (len <= 0xffff)	267	else if (len <= 0xffffU)
202	{	268	{
203	*enc->cur++ = major | 25;	269	*enc->cur++ = major | LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	270	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	271	*enc->cur++ = len;
206	}	272	}
207	else if (len <= 0xffffffff)	273	else if (len <= 0xffffffffU)
208	{	274	{
209	*enc->cur++ = major | 26;	275	*enc->cur++ = major | LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	276	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	277	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	278	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	279	*enc->cur++ = len;
214	}	280	}
215	else	281	else
216	{	282	{
217	*enc->cur++ = major | 27;	283	*enc->cur++ = major | LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	284	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	285	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	286	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	287	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	288	*enc->cur++ = len >> 24;
…		…
224	*enc->cur++ = len >> 8;	290	*enc->cur++ = len >> 8;
225	*enc->cur++ = len;	291	*enc->cur++ = len;
226	}	292	}
227	}	293	}
228		294
		295	// encodes a perl value into a CBOR integer
		296	ecb_inline void
		297	encode_int (enc_t *enc, SV *sv)
		298	{
		299	if (SvIsUV (sv))
		300	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
		301	else if (SvIVX (sv) >= 0)
		302	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
		303	else
		304	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
		305	}
		306
229	ecb_inline void	307	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	308	encode_tag (enc_t *enc, UV tag)
231	{	309	{
232	encode_uint (enc, 0xc0, tag);	310	encode_uint (enc, MAJOR_TAG, tag);
		311	}
		312
		313	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		314	ecb_noinline static void
		315	encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
		316	{
		317	STRLEN ulen = len;
		318	U8 *p, *pend = (U8 *)str + len;
		319
		320	for (p = (U8 *)str; p < pend; ++p)
		321	ulen += *p >> 7; // count set high bits
		322
		323	encode_uint (enc, MAJOR_TEXT, ulen);
		324
		325	need (enc, ulen);
		326	for (p = (U8 *)str; p < pend; ++p)
		327	if (*p < 0x80)
		328	*enc->cur++ = *p;
		329	else
		330	{
		331	*enc->cur++ = 0xc0 + (*p >> 6);
		332	*enc->cur++ = 0x80 + (*p & 63);
		333	}
233	}	334	}
234		335
235	ecb_inline void	336	ecb_inline void
236	encode_str (enc_t *enc, int utf8, char *str, STRLEN len)	337	encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
237	{	338	{
238	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	339	if (ecb_expect_false (upgrade_utf8))
		340	if (!utf8)
		341	{
		342	encode_str_utf8 (enc, utf8, str, len);
		343	return;
		344	}
		345
		346	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
239	need (enc, len);	347	need (enc, len);
240	memcpy (enc->cur, str, len);	348	memcpy (enc->cur, str, len);
241	enc->cur += len;	349	enc->cur += len;
242	}	350	}
243		351
244	static void	352	ecb_inline void
245	encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)	353	encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
246	{	354	{
247	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))	355	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
248	{	356	{
249	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	357	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
250		358
251	if (SvOK (*svp))	359	if (SvOK (*svp))
252	{	360	{
253	// already registered, use stringref	361	// already registered, use stringref
254	encode_tag (enc, CBOR_TAG_STRINGREF);	362	encode_tag (enc, CBOR_TAG_STRINGREF);
255	encode_uint (enc, 0x00, SvUV (*svp));	363	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
256	return;	364	return;
257	}	365	}
258	else if (len >= minimum_string_length (enc->stringref_idx))	366	else if (len >= minimum_string_length (enc->stringref_idx))
259	{	367	{
260	// register only	368	// register only
261	sv_setuv (*svp, enc->stringref_idx);	369	sv_setuv (*svp, enc->stringref_idx);
262	++enc->stringref_idx;	370	++enc->stringref_idx;
263	}	371	}
264	}	372	}
265		373
266	encode_str (enc, utf8, str, len);	374	encode_str (enc, upgrade_utf8, utf8, str, len);
267	}	375	}
268		376
269	static void encode_sv (enc_t *enc, SV *sv);	377	ecb_inline void
		378	encode_float16 (enc_t *enc, NV nv)
		379	{
		380	need (enc, 1+2);
		381
		382	*enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
		383
		384	uint16_t fp = ecb_float_to_binary16 (nv);
		385
		386	if (!ecb_big_endian ())
		387	fp = ecb_bswap16 (fp);
		388
		389	memcpy (enc->cur, &fp, 2);
		390	enc->cur += 2;
		391	}
		392
		393	ecb_inline void
		394	encode_float32 (enc_t *enc, NV nv)
		395	{
		396	need (enc, 1+4);
		397
		398	*enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
		399
		400	uint32_t fp = ecb_float_to_binary32 (nv);
		401
		402	if (!ecb_big_endian ())
		403	fp = ecb_bswap32 (fp);
		404
		405	memcpy (enc->cur, &fp, 4);
		406	enc->cur += 4;
		407	}
		408
		409	ecb_inline void
		410	encode_float64 (enc_t *enc, NV nv)
		411	{
		412	need (enc, 1+8);
		413
		414	*enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
		415
		416	uint64_t fp = ecb_double_to_binary64 (nv);
		417
		418	if (!ecb_big_endian ())
		419	fp = ecb_bswap64 (fp);
		420
		421	memcpy (enc->cur, &fp, 8);
		422	enc->cur += 8;
		423	}
		424
		425	ecb_inline void
		426	encode_bool (enc_t *enc, int istrue)
		427	{
		428	encode_ch (enc, istrue ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
		429	}
		430
		431	// encodes an arrayref containing key-value pairs as CBOR map
		432	ecb_inline void
		433	encode_array_as_map (enc_t *enc, SV *sv)
		434	{
		435	if (enc->depth >= enc->cbor.max_depth)
		436	croak (ERR_NESTING_EXCEEDED);
		437
		438	++enc->depth;
		439
		440	// as_map does error checking for us, but we re-check in case
		441	// things have changed.
		442
		443	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
		444	croak ("CBOR::XS::as_map requires an array reference (did you change the array after calling as_map?)");
		445
		446	AV *av = (AV *)SvRV (sv);
		447	int i, len = av_len (av);
		448
		449	if (!(len & 1))
		450	croak ("CBOR::XS::as_map requires an even number of elements (did you change the array after calling as_map?)");
		451
		452	encode_uint (enc, MAJOR_MAP, (len + 1) >> 1);
		453
		454	for (i = 0; i <= len; ++i)
		455	{
		456	SV **svp = av_fetch (av, i, 0);
		457	encode_sv (enc, svp ? *svp : &PL_sv_undef);
		458	}
		459
		460	--enc->depth;
		461	}
		462
		463	ecb_inline void
		464	encode_forced (enc_t *enc, UV type, SV *sv)
		465	{
		466	switch (type)
		467	{
		468	case AS_CBOR:
		469	{
		470	STRLEN len;
		471	char *str = SvPVbyte (sv, len);
		472
		473	need (enc, len);
		474	memcpy (enc->cur, str, len);
		475	enc->cur += len;
		476	}
		477	break;
		478
		479	case AS_BYTES:
		480	{
		481	STRLEN len;
		482	char *str = SvPVbyte (sv, len);
		483	encode_strref (enc, 0, 0, str, len);
		484	}
		485	break;
		486
		487	case AS_TEXT:
		488	{
		489	STRLEN len;
		490	char *str = SvPVutf8 (sv, len);
		491	encode_strref (enc, 1, 1, str, len);
		492	}
		493	break;
		494
		495	case AS_INT: encode_int (enc, sv); break;
		496
		497	case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
		498	case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
		499	case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
		500
		501	case AS_MAP: encode_array_as_map (enc, sv); break;
		502
		503	default:
		504	croak ("encountered malformed CBOR::XS::Tagged object");
		505	}
		506	}
270		507
271	static void	508	static void
272	encode_av (enc_t *enc, AV *av)	509	encode_av (enc_t *enc, AV *av)
273	{	510	{
274	int i, len = av_len (av);	511	int i, len = av_len (av);
…		…
276	if (enc->depth >= enc->cbor.max_depth)	513	if (enc->depth >= enc->cbor.max_depth)
277	croak (ERR_NESTING_EXCEEDED);	514	croak (ERR_NESTING_EXCEEDED);
278		515
279	++enc->depth;	516	++enc->depth;
280		517
281	encode_uint (enc, 0x80, len + 1);	518	encode_uint (enc, MAJOR_ARRAY, len + 1);
282		519
		520	if (ecb_expect_false (SvMAGICAL (av)))
283	for (i = 0; i <= len; ++i)	521	for (i = 0; i <= len; ++i)
284	{	522	{
285	SV **svp = av_fetch (av, i, 0);	523	SV **svp = av_fetch (av, i, 0);
286	encode_sv (enc, svp ? *svp : &PL_sv_undef);	524	encode_sv (enc, svp ? *svp : &PL_sv_undef);
287	}	525	}
		526	else
		527	for (i = 0; i <= len; ++i)
		528	{
		529	SV *sv = AvARRAY (av)[i];
		530	encode_sv (enc, sv ? sv : &PL_sv_undef);
		531	}
288		532
289	--enc->depth;	533	--enc->depth;
290	}	534	}
291		535
292	static void	536	static void
…		…
300	++enc->depth;	544	++enc->depth;
301		545
302	int pairs = hv_iterinit (hv);	546	int pairs = hv_iterinit (hv);
303	int mg = SvMAGICAL (hv);	547	int mg = SvMAGICAL (hv);
304		548
305	if (mg)	549	if (ecb_expect_false (mg))
306	encode_ch (enc, 0xa0 | 31);	550	encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
307	else	551	else
308	encode_uint (enc, 0xa0, pairs);	552	encode_uint (enc, MAJOR_MAP, pairs);
309		553
310	while ((he = hv_iternext (hv)))	554	while ((he = hv_iternext (hv)))
311	{	555	{
312	if (HeKLEN (he) == HEf_SVKEY)	556	if (HeKLEN (he) == HEf_SVKEY)
313	encode_sv (enc, HeSVKEY (he));	557	encode_sv (enc, HeSVKEY (he));
314	else	558	else
315	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	559	encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
316		560
317	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	561	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
318	}	562	}
319		563
320	if (mg)	564	if (ecb_expect_false (mg))
321	encode_ch (enc, 0xe0 | 31);	565	encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
322		566
323	--enc->depth;	567	--enc->depth;
324	}	568	}
325		569
326	// encode objects, arrays and special \0=false and \1=true values.	570	// encode objects, arrays and special \0=false and \1=true values.
…		…
345		589
346	HV *stash = SvSTASH (sv);	590	HV *stash = SvSTASH (sv);
347		591
348	if (stash == boolean_stash)	592	if (stash == boolean_stash)
349	{	593	{
350	encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);	594	encode_bool (enc, SvIV (sv));
351	return;	595	return;
352	}	596	}
353	else if (stash == error_stash)	597	else if (stash == error_stash)
354	{	598	{
355	encode_ch (enc, 0xe0 | 23);	599	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
356	return;	600	return;
357	}	601	}
358	else if (stash == tagged_stash)	602	else if (stash == tagged_stash)
359	{	603	{
360	if (svt != SVt_PVAV)	604	if (svt != SVt_PVAV)
361	croak ("encountered CBOR::XS::Tagged object that isn't an array");	605	croak ("encountered CBOR::XS::Tagged object that isn't an array");
362		606
		607	switch (av_len ((AV *)sv))
		608	{
		609	case 2-1:
		610	// actually a tagged value
363	encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));	611	encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
364	encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));	612	encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
		613	break;
		614
		615	case 3-1:
		616	// a forced type [value, type, undef]
		617	encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
		618	break;
		619
		620	default:
		621	croak ("encountered malformed CBOR::XS::Tagged object");
		622	}
365		623
366	return;	624	return;
367	}	625	}
368	}	626	}
369		627
370	if (ecb_expect_false (SvREFCNT (sv) > 1)	628	if (ecb_expect_false (SvREFCNT (sv) > 1)
371	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))	629	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
372	{	630	{
373	if (!enc->shareable)	631	if (ecb_expect_false (!enc->shareable))
374	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());	632	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
375		633
376	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);	634	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
377		635
378	if (SvOK (*svp))	636	if (SvOK (*svp))
379	{	637	{
380	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);	638	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381	encode_uint (enc, 0x00, SvUV (*svp));	639	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
382	return;	640	return;
383	}	641	}
384	else	642	else
385	{	643	{
386	sv_setuv (*svp, enc->shareable_idx);	644	sv_setuv (*svp, enc->shareable_idx);
…		…
392	if (ecb_expect_false (SvOBJECT (sv)))	650	if (ecb_expect_false (SvOBJECT (sv)))
393	{	651	{
394	HV *stash = SvSTASH (sv);	652	HV *stash = SvSTASH (sv);
395	GV *method;	653	GV *method;
396		654
		655	if (enc->cbor.flags & F_FORBID_OBJECTS)
		656	croak ("encountered object '%s', but forbid_objects is enabled",
		657	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
397	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	658	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
398	{	659	{
399	dSP;	660	dSP;
400		661
401	ENTER; SAVETMPS; PUSHMARK (SP);	662	ENTER; SAVETMPS;
		663	PUSHMARK (SP);
402	// we re-bless the reference to get overload and other niceties right	664	// we re-bless the reference to get overload and other niceties right
403	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	665	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
404		666
405	PUTBACK;	667	PUTBACK;
406	// G_SCALAR ensures that return value is 1	668	// G_SCALAR ensures that return value is 1
…		…
419	}	681	}
420	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	682	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
421	{	683	{
422	dSP;	684	dSP;
423		685
424	ENTER; SAVETMPS; PUSHMARK (SP);	686	ENTER; SAVETMPS;
		687	PUSHMARK (SP);
425	EXTEND (SP, 2);	688	EXTEND (SP, 2);
426	// we re-bless the reference to get overload and other niceties right	689	// we re-bless the reference to get overload and other niceties right
427	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	690	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
428	PUSHs (sv_cbor);	691	PUSHs (sv_cbor);
429		692
…		…
434	// catch this surprisingly common error	697	// catch this surprisingly common error
435	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	698	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));	699	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
437		700
438	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	701	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
439	encode_uint (enc, 0x80, count + 1);	702	encode_uint (enc, MAJOR_ARRAY, count + 1);
440	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	703	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
441		704
442	while (count)	705	{
		706	int i;
		707
		708	for (i = 0; i < count; ++i)
443	encode_sv (enc, SP[1 - count--]);	709	encode_sv (enc, SP[i + 1 - count]);
		710
		711	SP -= count;
		712	}
444		713
445	PUTBACK;	714	PUTBACK;
446		715
447	FREETMPS; LEAVE;	716	FREETMPS; LEAVE;
448	}	717	}
…		…
466	{	735	{
467	double nv = SvNVX (sv);	736	double nv = SvNVX (sv);
468		737
469	need (enc, 9);	738	need (enc, 9);
470		739
471	if (ecb_expect_false (nv == (U32)nv))	740	if (ecb_expect_false (nv == (NV)(U32)nv))
472	encode_uint (enc, 0x00, (U32)nv);	741	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
473	//TODO: maybe I32?	742	//TODO: maybe I32?
474	else if (ecb_expect_false (nv == (float)nv))	743	else if (ecb_expect_false (nv == (float)nv))
475	{	744	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	745	else
487	{	746	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	}	747	}
499		748
500	static void	749	static void
501	encode_sv (enc_t *enc, SV *sv)	750	encode_sv (enc_t *enc, SV *sv)
502	{	751	{
…		…
504		753
505	if (SvPOKp (sv))	754	if (SvPOKp (sv))
506	{	755	{
507	STRLEN len;	756	STRLEN len;
508	char *str = SvPV (sv, len);	757	char *str = SvPV (sv, len);
509	encode_strref (enc, SvUTF8 (sv), str, len);	758	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
510	}	759	}
511	else if (SvNOKp (sv))	760	else if (SvNOKp (sv))
512	encode_nv (enc, sv);	761	encode_nv (enc, sv);
513	else if (SvIOKp (sv))	762	else if (SvIOKp (sv))
514	{	763	encode_int (enc, 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))	764	else if (SvROK (sv))
523	encode_rv (enc, SvRV (sv));	765	encode_rv (enc, SvRV (sv));
524	else if (!SvOK (sv))	766	else if (!SvOK (sv))
525	encode_ch (enc, 0xe0 | 22);	767	encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
526	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	768	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
527	encode_ch (enc, 0xe0 | 23);	769	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
528	else	770	else
529	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	771	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
530	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	772	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
531	}	773	}
532		774
533	static SV *	775	static SV *
534	encode_cbor (SV *scalar, CBOR *cbor)	776	encode_cbor (SV *scalar, CBOR *cbor)
535	{	777	{
536	enc_t enc = { };	778	enc_t enc = { 0 };
537		779
538	enc.cbor = *cbor;	780	enc.cbor = *cbor;
539	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	781	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
540	enc.cur = SvPVX (enc.sv);	782	enc.cur = SvPVX (enc.sv);
541	enc.end = SvEND (enc.sv);	783	enc.end = SvEND (enc.sv);
542		784
543	SvPOK_only (enc.sv);	785	SvPOK_only (enc.sv);
544		786
545	if (cbor->flags & F_PACK_STRINGS)	787	if (cbor->flags & F_PACK_STRINGS)
546	{	788	{
…		…
573	U32 depth; // recursion depth	815	U32 depth; // recursion depth
574	U32 maxdepth; // recursion depth limit	816	U32 maxdepth; // recursion depth limit
575	AV *shareable;	817	AV *shareable;
576	AV *stringref;	818	AV *stringref;
577	SV *decode_tagged;	819	SV *decode_tagged;
		820	SV *err_sv; // optional sv for error, needs to be freed
578	} dec_t;	821	} dec_t;
579		822
580	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	823	// set dec->err to ERRSV
		824	ecb_cold static void
		825	err_errsv (dec_t *dec)
		826	{
		827	if (!dec->err)
		828	{
		829	dec->err_sv = newSVsv (ERRSV);
581		830
582	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	831	// chop off the trailing \n
		832	SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
		833	*SvEND (dec->err_sv) = 0;
583		834
		835	dec->err = SvPVutf8_nolen (dec->err_sv);
		836	}
		837	}
		838
		839	// the following functions are used to reduce code size and help the compiler to optimise
		840	ecb_cold static void
		841	err_set (dec_t *dec, const char *reason)
		842	{
		843	if (!dec->err)
		844	dec->err = reason;
		845	}
		846
		847	ecb_cold static void
		848	err_unexpected_end (dec_t *dec)
		849	{
		850	err_set (dec, "unexpected end of CBOR data");
		851	}
		852
		853	#define ERR_DO(do) SB do; goto fail; SE
		854	#define ERR(reason) ERR_DO (err_set (dec, reason))
		855	#define ERR_ERRSV ERR_DO (err_errsv (dec))
		856
		857	#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
		858
584	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	859	#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	860	#define DEC_DEC_DEPTH --dec->depth
586		861
587	static UV	862	static UV
588	decode_uint (dec_t *dec)	863	decode_uint (dec_t *dec)
589	{	864	{
590	switch (*dec->cur & 31)	865	U8 m = *dec->cur & MINOR_MASK;
591	{	866	++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		867
597	case 24:	868	if (ecb_expect_true (m < LENGTH_EXT1))
		869	return m;
		870	else if (ecb_expect_true (m == LENGTH_EXT1))
		871	{
598	WANT (2);	872	WANT (1);
599	dec->cur += 2;	873	dec->cur += 1;
600	return dec->cur[-1];	874	return dec->cur[-1];
601		875	}
602	case 25:	876	else if (ecb_expect_true (m == LENGTH_EXT2))
		877	{
603	WANT (3);	878	WANT (2);
604	dec->cur += 3;	879	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	880	return (((UV)dec->cur[-2]) << 8)
606	| ((UV)dec->cur[-1]);	881	| ((UV)dec->cur[-1]);
607		882	}
608	case 26:	883	else if (ecb_expect_true (m == LENGTH_EXT4))
		884	{
609	WANT (5);	885	WANT (4);
610	dec->cur += 5;	886	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	887	return (((UV)dec->cur[-4]) << 24)
612	| (((UV)dec->cur[-3]) << 16)	888	| (((UV)dec->cur[-3]) << 16)
613	| (((UV)dec->cur[-2]) << 8)	889	| (((UV)dec->cur[-2]) << 8)
614	| ((UV)dec->cur[-1]);	890	| ((UV)dec->cur[-1]);
615		891	}
616	case 27:	892	else if (ecb_expect_true (m == LENGTH_EXT8))
		893	{
617	WANT (9);	894	WANT (8);
618	dec->cur += 9;	895	dec->cur += 8;
619		896
620	return	897	return
621	#if UVSIZE < 8	898	#if UVSIZE < 8
622	0	899	0
623	#else	900	#else
624	(((UV)dec->cur[-8]) << 56)	901	(((UV)dec->cur[-8]) << 56)
625	| (((UV)dec->cur[-7]) << 48)	902	| (((UV)dec->cur[-7]) << 48)
626	| (((UV)dec->cur[-6]) << 40)	903	| (((UV)dec->cur[-6]) << 40)
627	| (((UV)dec->cur[-5]) << 32)	904	| (((UV)dec->cur[-5]) << 32)
628	#endif	905	#endif
629	| (((UV)dec->cur[-4]) << 24)	906	| (((UV)dec->cur[-4]) << 24)
630	| (((UV)dec->cur[-3]) << 16)	907	| (((UV)dec->cur[-3]) << 16)
631	| (((UV)dec->cur[-2]) << 8)	908	| (((UV)dec->cur[-2]) << 8)
632	| ((UV)dec->cur[-1]);	909	| ((UV)dec->cur[-1]);
633		910	}
634	default:	911	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	912	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		913
638	fail:	914	fail:
639	return 0;	915	return 0;
640	}	916	}
641		917
…		…
646	{	922	{
647	AV *av = newAV ();	923	AV *av = newAV ();
648		924
649	DEC_INC_DEPTH;	925	DEC_INC_DEPTH;
650		926
651	if ((*dec->cur & 31) == 31)	927	if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
652	{	928	{
653	++dec->cur;	929	++dec->cur;
654		930
655	for (;;)	931	for (;;)
656	{	932	{
657	WANT (1);	933	WANT (1);
658		934
659	if (*dec->cur == (0xe0 | 31))	935	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
660	{	936	{
661	++dec->cur;	937	++dec->cur;
662	break;	938	break;
663	}	939	}
664		940
665	av_push (av, decode_sv (dec));	941	av_push (av, decode_sv (dec));
666	}	942	}
667	}	943	}
668	else	944	else
669	{	945	{
670	int i, len = decode_uint (dec);	946	UV i, len = decode_uint (dec);
671		947
		948	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);	949	av_fill (av, len - 1);
673		950
674	for (i = 0; i < len; ++i)	951	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	952	AvARRAY (av)[i] = decode_sv (dec);
676	}	953	}
677		954
678	DEC_DEC_DEPTH;	955	DEC_DEC_DEPTH;
679	return newRV_noinc ((SV *)av);	956	return newRV_noinc ((SV *)av);
680		957
681	fail:	958	fail:
682	SvREFCNT_dec (av);	959	SvREFCNT_dec_NN (av);
683	DEC_DEC_DEPTH;	960	DEC_DEC_DEPTH;
684	return &PL_sv_undef;	961	return &PL_sv_undef;
685	}	962	}
686		963
687	static void	964	static void
…		…
689	{	966	{
690	// for speed reasons, we specialcase single-string	967	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	968	// byte or utf-8 strings as keys, but only when !stringref
692		969
693	if (ecb_expect_true (!dec->stringref))	970	if (ecb_expect_true (!dec->stringref))
694	if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)	971	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	972	{
696	I32 len = decode_uint (dec);	973	STRLEN len = decode_uint (dec);
697	char *key = (char *)dec->cur;	974	char *key = (char *)dec->cur;
698		975
		976	WANT (len);
699	dec->cur += len;	977	dec->cur += len;
700		978
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);	979	hv_store (hv, key, len, decode_sv (dec), 0);
705		980
706	return;	981	return;
707	}	982	}
708	else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)	983	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	984	{
710	I32 len = decode_uint (dec);	985	STRLEN len = decode_uint (dec);
711	char *key = (char *)dec->cur;	986	char *key = (char *)dec->cur;
712		987
		988	WANT (len);
713	dec->cur += len;	989	dec->cur += len;
714		990
715	if (ecb_expect_false (dec->stringref))	991	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	992	if (!is_utf8_string (key, len))
		993	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
717		994
718	hv_store (hv, key, -len, decode_sv (dec), 0);	995	hv_store (hv, key, -len, decode_sv (dec), 0);
719		996
720	return;	997	return;
721	}	998	}
722		999
723	SV *k = decode_sv (dec);	1000	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	1001	SV *v = decode_sv (dec);
725		1002
		1003	// we leak memory if uncaught exceptions are thrown by random magical
		1004	// methods, and this is hopefully the only place where it can happen,
		1005	// so if there is a chance of an exception, take the very slow path.
		1006	// since catching exceptions is "undocumented/internal/forbidden" by
		1007	// the new p5p powers, we need to call out to a perl function :/
		1008	if (ecb_expect_false (SvAMAGIC (k)))
		1009	{
		1010	dSP;
		1011
		1012	ENTER; SAVETMPS;
		1013	PUSHMARK (SP);
		1014	EXTEND (SP, 3);
		1015	PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
		1016	PUSHs (sv_2mortal (k));
		1017	PUSHs (sv_2mortal (v));
		1018
		1019	PUTBACK;
		1020	call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
		1021	SPAGAIN;
		1022
		1023	FREETMPS; LEAVE;
		1024
		1025	if (SvTRUE (ERRSV))
		1026	ERR_ERRSV;
		1027
		1028	return;
		1029	}
		1030
726	hv_store_ent (hv, k, v, 0);	1031	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	1032	SvREFCNT_dec_NN (k);
		1033
		1034	fail:
		1035	;
728	}	1036	}
729		1037
730	static SV *	1038	static SV *
731	decode_hv (dec_t *dec)	1039	decode_hv (dec_t *dec)
732	{	1040	{
733	HV *hv = newHV ();	1041	HV *hv = newHV ();
734		1042
735	DEC_INC_DEPTH;	1043	DEC_INC_DEPTH;
736		1044
737	if ((*dec->cur & 31) == 31)	1045	if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
738	{	1046	{
739	++dec->cur;	1047	++dec->cur;
740		1048
741	for (;;)	1049	for (;;)
742	{	1050	{
743	WANT (1);	1051	WANT (1);
744		1052
745	if (*dec->cur == (0xe0 | 31))	1053	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
746	{	1054	{
747	++dec->cur;	1055	++dec->cur;
748	break;	1056	break;
749	}	1057	}
750		1058
751	decode_he (dec, hv);	1059	decode_he (dec, hv);
752	}	1060	}
753	}	1061	}
754	else	1062	else
755	{	1063	{
756	int pairs = decode_uint (dec);	1064	UV pairs = decode_uint (dec);
		1065
		1066	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
757		1067
758	while (pairs--)	1068	while (pairs--)
759	decode_he (dec, hv);	1069	decode_he (dec, hv);
760	}	1070	}
761		1071
762	DEC_DEC_DEPTH;	1072	DEC_DEC_DEPTH;
763	return newRV_noinc ((SV *)hv);	1073	return newRV_noinc ((SV *)hv);
764		1074
765	fail:	1075	fail:
766	SvREFCNT_dec (hv);	1076	SvREFCNT_dec_NN (hv);
767	DEC_DEC_DEPTH;	1077	DEC_DEC_DEPTH;
768	return &PL_sv_undef;	1078	return &PL_sv_undef;
769	}	1079	}
770		1080
771	static SV *	1081	static SV *
772	decode_str (dec_t *dec, int utf8)	1082	decode_str (dec_t *dec, int utf8)
773	{	1083	{
774	SV *sv = 0;	1084	SV *sv = 0;
775		1085
776	if ((*dec->cur & 31) == 31)	1086	if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
777	{	1087	{
778	// indefinite length strings	1088	// indefinite length strings
779	++dec->cur;	1089	++dec->cur;
780		1090
781	unsigned char major = *dec->cur & 0xe0;	1091	U8 major = *dec->cur & MAJOR_MISC;
782		1092
783	sv = newSVpvn ("", 0);	1093	sv = newSVpvn ("", 0);
784		1094
785	for (;;)	1095	for (;;)
786	{	1096	{
787	WANT (1);	1097	WANT (1);
788		1098
789	if ((*dec->cur ^ major) >= 31)	1099	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 | 31))	1100	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
791	{	1101	{
792	++dec->cur;	1102	++dec->cur;
793	break;	1103	break;
794	}	1104	}
795	else	1105	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	1124	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	1125	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	1126	}
817		1127
818	if (utf8)	1128	if (utf8)
		1129	{
		1130	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		1131	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		1132	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		1133
819	SvUTF8_on (sv);	1134	SvUTF8_on (sv);
		1135	}
820		1136
821	return sv;	1137	return sv;
822		1138
823	fail:	1139	fail:
824	SvREFCNT_dec (sv);	1140	SvREFCNT_dec (sv);
…		…
843	sv = newRV_noinc (decode_sv (dec));	1159	sv = newRV_noinc (decode_sv (dec));
844	break;	1160	break;
845		1161
846	case CBOR_TAG_STRINGREF_NAMESPACE:	1162	case CBOR_TAG_STRINGREF_NAMESPACE:
847	{	1163	{
		1164	// do not use SAVETMPS/FREETMPS, as these will
		1165	// erase mortalised caches, e.g. "shareable"
848	ENTER; SAVETMPS;	1166	ENTER;
849		1167
850	SAVESPTR (dec->stringref);	1168	SAVESPTR (dec->stringref);
851	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());	1169	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
852		1170
853	sv = decode_sv (dec);	1171	sv = decode_sv (dec);
854		1172
855	FREETMPS; LEAVE;	1173	LEAVE;
856	}	1174	}
857	break;	1175	break;
858		1176
859	case CBOR_TAG_STRINGREF:	1177	case CBOR_TAG_STRINGREF:
860	{	1178	{
861	if ((*dec->cur >> 5) != 0)	1179	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1180	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		1181
864	UV idx = decode_uint (dec);	1182	UV idx = decode_uint (dec);
865		1183
866	if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))	1184	if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
867	ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");	1185	ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
868		1186
869	sv = newSVsv (AvARRAY (dec->stringref)[idx]);	1187	sv = newSVsv (AvARRAY (dec->stringref)[idx]);
870	}	1188	}
871	break;	1189	break;
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	1191	case CBOR_TAG_VALUE_SHAREABLE:
874	{	1192	{
875	if (ecb_expect_false (!dec->shareable))	1193	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());	1194	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
877		1195
		1196	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1197	{
878	sv = newSV (0);	1198	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1199	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		1200
881	SV *osv = decode_sv (dec);	1201	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	1202	sv_setsv (sv, osv);
883	SvREFCNT_dec_NN (osv);	1203	SvREFCNT_dec_NN (osv);
		1204	}
		1205	else
		1206	{
		1207	av_push (dec->shareable, &PL_sv_undef);
		1208	int idx = AvFILLp (dec->shareable);
		1209	sv = decode_sv (dec);
		1210	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		1211	}
884	}	1212	}
885	break;	1213	break;
886		1214
887	case CBOR_TAG_VALUE_SHAREDREF:	1215	case CBOR_TAG_VALUE_SHAREDREF:
888	{	1216	{
889	if ((*dec->cur >> 5) != 0)	1217	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1218	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		1219
892	UV idx = decode_uint (dec);	1220	UV idx = decode_uint (dec);
893		1221
894	if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))	1222	if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1223	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		1224
897	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1225	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1226
		1227	if (sv == &PL_sv_undef)
		1228	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
898	}	1229	}
899	break;	1230	break;
900		1231
901	case CBOR_TAG_PERL_OBJECT:	1232	case CBOR_TAG_PERL_OBJECT:
902	{	1233	{
		1234	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1235	goto filter;
		1236
903	sv = decode_sv (dec);	1237	sv = decode_sv (dec);
904		1238
905	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)	1239	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
906	ERR ("corrupted CBOR data (non-array perl object)");	1240	ERR ("corrupted CBOR data (non-array perl object)");
907		1241
…		…
917	if (!method)	1251	if (!method)
918	ERR ("cannot decode perl-object (package does not have a THAW method)");	1252	ERR ("cannot decode perl-object (package does not have a THAW method)");
919		1253
920	dSP;	1254	dSP;
921		1255
922	ENTER; SAVETMPS; PUSHMARK (SP);	1256	ENTER; SAVETMPS;
		1257	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1258	EXTEND (SP, len + 1);
924	// we re-bless the reference to get overload and other niceties right	1259	// we re-bless the reference to get overload and other niceties right
925	PUSHs (*av_fetch (av, 0, 1));	1260	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1261	PUSHs (sv_cbor);
927		1262
…		…
935	SPAGAIN;	1270	SPAGAIN;
936		1271
937	if (SvTRUE (ERRSV))	1272	if (SvTRUE (ERRSV))
938	{	1273	{
939	FREETMPS; LEAVE;	1274	FREETMPS; LEAVE;
940	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1275	ERR_ERRSV;
941	}	1276	}
942		1277
943	SvREFCNT_dec (sv);	1278	SvREFCNT_dec_NN (sv);
944	sv = SvREFCNT_inc (POPs);	1279	sv = SvREFCNT_inc (POPs);
945		1280
946	PUTBACK;	1281	PUTBACK;
947		1282
948	FREETMPS; LEAVE;	1283	FREETMPS; LEAVE;
949	}	1284	}
950	break;	1285	break;
951		1286
952	default:	1287	default:
		1288	filter:
953	{	1289	{
		1290	SV *tag_sv = newSVuv (tag);
		1291
954	sv = decode_sv (dec);	1292	sv = decode_sv (dec);
955		1293
956	dSP;	1294	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1295	ENTER; SAVETMPS;
		1296	PUSHMARK (SP);
958	EXTEND (SP, 2);	1297	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1298	PUSHs (tag_sv);
960	PUSHs (sv);	1299	PUSHs (sv);
961		1300
962	PUTBACK;	1301	PUTBACK;
963	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);	1302	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
964	SPAGAIN;	1303	SPAGAIN;
965		1304
966	if (SvTRUE (ERRSV))	1305	if (SvTRUE (ERRSV))
967	{	1306	{
		1307	SvREFCNT_dec_NN (tag_sv);
968	FREETMPS; LEAVE;	1308	FREETMPS; LEAVE;
969	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1309	ERR_ERRSV;
970	}	1310	}
971		1311
972	if (count)	1312	if (count)
973	{	1313	{
		1314	SvREFCNT_dec_NN (tag_sv);
974	SvREFCNT_dec (sv);	1315	SvREFCNT_dec_NN (sv);
975	sv = SvREFCNT_inc (POPs);	1316	sv = SvREFCNT_inc_NN (TOPs);
		1317	SP -= count;
976	}	1318	}
977	else	1319	else
978	{	1320	{
979	AV *av = newAV ();	1321	AV *av = newAV ();
980	av_push (av, newSVuv (tag));	1322	av_push (av, tag_sv);
981	av_push (av, sv);	1323	av_push (av, sv);
982		1324
983	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	1325	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
984	? cbor_tagged_stash	1326	? cbor_tagged_stash
985	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1327	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
1003	static SV *	1345	static SV *
1004	decode_sv (dec_t *dec)	1346	decode_sv (dec_t *dec)
1005	{	1347	{
1006	WANT (1);	1348	WANT (1);
1007		1349
1008	switch (*dec->cur >> 5)	1350	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1351	{
1010	case 0: // unsigned int	1352	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1353	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1354	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1355	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1356	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1357	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1358	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1359
1018	case 4: // array	1360	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)	1361	switch (*dec->cur++ & MINOR_MASK)
1026	{	1362	{
1027	case 20:	1363	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1364	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1365	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1366	#endif
1031	return newSVsv (types_false);	1367	return newSVsv (types_false);
1032	case 21:	1368	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1369	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1370	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1371	#endif
1036	return newSVsv (types_true);	1372	return newSVsv (types_true);
1037	case 22:	1373	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1374	return newSVsv (&PL_sv_undef);
1039	case 23:	1375	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1376	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1377	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1378	#endif
1043	return newSVsv (types_error);	1379	return newSVsv (types_error);
1044		1380
1045	case 25:	1381	case MISC_FLOAT16:
1046	{	1382	{
1047	WANT (2);	1383	WANT (2);
1048		1384
1049	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];	1385	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1050	dec->cur += 2;	1386	dec->cur += 2;
1051		1387
1052	return newSVnv (ecb_binary16_to_float (fp));	1388	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1389	}
1054		1390
1055	case 26:	1391	case MISC_FLOAT32:
1056	{	1392	{
1057	uint32_t fp;	1393	uint32_t fp;
1058	WANT (4);	1394	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1395	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1396	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1399	fp = ecb_bswap32 (fp);
1064		1400
1065	return newSVnv (ecb_binary32_to_float (fp));	1401	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1402	}
1067		1403
1068	case 27:	1404	case MISC_FLOAT64:
1069	{	1405	{
1070	uint64_t fp;	1406	uint64_t fp;
1071	WANT (8);	1407	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1408	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1409	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1412	fp = ecb_bswap64 (fp);
1077		1413
1078	return newSVnv (ecb_binary64_to_double (fp));	1414	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1415	}
1080		1416
1081	// 0..19 unassigned	1417	// 0..19 unassigned simple
1082	// 24 reserved + unassigned (reserved values are not encodable)	1418	// 24 reserved + unassigned simple (reserved values are not encodable)
		1419	// 28-30 unassigned misc
		1420	// 31 break code
1083	default:	1421	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1422	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1423	}
1086		1424
1087	break;	1425	break;
1088	}	1426	}
1089		1427
…		…
1092	}	1430	}
1093		1431
1094	static SV *	1432	static SV *
1095	decode_cbor (SV *string, CBOR *cbor, char **offset_return)	1433	decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1096	{	1434	{
1097	dec_t dec = { };	1435	dec_t dec = { 0 };
1098	SV *sv;	1436	SV *sv;
1099	STRLEN len;	1437	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1438	char *data = SvPVbyte (string, len);
1101		1439
1102	if (len > cbor->max_size && cbor->max_size)	1440	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1454	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1455	dec.err = "garbage after CBOR object";
1118		1456
1119	if (dec.err)	1457	if (dec.err)
1120	{	1458	{
		1459	if (dec.shareable)
		1460	{
		1461	// need to break cyclic links, which would all be in shareable
		1462	int i;
		1463	SV **svp;
		1464
		1465	for (i = av_len (dec.shareable) + 1; i--; )
		1466	if ((svp = av_fetch (dec.shareable, i, 0)))
		1467	sv_setsv (*svp, &PL_sv_undef);
		1468	}
		1469
1121	SvREFCNT_dec (sv);	1470	SvREFCNT_dec_NN (sv);
		1471
		1472	if (dec.err_sv)
		1473	sv_2mortal (dec.err_sv);
		1474
1122	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);	1475	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1123	}	1476	}
1124		1477
1125	sv = sv_2mortal (sv);	1478	sv = sv_2mortal (sv);
1126		1479
1127	return sv;	1480	return sv;
1128	}	1481	}
1129		1482
		1483	/////////////////////////////////////////////////////////////////////////////
		1484	// incremental parser
		1485
		1486	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1487
		1488	// returns 0 for notyet, 1 for success or error
		1489	static int
		1490	incr_parse (CBOR *self, SV *cborstr)
		1491	{
		1492	STRLEN cur;
		1493	SvPV (cborstr, cur);
		1494
		1495	while (ecb_expect_true (self->incr_need <= cur))
		1496	{
		1497	// table of integer count bytes
		1498	static I8 incr_len[MINOR_MASK + 1] = {
		1499	0, 0, 0, 0, 0, 0, 0, 0,
		1500	0, 0, 0, 0, 0, 0, 0, 0,
		1501	0, 0, 0, 0, 0, 0, 0, 0,
		1502	1, 2, 4, 8,-1,-1,-1,-2
		1503	};
		1504
		1505	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1506	U8 m = *p & MINOR_MASK;
		1507	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1508	I8 ilen = incr_len[m];
		1509
		1510	self->incr_need = self->incr_pos + 1;
		1511
		1512	if (ecb_expect_false (ilen < 0))
		1513	{
		1514	if (m != MINOR_INDEF)
		1515	return 1; // error
		1516
		1517	if (*p == (MAJOR_MISC | MINOR_INDEF))
		1518	{
		1519	if (count >= 0)
		1520	return 1; // error
		1521
		1522	count = 1;
		1523	}
		1524	else
		1525	{
		1526	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1527	count = -1;
		1528	}
		1529	}
		1530	else
		1531	{
		1532	self->incr_need += ilen;
		1533	if (ecb_expect_false (self->incr_need > cur))
		1534	return 0;
		1535
		1536	int major = *p >> MAJOR_SHIFT;
		1537
		1538	switch (major)
		1539	{
		1540	case MAJOR_TAG >> MAJOR_SHIFT:
		1541	++count; // tags merely prefix another value
		1542	break;
		1543
		1544	case MAJOR_BYTES >> MAJOR_SHIFT:
		1545	case MAJOR_TEXT >> MAJOR_SHIFT:
		1546	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1547	case MAJOR_MAP >> MAJOR_SHIFT:
		1548	{
		1549	UV len;
		1550
		1551	if (ecb_expect_false (ilen))
		1552	{
		1553	len = 0;
		1554
		1555	do {
		1556	len = (len << 8) | *++p;
		1557	} while (--ilen);
		1558	}
		1559	else
		1560	len = m;
		1561
		1562	switch (major)
		1563	{
		1564	case MAJOR_BYTES >> MAJOR_SHIFT:
		1565	case MAJOR_TEXT >> MAJOR_SHIFT:
		1566	self->incr_need += len;
		1567	if (ecb_expect_false (self->incr_need > cur))
		1568	return 0;
		1569
		1570	break;
		1571
		1572	case MAJOR_MAP >> MAJOR_SHIFT:
		1573	len <<= 1;
		1574	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1575	if (len)
		1576	{
		1577	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1578	count = len + 1;
		1579	}
		1580	break;
		1581	}
		1582	}
		1583	}
		1584	}
		1585
		1586	self->incr_pos = self->incr_need;
		1587
		1588	if (count > 0)
		1589	{
		1590	while (!--count)
		1591	{
		1592	if (!AvFILLp (self->incr_count))
		1593	return 1; // done
		1594
		1595	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1596	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1597	}
		1598
		1599	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1600	}
		1601	}
		1602
		1603	return 0;
		1604	}
		1605
		1606
1130	/////////////////////////////////////////////////////////////////////////////	1607	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1608	// XS interface functions
1132		1609
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1610	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1611
…		…
1146		1623
1147	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1624	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1148		1625
1149	sv_cbor = newSVpv ("CBOR", 0);	1626	sv_cbor = newSVpv ("CBOR", 0);
1150	SvREADONLY_on (sv_cbor);	1627	SvREADONLY_on (sv_cbor);
		1628
		1629	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1151	}	1630	}
1152		1631
1153	PROTOTYPES: DISABLE	1632	PROTOTYPES: DISABLE
1154		1633
1155	void CLONE (...)	1634	void CLONE (...)
…		…
1174	void shrink (CBOR *self, int enable = 1)	1653	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1654	ALIAS:
1176	shrink = F_SHRINK	1655	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1656	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1657	allow_sharing = F_ALLOW_SHARING
		1658	allow_cycles = F_ALLOW_CYCLES
		1659	forbid_objects = F_FORBID_OBJECTS
1179	pack_strings = F_PACK_STRINGS	1660	pack_strings = F_PACK_STRINGS
		1661	text_keys = F_TEXT_KEYS
		1662	text_strings = F_TEXT_STRINGS
		1663	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1664	PPCODE:
1181	{	1665	{
1182	if (enable)	1666	if (enable)
1183	self->flags |= ix;	1667	self->flags |= ix;
1184	else	1668	else
…		…
1190	void get_shrink (CBOR *self)	1674	void get_shrink (CBOR *self)
1191	ALIAS:	1675	ALIAS:
1192	get_shrink = F_SHRINK	1676	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1677	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1678	get_allow_sharing = F_ALLOW_SHARING
		1679	get_allow_cycles = F_ALLOW_CYCLES
		1680	get_forbid_objects = F_FORBID_OBJECTS
1195	get_pack_strings = F_PACK_STRINGS	1681	get_pack_strings = F_PACK_STRINGS
		1682	get_text_keys = F_TEXT_KEYS
		1683	get_text_strings = F_TEXT_STRINGS
		1684	get_validate_utf8 = F_VALIDATE_UTF8
1196	PPCODE:	1685	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1686	XPUSHs (boolSV (self->flags & ix));
1198		1687
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1688	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1689	PPCODE:
…		…
1249	EXTEND (SP, 2);	1738	EXTEND (SP, 2);
1250	PUSHs (sv);	1739	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1740	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1741	}
1253		1742
		1743	void incr_parse (CBOR *self, SV *cborstr)
		1744	ALIAS:
		1745	incr_parse_multiple = 1
		1746	PPCODE:
		1747	{
		1748	if (SvUTF8 (cborstr))
		1749	sv_utf8_downgrade (cborstr, 0);
		1750
		1751	if (!self->incr_count)
		1752	{
		1753	self->incr_count = newAV ();
		1754	self->incr_pos = 0;
		1755	self->incr_need = 1;
		1756
		1757	av_push (self->incr_count, newSViv (1));
		1758	}
		1759
		1760	do
		1761	{
		1762	if (!incr_parse (self, cborstr))
		1763	{
		1764	if (self->incr_need > self->max_size && self->max_size)
		1765	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1766	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1767
		1768	break;
		1769	}
		1770
		1771	SV *sv;
		1772	char *offset;
		1773
		1774	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1775	XPUSHs (sv);
		1776
		1777	sv_chop (cborstr, offset);
		1778
		1779	av_clear (self->incr_count);
		1780	av_push (self->incr_count, newSViv (1));
		1781
		1782	self->incr_pos = 0;
		1783	self->incr_need = self->incr_pos + 1;
		1784	}
		1785	while (ix);
		1786	}
		1787
		1788	void incr_reset (CBOR *self)
		1789	CODE:
		1790	{
		1791	SvREFCNT_dec (self->incr_count);
		1792	self->incr_count = 0;
		1793	}
		1794
1254	void DESTROY (CBOR *self)	1795	void DESTROY (CBOR *self)
1255	PPCODE:	1796	PPCODE:
1256	cbor_free (self);	1797	cbor_free (self);
1257		1798
1258	PROTOTYPES: ENABLE	1799	PROTOTYPES: ENABLE
1259		1800
1260	void encode_cbor (SV *scalar)	1801	void encode_cbor (SV *scalar)
		1802	ALIAS:
		1803	encode_cbor = 0
		1804	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1805	PPCODE:
1262	{	1806	{
1263	CBOR cbor;	1807	CBOR cbor;
1264	cbor_init (&cbor);	1808	cbor_init (&cbor);
		1809	cbor.flags |= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1810	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1811	XPUSHs (scalar);
1267	}	1812	}
1268		1813
1269	void decode_cbor (SV *cborstr)	1814	void decode_cbor (SV *cborstr)
…		…
1273	cbor_init (&cbor);	1818	cbor_init (&cbor);
1274	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;	1819	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1275	XPUSHs (cborstr);	1820	XPUSHs (cborstr);
1276	}	1821	}
1277		1822
		1823	#ifdef __AFL_COMPILER
		1824
		1825	void
		1826	afl_init ()
		1827	CODE:
		1828	__AFL_INIT ();
		1829
		1830	int
		1831	afl_loop (unsigned int count = 10000)
		1832	CODE:
		1833	RETVAL = __AFL_LOOP (count);
		1834	OUTPUT:
		1835	RETVAL
		1836
		1837	#endif
		1838

Diff Legend

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