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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs.
Revision 1.44 by root, Tue Feb 18 22:12:12 2014 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_VALIDATE_UTF8 0x00000020UL
63		106
64	#define INIT_SIZE 32 // initial scalar size to be allocated	107	#define INIT_SIZE 32 // initial scalar size to be allocated
65		108
66	#define SB do {	109	#define SB do {
67	#define SE } while (0)	110	#define SE } while (0)
…		…
86	typedef struct {	129	typedef struct {
87	U32 flags;	130	U32 flags;
88	U32 max_depth;	131	U32 max_depth;
89	STRLEN max_size;	132	STRLEN max_size;
90	SV *filter;	133	SV *filter;
		134
		135	// for the incremental parser
		136	STRLEN incr_pos; // the current offset into the text
		137	STRLEN incr_need; // minimum bytes needed to decode
		138	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91	} CBOR;	139	} CBOR;
92		140
93	ecb_inline void	141	ecb_inline void
94	cbor_init (CBOR *cbor)	142	cbor_init (CBOR *cbor)
95	{	143	{
…		…
99		147
100	ecb_inline void	148	ecb_inline void
101	cbor_free (CBOR *cbor)	149	cbor_free (CBOR *cbor)
102	{	150	{
103	SvREFCNT_dec (cbor->filter);	151	SvREFCNT_dec (cbor->filter);
		152	SvREFCNT_dec (cbor->incr_count);
104	}	153	}
105		154
106	/////////////////////////////////////////////////////////////////////////////	155	/////////////////////////////////////////////////////////////////////////////
107	// utility functions	156	// utility functions
108		157
…		…
138	{	187	{
139	return idx > 23	188	return idx > 23
140	? idx > 0xffU	189	? idx > 0xffU
141	? idx > 0xffffU	190	? idx > 0xffffU
142	? idx > 0xffffffffU	191	? idx > 0xffffffffU
		192	? 11
143	? 7	193	: 7
144	: 6
145	: 5	194	: 5
146	: 4	195	: 4
147	: 3;	196	: 3;
148	}	197	}
149		198
…		…
186	static void	235	static void
187	encode_uint (enc_t *enc, int major, UV len)	236	encode_uint (enc_t *enc, int major, UV len)
188	{	237	{
189	need (enc, 9);	238	need (enc, 9);
190		239
191	if (len < 24)	240	if (ecb_expect_true (len < LENGTH_EXT1))
192	*enc->cur++ = major \| len;	241	*enc->cur++ = major \| len;
193	else if (len <= 0xff)	242	else if (ecb_expect_true (len <= 0xffU))
194	{	243	{
195	*enc->cur++ = major \| 24;	244	*enc->cur++ = major \| LENGTH_EXT1;
196	*enc->cur++ = len;	245	*enc->cur++ = len;
197	}	246	}
198	else if (len <= 0xffff)	247	else if (len <= 0xffffU)
199	{	248	{
200	*enc->cur++ = major \| 25;	249	*enc->cur++ = major \| LENGTH_EXT2;
201	*enc->cur++ = len >> 8;	250	*enc->cur++ = len >> 8;
202	*enc->cur++ = len;	251	*enc->cur++ = len;
203	}	252	}
204	else if (len <= 0xffffffff)	253	else if (len <= 0xffffffffU)
205	{	254	{
206	*enc->cur++ = major \| 26;	255	*enc->cur++ = major \| LENGTH_EXT4;
207	*enc->cur++ = len >> 24;	256	*enc->cur++ = len >> 24;
208	*enc->cur++ = len >> 16;	257	*enc->cur++ = len >> 16;
209	*enc->cur++ = len >> 8;	258	*enc->cur++ = len >> 8;
210	*enc->cur++ = len;	259	*enc->cur++ = len;
211	}	260	}
212	else	261	else
213	{	262	{
214	*enc->cur++ = major \| 27;	263	*enc->cur++ = major \| LENGTH_EXT8;
215	*enc->cur++ = len >> 56;	264	*enc->cur++ = len >> 56;
216	*enc->cur++ = len >> 48;	265	*enc->cur++ = len >> 48;
217	*enc->cur++ = len >> 40;	266	*enc->cur++ = len >> 40;
218	*enc->cur++ = len >> 32;	267	*enc->cur++ = len >> 32;
219	*enc->cur++ = len >> 24;	268	*enc->cur++ = len >> 24;
…		…
224	}	273	}
225		274
226	ecb_inline void	275	ecb_inline void
227	encode_tag (enc_t *enc, UV tag)	276	encode_tag (enc_t *enc, UV tag)
228	{	277	{
229	encode_uint (enc, 0xc0, tag);	278	encode_uint (enc, MAJOR_TAG, tag);
		279	}
		280
		281	ecb_inline void
		282	encode_str (enc_t enc, int utf8, char str, STRLEN len)
		283	{
		284	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		285	need (enc, len);
		286	memcpy (enc->cur, str, len);
		287	enc->cur += len;
230	}	288	}
231		289
232	static void	290	static void
233	encode_str (enc_t enc, int utf8, char str, STRLEN len)	291	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
234	{	292	{
235	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	293	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
236	{	294	{
237	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	295	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
238		296
239	if (SvOK (*svp))	297	if (SvOK (*svp))
240	{	298	{
241	// already registered, use stringref	299	// already registered, use stringref
242	encode_tag (enc, CBOR_TAG_STRINGREF);	300	encode_tag (enc, CBOR_TAG_STRINGREF);
243	encode_uint (enc, 0x00, SvUV (*svp));	301	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
244	return;	302	return;
245	}	303	}
246	else if (len >= minimum_string_length (enc->stringref_idx))	304	else if (len >= minimum_string_length (enc->stringref_idx))
247	{	305	{
248	// register only	306	// register only
249	sv_setuv (*svp, enc->stringref_idx);	307	sv_setuv (*svp, enc->stringref_idx);
250	++enc->stringref_idx;	308	++enc->stringref_idx;
251	}	309	}
252	}	310	}
253		311
254	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	312	encode_str (enc, utf8, str, len);
255	need (enc, len);
256	memcpy (enc->cur, str, len);
257	enc->cur += len;
258	}	313	}
259		314
260	static void encode_sv (enc_t enc, SV sv);	315	static void encode_sv (enc_t enc, SV sv);
261		316
262	static void	317	static void
…		…
267	if (enc->depth >= enc->cbor.max_depth)	322	if (enc->depth >= enc->cbor.max_depth)
268	croak (ERR_NESTING_EXCEEDED);	323	croak (ERR_NESTING_EXCEEDED);
269		324
270	++enc->depth;	325	++enc->depth;
271		326
272	encode_uint (enc, 0x80, len + 1);	327	encode_uint (enc, MAJOR_ARRAY, len + 1);
273		328
274	for (i = 0; i <= len; ++i)	329	for (i = 0; i <= len; ++i)
275	{	330	{
276	SV **svp = av_fetch (av, i, 0);	331	SV **svp = av_fetch (av, i, 0);
277	encode_sv (enc, svp ? *svp : &PL_sv_undef);	332	encode_sv (enc, svp ? *svp : &PL_sv_undef);
…		…
292		347
293	int pairs = hv_iterinit (hv);	348	int pairs = hv_iterinit (hv);
294	int mg = SvMAGICAL (hv);	349	int mg = SvMAGICAL (hv);
295		350
296	if (mg)	351	if (mg)
297	encode_ch (enc, 0xa0 \| 31);	352	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
298	else	353	else
299	encode_uint (enc, 0xa0, pairs);	354	encode_uint (enc, MAJOR_MAP, pairs);
300		355
301	while ((he = hv_iternext (hv)))	356	while ((he = hv_iternext (hv)))
302	{	357	{
303	if (HeKLEN (he) == HEf_SVKEY)	358	if (HeKLEN (he) == HEf_SVKEY)
304	encode_sv (enc, HeSVKEY (he));	359	encode_sv (enc, HeSVKEY (he));
305	else	360	else
306	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	361	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
307		362
308	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	363	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
309	}	364	}
310		365
311	if (mg)	366	if (mg)
312	encode_ch (enc, 0xe0 \| 31);	367	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
313		368
314	--enc->depth;	369	--enc->depth;
315	}	370	}
316		371
317	// encode objects, arrays and special \0=false and \1=true values.	372	// encode objects, arrays and special \0=false and \1=true values.
318	static void	373	static void
319	encode_rv (enc_t enc, SV sv)	374	encode_rv (enc_t enc, SV sv)
320	{	375	{
321	SvGETMAGIC (sv);	376	SvGETMAGIC (sv);
322
323	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
324	&& ecb_expect_false (SvREFCNT (sv) > 1))
325	{
326	if (!enc->shareable)
327	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
328
329	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
330
331	if (SvOK (*svp))
332	{
333	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
334	encode_uint (enc, 0x00, SvUV (*svp));
335	return;
336	}
337	else
338	{
339	sv_setuv (*svp, enc->shareable_idx);
340	++enc->shareable_idx;
341	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
342	}
343	}
344		377
345	svtype svt = SvTYPE (sv);	378	svtype svt = SvTYPE (sv);
346		379
347	if (ecb_expect_false (SvOBJECT (sv)))	380	if (ecb_expect_false (SvOBJECT (sv)))
348	{	381	{
…		…
355	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	388	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
356	? cbor_tagged_stash	389	? cbor_tagged_stash
357	: gv_stashpv ("CBOR::XS::Tagged" , 1);	390	: gv_stashpv ("CBOR::XS::Tagged" , 1);
358		391
359	HV *stash = SvSTASH (sv);	392	HV *stash = SvSTASH (sv);
360	GV *method;
361		393
362	if (stash == boolean_stash)	394	if (stash == boolean_stash)
363	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	395	{
		396	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		397	return;
		398	}
364	else if (stash == error_stash)	399	else if (stash == error_stash)
365	encode_ch (enc, 0xe0 \| 23);	400	{
		401	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		402	return;
		403	}
366	else if (stash == tagged_stash)	404	else if (stash == tagged_stash)
367	{	405	{
368	if (svt != SVt_PVAV)	406	if (svt != SVt_PVAV)
369	croak ("encountered CBOR::XS::Tagged object that isn't an array");	407	croak ("encountered CBOR::XS::Tagged object that isn't an array");
370		408
371	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	409	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
372	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	410	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		411
		412	return;
		413	}
		414	}
		415
		416	if (ecb_expect_false (SvREFCNT (sv) > 1)
		417	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		418	{
		419	if (!enc->shareable)
		420	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		421
		422	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		423
		424	if (SvOK (*svp))
373	}	425	{
		426	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		427	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		428	return;
		429	}
		430	else
		431	{
		432	sv_setuv (*svp, enc->shareable_idx);
		433	++enc->shareable_idx;
		434	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		435	}
		436	}
		437
		438	if (ecb_expect_false (SvOBJECT (sv)))
		439	{
		440	HV *stash = SvSTASH (sv);
		441	GV *method;
		442
374	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	443	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
375	{	444	{
376	dSP;	445	dSP;
377		446
378	ENTER; SAVETMPS; PUSHMARK (SP);	447	ENTER; SAVETMPS; PUSHMARK (SP);
379	// we re-bless the reference to get overload and other niceties right	448	// we re-bless the reference to get overload and other niceties right
…		…
411	// catch this surprisingly common error	480	// catch this surprisingly common error
412	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	481	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
413	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	482	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
414		483
415	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	484	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
416	encode_uint (enc, 0x80, count + 1);	485	encode_uint (enc, MAJOR_ARRAY, count + 1);
417	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	486	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
418		487
419	while (count)	488	while (count)
420	encode_sv (enc, SP[1 - count--]);	489	encode_sv (enc, SP[1 - count--]);
421		490
422	PUTBACK;	491	PUTBACK;
…		…
443	{	512	{
444	double nv = SvNVX (sv);	513	double nv = SvNVX (sv);
445		514
446	need (enc, 9);	515	need (enc, 9);
447		516
448	if (ecb_expect_false (nv == (U32)nv))	517	if (ecb_expect_false (nv == (NV)(U32)nv))
449	encode_uint (enc, 0x00, (U32)nv);	518	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
450	//TODO: maybe I32?	519	//TODO: maybe I32?
451	else if (ecb_expect_false (nv == (float)nv))	520	else if (ecb_expect_false (nv == (float)nv))
452	{	521	{
453	uint32_t fp = ecb_float_to_binary32 (nv);	522	uint32_t fp = ecb_float_to_binary32 (nv);
454		523
455	*enc->cur++ = 0xe0 \| 26;	524	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
456		525
457	if (!ecb_big_endian ())	526	if (!ecb_big_endian ())
458	fp = ecb_bswap32 (fp);	527	fp = ecb_bswap32 (fp);
459		528
460	memcpy (enc->cur, &fp, 4);	529	memcpy (enc->cur, &fp, 4);
…		…
462	}	531	}
463	else	532	else
464	{	533	{
465	uint64_t fp = ecb_double_to_binary64 (nv);	534	uint64_t fp = ecb_double_to_binary64 (nv);
466		535
467	*enc->cur++ = 0xe0 \| 27;	536	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
468		537
469	if (!ecb_big_endian ())	538	if (!ecb_big_endian ())
470	fp = ecb_bswap64 (fp);	539	fp = ecb_bswap64 (fp);
471		540
472	memcpy (enc->cur, &fp, 8);	541	memcpy (enc->cur, &fp, 8);
…		…
481		550
482	if (SvPOKp (sv))	551	if (SvPOKp (sv))
483	{	552	{
484	STRLEN len;	553	STRLEN len;
485	char *str = SvPV (sv, len);	554	char *str = SvPV (sv, len);
486	encode_str (enc, SvUTF8 (sv), str, len);	555	encode_strref (enc, SvUTF8 (sv), str, len);
487	}	556	}
488	else if (SvNOKp (sv))	557	else if (SvNOKp (sv))
489	encode_nv (enc, sv);	558	encode_nv (enc, sv);
490	else if (SvIOKp (sv))	559	else if (SvIOKp (sv))
491	{	560	{
492	if (SvIsUV (sv))	561	if (SvIsUV (sv))
493	encode_uint (enc, 0x00, SvUVX (sv));	562	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
494	else if (SvIVX (sv) >= 0)	563	else if (SvIVX (sv) >= 0)
495	encode_uint (enc, 0x00, SvIVX (sv));	564	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
496	else	565	else
497	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	566	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
498	}	567	}
499	else if (SvROK (sv))	568	else if (SvROK (sv))
500	encode_rv (enc, SvRV (sv));	569	encode_rv (enc, SvRV (sv));
501	else if (!SvOK (sv))	570	else if (!SvOK (sv))
502	encode_ch (enc, 0xe0 \| 22);	571	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
503	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	572	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
504	encode_ch (enc, 0xe0 \| 23);	573	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
505	else	574	else
506	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	575	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
507	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	576	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
508	}	577	}
509		578
…		…
517	enc.cur = SvPVX (enc.sv);	586	enc.cur = SvPVX (enc.sv);
518	enc.end = SvEND (enc.sv);	587	enc.end = SvEND (enc.sv);
519		588
520	SvPOK_only (enc.sv);	589	SvPOK_only (enc.sv);
521		590
522	if (cbor->flags & F_ALLOW_STRINGREF)	591	if (cbor->flags & F_PACK_STRINGS)
523	{	592	{
524	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	593	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
525	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	594	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
526	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	595	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
527	}	596	}
…		…
562	#define DEC_DEC_DEPTH --dec->depth	631	#define DEC_DEC_DEPTH --dec->depth
563		632
564	static UV	633	static UV
565	decode_uint (dec_t *dec)	634	decode_uint (dec_t *dec)
566	{	635	{
567	switch (*dec->cur & 31)	636	U8 m = *dec->cur & MINOR_MASK;
568	{	637	++dec->cur;
569	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
570	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
571	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
572	return *dec->cur++ & 31;
573		638
574	case 24:	639	if (ecb_expect_true (m < LENGTH_EXT1))
		640	return m;
		641	else if (ecb_expect_true (m == LENGTH_EXT1))
		642	{
575	WANT (2);	643	WANT (1);
576	dec->cur += 2;	644	dec->cur += 1;
577	return dec->cur[-1];	645	return dec->cur[-1];
578		646	}
579	case 25:	647	else if (ecb_expect_true (m == LENGTH_EXT2))
		648	{
580	WANT (3);	649	WANT (2);
581	dec->cur += 3;	650	dec->cur += 2;
582	return (((UV)dec->cur[-2]) << 8)	651	return (((UV)dec->cur[-2]) << 8)
583	\| ((UV)dec->cur[-1]);	652	\| ((UV)dec->cur[-1]);
584		653	}
585	case 26:	654	else if (ecb_expect_true (m == LENGTH_EXT4))
		655	{
586	WANT (5);	656	WANT (4);
587	dec->cur += 5;	657	dec->cur += 4;
588	return (((UV)dec->cur[-4]) << 24)	658	return (((UV)dec->cur[-4]) << 24)
589	\| (((UV)dec->cur[-3]) << 16)	659	\| (((UV)dec->cur[-3]) << 16)
590	\| (((UV)dec->cur[-2]) << 8)	660	\| (((UV)dec->cur[-2]) << 8)
591	\| ((UV)dec->cur[-1]);	661	\| ((UV)dec->cur[-1]);
592		662	}
593	case 27:	663	else if (ecb_expect_true (m == LENGTH_EXT8))
		664	{
594	WANT (9);	665	WANT (8);
595	dec->cur += 9;	666	dec->cur += 8;
		667
		668	return
		669	#if UVSIZE < 8
		670	0
		671	#else
596	return (((UV)dec->cur[-8]) << 56)	672	(((UV)dec->cur[-8]) << 56)
597	\| (((UV)dec->cur[-7]) << 48)	673	\| (((UV)dec->cur[-7]) << 48)
598	\| (((UV)dec->cur[-6]) << 40)	674	\| (((UV)dec->cur[-6]) << 40)
599	\| (((UV)dec->cur[-5]) << 32)	675	\| (((UV)dec->cur[-5]) << 32)
		676	#endif
600	\| (((UV)dec->cur[-4]) << 24)	677	\| (((UV)dec->cur[-4]) << 24)
601	\| (((UV)dec->cur[-3]) << 16)	678	\| (((UV)dec->cur[-3]) << 16)
602	\| (((UV)dec->cur[-2]) << 8)	679	\| (((UV)dec->cur[-2]) << 8)
603	\| ((UV)dec->cur[-1]);	680	\| ((UV)dec->cur[-1]);
604		681	}
605	default:	682	else
606	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	683	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
607	}
608		684
609	fail:	685	fail:
610	return 0;	686	return 0;
611	}	687	}
612		688
…		…
617	{	693	{
618	AV *av = newAV ();	694	AV *av = newAV ();
619		695
620	DEC_INC_DEPTH;	696	DEC_INC_DEPTH;
621		697
622	if ((*dec->cur & 31) == 31)	698	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
623	{	699	{
624	++dec->cur;	700	++dec->cur;
625		701
626	for (;;)	702	for (;;)
627	{	703	{
628	WANT (1);	704	WANT (1);
629		705
630	if (*dec->cur == (0xe0 \| 31))	706	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
631	{	707	{
632	++dec->cur;	708	++dec->cur;
633	break;	709	break;
634	}	710	}
635		711
…		…
638	}	714	}
639	else	715	else
640	{	716	{
641	int i, len = decode_uint (dec);	717	int i, len = decode_uint (dec);
642		718
		719	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
643	av_fill (av, len - 1);	720	av_fill (av, len - 1);
644		721
645	for (i = 0; i < len; ++i)	722	for (i = 0; i < len; ++i)
646	AvARRAY (av)[i] = decode_sv (dec);	723	AvARRAY (av)[i] = decode_sv (dec);
647	}	724	}
…		…
660	{	737	{
661	// for speed reasons, we specialcase single-string	738	// for speed reasons, we specialcase single-string
662	// byte or utf-8 strings as keys, but only when !stringref	739	// byte or utf-8 strings as keys, but only when !stringref
663		740
664	if (ecb_expect_true (!dec->stringref))	741	if (ecb_expect_true (!dec->stringref))
665	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	742	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
666	{	743	{
667	I32 len = decode_uint (dec);	744	I32 len = decode_uint (dec);
668	char key = (char )dec->cur;	745	char key = (char )dec->cur;
669		746
670	dec->cur += len;	747	dec->cur += len;
671		748
672	if (ecb_expect_false (dec->stringref))
673	av_push (dec->stringref, newSVpvn (key, len));
674
675	hv_store (hv, key, len, decode_sv (dec), 0);	749	hv_store (hv, key, len, decode_sv (dec), 0);
676		750
677	return;	751	return;
678	}	752	}
679	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	753	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
680	{	754	{
681	I32 len = decode_uint (dec);	755	I32 len = decode_uint (dec);
682	char key = (char )dec->cur;	756	char key = (char )dec->cur;
683		757
684	dec->cur += len;	758	dec->cur += len;
685		759
686	if (ecb_expect_false (dec->stringref))	760	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
687	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	761	if (!is_utf8_string (key, len))
		762	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
688		763
689	hv_store (hv, key, -len, decode_sv (dec), 0);	764	hv_store (hv, key, -len, decode_sv (dec), 0);
690		765
691	return;	766	return;
692	}	767	}
…		…
694	SV *k = decode_sv (dec);	769	SV *k = decode_sv (dec);
695	SV *v = decode_sv (dec);	770	SV *v = decode_sv (dec);
696		771
697	hv_store_ent (hv, k, v, 0);	772	hv_store_ent (hv, k, v, 0);
698	SvREFCNT_dec (k);	773	SvREFCNT_dec (k);
		774
		775	fail:
		776	;
699	}	777	}
700		778
701	static SV *	779	static SV *
702	decode_hv (dec_t *dec)	780	decode_hv (dec_t *dec)
703	{	781	{
704	HV *hv = newHV ();	782	HV *hv = newHV ();
705		783
706	DEC_INC_DEPTH;	784	DEC_INC_DEPTH;
707		785
708	if ((*dec->cur & 31) == 31)	786	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
709	{	787	{
710	++dec->cur;	788	++dec->cur;
711		789
712	for (;;)	790	for (;;)
713	{	791	{
714	WANT (1);	792	WANT (1);
715		793
716	if (*dec->cur == (0xe0 \| 31))	794	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
717	{	795	{
718	++dec->cur;	796	++dec->cur;
719	break;	797	break;
720	}	798	}
721		799
…		…
742	static SV *	820	static SV *
743	decode_str (dec_t *dec, int utf8)	821	decode_str (dec_t *dec, int utf8)
744	{	822	{
745	SV *sv = 0;	823	SV *sv = 0;
746		824
747	if ((*dec->cur & 31) == 31)	825	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
748	{	826	{
		827	// indefinite length strings
749	++dec->cur;	828	++dec->cur;
750		829
		830	U8 major = *dec->cur & MAJOR_MISC;
		831
751	sv = newSVpvn ("", 0);	832	sv = newSVpvn ("", 0);
752		833
753	// not very fast, and certainly not robust against illegal input
754	for (;;)	834	for (;;)
755	{	835	{
756	WANT (1);	836	WANT (1);
757		837
758	if (*dec->cur == (0xe0 \| 31))	838	if ((*dec->cur - major) > LENGTH_EXT8)
		839	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
759	{	840	{
760	++dec->cur;	841	++dec->cur;
761	break;	842	break;
762	}	843	}
		844	else
		845	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
763		846
764	sv_catsv (sv, decode_sv (dec));	847	STRLEN len = decode_uint (dec);
		848
		849	WANT (len);
		850	sv_catpvn (sv, dec->cur, len);
		851	dec->cur += len;
765	}	852	}
766	}	853	}
767	else	854	else
768	{	855	{
769	STRLEN len = decode_uint (dec);	856	STRLEN len = decode_uint (dec);
…		…
776	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	863	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
777	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	864	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
778	}	865	}
779		866
780	if (utf8)	867	if (utf8)
		868	{
		869	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		870	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		871	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		872
781	SvUTF8_on (sv);	873	SvUTF8_on (sv);
		874	}
782		875
783	return sv;	876	return sv;
784		877
785	fail:	878	fail:
786	SvREFCNT_dec (sv);	879	SvREFCNT_dec (sv);
…		…
818	}	911	}
819	break;	912	break;
820		913
821	case CBOR_TAG_STRINGREF:	914	case CBOR_TAG_STRINGREF:
822	{	915	{
823	if ((*dec->cur >> 5) != 0)	916	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
824	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	917	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
825		918
826	UV idx = decode_uint (dec);	919	UV idx = decode_uint (dec);
827		920
828	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	921	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
835	case CBOR_TAG_VALUE_SHAREABLE:	928	case CBOR_TAG_VALUE_SHAREABLE:
836	{	929	{
837	if (ecb_expect_false (!dec->shareable))	930	if (ecb_expect_false (!dec->shareable))
838	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	931	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
839		932
		933	if (dec->cbor.flags & F_ALLOW_CYCLES)
		934	{
840	sv = newSV (0);	935	sv = newSV (0);
841	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	936	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
842		937
843	SV *osv = decode_sv (dec);	938	SV *osv = decode_sv (dec);
844	sv_setsv (sv, osv);	939	sv_setsv (sv, osv);
845	SvREFCNT_dec_NN (osv);	940	SvREFCNT_dec_NN (osv);
		941	}
		942	else
		943	{
		944	av_push (dec->shareable, &PL_sv_undef);
		945	int idx = AvFILLp (dec->shareable);
		946	sv = decode_sv (dec);
		947	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		948	}
846	}	949	}
847	break;	950	break;
848		951
849	case CBOR_TAG_VALUE_SHAREDREF:	952	case CBOR_TAG_VALUE_SHAREDREF:
850	{	953	{
851	if ((*dec->cur >> 5) != 0)	954	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
852	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	955	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
853		956
854	UV idx = decode_uint (dec);	957	UV idx = decode_uint (dec);
855		958
856	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	959	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
857	ERR ("corrupted CBOR data (sharedref index out of bounds)");	960	ERR ("corrupted CBOR data (sharedref index out of bounds)");
858		961
859	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	962	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		963
		964	if (sv == &PL_sv_undef)
		965	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
860	}	966	}
861	break;	967	break;
862		968
863	case CBOR_TAG_PERL_OBJECT:	969	case CBOR_TAG_PERL_OBJECT:
864	{	970	{
…		…
965	static SV *	1071	static SV *
966	decode_sv (dec_t *dec)	1072	decode_sv (dec_t *dec)
967	{	1073	{
968	WANT (1);	1074	WANT (1);
969		1075
970	switch (*dec->cur >> 5)	1076	switch (*dec->cur >> MAJOR_SHIFT)
971	{	1077	{
972	case 0: // unsigned int	1078	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
973	return newSVuv (decode_uint (dec));	1079	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
974	case 1: // negative int	1080	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
975	return newSViv (-1 - (IV)decode_uint (dec));	1081	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
976	case 2: // octet string	1082	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
977	return decode_str (dec, 0);	1083	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
978	case 3: // utf-8 string	1084	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
979	return decode_str (dec, 1);	1085
980	case 4: // array	1086	case MAJOR_MISC >> MAJOR_SHIFT:
981	return decode_av (dec);
982	case 5: // map
983	return decode_hv (dec);
984	case 6: // tag
985	return decode_tagged (dec);
986	case 7: // misc
987	switch (*dec->cur++ & 31)	1087	switch (*dec->cur++ & MINOR_MASK)
988	{	1088	{
989	case 20:	1089	case SIMPLE_FALSE:
990	#if CBOR_SLOW	1090	#if CBOR_SLOW
991	types_false = get_bool ("Types::Serialiser::false");	1091	types_false = get_bool ("Types::Serialiser::false");
992	#endif	1092	#endif
993	return newSVsv (types_false);	1093	return newSVsv (types_false);
994	case 21:	1094	case SIMPLE_TRUE:
995	#if CBOR_SLOW	1095	#if CBOR_SLOW
996	types_true = get_bool ("Types::Serialiser::true");	1096	types_true = get_bool ("Types::Serialiser::true");
997	#endif	1097	#endif
998	return newSVsv (types_true);	1098	return newSVsv (types_true);
999	case 22:	1099	case SIMPLE_NULL:
1000	return newSVsv (&PL_sv_undef);	1100	return newSVsv (&PL_sv_undef);
1001	case 23:	1101	case SIMPLE_UNDEF:
1002	#if CBOR_SLOW	1102	#if CBOR_SLOW
1003	types_error = get_bool ("Types::Serialiser::error");	1103	types_error = get_bool ("Types::Serialiser::error");
1004	#endif	1104	#endif
1005	return newSVsv (types_error);	1105	return newSVsv (types_error);
1006		1106
1007	case 25:	1107	case MISC_FLOAT16:
1008	{	1108	{
1009	WANT (2);	1109	WANT (2);
1010		1110
1011	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1111	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1012	dec->cur += 2;	1112	dec->cur += 2;
1013		1113
1014	return newSVnv (ecb_binary16_to_float (fp));	1114	return newSVnv (ecb_binary16_to_float (fp));
1015	}	1115	}
1016		1116
1017	case 26:	1117	case MISC_FLOAT32:
1018	{	1118	{
1019	uint32_t fp;	1119	uint32_t fp;
1020	WANT (4);	1120	WANT (4);
1021	memcpy (&fp, dec->cur, 4);	1121	memcpy (&fp, dec->cur, 4);
1022	dec->cur += 4;	1122	dec->cur += 4;
…		…
1025	fp = ecb_bswap32 (fp);	1125	fp = ecb_bswap32 (fp);
1026		1126
1027	return newSVnv (ecb_binary32_to_float (fp));	1127	return newSVnv (ecb_binary32_to_float (fp));
1028	}	1128	}
1029		1129
1030	case 27:	1130	case MISC_FLOAT64:
1031	{	1131	{
1032	uint64_t fp;	1132	uint64_t fp;
1033	WANT (8);	1133	WANT (8);
1034	memcpy (&fp, dec->cur, 8);	1134	memcpy (&fp, dec->cur, 8);
1035	dec->cur += 8;	1135	dec->cur += 8;
…		…
1038	fp = ecb_bswap64 (fp);	1138	fp = ecb_bswap64 (fp);
1039		1139
1040	return newSVnv (ecb_binary64_to_double (fp));	1140	return newSVnv (ecb_binary64_to_double (fp));
1041	}	1141	}
1042		1142
1043	// 0..19 unassigned	1143	// 0..19 unassigned simple
1044	// 24 reserved + unassigned (reserved values are not encodable)	1144	// 24 reserved + unassigned simple (reserved values are not encodable)
		1145	// 28-30 unassigned misc
		1146	// 31 break code
1045	default:	1147	default:
1046	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1148	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1047	}	1149	}
1048		1150
1049	break;	1151	break;
1050	}	1152	}
1051		1153
…		…
1078	if (dec.cur != dec.end && !dec.err)	1180	if (dec.cur != dec.end && !dec.err)
1079	dec.err = "garbage after CBOR object";	1181	dec.err = "garbage after CBOR object";
1080		1182
1081	if (dec.err)	1183	if (dec.err)
1082	{	1184	{
		1185	if (dec.shareable)
		1186	{
		1187	// need to break cyclic links, which whould all be in shareable
		1188	int i;
		1189	SV **svp;
		1190
		1191	for (i = av_len (dec.shareable) + 1; i--; )
		1192	if ((svp = av_fetch (dec.shareable, i, 0)))
		1193	sv_setsv (*svp, &PL_sv_undef);
		1194	}
		1195
1083	SvREFCNT_dec (sv);	1196	SvREFCNT_dec (sv);
1084	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1197	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1085	}	1198	}
1086		1199
1087	sv = sv_2mortal (sv);	1200	sv = sv_2mortal (sv);
1088		1201
1089	return sv;	1202	return sv;
1090	}	1203	}
1091		1204
		1205	/////////////////////////////////////////////////////////////////////////////
		1206	// incremental parser
		1207
		1208	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1209
		1210	// returns 0 for notyet, 1 for success or error
		1211	static int
		1212	incr_parse (CBOR self, SV cborstr)
		1213	{
		1214	STRLEN cur;
		1215	SvPV (cborstr, cur);
		1216
		1217	while (ecb_expect_true (self->incr_need <= cur))
		1218	{
		1219	// table of integer count bytes
		1220	static I8 incr_len[MINOR_MASK + 1] = {
		1221	0, 0, 0, 0, 0, 0, 0, 0,
		1222	0, 0, 0, 0, 0, 0, 0, 0,
		1223	0, 0, 0, 0, 0, 0, 0, 0,
		1224	1, 2, 4, 8,-1,-1,-1,-2
		1225	};
		1226
		1227	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1228	U8 m = *p & MINOR_MASK;
		1229	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1230	I8 ilen = incr_len[m];
		1231
		1232	self->incr_need = self->incr_pos + 1;
		1233
		1234	if (ecb_expect_false (ilen < 0))
		1235	{
		1236	if (m != MINOR_INDEF)
		1237	return 1; // error
		1238
		1239	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1240	{
		1241	if (count >= 0)
		1242	return 1; // error
		1243
		1244	count = 1;
		1245	}
		1246	else
		1247	{
		1248	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1249	count = -1;
		1250	}
		1251	}
		1252	else
		1253	{
		1254	self->incr_need += ilen;
		1255	if (ecb_expect_false (self->incr_need > cur))
		1256	return 0;
		1257
		1258	int major = *p >> MAJOR_SHIFT;
		1259
		1260	switch (major)
		1261	{
		1262	case MAJOR_BYTES >> MAJOR_SHIFT:
		1263	case MAJOR_TEXT >> MAJOR_SHIFT:
		1264	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1265	case MAJOR_MAP >> MAJOR_SHIFT:
		1266	{
		1267	UV len;
		1268
		1269	if (ecb_expect_false (ilen))
		1270	{
		1271	len = 0;
		1272
		1273	do {
		1274	len = (len << 8) \| *++p;
		1275	} while (--ilen);
		1276	}
		1277	else
		1278	len = m;
		1279
		1280	switch (major)
		1281	{
		1282	case MAJOR_BYTES >> MAJOR_SHIFT:
		1283	case MAJOR_TEXT >> MAJOR_SHIFT:
		1284	self->incr_need += len;
		1285	if (ecb_expect_false (self->incr_need > cur))
		1286	return 0;
		1287
		1288	break;
		1289
		1290	case MAJOR_MAP >> MAJOR_SHIFT:
		1291	len <<= 1;
		1292	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1293	if (len)
		1294	{
		1295	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1296	count = len + 1;
		1297	}
		1298	break;
		1299	}
		1300	}
		1301	}
		1302	}
		1303
		1304	self->incr_pos = self->incr_need;
		1305
		1306	if (count > 0)
		1307	{
		1308	while (!--count)
		1309	{
		1310	if (!AvFILLp (self->incr_count))
		1311	return 1; // done
		1312
		1313	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1314	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1315	}
		1316
		1317	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1318	}
		1319	}
		1320
		1321	return 0;
		1322	}
		1323
		1324
1092	/////////////////////////////////////////////////////////////////////////////	1325	/////////////////////////////////////////////////////////////////////////////
1093	// XS interface functions	1326	// XS interface functions
1094		1327
1095	MODULE = CBOR::XS PACKAGE = CBOR::XS	1328	MODULE = CBOR::XS PACKAGE = CBOR::XS
1096		1329
…		…
1136	void shrink (CBOR *self, int enable = 1)	1369	void shrink (CBOR *self, int enable = 1)
1137	ALIAS:	1370	ALIAS:
1138	shrink = F_SHRINK	1371	shrink = F_SHRINK
1139	allow_unknown = F_ALLOW_UNKNOWN	1372	allow_unknown = F_ALLOW_UNKNOWN
1140	allow_sharing = F_ALLOW_SHARING	1373	allow_sharing = F_ALLOW_SHARING
1141	allow_stringref = F_ALLOW_STRINGREF	1374	allow_cycles = F_ALLOW_CYCLES
		1375	pack_strings = F_PACK_STRINGS
		1376	validate_utf8 = F_VALIDATE_UTF8
1142	PPCODE:	1377	PPCODE:
1143	{	1378	{
1144	if (enable)	1379	if (enable)
1145	self->flags \|= ix;	1380	self->flags \|= ix;
1146	else	1381	else
…		…
1152	void get_shrink (CBOR *self)	1387	void get_shrink (CBOR *self)
1153	ALIAS:	1388	ALIAS:
1154	get_shrink = F_SHRINK	1389	get_shrink = F_SHRINK
1155	get_allow_unknown = F_ALLOW_UNKNOWN	1390	get_allow_unknown = F_ALLOW_UNKNOWN
1156	get_allow_sharing = F_ALLOW_SHARING	1391	get_allow_sharing = F_ALLOW_SHARING
1157	get_allow_stringref = F_ALLOW_STRINGREF	1392	get_allow_cycles = F_ALLOW_CYCLES
		1393	get_pack_strings = F_PACK_STRINGS
		1394	get_validate_utf8 = F_VALIDATE_UTF8
1158	PPCODE:	1395	PPCODE:
1159	XPUSHs (boolSV (self->flags & ix));	1396	XPUSHs (boolSV (self->flags & ix));
1160		1397
1161	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1398	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1162	PPCODE:	1399	PPCODE:
…		…
1211	EXTEND (SP, 2);	1448	EXTEND (SP, 2);
1212	PUSHs (sv);	1449	PUSHs (sv);
1213	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1450	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1214	}	1451	}
1215		1452
		1453	void incr_parse (CBOR self, SV cborstr)
		1454	ALIAS:
		1455	incr_parse_multiple = 1
		1456	PPCODE:
		1457	{
		1458	if (SvUTF8 (cborstr))
		1459	sv_utf8_downgrade (cborstr, 0);
		1460
		1461	if (!self->incr_count)
		1462	{
		1463	self->incr_count = newAV ();
		1464	self->incr_pos = 0;
		1465	self->incr_need = 1;
		1466
		1467	av_push (self->incr_count, newSViv (1));
		1468	}
		1469
		1470	do
		1471	{
		1472	if (!incr_parse (self, cborstr))
		1473	{
		1474	if (self->incr_need > self->max_size && self->max_size)
		1475	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1476	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1477
		1478	break;
		1479	}
		1480
		1481	SV *sv;
		1482	char *offset;
		1483
		1484	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1485	XPUSHs (sv);
		1486
		1487	sv_chop (cborstr, offset);
		1488
		1489	av_clear (self->incr_count);
		1490	av_push (self->incr_count, newSViv (1));
		1491
		1492	self->incr_pos = 0;
		1493	self->incr_need = self->incr_pos + 1;
		1494	}
		1495	while (ix);
		1496	}
		1497
		1498	void incr_reset (CBOR *self)
		1499	CODE:
		1500	{
		1501	SvREFCNT_dec (self->incr_count);
		1502	self->incr_count = 0;
		1503	}
		1504
1216	void DESTROY (CBOR *self)	1505	void DESTROY (CBOR *self)
1217	PPCODE:	1506	PPCODE:
1218	cbor_free (self);	1507	cbor_free (self);
1219		1508
1220	PROTOTYPES: ENABLE	1509	PROTOTYPES: ENABLE
1221		1510
1222	void encode_cbor (SV *scalar)	1511	void encode_cbor (SV *scalar)
		1512	ALIAS:
		1513	encode_cbor = 0
		1514	encode_cbor_sharing = F_ALLOW_SHARING
1223	PPCODE:	1515	PPCODE:
1224	{	1516	{
1225	CBOR cbor;	1517	CBOR cbor;
1226	cbor_init (&cbor);	1518	cbor_init (&cbor);
		1519	cbor.flags \|= ix;
1227	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1520	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1228	XPUSHs (scalar);	1521	XPUSHs (scalar);
1229	}	1522	}
1230		1523
1231	void decode_cbor (SV *cborstr)	1524	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs. Revision 1.44 by root, Tue Feb 18 22:12:12 2014 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs.
Revision 1.44 by root, Tue Feb 18 22:12:12 2014 UTC