[ViewVC] Diff of: 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 1.61 by root, Sat Nov 26 02:10:19 2016 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))
…		…
23	#endif	25	#endif
24	#ifndef SvREFCNT_dec_NN	26	#ifndef SvREFCNT_dec_NN
25	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)	27	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26	#endif	28	#endif
27		29
		30	// known major and minor types
		31	enum cbor_type
		32	{
		33	MAJOR_SHIFT = 5,
		34	MINOR_MASK = 0x1f,
		35
		36	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		37	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		38	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		39	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		40	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		41	MAJOR_MAP = 5 << MAJOR_SHIFT,
		42	MAJOR_TAG = 6 << MAJOR_SHIFT,
		43	MAJOR_MISC = 7 << MAJOR_SHIFT,
		44
		45	// INT/STRING/ARRAY/MAP subtypes
		46	LENGTH_EXT1 = 24,
		47	LENGTH_EXT2 = 25,
		48	LENGTH_EXT4 = 26,
		49	LENGTH_EXT8 = 27,
		50
		51	// SIMPLE types (effectively MISC subtypes)
		52	SIMPLE_FALSE = 20,
		53	SIMPLE_TRUE = 21,
		54	SIMPLE_NULL = 22,
		55	SIMPLE_UNDEF = 23,
		56
		57	// MISC subtype (unused)
		58	MISC_EXT1 = 24,
		59	MISC_FLOAT16 = 25,
		60	MISC_FLOAT32 = 26,
		61	MISC_FLOAT64 = 27,
		62
		63	// BYTES/TEXT/ARRAY/MAP
		64	MINOR_INDEF = 31,
		65	};
		66
28	// known tags	67	// known tags
29	enum cbor_tag	68	enum cbor_tag
30	{	69	{
31	// extensions	70	// extensions
32	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref	71	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
33	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object	72	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
34	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object	73	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
35	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
36	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing	75	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
37	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref	76	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
38	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	77	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
39		78
40	// rfc7049	79	// rfc7049
41	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	80	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
42	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	81	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
43	CBOR_TAG_POS_BIGNUM = 2, // byte string	82	CBOR_TAG_POS_BIGNUM = 2, // byte string
44	CBOR_TAG_NEG_BIGNUM = 3, // byte string	83	CBOR_TAG_NEG_BIGNUM = 3, // byte string
45	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	84	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
46	CBOR_TAG_BIGFLOAT = 5, // array	85	CBOR_TAG_BIGFLOAT = 5, // array
47		86
48	CBOR_TAG_CONV_B64U = 21, // base64url, any	87	CBOR_TAG_CONV_B64U = 21, // base64url, any
49	CBOR_TAG_CONV_B64 = 22, // base64, any	88	CBOR_TAG_CONV_B64 = 22, // base64, any
50	CBOR_TAG_CONV_HEX = 23, // base16, any	89	CBOR_TAG_CONV_HEX = 23, // base16, any
51	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	90	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
52		91
53	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	92	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
54	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	93	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
55	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	94	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
56	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	95	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
57	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	96	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
58		97
59	CBOR_TAG_MAGIC = 55799 // self-describe cbor	98	CBOR_TAG_MAGIC = 55799, // self-describe cbor
60	};	99	};
61		100
62	#define F_SHRINK 0x00000001UL	101	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	102	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL	103	#define F_ALLOW_SHARING 0x00000004UL
		104	#define F_ALLOW_CYCLES 0x00000008UL
		105	#define F_FORBID_OBJECTS 0x00000010UL
65	#define F_PACK_STRINGS 0x00000008UL	106	#define F_PACK_STRINGS 0x00000020UL
		107	#define F_TEXT_KEYS 0x00000040UL
		108	#define F_TEXT_STRINGS 0x00000080UL
		109	#define F_VALIDATE_UTF8 0x00000100UL
66		110
67	#define INIT_SIZE 32 // initial scalar size to be allocated	111	#define INIT_SIZE 32 // initial scalar size to be allocated
68		112
69	#define SB do {	113	#define SB do {
70	#define SE } while (0)	114	#define SE } while (0)
…		…
89	typedef struct {	133	typedef struct {
90	U32 flags;	134	U32 flags;
91	U32 max_depth;	135	U32 max_depth;
92	STRLEN max_size;	136	STRLEN max_size;
93	SV *filter;	137	SV *filter;
		138
		139	// for the incremental parser
		140	STRLEN incr_pos; // the current offset into the text
		141	STRLEN incr_need; // minimum bytes needed to decode
		142	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
94	} CBOR;	143	} CBOR;
95		144
96	ecb_inline void	145	ecb_inline void
97	cbor_init (CBOR *cbor)	146	cbor_init (CBOR *cbor)
98	{	147	{
…		…
102		151
103	ecb_inline void	152	ecb_inline void
104	cbor_free (CBOR *cbor)	153	cbor_free (CBOR *cbor)
105	{	154	{
106	SvREFCNT_dec (cbor->filter);	155	SvREFCNT_dec (cbor->filter);
		156	SvREFCNT_dec (cbor->incr_count);
107	}	157	}
108		158
109	/////////////////////////////////////////////////////////////////////////////	159	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	160	// utility functions
111		161
…		…
137		187
138	// minimum length of a string to be registered for stringref	188	// minimum length of a string to be registered for stringref
139	ecb_inline int	189	ecb_inline int
140	minimum_string_length (UV idx)	190	minimum_string_length (UV idx)
141	{	191	{
142	return idx > 23	192
143	? idx > 0xffU	193	return idx <= 23 ? 3
144	? idx > 0xffffU	194	: idx <= 0xffU ? 4
		195	: idx <= 0xffffU ? 5
145	? idx > 0xffffffffU	196	: idx <= 0xffffffffU ? 7
146	? 11	197	: 11;
147	: 7
148	: 5
149	: 4
150	: 3;
151	}	198	}
152		199
153	/////////////////////////////////////////////////////////////////////////////	200	/////////////////////////////////////////////////////////////////////////////
154	// encoder	201	// encoder
155		202
…		…
168	} enc_t;	215	} enc_t;
169		216
170	ecb_inline void	217	ecb_inline void
171	need (enc_t *enc, STRLEN len)	218	need (enc_t *enc, STRLEN len)
172	{	219	{
173	if (ecb_expect_false (enc->cur + len >= enc->end))	220	if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
174	{	221	{
175	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);	222	STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
176	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);	223	SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
177	enc->cur = SvPVX (enc->sv) + cur;	224	enc->cur = SvPVX (enc->sv) + cur;
178	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;	225	enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
…		…
189	static void	236	static void
190	encode_uint (enc_t *enc, int major, UV len)	237	encode_uint (enc_t *enc, int major, UV len)
191	{	238	{
192	need (enc, 9);	239	need (enc, 9);
193		240
194	if (ecb_expect_true (len < 24))	241	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major \| len;	242	*enc->cur++ = major \| len;
196	else if (ecb_expect_true (len <= 0xff))	243	else if (ecb_expect_true (len <= 0xffU))
197	{	244	{
198	*enc->cur++ = major \| 24;	245	*enc->cur++ = major \| LENGTH_EXT1;
199	*enc->cur++ = len;	246	*enc->cur++ = len;
200	}	247	}
201	else if (len <= 0xffff)	248	else if (len <= 0xffffU)
202	{	249	{
203	*enc->cur++ = major \| 25;	250	*enc->cur++ = major \| LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	251	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	252	*enc->cur++ = len;
206	}	253	}
207	else if (len <= 0xffffffff)	254	else if (len <= 0xffffffffU)
208	{	255	{
209	*enc->cur++ = major \| 26;	256	*enc->cur++ = major \| LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	257	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	258	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	259	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	260	*enc->cur++ = len;
214	}	261	}
215	else	262	else
216	{	263	{
217	*enc->cur++ = major \| 27;	264	*enc->cur++ = major \| LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	265	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	266	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	267	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	268	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	269	*enc->cur++ = len >> 24;
…		…
227	}	274	}
228		275
229	ecb_inline void	276	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	277	encode_tag (enc_t *enc, UV tag)
231	{	278	{
232	encode_uint (enc, 0xc0, tag);	279	encode_uint (enc, MAJOR_TAG, tag);
		280	}
		281
		282	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		283	ecb_noinline static void
		284	encode_str_utf8 (enc_t enc, int utf8, char str, STRLEN len)
		285	{
		286	STRLEN ulen = len;
		287	U8 p, pend = (U8 *)str + len;
		288
		289	for (p = (U8 *)str; p < pend; ++p)
		290	ulen += *p >> 7; // count set high bits
		291
		292	encode_uint (enc, MAJOR_TEXT, ulen);
		293
		294	need (enc, ulen);
		295	for (p = (U8 *)str; p < pend; ++p)
		296	if (*p < 0x80)
		297	enc->cur++ = p;
		298	else
		299	{
		300	enc->cur++ = 0xc0 + (p >> 6);
		301	enc->cur++ = 0x80 + (p & 63);
		302	}
233	}	303	}
234		304
235	ecb_inline void	305	ecb_inline void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	306	encode_str (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
237	{	307	{
238	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	308	if (ecb_expect_false (upgrade_utf8))
		309	if (!utf8)
		310	{
		311	encode_str_utf8 (enc, utf8, str, len);
		312	return;
		313	}
		314
		315	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
239	need (enc, len);	316	need (enc, len);
240	memcpy (enc->cur, str, len);	317	memcpy (enc->cur, str, len);
241	enc->cur += len;	318	enc->cur += len;
242	}	319	}
243		320
244	static void	321	ecb_inline void
245	encode_strref (enc_t enc, int utf8, char str, STRLEN len)	322	encode_strref (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
246	{	323	{
247	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))	324	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
248	{	325	{
249	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	326	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
250		327
251	if (SvOK (*svp))	328	if (SvOK (*svp))
252	{	329	{
253	// already registered, use stringref	330	// already registered, use stringref
254	encode_tag (enc, CBOR_TAG_STRINGREF);	331	encode_tag (enc, CBOR_TAG_STRINGREF);
255	encode_uint (enc, 0x00, SvUV (*svp));	332	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
256	return;	333	return;
257	}	334	}
258	else if (len >= minimum_string_length (enc->stringref_idx))	335	else if (len >= minimum_string_length (enc->stringref_idx))
259	{	336	{
260	// register only	337	// register only
261	sv_setuv (*svp, enc->stringref_idx);	338	sv_setuv (*svp, enc->stringref_idx);
262	++enc->stringref_idx;	339	++enc->stringref_idx;
263	}	340	}
264	}	341	}
265		342
266	encode_str (enc, utf8, str, len);	343	encode_str (enc, upgrade_utf8, utf8, str, len);
267	}	344	}
268		345
269	static void encode_sv (enc_t enc, SV sv);	346	static void encode_sv (enc_t enc, SV sv);
270		347
271	static void	348	static void
…		…
276	if (enc->depth >= enc->cbor.max_depth)	353	if (enc->depth >= enc->cbor.max_depth)
277	croak (ERR_NESTING_EXCEEDED);	354	croak (ERR_NESTING_EXCEEDED);
278		355
279	++enc->depth;	356	++enc->depth;
280		357
281	encode_uint (enc, 0x80, len + 1);	358	encode_uint (enc, MAJOR_ARRAY, len + 1);
282		359
		360	if (SvMAGICAL (av))
283	for (i = 0; i <= len; ++i)	361	for (i = 0; i <= len; ++i)
284	{	362	{
285	SV **svp = av_fetch (av, i, 0);	363	SV **svp = av_fetch (av, i, 0);
286	encode_sv (enc, svp ? *svp : &PL_sv_undef);	364	encode_sv (enc, svp ? *svp : &PL_sv_undef);
287	}	365	}
		366	else
		367	for (i = 0; i <= len; ++i)
		368	{
		369	SV *sv = AvARRAY (av)[i];
		370	encode_sv (enc, sv ? sv : &PL_sv_undef);
		371	}
288		372
289	--enc->depth;	373	--enc->depth;
290	}	374	}
291		375
292	static void	376	static void
…		…
301		385
302	int pairs = hv_iterinit (hv);	386	int pairs = hv_iterinit (hv);
303	int mg = SvMAGICAL (hv);	387	int mg = SvMAGICAL (hv);
304		388
305	if (mg)	389	if (mg)
306	encode_ch (enc, 0xa0 \| 31);	390	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
307	else	391	else
308	encode_uint (enc, 0xa0, pairs);	392	encode_uint (enc, MAJOR_MAP, pairs);
309		393
310	while ((he = hv_iternext (hv)))	394	while ((he = hv_iternext (hv)))
311	{	395	{
312	if (HeKLEN (he) == HEf_SVKEY)	396	if (HeKLEN (he) == HEf_SVKEY)
313	encode_sv (enc, HeSVKEY (he));	397	encode_sv (enc, HeSVKEY (he));
314	else	398	else
315	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	399	encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS \| F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
316		400
317	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	401	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
318	}	402	}
319		403
320	if (mg)	404	if (mg)
321	encode_ch (enc, 0xe0 \| 31);	405	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
322		406
323	--enc->depth;	407	--enc->depth;
324	}	408	}
325		409
326	// encode objects, arrays and special \0=false and \1=true values.	410	// encode objects, arrays and special \0=false and \1=true values.
…		…
345		429
346	HV *stash = SvSTASH (sv);	430	HV *stash = SvSTASH (sv);
347		431
348	if (stash == boolean_stash)	432	if (stash == boolean_stash)
349	{	433	{
350	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	434	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
351	return;	435	return;
352	}	436	}
353	else if (stash == error_stash)	437	else if (stash == error_stash)
354	{	438	{
355	encode_ch (enc, 0xe0 \| 23);	439	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
356	return;	440	return;
357	}	441	}
358	else if (stash == tagged_stash)	442	else if (stash == tagged_stash)
359	{	443	{
360	if (svt != SVt_PVAV)	444	if (svt != SVt_PVAV)
361	croak ("encountered CBOR::XS::Tagged object that isn't an array");	445	croak ("encountered CBOR::XS::Tagged object that isn't an array");
362		446
363	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	447	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
364	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	448	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
365		449
366	return;	450	return;
367	}	451	}
368	}	452	}
…		…
376	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);	460	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
377		461
378	if (SvOK (*svp))	462	if (SvOK (*svp))
379	{	463	{
380	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);	464	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381	encode_uint (enc, 0x00, SvUV (*svp));	465	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
382	return;	466	return;
383	}	467	}
384	else	468	else
385	{	469	{
386	sv_setuv (*svp, enc->shareable_idx);	470	sv_setuv (*svp, enc->shareable_idx);
…		…
392	if (ecb_expect_false (SvOBJECT (sv)))	476	if (ecb_expect_false (SvOBJECT (sv)))
393	{	477	{
394	HV *stash = SvSTASH (sv);	478	HV *stash = SvSTASH (sv);
395	GV *method;	479	GV *method;
396		480
		481	if (enc->cbor.flags & F_FORBID_OBJECTS)
		482	croak ("encountered object '%s', but forbid_objects is enabled",
		483	SvPV_nolen (sv_2mortal (newRV_inc (sv))));
397	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	484	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
398	{	485	{
399	dSP;	486	dSP;
400		487
401	ENTER; SAVETMPS; PUSHMARK (SP);	488	ENTER; SAVETMPS;
		489	PUSHMARK (SP);
402	// we re-bless the reference to get overload and other niceties right	490	// we re-bless the reference to get overload and other niceties right
403	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	491	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
404		492
405	PUTBACK;	493	PUTBACK;
406	// G_SCALAR ensures that return value is 1	494	// G_SCALAR ensures that return value is 1
…		…
419	}	507	}
420	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	508	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
421	{	509	{
422	dSP;	510	dSP;
423		511
424	ENTER; SAVETMPS; PUSHMARK (SP);	512	ENTER; SAVETMPS;
		513	SAVESTACK_POS ();
		514	PUSHMARK (SP);
425	EXTEND (SP, 2);	515	EXTEND (SP, 2);
426	// we re-bless the reference to get overload and other niceties right	516	// we re-bless the reference to get overload and other niceties right
427	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	517	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
428	PUSHs (sv_cbor);	518	PUSHs (sv_cbor);
429		519
…		…
434	// catch this surprisingly common error	524	// catch this surprisingly common error
435	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	525	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));	526	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
437		527
438	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	528	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
439	encode_uint (enc, 0x80, count + 1);	529	encode_uint (enc, MAJOR_ARRAY, count + 1);
440	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	530	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
441		531
442	while (count)	532	while (count)
443	encode_sv (enc, SP[1 - count--]);	533	encode_sv (enc, SP[1 - count--]);
444		534
445	PUTBACK;	535	PUTBACK;
…		…
466	{	556	{
467	double nv = SvNVX (sv);	557	double nv = SvNVX (sv);
468		558
469	need (enc, 9);	559	need (enc, 9);
470		560
471	if (ecb_expect_false (nv == (U32)nv))	561	if (ecb_expect_false (nv == (NV)(U32)nv))
472	encode_uint (enc, 0x00, (U32)nv);	562	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
473	//TODO: maybe I32?	563	//TODO: maybe I32?
474	else if (ecb_expect_false (nv == (float)nv))	564	else if (ecb_expect_false (nv == (float)nv))
475	{	565	{
476	uint32_t fp = ecb_float_to_binary32 (nv);	566	uint32_t fp = ecb_float_to_binary32 (nv);
477		567
478	*enc->cur++ = 0xe0 \| 26;	568	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
479		569
480	if (!ecb_big_endian ())	570	if (!ecb_big_endian ())
481	fp = ecb_bswap32 (fp);	571	fp = ecb_bswap32 (fp);
482		572
483	memcpy (enc->cur, &fp, 4);	573	memcpy (enc->cur, &fp, 4);
…		…
485	}	575	}
486	else	576	else
487	{	577	{
488	uint64_t fp = ecb_double_to_binary64 (nv);	578	uint64_t fp = ecb_double_to_binary64 (nv);
489		579
490	*enc->cur++ = 0xe0 \| 27;	580	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
491		581
492	if (!ecb_big_endian ())	582	if (!ecb_big_endian ())
493	fp = ecb_bswap64 (fp);	583	fp = ecb_bswap64 (fp);
494		584
495	memcpy (enc->cur, &fp, 8);	585	memcpy (enc->cur, &fp, 8);
…		…
504		594
505	if (SvPOKp (sv))	595	if (SvPOKp (sv))
506	{	596	{
507	STRLEN len;	597	STRLEN len;
508	char *str = SvPV (sv, len);	598	char *str = SvPV (sv, len);
509	encode_strref (enc, SvUTF8 (sv), str, len);	599	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
510	}	600	}
511	else if (SvNOKp (sv))	601	else if (SvNOKp (sv))
512	encode_nv (enc, sv);	602	encode_nv (enc, sv);
513	else if (SvIOKp (sv))	603	else if (SvIOKp (sv))
514	{	604	{
515	if (SvIsUV (sv))	605	if (SvIsUV (sv))
516	encode_uint (enc, 0x00, SvUVX (sv));	606	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
517	else if (SvIVX (sv) >= 0)	607	else if (SvIVX (sv) >= 0)
518	encode_uint (enc, 0x00, SvIVX (sv));	608	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
519	else	609	else
520	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	610	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
521	}	611	}
522	else if (SvROK (sv))	612	else if (SvROK (sv))
523	encode_rv (enc, SvRV (sv));	613	encode_rv (enc, SvRV (sv));
524	else if (!SvOK (sv))	614	else if (!SvOK (sv))
525	encode_ch (enc, 0xe0 \| 22);	615	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
526	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	616	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
527	encode_ch (enc, 0xe0 \| 23);	617	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
528	else	618	else
529	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	619	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
530	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	620	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
531	}	621	}
532		622
533	static SV *	623	static SV *
534	encode_cbor (SV scalar, CBOR cbor)	624	encode_cbor (SV scalar, CBOR cbor)
535	{	625	{
536	enc_t enc = { };	626	enc_t enc = { 0 };
537		627
538	enc.cbor = *cbor;	628	enc.cbor = *cbor;
539	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	629	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
540	enc.cur = SvPVX (enc.sv);	630	enc.cur = SvPVX (enc.sv);
541	enc.end = SvEND (enc.sv);	631	enc.end = SvEND (enc.sv);
…		…
573	U32 depth; // recursion depth	663	U32 depth; // recursion depth
574	U32 maxdepth; // recursion depth limit	664	U32 maxdepth; // recursion depth limit
575	AV *shareable;	665	AV *shareable;
576	AV *stringref;	666	AV *stringref;
577	SV *decode_tagged;	667	SV *decode_tagged;
		668	SV *err_sv; // optional sv for error, needs to be freed
578	} dec_t;	669	} dec_t;
579		670
580	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	671	// set dec->err to ERRSV
		672	ecb_cold static void
		673	err_errsv (dec_t *dec)
		674	{
		675	if (!dec->err)
		676	{
		677	dec->err_sv = newSVsv (ERRSV);
581		678
582	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	679	// chop off the trailing \n
		680	SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
		681	*SvEND (dec->err_sv) = 0;
583		682
		683	dec->err = SvPVutf8_nolen (dec->err_sv);
		684	}
		685	}
		686
		687	// the following functions are used to reduce code size and help the compiler to optimise
		688	ecb_cold static void
		689	err_set (dec_t dec, const char reason)
		690	{
		691	if (!dec->err)
		692	dec->err = reason;
		693	}
		694
		695	ecb_cold static void
		696	err_unexpected_end (dec_t *dec)
		697	{
		698	err_set (dec, "unexpected end of CBOR data");
		699	}
		700
		701	ecb_cold static void
		702	err_nesting_exceeded (dec_t *dec)
		703	{
		704	err_set (dec, ERR_NESTING_EXCEEDED);
		705	}
		706
		707	#define ERR_DO(do) SB do; goto fail; SE
		708	#define ERR(reason) ERR_DO (err_set (dec, reason))
		709	#define ERR_ERRSV ERR_DO (err_errsv (dec))
		710
		711	#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
		712
584	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	713	#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	714	#define DEC_DEC_DEPTH --dec->depth
586		715
587	static UV	716	static UV
588	decode_uint (dec_t *dec)	717	decode_uint (dec_t *dec)
589	{	718	{
590	switch (*dec->cur & 31)	719	U8 m = *dec->cur & MINOR_MASK;
591	{	720	++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		721
597	case 24:	722	if (ecb_expect_true (m < LENGTH_EXT1))
		723	return m;
		724	else if (ecb_expect_true (m == LENGTH_EXT1))
		725	{
598	WANT (2);	726	WANT (1);
599	dec->cur += 2;	727	dec->cur += 1;
600	return dec->cur[-1];	728	return dec->cur[-1];
601		729	}
602	case 25:	730	else if (ecb_expect_true (m == LENGTH_EXT2))
		731	{
603	WANT (3);	732	WANT (2);
604	dec->cur += 3;	733	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	734	return (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	735	\| ((UV)dec->cur[-1]);
607		736	}
608	case 26:	737	else if (ecb_expect_true (m == LENGTH_EXT4))
		738	{
609	WANT (5);	739	WANT (4);
610	dec->cur += 5;	740	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	741	return (((UV)dec->cur[-4]) << 24)
612	\| (((UV)dec->cur[-3]) << 16)	742	\| (((UV)dec->cur[-3]) << 16)
613	\| (((UV)dec->cur[-2]) << 8)	743	\| (((UV)dec->cur[-2]) << 8)
614	\| ((UV)dec->cur[-1]);	744	\| ((UV)dec->cur[-1]);
615		745	}
616	case 27:	746	else if (ecb_expect_true (m == LENGTH_EXT8))
		747	{
617	WANT (9);	748	WANT (8);
618	dec->cur += 9;	749	dec->cur += 8;
619		750
620	return	751	return
621	#if UVSIZE < 8	752	#if UVSIZE < 8
622	0	753	0
623	#else	754	#else
624	(((UV)dec->cur[-8]) << 56)	755	(((UV)dec->cur[-8]) << 56)
625	\| (((UV)dec->cur[-7]) << 48)	756	\| (((UV)dec->cur[-7]) << 48)
626	\| (((UV)dec->cur[-6]) << 40)	757	\| (((UV)dec->cur[-6]) << 40)
627	\| (((UV)dec->cur[-5]) << 32)	758	\| (((UV)dec->cur[-5]) << 32)
628	#endif	759	#endif
629	\| (((UV)dec->cur[-4]) << 24)	760	\| (((UV)dec->cur[-4]) << 24)
630	\| (((UV)dec->cur[-3]) << 16)	761	\| (((UV)dec->cur[-3]) << 16)
631	\| (((UV)dec->cur[-2]) << 8)	762	\| (((UV)dec->cur[-2]) << 8)
632	\| ((UV)dec->cur[-1]);	763	\| ((UV)dec->cur[-1]);
633		764	}
634	default:	765	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	766	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		767
638	fail:	768	fail:
639	return 0;	769	return 0;
640	}	770	}
641		771
…		…
646	{	776	{
647	AV *av = newAV ();	777	AV *av = newAV ();
648		778
649	DEC_INC_DEPTH;	779	DEC_INC_DEPTH;
650		780
651	if ((*dec->cur & 31) == 31)	781	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
652	{	782	{
653	++dec->cur;	783	++dec->cur;
654		784
655	for (;;)	785	for (;;)
656	{	786	{
657	WANT (1);	787	WANT (1);
658		788
659	if (*dec->cur == (0xe0 \| 31))	789	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
660	{	790	{
661	++dec->cur;	791	++dec->cur;
662	break;	792	break;
663	}	793	}
664		794
665	av_push (av, decode_sv (dec));	795	av_push (av, decode_sv (dec));
666	}	796	}
667	}	797	}
668	else	798	else
669	{	799	{
670	int i, len = decode_uint (dec);	800	UV i, len = decode_uint (dec);
671		801
		802	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);	803	av_fill (av, len - 1);
673		804
674	for (i = 0; i < len; ++i)	805	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	806	AvARRAY (av)[i] = decode_sv (dec);
676	}	807	}
…		…
689	{	820	{
690	// for speed reasons, we specialcase single-string	821	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	822	// byte or utf-8 strings as keys, but only when !stringref
692		823
693	if (ecb_expect_true (!dec->stringref))	824	if (ecb_expect_true (!dec->stringref))
694	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	825	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	826	{
696	I32 len = decode_uint (dec);	827	STRLEN len = decode_uint (dec);
697	char key = (char )dec->cur;	828	char key = (char )dec->cur;
698		829
		830	WANT (len);
699	dec->cur += len;	831	dec->cur += len;
700		832
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);	833	hv_store (hv, key, len, decode_sv (dec), 0);
705		834
706	return;	835	return;
707	}	836	}
708	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	837	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	838	{
710	I32 len = decode_uint (dec);	839	STRLEN len = decode_uint (dec);
711	char key = (char )dec->cur;	840	char key = (char )dec->cur;
712		841
		842	WANT (len);
713	dec->cur += len;	843	dec->cur += len;
714		844
715	if (ecb_expect_false (dec->stringref))	845	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	846	if (!is_utf8_string (key, len))
		847	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
717		848
718	hv_store (hv, key, -len, decode_sv (dec), 0);	849	hv_store (hv, key, -len, decode_sv (dec), 0);
719		850
720	return;	851	return;
721	}	852	}
722		853
723	SV *k = decode_sv (dec);	854	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	855	SV *v = decode_sv (dec);
725		856
		857	// we leak memory if uncaught exceptions are thrown by random magical
		858	// methods, and this is hopefully the only place where it can happen,
		859	// so if there is a chance of an exception, take the very slow path.
		860	// since catching exceptions is "undocumented/internal/forbidden" by
		861	// the new p5p powers, we need to call out to a perl function :/
		862	if (ecb_expect_false (SvAMAGIC (k)))
		863	{
		864	dSP;
		865
		866	ENTER; SAVETMPS;
		867	PUSHMARK (SP);
		868	EXTEND (SP, 3);
		869	PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
		870	PUSHs (sv_2mortal (k));
		871	PUSHs (sv_2mortal (v));
		872
		873	PUTBACK;
		874	call_pv ("CBOR::XS::_hv_store", G_VOID \| G_DISCARD \| G_EVAL);
		875	SPAGAIN;
		876
		877	FREETMPS; LEAVE;
		878
		879	if (SvTRUE (ERRSV))
		880	ERR_ERRSV;
		881
		882	return;
		883	}
		884
726	hv_store_ent (hv, k, v, 0);	885	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	886	SvREFCNT_dec (k);
		887
		888	fail:
		889	;
728	}	890	}
729		891
730	static SV *	892	static SV *
731	decode_hv (dec_t *dec)	893	decode_hv (dec_t *dec)
732	{	894	{
733	HV *hv = newHV ();	895	HV *hv = newHV ();
734		896
735	DEC_INC_DEPTH;	897	DEC_INC_DEPTH;
736		898
737	if ((*dec->cur & 31) == 31)	899	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
738	{	900	{
739	++dec->cur;	901	++dec->cur;
740		902
741	for (;;)	903	for (;;)
742	{	904	{
743	WANT (1);	905	WANT (1);
744		906
745	if (*dec->cur == (0xe0 \| 31))	907	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
746	{	908	{
747	++dec->cur;	909	++dec->cur;
748	break;	910	break;
749	}	911	}
750		912
751	decode_he (dec, hv);	913	decode_he (dec, hv);
752	}	914	}
753	}	915	}
754	else	916	else
755	{	917	{
756	int pairs = decode_uint (dec);	918	UV pairs = decode_uint (dec);
		919
		920	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
757		921
758	while (pairs--)	922	while (pairs--)
759	decode_he (dec, hv);	923	decode_he (dec, hv);
760	}	924	}
761		925
…		…
771	static SV *	935	static SV *
772	decode_str (dec_t *dec, int utf8)	936	decode_str (dec_t *dec, int utf8)
773	{	937	{
774	SV *sv = 0;	938	SV *sv = 0;
775		939
776	if ((*dec->cur & 31) == 31)	940	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
777	{	941	{
778	// indefinite length strings	942	// indefinite length strings
779	++dec->cur;	943	++dec->cur;
780		944
781	unsigned char major = *dec->cur & 0xe0;	945	U8 major = *dec->cur & MAJOR_MISC;
782		946
783	sv = newSVpvn ("", 0);	947	sv = newSVpvn ("", 0);
784		948
785	for (;;)	949	for (;;)
786	{	950	{
787	WANT (1);	951	WANT (1);
788		952
789	if ((*dec->cur ^ major) >= 31)	953	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 \| 31))	954	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
791	{	955	{
792	++dec->cur;	956	++dec->cur;
793	break;	957	break;
794	}	958	}
795	else	959	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	978	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	979	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	980	}
817		981
818	if (utf8)	982	if (utf8)
		983	{
		984	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		985	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		986	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		987
819	SvUTF8_on (sv);	988	SvUTF8_on (sv);
		989	}
820		990
821	return sv;	991	return sv;
822		992
823	fail:	993	fail:
824	SvREFCNT_dec (sv);	994	SvREFCNT_dec (sv);
…		…
843	sv = newRV_noinc (decode_sv (dec));	1013	sv = newRV_noinc (decode_sv (dec));
844	break;	1014	break;
845		1015
846	case CBOR_TAG_STRINGREF_NAMESPACE:	1016	case CBOR_TAG_STRINGREF_NAMESPACE:
847	{	1017	{
		1018	// do nmot use SAVETMPS/FREETMPS, as these will
		1019	// erase mortalised caches, e.g. "shareable"
848	ENTER; SAVETMPS;	1020	ENTER;
849		1021
850	SAVESPTR (dec->stringref);	1022	SAVESPTR (dec->stringref);
851	dec->stringref = (AV )sv_2mortal ((SV )newAV ());	1023	dec->stringref = (AV )sv_2mortal ((SV )newAV ());
852		1024
853	sv = decode_sv (dec);	1025	sv = decode_sv (dec);
854		1026
855	FREETMPS; LEAVE;	1027	LEAVE;
856	}	1028	}
857	break;	1029	break;
858		1030
859	case CBOR_TAG_STRINGREF:	1031	case CBOR_TAG_STRINGREF:
860	{	1032	{
861	if ((*dec->cur >> 5) != 0)	1033	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1034	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		1035
864	UV idx = decode_uint (dec);	1036	UV idx = decode_uint (dec);
865		1037
866	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	1038	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	1045	case CBOR_TAG_VALUE_SHAREABLE:
874	{	1046	{
875	if (ecb_expect_false (!dec->shareable))	1047	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	1048	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
877		1049
		1050	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1051	{
878	sv = newSV (0);	1052	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1053	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		1054
881	SV *osv = decode_sv (dec);	1055	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	1056	sv_setsv (sv, osv);
883	SvREFCNT_dec_NN (osv);	1057	SvREFCNT_dec_NN (osv);
		1058	}
		1059	else
		1060	{
		1061	av_push (dec->shareable, &PL_sv_undef);
		1062	int idx = AvFILLp (dec->shareable);
		1063	sv = decode_sv (dec);
		1064	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		1065	}
884	}	1066	}
885	break;	1067	break;
886		1068
887	case CBOR_TAG_VALUE_SHAREDREF:	1069	case CBOR_TAG_VALUE_SHAREDREF:
888	{	1070	{
889	if ((*dec->cur >> 5) != 0)	1071	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1072	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		1073
892	UV idx = decode_uint (dec);	1074	UV idx = decode_uint (dec);
893		1075
894	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	1076	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1077	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		1078
897	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1079	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1080
		1081	if (sv == &PL_sv_undef)
		1082	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
898	}	1083	}
899	break;	1084	break;
900		1085
901	case CBOR_TAG_PERL_OBJECT:	1086	case CBOR_TAG_PERL_OBJECT:
902	{	1087	{
		1088	if (dec->cbor.flags & F_FORBID_OBJECTS)
		1089	goto filter;
		1090
903	sv = decode_sv (dec);	1091	sv = decode_sv (dec);
904		1092
905	if (!SvROK (sv) \|\| SvTYPE (SvRV (sv)) != SVt_PVAV)	1093	if (!SvROK (sv) \|\| SvTYPE (SvRV (sv)) != SVt_PVAV)
906	ERR ("corrupted CBOR data (non-array perl object)");	1094	ERR ("corrupted CBOR data (non-array perl object)");
907		1095
…		…
917	if (!method)	1105	if (!method)
918	ERR ("cannot decode perl-object (package does not have a THAW method)");	1106	ERR ("cannot decode perl-object (package does not have a THAW method)");
919		1107
920	dSP;	1108	dSP;
921		1109
922	ENTER; SAVETMPS; PUSHMARK (SP);	1110	ENTER; SAVETMPS;
		1111	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1112	EXTEND (SP, len + 1);
924	// we re-bless the reference to get overload and other niceties right	1113	// we re-bless the reference to get overload and other niceties right
925	PUSHs (*av_fetch (av, 0, 1));	1114	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1115	PUSHs (sv_cbor);
927		1116
…		…
935	SPAGAIN;	1124	SPAGAIN;
936		1125
937	if (SvTRUE (ERRSV))	1126	if (SvTRUE (ERRSV))
938	{	1127	{
939	FREETMPS; LEAVE;	1128	FREETMPS; LEAVE;
940	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1129	ERR_ERRSV;
941	}	1130	}
942		1131
943	SvREFCNT_dec (sv);	1132	SvREFCNT_dec (sv);
944	sv = SvREFCNT_inc (POPs);	1133	sv = SvREFCNT_inc (POPs);
945		1134
…		…
948	FREETMPS; LEAVE;	1137	FREETMPS; LEAVE;
949	}	1138	}
950	break;	1139	break;
951		1140
952	default:	1141	default:
		1142	filter:
953	{	1143	{
		1144	SV *tag_sv = newSVuv (tag);
		1145
954	sv = decode_sv (dec);	1146	sv = decode_sv (dec);
955		1147
956	dSP;	1148	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1149	ENTER; SAVETMPS;
		1150	SAVESTACK_POS ();
		1151	PUSHMARK (SP);
958	EXTEND (SP, 2);	1152	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1153	PUSHs (tag_sv);
960	PUSHs (sv);	1154	PUSHs (sv);
961		1155
962	PUTBACK;	1156	PUTBACK;
963	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);	1157	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
964	SPAGAIN;	1158	SPAGAIN;
965		1159
966	if (SvTRUE (ERRSV))	1160	if (SvTRUE (ERRSV))
967	{	1161	{
		1162	SvREFCNT_dec (tag_sv);
968	FREETMPS; LEAVE;	1163	FREETMPS; LEAVE;
969	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1164	ERR_ERRSV;
970	}	1165	}
971		1166
972	if (count)	1167	if (count)
973	{	1168	{
		1169	SvREFCNT_dec (tag_sv);
974	SvREFCNT_dec (sv);	1170	SvREFCNT_dec (sv);
975	sv = SvREFCNT_inc (POPs);	1171	sv = SvREFCNT_inc (POPs);
976	}	1172	}
977	else	1173	else
978	{	1174	{
979	AV *av = newAV ();	1175	AV *av = newAV ();
980	av_push (av, newSVuv (tag));	1176	av_push (av, tag_sv);
981	av_push (av, sv);	1177	av_push (av, sv);
982		1178
983	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1179	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
984	? cbor_tagged_stash	1180	? cbor_tagged_stash
985	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1181	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
1003	static SV *	1199	static SV *
1004	decode_sv (dec_t *dec)	1200	decode_sv (dec_t *dec)
1005	{	1201	{
1006	WANT (1);	1202	WANT (1);
1007		1203
1008	switch (*dec->cur >> 5)	1204	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1205	{
1010	case 0: // unsigned int	1206	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1207	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1208	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1209	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1210	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1211	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1212	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1213
1018	case 4: // array	1214	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)	1215	switch (*dec->cur++ & MINOR_MASK)
1026	{	1216	{
1027	case 20:	1217	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1218	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1219	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1220	#endif
1031	return newSVsv (types_false);	1221	return newSVsv (types_false);
1032	case 21:	1222	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1223	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1224	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1225	#endif
1036	return newSVsv (types_true);	1226	return newSVsv (types_true);
1037	case 22:	1227	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1228	return newSVsv (&PL_sv_undef);
1039	case 23:	1229	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1230	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1231	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1232	#endif
1043	return newSVsv (types_error);	1233	return newSVsv (types_error);
1044		1234
1045	case 25:	1235	case MISC_FLOAT16:
1046	{	1236	{
1047	WANT (2);	1237	WANT (2);
1048		1238
1049	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1239	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1050	dec->cur += 2;	1240	dec->cur += 2;
1051		1241
1052	return newSVnv (ecb_binary16_to_float (fp));	1242	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1243	}
1054		1244
1055	case 26:	1245	case MISC_FLOAT32:
1056	{	1246	{
1057	uint32_t fp;	1247	uint32_t fp;
1058	WANT (4);	1248	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1249	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1250	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1253	fp = ecb_bswap32 (fp);
1064		1254
1065	return newSVnv (ecb_binary32_to_float (fp));	1255	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1256	}
1067		1257
1068	case 27:	1258	case MISC_FLOAT64:
1069	{	1259	{
1070	uint64_t fp;	1260	uint64_t fp;
1071	WANT (8);	1261	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1262	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1263	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1266	fp = ecb_bswap64 (fp);
1077		1267
1078	return newSVnv (ecb_binary64_to_double (fp));	1268	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1269	}
1080		1270
1081	// 0..19 unassigned	1271	// 0..19 unassigned simple
1082	// 24 reserved + unassigned (reserved values are not encodable)	1272	// 24 reserved + unassigned simple (reserved values are not encodable)
		1273	// 28-30 unassigned misc
		1274	// 31 break code
1083	default:	1275	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1276	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1277	}
1086		1278
1087	break;	1279	break;
1088	}	1280	}
1089		1281
…		…
1092	}	1284	}
1093		1285
1094	static SV *	1286	static SV *
1095	decode_cbor (SV string, CBOR cbor, char **offset_return)	1287	decode_cbor (SV string, CBOR cbor, char **offset_return)
1096	{	1288	{
1097	dec_t dec = { };	1289	dec_t dec = { 0 };
1098	SV *sv;	1290	SV *sv;
1099	STRLEN len;	1291	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1292	char *data = SvPVbyte (string, len);
1101		1293
1102	if (len > cbor->max_size && cbor->max_size)	1294	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1308	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1309	dec.err = "garbage after CBOR object";
1118		1310
1119	if (dec.err)	1311	if (dec.err)
1120	{	1312	{
		1313	if (dec.shareable)
		1314	{
		1315	// need to break cyclic links, which would all be in shareable
		1316	int i;
		1317	SV **svp;
		1318
		1319	for (i = av_len (dec.shareable) + 1; i--; )
		1320	if ((svp = av_fetch (dec.shareable, i, 0)))
		1321	sv_setsv (*svp, &PL_sv_undef);
		1322	}
		1323
1121	SvREFCNT_dec (sv);	1324	SvREFCNT_dec (sv);
		1325
		1326	if (dec.err_sv)
		1327	sv_2mortal (dec.err_sv);
		1328
1122	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1329	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1123	}	1330	}
1124		1331
1125	sv = sv_2mortal (sv);	1332	sv = sv_2mortal (sv);
1126		1333
1127	return sv;	1334	return sv;
1128	}	1335	}
1129		1336
		1337	/////////////////////////////////////////////////////////////////////////////
		1338	// incremental parser
		1339
		1340	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1341
		1342	// returns 0 for notyet, 1 for success or error
		1343	static int
		1344	incr_parse (CBOR self, SV cborstr)
		1345	{
		1346	STRLEN cur;
		1347	SvPV (cborstr, cur);
		1348
		1349	while (ecb_expect_true (self->incr_need <= cur))
		1350	{
		1351	// table of integer count bytes
		1352	static I8 incr_len[MINOR_MASK + 1] = {
		1353	0, 0, 0, 0, 0, 0, 0, 0,
		1354	0, 0, 0, 0, 0, 0, 0, 0,
		1355	0, 0, 0, 0, 0, 0, 0, 0,
		1356	1, 2, 4, 8,-1,-1,-1,-2
		1357	};
		1358
		1359	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1360	U8 m = *p & MINOR_MASK;
		1361	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1362	I8 ilen = incr_len[m];
		1363
		1364	self->incr_need = self->incr_pos + 1;
		1365
		1366	if (ecb_expect_false (ilen < 0))
		1367	{
		1368	if (m != MINOR_INDEF)
		1369	return 1; // error
		1370
		1371	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1372	{
		1373	if (count >= 0)
		1374	return 1; // error
		1375
		1376	count = 1;
		1377	}
		1378	else
		1379	{
		1380	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1381	count = -1;
		1382	}
		1383	}
		1384	else
		1385	{
		1386	self->incr_need += ilen;
		1387	if (ecb_expect_false (self->incr_need > cur))
		1388	return 0;
		1389
		1390	int major = *p >> MAJOR_SHIFT;
		1391
		1392	switch (major)
		1393	{
		1394	case MAJOR_TAG >> MAJOR_SHIFT:
		1395	++count; // tags merely prefix another value
		1396	break;
		1397
		1398	case MAJOR_BYTES >> MAJOR_SHIFT:
		1399	case MAJOR_TEXT >> MAJOR_SHIFT:
		1400	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1401	case MAJOR_MAP >> MAJOR_SHIFT:
		1402	{
		1403	UV len;
		1404
		1405	if (ecb_expect_false (ilen))
		1406	{
		1407	len = 0;
		1408
		1409	do {
		1410	len = (len << 8) \| *++p;
		1411	} while (--ilen);
		1412	}
		1413	else
		1414	len = m;
		1415
		1416	switch (major)
		1417	{
		1418	case MAJOR_BYTES >> MAJOR_SHIFT:
		1419	case MAJOR_TEXT >> MAJOR_SHIFT:
		1420	self->incr_need += len;
		1421	if (ecb_expect_false (self->incr_need > cur))
		1422	return 0;
		1423
		1424	break;
		1425
		1426	case MAJOR_MAP >> MAJOR_SHIFT:
		1427	len <<= 1;
		1428	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1429	if (len)
		1430	{
		1431	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1432	count = len + 1;
		1433	}
		1434	break;
		1435	}
		1436	}
		1437	}
		1438	}
		1439
		1440	self->incr_pos = self->incr_need;
		1441
		1442	if (count > 0)
		1443	{
		1444	while (!--count)
		1445	{
		1446	if (!AvFILLp (self->incr_count))
		1447	return 1; // done
		1448
		1449	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1450	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1451	}
		1452
		1453	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1454	}
		1455	}
		1456
		1457	return 0;
		1458	}
		1459
		1460
1130	/////////////////////////////////////////////////////////////////////////////	1461	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1462	// XS interface functions
1132		1463
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1464	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1465
…		…
1146		1477
1147	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1478	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1148		1479
1149	sv_cbor = newSVpv ("CBOR", 0);	1480	sv_cbor = newSVpv ("CBOR", 0);
1150	SvREADONLY_on (sv_cbor);	1481	SvREADONLY_on (sv_cbor);
		1482
		1483	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1151	}	1484	}
1152		1485
1153	PROTOTYPES: DISABLE	1486	PROTOTYPES: DISABLE
1154		1487
1155	void CLONE (...)	1488	void CLONE (...)
…		…
1174	void shrink (CBOR *self, int enable = 1)	1507	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1508	ALIAS:
1176	shrink = F_SHRINK	1509	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1510	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1511	allow_sharing = F_ALLOW_SHARING
		1512	allow_cycles = F_ALLOW_CYCLES
		1513	forbid_objects = F_FORBID_OBJECTS
1179	pack_strings = F_PACK_STRINGS	1514	pack_strings = F_PACK_STRINGS
		1515	text_keys = F_TEXT_KEYS
		1516	text_strings = F_TEXT_STRINGS
		1517	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1518	PPCODE:
1181	{	1519	{
1182	if (enable)	1520	if (enable)
1183	self->flags \|= ix;	1521	self->flags \|= ix;
1184	else	1522	else
…		…
1190	void get_shrink (CBOR *self)	1528	void get_shrink (CBOR *self)
1191	ALIAS:	1529	ALIAS:
1192	get_shrink = F_SHRINK	1530	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1531	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1532	get_allow_sharing = F_ALLOW_SHARING
		1533	get_allow_cycles = F_ALLOW_CYCLES
		1534	get_forbid_objects = F_FORBID_OBJECTS
1195	get_pack_strings = F_PACK_STRINGS	1535	get_pack_strings = F_PACK_STRINGS
		1536	get_text_keys = F_TEXT_KEYS
		1537	get_text_strings = F_TEXT_STRINGS
		1538	get_validate_utf8 = F_VALIDATE_UTF8
1196	PPCODE:	1539	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1540	XPUSHs (boolSV (self->flags & ix));
1198		1541
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1542	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1543	PPCODE:
…		…
1249	EXTEND (SP, 2);	1592	EXTEND (SP, 2);
1250	PUSHs (sv);	1593	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1594	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1595	}
1253		1596
		1597	void incr_parse (CBOR self, SV cborstr)
		1598	ALIAS:
		1599	incr_parse_multiple = 1
		1600	PPCODE:
		1601	{
		1602	if (SvUTF8 (cborstr))
		1603	sv_utf8_downgrade (cborstr, 0);
		1604
		1605	if (!self->incr_count)
		1606	{
		1607	self->incr_count = newAV ();
		1608	self->incr_pos = 0;
		1609	self->incr_need = 1;
		1610
		1611	av_push (self->incr_count, newSViv (1));
		1612	}
		1613
		1614	do
		1615	{
		1616	if (!incr_parse (self, cborstr))
		1617	{
		1618	if (self->incr_need > self->max_size && self->max_size)
		1619	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1620	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1621
		1622	break;
		1623	}
		1624
		1625	SV *sv;
		1626	char *offset;
		1627
		1628	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1629	XPUSHs (sv);
		1630
		1631	sv_chop (cborstr, offset);
		1632
		1633	av_clear (self->incr_count);
		1634	av_push (self->incr_count, newSViv (1));
		1635
		1636	self->incr_pos = 0;
		1637	self->incr_need = self->incr_pos + 1;
		1638	}
		1639	while (ix);
		1640	}
		1641
		1642	void incr_reset (CBOR *self)
		1643	CODE:
		1644	{
		1645	SvREFCNT_dec (self->incr_count);
		1646	self->incr_count = 0;
		1647	}
		1648
1254	void DESTROY (CBOR *self)	1649	void DESTROY (CBOR *self)
1255	PPCODE:	1650	PPCODE:
1256	cbor_free (self);	1651	cbor_free (self);
1257		1652
1258	PROTOTYPES: ENABLE	1653	PROTOTYPES: ENABLE
1259		1654
1260	void encode_cbor (SV *scalar)	1655	void encode_cbor (SV *scalar)
		1656	ALIAS:
		1657	encode_cbor = 0
		1658	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1659	PPCODE:
1262	{	1660	{
1263	CBOR cbor;	1661	CBOR cbor;
1264	cbor_init (&cbor);	1662	cbor_init (&cbor);
		1663	cbor.flags \|= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1664	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1665	XPUSHs (scalar);
1267	}	1666	}
1268		1667
1269	void decode_cbor (SV *cborstr)	1668	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs. Revision 1.61 by root, Sat Nov 26 02:10:19 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs.
Revision 1.61 by root, Sat Nov 26 02:10:19 2016 UTC