ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.29 by root, Thu Nov 28 09:13:12 2013 UTC vs.
Revision 1.71 by root, Fri Mar 19 17:30:27 2021 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 //TODO	120	#define F_ALLOW_SHARING 0x00000004UL
65	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	121	#define F_ALLOW_CYCLES 0x00000008UL
		122	#define F_FORBID_OBJECTS 0x00000010UL
		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
…		…
134	#endif	201	#endif
135	}	202	}
136	}	203	}
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 STRLEN
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 (len < 24)	260	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major | len;	261	*enc->cur++ = major | len;
196	else if (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);
233	}	311	}
234		312
235	static void	313	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		314	ecb_noinline static void
236	encode_str (enc_t *enc, int utf8, char *str, STRLEN len)	315	encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
237	{	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	}
		334	}
		335
		336	ecb_inline void
		337	encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
		338	{
		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);
		347	need (enc, len);
		348	memcpy (enc->cur, str, len);
		349	enc->cur += len;
		350	}
		351
		352	ecb_inline void
		353	encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
		354	{
238	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	355	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
239	{	356	{
240	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	357	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
241		358
242	if (SvOK (*svp))	359	if (SvOK (*svp))
243	{	360	{
244	// already registered, use stringref	361	// already registered, use stringref
245	encode_tag (enc, CBOR_TAG_STRINGREF);	362	encode_tag (enc, CBOR_TAG_STRINGREF);
246	encode_uint (enc, 0x00, SvUV (*svp));	363	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
247	return;	364	return;
248	}	365	}
249	else if (len >= minimum_string_length (enc->stringref_idx))	366	else if (len >= minimum_string_length (enc->stringref_idx))
250	{	367	{
251	// register only	368	// register only
252	sv_setuv (*svp, enc->stringref_idx);	369	sv_setuv (*svp, enc->stringref_idx);
253	++enc->stringref_idx;	370	++enc->stringref_idx;
254	}	371	}
255	}	372	}
256		373
257	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	374	encode_str (enc, upgrade_utf8, utf8, str, len);
		375	}
		376
		377	ecb_inline void
		378	encode_float16 (enc_t *enc, NV nv)
		379	{
258	need (enc, len);	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
259	memcpy (enc->cur, str, len);	389	memcpy (enc->cur, &fp, 2);
260	enc->cur += len;	390	enc->cur += 2;
261	}	391	}
262		392
263	static void encode_sv (enc_t *enc, SV *sv);	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	}
264		507
265	static void	508	static void
266	encode_av (enc_t *enc, AV *av)	509	encode_av (enc_t *enc, AV *av)
267	{	510	{
268	int i, len = av_len (av);	511	int i, len = av_len (av);
…		…
270	if (enc->depth >= enc->cbor.max_depth)	513	if (enc->depth >= enc->cbor.max_depth)
271	croak (ERR_NESTING_EXCEEDED);	514	croak (ERR_NESTING_EXCEEDED);
272		515
273	++enc->depth;	516	++enc->depth;
274		517
275	encode_uint (enc, 0x80, len + 1);	518	encode_uint (enc, MAJOR_ARRAY, len + 1);
276		519
		520	if (ecb_expect_false (SvMAGICAL (av)))
277	for (i = 0; i <= len; ++i)	521	for (i = 0; i <= len; ++i)
278	{	522	{
279	SV **svp = av_fetch (av, i, 0);	523	SV **svp = av_fetch (av, i, 0);
280	encode_sv (enc, svp ? *svp : &PL_sv_undef);	524	encode_sv (enc, svp ? *svp : &PL_sv_undef);
281	}	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	}
282		532
283	--enc->depth;	533	--enc->depth;
284	}	534	}
285		535
286	static void	536	static void
…		…
294	++enc->depth;	544	++enc->depth;
295		545
296	int pairs = hv_iterinit (hv);	546	int pairs = hv_iterinit (hv);
297	int mg = SvMAGICAL (hv);	547	int mg = SvMAGICAL (hv);
298		548
299	if (mg)	549	if (ecb_expect_false (mg))
300	encode_ch (enc, 0xa0 | 31);	550	encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
301	else	551	else
302	encode_uint (enc, 0xa0, pairs);	552	encode_uint (enc, MAJOR_MAP, pairs);
303		553
304	while ((he = hv_iternext (hv)))	554	while ((he = hv_iternext (hv)))
305	{	555	{
306	if (HeKLEN (he) == HEf_SVKEY)	556	if (HeKLEN (he) == HEf_SVKEY)
307	encode_sv (enc, HeSVKEY (he));	557	encode_sv (enc, HeSVKEY (he));
308	else	558	else
309	encode_str (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));
310		560
311	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));
312	}	562	}
313		563
314	if (mg)	564	if (ecb_expect_false (mg))
315	encode_ch (enc, 0xe0 | 31);	565	encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
316		566
317	--enc->depth;	567	--enc->depth;
318	}	568	}
319		569
320	// encode objects, arrays and special \0=false and \1=true values.	570	// encode objects, arrays and special \0=false and \1=true values.
321	static void	571	static void
322	encode_rv (enc_t *enc, SV *sv)	572	encode_rv (enc_t *enc, SV *sv)
323	{	573	{
324	SvGETMAGIC (sv);	574	SvGETMAGIC (sv);
325
326	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
327	&& ecb_expect_false (SvREFCNT (sv) > 1))
328	{
329	if (!enc->shareable)
330	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
331
332	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
333
334	if (SvOK (*svp))
335	{
336	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
337	encode_uint (enc, 0x00, SvUV (*svp));
338	return;
339	}
340	else
341	{
342	sv_setuv (*svp, enc->shareable_idx);
343	++enc->shareable_idx;
344	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
345	}
346	}
347		575
348	svtype svt = SvTYPE (sv);	576	svtype svt = SvTYPE (sv);
349		577
350	if (ecb_expect_false (SvOBJECT (sv)))	578	if (ecb_expect_false (SvOBJECT (sv)))
351	{	579	{
…		…
358	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	586	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
359	? cbor_tagged_stash	587	? cbor_tagged_stash
360	: gv_stashpv ("CBOR::XS::Tagged" , 1);	588	: gv_stashpv ("CBOR::XS::Tagged" , 1);
361		589
362	HV *stash = SvSTASH (sv);	590	HV *stash = SvSTASH (sv);
363	GV *method;
364		591
365	if (stash == boolean_stash)	592	if (stash == boolean_stash)
366	encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);	593	{
		594	encode_bool (enc, SvIV (sv));
		595	return;
		596	}
367	else if (stash == error_stash)	597	else if (stash == error_stash)
368	encode_ch (enc, 0xe0 | 23);	598	{
		599	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
		600	return;
		601	}
369	else if (stash == tagged_stash)	602	else if (stash == tagged_stash)
370	{	603	{
371	if (svt != SVt_PVAV)	604	if (svt != SVt_PVAV)
372	croak ("encountered CBOR::XS::Tagged object that isn't an array");	605	croak ("encountered CBOR::XS::Tagged object that isn't an array");
373		606
		607	switch (av_len ((AV *)sv))
		608	{
		609	case 2-1:
		610	// actually a tagged value
374	encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));	611	encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
375	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	}
		623
		624	return;
		625	}
		626	}
		627
		628	if (ecb_expect_false (SvREFCNT (sv) > 1)
		629	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		630	{
		631	if (ecb_expect_false (!enc->shareable))
		632	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
		633
		634	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
		635
		636	if (SvOK (*svp))
376	}	637	{
		638	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		639	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		640	return;
		641	}
		642	else
		643	{
		644	sv_setuv (*svp, enc->shareable_idx);
		645	++enc->shareable_idx;
		646	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		647	}
		648	}
		649
		650	if (ecb_expect_false (SvOBJECT (sv)))
		651	{
		652	HV *stash = SvSTASH (sv);
		653	GV *method;
		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))));
377	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	658	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
378	{	659	{
379	dSP;	660	dSP;
380		661
381	ENTER; SAVETMPS; PUSHMARK (SP);	662	ENTER; SAVETMPS;
		663	PUSHMARK (SP);
382	// 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
383	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	665	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
384		666
385	PUTBACK;	667	PUTBACK;
386	// G_SCALAR ensures that return value is 1	668	// G_SCALAR ensures that return value is 1
…		…
399	}	681	}
400	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	682	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
401	{	683	{
402	dSP;	684	dSP;
403		685
404	ENTER; SAVETMPS; PUSHMARK (SP);	686	ENTER; SAVETMPS;
		687	PUSHMARK (SP);
405	EXTEND (SP, 2);	688	EXTEND (SP, 2);
406	// 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
407	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	690	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
408	PUSHs (sv_cbor);	691	PUSHs (sv_cbor);
409		692
…		…
414	// catch this surprisingly common error	697	// catch this surprisingly common error
415	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	698	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
416	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));
417		700
418	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	701	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
419	encode_uint (enc, 0x80, count + 1);	702	encode_uint (enc, MAJOR_ARRAY, count + 1);
420	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	703	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
421		704
422	while (count)	705	{
		706	int i;
		707
		708	for (i = 0; i < count; ++i)
423	encode_sv (enc, SP[1 - count--]);	709	encode_sv (enc, SP[i + 1 - count]);
		710
		711	SP -= count;
		712	}
424		713
425	PUTBACK;	714	PUTBACK;
426		715
427	FREETMPS; LEAVE;	716	FREETMPS; LEAVE;
428	}	717	}
…		…
446	{	735	{
447	double nv = SvNVX (sv);	736	double nv = SvNVX (sv);
448		737
449	need (enc, 9);	738	need (enc, 9);
450		739
451	if (ecb_expect_false (nv == (U32)nv))	740	if (ecb_expect_false (nv == (NV)(U32)nv))
452	encode_uint (enc, 0x00, (U32)nv);	741	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
453	//TODO: maybe I32?	742	//TODO: maybe I32?
454	else if (ecb_expect_false (nv == (float)nv))	743	else if (ecb_expect_false (nv == (float)nv))
455	{	744	encode_float32 (enc, nv);
456	uint32_t fp = ecb_float_to_binary32 (nv);
457
458	*enc->cur++ = 0xe0 | 26;
459
460	if (!ecb_big_endian ())
461	fp = ecb_bswap32 (fp);
462
463	memcpy (enc->cur, &fp, 4);
464	enc->cur += 4;
465	}
466	else	745	else
467	{	746	encode_float64 (enc, nv);
468	uint64_t fp = ecb_double_to_binary64 (nv);
469
470	*enc->cur++ = 0xe0 | 27;
471
472	if (!ecb_big_endian ())
473	fp = ecb_bswap64 (fp);
474
475	memcpy (enc->cur, &fp, 8);
476	enc->cur += 8;
477	}
478	}	747	}
479		748
480	static void	749	static void
481	encode_sv (enc_t *enc, SV *sv)	750	encode_sv (enc_t *enc, SV *sv)
482	{	751	{
…		…
484		753
485	if (SvPOKp (sv))	754	if (SvPOKp (sv))
486	{	755	{
487	STRLEN len;	756	STRLEN len;
488	char *str = SvPV (sv, len);	757	char *str = SvPV (sv, len);
489	encode_str (enc, SvUTF8 (sv), str, len);	758	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
490	}	759	}
491	else if (SvNOKp (sv))	760	else if (SvNOKp (sv))
492	encode_nv (enc, sv);	761	encode_nv (enc, sv);
493	else if (SvIOKp (sv))	762	else if (SvIOKp (sv))
494	{	763	encode_int (enc, sv);
495	if (SvIsUV (sv))
496	encode_uint (enc, 0x00, SvUVX (sv));
497	else if (SvIVX (sv) >= 0)
498	encode_uint (enc, 0x00, SvIVX (sv));
499	else
500	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
501	}
502	else if (SvROK (sv))	764	else if (SvROK (sv))
503	encode_rv (enc, SvRV (sv));	765	encode_rv (enc, SvRV (sv));
504	else if (!SvOK (sv))	766	else if (!SvOK (sv))
505	encode_ch (enc, 0xe0 | 22);	767	encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
506	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	768	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
507	encode_ch (enc, 0xe0 | 23);	769	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
508	else	770	else
509	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",
510	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	772	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
511	}	773	}
512		774
513	static SV *	775	static SV *
514	encode_cbor (SV *scalar, CBOR *cbor)	776	encode_cbor (SV *scalar, CBOR *cbor)
515	{	777	{
516	enc_t enc = { };	778	enc_t enc = { 0 };
517		779
518	enc.cbor = *cbor;	780	enc.cbor = *cbor;
519	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	781	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
520	enc.cur = SvPVX (enc.sv);	782	enc.cur = SvPVX (enc.sv);
521	enc.end = SvEND (enc.sv);	783	enc.end = SvEND (enc.sv);
522		784
523	SvPOK_only (enc.sv);	785	SvPOK_only (enc.sv);
524		786
525	if (cbor->flags & F_ALLOW_STRINGREF)	787	if (cbor->flags & F_PACK_STRINGS)
526	{	788	{
527	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	789	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
528	enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());	790	enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
529	enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());	791	enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
530	}	792	}
…		…
553	U32 depth; // recursion depth	815	U32 depth; // recursion depth
554	U32 maxdepth; // recursion depth limit	816	U32 maxdepth; // recursion depth limit
555	AV *shareable;	817	AV *shareable;
556	AV *stringref;	818	AV *stringref;
557	SV *decode_tagged;	819	SV *decode_tagged;
		820	SV *err_sv; // optional sv for error, needs to be freed
558	} dec_t;	821	} dec_t;
559		822
560	#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);
561		830
562	#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;
563		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
564	#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)
565	#define DEC_DEC_DEPTH --dec->depth	860	#define DEC_DEC_DEPTH --dec->depth
566		861
567	static UV	862	static UV
568	decode_uint (dec_t *dec)	863	decode_uint (dec_t *dec)
569	{	864	{
570	switch (*dec->cur & 31)	865	U8 m = *dec->cur & MINOR_MASK;
571	{	866	++dec->cur;
572	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
573	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
574	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
575	return *dec->cur++ & 31;
576		867
577	case 24:	868	if (ecb_expect_true (m < LENGTH_EXT1))
		869	return m;
		870	else if (ecb_expect_true (m == LENGTH_EXT1))
		871	{
578	WANT (2);	872	WANT (1);
579	dec->cur += 2;	873	dec->cur += 1;
580	return dec->cur[-1];	874	return dec->cur[-1];
581		875	}
582	case 25:	876	else if (ecb_expect_true (m == LENGTH_EXT2))
		877	{
583	WANT (3);	878	WANT (2);
584	dec->cur += 3;	879	dec->cur += 2;
585	return (((UV)dec->cur[-2]) << 8)	880	return (((UV)dec->cur[-2]) << 8)
586	| ((UV)dec->cur[-1]);	881	| ((UV)dec->cur[-1]);
587		882	}
588	case 26:	883	else if (ecb_expect_true (m == LENGTH_EXT4))
		884	{
589	WANT (5);	885	WANT (4);
590	dec->cur += 5;	886	dec->cur += 4;
591	return (((UV)dec->cur[-4]) << 24)	887	return (((UV)dec->cur[-4]) << 24)
592	| (((UV)dec->cur[-3]) << 16)	888	| (((UV)dec->cur[-3]) << 16)
593	| (((UV)dec->cur[-2]) << 8)	889	| (((UV)dec->cur[-2]) << 8)
594	| ((UV)dec->cur[-1]);	890	| ((UV)dec->cur[-1]);
595		891	}
596	case 27:	892	else if (ecb_expect_true (m == LENGTH_EXT8))
		893	{
597	WANT (9);	894	WANT (8);
598	dec->cur += 9;	895	dec->cur += 8;
		896
		897	return
		898	#if UVSIZE < 8
		899	0
		900	#else
599	return (((UV)dec->cur[-8]) << 56)	901	(((UV)dec->cur[-8]) << 56)
600	| (((UV)dec->cur[-7]) << 48)	902	| (((UV)dec->cur[-7]) << 48)
601	| (((UV)dec->cur[-6]) << 40)	903	| (((UV)dec->cur[-6]) << 40)
602	| (((UV)dec->cur[-5]) << 32)	904	| (((UV)dec->cur[-5]) << 32)
		905	#endif
603	| (((UV)dec->cur[-4]) << 24)	906	| (((UV)dec->cur[-4]) << 24)
604	| (((UV)dec->cur[-3]) << 16)	907	| (((UV)dec->cur[-3]) << 16)
605	| (((UV)dec->cur[-2]) << 8)	908	| (((UV)dec->cur[-2]) << 8)
606	| ((UV)dec->cur[-1]);	909	| ((UV)dec->cur[-1]);
607		910	}
608	default:	911	else
609	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	912	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
610	}
611		913
612	fail:	914	fail:
613	return 0;	915	return 0;
614	}	916	}
615		917
…		…
620	{	922	{
621	AV *av = newAV ();	923	AV *av = newAV ();
622		924
623	DEC_INC_DEPTH;	925	DEC_INC_DEPTH;
624		926
625	if ((*dec->cur & 31) == 31)	927	if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
626	{	928	{
627	++dec->cur;	929	++dec->cur;
628		930
629	for (;;)	931	for (;;)
630	{	932	{
631	WANT (1);	933	WANT (1);
632		934
633	if (*dec->cur == (0xe0 | 31))	935	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
634	{	936	{
635	++dec->cur;	937	++dec->cur;
636	break;	938	break;
637	}	939	}
638		940
639	av_push (av, decode_sv (dec));	941	av_push (av, decode_sv (dec));
640	}	942	}
641	}	943	}
642	else	944	else
643	{	945	{
644	int i, len = decode_uint (dec);	946	UV i, len = decode_uint (dec);
645		947
		948	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
646	av_fill (av, len - 1);	949	av_fill (av, len - 1);
647		950
648	for (i = 0; i < len; ++i)	951	for (i = 0; i < len; ++i)
649	AvARRAY (av)[i] = decode_sv (dec);	952	AvARRAY (av)[i] = decode_sv (dec);
650	}	953	}
651		954
652	DEC_DEC_DEPTH;	955	DEC_DEC_DEPTH;
653	return newRV_noinc ((SV *)av);	956	return newRV_noinc ((SV *)av);
654		957
655	fail:	958	fail:
656	SvREFCNT_dec (av);	959	SvREFCNT_dec_NN (av);
657	DEC_DEC_DEPTH;	960	DEC_DEC_DEPTH;
658	return &PL_sv_undef;	961	return &PL_sv_undef;
659	}	962	}
660		963
661	static void	964	static void
…		…
663	{	966	{
664	// for speed reasons, we specialcase single-string	967	// for speed reasons, we specialcase single-string
665	// byte or utf-8 strings as keys, but only when !stringref	968	// byte or utf-8 strings as keys, but only when !stringref
666		969
667	if (ecb_expect_true (!dec->stringref))	970	if (ecb_expect_true (!dec->stringref))
668	if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)	971	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
669	{	972	{
670	I32 len = decode_uint (dec);	973	STRLEN len = decode_uint (dec);
671	char *key = (char *)dec->cur;	974	char *key = (char *)dec->cur;
672		975
		976	WANT (len);
673	dec->cur += len;	977	dec->cur += len;
674		978
675	if (ecb_expect_false (dec->stringref))
676	av_push (dec->stringref, newSVpvn (key, len));
677
678	hv_store (hv, key, len, decode_sv (dec), 0);	979	hv_store (hv, key, len, decode_sv (dec), 0);
679		980
680	return;	981	return;
681	}	982	}
682	else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)	983	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
683	{	984	{
684	I32 len = decode_uint (dec);	985	STRLEN len = decode_uint (dec);
685	char *key = (char *)dec->cur;	986	char *key = (char *)dec->cur;
686		987
		988	WANT (len);
687	dec->cur += len;	989	dec->cur += len;
688		990
689	if (ecb_expect_false (dec->stringref))	991	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
690	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	992	if (!is_utf8_string ((U8 *)key, len))
		993	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
691		994
692	hv_store (hv, key, -len, decode_sv (dec), 0);	995	hv_store (hv, key, -len, decode_sv (dec), 0);
693		996
694	return;	997	return;
695	}	998	}
696		999
697	SV *k = decode_sv (dec);	1000	SV *k = decode_sv (dec);
698	SV *v = decode_sv (dec);	1001	SV *v = decode_sv (dec);
699		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
700	hv_store_ent (hv, k, v, 0);	1031	hv_store_ent (hv, k, v, 0);
701	SvREFCNT_dec (k);	1032	SvREFCNT_dec_NN (k);
		1033
		1034	fail:
		1035	;
702	}	1036	}
703		1037
704	static SV *	1038	static SV *
705	decode_hv (dec_t *dec)	1039	decode_hv (dec_t *dec)
706	{	1040	{
707	HV *hv = newHV ();	1041	HV *hv = newHV ();
708		1042
709	DEC_INC_DEPTH;	1043	DEC_INC_DEPTH;
710		1044
711	if ((*dec->cur & 31) == 31)	1045	if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
712	{	1046	{
713	++dec->cur;	1047	++dec->cur;
714		1048
715	for (;;)	1049	for (;;)
716	{	1050	{
717	WANT (1);	1051	WANT (1);
718		1052
719	if (*dec->cur == (0xe0 | 31))	1053	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
720	{	1054	{
721	++dec->cur;	1055	++dec->cur;
722	break;	1056	break;
723	}	1057	}
724		1058
725	decode_he (dec, hv);	1059	decode_he (dec, hv);
726	}	1060	}
727	}	1061	}
728	else	1062	else
729	{	1063	{
730	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
731		1067
732	while (pairs--)	1068	while (pairs--)
733	decode_he (dec, hv);	1069	decode_he (dec, hv);
734	}	1070	}
735		1071
736	DEC_DEC_DEPTH;	1072	DEC_DEC_DEPTH;
737	return newRV_noinc ((SV *)hv);	1073	return newRV_noinc ((SV *)hv);
738		1074
739	fail:	1075	fail:
740	SvREFCNT_dec (hv);	1076	SvREFCNT_dec_NN (hv);
741	DEC_DEC_DEPTH;	1077	DEC_DEC_DEPTH;
742	return &PL_sv_undef;	1078	return &PL_sv_undef;
743	}	1079	}
744		1080
745	static SV *	1081	static SV *
746	decode_str (dec_t *dec, int utf8)	1082	decode_str (dec_t *dec, int utf8)
747	{	1083	{
748	SV *sv = 0;	1084	SV *sv = 0;
749		1085
750	if ((*dec->cur & 31) == 31)	1086	if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
751	{	1087	{
		1088	// indefinite length strings
752	++dec->cur;	1089	++dec->cur;
753		1090
		1091	U8 major = *dec->cur & MAJOR_MISC;
		1092
754	sv = newSVpvn ("", 0);	1093	sv = newSVpvn ("", 0);
755		1094
756	// not very fast, and certainly not robust against illegal input
757	for (;;)	1095	for (;;)
758	{	1096	{
759	WANT (1);	1097	WANT (1);
760		1098
761	if (*dec->cur == (0xe0 | 31))	1099	if ((*dec->cur - major) > LENGTH_EXT8)
		1100	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
762	{	1101	{
763	++dec->cur;	1102	++dec->cur;
764	break;	1103	break;
765	}	1104	}
		1105	else
		1106	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
766		1107
767	sv_catsv (sv, decode_sv (dec));	1108	STRLEN len = decode_uint (dec);
		1109
		1110	WANT (len);
		1111	sv_catpvn (sv, dec->cur, len);
		1112	dec->cur += len;
768	}	1113	}
769	}	1114	}
770	else	1115	else
771	{	1116	{
772	STRLEN len = decode_uint (dec);	1117	STRLEN len = decode_uint (dec);
…		…
779	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	1124	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
780	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	1125	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
781	}	1126	}
782		1127
783	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
784	SvUTF8_on (sv);	1134	SvUTF8_on (sv);
		1135	}
785		1136
786	return sv;	1137	return sv;
787		1138
788	fail:	1139	fail:
789	SvREFCNT_dec (sv);	1140	SvREFCNT_dec (sv);
…		…
808	sv = newRV_noinc (decode_sv (dec));	1159	sv = newRV_noinc (decode_sv (dec));
809	break;	1160	break;
810		1161
811	case CBOR_TAG_STRINGREF_NAMESPACE:	1162	case CBOR_TAG_STRINGREF_NAMESPACE:
812	{	1163	{
		1164	// do not use SAVETMPS/FREETMPS, as these will
		1165	// erase mortalised caches, e.g. "shareable"
813	ENTER; SAVETMPS;	1166	ENTER;
814		1167
815	SAVESPTR (dec->stringref);	1168	SAVESPTR (dec->stringref);
816	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());	1169	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
817		1170
818	sv = decode_sv (dec);	1171	sv = decode_sv (dec);
819		1172
820	FREETMPS; LEAVE;	1173	LEAVE;
821	}	1174	}
822	break;	1175	break;
823		1176
824	case CBOR_TAG_STRINGREF:	1177	case CBOR_TAG_STRINGREF:
825	{	1178	{
826	if ((*dec->cur >> 5) != 0)	1179	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
827	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1180	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
828		1181
829	UV idx = decode_uint (dec);	1182	UV idx = decode_uint (dec);
830		1183
831	if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))	1184	if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
832	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)");
833		1186
834	sv = newSVsv (AvARRAY (dec->stringref)[idx]);	1187	sv = newSVsv (AvARRAY (dec->stringref)[idx]);
835	}	1188	}
836	break;	1189	break;
…		…
838	case CBOR_TAG_VALUE_SHAREABLE:	1191	case CBOR_TAG_VALUE_SHAREABLE:
839	{	1192	{
840	if (ecb_expect_false (!dec->shareable))	1193	if (ecb_expect_false (!dec->shareable))
841	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());	1194	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
842		1195
		1196	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1197	{
843	sv = newSV (0);	1198	sv = newSV (0);
844	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1199	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
845		1200
846	SV *osv = decode_sv (dec);	1201	SV *osv = decode_sv (dec);
847	sv_setsv (sv, osv);	1202	sv_setsv (sv, osv);
848	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	}
849	}	1212	}
850	break;	1213	break;
851		1214
852	case CBOR_TAG_VALUE_SHAREDREF:	1215	case CBOR_TAG_VALUE_SHAREDREF:
853	{	1216	{
854	if ((*dec->cur >> 5) != 0)	1217	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
855	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1218	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
856		1219
857	UV idx = decode_uint (dec);	1220	UV idx = decode_uint (dec);
858		1221
859	if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))	1222	if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
860	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1223	ERR ("corrupted CBOR data (sharedref index out of bounds)");
861		1224
862	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");
863	}	1229	}
864	break;	1230	break;
865		1231
866	case CBOR_TAG_PERL_OBJECT:	1232	case CBOR_TAG_PERL_OBJECT:
867	{	1233	{
		1234	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1235	goto filter;
		1236
868	sv = decode_sv (dec);	1237	sv = decode_sv (dec);
869		1238
870	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)	1239	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
871	ERR ("corrupted CBOR data (non-array perl object)");	1240	ERR ("corrupted CBOR data (non-array perl object)");
872		1241
…		…
882	if (!method)	1251	if (!method)
883	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)");
884		1253
885	dSP;	1254	dSP;
886		1255
887	ENTER; SAVETMPS; PUSHMARK (SP);	1256	ENTER; SAVETMPS;
		1257	PUSHMARK (SP);
888	EXTEND (SP, len + 1);	1258	EXTEND (SP, len + 1);
889	// 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
890	PUSHs (*av_fetch (av, 0, 1));	1260	PUSHs (*av_fetch (av, 0, 1));
891	PUSHs (sv_cbor);	1261	PUSHs (sv_cbor);
892		1262
…		…
900	SPAGAIN;	1270	SPAGAIN;
901		1271
902	if (SvTRUE (ERRSV))	1272	if (SvTRUE (ERRSV))
903	{	1273	{
904	FREETMPS; LEAVE;	1274	FREETMPS; LEAVE;
905	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1275	ERR_ERRSV;
906	}	1276	}
907		1277
908	SvREFCNT_dec (sv);	1278	SvREFCNT_dec_NN (sv);
909	sv = SvREFCNT_inc (POPs);	1279	sv = SvREFCNT_inc (POPs);
910		1280
911	PUTBACK;	1281	PUTBACK;
912		1282
913	FREETMPS; LEAVE;	1283	FREETMPS; LEAVE;
914	}	1284	}
915	break;	1285	break;
916		1286
917	default:	1287	default:
		1288	filter:
918	{	1289	{
		1290	SV *tag_sv = newSVuv (tag);
		1291
919	sv = decode_sv (dec);	1292	sv = decode_sv (dec);
920		1293
921	dSP;	1294	dSP;
922	ENTER; SAVETMPS; PUSHMARK (SP);	1295	ENTER; SAVETMPS;
		1296	PUSHMARK (SP);
923	EXTEND (SP, 2);	1297	EXTEND (SP, 2);
924	PUSHs (newSVuv (tag));	1298	PUSHs (tag_sv);
925	PUSHs (sv);	1299	PUSHs (sv);
926		1300
927	PUTBACK;	1301	PUTBACK;
928	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);
929	SPAGAIN;	1303	SPAGAIN;
930		1304
931	if (SvTRUE (ERRSV))	1305	if (SvTRUE (ERRSV))
932	{	1306	{
		1307	SvREFCNT_dec_NN (tag_sv);
933	FREETMPS; LEAVE;	1308	FREETMPS; LEAVE;
934	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1309	ERR_ERRSV;
935	}	1310	}
936		1311
937	if (count)	1312	if (count)
938	{	1313	{
		1314	SvREFCNT_dec_NN (tag_sv);
939	SvREFCNT_dec (sv);	1315	SvREFCNT_dec_NN (sv);
940	sv = SvREFCNT_inc (POPs);	1316	sv = SvREFCNT_inc_NN (TOPs);
		1317	SP -= count;
941	}	1318	}
942	else	1319	else
943	{	1320	{
944	AV *av = newAV ();	1321	AV *av = newAV ();
945	av_push (av, newSVuv (tag));	1322	av_push (av, tag_sv);
946	av_push (av, sv);	1323	av_push (av, sv);
947		1324
948	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	1325	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
949	? cbor_tagged_stash	1326	? cbor_tagged_stash
950	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1327	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
968	static SV *	1345	static SV *
969	decode_sv (dec_t *dec)	1346	decode_sv (dec_t *dec)
970	{	1347	{
971	WANT (1);	1348	WANT (1);
972		1349
973	switch (*dec->cur >> 5)	1350	switch (*dec->cur >> MAJOR_SHIFT)
974	{	1351	{
975	case 0: // unsigned int	1352	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
976	return newSVuv (decode_uint (dec));	1353	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
977	case 1: // negative int	1354	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
978	return newSViv (-1 - (IV)decode_uint (dec));	1355	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
979	case 2: // octet string	1356	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
980	return decode_str (dec, 0);	1357	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
981	case 3: // utf-8 string	1358	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
982	return decode_str (dec, 1);	1359
983	case 4: // array	1360	case MAJOR_MISC >> MAJOR_SHIFT:
984	return decode_av (dec);
985	case 5: // map
986	return decode_hv (dec);
987	case 6: // tag
988	return decode_tagged (dec);
989	case 7: // misc
990	switch (*dec->cur++ & 31)	1361	switch (*dec->cur++ & MINOR_MASK)
991	{	1362	{
992	case 20:	1363	case SIMPLE_FALSE:
993	#if CBOR_SLOW	1364	#if CBOR_SLOW
994	types_false = get_bool ("Types::Serialiser::false");	1365	types_false = get_bool ("Types::Serialiser::false");
995	#endif	1366	#endif
996	return newSVsv (types_false);	1367	return newSVsv (types_false);
997	case 21:	1368	case SIMPLE_TRUE:
998	#if CBOR_SLOW	1369	#if CBOR_SLOW
999	types_true = get_bool ("Types::Serialiser::true");	1370	types_true = get_bool ("Types::Serialiser::true");
1000	#endif	1371	#endif
1001	return newSVsv (types_true);	1372	return newSVsv (types_true);
1002	case 22:	1373	case SIMPLE_NULL:
1003	return newSVsv (&PL_sv_undef);	1374	return newSVsv (&PL_sv_undef);
1004	case 23:	1375	case SIMPLE_UNDEF:
1005	#if CBOR_SLOW	1376	#if CBOR_SLOW
1006	types_error = get_bool ("Types::Serialiser::error");	1377	types_error = get_bool ("Types::Serialiser::error");
1007	#endif	1378	#endif
1008	return newSVsv (types_error);	1379	return newSVsv (types_error);
1009		1380
1010	case 25:	1381	case MISC_FLOAT16:
1011	{	1382	{
1012	WANT (2);	1383	WANT (2);
1013		1384
1014	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];	1385	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1015	dec->cur += 2;	1386	dec->cur += 2;
1016		1387
1017	return newSVnv (ecb_binary16_to_float (fp));	1388	return newSVnv (ecb_binary16_to_float (fp));
1018	}	1389	}
1019		1390
1020	case 26:	1391	case MISC_FLOAT32:
1021	{	1392	{
1022	uint32_t fp;	1393	uint32_t fp;
1023	WANT (4);	1394	WANT (4);
1024	memcpy (&fp, dec->cur, 4);	1395	memcpy (&fp, dec->cur, 4);
1025	dec->cur += 4;	1396	dec->cur += 4;
…		…
1028	fp = ecb_bswap32 (fp);	1399	fp = ecb_bswap32 (fp);
1029		1400
1030	return newSVnv (ecb_binary32_to_float (fp));	1401	return newSVnv (ecb_binary32_to_float (fp));
1031	}	1402	}
1032		1403
1033	case 27:	1404	case MISC_FLOAT64:
1034	{	1405	{
1035	uint64_t fp;	1406	uint64_t fp;
1036	WANT (8);	1407	WANT (8);
1037	memcpy (&fp, dec->cur, 8);	1408	memcpy (&fp, dec->cur, 8);
1038	dec->cur += 8;	1409	dec->cur += 8;
…		…
1041	fp = ecb_bswap64 (fp);	1412	fp = ecb_bswap64 (fp);
1042		1413
1043	return newSVnv (ecb_binary64_to_double (fp));	1414	return newSVnv (ecb_binary64_to_double (fp));
1044	}	1415	}
1045		1416
1046	// 0..19 unassigned	1417	// 0..19 unassigned simple
1047	// 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
1048	default:	1421	default:
1049	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1422	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1050	}	1423	}
1051		1424
1052	break;	1425	break;
1053	}	1426	}
1054		1427
…		…
1057	}	1430	}
1058		1431
1059	static SV *	1432	static SV *
1060	decode_cbor (SV *string, CBOR *cbor, char **offset_return)	1433	decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1061	{	1434	{
1062	dec_t dec = { };	1435	dec_t dec = { 0 };
1063	SV *sv;	1436	SV *sv;
1064	STRLEN len;	1437	STRLEN len;
1065	char *data = SvPVbyte (string, len);	1438	char *data = SvPVbyte (string, len);
1066		1439
1067	if (len > cbor->max_size && cbor->max_size)	1440	if (len > cbor->max_size && cbor->max_size)
…		…
1081	if (dec.cur != dec.end && !dec.err)	1454	if (dec.cur != dec.end && !dec.err)
1082	dec.err = "garbage after CBOR object";	1455	dec.err = "garbage after CBOR object";
1083		1456
1084	if (dec.err)	1457	if (dec.err)
1085	{	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
1086	SvREFCNT_dec (sv);	1470	SvREFCNT_dec_NN (sv);
		1471
		1472	if (dec.err_sv)
		1473	sv_2mortal (dec.err_sv);
		1474
1087	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);
1088	}	1476	}
1089		1477
1090	sv = sv_2mortal (sv);	1478	sv = sv_2mortal (sv);
1091		1479
1092	return sv;	1480	return sv;
1093	}	1481	}
1094		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	/* FALLTHROUGH */
		1575	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1576	if (len)
		1577	{
		1578	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1579	count = len + 1;
		1580	}
		1581	break;
		1582	}
		1583	}
		1584	}
		1585	}
		1586
		1587	self->incr_pos = self->incr_need;
		1588
		1589	if (count > 0)
		1590	{
		1591	while (!--count)
		1592	{
		1593	if (!AvFILLp (self->incr_count))
		1594	return 1; // done
		1595
		1596	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1597	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1598	}
		1599
		1600	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1601	}
		1602	}
		1603
		1604	return 0;
		1605	}
		1606
		1607
1095	/////////////////////////////////////////////////////////////////////////////	1608	/////////////////////////////////////////////////////////////////////////////
1096	// XS interface functions	1609	// XS interface functions
1097		1610
1098	MODULE = CBOR::XS PACKAGE = CBOR::XS	1611	MODULE = CBOR::XS PACKAGE = CBOR::XS
1099		1612
…		…
1111		1624
1112	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1625	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1113		1626
1114	sv_cbor = newSVpv ("CBOR", 0);	1627	sv_cbor = newSVpv ("CBOR", 0);
1115	SvREADONLY_on (sv_cbor);	1628	SvREADONLY_on (sv_cbor);
		1629
		1630	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1116	}	1631	}
1117		1632
1118	PROTOTYPES: DISABLE	1633	PROTOTYPES: DISABLE
1119		1634
1120	void CLONE (...)	1635	void CLONE (...)
…		…
1139	void shrink (CBOR *self, int enable = 1)	1654	void shrink (CBOR *self, int enable = 1)
1140	ALIAS:	1655	ALIAS:
1141	shrink = F_SHRINK	1656	shrink = F_SHRINK
1142	allow_unknown = F_ALLOW_UNKNOWN	1657	allow_unknown = F_ALLOW_UNKNOWN
1143	allow_sharing = F_ALLOW_SHARING	1658	allow_sharing = F_ALLOW_SHARING
1144	allow_stringref = F_ALLOW_STRINGREF	1659	allow_cycles = F_ALLOW_CYCLES
		1660	forbid_objects = F_FORBID_OBJECTS
		1661	pack_strings = F_PACK_STRINGS
		1662	text_keys = F_TEXT_KEYS
		1663	text_strings = F_TEXT_STRINGS
		1664	validate_utf8 = F_VALIDATE_UTF8
1145	PPCODE:	1665	PPCODE:
1146	{	1666	{
1147	if (enable)	1667	if (enable)
1148	self->flags |= ix;	1668	self->flags |= ix;
1149	else	1669	else
…		…
1155	void get_shrink (CBOR *self)	1675	void get_shrink (CBOR *self)
1156	ALIAS:	1676	ALIAS:
1157	get_shrink = F_SHRINK	1677	get_shrink = F_SHRINK
1158	get_allow_unknown = F_ALLOW_UNKNOWN	1678	get_allow_unknown = F_ALLOW_UNKNOWN
1159	get_allow_sharing = F_ALLOW_SHARING	1679	get_allow_sharing = F_ALLOW_SHARING
1160	get_allow_stringref = F_ALLOW_STRINGREF	1680	get_allow_cycles = F_ALLOW_CYCLES
		1681	get_forbid_objects = F_FORBID_OBJECTS
		1682	get_pack_strings = F_PACK_STRINGS
		1683	get_text_keys = F_TEXT_KEYS
		1684	get_text_strings = F_TEXT_STRINGS
		1685	get_validate_utf8 = F_VALIDATE_UTF8
1161	PPCODE:	1686	PPCODE:
1162	XPUSHs (boolSV (self->flags & ix));	1687	XPUSHs (boolSV (self->flags & ix));
1163		1688
1164	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1689	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1165	PPCODE:	1690	PPCODE:
…		…
1214	EXTEND (SP, 2);	1739	EXTEND (SP, 2);
1215	PUSHs (sv);	1740	PUSHs (sv);
1216	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1741	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1217	}	1742	}
1218		1743
		1744	void incr_parse (CBOR *self, SV *cborstr)
		1745	ALIAS:
		1746	incr_parse_multiple = 1
		1747	PPCODE:
		1748	{
		1749	if (SvUTF8 (cborstr))
		1750	sv_utf8_downgrade (cborstr, 0);
		1751
		1752	if (!self->incr_count)
		1753	{
		1754	self->incr_count = newAV ();
		1755	self->incr_pos = 0;
		1756	self->incr_need = 1;
		1757
		1758	av_push (self->incr_count, newSViv (1));
		1759	}
		1760
		1761	do
		1762	{
		1763	if (!incr_parse (self, cborstr))
		1764	{
		1765	if (self->incr_need > self->max_size && self->max_size)
		1766	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1767	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1768
		1769	break;
		1770	}
		1771
		1772	SV *sv;
		1773	char *offset;
		1774
		1775	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1776	XPUSHs (sv);
		1777
		1778	sv_chop (cborstr, offset);
		1779
		1780	av_clear (self->incr_count);
		1781	av_push (self->incr_count, newSViv (1));
		1782
		1783	self->incr_pos = 0;
		1784	self->incr_need = self->incr_pos + 1;
		1785	}
		1786	while (ix);
		1787	}
		1788
		1789	void incr_reset (CBOR *self)
		1790	CODE:
		1791	{
		1792	SvREFCNT_dec (self->incr_count);
		1793	self->incr_count = 0;
		1794	}
		1795
1219	void DESTROY (CBOR *self)	1796	void DESTROY (CBOR *self)
1220	PPCODE:	1797	PPCODE:
1221	cbor_free (self);	1798	cbor_free (self);
1222		1799
1223	PROTOTYPES: ENABLE	1800	PROTOTYPES: ENABLE
1224		1801
1225	void encode_cbor (SV *scalar)	1802	void encode_cbor (SV *scalar)
		1803	ALIAS:
		1804	encode_cbor = 0
		1805	encode_cbor_sharing = F_ALLOW_SHARING
1226	PPCODE:	1806	PPCODE:
1227	{	1807	{
1228	CBOR cbor;	1808	CBOR cbor;
1229	cbor_init (&cbor);	1809	cbor_init (&cbor);
		1810	cbor.flags |= ix;
1230	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1811	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1231	XPUSHs (scalar);	1812	XPUSHs (scalar);
1232	}	1813	}
1233		1814
1234	void decode_cbor (SV *cborstr)	1815	void decode_cbor (SV *cborstr)
…		…
1238	cbor_init (&cbor);	1819	cbor_init (&cbor);
1239	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;	1820	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1240	XPUSHs (cborstr);	1821	XPUSHs (cborstr);
1241	}	1822	}
1242		1823
		1824	#ifdef __AFL_COMPILER
		1825
		1826	void
		1827	afl_init ()
		1828	CODE:
		1829	__AFL_INIT ();
		1830
		1831	int
		1832	afl_loop (unsigned int count = 10000)
		1833	CODE:
		1834	RETVAL = __AFL_LOOP (count);
		1835	OUTPUT:
		1836	RETVAL
		1837
		1838	#endif
		1839

Diff Legend

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