ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.26 by root, Fri Nov 22 09:40:13 2013 UTC vs.
Revision 1.69 by root, Mon Nov 30 20:38:25 2020 UTC

…		…
6	#include <string.h>	6	#include <string.h>
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
		11	#include <inttypes.h>
11		12
		13	#define ECB_NO_THREADS 1
12	#include "ecb.h"	14	#include "ecb.h"
13		15
14	// compatibility with perl <5.18	16	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	17	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	18	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	21	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	22	#endif
21	#ifndef HvNAMEUTF8	23	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	24	# define HvNAMEUTF8(hv) 0
23	#endif	25	#endif
		26	#ifndef SvREFCNT_inc_NN
		27	# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
		28	#endif
		29	#ifndef SvREFCNT_dec_NN
		30	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		31	#endif
		32
		33	// known major and minor types
		34	enum cbor_type
		35	{
		36	MAJOR_SHIFT = 5,
		37	MINOR_MASK = 0x1f,
		38
		39	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		40	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		41	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		42	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		43	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		44	MAJOR_MAP = 5 << MAJOR_SHIFT,
		45	MAJOR_TAG = 6 << MAJOR_SHIFT,
		46	MAJOR_MISC = 7 << MAJOR_SHIFT,
		47
		48	// INT/STRING/ARRAY/MAP subtypes
		49	LENGTH_EXT1 = 24,
		50	LENGTH_EXT2 = 25,
		51	LENGTH_EXT4 = 26,
		52	LENGTH_EXT8 = 27,
		53
		54	// SIMPLE types (effectively MISC subtypes)
		55	SIMPLE_FALSE = 20,
		56	SIMPLE_TRUE = 21,
		57	SIMPLE_NULL = 22,
		58	SIMPLE_UNDEF = 23,
		59
		60	// MISC subtype (unused)
		61	MISC_EXT1 = 24,
		62	MISC_FLOAT16 = 25,
		63	MISC_FLOAT32 = 26,
		64	MISC_FLOAT64 = 27,
		65
		66	// BYTES/TEXT/ARRAY/MAP
		67	MINOR_INDEF = 31,
		68	};
24		69
25	// known tags	70	// known tags
26	enum cbor_tag	71	enum cbor_tag
27	{	72	{
28	// inofficial extensions (pending iana registration)	73	// extensions
		74	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	75	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	76	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	77	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	78	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	79	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	80	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		81
37	// rfc7049	82	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	83	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	84	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	85	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	86	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	87	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	88	CBOR_TAG_BIGFLOAT = 5, // array
44		89
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	90	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	91	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	92	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	93	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		94
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	95	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	96	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	97	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	98	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	99	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		100
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	101	CBOR_TAG_MAGIC = 55799, // self-describe cbor
		102	};
		103
		104	// known forced types, also hardcoded in CBOR.pm
		105	enum
		106	{
		107	AS_CBOR = 0,
		108	AS_INT = 1,
		109	AS_BYTES = 2,
		110	AS_TEXT = 3,
		111	AS_FLOAT16 = 4,
		112	AS_FLOAT32 = 5,
		113	AS_FLOAT64 = 6,
		114	// possibly future enhancements: (generic) float, (generic) string
57	};	115	};
58		116
59	#define F_SHRINK 0x00000001UL	117	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	118	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	119	#define F_ALLOW_SHARING 0x00000004UL
62	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	120	#define F_ALLOW_CYCLES 0x00000008UL
		121	#define F_FORBID_OBJECTS 0x00000010UL
		122	#define F_PACK_STRINGS 0x00000020UL
		123	#define F_TEXT_KEYS 0x00000040UL
		124	#define F_TEXT_STRINGS 0x00000080UL
		125	#define F_VALIDATE_UTF8 0x00000100UL
63		126
64	#define INIT_SIZE 32 // initial scalar size to be allocated	127	#define INIT_SIZE 32 // initial scalar size to be allocated
65		128
66	#define SB do {	129	#define SB do {
67	#define SE } while (0)	130	#define SE } while (0)
…		…
79	# define CBOR_SLOW 0	142	# define CBOR_SLOW 0
80	# define CBOR_STASH cbor_stash	143	# define CBOR_STASH cbor_stash
81	#endif	144	#endif
82		145
83	static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::	146	static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
84	static SV *types_true, *types_false, *types_error, *sv_cbor;	147	static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
85		148
86	typedef struct {	149	typedef struct {
87	U32 flags;	150	U32 flags;
88	U32 max_depth;	151	U32 max_depth;
89	STRLEN max_size;	152	STRLEN max_size;
		153	SV *filter;
		154
		155	// for the incremental parser
		156	STRLEN incr_pos; // the current offset into the text
		157	STRLEN incr_need; // minimum bytes needed to decode
		158	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90	} CBOR;	159	} CBOR;
91		160
92	ecb_inline void	161	ecb_inline void
93	cbor_init (CBOR *cbor)	162	cbor_init (CBOR *cbor)
94	{	163	{
95	Zero (cbor, 1, CBOR);	164	Zero (cbor, 1, CBOR);
96	cbor->max_depth = 512;	165	cbor->max_depth = 512;
		166	}
		167
		168	ecb_inline void
		169	cbor_free (CBOR *cbor)
		170	{
		171	SvREFCNT_dec (cbor->filter);
		172	SvREFCNT_dec (cbor->incr_count);
97	}	173	}
98		174
99	/////////////////////////////////////////////////////////////////////////////	175	/////////////////////////////////////////////////////////////////////////////
100	// utility functions	176	// utility functions
101		177
…		…
127		203
128	// minimum length of a string to be registered for stringref	204	// minimum length of a string to be registered for stringref
129	ecb_inline int	205	ecb_inline int
130	minimum_string_length (UV idx)	206	minimum_string_length (UV idx)
131	{	207	{
132	return idx > 23	208	return idx <= 23 ? 3
133	? idx > 0xffU	209	: idx <= 0xffU ? 4
134	? idx > 0xffffU	210	: idx <= 0xffffU ? 5
135	? idx > 0xffffffffU	211	: idx <= 0xffffffffU ? 7
136	? 7	212	: 11;
137	: 6
138	: 5
139	: 4
140	: 3;
141	}	213	}
142		214
143	/////////////////////////////////////////////////////////////////////////////	215	/////////////////////////////////////////////////////////////////////////////
144	// encoder	216	// encoder
145		217
…		…
158	} enc_t;	230	} enc_t;
159		231
160	ecb_inline void	232	ecb_inline void
161	need (enc_t *enc, STRLEN len)	233	need (enc_t *enc, STRLEN len)
162	{	234	{
163	if (ecb_expect_false (enc->cur + len >= enc->end))	235	if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
164	{	236	{
165	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);	237	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
166	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);	238	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
167	enc->cur = SvPVX (enc->sv) + cur;	239	enc->cur = SvPVX (enc->sv) + cur;
168	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;	240	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
…		…
174	{	246	{
175	need (enc, 1);	247	need (enc, 1);
176	*enc->cur++ = ch;	248	*enc->cur++ = ch;
177	}	249	}
178		250
		251	// used for tags, intregers, element counts and so on
179	static void	252	static void
180	encode_uint (enc_t *enc, int major, UV len)	253	encode_uint (enc_t *enc, int major, UV len)
181	{	254	{
182	need (enc, 9);	255	need (enc, 9);
183		256
184	if (len < 24)	257	if (ecb_expect_true (len < LENGTH_EXT1))
185	*enc->cur++ = major | len;	258	*enc->cur++ = major | len;
186	else if (len <= 0xff)	259	else if (ecb_expect_true (len <= 0xffU))
187	{	260	{
188	*enc->cur++ = major | 24;	261	*enc->cur++ = major | LENGTH_EXT1;
189	*enc->cur++ = len;	262	*enc->cur++ = len;
190	}	263	}
191	else if (len <= 0xffff)	264	else if (len <= 0xffffU)
192	{	265	{
193	*enc->cur++ = major | 25;	266	*enc->cur++ = major | LENGTH_EXT2;
194	*enc->cur++ = len >> 8;	267	*enc->cur++ = len >> 8;
195	*enc->cur++ = len;	268	*enc->cur++ = len;
196	}	269	}
197	else if (len <= 0xffffffff)	270	else if (len <= 0xffffffffU)
198	{	271	{
199	*enc->cur++ = major | 26;	272	*enc->cur++ = major | LENGTH_EXT4;
200	*enc->cur++ = len >> 24;	273	*enc->cur++ = len >> 24;
201	*enc->cur++ = len >> 16;	274	*enc->cur++ = len >> 16;
202	*enc->cur++ = len >> 8;	275	*enc->cur++ = len >> 8;
203	*enc->cur++ = len;	276	*enc->cur++ = len;
204	}	277	}
205	else	278	else
206	{	279	{
207	*enc->cur++ = major | 27;	280	*enc->cur++ = major | LENGTH_EXT8;
208	*enc->cur++ = len >> 56;	281	*enc->cur++ = len >> 56;
209	*enc->cur++ = len >> 48;	282	*enc->cur++ = len >> 48;
210	*enc->cur++ = len >> 40;	283	*enc->cur++ = len >> 40;
211	*enc->cur++ = len >> 32;	284	*enc->cur++ = len >> 32;
212	*enc->cur++ = len >> 24;	285	*enc->cur++ = len >> 24;
…		…
214	*enc->cur++ = len >> 8;	287	*enc->cur++ = len >> 8;
215	*enc->cur++ = len;	288	*enc->cur++ = len;
216	}	289	}
217	}	290	}
218		291
		292	// encodes a perl value into a CBOR integer
		293	ecb_inline void
		294	encode_int (enc_t *enc, SV *sv)
		295	{
		296	if (SvIsUV (sv))
		297	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
		298	else if (SvIVX (sv) >= 0)
		299	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
		300	else
		301	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
		302	}
		303
219	ecb_inline void	304	ecb_inline void
220	encode_tag (enc_t *enc, UV tag)	305	encode_tag (enc_t *enc, UV tag)
221	{	306	{
222	encode_uint (enc, 0xc0, tag);	307	encode_uint (enc, MAJOR_TAG, tag);
223	}	308	}
224		309
225	static void	310	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		311	ecb_noinline static void
226	encode_str (enc_t *enc, int utf8, char *str, STRLEN len)	312	encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
227	{	313	{
		314	STRLEN ulen = len;
		315	U8 *p, *pend = (U8 *)str + len;
		316
		317	for (p = (U8 *)str; p < pend; ++p)
		318	ulen += *p >> 7; // count set high bits
		319
		320	encode_uint (enc, MAJOR_TEXT, ulen);
		321
		322	need (enc, ulen);
		323	for (p = (U8 *)str; p < pend; ++p)
		324	if (*p < 0x80)
		325	*enc->cur++ = *p;
		326	else
		327	{
		328	*enc->cur++ = 0xc0 + (*p >> 6);
		329	*enc->cur++ = 0x80 + (*p & 63);
		330	}
		331	}
		332
		333	ecb_inline void
		334	encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
		335	{
		336	if (ecb_expect_false (upgrade_utf8))
		337	if (!utf8)
		338	{
		339	encode_str_utf8 (enc, utf8, str, len);
		340	return;
		341	}
		342
		343	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		344	need (enc, len);
		345	memcpy (enc->cur, str, len);
		346	enc->cur += len;
		347	}
		348
		349	ecb_inline void
		350	encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
		351	{
228	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	352	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229	{	353	{
230	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	354	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231		355
232	if (SvOK (*svp))	356	if (SvOK (*svp))
233	{	357	{
234	// already registered, use stringref	358	// already registered, use stringref
235	encode_tag (enc, CBOR_TAG_STRINGREF);	359	encode_tag (enc, CBOR_TAG_STRINGREF);
236	encode_uint (enc, 0x00, SvUV (*svp));	360	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237	return;	361	return;
238	}	362	}
239	else if (len >= minimum_string_length (enc->stringref_idx))	363	else if (len >= minimum_string_length (enc->stringref_idx))
240	{	364	{
241	// register only	365	// register only
242	sv_setuv (*svp, enc->stringref_idx);	366	sv_setuv (*svp, enc->stringref_idx);
243	++enc->stringref_idx;	367	++enc->stringref_idx;
244	}	368	}
245	}	369	}
246		370
247	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	371	encode_str (enc, upgrade_utf8, utf8, str, len);
		372	}
		373
		374	ecb_inline void
		375	encode_float16 (enc_t *enc, NV nv)
		376	{
248	need (enc, len);	377	need (enc, 1+2);
		378
		379	*enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
		380
		381	uint16_t fp = ecb_float_to_binary16 (nv);
		382
		383	if (!ecb_big_endian ())
		384	fp = ecb_bswap16 (fp);
		385
249	memcpy (enc->cur, str, len);	386	memcpy (enc->cur, &fp, 2);
250	enc->cur += len;	387	enc->cur += 2;
		388	}
		389
		390	ecb_inline void
		391	encode_float32 (enc_t *enc, NV nv)
		392	{
		393	need (enc, 1+4);
		394
		395	*enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
		396
		397	uint32_t fp = ecb_float_to_binary32 (nv);
		398
		399	if (!ecb_big_endian ())
		400	fp = ecb_bswap32 (fp);
		401
		402	memcpy (enc->cur, &fp, 4);
		403	enc->cur += 4;
		404	}
		405
		406	ecb_inline void
		407	encode_float64 (enc_t *enc, NV nv)
		408	{
		409	need (enc, 1+8);
		410
		411	*enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
		412
		413	uint64_t fp = ecb_double_to_binary64 (nv);
		414
		415	if (!ecb_big_endian ())
		416	fp = ecb_bswap64 (fp);
		417
		418	memcpy (enc->cur, &fp, 8);
		419	enc->cur += 8;
		420	}
		421
		422	ecb_inline void
		423	encode_bool (enc_t *enc, int istrue)
		424	{
		425	encode_ch (enc, istrue ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
		426	}
		427
		428	ecb_inline void
		429	encode_forced (enc_t *enc, UV type, SV *sv)
		430	{
		431	switch (type)
		432	{
		433	case AS_CBOR:
		434	{
		435	STRLEN len;
		436	char *str = SvPVbyte (sv, len);
		437
		438	need (enc, len);
		439	memcpy (enc->cur, str, len);
		440	enc->cur += len;
		441	}
		442	break;
		443
		444	case AS_BYTES:
		445	{
		446	STRLEN len;
		447	char *str = SvPVbyte (sv, len);
		448	encode_strref (enc, 0, 0, str, len);
		449	}
		450	break;
		451
		452	case AS_TEXT:
		453	{
		454	STRLEN len;
		455	char *str = SvPVutf8 (sv, len);
		456	encode_strref (enc, 1, 1, str, len);
		457	}
		458	break;
		459
		460	case AS_INT: encode_int (enc, sv); break;
		461
		462	case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
		463	case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
		464	case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
		465
		466	default:
		467	croak ("encountered malformed CBOR::XS::Tagged object");
		468	}
251	}	469	}
252		470
253	static void encode_sv (enc_t *enc, SV *sv);	471	static void encode_sv (enc_t *enc, SV *sv);
254		472
255	static void	473	static void
…		…
260	if (enc->depth >= enc->cbor.max_depth)	478	if (enc->depth >= enc->cbor.max_depth)
261	croak (ERR_NESTING_EXCEEDED);	479	croak (ERR_NESTING_EXCEEDED);
262		480
263	++enc->depth;	481	++enc->depth;
264		482
265	encode_uint (enc, 0x80, len + 1);	483	encode_uint (enc, MAJOR_ARRAY, len + 1);
266		484
		485	if (ecb_expect_false (SvMAGICAL (av)))
267	for (i = 0; i <= len; ++i)	486	for (i = 0; i <= len; ++i)
268	{	487	{
269	SV **svp = av_fetch (av, i, 0);	488	SV **svp = av_fetch (av, i, 0);
270	encode_sv (enc, svp ? *svp : &PL_sv_undef);	489	encode_sv (enc, svp ? *svp : &PL_sv_undef);
271	}	490	}
		491	else
		492	for (i = 0; i <= len; ++i)
		493	{
		494	SV *sv = AvARRAY (av)[i];
		495	encode_sv (enc, sv ? sv : &PL_sv_undef);
		496	}
272		497
273	--enc->depth;	498	--enc->depth;
274	}	499	}
275		500
276	static void	501	static void
…		…
284	++enc->depth;	509	++enc->depth;
285		510
286	int pairs = hv_iterinit (hv);	511	int pairs = hv_iterinit (hv);
287	int mg = SvMAGICAL (hv);	512	int mg = SvMAGICAL (hv);
288		513
289	if (mg)	514	if (ecb_expect_false (mg))
290	encode_ch (enc, 0xa0 | 31);	515	encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291	else	516	else
292	encode_uint (enc, 0xa0, pairs);	517	encode_uint (enc, MAJOR_MAP, pairs);
293		518
294	while ((he = hv_iternext (hv)))	519	while ((he = hv_iternext (hv)))
295	{	520	{
296	if (HeKLEN (he) == HEf_SVKEY)	521	if (HeKLEN (he) == HEf_SVKEY)
297	encode_sv (enc, HeSVKEY (he));	522	encode_sv (enc, HeSVKEY (he));
298	else	523	else
299	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	524	encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300		525
301	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	526	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302	}	527	}
303		528
304	if (mg)	529	if (ecb_expect_false (mg))
305	encode_ch (enc, 0xe0 | 31);	530	encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306		531
307	--enc->depth;	532	--enc->depth;
308	}	533	}
309		534
310	// encode objects, arrays and special \0=false and \1=true values.	535	// encode objects, arrays and special \0=false and \1=true values.
311	static void	536	static void
312	encode_rv (enc_t *enc, SV *sv)	537	encode_rv (enc_t *enc, SV *sv)
313	{	538	{
314	SvGETMAGIC (sv);	539	SvGETMAGIC (sv);
315
316	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
317	&& ecb_expect_false (SvREFCNT (sv) > 1))
318	{
319	if (!enc->shareable)
320	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
321
322	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
323
324	if (SvOK (*svp))
325	{
326	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
327	encode_uint (enc, 0x00, SvUV (*svp));
328	return;
329	}
330	else
331	{
332	sv_setuv (*svp, enc->shareable_idx);
333	++enc->shareable_idx;
334	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
335	}
336	}
337		540
338	svtype svt = SvTYPE (sv);	541	svtype svt = SvTYPE (sv);
339		542
340	if (ecb_expect_false (SvOBJECT (sv)))	543	if (ecb_expect_false (SvOBJECT (sv)))
341	{	544	{
…		…
348	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	551	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
349	? cbor_tagged_stash	552	? cbor_tagged_stash
350	: gv_stashpv ("CBOR::XS::Tagged" , 1);	553	: gv_stashpv ("CBOR::XS::Tagged" , 1);
351		554
352	HV *stash = SvSTASH (sv);	555	HV *stash = SvSTASH (sv);
353	GV *method;
354		556
355	if (stash == boolean_stash)	557	if (stash == boolean_stash)
356	encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);	558	{
		559	encode_bool (enc, SvIV (sv));
		560	return;
		561	}
357	else if (stash == error_stash)	562	else if (stash == error_stash)
358	encode_ch (enc, 0xe0 | 23);	563	{
		564	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
		565	return;
		566	}
359	else if (stash == tagged_stash)	567	else if (stash == tagged_stash)
360	{	568	{
361	if (svt != SVt_PVAV)	569	if (svt != SVt_PVAV)
362	croak ("encountered CBOR::XS::Tagged object that isn't an array");	570	croak ("encountered CBOR::XS::Tagged object that isn't an array");
363		571
		572	switch (av_len ((AV *)sv))
		573	{
		574	case 2-1:
		575	// actually a tagged value
364	encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));	576	encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
365	encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));	577	encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
		578	break;
		579
		580	case 3-1:
		581	// a forced type [value, type, undef]
		582	encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
		583	break;
		584
		585	default:
		586	croak ("encountered malformed CBOR::XS::Tagged object");
		587	}
		588
		589	return;
		590	}
		591	}
		592
		593	if (ecb_expect_false (SvREFCNT (sv) > 1)
		594	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		595	{
		596	if (ecb_expect_false (!enc->shareable))
		597	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
		598
		599	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
		600
		601	if (SvOK (*svp))
366	}	602	{
		603	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		604	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		605	return;
		606	}
		607	else
		608	{
		609	sv_setuv (*svp, enc->shareable_idx);
		610	++enc->shareable_idx;
		611	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		612	}
		613	}
		614
		615	if (ecb_expect_false (SvOBJECT (sv)))
		616	{
		617	HV *stash = SvSTASH (sv);
		618	GV *method;
		619
		620	if (enc->cbor.flags & F_FORBID_OBJECTS)
		621	croak ("encountered object '%s', but forbid_objects is enabled",
		622	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
367	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	623	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
368	{	624	{
369	dSP;	625	dSP;
370		626
371	ENTER; SAVETMPS; PUSHMARK (SP);	627	ENTER; SAVETMPS;
		628	PUSHMARK (SP);
372	// we re-bless the reference to get overload and other niceties right	629	// we re-bless the reference to get overload and other niceties right
373	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	630	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
374		631
375	PUTBACK;	632	PUTBACK;
376	// G_SCALAR ensures that return value is 1	633	// G_SCALAR ensures that return value is 1
…		…
389	}	646	}
390	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	647	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
391	{	648	{
392	dSP;	649	dSP;
393		650
394	ENTER; SAVETMPS; PUSHMARK (SP);	651	ENTER; SAVETMPS;
		652	PUSHMARK (SP);
395	EXTEND (SP, 2);	653	EXTEND (SP, 2);
396	// we re-bless the reference to get overload and other niceties right	654	// we re-bless the reference to get overload and other niceties right
397	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	655	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
398	PUSHs (sv_cbor);	656	PUSHs (sv_cbor);
399		657
…		…
404	// catch this surprisingly common error	662	// catch this surprisingly common error
405	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	663	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
406	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	664	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
407		665
408	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	666	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
409	encode_uint (enc, 0x80, count + 1);	667	encode_uint (enc, MAJOR_ARRAY, count + 1);
410	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	668	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
411		669
412	while (count)	670	{
		671	int i;
		672
		673	for (i = 0; i < count; ++i)
413	encode_sv (enc, SP[1 - count--]);	674	encode_sv (enc, SP[i + 1 - count]);
		675
		676	SP -= count;
		677	}
414		678
415	PUTBACK;	679	PUTBACK;
416		680
417	FREETMPS; LEAVE;	681	FREETMPS; LEAVE;
418	}	682	}
…		…
436	{	700	{
437	double nv = SvNVX (sv);	701	double nv = SvNVX (sv);
438		702
439	need (enc, 9);	703	need (enc, 9);
440		704
441	if (ecb_expect_false (nv == (U32)nv))	705	if (ecb_expect_false (nv == (NV)(U32)nv))
442	encode_uint (enc, 0x00, (U32)nv);	706	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
443	//TODO: maybe I32?	707	//TODO: maybe I32?
444	else if (ecb_expect_false (nv == (float)nv))	708	else if (ecb_expect_false (nv == (float)nv))
445	{	709	encode_float32 (enc, nv);
446	uint32_t fp = ecb_float_to_binary32 (nv);
447
448	*enc->cur++ = 0xe0 | 26;
449
450	if (!ecb_big_endian ())
451	fp = ecb_bswap32 (fp);
452
453	memcpy (enc->cur, &fp, 4);
454	enc->cur += 4;
455	}
456	else	710	else
457	{	711	encode_float64 (enc, nv);
458	uint64_t fp = ecb_double_to_binary64 (nv);
459
460	*enc->cur++ = 0xe0 | 27;
461
462	if (!ecb_big_endian ())
463	fp = ecb_bswap64 (fp);
464
465	memcpy (enc->cur, &fp, 8);
466	enc->cur += 8;
467	}
468	}	712	}
469		713
470	static void	714	static void
471	encode_sv (enc_t *enc, SV *sv)	715	encode_sv (enc_t *enc, SV *sv)
472	{	716	{
…		…
474		718
475	if (SvPOKp (sv))	719	if (SvPOKp (sv))
476	{	720	{
477	STRLEN len;	721	STRLEN len;
478	char *str = SvPV (sv, len);	722	char *str = SvPV (sv, len);
479	encode_str (enc, SvUTF8 (sv), str, len);	723	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
480	}	724	}
481	else if (SvNOKp (sv))	725	else if (SvNOKp (sv))
482	encode_nv (enc, sv);	726	encode_nv (enc, sv);
483	else if (SvIOKp (sv))	727	else if (SvIOKp (sv))
484	{	728	encode_int (enc, sv);
485	if (SvIsUV (sv))
486	encode_uint (enc, 0x00, SvUVX (sv));
487	else if (SvIVX (sv) >= 0)
488	encode_uint (enc, 0x00, SvIVX (sv));
489	else
490	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
491	}
492	else if (SvROK (sv))	729	else if (SvROK (sv))
493	encode_rv (enc, SvRV (sv));	730	encode_rv (enc, SvRV (sv));
494	else if (!SvOK (sv))	731	else if (!SvOK (sv))
495	encode_ch (enc, 0xe0 | 22);	732	encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
496	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	733	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
497	encode_ch (enc, 0xe0 | 23);	734	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
498	else	735	else
499	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	736	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
500	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	737	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
501	}	738	}
502		739
503	static SV *	740	static SV *
504	encode_cbor (SV *scalar, CBOR *cbor)	741	encode_cbor (SV *scalar, CBOR *cbor)
505	{	742	{
506	enc_t enc = { };	743	enc_t enc = { 0 };
507		744
508	enc.cbor = *cbor;	745	enc.cbor = *cbor;
509	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	746	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
510	enc.cur = SvPVX (enc.sv);	747	enc.cur = SvPVX (enc.sv);
511	enc.end = SvEND (enc.sv);	748	enc.end = SvEND (enc.sv);
512		749
513	SvPOK_only (enc.sv);	750	SvPOK_only (enc.sv);
514		751
515	if (cbor->flags & F_ALLOW_STRINGREF)	752	if (cbor->flags & F_PACK_STRINGS)
516	{	753	{
517	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	754	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
518	enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());	755	enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
519	enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());	756	enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
520	}	757	}
…		…
542	CBOR cbor;	779	CBOR cbor;
543	U32 depth; // recursion depth	780	U32 depth; // recursion depth
544	U32 maxdepth; // recursion depth limit	781	U32 maxdepth; // recursion depth limit
545	AV *shareable;	782	AV *shareable;
546	AV *stringref;	783	AV *stringref;
		784	SV *decode_tagged;
		785	SV *err_sv; // optional sv for error, needs to be freed
547	} dec_t;	786	} dec_t;
548		787
549	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	788	// set dec->err to ERRSV
		789	ecb_cold static void
		790	err_errsv (dec_t *dec)
		791	{
		792	if (!dec->err)
		793	{
		794	dec->err_sv = newSVsv (ERRSV);
550		795
551	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	796	// chop off the trailing \n
		797	SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
		798	*SvEND (dec->err_sv) = 0;
552		799
		800	dec->err = SvPVutf8_nolen (dec->err_sv);
		801	}
		802	}
		803
		804	// the following functions are used to reduce code size and help the compiler to optimise
		805	ecb_cold static void
		806	err_set (dec_t *dec, const char *reason)
		807	{
		808	if (!dec->err)
		809	dec->err = reason;
		810	}
		811
		812	ecb_cold static void
		813	err_unexpected_end (dec_t *dec)
		814	{
		815	err_set (dec, "unexpected end of CBOR data");
		816	}
		817
		818	#define ERR_DO(do) SB do; goto fail; SE
		819	#define ERR(reason) ERR_DO (err_set (dec, reason))
		820	#define ERR_ERRSV ERR_DO (err_errsv (dec))
		821
		822	#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
		823
553	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	824	#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
554	#define DEC_DEC_DEPTH --dec->depth	825	#define DEC_DEC_DEPTH --dec->depth
555		826
556	static UV	827	static UV
557	decode_uint (dec_t *dec)	828	decode_uint (dec_t *dec)
558	{	829	{
559	switch (*dec->cur & 31)	830	U8 m = *dec->cur & MINOR_MASK;
560	{	831	++dec->cur;
561	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
562	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
563	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
564	return *dec->cur++ & 31;
565		832
566	case 24:	833	if (ecb_expect_true (m < LENGTH_EXT1))
		834	return m;
		835	else if (ecb_expect_true (m == LENGTH_EXT1))
		836	{
567	WANT (2);	837	WANT (1);
568	dec->cur += 2;	838	dec->cur += 1;
569	return dec->cur[-1];	839	return dec->cur[-1];
570		840	}
571	case 25:	841	else if (ecb_expect_true (m == LENGTH_EXT2))
		842	{
572	WANT (3);	843	WANT (2);
573	dec->cur += 3;	844	dec->cur += 2;
574	return (((UV)dec->cur[-2]) << 8)	845	return (((UV)dec->cur[-2]) << 8)
575	| ((UV)dec->cur[-1]);	846	| ((UV)dec->cur[-1]);
576		847	}
577	case 26:	848	else if (ecb_expect_true (m == LENGTH_EXT4))
		849	{
578	WANT (5);	850	WANT (4);
579	dec->cur += 5;	851	dec->cur += 4;
580	return (((UV)dec->cur[-4]) << 24)	852	return (((UV)dec->cur[-4]) << 24)
581	| (((UV)dec->cur[-3]) << 16)	853	| (((UV)dec->cur[-3]) << 16)
582	| (((UV)dec->cur[-2]) << 8)	854	| (((UV)dec->cur[-2]) << 8)
583	| ((UV)dec->cur[-1]);	855	| ((UV)dec->cur[-1]);
584		856	}
585	case 27:	857	else if (ecb_expect_true (m == LENGTH_EXT8))
		858	{
586	WANT (9);	859	WANT (8);
587	dec->cur += 9;	860	dec->cur += 8;
		861
		862	return
		863	#if UVSIZE < 8
		864	0
		865	#else
588	return (((UV)dec->cur[-8]) << 56)	866	(((UV)dec->cur[-8]) << 56)
589	| (((UV)dec->cur[-7]) << 48)	867	| (((UV)dec->cur[-7]) << 48)
590	| (((UV)dec->cur[-6]) << 40)	868	| (((UV)dec->cur[-6]) << 40)
591	| (((UV)dec->cur[-5]) << 32)	869	| (((UV)dec->cur[-5]) << 32)
		870	#endif
592	| (((UV)dec->cur[-4]) << 24)	871	| (((UV)dec->cur[-4]) << 24)
593	| (((UV)dec->cur[-3]) << 16)	872	| (((UV)dec->cur[-3]) << 16)
594	| (((UV)dec->cur[-2]) << 8)	873	| (((UV)dec->cur[-2]) << 8)
595	| ((UV)dec->cur[-1]);	874	| ((UV)dec->cur[-1]);
596		875	}
597	default:	876	else
598	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	877	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
599	}
600		878
601	fail:	879	fail:
602	return 0;	880	return 0;
603	}	881	}
604		882
…		…
609	{	887	{
610	AV *av = newAV ();	888	AV *av = newAV ();
611		889
612	DEC_INC_DEPTH;	890	DEC_INC_DEPTH;
613		891
614	if ((*dec->cur & 31) == 31)	892	if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
615	{	893	{
616	++dec->cur;	894	++dec->cur;
617		895
618	for (;;)	896	for (;;)
619	{	897	{
620	WANT (1);	898	WANT (1);
621		899
622	if (*dec->cur == (0xe0 | 31))	900	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
623	{	901	{
624	++dec->cur;	902	++dec->cur;
625	break;	903	break;
626	}	904	}
627		905
628	av_push (av, decode_sv (dec));	906	av_push (av, decode_sv (dec));
629	}	907	}
630	}	908	}
631	else	909	else
632	{	910	{
633	int i, len = decode_uint (dec);	911	UV i, len = decode_uint (dec);
634		912
		913	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
635	av_fill (av, len - 1);	914	av_fill (av, len - 1);
636		915
637	for (i = 0; i < len; ++i)	916	for (i = 0; i < len; ++i)
638	AvARRAY (av)[i] = decode_sv (dec);	917	AvARRAY (av)[i] = decode_sv (dec);
639	}	918	}
640		919
641	DEC_DEC_DEPTH;	920	DEC_DEC_DEPTH;
642	return newRV_noinc ((SV *)av);	921	return newRV_noinc ((SV *)av);
643		922
644	fail:	923	fail:
645	SvREFCNT_dec (av);	924	SvREFCNT_dec_NN (av);
646	DEC_DEC_DEPTH;	925	DEC_DEC_DEPTH;
647	return &PL_sv_undef;	926	return &PL_sv_undef;
648	}	927	}
649		928
650	static void	929	static void
…		…
652	{	931	{
653	// for speed reasons, we specialcase single-string	932	// for speed reasons, we specialcase single-string
654	// byte or utf-8 strings as keys, but only when !stringref	933	// byte or utf-8 strings as keys, but only when !stringref
655		934
656	if (ecb_expect_true (!dec->stringref))	935	if (ecb_expect_true (!dec->stringref))
657	if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)	936	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
658	{	937	{
659	I32 len = decode_uint (dec);	938	STRLEN len = decode_uint (dec);
660	char *key = (char *)dec->cur;	939	char *key = (char *)dec->cur;
661		940
		941	WANT (len);
662	dec->cur += len;	942	dec->cur += len;
663		943
664	if (ecb_expect_false (dec->stringref))
665	av_push (dec->stringref, newSVpvn (key, len));
666
667	hv_store (hv, key, len, decode_sv (dec), 0);	944	hv_store (hv, key, len, decode_sv (dec), 0);
668		945
669	return;	946	return;
670	}	947	}
671	else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)	948	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
672	{	949	{
673	I32 len = decode_uint (dec);	950	STRLEN len = decode_uint (dec);
674	char *key = (char *)dec->cur;	951	char *key = (char *)dec->cur;
675		952
		953	WANT (len);
676	dec->cur += len;	954	dec->cur += len;
677		955
678	if (ecb_expect_false (dec->stringref))	956	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
679	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	957	if (!is_utf8_string (key, len))
		958	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
680		959
681	hv_store (hv, key, -len, decode_sv (dec), 0);	960	hv_store (hv, key, -len, decode_sv (dec), 0);
682		961
683	return;	962	return;
684	}	963	}
685		964
686	SV *k = decode_sv (dec);	965	SV *k = decode_sv (dec);
687	SV *v = decode_sv (dec);	966	SV *v = decode_sv (dec);
688		967
		968	// we leak memory if uncaught exceptions are thrown by random magical
		969	// methods, and this is hopefully the only place where it can happen,
		970	// so if there is a chance of an exception, take the very slow path.
		971	// since catching exceptions is "undocumented/internal/forbidden" by
		972	// the new p5p powers, we need to call out to a perl function :/
		973	if (ecb_expect_false (SvAMAGIC (k)))
		974	{
		975	dSP;
		976
		977	ENTER; SAVETMPS;
		978	PUSHMARK (SP);
		979	EXTEND (SP, 3);
		980	PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
		981	PUSHs (sv_2mortal (k));
		982	PUSHs (sv_2mortal (v));
		983
		984	PUTBACK;
		985	call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
		986	SPAGAIN;
		987
		988	FREETMPS; LEAVE;
		989
		990	if (SvTRUE (ERRSV))
		991	ERR_ERRSV;
		992
		993	return;
		994	}
		995
689	hv_store_ent (hv, k, v, 0);	996	hv_store_ent (hv, k, v, 0);
690	SvREFCNT_dec (k);	997	SvREFCNT_dec_NN (k);
		998
		999	fail:
		1000	;
691	}	1001	}
692		1002
693	static SV *	1003	static SV *
694	decode_hv (dec_t *dec)	1004	decode_hv (dec_t *dec)
695	{	1005	{
696	HV *hv = newHV ();	1006	HV *hv = newHV ();
697		1007
698	DEC_INC_DEPTH;	1008	DEC_INC_DEPTH;
699		1009
700	if ((*dec->cur & 31) == 31)	1010	if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
701	{	1011	{
702	++dec->cur;	1012	++dec->cur;
703		1013
704	for (;;)	1014	for (;;)
705	{	1015	{
706	WANT (1);	1016	WANT (1);
707		1017
708	if (*dec->cur == (0xe0 | 31))	1018	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
709	{	1019	{
710	++dec->cur;	1020	++dec->cur;
711	break;	1021	break;
712	}	1022	}
713		1023
714	decode_he (dec, hv);	1024	decode_he (dec, hv);
715	}	1025	}
716	}	1026	}
717	else	1027	else
718	{	1028	{
719	int pairs = decode_uint (dec);	1029	UV pairs = decode_uint (dec);
		1030
		1031	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
720		1032
721	while (pairs--)	1033	while (pairs--)
722	decode_he (dec, hv);	1034	decode_he (dec, hv);
723	}	1035	}
724		1036
725	DEC_DEC_DEPTH;	1037	DEC_DEC_DEPTH;
726	return newRV_noinc ((SV *)hv);	1038	return newRV_noinc ((SV *)hv);
727		1039
728	fail:	1040	fail:
729	SvREFCNT_dec (hv);	1041	SvREFCNT_dec_NN (hv);
730	DEC_DEC_DEPTH;	1042	DEC_DEC_DEPTH;
731	return &PL_sv_undef;	1043	return &PL_sv_undef;
732	}	1044	}
733		1045
734	static SV *	1046	static SV *
735	decode_str (dec_t *dec, int utf8)	1047	decode_str (dec_t *dec, int utf8)
736	{	1048	{
737	SV *sv = 0;	1049	SV *sv = 0;
738		1050
739	if ((*dec->cur & 31) == 31)	1051	if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
740	{	1052	{
		1053	// indefinite length strings
741	++dec->cur;	1054	++dec->cur;
742		1055
		1056	U8 major = *dec->cur & MAJOR_MISC;
		1057
743	sv = newSVpvn ("", 0);	1058	sv = newSVpvn ("", 0);
744		1059
745	// not very fast, and certainly not robust against illegal input
746	for (;;)	1060	for (;;)
747	{	1061	{
748	WANT (1);	1062	WANT (1);
749		1063
750	if (*dec->cur == (0xe0 | 31))	1064	if ((*dec->cur - major) > LENGTH_EXT8)
		1065	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
751	{	1066	{
752	++dec->cur;	1067	++dec->cur;
753	break;	1068	break;
754	}	1069	}
		1070	else
		1071	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
755		1072
756	sv_catsv (sv, decode_sv (dec));	1073	STRLEN len = decode_uint (dec);
		1074
		1075	WANT (len);
		1076	sv_catpvn (sv, dec->cur, len);
		1077	dec->cur += len;
757	}	1078	}
758	}	1079	}
759	else	1080	else
760	{	1081	{
761	STRLEN len = decode_uint (dec);	1082	STRLEN len = decode_uint (dec);
…		…
768	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	1089	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
769	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	1090	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770	}	1091	}
771		1092
772	if (utf8)	1093	if (utf8)
		1094	{
		1095	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		1096	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		1097	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		1098
773	SvUTF8_on (sv);	1099	SvUTF8_on (sv);
		1100	}
774		1101
775	return sv;	1102	return sv;
776		1103
777	fail:	1104	fail:
778	SvREFCNT_dec (sv);	1105	SvREFCNT_dec (sv);
…		…
797	sv = newRV_noinc (decode_sv (dec));	1124	sv = newRV_noinc (decode_sv (dec));
798	break;	1125	break;
799		1126
800	case CBOR_TAG_STRINGREF_NAMESPACE:	1127	case CBOR_TAG_STRINGREF_NAMESPACE:
801	{	1128	{
		1129	// do not use SAVETMPS/FREETMPS, as these will
		1130	// erase mortalised caches, e.g. "shareable"
802	ENTER; SAVETMPS;	1131	ENTER;
803		1132
804	SAVESPTR (dec->stringref);	1133	SAVESPTR (dec->stringref);
805	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());	1134	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
806		1135
807	sv = decode_sv (dec);	1136	sv = decode_sv (dec);
808		1137
809	FREETMPS; LEAVE;	1138	LEAVE;
810	}	1139	}
811	break;	1140	break;
812		1141
813	case CBOR_TAG_STRINGREF:	1142	case CBOR_TAG_STRINGREF:
814	{	1143	{
815	if ((*dec->cur >> 5) != 0)	1144	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
816	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1145	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
817		1146
818	UV idx = decode_uint (dec);	1147	UV idx = decode_uint (dec);
819		1148
820	if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))	1149	if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
821	ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");	1150	ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
822		1151
823	sv = newSVsv (AvARRAY (dec->stringref)[idx]);	1152	sv = newSVsv (AvARRAY (dec->stringref)[idx]);
824	}	1153	}
825	break;	1154	break;
…		…
827	case CBOR_TAG_VALUE_SHAREABLE:	1156	case CBOR_TAG_VALUE_SHAREABLE:
828	{	1157	{
829	if (ecb_expect_false (!dec->shareable))	1158	if (ecb_expect_false (!dec->shareable))
830	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());	1159	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
831		1160
		1161	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1162	{
832	sv = newSV (0);	1163	sv = newSV (0);
833	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1164	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
834		1165
835	SV *osv = decode_sv (dec);	1166	SV *osv = decode_sv (dec);
836	sv_setsv (sv, osv);	1167	sv_setsv (sv, osv);
837	SvREFCNT_dec_NN (osv);	1168	SvREFCNT_dec_NN (osv);
		1169	}
		1170	else
		1171	{
		1172	av_push (dec->shareable, &PL_sv_undef);
		1173	int idx = AvFILLp (dec->shareable);
		1174	sv = decode_sv (dec);
		1175	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		1176	}
838	}	1177	}
839	break;	1178	break;
840		1179
841	case CBOR_TAG_VALUE_SHAREDREF:	1180	case CBOR_TAG_VALUE_SHAREDREF:
842	{	1181	{
843	if ((*dec->cur >> 5) != 0)	1182	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
844	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1183	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
845		1184
846	UV idx = decode_uint (dec);	1185	UV idx = decode_uint (dec);
847		1186
848	if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))	1187	if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
849	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1188	ERR ("corrupted CBOR data (sharedref index out of bounds)");
850		1189
851	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1190	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1191
		1192	if (sv == &PL_sv_undef)
		1193	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
852	}	1194	}
853	break;	1195	break;
854		1196
855	case CBOR_TAG_PERL_OBJECT:	1197	case CBOR_TAG_PERL_OBJECT:
856	{	1198	{
		1199	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1200	goto filter;
		1201
857	sv = decode_sv (dec);	1202	sv = decode_sv (dec);
858		1203
859	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)	1204	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
860	ERR ("corrupted CBOR data (non-array perl object)");	1205	ERR ("corrupted CBOR data (non-array perl object)");
861		1206
…		…
871	if (!method)	1216	if (!method)
872	ERR ("cannot decode perl-object (package does not have a THAW method)");	1217	ERR ("cannot decode perl-object (package does not have a THAW method)");
873		1218
874	dSP;	1219	dSP;
875		1220
876	ENTER; SAVETMPS; PUSHMARK (SP);	1221	ENTER; SAVETMPS;
		1222	PUSHMARK (SP);
877	EXTEND (SP, len + 1);	1223	EXTEND (SP, len + 1);
878	// we re-bless the reference to get overload and other niceties right	1224	// we re-bless the reference to get overload and other niceties right
879	PUSHs (*av_fetch (av, 0, 1));	1225	PUSHs (*av_fetch (av, 0, 1));
880	PUSHs (sv_cbor);	1226	PUSHs (sv_cbor);
881		1227
…		…
889	SPAGAIN;	1235	SPAGAIN;
890		1236
891	if (SvTRUE (ERRSV))	1237	if (SvTRUE (ERRSV))
892	{	1238	{
893	FREETMPS; LEAVE;	1239	FREETMPS; LEAVE;
894	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1240	ERR_ERRSV;
895	}	1241	}
896		1242
897	SvREFCNT_dec (sv);	1243	SvREFCNT_dec_NN (sv);
898	sv = SvREFCNT_inc (POPs);	1244	sv = SvREFCNT_inc (POPs);
899		1245
900	PUTBACK;	1246	PUTBACK;
901		1247
902	FREETMPS; LEAVE;	1248	FREETMPS; LEAVE;
903	}	1249	}
904	break;	1250	break;
905		1251
906	default:	1252	default:
		1253	filter:
907	{	1254	{
		1255	SV *tag_sv = newSVuv (tag);
		1256
908	sv = decode_sv (dec);	1257	sv = decode_sv (dec);
909		1258
		1259	dSP;
		1260	ENTER; SAVETMPS;
		1261	PUSHMARK (SP);
		1262	EXTEND (SP, 2);
		1263	PUSHs (tag_sv);
		1264	PUSHs (sv);
		1265
		1266	PUTBACK;
		1267	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
		1268	SPAGAIN;
		1269
		1270	if (SvTRUE (ERRSV))
		1271	{
		1272	SvREFCNT_dec_NN (tag_sv);
		1273	FREETMPS; LEAVE;
		1274	ERR_ERRSV;
		1275	}
		1276
		1277	if (count)
		1278	{
		1279	SvREFCNT_dec_NN (tag_sv);
		1280	SvREFCNT_dec_NN (sv);
		1281	sv = SvREFCNT_inc_NN (TOPs);
		1282	SP -= count;
		1283	}
		1284	else
		1285	{
910	AV *av = newAV ();	1286	AV *av = newAV ();
911	av_push (av, newSVuv (tag));	1287	av_push (av, tag_sv);
912	av_push (av, sv);	1288	av_push (av, sv);
913		1289
914	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	1290	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
915	? cbor_tagged_stash	1291	? cbor_tagged_stash
916	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1292	: gv_stashpv ("CBOR::XS::Tagged" , 1);
917
918	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1293	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1294	}
		1295
		1296	PUTBACK;
		1297
		1298	FREETMPS; LEAVE;
919	}	1299	}
920	break;	1300	break;
921	}	1301	}
922		1302
923	return sv;	1303	return sv;
…		…
930	static SV *	1310	static SV *
931	decode_sv (dec_t *dec)	1311	decode_sv (dec_t *dec)
932	{	1312	{
933	WANT (1);	1313	WANT (1);
934		1314
935	switch (*dec->cur >> 5)	1315	switch (*dec->cur >> MAJOR_SHIFT)
936	{	1316	{
937	case 0: // unsigned int	1317	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
938	return newSVuv (decode_uint (dec));	1318	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
939	case 1: // negative int	1319	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
940	return newSViv (-1 - (IV)decode_uint (dec));	1320	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
941	case 2: // octet string	1321	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
942	return decode_str (dec, 0);	1322	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
943	case 3: // utf-8 string	1323	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
944	return decode_str (dec, 1);	1324
945	case 4: // array	1325	case MAJOR_MISC >> MAJOR_SHIFT:
946	return decode_av (dec);
947	case 5: // map
948	return decode_hv (dec);
949	case 6: // tag
950	return decode_tagged (dec);
951	case 7: // misc
952	switch (*dec->cur++ & 31)	1326	switch (*dec->cur++ & MINOR_MASK)
953	{	1327	{
954	case 20:	1328	case SIMPLE_FALSE:
955	#if CBOR_SLOW	1329	#if CBOR_SLOW
956	types_false = get_bool ("Types::Serialiser::false");	1330	types_false = get_bool ("Types::Serialiser::false");
957	#endif	1331	#endif
958	return newSVsv (types_false);	1332	return newSVsv (types_false);
959	case 21:	1333	case SIMPLE_TRUE:
960	#if CBOR_SLOW	1334	#if CBOR_SLOW
961	types_true = get_bool ("Types::Serialiser::true");	1335	types_true = get_bool ("Types::Serialiser::true");
962	#endif	1336	#endif
963	return newSVsv (types_true);	1337	return newSVsv (types_true);
964	case 22:	1338	case SIMPLE_NULL:
965	return newSVsv (&PL_sv_undef);	1339	return newSVsv (&PL_sv_undef);
966	case 23:	1340	case SIMPLE_UNDEF:
967	#if CBOR_SLOW	1341	#if CBOR_SLOW
968	types_error = get_bool ("Types::Serialiser::error");	1342	types_error = get_bool ("Types::Serialiser::error");
969	#endif	1343	#endif
970	return newSVsv (types_error);	1344	return newSVsv (types_error);
971		1345
972	case 25:	1346	case MISC_FLOAT16:
973	{	1347	{
974	WANT (2);	1348	WANT (2);
975		1349
976	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];	1350	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
977	dec->cur += 2;	1351	dec->cur += 2;
978		1352
979	return newSVnv (ecb_binary16_to_float (fp));	1353	return newSVnv (ecb_binary16_to_float (fp));
980	}	1354	}
981		1355
982	case 26:	1356	case MISC_FLOAT32:
983	{	1357	{
984	uint32_t fp;	1358	uint32_t fp;
985	WANT (4);	1359	WANT (4);
986	memcpy (&fp, dec->cur, 4);	1360	memcpy (&fp, dec->cur, 4);
987	dec->cur += 4;	1361	dec->cur += 4;
…		…
990	fp = ecb_bswap32 (fp);	1364	fp = ecb_bswap32 (fp);
991		1365
992	return newSVnv (ecb_binary32_to_float (fp));	1366	return newSVnv (ecb_binary32_to_float (fp));
993	}	1367	}
994		1368
995	case 27:	1369	case MISC_FLOAT64:
996	{	1370	{
997	uint64_t fp;	1371	uint64_t fp;
998	WANT (8);	1372	WANT (8);
999	memcpy (&fp, dec->cur, 8);	1373	memcpy (&fp, dec->cur, 8);
1000	dec->cur += 8;	1374	dec->cur += 8;
…		…
1003	fp = ecb_bswap64 (fp);	1377	fp = ecb_bswap64 (fp);
1004		1378
1005	return newSVnv (ecb_binary64_to_double (fp));	1379	return newSVnv (ecb_binary64_to_double (fp));
1006	}	1380	}
1007		1381
1008	// 0..19 unassigned	1382	// 0..19 unassigned simple
1009	// 24 reserved + unassigned (reserved values are not encodable)	1383	// 24 reserved + unassigned simple (reserved values are not encodable)
		1384	// 28-30 unassigned misc
		1385	// 31 break code
1010	default:	1386	default:
1011	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1387	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1012	}	1388	}
1013		1389
1014	break;	1390	break;
1015	}	1391	}
1016		1392
…		…
1019	}	1395	}
1020		1396
1021	static SV *	1397	static SV *
1022	decode_cbor (SV *string, CBOR *cbor, char **offset_return)	1398	decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1023	{	1399	{
1024	dec_t dec = { };	1400	dec_t dec = { 0 };
1025	SV *sv;	1401	SV *sv;
1026	STRLEN len;	1402	STRLEN len;
1027	char *data = SvPVbyte (string, len);	1403	char *data = SvPVbyte (string, len);
1028		1404
1029	if (len > cbor->max_size && cbor->max_size)	1405	if (len > cbor->max_size && cbor->max_size)
…		…
1043	if (dec.cur != dec.end && !dec.err)	1419	if (dec.cur != dec.end && !dec.err)
1044	dec.err = "garbage after CBOR object";	1420	dec.err = "garbage after CBOR object";
1045		1421
1046	if (dec.err)	1422	if (dec.err)
1047	{	1423	{
		1424	if (dec.shareable)
		1425	{
		1426	// need to break cyclic links, which would all be in shareable
		1427	int i;
		1428	SV **svp;
		1429
		1430	for (i = av_len (dec.shareable) + 1; i--; )
		1431	if ((svp = av_fetch (dec.shareable, i, 0)))
		1432	sv_setsv (*svp, &PL_sv_undef);
		1433	}
		1434
1048	SvREFCNT_dec (sv);	1435	SvREFCNT_dec_NN (sv);
		1436
		1437	if (dec.err_sv)
		1438	sv_2mortal (dec.err_sv);
		1439
1049	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);	1440	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1050	}	1441	}
1051		1442
1052	sv = sv_2mortal (sv);	1443	sv = sv_2mortal (sv);
1053		1444
1054	return sv;	1445	return sv;
1055	}	1446	}
1056		1447
		1448	/////////////////////////////////////////////////////////////////////////////
		1449	// incremental parser
		1450
		1451	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1452
		1453	// returns 0 for notyet, 1 for success or error
		1454	static int
		1455	incr_parse (CBOR *self, SV *cborstr)
		1456	{
		1457	STRLEN cur;
		1458	SvPV (cborstr, cur);
		1459
		1460	while (ecb_expect_true (self->incr_need <= cur))
		1461	{
		1462	// table of integer count bytes
		1463	static I8 incr_len[MINOR_MASK + 1] = {
		1464	0, 0, 0, 0, 0, 0, 0, 0,
		1465	0, 0, 0, 0, 0, 0, 0, 0,
		1466	0, 0, 0, 0, 0, 0, 0, 0,
		1467	1, 2, 4, 8,-1,-1,-1,-2
		1468	};
		1469
		1470	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1471	U8 m = *p & MINOR_MASK;
		1472	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1473	I8 ilen = incr_len[m];
		1474
		1475	self->incr_need = self->incr_pos + 1;
		1476
		1477	if (ecb_expect_false (ilen < 0))
		1478	{
		1479	if (m != MINOR_INDEF)
		1480	return 1; // error
		1481
		1482	if (*p == (MAJOR_MISC | MINOR_INDEF))
		1483	{
		1484	if (count >= 0)
		1485	return 1; // error
		1486
		1487	count = 1;
		1488	}
		1489	else
		1490	{
		1491	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1492	count = -1;
		1493	}
		1494	}
		1495	else
		1496	{
		1497	self->incr_need += ilen;
		1498	if (ecb_expect_false (self->incr_need > cur))
		1499	return 0;
		1500
		1501	int major = *p >> MAJOR_SHIFT;
		1502
		1503	switch (major)
		1504	{
		1505	case MAJOR_TAG >> MAJOR_SHIFT:
		1506	++count; // tags merely prefix another value
		1507	break;
		1508
		1509	case MAJOR_BYTES >> MAJOR_SHIFT:
		1510	case MAJOR_TEXT >> MAJOR_SHIFT:
		1511	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1512	case MAJOR_MAP >> MAJOR_SHIFT:
		1513	{
		1514	UV len;
		1515
		1516	if (ecb_expect_false (ilen))
		1517	{
		1518	len = 0;
		1519
		1520	do {
		1521	len = (len << 8) | *++p;
		1522	} while (--ilen);
		1523	}
		1524	else
		1525	len = m;
		1526
		1527	switch (major)
		1528	{
		1529	case MAJOR_BYTES >> MAJOR_SHIFT:
		1530	case MAJOR_TEXT >> MAJOR_SHIFT:
		1531	self->incr_need += len;
		1532	if (ecb_expect_false (self->incr_need > cur))
		1533	return 0;
		1534
		1535	break;
		1536
		1537	case MAJOR_MAP >> MAJOR_SHIFT:
		1538	len <<= 1;
		1539	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1540	if (len)
		1541	{
		1542	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1543	count = len + 1;
		1544	}
		1545	break;
		1546	}
		1547	}
		1548	}
		1549	}
		1550
		1551	self->incr_pos = self->incr_need;
		1552
		1553	if (count > 0)
		1554	{
		1555	while (!--count)
		1556	{
		1557	if (!AvFILLp (self->incr_count))
		1558	return 1; // done
		1559
		1560	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1561	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1562	}
		1563
		1564	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1565	}
		1566	}
		1567
		1568	return 0;
		1569	}
		1570
		1571
1057	/////////////////////////////////////////////////////////////////////////////	1572	/////////////////////////////////////////////////////////////////////////////
1058	// XS interface functions	1573	// XS interface functions
1059		1574
1060	MODULE = CBOR::XS PACKAGE = CBOR::XS	1575	MODULE = CBOR::XS PACKAGE = CBOR::XS
1061		1576
…		…
1069		1584
1070	types_true = get_bool ("Types::Serialiser::true" );	1585	types_true = get_bool ("Types::Serialiser::true" );
1071	types_false = get_bool ("Types::Serialiser::false");	1586	types_false = get_bool ("Types::Serialiser::false");
1072	types_error = get_bool ("Types::Serialiser::error");	1587	types_error = get_bool ("Types::Serialiser::error");
1073		1588
		1589	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
		1590
1074	sv_cbor = newSVpv ("CBOR", 0);	1591	sv_cbor = newSVpv ("CBOR", 0);
1075	SvREADONLY_on (sv_cbor);	1592	SvREADONLY_on (sv_cbor);
		1593
		1594	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1076	}	1595	}
1077		1596
1078	PROTOTYPES: DISABLE	1597	PROTOTYPES: DISABLE
1079		1598
1080	void CLONE (...)	1599	void CLONE (...)
…		…
1099	void shrink (CBOR *self, int enable = 1)	1618	void shrink (CBOR *self, int enable = 1)
1100	ALIAS:	1619	ALIAS:
1101	shrink = F_SHRINK	1620	shrink = F_SHRINK
1102	allow_unknown = F_ALLOW_UNKNOWN	1621	allow_unknown = F_ALLOW_UNKNOWN
1103	allow_sharing = F_ALLOW_SHARING	1622	allow_sharing = F_ALLOW_SHARING
1104	allow_stringref = F_ALLOW_STRINGREF	1623	allow_cycles = F_ALLOW_CYCLES
		1624	forbid_objects = F_FORBID_OBJECTS
		1625	pack_strings = F_PACK_STRINGS
		1626	text_keys = F_TEXT_KEYS
		1627	text_strings = F_TEXT_STRINGS
		1628	validate_utf8 = F_VALIDATE_UTF8
1105	PPCODE:	1629	PPCODE:
1106	{	1630	{
1107	if (enable)	1631	if (enable)
1108	self->flags |= ix;	1632	self->flags |= ix;
1109	else	1633	else
…		…
1115	void get_shrink (CBOR *self)	1639	void get_shrink (CBOR *self)
1116	ALIAS:	1640	ALIAS:
1117	get_shrink = F_SHRINK	1641	get_shrink = F_SHRINK
1118	get_allow_unknown = F_ALLOW_UNKNOWN	1642	get_allow_unknown = F_ALLOW_UNKNOWN
1119	get_allow_sharing = F_ALLOW_SHARING	1643	get_allow_sharing = F_ALLOW_SHARING
1120	get_allow_stringref = F_ALLOW_STRINGREF	1644	get_allow_cycles = F_ALLOW_CYCLES
		1645	get_forbid_objects = F_FORBID_OBJECTS
		1646	get_pack_strings = F_PACK_STRINGS
		1647	get_text_keys = F_TEXT_KEYS
		1648	get_text_strings = F_TEXT_STRINGS
		1649	get_validate_utf8 = F_VALIDATE_UTF8
1121	PPCODE:	1650	PPCODE:
1122	XPUSHs (boolSV (self->flags & ix));	1651	XPUSHs (boolSV (self->flags & ix));
1123		1652
1124	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1653	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1125	PPCODE:	1654	PPCODE:
…		…
1138	XPUSHs (ST (0));	1667	XPUSHs (ST (0));
1139		1668
1140	int get_max_size (CBOR *self)	1669	int get_max_size (CBOR *self)
1141	CODE:	1670	CODE:
1142	RETVAL = self->max_size;	1671	RETVAL = self->max_size;
		1672	OUTPUT:
		1673	RETVAL
		1674
		1675	void filter (CBOR *self, SV *filter = 0)
		1676	PPCODE:
		1677	SvREFCNT_dec (self->filter);
		1678	self->filter = filter ? newSVsv (filter) : filter;
		1679	XPUSHs (ST (0));
		1680
		1681	SV *get_filter (CBOR *self)
		1682	CODE:
		1683	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1143	OUTPUT:	1684	OUTPUT:
1144	RETVAL	1685	RETVAL
1145		1686
1146	void encode (CBOR *self, SV *scalar)	1687	void encode (CBOR *self, SV *scalar)
1147	PPCODE:	1688	PPCODE:
…		…
1162	EXTEND (SP, 2);	1703	EXTEND (SP, 2);
1163	PUSHs (sv);	1704	PUSHs (sv);
1164	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1705	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1165	}	1706	}
1166		1707
		1708	void incr_parse (CBOR *self, SV *cborstr)
		1709	ALIAS:
		1710	incr_parse_multiple = 1
		1711	PPCODE:
		1712	{
		1713	if (SvUTF8 (cborstr))
		1714	sv_utf8_downgrade (cborstr, 0);
		1715
		1716	if (!self->incr_count)
		1717	{
		1718	self->incr_count = newAV ();
		1719	self->incr_pos = 0;
		1720	self->incr_need = 1;
		1721
		1722	av_push (self->incr_count, newSViv (1));
		1723	}
		1724
		1725	do
		1726	{
		1727	if (!incr_parse (self, cborstr))
		1728	{
		1729	if (self->incr_need > self->max_size && self->max_size)
		1730	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1731	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1732
		1733	break;
		1734	}
		1735
		1736	SV *sv;
		1737	char *offset;
		1738
		1739	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1740	XPUSHs (sv);
		1741
		1742	sv_chop (cborstr, offset);
		1743
		1744	av_clear (self->incr_count);
		1745	av_push (self->incr_count, newSViv (1));
		1746
		1747	self->incr_pos = 0;
		1748	self->incr_need = self->incr_pos + 1;
		1749	}
		1750	while (ix);
		1751	}
		1752
		1753	void incr_reset (CBOR *self)
		1754	CODE:
		1755	{
		1756	SvREFCNT_dec (self->incr_count);
		1757	self->incr_count = 0;
		1758	}
		1759
		1760	void DESTROY (CBOR *self)
		1761	PPCODE:
		1762	cbor_free (self);
		1763
1167	PROTOTYPES: ENABLE	1764	PROTOTYPES: ENABLE
1168		1765
1169	void encode_cbor (SV *scalar)	1766	void encode_cbor (SV *scalar)
		1767	ALIAS:
		1768	encode_cbor = 0
		1769	encode_cbor_sharing = F_ALLOW_SHARING
1170	PPCODE:	1770	PPCODE:
1171	{	1771	{
1172	CBOR cbor;	1772	CBOR cbor;
1173	cbor_init (&cbor);	1773	cbor_init (&cbor);
		1774	cbor.flags |= ix;
1174	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1775	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1175	XPUSHs (scalar);	1776	XPUSHs (scalar);
1176	}	1777	}
1177		1778
1178	void decode_cbor (SV *cborstr)	1779	void decode_cbor (SV *cborstr)
…		…
1182	cbor_init (&cbor);	1783	cbor_init (&cbor);
1183	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;	1784	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1184	XPUSHs (cborstr);	1785	XPUSHs (cborstr);
1185	}	1786	}
1186		1787
		1788	#ifdef __AFL_COMPILER
		1789
		1790	void
		1791	afl_init ()
		1792	CODE:
		1793	__AFL_INIT ();
		1794
		1795	int
		1796	afl_loop (unsigned int count = 10000)
		1797	CODE:
		1798	RETVAL = __AFL_LOOP (count);
		1799	OUTPUT:
		1800	RETVAL
		1801
		1802	#endif
		1803

Diff Legend

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