ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs.
Revision 1.69 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
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	// possibly future enhancements: (generic) float, (generic) string
60	};	115	};
61		116
62	#define F_SHRINK 0x00000001UL	117	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	118	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL	119	#define F_ALLOW_SHARING 0x00000004UL
		120	#define F_ALLOW_CYCLES 0x00000008UL
		121	#define F_FORBID_OBJECTS 0x00000010UL
65	#define F_PACK_STRINGS 0x00000008UL	122	#define F_PACK_STRINGS 0x00000020UL
		123	#define F_TEXT_KEYS 0x00000040UL
		124	#define F_TEXT_STRINGS 0x00000080UL
		125	#define F_VALIDATE_UTF8 0x00000100UL
66		126
67	#define INIT_SIZE 32 // initial scalar size to be allocated	127	#define INIT_SIZE 32 // initial scalar size to be allocated
68		128
69	#define SB do {	129	#define SB do {
70	#define SE } while (0)	130	#define SE } while (0)
…		…
89	typedef struct {	149	typedef struct {
90	U32 flags;	150	U32 flags;
91	U32 max_depth;	151	U32 max_depth;
92	STRLEN max_size;	152	STRLEN max_size;
93	SV *filter;	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.
94	} CBOR;	159	} CBOR;
95		160
96	ecb_inline void	161	ecb_inline void
97	cbor_init (CBOR *cbor)	162	cbor_init (CBOR *cbor)
98	{	163	{
…		…
102		167
103	ecb_inline void	168	ecb_inline void
104	cbor_free (CBOR *cbor)	169	cbor_free (CBOR *cbor)
105	{	170	{
106	SvREFCNT_dec (cbor->filter);	171	SvREFCNT_dec (cbor->filter);
		172	SvREFCNT_dec (cbor->incr_count);
107	}	173	}
108		174
109	/////////////////////////////////////////////////////////////////////////////	175	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	176	// utility functions
111		177
…		…
137		203
138	// minimum length of a string to be registered for stringref	204	// minimum length of a string to be registered for stringref
139	ecb_inline int	205	ecb_inline int
140	minimum_string_length (UV idx)	206	minimum_string_length (UV idx)
141	{	207	{
142	return idx > 23	208	return idx <= 23 ? 3
143	? idx > 0xffU	209	: idx <= 0xffU ? 4
144	? idx > 0xffffU	210	: idx <= 0xffffU ? 5
145	? idx > 0xffffffffU	211	: idx <= 0xffffffffU ? 7
146	? 11	212	: 11;
147	: 7
148	: 5
149	: 4
150	: 3;
151	}	213	}
152		214
153	/////////////////////////////////////////////////////////////////////////////	215	/////////////////////////////////////////////////////////////////////////////
154	// encoder	216	// encoder
155		217
…		…
168	} enc_t;	230	} enc_t;
169		231
170	ecb_inline void	232	ecb_inline void
171	need (enc_t *enc, STRLEN len)	233	need (enc_t *enc, STRLEN len)
172	{	234	{
173	if (ecb_expect_false (enc->cur + len >= enc->end))	235	if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
174	{	236	{
175	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);	237	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
176	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);	238	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
177	enc->cur = SvPVX (enc->sv) + cur;	239	enc->cur = SvPVX (enc->sv) + cur;
178	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;	240	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
…		…
184	{	246	{
185	need (enc, 1);	247	need (enc, 1);
186	*enc->cur++ = ch;	248	*enc->cur++ = ch;
187	}	249	}
188		250
		251	// used for tags, intregers, element counts and so on
189	static void	252	static void
190	encode_uint (enc_t *enc, int major, UV len)	253	encode_uint (enc_t *enc, int major, UV len)
191	{	254	{
192	need (enc, 9);	255	need (enc, 9);
193		256
194	if (ecb_expect_true (len < 24))	257	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major | len;	258	*enc->cur++ = major | len;
196	else if (ecb_expect_true (len <= 0xff))	259	else if (ecb_expect_true (len <= 0xffU))
197	{	260	{
198	*enc->cur++ = major | 24;	261	*enc->cur++ = major | LENGTH_EXT1;
199	*enc->cur++ = len;	262	*enc->cur++ = len;
200	}	263	}
201	else if (len <= 0xffff)	264	else if (len <= 0xffffU)
202	{	265	{
203	*enc->cur++ = major | 25;	266	*enc->cur++ = major | LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	267	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	268	*enc->cur++ = len;
206	}	269	}
207	else if (len <= 0xffffffff)	270	else if (len <= 0xffffffffU)
208	{	271	{
209	*enc->cur++ = major | 26;	272	*enc->cur++ = major | LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	273	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	274	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	275	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	276	*enc->cur++ = len;
214	}	277	}
215	else	278	else
216	{	279	{
217	*enc->cur++ = major | 27;	280	*enc->cur++ = major | LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	281	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	282	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	283	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	284	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	285	*enc->cur++ = len >> 24;
…		…
224	*enc->cur++ = len >> 8;	287	*enc->cur++ = len >> 8;
225	*enc->cur++ = len;	288	*enc->cur++ = len;
226	}	289	}
227	}	290	}
228		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
229	ecb_inline void	304	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	305	encode_tag (enc_t *enc, UV tag)
231	{	306	{
232	encode_uint (enc, 0xc0, tag);	307	encode_uint (enc, MAJOR_TAG, tag);
		308	}
		309
		310	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		311	ecb_noinline static void
		312	encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
		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	}
233	}	331	}
234		332
235	ecb_inline void	333	ecb_inline void
236	encode_str (enc_t *enc, int utf8, char *str, STRLEN len)	334	encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
237	{	335	{
238	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	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);
239	need (enc, len);	344	need (enc, len);
240	memcpy (enc->cur, str, len);	345	memcpy (enc->cur, str, len);
241	enc->cur += len;	346	enc->cur += len;
242	}	347	}
243		348
244	static void	349	ecb_inline void
245	encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)	350	encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
246	{	351	{
247	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))	352	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
248	{	353	{
249	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	354	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
250		355
251	if (SvOK (*svp))	356	if (SvOK (*svp))
252	{	357	{
253	// already registered, use stringref	358	// already registered, use stringref
254	encode_tag (enc, CBOR_TAG_STRINGREF);	359	encode_tag (enc, CBOR_TAG_STRINGREF);
255	encode_uint (enc, 0x00, SvUV (*svp));	360	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
256	return;	361	return;
257	}	362	}
258	else if (len >= minimum_string_length (enc->stringref_idx))	363	else if (len >= minimum_string_length (enc->stringref_idx))
259	{	364	{
260	// register only	365	// register only
261	sv_setuv (*svp, enc->stringref_idx);	366	sv_setuv (*svp, enc->stringref_idx);
262	++enc->stringref_idx;	367	++enc->stringref_idx;
263	}	368	}
264	}	369	}
265		370
266	encode_str (enc, utf8, str, 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	{
		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
		386	memcpy (enc->cur, &fp, 2);
		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	}
267	}	469	}
268		470
269	static void encode_sv (enc_t *enc, SV *sv);	471	static void encode_sv (enc_t *enc, SV *sv);
270		472
271	static void	473	static void
…		…
276	if (enc->depth >= enc->cbor.max_depth)	478	if (enc->depth >= enc->cbor.max_depth)
277	croak (ERR_NESTING_EXCEEDED);	479	croak (ERR_NESTING_EXCEEDED);
278		480
279	++enc->depth;	481	++enc->depth;
280		482
281	encode_uint (enc, 0x80, len + 1);	483	encode_uint (enc, MAJOR_ARRAY, len + 1);
282		484
		485	if (ecb_expect_false (SvMAGICAL (av)))
283	for (i = 0; i <= len; ++i)	486	for (i = 0; i <= len; ++i)
284	{	487	{
285	SV **svp = av_fetch (av, i, 0);	488	SV **svp = av_fetch (av, i, 0);
286	encode_sv (enc, svp ? *svp : &PL_sv_undef);	489	encode_sv (enc, svp ? *svp : &PL_sv_undef);
287	}	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	}
288		497
289	--enc->depth;	498	--enc->depth;
290	}	499	}
291		500
292	static void	501	static void
…		…
300	++enc->depth;	509	++enc->depth;
301		510
302	int pairs = hv_iterinit (hv);	511	int pairs = hv_iterinit (hv);
303	int mg = SvMAGICAL (hv);	512	int mg = SvMAGICAL (hv);
304		513
305	if (mg)	514	if (ecb_expect_false (mg))
306	encode_ch (enc, 0xa0 | 31);	515	encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
307	else	516	else
308	encode_uint (enc, 0xa0, pairs);	517	encode_uint (enc, MAJOR_MAP, pairs);
309		518
310	while ((he = hv_iternext (hv)))	519	while ((he = hv_iternext (hv)))
311	{	520	{
312	if (HeKLEN (he) == HEf_SVKEY)	521	if (HeKLEN (he) == HEf_SVKEY)
313	encode_sv (enc, HeSVKEY (he));	522	encode_sv (enc, HeSVKEY (he));
314	else	523	else
315	encode_strref (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));
316		525
317	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));
318	}	527	}
319		528
320	if (mg)	529	if (ecb_expect_false (mg))
321	encode_ch (enc, 0xe0 | 31);	530	encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
322		531
323	--enc->depth;	532	--enc->depth;
324	}	533	}
325		534
326	// encode objects, arrays and special \0=false and \1=true values.	535	// encode objects, arrays and special \0=false and \1=true values.
…		…
345		554
346	HV *stash = SvSTASH (sv);	555	HV *stash = SvSTASH (sv);
347		556
348	if (stash == boolean_stash)	557	if (stash == boolean_stash)
349	{	558	{
350	encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);	559	encode_bool (enc, SvIV (sv));
351	return;	560	return;
352	}	561	}
353	else if (stash == error_stash)	562	else if (stash == error_stash)
354	{	563	{
355	encode_ch (enc, 0xe0 | 23);	564	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
356	return;	565	return;
357	}	566	}
358	else if (stash == tagged_stash)	567	else if (stash == tagged_stash)
359	{	568	{
360	if (svt != SVt_PVAV)	569	if (svt != SVt_PVAV)
361	croak ("encountered CBOR::XS::Tagged object that isn't an array");	570	croak ("encountered CBOR::XS::Tagged object that isn't an array");
362		571
		572	switch (av_len ((AV *)sv))
		573	{
		574	case 2-1:
		575	// actually a tagged value
363	encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));	576	encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
364	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	}
365		588
366	return;	589	return;
367	}	590	}
368	}	591	}
369		592
370	if (ecb_expect_false (SvREFCNT (sv) > 1)	593	if (ecb_expect_false (SvREFCNT (sv) > 1)
371	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))	594	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
372	{	595	{
373	if (!enc->shareable)	596	if (ecb_expect_false (!enc->shareable))
374	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());	597	enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
375		598
376	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);	599	SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
377		600
378	if (SvOK (*svp))	601	if (SvOK (*svp))
379	{	602	{
380	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);	603	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381	encode_uint (enc, 0x00, SvUV (*svp));	604	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
382	return;	605	return;
383	}	606	}
384	else	607	else
385	{	608	{
386	sv_setuv (*svp, enc->shareable_idx);	609	sv_setuv (*svp, enc->shareable_idx);
…		…
392	if (ecb_expect_false (SvOBJECT (sv)))	615	if (ecb_expect_false (SvOBJECT (sv)))
393	{	616	{
394	HV *stash = SvSTASH (sv);	617	HV *stash = SvSTASH (sv);
395	GV *method;	618	GV *method;
396		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))));
397	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	623	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
398	{	624	{
399	dSP;	625	dSP;
400		626
401	ENTER; SAVETMPS; PUSHMARK (SP);	627	ENTER; SAVETMPS;
		628	PUSHMARK (SP);
402	// 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
403	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	630	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
404		631
405	PUTBACK;	632	PUTBACK;
406	// G_SCALAR ensures that return value is 1	633	// G_SCALAR ensures that return value is 1
…		…
419	}	646	}
420	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	647	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
421	{	648	{
422	dSP;	649	dSP;
423		650
424	ENTER; SAVETMPS; PUSHMARK (SP);	651	ENTER; SAVETMPS;
		652	PUSHMARK (SP);
425	EXTEND (SP, 2);	653	EXTEND (SP, 2);
426	// 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
427	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	655	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
428	PUSHs (sv_cbor);	656	PUSHs (sv_cbor);
429		657
…		…
434	// catch this surprisingly common error	662	// catch this surprisingly common error
435	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	663	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
436	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	664	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
437		665
438	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	666	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
439	encode_uint (enc, 0x80, count + 1);	667	encode_uint (enc, MAJOR_ARRAY, count + 1);
440	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	668	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
441		669
442	while (count)	670	{
		671	int i;
		672
		673	for (i = 0; i < count; ++i)
443	encode_sv (enc, SP[1 - count--]);	674	encode_sv (enc, SP[i + 1 - count]);
		675
		676	SP -= count;
		677	}
444		678
445	PUTBACK;	679	PUTBACK;
446		680
447	FREETMPS; LEAVE;	681	FREETMPS; LEAVE;
448	}	682	}
…		…
466	{	700	{
467	double nv = SvNVX (sv);	701	double nv = SvNVX (sv);
468		702
469	need (enc, 9);	703	need (enc, 9);
470		704
471	if (ecb_expect_false (nv == (U32)nv))	705	if (ecb_expect_false (nv == (NV)(U32)nv))
472	encode_uint (enc, 0x00, (U32)nv);	706	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
473	//TODO: maybe I32?	707	//TODO: maybe I32?
474	else if (ecb_expect_false (nv == (float)nv))	708	else if (ecb_expect_false (nv == (float)nv))
475	{	709	encode_float32 (enc, nv);
476	uint32_t fp = ecb_float_to_binary32 (nv);
477
478	*enc->cur++ = 0xe0 | 26;
479
480	if (!ecb_big_endian ())
481	fp = ecb_bswap32 (fp);
482
483	memcpy (enc->cur, &fp, 4);
484	enc->cur += 4;
485	}
486	else	710	else
487	{	711	encode_float64 (enc, nv);
488	uint64_t fp = ecb_double_to_binary64 (nv);
489
490	*enc->cur++ = 0xe0 | 27;
491
492	if (!ecb_big_endian ())
493	fp = ecb_bswap64 (fp);
494
495	memcpy (enc->cur, &fp, 8);
496	enc->cur += 8;
497	}
498	}	712	}
499		713
500	static void	714	static void
501	encode_sv (enc_t *enc, SV *sv)	715	encode_sv (enc_t *enc, SV *sv)
502	{	716	{
…		…
504		718
505	if (SvPOKp (sv))	719	if (SvPOKp (sv))
506	{	720	{
507	STRLEN len;	721	STRLEN len;
508	char *str = SvPV (sv, len);	722	char *str = SvPV (sv, len);
509	encode_strref (enc, SvUTF8 (sv), str, len);	723	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
510	}	724	}
511	else if (SvNOKp (sv))	725	else if (SvNOKp (sv))
512	encode_nv (enc, sv);	726	encode_nv (enc, sv);
513	else if (SvIOKp (sv))	727	else if (SvIOKp (sv))
514	{	728	encode_int (enc, sv);
515	if (SvIsUV (sv))
516	encode_uint (enc, 0x00, SvUVX (sv));
517	else if (SvIVX (sv) >= 0)
518	encode_uint (enc, 0x00, SvIVX (sv));
519	else
520	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
521	}
522	else if (SvROK (sv))	729	else if (SvROK (sv))
523	encode_rv (enc, SvRV (sv));	730	encode_rv (enc, SvRV (sv));
524	else if (!SvOK (sv))	731	else if (!SvOK (sv))
525	encode_ch (enc, 0xe0 | 22);	732	encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
526	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	733	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
527	encode_ch (enc, 0xe0 | 23);	734	encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
528	else	735	else
529	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",
530	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	737	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
531	}	738	}
532		739
533	static SV *	740	static SV *
534	encode_cbor (SV *scalar, CBOR *cbor)	741	encode_cbor (SV *scalar, CBOR *cbor)
535	{	742	{
536	enc_t enc = { };	743	enc_t enc = { 0 };
537		744
538	enc.cbor = *cbor;	745	enc.cbor = *cbor;
539	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	746	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
540	enc.cur = SvPVX (enc.sv);	747	enc.cur = SvPVX (enc.sv);
541	enc.end = SvEND (enc.sv);	748	enc.end = SvEND (enc.sv);
542		749
543	SvPOK_only (enc.sv);	750	SvPOK_only (enc.sv);
544		751
545	if (cbor->flags & F_PACK_STRINGS)	752	if (cbor->flags & F_PACK_STRINGS)
546	{	753	{
…		…
573	U32 depth; // recursion depth	780	U32 depth; // recursion depth
574	U32 maxdepth; // recursion depth limit	781	U32 maxdepth; // recursion depth limit
575	AV *shareable;	782	AV *shareable;
576	AV *stringref;	783	AV *stringref;
577	SV *decode_tagged;	784	SV *decode_tagged;
		785	SV *err_sv; // optional sv for error, needs to be freed
578	} dec_t;	786	} dec_t;
579		787
580	#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);
581		795
582	#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;
583		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
584	#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)
585	#define DEC_DEC_DEPTH --dec->depth	825	#define DEC_DEC_DEPTH --dec->depth
586		826
587	static UV	827	static UV
588	decode_uint (dec_t *dec)	828	decode_uint (dec_t *dec)
589	{	829	{
590	switch (*dec->cur & 31)	830	U8 m = *dec->cur & MINOR_MASK;
591	{	831	++dec->cur;
592	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
593	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
594	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
595	return *dec->cur++ & 31;
596		832
597	case 24:	833	if (ecb_expect_true (m < LENGTH_EXT1))
		834	return m;
		835	else if (ecb_expect_true (m == LENGTH_EXT1))
		836	{
598	WANT (2);	837	WANT (1);
599	dec->cur += 2;	838	dec->cur += 1;
600	return dec->cur[-1];	839	return dec->cur[-1];
601		840	}
602	case 25:	841	else if (ecb_expect_true (m == LENGTH_EXT2))
		842	{
603	WANT (3);	843	WANT (2);
604	dec->cur += 3;	844	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	845	return (((UV)dec->cur[-2]) << 8)
606	| ((UV)dec->cur[-1]);	846	| ((UV)dec->cur[-1]);
607		847	}
608	case 26:	848	else if (ecb_expect_true (m == LENGTH_EXT4))
		849	{
609	WANT (5);	850	WANT (4);
610	dec->cur += 5;	851	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	852	return (((UV)dec->cur[-4]) << 24)
612	| (((UV)dec->cur[-3]) << 16)	853	| (((UV)dec->cur[-3]) << 16)
613	| (((UV)dec->cur[-2]) << 8)	854	| (((UV)dec->cur[-2]) << 8)
614	| ((UV)dec->cur[-1]);	855	| ((UV)dec->cur[-1]);
615		856	}
616	case 27:	857	else if (ecb_expect_true (m == LENGTH_EXT8))
		858	{
617	WANT (9);	859	WANT (8);
618	dec->cur += 9;	860	dec->cur += 8;
619		861
620	return	862	return
621	#if UVSIZE < 8	863	#if UVSIZE < 8
622	0	864	0
623	#else	865	#else
624	(((UV)dec->cur[-8]) << 56)	866	(((UV)dec->cur[-8]) << 56)
625	| (((UV)dec->cur[-7]) << 48)	867	| (((UV)dec->cur[-7]) << 48)
626	| (((UV)dec->cur[-6]) << 40)	868	| (((UV)dec->cur[-6]) << 40)
627	| (((UV)dec->cur[-5]) << 32)	869	| (((UV)dec->cur[-5]) << 32)
628	#endif	870	#endif
629	| (((UV)dec->cur[-4]) << 24)	871	| (((UV)dec->cur[-4]) << 24)
630	| (((UV)dec->cur[-3]) << 16)	872	| (((UV)dec->cur[-3]) << 16)
631	| (((UV)dec->cur[-2]) << 8)	873	| (((UV)dec->cur[-2]) << 8)
632	| ((UV)dec->cur[-1]);	874	| ((UV)dec->cur[-1]);
633		875	}
634	default:	876	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	877	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		878
638	fail:	879	fail:
639	return 0;	880	return 0;
640	}	881	}
641		882
…		…
646	{	887	{
647	AV *av = newAV ();	888	AV *av = newAV ();
648		889
649	DEC_INC_DEPTH;	890	DEC_INC_DEPTH;
650		891
651	if ((*dec->cur & 31) == 31)	892	if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
652	{	893	{
653	++dec->cur;	894	++dec->cur;
654		895
655	for (;;)	896	for (;;)
656	{	897	{
657	WANT (1);	898	WANT (1);
658		899
659	if (*dec->cur == (0xe0 | 31))	900	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
660	{	901	{
661	++dec->cur;	902	++dec->cur;
662	break;	903	break;
663	}	904	}
664		905
665	av_push (av, decode_sv (dec));	906	av_push (av, decode_sv (dec));
666	}	907	}
667	}	908	}
668	else	909	else
669	{	910	{
670	int i, len = decode_uint (dec);	911	UV i, len = decode_uint (dec);
671		912
		913	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
672	av_fill (av, len - 1);	914	av_fill (av, len - 1);
673		915
674	for (i = 0; i < len; ++i)	916	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	917	AvARRAY (av)[i] = decode_sv (dec);
676	}	918	}
677		919
678	DEC_DEC_DEPTH;	920	DEC_DEC_DEPTH;
679	return newRV_noinc ((SV *)av);	921	return newRV_noinc ((SV *)av);
680		922
681	fail:	923	fail:
682	SvREFCNT_dec (av);	924	SvREFCNT_dec_NN (av);
683	DEC_DEC_DEPTH;	925	DEC_DEC_DEPTH;
684	return &PL_sv_undef;	926	return &PL_sv_undef;
685	}	927	}
686		928
687	static void	929	static void
…		…
689	{	931	{
690	// for speed reasons, we specialcase single-string	932	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	933	// byte or utf-8 strings as keys, but only when !stringref
692		934
693	if (ecb_expect_true (!dec->stringref))	935	if (ecb_expect_true (!dec->stringref))
694	if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)	936	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	937	{
696	I32 len = decode_uint (dec);	938	STRLEN len = decode_uint (dec);
697	char *key = (char *)dec->cur;	939	char *key = (char *)dec->cur;
698		940
		941	WANT (len);
699	dec->cur += len;	942	dec->cur += len;
700		943
701	if (ecb_expect_false (dec->stringref))
702	av_push (dec->stringref, newSVpvn (key, len));
703
704	hv_store (hv, key, len, decode_sv (dec), 0);	944	hv_store (hv, key, len, decode_sv (dec), 0);
705		945
706	return;	946	return;
707	}	947	}
708	else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)	948	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	949	{
710	I32 len = decode_uint (dec);	950	STRLEN len = decode_uint (dec);
711	char *key = (char *)dec->cur;	951	char *key = (char *)dec->cur;
712		952
		953	WANT (len);
713	dec->cur += len;	954	dec->cur += len;
714		955
715	if (ecb_expect_false (dec->stringref))	956	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	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)");
717		959
718	hv_store (hv, key, -len, decode_sv (dec), 0);	960	hv_store (hv, key, -len, decode_sv (dec), 0);
719		961
720	return;	962	return;
721	}	963	}
722		964
723	SV *k = decode_sv (dec);	965	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	966	SV *v = decode_sv (dec);
725		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
726	hv_store_ent (hv, k, v, 0);	996	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	997	SvREFCNT_dec_NN (k);
		998
		999	fail:
		1000	;
728	}	1001	}
729		1002
730	static SV *	1003	static SV *
731	decode_hv (dec_t *dec)	1004	decode_hv (dec_t *dec)
732	{	1005	{
733	HV *hv = newHV ();	1006	HV *hv = newHV ();
734		1007
735	DEC_INC_DEPTH;	1008	DEC_INC_DEPTH;
736		1009
737	if ((*dec->cur & 31) == 31)	1010	if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
738	{	1011	{
739	++dec->cur;	1012	++dec->cur;
740		1013
741	for (;;)	1014	for (;;)
742	{	1015	{
743	WANT (1);	1016	WANT (1);
744		1017
745	if (*dec->cur == (0xe0 | 31))	1018	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
746	{	1019	{
747	++dec->cur;	1020	++dec->cur;
748	break;	1021	break;
749	}	1022	}
750		1023
751	decode_he (dec, hv);	1024	decode_he (dec, hv);
752	}	1025	}
753	}	1026	}
754	else	1027	else
755	{	1028	{
756	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
757		1032
758	while (pairs--)	1033	while (pairs--)
759	decode_he (dec, hv);	1034	decode_he (dec, hv);
760	}	1035	}
761		1036
762	DEC_DEC_DEPTH;	1037	DEC_DEC_DEPTH;
763	return newRV_noinc ((SV *)hv);	1038	return newRV_noinc ((SV *)hv);
764		1039
765	fail:	1040	fail:
766	SvREFCNT_dec (hv);	1041	SvREFCNT_dec_NN (hv);
767	DEC_DEC_DEPTH;	1042	DEC_DEC_DEPTH;
768	return &PL_sv_undef;	1043	return &PL_sv_undef;
769	}	1044	}
770		1045
771	static SV *	1046	static SV *
772	decode_str (dec_t *dec, int utf8)	1047	decode_str (dec_t *dec, int utf8)
773	{	1048	{
774	SV *sv = 0;	1049	SV *sv = 0;
775		1050
776	if ((*dec->cur & 31) == 31)	1051	if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
777	{	1052	{
778	// indefinite length strings	1053	// indefinite length strings
779	++dec->cur;	1054	++dec->cur;
780		1055
781	unsigned char major = *dec->cur & 0xe0;	1056	U8 major = *dec->cur & MAJOR_MISC;
782		1057
783	sv = newSVpvn ("", 0);	1058	sv = newSVpvn ("", 0);
784		1059
785	for (;;)	1060	for (;;)
786	{	1061	{
787	WANT (1);	1062	WANT (1);
788		1063
789	if ((*dec->cur ^ major) >= 31)	1064	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 | 31))	1065	if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
791	{	1066	{
792	++dec->cur;	1067	++dec->cur;
793	break;	1068	break;
794	}	1069	}
795	else	1070	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	1089	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	1090	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	1091	}
817		1092
818	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
819	SvUTF8_on (sv);	1099	SvUTF8_on (sv);
		1100	}
820		1101
821	return sv;	1102	return sv;
822		1103
823	fail:	1104	fail:
824	SvREFCNT_dec (sv);	1105	SvREFCNT_dec (sv);
…		…
843	sv = newRV_noinc (decode_sv (dec));	1124	sv = newRV_noinc (decode_sv (dec));
844	break;	1125	break;
845		1126
846	case CBOR_TAG_STRINGREF_NAMESPACE:	1127	case CBOR_TAG_STRINGREF_NAMESPACE:
847	{	1128	{
		1129	// do not use SAVETMPS/FREETMPS, as these will
		1130	// erase mortalised caches, e.g. "shareable"
848	ENTER; SAVETMPS;	1131	ENTER;
849		1132
850	SAVESPTR (dec->stringref);	1133	SAVESPTR (dec->stringref);
851	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());	1134	dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
852		1135
853	sv = decode_sv (dec);	1136	sv = decode_sv (dec);
854		1137
855	FREETMPS; LEAVE;	1138	LEAVE;
856	}	1139	}
857	break;	1140	break;
858		1141
859	case CBOR_TAG_STRINGREF:	1142	case CBOR_TAG_STRINGREF:
860	{	1143	{
861	if ((*dec->cur >> 5) != 0)	1144	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1145	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		1146
864	UV idx = decode_uint (dec);	1147	UV idx = decode_uint (dec);
865		1148
866	if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))	1149	if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
867	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)");
868		1151
869	sv = newSVsv (AvARRAY (dec->stringref)[idx]);	1152	sv = newSVsv (AvARRAY (dec->stringref)[idx]);
870	}	1153	}
871	break;	1154	break;
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	1156	case CBOR_TAG_VALUE_SHAREABLE:
874	{	1157	{
875	if (ecb_expect_false (!dec->shareable))	1158	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());	1159	dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
877		1160
		1161	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1162	{
878	sv = newSV (0);	1163	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1164	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		1165
881	SV *osv = decode_sv (dec);	1166	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	1167	sv_setsv (sv, osv);
883	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	}
884	}	1177	}
885	break;	1178	break;
886		1179
887	case CBOR_TAG_VALUE_SHAREDREF:	1180	case CBOR_TAG_VALUE_SHAREDREF:
888	{	1181	{
889	if ((*dec->cur >> 5) != 0)	1182	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1183	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		1184
892	UV idx = decode_uint (dec);	1185	UV idx = decode_uint (dec);
893		1186
894	if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))	1187	if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1188	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		1189
897	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");
898	}	1194	}
899	break;	1195	break;
900		1196
901	case CBOR_TAG_PERL_OBJECT:	1197	case CBOR_TAG_PERL_OBJECT:
902	{	1198	{
		1199	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1200	goto filter;
		1201
903	sv = decode_sv (dec);	1202	sv = decode_sv (dec);
904		1203
905	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)	1204	if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
906	ERR ("corrupted CBOR data (non-array perl object)");	1205	ERR ("corrupted CBOR data (non-array perl object)");
907		1206
…		…
917	if (!method)	1216	if (!method)
918	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)");
919		1218
920	dSP;	1219	dSP;
921		1220
922	ENTER; SAVETMPS; PUSHMARK (SP);	1221	ENTER; SAVETMPS;
		1222	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1223	EXTEND (SP, len + 1);
924	// 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
925	PUSHs (*av_fetch (av, 0, 1));	1225	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1226	PUSHs (sv_cbor);
927		1227
…		…
935	SPAGAIN;	1235	SPAGAIN;
936		1236
937	if (SvTRUE (ERRSV))	1237	if (SvTRUE (ERRSV))
938	{	1238	{
939	FREETMPS; LEAVE;	1239	FREETMPS; LEAVE;
940	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1240	ERR_ERRSV;
941	}	1241	}
942		1242
943	SvREFCNT_dec (sv);	1243	SvREFCNT_dec_NN (sv);
944	sv = SvREFCNT_inc (POPs);	1244	sv = SvREFCNT_inc (POPs);
945		1245
946	PUTBACK;	1246	PUTBACK;
947		1247
948	FREETMPS; LEAVE;	1248	FREETMPS; LEAVE;
949	}	1249	}
950	break;	1250	break;
951		1251
952	default:	1252	default:
		1253	filter:
953	{	1254	{
		1255	SV *tag_sv = newSVuv (tag);
		1256
954	sv = decode_sv (dec);	1257	sv = decode_sv (dec);
955		1258
956	dSP;	1259	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1260	ENTER; SAVETMPS;
		1261	PUSHMARK (SP);
958	EXTEND (SP, 2);	1262	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1263	PUSHs (tag_sv);
960	PUSHs (sv);	1264	PUSHs (sv);
961		1265
962	PUTBACK;	1266	PUTBACK;
963	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);	1267	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
964	SPAGAIN;	1268	SPAGAIN;
965		1269
966	if (SvTRUE (ERRSV))	1270	if (SvTRUE (ERRSV))
967	{	1271	{
		1272	SvREFCNT_dec_NN (tag_sv);
968	FREETMPS; LEAVE;	1273	FREETMPS; LEAVE;
969	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1274	ERR_ERRSV;
970	}	1275	}
971		1276
972	if (count)	1277	if (count)
973	{	1278	{
		1279	SvREFCNT_dec_NN (tag_sv);
974	SvREFCNT_dec (sv);	1280	SvREFCNT_dec_NN (sv);
975	sv = SvREFCNT_inc (POPs);	1281	sv = SvREFCNT_inc_NN (TOPs);
		1282	SP -= count;
976	}	1283	}
977	else	1284	else
978	{	1285	{
979	AV *av = newAV ();	1286	AV *av = newAV ();
980	av_push (av, newSVuv (tag));	1287	av_push (av, tag_sv);
981	av_push (av, sv);	1288	av_push (av, sv);
982		1289
983	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash	1290	HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
984	? cbor_tagged_stash	1291	? cbor_tagged_stash
985	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1292	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
1003	static SV *	1310	static SV *
1004	decode_sv (dec_t *dec)	1311	decode_sv (dec_t *dec)
1005	{	1312	{
1006	WANT (1);	1313	WANT (1);
1007		1314
1008	switch (*dec->cur >> 5)	1315	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1316	{
1010	case 0: // unsigned int	1317	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1318	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1319	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1320	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1321	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1322	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1323	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1324
1018	case 4: // array	1325	case MAJOR_MISC >> MAJOR_SHIFT:
1019	return decode_av (dec);
1020	case 5: // map
1021	return decode_hv (dec);
1022	case 6: // tag
1023	return decode_tagged (dec);
1024	case 7: // misc
1025	switch (*dec->cur++ & 31)	1326	switch (*dec->cur++ & MINOR_MASK)
1026	{	1327	{
1027	case 20:	1328	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1329	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1330	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1331	#endif
1031	return newSVsv (types_false);	1332	return newSVsv (types_false);
1032	case 21:	1333	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1334	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1335	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1336	#endif
1036	return newSVsv (types_true);	1337	return newSVsv (types_true);
1037	case 22:	1338	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1339	return newSVsv (&PL_sv_undef);
1039	case 23:	1340	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1341	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1342	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1343	#endif
1043	return newSVsv (types_error);	1344	return newSVsv (types_error);
1044		1345
1045	case 25:	1346	case MISC_FLOAT16:
1046	{	1347	{
1047	WANT (2);	1348	WANT (2);
1048		1349
1049	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];	1350	uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1050	dec->cur += 2;	1351	dec->cur += 2;
1051		1352
1052	return newSVnv (ecb_binary16_to_float (fp));	1353	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1354	}
1054		1355
1055	case 26:	1356	case MISC_FLOAT32:
1056	{	1357	{
1057	uint32_t fp;	1358	uint32_t fp;
1058	WANT (4);	1359	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1360	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1361	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1364	fp = ecb_bswap32 (fp);
1064		1365
1065	return newSVnv (ecb_binary32_to_float (fp));	1366	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1367	}
1067		1368
1068	case 27:	1369	case MISC_FLOAT64:
1069	{	1370	{
1070	uint64_t fp;	1371	uint64_t fp;
1071	WANT (8);	1372	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1373	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1374	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1377	fp = ecb_bswap64 (fp);
1077		1378
1078	return newSVnv (ecb_binary64_to_double (fp));	1379	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1380	}
1080		1381
1081	// 0..19 unassigned	1382	// 0..19 unassigned simple
1082	// 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
1083	default:	1386	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1387	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1388	}
1086		1389
1087	break;	1390	break;
1088	}	1391	}
1089		1392
…		…
1092	}	1395	}
1093		1396
1094	static SV *	1397	static SV *
1095	decode_cbor (SV *string, CBOR *cbor, char **offset_return)	1398	decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1096	{	1399	{
1097	dec_t dec = { };	1400	dec_t dec = { 0 };
1098	SV *sv;	1401	SV *sv;
1099	STRLEN len;	1402	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1403	char *data = SvPVbyte (string, len);
1101		1404
1102	if (len > cbor->max_size && cbor->max_size)	1405	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1419	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1420	dec.err = "garbage after CBOR object";
1118		1421
1119	if (dec.err)	1422	if (dec.err)
1120	{	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
1121	SvREFCNT_dec (sv);	1435	SvREFCNT_dec_NN (sv);
		1436
		1437	if (dec.err_sv)
		1438	sv_2mortal (dec.err_sv);
		1439
1122	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);
1123	}	1441	}
1124		1442
1125	sv = sv_2mortal (sv);	1443	sv = sv_2mortal (sv);
1126		1444
1127	return sv;	1445	return sv;
1128	}	1446	}
1129		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
1130	/////////////////////////////////////////////////////////////////////////////	1572	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1573	// XS interface functions
1132		1574
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1575	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1576
…		…
1146		1588
1147	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1589	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1148		1590
1149	sv_cbor = newSVpv ("CBOR", 0);	1591	sv_cbor = newSVpv ("CBOR", 0);
1150	SvREADONLY_on (sv_cbor);	1592	SvREADONLY_on (sv_cbor);
		1593
		1594	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1151	}	1595	}
1152		1596
1153	PROTOTYPES: DISABLE	1597	PROTOTYPES: DISABLE
1154		1598
1155	void CLONE (...)	1599	void CLONE (...)
…		…
1174	void shrink (CBOR *self, int enable = 1)	1618	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1619	ALIAS:
1176	shrink = F_SHRINK	1620	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1621	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1622	allow_sharing = F_ALLOW_SHARING
		1623	allow_cycles = F_ALLOW_CYCLES
		1624	forbid_objects = F_FORBID_OBJECTS
1179	pack_strings = F_PACK_STRINGS	1625	pack_strings = F_PACK_STRINGS
		1626	text_keys = F_TEXT_KEYS
		1627	text_strings = F_TEXT_STRINGS
		1628	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1629	PPCODE:
1181	{	1630	{
1182	if (enable)	1631	if (enable)
1183	self->flags |= ix;	1632	self->flags |= ix;
1184	else	1633	else
…		…
1190	void get_shrink (CBOR *self)	1639	void get_shrink (CBOR *self)
1191	ALIAS:	1640	ALIAS:
1192	get_shrink = F_SHRINK	1641	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1642	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1643	get_allow_sharing = F_ALLOW_SHARING
		1644	get_allow_cycles = F_ALLOW_CYCLES
		1645	get_forbid_objects = F_FORBID_OBJECTS
1195	get_pack_strings = F_PACK_STRINGS	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
1196	PPCODE:	1650	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1651	XPUSHs (boolSV (self->flags & ix));
1198		1652
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1653	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1654	PPCODE:
…		…
1249	EXTEND (SP, 2);	1703	EXTEND (SP, 2);
1250	PUSHs (sv);	1704	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1705	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1706	}
1253		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
1254	void DESTROY (CBOR *self)	1760	void DESTROY (CBOR *self)
1255	PPCODE:	1761	PPCODE:
1256	cbor_free (self);	1762	cbor_free (self);
1257		1763
1258	PROTOTYPES: ENABLE	1764	PROTOTYPES: ENABLE
1259		1765
1260	void encode_cbor (SV *scalar)	1766	void encode_cbor (SV *scalar)
		1767	ALIAS:
		1768	encode_cbor = 0
		1769	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1770	PPCODE:
1262	{	1771	{
1263	CBOR cbor;	1772	CBOR cbor;
1264	cbor_init (&cbor);	1773	cbor_init (&cbor);
		1774	cbor.flags |= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1775	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1776	XPUSHs (scalar);
1267	}	1777	}
1268		1778
1269	void decode_cbor (SV *cborstr)	1779	void decode_cbor (SV *cborstr)
…		…
1273	cbor_init (&cbor);	1783	cbor_init (&cbor);
1274	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;	1784	PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1275	XPUSHs (cborstr);	1785	XPUSHs (cborstr);
1276	}	1786	}
1277		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