[ViewVC] Diff of: cvs/CBOR-XS/XS.xs

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs.
Revision 1.57 by root, Fri Nov 25 12:16:12 2016 UTC

…		…
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		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	20	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	21	#endif
21	#ifndef HvNAMEUTF8	22	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	23	# define HvNAMEUTF8(hv) 0
23	#endif	24	#endif
		25	#ifndef SvREFCNT_dec_NN
		26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		27	#endif
		28
		29	// known major and minor types
		30	enum cbor_type
		31	{
		32	MAJOR_SHIFT = 5,
		33	MINOR_MASK = 0x1f,
		34
		35	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		36	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		37	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		38	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		39	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		40	MAJOR_MAP = 5 << MAJOR_SHIFT,
		41	MAJOR_TAG = 6 << MAJOR_SHIFT,
		42	MAJOR_MISC = 7 << MAJOR_SHIFT,
		43
		44	// INT/STRING/ARRAY/MAP subtypes
		45	LENGTH_EXT1 = 24,
		46	LENGTH_EXT2 = 25,
		47	LENGTH_EXT4 = 26,
		48	LENGTH_EXT8 = 27,
		49
		50	// SIMPLE types (effectively MISC subtypes)
		51	SIMPLE_FALSE = 20,
		52	SIMPLE_TRUE = 21,
		53	SIMPLE_NULL = 22,
		54	SIMPLE_UNDEF = 23,
		55
		56	// MISC subtype (unused)
		57	MISC_EXT1 = 24,
		58	MISC_FLOAT16 = 25,
		59	MISC_FLOAT32 = 26,
		60	MISC_FLOAT64 = 27,
		61
		62	// BYTES/TEXT/ARRAY/MAP
		63	MINOR_INDEF = 31,
		64	};
24		65
25	// known tags	66	// known tags
26	enum cbor_tag	67	enum cbor_tag
27	{	68	{
28	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	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	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		77
37	// rfc7049	78	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
44		85
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		90
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		96
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
57	};	98	};
58		99
59	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
62	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	103	#define F_ALLOW_CYCLES 0x00000008UL
		104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_TEXT_KEYS 0x00000020UL
		106	#define F_TEXT_STRINGS 0x00000040UL
		107	#define F_VALIDATE_UTF8 0x00000080UL
63		108
64	#define INIT_SIZE 32 // initial scalar size to be allocated	109	#define INIT_SIZE 32 // initial scalar size to be allocated
65		110
66	#define SB do {	111	#define SB do {
67	#define SE } while (0)	112	#define SE } while (0)
…		…
79	# define CBOR_SLOW 0	124	# define CBOR_SLOW 0
80	# define CBOR_STASH cbor_stash	125	# define CBOR_STASH cbor_stash
81	#endif	126	#endif
82		127
83	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::	128	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;	129	static SV types_true, types_false, types_error, sv_cbor, *default_filter;
85		130
86	typedef struct {	131	typedef struct {
87	U32 flags;	132	U32 flags;
88	U32 max_depth;	133	U32 max_depth;
89	STRLEN max_size;	134	STRLEN max_size;
		135	SV *filter;
		136
		137	// for the incremental parser
		138	STRLEN incr_pos; // the current offset into the text
		139	STRLEN incr_need; // minimum bytes needed to decode
		140	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90	} CBOR;	141	} CBOR;
91		142
92	ecb_inline void	143	ecb_inline void
93	cbor_init (CBOR *cbor)	144	cbor_init (CBOR *cbor)
94	{	145	{
95	Zero (cbor, 1, CBOR);	146	Zero (cbor, 1, CBOR);
96	cbor->max_depth = 512;	147	cbor->max_depth = 512;
		148	}
		149
		150	ecb_inline void
		151	cbor_free (CBOR *cbor)
		152	{
		153	SvREFCNT_dec (cbor->filter);
		154	SvREFCNT_dec (cbor->incr_count);
97	}	155	}
98		156
99	/////////////////////////////////////////////////////////////////////////////	157	/////////////////////////////////////////////////////////////////////////////
100	// utility functions	158	// utility functions
101		159
…		…
131	{	189	{
132	return idx > 23	190	return idx > 23
133	? idx > 0xffU	191	? idx > 0xffU
134	? idx > 0xffffU	192	? idx > 0xffffU
135	? idx > 0xffffffffU	193	? idx > 0xffffffffU
		194	? 11
136	? 7	195	: 7
137	: 6
138	: 5	196	: 5
139	: 4	197	: 4
140	: 3;	198	: 3;
141	}	199	}
142		200
…		…
179	static void	237	static void
180	encode_uint (enc_t *enc, int major, UV len)	238	encode_uint (enc_t *enc, int major, UV len)
181	{	239	{
182	need (enc, 9);	240	need (enc, 9);
183		241
184	if (len < 24)	242	if (ecb_expect_true (len < LENGTH_EXT1))
185	*enc->cur++ = major \| len;	243	*enc->cur++ = major \| len;
186	else if (len <= 0xff)	244	else if (ecb_expect_true (len <= 0xffU))
187	{	245	{
188	*enc->cur++ = major \| 24;	246	*enc->cur++ = major \| LENGTH_EXT1;
189	*enc->cur++ = len;	247	*enc->cur++ = len;
190	}	248	}
191	else if (len <= 0xffff)	249	else if (len <= 0xffffU)
192	{	250	{
193	*enc->cur++ = major \| 25;	251	*enc->cur++ = major \| LENGTH_EXT2;
194	*enc->cur++ = len >> 8;	252	*enc->cur++ = len >> 8;
195	*enc->cur++ = len;	253	*enc->cur++ = len;
196	}	254	}
197	else if (len <= 0xffffffff)	255	else if (len <= 0xffffffffU)
198	{	256	{
199	*enc->cur++ = major \| 26;	257	*enc->cur++ = major \| LENGTH_EXT4;
200	*enc->cur++ = len >> 24;	258	*enc->cur++ = len >> 24;
201	*enc->cur++ = len >> 16;	259	*enc->cur++ = len >> 16;
202	*enc->cur++ = len >> 8;	260	*enc->cur++ = len >> 8;
203	*enc->cur++ = len;	261	*enc->cur++ = len;
204	}	262	}
205	else	263	else
206	{	264	{
207	*enc->cur++ = major \| 27;	265	*enc->cur++ = major \| LENGTH_EXT8;
208	*enc->cur++ = len >> 56;	266	*enc->cur++ = len >> 56;
209	*enc->cur++ = len >> 48;	267	*enc->cur++ = len >> 48;
210	*enc->cur++ = len >> 40;	268	*enc->cur++ = len >> 40;
211	*enc->cur++ = len >> 32;	269	*enc->cur++ = len >> 32;
212	*enc->cur++ = len >> 24;	270	*enc->cur++ = len >> 24;
…		…
217	}	275	}
218		276
219	ecb_inline void	277	ecb_inline void
220	encode_tag (enc_t *enc, UV tag)	278	encode_tag (enc_t *enc, UV tag)
221	{	279	{
222	encode_uint (enc, 0xc0, tag);	280	encode_uint (enc, MAJOR_TAG, tag);
223	}	281	}
224		282
225	static void	283	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		284	ecb_noinline static void
226	encode_str (enc_t enc, int utf8, char str, STRLEN len)	285	encode_str_utf8 (enc_t enc, int utf8, char str, STRLEN len)
227	{	286	{
		287	STRLEN ulen = len;
		288	U8 p, pend = (U8 *)str + len;
		289
		290	for (p = (U8 *)str; p < pend; ++p)
		291	ulen += *p >> 7; // count set high bits
		292
		293	encode_uint (enc, MAJOR_TEXT, ulen);
		294
		295	need (enc, ulen);
		296	for (p = (U8 *)str; p < pend; ++p)
		297	if (*p < 0x80)
		298	enc->cur++ = p;
		299	else
		300	{
		301	enc->cur++ = 0xc0 + (p >> 6);
		302	enc->cur++ = 0x80 + (p & 63);
		303	}
		304	}
		305
		306	ecb_inline void
		307	encode_str (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
		308	{
		309	if (ecb_expect_false (upgrade_utf8))
		310	if (!utf8)
		311	{
		312	encode_str_utf8 (enc, utf8, str, len);
		313	return;
		314	}
		315
		316	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		317	need (enc, len);
		318	memcpy (enc->cur, str, len);
		319	enc->cur += len;
		320	}
		321
		322	ecb_inline void
		323	encode_strref (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
		324	{
228	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	325	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229	{	326	{
230	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	327	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231		328
232	if (SvOK (*svp))	329	if (SvOK (*svp))
233	{	330	{
234	// already registered, use stringref	331	// already registered, use stringref
235	encode_tag (enc, CBOR_TAG_STRINGREF);	332	encode_tag (enc, CBOR_TAG_STRINGREF);
236	encode_uint (enc, 0x00, SvUV (*svp));	333	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237	return;	334	return;
238	}	335	}
239	else if (len >= minimum_string_length (enc->stringref_idx))	336	else if (len >= minimum_string_length (enc->stringref_idx))
240	{	337	{
241	// register only	338	// register only
242	sv_setuv (*svp, enc->stringref_idx);	339	sv_setuv (*svp, enc->stringref_idx);
243	++enc->stringref_idx;	340	++enc->stringref_idx;
244	}	341	}
245	}	342	}
246		343
247	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	344	encode_str (enc, upgrade_utf8, utf8, str, len);
248	need (enc, len);
249	memcpy (enc->cur, str, len);
250	enc->cur += len;
251	}	345	}
252		346
253	static void encode_sv (enc_t enc, SV sv);	347	static void encode_sv (enc_t enc, SV sv);
254		348
255	static void	349	static void
…		…
260	if (enc->depth >= enc->cbor.max_depth)	354	if (enc->depth >= enc->cbor.max_depth)
261	croak (ERR_NESTING_EXCEEDED);	355	croak (ERR_NESTING_EXCEEDED);
262		356
263	++enc->depth;	357	++enc->depth;
264		358
265	encode_uint (enc, 0x80, len + 1);	359	encode_uint (enc, MAJOR_ARRAY, len + 1);
266		360
		361	if (SvMAGICAL (av))
267	for (i = 0; i <= len; ++i)	362	for (i = 0; i <= len; ++i)
268	{	363	{
269	SV **svp = av_fetch (av, i, 0);	364	SV **svp = av_fetch (av, i, 0);
270	encode_sv (enc, svp ? *svp : &PL_sv_undef);	365	encode_sv (enc, svp ? *svp : &PL_sv_undef);
271	}	366	}
		367	else
		368	for (i = 0; i <= len; ++i)
		369	{
		370	SV *sv = AvARRAY (av)[i];
		371	encode_sv (enc, sv ? sv : &PL_sv_undef);
		372	}
272		373
273	--enc->depth;	374	--enc->depth;
274	}
275
276	ecb_inline void
277	encode_he (enc_t enc, HE he)
278	{
279	}	375	}
280		376
281	static void	377	static void
282	encode_hv (enc_t enc, HV hv)	378	encode_hv (enc_t enc, HV hv)
283	{	379	{
…		…
290		386
291	int pairs = hv_iterinit (hv);	387	int pairs = hv_iterinit (hv);
292	int mg = SvMAGICAL (hv);	388	int mg = SvMAGICAL (hv);
293		389
294	if (mg)	390	if (mg)
295	encode_ch (enc, 0xa0 \| 31);	391	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
296	else	392	else
297	encode_uint (enc, 0xa0, pairs);	393	encode_uint (enc, MAJOR_MAP, pairs);
298		394
299	while ((he = hv_iternext (hv)))	395	while ((he = hv_iternext (hv)))
300	{	396	{
301	if (HeKLEN (he) == HEf_SVKEY)	397	if (HeKLEN (he) == HEf_SVKEY)
302	encode_sv (enc, HeSVKEY (he));	398	encode_sv (enc, HeSVKEY (he));
303	else	399	else
304	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	400	encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS \| F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305		401
306	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	402	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307	}	403	}
308		404
309	if (mg)	405	if (mg)
310	encode_ch (enc, 0xe0 \| 31);	406	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
311		407
312	--enc->depth;	408	--enc->depth;
313	}	409	}
314		410
315	// encode objects, arrays and special \0=false and \1=true values.	411	// encode objects, arrays and special \0=false and \1=true values.
316	static void	412	static void
317	encode_rv (enc_t enc, SV sv)	413	encode_rv (enc_t enc, SV sv)
318	{	414	{
319	SvGETMAGIC (sv);	415	SvGETMAGIC (sv);
320
321	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
322	&& ecb_expect_false (SvREFCNT (sv) > 1))
323	{
324	if (!enc->shareable)
325	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
326
327	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
328
329	if (SvOK (*svp))
330	{
331	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
332	encode_uint (enc, 0x00, SvUV (*svp));
333	return;
334	}
335	else
336	{
337	sv_setuv (*svp, enc->shareable_idx);
338	++enc->shareable_idx;
339	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
340	}
341	}
342		416
343	svtype svt = SvTYPE (sv);	417	svtype svt = SvTYPE (sv);
344		418
345	if (ecb_expect_false (SvOBJECT (sv)))	419	if (ecb_expect_false (SvOBJECT (sv)))
346	{	420	{
…		…
353	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	427	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
354	? cbor_tagged_stash	428	? cbor_tagged_stash
355	: gv_stashpv ("CBOR::XS::Tagged" , 1);	429	: gv_stashpv ("CBOR::XS::Tagged" , 1);
356		430
357	HV *stash = SvSTASH (sv);	431	HV *stash = SvSTASH (sv);
358	GV *method;
359		432
360	if (stash == boolean_stash)	433	if (stash == boolean_stash)
361	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	434	{
		435	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		436	return;
		437	}
362	else if (stash == error_stash)	438	else if (stash == error_stash)
363	encode_ch (enc, 0xe0 \| 23);	439	{
		440	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		441	return;
		442	}
364	else if (stash == tagged_stash)	443	else if (stash == tagged_stash)
365	{	444	{
366	if (svt != SVt_PVAV)	445	if (svt != SVt_PVAV)
367	croak ("encountered CBOR::XS::Tagged object that isn't an array");	446	croak ("encountered CBOR::XS::Tagged object that isn't an array");
368		447
369	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	448	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
370	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	449	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		450
		451	return;
		452	}
		453	}
		454
		455	if (ecb_expect_false (SvREFCNT (sv) > 1)
		456	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		457	{
		458	if (!enc->shareable)
		459	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		460
		461	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		462
		463	if (SvOK (*svp))
371	}	464	{
		465	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		466	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		467	return;
		468	}
		469	else
		470	{
		471	sv_setuv (*svp, enc->shareable_idx);
		472	++enc->shareable_idx;
		473	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		474	}
		475	}
		476
		477	if (ecb_expect_false (SvOBJECT (sv)))
		478	{
		479	HV *stash = SvSTASH (sv);
		480	GV *method;
		481
372	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	482	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373	{	483	{
374	dSP;	484	dSP;
375		485
376	ENTER; SAVETMPS; PUSHMARK (SP);	486	ENTER; SAVETMPS;
		487	PUSHMARK (SP);
377	// we re-bless the reference to get overload and other niceties right	488	// we re-bless the reference to get overload and other niceties right
378	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	489	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379		490
380	PUTBACK;	491	PUTBACK;
381	// G_SCALAR ensures that return value is 1	492	// G_SCALAR ensures that return value is 1
…		…
394	}	505	}
395	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	506	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396	{	507	{
397	dSP;	508	dSP;
398		509
399	ENTER; SAVETMPS; PUSHMARK (SP);	510	ENTER; SAVETMPS;
		511	SAVESTACK_POS ();
		512	PUSHMARK (SP);
400	EXTEND (SP, 2);	513	EXTEND (SP, 2);
401	// we re-bless the reference to get overload and other niceties right	514	// we re-bless the reference to get overload and other niceties right
402	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	515	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403	PUSHs (sv_cbor);	516	PUSHs (sv_cbor);
404		517
…		…
409	// catch this surprisingly common error	522	// catch this surprisingly common error
410	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	523	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
411	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	524	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412		525
413	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	526	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414	encode_uint (enc, 0x80, count + 1);	527	encode_uint (enc, MAJOR_ARRAY, count + 1);
415	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	528	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416		529
417	while (count)	530	while (count)
418	encode_sv (enc, SP[1 - count--]);	531	encode_sv (enc, SP[1 - count--]);
419		532
420	PUTBACK;	533	PUTBACK;
…		…
441	{	554	{
442	double nv = SvNVX (sv);	555	double nv = SvNVX (sv);
443		556
444	need (enc, 9);	557	need (enc, 9);
445		558
446	if (ecb_expect_false (nv == (U32)nv))	559	if (ecb_expect_false (nv == (NV)(U32)nv))
447	encode_uint (enc, 0x00, (U32)nv);	560	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448	//TODO: maybe I32?	561	//TODO: maybe I32?
449	else if (ecb_expect_false (nv == (float)nv))	562	else if (ecb_expect_false (nv == (float)nv))
450	{	563	{
451	uint32_t fp = ecb_float_to_binary32 (nv);	564	uint32_t fp = ecb_float_to_binary32 (nv);
452		565
453	*enc->cur++ = 0xe0 \| 26;	566	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
454		567
455	if (!ecb_big_endian ())	568	if (!ecb_big_endian ())
456	fp = ecb_bswap32 (fp);	569	fp = ecb_bswap32 (fp);
457		570
458	memcpy (enc->cur, &fp, 4);	571	memcpy (enc->cur, &fp, 4);
…		…
460	}	573	}
461	else	574	else
462	{	575	{
463	uint64_t fp = ecb_double_to_binary64 (nv);	576	uint64_t fp = ecb_double_to_binary64 (nv);
464		577
465	*enc->cur++ = 0xe0 \| 27;	578	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
466		579
467	if (!ecb_big_endian ())	580	if (!ecb_big_endian ())
468	fp = ecb_bswap64 (fp);	581	fp = ecb_bswap64 (fp);
469		582
470	memcpy (enc->cur, &fp, 8);	583	memcpy (enc->cur, &fp, 8);
…		…
479		592
480	if (SvPOKp (sv))	593	if (SvPOKp (sv))
481	{	594	{
482	STRLEN len;	595	STRLEN len;
483	char *str = SvPV (sv, len);	596	char *str = SvPV (sv, len);
484	encode_str (enc, SvUTF8 (sv), str, len);	597	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
485	}	598	}
486	else if (SvNOKp (sv))	599	else if (SvNOKp (sv))
487	encode_nv (enc, sv);	600	encode_nv (enc, sv);
488	else if (SvIOKp (sv))	601	else if (SvIOKp (sv))
489	{	602	{
490	if (SvIsUV (sv))	603	if (SvIsUV (sv))
491	encode_uint (enc, 0x00, SvUVX (sv));	604	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
492	else if (SvIVX (sv) >= 0)	605	else if (SvIVX (sv) >= 0)
493	encode_uint (enc, 0x00, SvIVX (sv));	606	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
494	else	607	else
495	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	608	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
496	}	609	}
497	else if (SvROK (sv))	610	else if (SvROK (sv))
498	encode_rv (enc, SvRV (sv));	611	encode_rv (enc, SvRV (sv));
499	else if (!SvOK (sv))	612	else if (!SvOK (sv))
500	encode_ch (enc, 0xe0 \| 22);	613	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
501	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	614	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502	encode_ch (enc, 0xe0 \| 23);	615	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
503	else	616	else
504	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	617	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	618	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506	}	619	}
507		620
508	static SV *	621	static SV *
509	encode_cbor (SV scalar, CBOR cbor)	622	encode_cbor (SV scalar, CBOR cbor)
510	{	623	{
511	enc_t enc = { };	624	enc_t enc = { 0 };
512		625
513	enc.cbor = *cbor;	626	enc.cbor = *cbor;
514	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	627	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515	enc.cur = SvPVX (enc.sv);	628	enc.cur = SvPVX (enc.sv);
516	enc.end = SvEND (enc.sv);	629	enc.end = SvEND (enc.sv);
517		630
518	SvPOK_only (enc.sv);	631	SvPOK_only (enc.sv);
519		632
520	if (cbor->flags & F_ALLOW_STRINGREF)	633	if (cbor->flags & F_PACK_STRINGS)
521	{	634	{
522	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	635	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	636	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
524	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	637	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
525	}	638	}
…		…
547	CBOR cbor;	660	CBOR cbor;
548	U32 depth; // recursion depth	661	U32 depth; // recursion depth
549	U32 maxdepth; // recursion depth limit	662	U32 maxdepth; // recursion depth limit
550	AV *shareable;	663	AV *shareable;
551	AV *stringref;	664	AV *stringref;
		665	SV *decode_tagged;
552	} dec_t;	666	} dec_t;
553		667
554	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	668	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
555		669
556	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	670	#define WANT(len) if (ecb_expect_false ((UV)(dec->end - dec->cur) < (UV)len)) ERR ("unexpected end of CBOR data")
557		671
558	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	672	#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
559	#define DEC_DEC_DEPTH --dec->depth	673	#define DEC_DEC_DEPTH --dec->depth
560		674
561	static UV	675	static UV
562	decode_uint (dec_t *dec)	676	decode_uint (dec_t *dec)
563	{	677	{
564	switch (*dec->cur & 31)	678	U8 m = *dec->cur & MINOR_MASK;
565	{	679	++dec->cur;
566	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
567	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
568	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
569	return *dec->cur++ & 31;
570		680
571	case 24:	681	if (ecb_expect_true (m < LENGTH_EXT1))
		682	return m;
		683	else if (ecb_expect_true (m == LENGTH_EXT1))
		684	{
572	WANT (2);	685	WANT (1);
573	dec->cur += 2;	686	dec->cur += 1;
574	return dec->cur[-1];	687	return dec->cur[-1];
575		688	}
576	case 25:	689	else if (ecb_expect_true (m == LENGTH_EXT2))
		690	{
577	WANT (3);	691	WANT (2);
578	dec->cur += 3;	692	dec->cur += 2;
579	return (((UV)dec->cur[-2]) << 8)	693	return (((UV)dec->cur[-2]) << 8)
580	\| ((UV)dec->cur[-1]);	694	\| ((UV)dec->cur[-1]);
581		695	}
582	case 26:	696	else if (ecb_expect_true (m == LENGTH_EXT4))
		697	{
583	WANT (5);	698	WANT (4);
584	dec->cur += 5;	699	dec->cur += 4;
585	return (((UV)dec->cur[-4]) << 24)	700	return (((UV)dec->cur[-4]) << 24)
586	\| (((UV)dec->cur[-3]) << 16)	701	\| (((UV)dec->cur[-3]) << 16)
587	\| (((UV)dec->cur[-2]) << 8)	702	\| (((UV)dec->cur[-2]) << 8)
588	\| ((UV)dec->cur[-1]);	703	\| ((UV)dec->cur[-1]);
589		704	}
590	case 27:	705	else if (ecb_expect_true (m == LENGTH_EXT8))
		706	{
591	WANT (9);	707	WANT (8);
592	dec->cur += 9;	708	dec->cur += 8;
		709
		710	return
		711	#if UVSIZE < 8
		712	0
		713	#else
593	return (((UV)dec->cur[-8]) << 56)	714	(((UV)dec->cur[-8]) << 56)
594	\| (((UV)dec->cur[-7]) << 48)	715	\| (((UV)dec->cur[-7]) << 48)
595	\| (((UV)dec->cur[-6]) << 40)	716	\| (((UV)dec->cur[-6]) << 40)
596	\| (((UV)dec->cur[-5]) << 32)	717	\| (((UV)dec->cur[-5]) << 32)
		718	#endif
597	\| (((UV)dec->cur[-4]) << 24)	719	\| (((UV)dec->cur[-4]) << 24)
598	\| (((UV)dec->cur[-3]) << 16)	720	\| (((UV)dec->cur[-3]) << 16)
599	\| (((UV)dec->cur[-2]) << 8)	721	\| (((UV)dec->cur[-2]) << 8)
600	\| ((UV)dec->cur[-1]);	722	\| ((UV)dec->cur[-1]);
601		723	}
602	default:	724	else
603	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	725	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604	}
605		726
606	fail:	727	fail:
607	return 0;	728	return 0;
608	}	729	}
609		730
…		…
614	{	735	{
615	AV *av = newAV ();	736	AV *av = newAV ();
616		737
617	DEC_INC_DEPTH;	738	DEC_INC_DEPTH;
618		739
619	if ((*dec->cur & 31) == 31)	740	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
620	{	741	{
621	++dec->cur;	742	++dec->cur;
622		743
623	for (;;)	744	for (;;)
624	{	745	{
625	WANT (1);	746	WANT (1);
626		747
627	if (*dec->cur == (0xe0 \| 31))	748	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
628	{	749	{
629	++dec->cur;	750	++dec->cur;
630	break;	751	break;
631	}	752	}
632		753
633	av_push (av, decode_sv (dec));	754	av_push (av, decode_sv (dec));
634	}	755	}
635	}	756	}
636	else	757	else
637	{	758	{
638	int i, len = decode_uint (dec);	759	UV i, len = decode_uint (dec);
639		760
		761	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
640	av_fill (av, len - 1);	762	av_fill (av, len - 1);
641		763
642	for (i = 0; i < len; ++i)	764	for (i = 0; i < len; ++i)
643	AvARRAY (av)[i] = decode_sv (dec);	765	AvARRAY (av)[i] = decode_sv (dec);
644	}	766	}
…		…
657	{	779	{
658	// for speed reasons, we specialcase single-string	780	// for speed reasons, we specialcase single-string
659	// byte or utf-8 strings as keys, but only when !stringref	781	// byte or utf-8 strings as keys, but only when !stringref
660		782
661	if (ecb_expect_true (!dec->stringref))	783	if (ecb_expect_true (!dec->stringref))
662	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	784	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663	{	785	{
664	I32 len = decode_uint (dec);	786	STRLEN len = decode_uint (dec);
665	char key = (char )dec->cur;	787	char key = (char )dec->cur;
666		788
		789	WANT (len);
667	dec->cur += len;	790	dec->cur += len;
668		791
669	if (ecb_expect_false (dec->stringref))
670	av_push (dec->stringref, newSVpvn (key, len));
671
672	hv_store (hv, key, len, decode_sv (dec), 0);	792	hv_store (hv, key, len, decode_sv (dec), 0);
673		793
674	return;	794	return;
675	}	795	}
676	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	796	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677	{	797	{
678	I32 len = decode_uint (dec);	798	STRLEN len = decode_uint (dec);
679	char key = (char )dec->cur;	799	char key = (char )dec->cur;
680		800
		801	WANT (len);
681	dec->cur += len;	802	dec->cur += len;
682		803
683	if (ecb_expect_false (dec->stringref))	804	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	805	if (!is_utf8_string (key, len))
		806	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685		807
686	hv_store (hv, key, -len, decode_sv (dec), 0);	808	hv_store (hv, key, -len, decode_sv (dec), 0);
687		809
688	return;	810	return;
689	}	811	}
…		…
691	SV *k = decode_sv (dec);	813	SV *k = decode_sv (dec);
692	SV *v = decode_sv (dec);	814	SV *v = decode_sv (dec);
693		815
694	hv_store_ent (hv, k, v, 0);	816	hv_store_ent (hv, k, v, 0);
695	SvREFCNT_dec (k);	817	SvREFCNT_dec (k);
		818
		819	fail:
		820	;
696	}	821	}
697		822
698	static SV *	823	static SV *
699	decode_hv (dec_t *dec)	824	decode_hv (dec_t *dec)
700	{	825	{
701	HV *hv = newHV ();	826	HV *hv = newHV ();
702		827
703	DEC_INC_DEPTH;	828	DEC_INC_DEPTH;
704		829
705	if ((*dec->cur & 31) == 31)	830	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
706	{	831	{
707	++dec->cur;	832	++dec->cur;
708		833
709	for (;;)	834	for (;;)
710	{	835	{
711	WANT (1);	836	WANT (1);
712		837
713	if (*dec->cur == (0xe0 \| 31))	838	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
714	{	839	{
715	++dec->cur;	840	++dec->cur;
716	break;	841	break;
717	}	842	}
718		843
719	decode_he (dec, hv);	844	decode_he (dec, hv);
720	}	845	}
721	}	846	}
722	else	847	else
723	{	848	{
724	int pairs = decode_uint (dec);	849	UV pairs = decode_uint (dec);
		850
		851	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
725		852
726	while (pairs--)	853	while (pairs--)
727	decode_he (dec, hv);	854	decode_he (dec, hv);
728	}	855	}
729		856
…		…
739	static SV *	866	static SV *
740	decode_str (dec_t *dec, int utf8)	867	decode_str (dec_t *dec, int utf8)
741	{	868	{
742	SV *sv = 0;	869	SV *sv = 0;
743		870
744	if ((*dec->cur & 31) == 31)	871	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
745	{	872	{
		873	// indefinite length strings
746	++dec->cur;	874	++dec->cur;
747		875
		876	U8 major = *dec->cur & MAJOR_MISC;
		877
748	sv = newSVpvn ("", 0);	878	sv = newSVpvn ("", 0);
749		879
750	// not very fast, and certainly not robust against illegal input
751	for (;;)	880	for (;;)
752	{	881	{
753	WANT (1);	882	WANT (1);
754		883
755	if (*dec->cur == (0xe0 \| 31))	884	if ((*dec->cur - major) > LENGTH_EXT8)
		885	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
756	{	886	{
757	++dec->cur;	887	++dec->cur;
758	break;	888	break;
759	}	889	}
		890	else
		891	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760		892
761	sv_catsv (sv, decode_sv (dec));	893	STRLEN len = decode_uint (dec);
		894
		895	WANT (len);
		896	sv_catpvn (sv, dec->cur, len);
		897	dec->cur += len;
762	}	898	}
763	}	899	}
764	else	900	else
765	{	901	{
766	STRLEN len = decode_uint (dec);	902	STRLEN len = decode_uint (dec);
767		903
768	WANT (len);	904	WANT (len);
769	sv = newSVpvn (dec->cur, len);	905	sv = newSVpvn (dec->cur, len);
770	dec->cur += len;	906	dec->cur += len;
		907
		908	if (ecb_expect_false (dec->stringref)
		909	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
		910	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
771	}	911	}
772		912
773	if (utf8)	913	if (utf8)
		914	{
		915	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		916	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		917	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		918
774	SvUTF8_on (sv);	919	SvUTF8_on (sv);
775		920	}
776	if (ecb_expect_false (dec->stringref)
777	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
778	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
779		921
780	return sv;	922	return sv;
781		923
782	fail:	924	fail:
783	SvREFCNT_dec (sv);	925	SvREFCNT_dec (sv);
…		…
802	sv = newRV_noinc (decode_sv (dec));	944	sv = newRV_noinc (decode_sv (dec));
803	break;	945	break;
804		946
805	case CBOR_TAG_STRINGREF_NAMESPACE:	947	case CBOR_TAG_STRINGREF_NAMESPACE:
806	{	948	{
		949	// do nmot use SAVETMPS/FREETMPS, as these will
		950	// erase mortalised caches, e.g. "shareable"
807	ENTER; SAVETMPS;	951	ENTER;
808		952
809	SAVESPTR (dec->stringref);	953	SAVESPTR (dec->stringref);
810	dec->stringref = (AV )sv_2mortal ((SV )newAV ());	954	dec->stringref = (AV )sv_2mortal ((SV )newAV ());
811		955
812	sv = decode_sv (dec);	956	sv = decode_sv (dec);
813		957
814	FREETMPS; LEAVE;	958	LEAVE;
815	}	959	}
816	break;	960	break;
817		961
818	case CBOR_TAG_STRINGREF:	962	case CBOR_TAG_STRINGREF:
819	{	963	{
820	if ((*dec->cur >> 5) != 0)	964	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	965	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822		966
823	UV idx = decode_uint (dec);	967	UV idx = decode_uint (dec);
824		968
825	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	969	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
832	case CBOR_TAG_VALUE_SHAREABLE:	976	case CBOR_TAG_VALUE_SHAREABLE:
833	{	977	{
834	if (ecb_expect_false (!dec->shareable))	978	if (ecb_expect_false (!dec->shareable))
835	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	979	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
836		980
		981	if (dec->cbor.flags & F_ALLOW_CYCLES)
		982	{
837	sv = newSV (0);	983	sv = newSV (0);
838	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	984	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839		985
840	SV *osv = decode_sv (dec);	986	SV *osv = decode_sv (dec);
841	sv_setsv (sv, osv);	987	sv_setsv (sv, osv);
842	SvREFCNT_dec_NN (osv);	988	SvREFCNT_dec_NN (osv);
		989	}
		990	else
		991	{
		992	av_push (dec->shareable, &PL_sv_undef);
		993	int idx = AvFILLp (dec->shareable);
		994	sv = decode_sv (dec);
		995	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		996	}
843	}	997	}
844	break;	998	break;
845		999
846	case CBOR_TAG_VALUE_SHAREDREF:	1000	case CBOR_TAG_VALUE_SHAREDREF:
847	{	1001	{
848	if ((*dec->cur >> 5) != 0)	1002	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1003	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850		1004
851	UV idx = decode_uint (dec);	1005	UV idx = decode_uint (dec);
852		1006
853	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	1007	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
854	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1008	ERR ("corrupted CBOR data (sharedref index out of bounds)");
855		1009
856	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1010	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1011
		1012	if (sv == &PL_sv_undef)
		1013	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857	}	1014	}
858	break;	1015	break;
859		1016
860	case CBOR_TAG_PERL_OBJECT:	1017	case CBOR_TAG_PERL_OBJECT:
861	{	1018	{
…		…
876	if (!method)	1033	if (!method)
877	ERR ("cannot decode perl-object (package does not have a THAW method)");	1034	ERR ("cannot decode perl-object (package does not have a THAW method)");
878		1035
879	dSP;	1036	dSP;
880		1037
881	ENTER; SAVETMPS; PUSHMARK (SP);	1038	ENTER; SAVETMPS;
		1039	PUSHMARK (SP);
882	EXTEND (SP, len + 1);	1040	EXTEND (SP, len + 1);
883	// we re-bless the reference to get overload and other niceties right	1041	// we re-bless the reference to get overload and other niceties right
884	PUSHs (*av_fetch (av, 0, 1));	1042	PUSHs (*av_fetch (av, 0, 1));
885	PUSHs (sv_cbor);	1043	PUSHs (sv_cbor);
886		1044
…		…
910		1068
911	default:	1069	default:
912	{	1070	{
913	sv = decode_sv (dec);	1071	sv = decode_sv (dec);
914		1072
		1073	dSP;
		1074	ENTER; SAVETMPS;
		1075	SAVESTACK_POS ();
		1076	PUSHMARK (SP);
		1077	EXTEND (SP, 2);
		1078	PUSHs (newSVuv (tag));
		1079	PUSHs (sv);
		1080
		1081	PUTBACK;
		1082	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
		1083	SPAGAIN;
		1084
		1085	if (SvTRUE (ERRSV))
		1086	{
		1087	FREETMPS; LEAVE;
		1088	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
		1089	}
		1090
		1091	if (count)
		1092	{
		1093	SvREFCNT_dec (sv);
		1094	sv = SvREFCNT_inc (POPs);
		1095	}
		1096	else
		1097	{
915	AV *av = newAV ();	1098	AV *av = newAV ();
916	av_push (av, newSVuv (tag));	1099	av_push (av, newSVuv (tag));
917	av_push (av, sv);	1100	av_push (av, sv);
918		1101
919	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1102	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
920	? cbor_tagged_stash	1103	? cbor_tagged_stash
921	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1104	: gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1105	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1106	}
		1107
		1108	PUTBACK;
		1109
		1110	FREETMPS; LEAVE;
924	}	1111	}
925	break;	1112	break;
926	}	1113	}
927		1114
928	return sv;	1115	return sv;
…		…
935	static SV *	1122	static SV *
936	decode_sv (dec_t *dec)	1123	decode_sv (dec_t *dec)
937	{	1124	{
938	WANT (1);	1125	WANT (1);
939		1126
940	switch (*dec->cur >> 5)	1127	switch (*dec->cur >> MAJOR_SHIFT)
941	{	1128	{
942	case 0: // unsigned int	1129	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943	return newSVuv (decode_uint (dec));	1130	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944	case 1: // negative int	1131	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945	return newSViv (-1 - (IV)decode_uint (dec));	1132	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946	case 2: // octet string	1133	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947	return decode_str (dec, 0);	1134	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948	case 3: // utf-8 string	1135	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949	return decode_str (dec, 1);	1136
950	case 4: // array	1137	case MAJOR_MISC >> MAJOR_SHIFT:
951	return decode_av (dec);
952	case 5: // map
953	return decode_hv (dec);
954	case 6: // tag
955	return decode_tagged (dec);
956	case 7: // misc
957	switch (*dec->cur++ & 31)	1138	switch (*dec->cur++ & MINOR_MASK)
958	{	1139	{
959	case 20:	1140	case SIMPLE_FALSE:
960	#if CBOR_SLOW	1141	#if CBOR_SLOW
961	types_false = get_bool ("Types::Serialiser::false");	1142	types_false = get_bool ("Types::Serialiser::false");
962	#endif	1143	#endif
963	return newSVsv (types_false);	1144	return newSVsv (types_false);
964	case 21:	1145	case SIMPLE_TRUE:
965	#if CBOR_SLOW	1146	#if CBOR_SLOW
966	types_true = get_bool ("Types::Serialiser::true");	1147	types_true = get_bool ("Types::Serialiser::true");
967	#endif	1148	#endif
968	return newSVsv (types_true);	1149	return newSVsv (types_true);
969	case 22:	1150	case SIMPLE_NULL:
970	return newSVsv (&PL_sv_undef);	1151	return newSVsv (&PL_sv_undef);
971	case 23:	1152	case SIMPLE_UNDEF:
972	#if CBOR_SLOW	1153	#if CBOR_SLOW
973	types_error = get_bool ("Types::Serialiser::error");	1154	types_error = get_bool ("Types::Serialiser::error");
974	#endif	1155	#endif
975	return newSVsv (types_error);	1156	return newSVsv (types_error);
976		1157
977	case 25:	1158	case MISC_FLOAT16:
978	{	1159	{
979	WANT (2);	1160	WANT (2);
980		1161
981	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1162	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
982	dec->cur += 2;	1163	dec->cur += 2;
983		1164
984	return newSVnv (ecb_binary16_to_float (fp));	1165	return newSVnv (ecb_binary16_to_float (fp));
985	}	1166	}
986		1167
987	case 26:	1168	case MISC_FLOAT32:
988	{	1169	{
989	uint32_t fp;	1170	uint32_t fp;
990	WANT (4);	1171	WANT (4);
991	memcpy (&fp, dec->cur, 4);	1172	memcpy (&fp, dec->cur, 4);
992	dec->cur += 4;	1173	dec->cur += 4;
…		…
995	fp = ecb_bswap32 (fp);	1176	fp = ecb_bswap32 (fp);
996		1177
997	return newSVnv (ecb_binary32_to_float (fp));	1178	return newSVnv (ecb_binary32_to_float (fp));
998	}	1179	}
999		1180
1000	case 27:	1181	case MISC_FLOAT64:
1001	{	1182	{
1002	uint64_t fp;	1183	uint64_t fp;
1003	WANT (8);	1184	WANT (8);
1004	memcpy (&fp, dec->cur, 8);	1185	memcpy (&fp, dec->cur, 8);
1005	dec->cur += 8;	1186	dec->cur += 8;
…		…
1008	fp = ecb_bswap64 (fp);	1189	fp = ecb_bswap64 (fp);
1009		1190
1010	return newSVnv (ecb_binary64_to_double (fp));	1191	return newSVnv (ecb_binary64_to_double (fp));
1011	}	1192	}
1012		1193
1013	// 0..19 unassigned	1194	// 0..19 unassigned simple
1014	// 24 reserved + unassigned (reserved values are not encodable)	1195	// 24 reserved + unassigned simple (reserved values are not encodable)
		1196	// 28-30 unassigned misc
		1197	// 31 break code
1015	default:	1198	default:
1016	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1199	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017	}	1200	}
1018		1201
1019	break;	1202	break;
1020	}	1203	}
1021		1204
…		…
1024	}	1207	}
1025		1208
1026	static SV *	1209	static SV *
1027	decode_cbor (SV string, CBOR cbor, char **offset_return)	1210	decode_cbor (SV string, CBOR cbor, char **offset_return)
1028	{	1211	{
1029	dec_t dec = { };	1212	dec_t dec = { 0 };
1030	SV *sv;	1213	SV *sv;
1031	STRLEN len;	1214	STRLEN len;
1032	char *data = SvPVbyte (string, len);	1215	char *data = SvPVbyte (string, len);
1033		1216
1034	if (len > cbor->max_size && cbor->max_size)	1217	if (len > cbor->max_size && cbor->max_size)
…		…
1048	if (dec.cur != dec.end && !dec.err)	1231	if (dec.cur != dec.end && !dec.err)
1049	dec.err = "garbage after CBOR object";	1232	dec.err = "garbage after CBOR object";
1050		1233
1051	if (dec.err)	1234	if (dec.err)
1052	{	1235	{
		1236	if (dec.shareable)
		1237	{
		1238	// need to break cyclic links, which whould all be in shareable
		1239	int i;
		1240	SV **svp;
		1241
		1242	for (i = av_len (dec.shareable) + 1; i--; )
		1243	if ((svp = av_fetch (dec.shareable, i, 0)))
		1244	sv_setsv (*svp, &PL_sv_undef);
		1245	}
		1246
1053	SvREFCNT_dec (sv);	1247	SvREFCNT_dec (sv);
1054	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1248	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1055	}	1249	}
1056		1250
1057	sv = sv_2mortal (sv);	1251	sv = sv_2mortal (sv);
1058		1252
1059	return sv;	1253	return sv;
1060	}	1254	}
1061		1255
		1256	/////////////////////////////////////////////////////////////////////////////
		1257	// incremental parser
		1258
		1259	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1260
		1261	// returns 0 for notyet, 1 for success or error
		1262	static int
		1263	incr_parse (CBOR self, SV cborstr)
		1264	{
		1265	STRLEN cur;
		1266	SvPV (cborstr, cur);
		1267
		1268	while (ecb_expect_true (self->incr_need <= cur))
		1269	{
		1270	// table of integer count bytes
		1271	static I8 incr_len[MINOR_MASK + 1] = {
		1272	0, 0, 0, 0, 0, 0, 0, 0,
		1273	0, 0, 0, 0, 0, 0, 0, 0,
		1274	0, 0, 0, 0, 0, 0, 0, 0,
		1275	1, 2, 4, 8,-1,-1,-1,-2
		1276	};
		1277
		1278	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1279	U8 m = *p & MINOR_MASK;
		1280	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1281	I8 ilen = incr_len[m];
		1282
		1283	self->incr_need = self->incr_pos + 1;
		1284
		1285	if (ecb_expect_false (ilen < 0))
		1286	{
		1287	if (m != MINOR_INDEF)
		1288	return 1; // error
		1289
		1290	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1291	{
		1292	if (count >= 0)
		1293	return 1; // error
		1294
		1295	count = 1;
		1296	}
		1297	else
		1298	{
		1299	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1300	count = -1;
		1301	}
		1302	}
		1303	else
		1304	{
		1305	self->incr_need += ilen;
		1306	if (ecb_expect_false (self->incr_need > cur))
		1307	return 0;
		1308
		1309	int major = *p >> MAJOR_SHIFT;
		1310
		1311	switch (major)
		1312	{
		1313	case MAJOR_TAG >> MAJOR_SHIFT:
		1314	++count; // tags merely prefix another value
		1315	break;
		1316
		1317	case MAJOR_BYTES >> MAJOR_SHIFT:
		1318	case MAJOR_TEXT >> MAJOR_SHIFT:
		1319	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1320	case MAJOR_MAP >> MAJOR_SHIFT:
		1321	{
		1322	UV len;
		1323
		1324	if (ecb_expect_false (ilen))
		1325	{
		1326	len = 0;
		1327
		1328	do {
		1329	len = (len << 8) \| *++p;
		1330	} while (--ilen);
		1331	}
		1332	else
		1333	len = m;
		1334
		1335	switch (major)
		1336	{
		1337	case MAJOR_BYTES >> MAJOR_SHIFT:
		1338	case MAJOR_TEXT >> MAJOR_SHIFT:
		1339	self->incr_need += len;
		1340	if (ecb_expect_false (self->incr_need > cur))
		1341	return 0;
		1342
		1343	break;
		1344
		1345	case MAJOR_MAP >> MAJOR_SHIFT:
		1346	len <<= 1;
		1347	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1348	if (len)
		1349	{
		1350	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1351	count = len + 1;
		1352	}
		1353	break;
		1354	}
		1355	}
		1356	}
		1357	}
		1358
		1359	self->incr_pos = self->incr_need;
		1360
		1361	if (count > 0)
		1362	{
		1363	while (!--count)
		1364	{
		1365	if (!AvFILLp (self->incr_count))
		1366	return 1; // done
		1367
		1368	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1369	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1370	}
		1371
		1372	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1373	}
		1374	}
		1375
		1376	return 0;
		1377	}
		1378
		1379
1062	/////////////////////////////////////////////////////////////////////////////	1380	/////////////////////////////////////////////////////////////////////////////
1063	// XS interface functions	1381	// XS interface functions
1064		1382
1065	MODULE = CBOR::XS PACKAGE = CBOR::XS	1383	MODULE = CBOR::XS PACKAGE = CBOR::XS
1066		1384
…		…
1074		1392
1075	types_true = get_bool ("Types::Serialiser::true" );	1393	types_true = get_bool ("Types::Serialiser::true" );
1076	types_false = get_bool ("Types::Serialiser::false");	1394	types_false = get_bool ("Types::Serialiser::false");
1077	types_error = get_bool ("Types::Serialiser::error");	1395	types_error = get_bool ("Types::Serialiser::error");
1078		1396
		1397	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
		1398
1079	sv_cbor = newSVpv ("CBOR", 0);	1399	sv_cbor = newSVpv ("CBOR", 0);
1080	SvREADONLY_on (sv_cbor);	1400	SvREADONLY_on (sv_cbor);
		1401
		1402	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1081	}	1403	}
1082		1404
1083	PROTOTYPES: DISABLE	1405	PROTOTYPES: DISABLE
1084		1406
1085	void CLONE (...)	1407	void CLONE (...)
…		…
1104	void shrink (CBOR *self, int enable = 1)	1426	void shrink (CBOR *self, int enable = 1)
1105	ALIAS:	1427	ALIAS:
1106	shrink = F_SHRINK	1428	shrink = F_SHRINK
1107	allow_unknown = F_ALLOW_UNKNOWN	1429	allow_unknown = F_ALLOW_UNKNOWN
1108	allow_sharing = F_ALLOW_SHARING	1430	allow_sharing = F_ALLOW_SHARING
1109	allow_stringref = F_ALLOW_STRINGREF	1431	allow_cycles = F_ALLOW_CYCLES
		1432	pack_strings = F_PACK_STRINGS
		1433	text_keys = F_TEXT_KEYS
		1434	text_strings = F_TEXT_STRINGS
		1435	validate_utf8 = F_VALIDATE_UTF8
1110	PPCODE:	1436	PPCODE:
1111	{	1437	{
1112	if (enable)	1438	if (enable)
1113	self->flags \|= ix;	1439	self->flags \|= ix;
1114	else	1440	else
…		…
1120	void get_shrink (CBOR *self)	1446	void get_shrink (CBOR *self)
1121	ALIAS:	1447	ALIAS:
1122	get_shrink = F_SHRINK	1448	get_shrink = F_SHRINK
1123	get_allow_unknown = F_ALLOW_UNKNOWN	1449	get_allow_unknown = F_ALLOW_UNKNOWN
1124	get_allow_sharing = F_ALLOW_SHARING	1450	get_allow_sharing = F_ALLOW_SHARING
1125	get_allow_stringref = F_ALLOW_STRINGREF	1451	get_allow_cycles = F_ALLOW_CYCLES
		1452	get_pack_strings = F_PACK_STRINGS
		1453	get_text_keys = F_TEXT_KEYS
		1454	get_text_strings = F_TEXT_STRINGS
		1455	get_validate_utf8 = F_VALIDATE_UTF8
1126	PPCODE:	1456	PPCODE:
1127	XPUSHs (boolSV (self->flags & ix));	1457	XPUSHs (boolSV (self->flags & ix));
1128		1458
1129	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1459	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130	PPCODE:	1460	PPCODE:
…		…
1143	XPUSHs (ST (0));	1473	XPUSHs (ST (0));
1144		1474
1145	int get_max_size (CBOR *self)	1475	int get_max_size (CBOR *self)
1146	CODE:	1476	CODE:
1147	RETVAL = self->max_size;	1477	RETVAL = self->max_size;
		1478	OUTPUT:
		1479	RETVAL
		1480
		1481	void filter (CBOR self, SV filter = 0)
		1482	PPCODE:
		1483	SvREFCNT_dec (self->filter);
		1484	self->filter = filter ? newSVsv (filter) : filter;
		1485	XPUSHs (ST (0));
		1486
		1487	SV get_filter (CBOR self)
		1488	CODE:
		1489	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148	OUTPUT:	1490	OUTPUT:
1149	RETVAL	1491	RETVAL
1150		1492
1151	void encode (CBOR self, SV scalar)	1493	void encode (CBOR self, SV scalar)
1152	PPCODE:	1494	PPCODE:
…		…
1167	EXTEND (SP, 2);	1509	EXTEND (SP, 2);
1168	PUSHs (sv);	1510	PUSHs (sv);
1169	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1511	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170	}	1512	}
1171		1513
		1514	void incr_parse (CBOR self, SV cborstr)
		1515	ALIAS:
		1516	incr_parse_multiple = 1
		1517	PPCODE:
		1518	{
		1519	if (SvUTF8 (cborstr))
		1520	sv_utf8_downgrade (cborstr, 0);
		1521
		1522	if (!self->incr_count)
		1523	{
		1524	self->incr_count = newAV ();
		1525	self->incr_pos = 0;
		1526	self->incr_need = 1;
		1527
		1528	av_push (self->incr_count, newSViv (1));
		1529	}
		1530
		1531	do
		1532	{
		1533	if (!incr_parse (self, cborstr))
		1534	{
		1535	if (self->incr_need > self->max_size && self->max_size)
		1536	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1537	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1538
		1539	break;
		1540	}
		1541
		1542	SV *sv;
		1543	char *offset;
		1544
		1545	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1546	XPUSHs (sv);
		1547
		1548	sv_chop (cborstr, offset);
		1549
		1550	av_clear (self->incr_count);
		1551	av_push (self->incr_count, newSViv (1));
		1552
		1553	self->incr_pos = 0;
		1554	self->incr_need = self->incr_pos + 1;
		1555	}
		1556	while (ix);
		1557	}
		1558
		1559	void incr_reset (CBOR *self)
		1560	CODE:
		1561	{
		1562	SvREFCNT_dec (self->incr_count);
		1563	self->incr_count = 0;
		1564	}
		1565
		1566	void DESTROY (CBOR *self)
		1567	PPCODE:
		1568	cbor_free (self);
		1569
1172	PROTOTYPES: ENABLE	1570	PROTOTYPES: ENABLE
1173		1571
1174	void encode_cbor (SV *scalar)	1572	void encode_cbor (SV *scalar)
		1573	ALIAS:
		1574	encode_cbor = 0
		1575	encode_cbor_sharing = F_ALLOW_SHARING
1175	PPCODE:	1576	PPCODE:
1176	{	1577	{
1177	CBOR cbor;	1578	CBOR cbor;
1178	cbor_init (&cbor);	1579	cbor_init (&cbor);
		1580	cbor.flags \|= ix;
1179	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1581	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180	XPUSHs (scalar);	1582	XPUSHs (scalar);
1181	}	1583	}
1182		1584
1183	void decode_cbor (SV *cborstr)	1585	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs. Revision 1.57 by root, Fri Nov 25 12:16:12 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs.
Revision 1.57 by root, Fri Nov 25 12:16:12 2016 UTC