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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.28 by root, Sat Nov 23 18:30:59 2013 UTC vs.
Revision 1.45 by root, Sun Dec 14 05:48:39 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))
…		…
23	#endif	24	#endif
24	#ifndef SvREFCNT_dec_NN	25	#ifndef SvREFCNT_dec_NN
25	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)	26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26	#endif	27	#endif
27		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	};
		65
28	// known tags	66	// known tags
29	enum cbor_tag	67	enum cbor_tag
30	{	68	{
31	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
32	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
33	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
34	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
35	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
36	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
37	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
38	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
39		77
40	// rfc7049	78	// rfc7049
41	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
42	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
43	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
44	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
45	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
46	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
47		85
48	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
49	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
50	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
51	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
52		90
53	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
54	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
55	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
56	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
57	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
58		96
59	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
60	};	98	};
61		99
62	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
65	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	103	#define F_ALLOW_CYCLES 0x00000008UL
		104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_VALIDATE_UTF8 0x00000020UL
66		106
67	#define INIT_SIZE 32 // initial scalar size to be allocated	107	#define INIT_SIZE 32 // initial scalar size to be allocated
68		108
69	#define SB do {	109	#define SB do {
70	#define SE } while (0)	110	#define SE } while (0)
…		…
89	typedef struct {	129	typedef struct {
90	U32 flags;	130	U32 flags;
91	U32 max_depth;	131	U32 max_depth;
92	STRLEN max_size;	132	STRLEN max_size;
93	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.
94	} CBOR;	139	} CBOR;
95		140
96	ecb_inline void	141	ecb_inline void
97	cbor_init (CBOR *cbor)	142	cbor_init (CBOR *cbor)
98	{	143	{
…		…
102		147
103	ecb_inline void	148	ecb_inline void
104	cbor_free (CBOR *cbor)	149	cbor_free (CBOR *cbor)
105	{	150	{
106	SvREFCNT_dec (cbor->filter);	151	SvREFCNT_dec (cbor->filter);
		152	SvREFCNT_dec (cbor->incr_count);
107	}	153	}
108		154
109	/////////////////////////////////////////////////////////////////////////////	155	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	156	// utility functions
111		157
…		…
141	{	187	{
142	return idx > 23	188	return idx > 23
143	? idx > 0xffU	189	? idx > 0xffU
144	? idx > 0xffffU	190	? idx > 0xffffU
145	? idx > 0xffffffffU	191	? idx > 0xffffffffU
		192	? 11
146	? 7	193	: 7
147	: 6
148	: 5	194	: 5
149	: 4	195	: 4
150	: 3;	196	: 3;
151	}	197	}
152		198
…		…
189	static void	235	static void
190	encode_uint (enc_t *enc, int major, UV len)	236	encode_uint (enc_t *enc, int major, UV len)
191	{	237	{
192	need (enc, 9);	238	need (enc, 9);
193		239
194	if (len < 24)	240	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major \| len;	241	*enc->cur++ = major \| len;
196	else if (len <= 0xff)	242	else if (ecb_expect_true (len <= 0xffU))
197	{	243	{
198	*enc->cur++ = major \| 24;	244	*enc->cur++ = major \| LENGTH_EXT1;
199	*enc->cur++ = len;	245	*enc->cur++ = len;
200	}	246	}
201	else if (len <= 0xffff)	247	else if (len <= 0xffffU)
202	{	248	{
203	*enc->cur++ = major \| 25;	249	*enc->cur++ = major \| LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	250	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	251	*enc->cur++ = len;
206	}	252	}
207	else if (len <= 0xffffffff)	253	else if (len <= 0xffffffffU)
208	{	254	{
209	*enc->cur++ = major \| 26;	255	*enc->cur++ = major \| LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	256	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	257	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	258	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	259	*enc->cur++ = len;
214	}	260	}
215	else	261	else
216	{	262	{
217	*enc->cur++ = major \| 27;	263	*enc->cur++ = major \| LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	264	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	265	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	266	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	267	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	268	*enc->cur++ = len >> 24;
…		…
227	}	273	}
228		274
229	ecb_inline void	275	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	276	encode_tag (enc_t *enc, UV tag)
231	{	277	{
232	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;
233	}	288	}
234		289
235	static void	290	static void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	291	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
237	{	292	{
238	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	293	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
239	{	294	{
240	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	295	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
241		296
242	if (SvOK (*svp))	297	if (SvOK (*svp))
243	{	298	{
244	// already registered, use stringref	299	// already registered, use stringref
245	encode_tag (enc, CBOR_TAG_STRINGREF);	300	encode_tag (enc, CBOR_TAG_STRINGREF);
246	encode_uint (enc, 0x00, SvUV (*svp));	301	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
247	return;	302	return;
248	}	303	}
249	else if (len >= minimum_string_length (enc->stringref_idx))	304	else if (len >= minimum_string_length (enc->stringref_idx))
250	{	305	{
251	// register only	306	// register only
252	sv_setuv (*svp, enc->stringref_idx);	307	sv_setuv (*svp, enc->stringref_idx);
253	++enc->stringref_idx;	308	++enc->stringref_idx;
254	}	309	}
255	}	310	}
256		311
257	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	312	encode_str (enc, utf8, str, len);
258	need (enc, len);
259	memcpy (enc->cur, str, len);
260	enc->cur += len;
261	}	313	}
262		314
263	static void encode_sv (enc_t enc, SV sv);	315	static void encode_sv (enc_t enc, SV sv);
264		316
265	static void	317	static void
…		…
270	if (enc->depth >= enc->cbor.max_depth)	322	if (enc->depth >= enc->cbor.max_depth)
271	croak (ERR_NESTING_EXCEEDED);	323	croak (ERR_NESTING_EXCEEDED);
272		324
273	++enc->depth;	325	++enc->depth;
274		326
275	encode_uint (enc, 0x80, len + 1);	327	encode_uint (enc, MAJOR_ARRAY, len + 1);
276		328
		329	if (SvMAGICAL (av))
277	for (i = 0; i <= len; ++i)	330	for (i = 0; i <= len; ++i)
278	{	331	{
279	SV **svp = av_fetch (av, i, 0);	332	SV **svp = av_fetch (av, i, 0);
280	encode_sv (enc, svp ? *svp : &PL_sv_undef);	333	encode_sv (enc, svp ? *svp : &PL_sv_undef);
281	}	334	}
		335	else
		336	for (i = 0; i <= len; ++i)
		337	encode_sv (enc, AvARRAY (av)[i]);
282		338
283	--enc->depth;	339	--enc->depth;
284	}	340	}
285		341
286	static void	342	static void
…		…
295		351
296	int pairs = hv_iterinit (hv);	352	int pairs = hv_iterinit (hv);
297	int mg = SvMAGICAL (hv);	353	int mg = SvMAGICAL (hv);
298		354
299	if (mg)	355	if (mg)
300	encode_ch (enc, 0xa0 \| 31);	356	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
301	else	357	else
302	encode_uint (enc, 0xa0, pairs);	358	encode_uint (enc, MAJOR_MAP, pairs);
303		359
304	while ((he = hv_iternext (hv)))	360	while ((he = hv_iternext (hv)))
305	{	361	{
306	if (HeKLEN (he) == HEf_SVKEY)	362	if (HeKLEN (he) == HEf_SVKEY)
307	encode_sv (enc, HeSVKEY (he));	363	encode_sv (enc, HeSVKEY (he));
308	else	364	else
309	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	365	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
310		366
311	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	367	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
312	}	368	}
313		369
314	if (mg)	370	if (mg)
315	encode_ch (enc, 0xe0 \| 31);	371	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
316		372
317	--enc->depth;	373	--enc->depth;
318	}	374	}
319		375
320	// encode objects, arrays and special \0=false and \1=true values.	376	// encode objects, arrays and special \0=false and \1=true values.
321	static void	377	static void
322	encode_rv (enc_t enc, SV sv)	378	encode_rv (enc_t enc, SV sv)
323	{	379	{
324	SvGETMAGIC (sv);	380	SvGETMAGIC (sv);
325
326	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
327	&& ecb_expect_false (SvREFCNT (sv) > 1))
328	{
329	if (!enc->shareable)
330	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
331
332	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
333
334	if (SvOK (*svp))
335	{
336	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
337	encode_uint (enc, 0x00, SvUV (*svp));
338	return;
339	}
340	else
341	{
342	sv_setuv (*svp, enc->shareable_idx);
343	++enc->shareable_idx;
344	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
345	}
346	}
347		381
348	svtype svt = SvTYPE (sv);	382	svtype svt = SvTYPE (sv);
349		383
350	if (ecb_expect_false (SvOBJECT (sv)))	384	if (ecb_expect_false (SvOBJECT (sv)))
351	{	385	{
…		…
358	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	392	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
359	? cbor_tagged_stash	393	? cbor_tagged_stash
360	: gv_stashpv ("CBOR::XS::Tagged" , 1);	394	: gv_stashpv ("CBOR::XS::Tagged" , 1);
361		395
362	HV *stash = SvSTASH (sv);	396	HV *stash = SvSTASH (sv);
363	GV *method;
364		397
365	if (stash == boolean_stash)	398	if (stash == boolean_stash)
366	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	399	{
		400	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		401	return;
		402	}
367	else if (stash == error_stash)	403	else if (stash == error_stash)
368	encode_ch (enc, 0xe0 \| 23);	404	{
		405	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		406	return;
		407	}
369	else if (stash == tagged_stash)	408	else if (stash == tagged_stash)
370	{	409	{
371	if (svt != SVt_PVAV)	410	if (svt != SVt_PVAV)
372	croak ("encountered CBOR::XS::Tagged object that isn't an array");	411	croak ("encountered CBOR::XS::Tagged object that isn't an array");
373		412
374	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	413	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
375	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	414	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		415
		416	return;
		417	}
		418	}
		419
		420	if (ecb_expect_false (SvREFCNT (sv) > 1)
		421	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		422	{
		423	if (!enc->shareable)
		424	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		425
		426	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		427
		428	if (SvOK (*svp))
376	}	429	{
		430	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		431	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		432	return;
		433	}
		434	else
		435	{
		436	sv_setuv (*svp, enc->shareable_idx);
		437	++enc->shareable_idx;
		438	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		439	}
		440	}
		441
		442	if (ecb_expect_false (SvOBJECT (sv)))
		443	{
		444	HV *stash = SvSTASH (sv);
		445	GV *method;
		446
377	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	447	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
378	{	448	{
379	dSP;	449	dSP;
380		450
381	ENTER; SAVETMPS; PUSHMARK (SP);	451	ENTER; SAVETMPS; PUSHMARK (SP);
382	// we re-bless the reference to get overload and other niceties right	452	// we re-bless the reference to get overload and other niceties right
…		…
414	// catch this surprisingly common error	484	// catch this surprisingly common error
415	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	485	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
416	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	486	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
417		487
418	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	488	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
419	encode_uint (enc, 0x80, count + 1);	489	encode_uint (enc, MAJOR_ARRAY, count + 1);
420	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	490	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
421		491
422	while (count)	492	while (count)
423	encode_sv (enc, SP[1 - count--]);	493	encode_sv (enc, SP[1 - count--]);
424		494
425	PUTBACK;	495	PUTBACK;
…		…
446	{	516	{
447	double nv = SvNVX (sv);	517	double nv = SvNVX (sv);
448		518
449	need (enc, 9);	519	need (enc, 9);
450		520
451	if (ecb_expect_false (nv == (U32)nv))	521	if (ecb_expect_false (nv == (NV)(U32)nv))
452	encode_uint (enc, 0x00, (U32)nv);	522	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
453	//TODO: maybe I32?	523	//TODO: maybe I32?
454	else if (ecb_expect_false (nv == (float)nv))	524	else if (ecb_expect_false (nv == (float)nv))
455	{	525	{
456	uint32_t fp = ecb_float_to_binary32 (nv);	526	uint32_t fp = ecb_float_to_binary32 (nv);
457		527
458	*enc->cur++ = 0xe0 \| 26;	528	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
459		529
460	if (!ecb_big_endian ())	530	if (!ecb_big_endian ())
461	fp = ecb_bswap32 (fp);	531	fp = ecb_bswap32 (fp);
462		532
463	memcpy (enc->cur, &fp, 4);	533	memcpy (enc->cur, &fp, 4);
…		…
465	}	535	}
466	else	536	else
467	{	537	{
468	uint64_t fp = ecb_double_to_binary64 (nv);	538	uint64_t fp = ecb_double_to_binary64 (nv);
469		539
470	*enc->cur++ = 0xe0 \| 27;	540	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
471		541
472	if (!ecb_big_endian ())	542	if (!ecb_big_endian ())
473	fp = ecb_bswap64 (fp);	543	fp = ecb_bswap64 (fp);
474		544
475	memcpy (enc->cur, &fp, 8);	545	memcpy (enc->cur, &fp, 8);
…		…
484		554
485	if (SvPOKp (sv))	555	if (SvPOKp (sv))
486	{	556	{
487	STRLEN len;	557	STRLEN len;
488	char *str = SvPV (sv, len);	558	char *str = SvPV (sv, len);
489	encode_str (enc, SvUTF8 (sv), str, len);	559	encode_strref (enc, SvUTF8 (sv), str, len);
490	}	560	}
491	else if (SvNOKp (sv))	561	else if (SvNOKp (sv))
492	encode_nv (enc, sv);	562	encode_nv (enc, sv);
493	else if (SvIOKp (sv))	563	else if (SvIOKp (sv))
494	{	564	{
495	if (SvIsUV (sv))	565	if (SvIsUV (sv))
496	encode_uint (enc, 0x00, SvUVX (sv));	566	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
497	else if (SvIVX (sv) >= 0)	567	else if (SvIVX (sv) >= 0)
498	encode_uint (enc, 0x00, SvIVX (sv));	568	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
499	else	569	else
500	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	570	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
501	}	571	}
502	else if (SvROK (sv))	572	else if (SvROK (sv))
503	encode_rv (enc, SvRV (sv));	573	encode_rv (enc, SvRV (sv));
504	else if (!SvOK (sv))	574	else if (!SvOK (sv))
505	encode_ch (enc, 0xe0 \| 22);	575	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
506	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	576	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
507	encode_ch (enc, 0xe0 \| 23);	577	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
508	else	578	else
509	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	579	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
510	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	580	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
511	}	581	}
512		582
…		…
520	enc.cur = SvPVX (enc.sv);	590	enc.cur = SvPVX (enc.sv);
521	enc.end = SvEND (enc.sv);	591	enc.end = SvEND (enc.sv);
522		592
523	SvPOK_only (enc.sv);	593	SvPOK_only (enc.sv);
524		594
525	if (cbor->flags & F_ALLOW_STRINGREF)	595	if (cbor->flags & F_PACK_STRINGS)
526	{	596	{
527	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	597	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
528	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	598	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
529	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	599	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
530	}	600	}
…		…
565	#define DEC_DEC_DEPTH --dec->depth	635	#define DEC_DEC_DEPTH --dec->depth
566		636
567	static UV	637	static UV
568	decode_uint (dec_t *dec)	638	decode_uint (dec_t *dec)
569	{	639	{
570	switch (*dec->cur & 31)	640	U8 m = *dec->cur & MINOR_MASK;
571	{	641	++dec->cur;
572	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
573	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
574	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
575	return *dec->cur++ & 31;
576		642
577	case 24:	643	if (ecb_expect_true (m < LENGTH_EXT1))
		644	return m;
		645	else if (ecb_expect_true (m == LENGTH_EXT1))
		646	{
578	WANT (2);	647	WANT (1);
579	dec->cur += 2;	648	dec->cur += 1;
580	return dec->cur[-1];	649	return dec->cur[-1];
581		650	}
582	case 25:	651	else if (ecb_expect_true (m == LENGTH_EXT2))
		652	{
583	WANT (3);	653	WANT (2);
584	dec->cur += 3;	654	dec->cur += 2;
585	return (((UV)dec->cur[-2]) << 8)	655	return (((UV)dec->cur[-2]) << 8)
586	\| ((UV)dec->cur[-1]);	656	\| ((UV)dec->cur[-1]);
587		657	}
588	case 26:	658	else if (ecb_expect_true (m == LENGTH_EXT4))
		659	{
589	WANT (5);	660	WANT (4);
590	dec->cur += 5;	661	dec->cur += 4;
591	return (((UV)dec->cur[-4]) << 24)	662	return (((UV)dec->cur[-4]) << 24)
592	\| (((UV)dec->cur[-3]) << 16)	663	\| (((UV)dec->cur[-3]) << 16)
593	\| (((UV)dec->cur[-2]) << 8)	664	\| (((UV)dec->cur[-2]) << 8)
594	\| ((UV)dec->cur[-1]);	665	\| ((UV)dec->cur[-1]);
595		666	}
596	case 27:	667	else if (ecb_expect_true (m == LENGTH_EXT8))
		668	{
597	WANT (9);	669	WANT (8);
598	dec->cur += 9;	670	dec->cur += 8;
		671
		672	return
		673	#if UVSIZE < 8
		674	0
		675	#else
599	return (((UV)dec->cur[-8]) << 56)	676	(((UV)dec->cur[-8]) << 56)
600	\| (((UV)dec->cur[-7]) << 48)	677	\| (((UV)dec->cur[-7]) << 48)
601	\| (((UV)dec->cur[-6]) << 40)	678	\| (((UV)dec->cur[-6]) << 40)
602	\| (((UV)dec->cur[-5]) << 32)	679	\| (((UV)dec->cur[-5]) << 32)
		680	#endif
603	\| (((UV)dec->cur[-4]) << 24)	681	\| (((UV)dec->cur[-4]) << 24)
604	\| (((UV)dec->cur[-3]) << 16)	682	\| (((UV)dec->cur[-3]) << 16)
605	\| (((UV)dec->cur[-2]) << 8)	683	\| (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	684	\| ((UV)dec->cur[-1]);
607		685	}
608	default:	686	else
609	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	687	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
610	}
611		688
612	fail:	689	fail:
613	return 0;	690	return 0;
614	}	691	}
615		692
…		…
620	{	697	{
621	AV *av = newAV ();	698	AV *av = newAV ();
622		699
623	DEC_INC_DEPTH;	700	DEC_INC_DEPTH;
624		701
625	if ((*dec->cur & 31) == 31)	702	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
626	{	703	{
627	++dec->cur;	704	++dec->cur;
628		705
629	for (;;)	706	for (;;)
630	{	707	{
631	WANT (1);	708	WANT (1);
632		709
633	if (*dec->cur == (0xe0 \| 31))	710	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
634	{	711	{
635	++dec->cur;	712	++dec->cur;
636	break;	713	break;
637	}	714	}
638		715
…		…
641	}	718	}
642	else	719	else
643	{	720	{
644	int i, len = decode_uint (dec);	721	int i, len = decode_uint (dec);
645		722
		723	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
646	av_fill (av, len - 1);	724	av_fill (av, len - 1);
647		725
648	for (i = 0; i < len; ++i)	726	for (i = 0; i < len; ++i)
649	AvARRAY (av)[i] = decode_sv (dec);	727	AvARRAY (av)[i] = decode_sv (dec);
650	}	728	}
…		…
663	{	741	{
664	// for speed reasons, we specialcase single-string	742	// for speed reasons, we specialcase single-string
665	// byte or utf-8 strings as keys, but only when !stringref	743	// byte or utf-8 strings as keys, but only when !stringref
666		744
667	if (ecb_expect_true (!dec->stringref))	745	if (ecb_expect_true (!dec->stringref))
668	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	746	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
669	{	747	{
670	I32 len = decode_uint (dec);	748	I32 len = decode_uint (dec);
671	char key = (char )dec->cur;	749	char key = (char )dec->cur;
672		750
673	dec->cur += len;	751	dec->cur += len;
674		752
675	if (ecb_expect_false (dec->stringref))
676	av_push (dec->stringref, newSVpvn (key, len));
677
678	hv_store (hv, key, len, decode_sv (dec), 0);	753	hv_store (hv, key, len, decode_sv (dec), 0);
679		754
680	return;	755	return;
681	}	756	}
682	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	757	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
683	{	758	{
684	I32 len = decode_uint (dec);	759	I32 len = decode_uint (dec);
685	char key = (char )dec->cur;	760	char key = (char )dec->cur;
686		761
687	dec->cur += len;	762	dec->cur += len;
688		763
689	if (ecb_expect_false (dec->stringref))	764	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
690	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	765	if (!is_utf8_string (key, len))
		766	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
691		767
692	hv_store (hv, key, -len, decode_sv (dec), 0);	768	hv_store (hv, key, -len, decode_sv (dec), 0);
693		769
694	return;	770	return;
695	}	771	}
…		…
697	SV *k = decode_sv (dec);	773	SV *k = decode_sv (dec);
698	SV *v = decode_sv (dec);	774	SV *v = decode_sv (dec);
699		775
700	hv_store_ent (hv, k, v, 0);	776	hv_store_ent (hv, k, v, 0);
701	SvREFCNT_dec (k);	777	SvREFCNT_dec (k);
		778
		779	fail:
		780	;
702	}	781	}
703		782
704	static SV *	783	static SV *
705	decode_hv (dec_t *dec)	784	decode_hv (dec_t *dec)
706	{	785	{
707	HV *hv = newHV ();	786	HV *hv = newHV ();
708		787
709	DEC_INC_DEPTH;	788	DEC_INC_DEPTH;
710		789
711	if ((*dec->cur & 31) == 31)	790	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
712	{	791	{
713	++dec->cur;	792	++dec->cur;
714		793
715	for (;;)	794	for (;;)
716	{	795	{
717	WANT (1);	796	WANT (1);
718		797
719	if (*dec->cur == (0xe0 \| 31))	798	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
720	{	799	{
721	++dec->cur;	800	++dec->cur;
722	break;	801	break;
723	}	802	}
724		803
…		…
745	static SV *	824	static SV *
746	decode_str (dec_t *dec, int utf8)	825	decode_str (dec_t *dec, int utf8)
747	{	826	{
748	SV *sv = 0;	827	SV *sv = 0;
749		828
750	if ((*dec->cur & 31) == 31)	829	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
751	{	830	{
		831	// indefinite length strings
752	++dec->cur;	832	++dec->cur;
753		833
		834	U8 major = *dec->cur & MAJOR_MISC;
		835
754	sv = newSVpvn ("", 0);	836	sv = newSVpvn ("", 0);
755		837
756	// not very fast, and certainly not robust against illegal input
757	for (;;)	838	for (;;)
758	{	839	{
759	WANT (1);	840	WANT (1);
760		841
761	if (*dec->cur == (0xe0 \| 31))	842	if ((*dec->cur - major) > LENGTH_EXT8)
		843	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
762	{	844	{
763	++dec->cur;	845	++dec->cur;
764	break;	846	break;
765	}	847	}
		848	else
		849	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
766		850
767	sv_catsv (sv, decode_sv (dec));	851	STRLEN len = decode_uint (dec);
		852
		853	WANT (len);
		854	sv_catpvn (sv, dec->cur, len);
		855	dec->cur += len;
768	}	856	}
769	}	857	}
770	else	858	else
771	{	859	{
772	STRLEN len = decode_uint (dec);	860	STRLEN len = decode_uint (dec);
…		…
779	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	867	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
780	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	868	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
781	}	869	}
782		870
783	if (utf8)	871	if (utf8)
		872	{
		873	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		874	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		875	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		876
784	SvUTF8_on (sv);	877	SvUTF8_on (sv);
		878	}
785		879
786	return sv;	880	return sv;
787		881
788	fail:	882	fail:
789	SvREFCNT_dec (sv);	883	SvREFCNT_dec (sv);
…		…
821	}	915	}
822	break;	916	break;
823		917
824	case CBOR_TAG_STRINGREF:	918	case CBOR_TAG_STRINGREF:
825	{	919	{
826	if ((*dec->cur >> 5) != 0)	920	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
827	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	921	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
828		922
829	UV idx = decode_uint (dec);	923	UV idx = decode_uint (dec);
830		924
831	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	925	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
838	case CBOR_TAG_VALUE_SHAREABLE:	932	case CBOR_TAG_VALUE_SHAREABLE:
839	{	933	{
840	if (ecb_expect_false (!dec->shareable))	934	if (ecb_expect_false (!dec->shareable))
841	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	935	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
842		936
		937	if (dec->cbor.flags & F_ALLOW_CYCLES)
		938	{
843	sv = newSV (0);	939	sv = newSV (0);
844	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	940	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
845		941
846	SV *osv = decode_sv (dec);	942	SV *osv = decode_sv (dec);
847	sv_setsv (sv, osv);	943	sv_setsv (sv, osv);
848	SvREFCNT_dec_NN (osv);	944	SvREFCNT_dec_NN (osv);
		945	}
		946	else
		947	{
		948	av_push (dec->shareable, &PL_sv_undef);
		949	int idx = AvFILLp (dec->shareable);
		950	sv = decode_sv (dec);
		951	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		952	}
849	}	953	}
850	break;	954	break;
851		955
852	case CBOR_TAG_VALUE_SHAREDREF:	956	case CBOR_TAG_VALUE_SHAREDREF:
853	{	957	{
854	if ((*dec->cur >> 5) != 0)	958	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
855	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	959	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
856		960
857	UV idx = decode_uint (dec);	961	UV idx = decode_uint (dec);
858		962
859	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	963	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
860	ERR ("corrupted CBOR data (sharedref index out of bounds)");	964	ERR ("corrupted CBOR data (sharedref index out of bounds)");
861		965
862	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	966	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		967
		968	if (sv == &PL_sv_undef)
		969	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
863	}	970	}
864	break;	971	break;
865		972
866	case CBOR_TAG_PERL_OBJECT:	973	case CBOR_TAG_PERL_OBJECT:
867	{	974	{
…		…
968	static SV *	1075	static SV *
969	decode_sv (dec_t *dec)	1076	decode_sv (dec_t *dec)
970	{	1077	{
971	WANT (1);	1078	WANT (1);
972		1079
973	switch (*dec->cur >> 5)	1080	switch (*dec->cur >> MAJOR_SHIFT)
974	{	1081	{
975	case 0: // unsigned int	1082	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
976	return newSVuv (decode_uint (dec));	1083	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
977	case 1: // negative int	1084	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
978	return newSViv (-1 - (IV)decode_uint (dec));	1085	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
979	case 2: // octet string	1086	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
980	return decode_str (dec, 0);	1087	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
981	case 3: // utf-8 string	1088	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
982	return decode_str (dec, 1);	1089
983	case 4: // array	1090	case MAJOR_MISC >> MAJOR_SHIFT:
984	return decode_av (dec);
985	case 5: // map
986	return decode_hv (dec);
987	case 6: // tag
988	return decode_tagged (dec);
989	case 7: // misc
990	switch (*dec->cur++ & 31)	1091	switch (*dec->cur++ & MINOR_MASK)
991	{	1092	{
992	case 20:	1093	case SIMPLE_FALSE:
993	#if CBOR_SLOW	1094	#if CBOR_SLOW
994	types_false = get_bool ("Types::Serialiser::false");	1095	types_false = get_bool ("Types::Serialiser::false");
995	#endif	1096	#endif
996	return newSVsv (types_false);	1097	return newSVsv (types_false);
997	case 21:	1098	case SIMPLE_TRUE:
998	#if CBOR_SLOW	1099	#if CBOR_SLOW
999	types_true = get_bool ("Types::Serialiser::true");	1100	types_true = get_bool ("Types::Serialiser::true");
1000	#endif	1101	#endif
1001	return newSVsv (types_true);	1102	return newSVsv (types_true);
1002	case 22:	1103	case SIMPLE_NULL:
1003	return newSVsv (&PL_sv_undef);	1104	return newSVsv (&PL_sv_undef);
1004	case 23:	1105	case SIMPLE_UNDEF:
1005	#if CBOR_SLOW	1106	#if CBOR_SLOW
1006	types_error = get_bool ("Types::Serialiser::error");	1107	types_error = get_bool ("Types::Serialiser::error");
1007	#endif	1108	#endif
1008	return newSVsv (types_error);	1109	return newSVsv (types_error);
1009		1110
1010	case 25:	1111	case MISC_FLOAT16:
1011	{	1112	{
1012	WANT (2);	1113	WANT (2);
1013		1114
1014	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1115	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1015	dec->cur += 2;	1116	dec->cur += 2;
1016		1117
1017	return newSVnv (ecb_binary16_to_float (fp));	1118	return newSVnv (ecb_binary16_to_float (fp));
1018	}	1119	}
1019		1120
1020	case 26:	1121	case MISC_FLOAT32:
1021	{	1122	{
1022	uint32_t fp;	1123	uint32_t fp;
1023	WANT (4);	1124	WANT (4);
1024	memcpy (&fp, dec->cur, 4);	1125	memcpy (&fp, dec->cur, 4);
1025	dec->cur += 4;	1126	dec->cur += 4;
…		…
1028	fp = ecb_bswap32 (fp);	1129	fp = ecb_bswap32 (fp);
1029		1130
1030	return newSVnv (ecb_binary32_to_float (fp));	1131	return newSVnv (ecb_binary32_to_float (fp));
1031	}	1132	}
1032		1133
1033	case 27:	1134	case MISC_FLOAT64:
1034	{	1135	{
1035	uint64_t fp;	1136	uint64_t fp;
1036	WANT (8);	1137	WANT (8);
1037	memcpy (&fp, dec->cur, 8);	1138	memcpy (&fp, dec->cur, 8);
1038	dec->cur += 8;	1139	dec->cur += 8;
…		…
1041	fp = ecb_bswap64 (fp);	1142	fp = ecb_bswap64 (fp);
1042		1143
1043	return newSVnv (ecb_binary64_to_double (fp));	1144	return newSVnv (ecb_binary64_to_double (fp));
1044	}	1145	}
1045		1146
1046	// 0..19 unassigned	1147	// 0..19 unassigned simple
1047	// 24 reserved + unassigned (reserved values are not encodable)	1148	// 24 reserved + unassigned simple (reserved values are not encodable)
		1149	// 28-30 unassigned misc
		1150	// 31 break code
1048	default:	1151	default:
1049	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1152	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1050	}	1153	}
1051		1154
1052	break;	1155	break;
1053	}	1156	}
1054		1157
…		…
1081	if (dec.cur != dec.end && !dec.err)	1184	if (dec.cur != dec.end && !dec.err)
1082	dec.err = "garbage after CBOR object";	1185	dec.err = "garbage after CBOR object";
1083		1186
1084	if (dec.err)	1187	if (dec.err)
1085	{	1188	{
		1189	if (dec.shareable)
		1190	{
		1191	// need to break cyclic links, which whould all be in shareable
		1192	int i;
		1193	SV **svp;
		1194
		1195	for (i = av_len (dec.shareable) + 1; i--; )
		1196	if ((svp = av_fetch (dec.shareable, i, 0)))
		1197	sv_setsv (*svp, &PL_sv_undef);
		1198	}
		1199
1086	SvREFCNT_dec (sv);	1200	SvREFCNT_dec (sv);
1087	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1201	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1088	}	1202	}
1089		1203
1090	sv = sv_2mortal (sv);	1204	sv = sv_2mortal (sv);
1091		1205
1092	return sv;	1206	return sv;
1093	}	1207	}
1094		1208
		1209	/////////////////////////////////////////////////////////////////////////////
		1210	// incremental parser
		1211
		1212	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1213
		1214	// returns 0 for notyet, 1 for success or error
		1215	static int
		1216	incr_parse (CBOR self, SV cborstr)
		1217	{
		1218	STRLEN cur;
		1219	SvPV (cborstr, cur);
		1220
		1221	while (ecb_expect_true (self->incr_need <= cur))
		1222	{
		1223	// table of integer count bytes
		1224	static I8 incr_len[MINOR_MASK + 1] = {
		1225	0, 0, 0, 0, 0, 0, 0, 0,
		1226	0, 0, 0, 0, 0, 0, 0, 0,
		1227	0, 0, 0, 0, 0, 0, 0, 0,
		1228	1, 2, 4, 8,-1,-1,-1,-2
		1229	};
		1230
		1231	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1232	U8 m = *p & MINOR_MASK;
		1233	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1234	I8 ilen = incr_len[m];
		1235
		1236	self->incr_need = self->incr_pos + 1;
		1237
		1238	if (ecb_expect_false (ilen < 0))
		1239	{
		1240	if (m != MINOR_INDEF)
		1241	return 1; // error
		1242
		1243	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1244	{
		1245	if (count >= 0)
		1246	return 1; // error
		1247
		1248	count = 1;
		1249	}
		1250	else
		1251	{
		1252	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1253	count = -1;
		1254	}
		1255	}
		1256	else
		1257	{
		1258	self->incr_need += ilen;
		1259	if (ecb_expect_false (self->incr_need > cur))
		1260	return 0;
		1261
		1262	int major = *p >> MAJOR_SHIFT;
		1263
		1264	switch (major)
		1265	{
		1266	case MAJOR_BYTES >> MAJOR_SHIFT:
		1267	case MAJOR_TEXT >> MAJOR_SHIFT:
		1268	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1269	case MAJOR_MAP >> MAJOR_SHIFT:
		1270	{
		1271	UV len;
		1272
		1273	if (ecb_expect_false (ilen))
		1274	{
		1275	len = 0;
		1276
		1277	do {
		1278	len = (len << 8) \| *++p;
		1279	} while (--ilen);
		1280	}
		1281	else
		1282	len = m;
		1283
		1284	switch (major)
		1285	{
		1286	case MAJOR_BYTES >> MAJOR_SHIFT:
		1287	case MAJOR_TEXT >> MAJOR_SHIFT:
		1288	self->incr_need += len;
		1289	if (ecb_expect_false (self->incr_need > cur))
		1290	return 0;
		1291
		1292	break;
		1293
		1294	case MAJOR_MAP >> MAJOR_SHIFT:
		1295	len <<= 1;
		1296	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1297	if (len)
		1298	{
		1299	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1300	count = len + 1;
		1301	}
		1302	break;
		1303	}
		1304	}
		1305	}
		1306	}
		1307
		1308	self->incr_pos = self->incr_need;
		1309
		1310	if (count > 0)
		1311	{
		1312	while (!--count)
		1313	{
		1314	if (!AvFILLp (self->incr_count))
		1315	return 1; // done
		1316
		1317	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1318	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1319	}
		1320
		1321	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1322	}
		1323	}
		1324
		1325	return 0;
		1326	}
		1327
		1328
1095	/////////////////////////////////////////////////////////////////////////////	1329	/////////////////////////////////////////////////////////////////////////////
1096	// XS interface functions	1330	// XS interface functions
1097		1331
1098	MODULE = CBOR::XS PACKAGE = CBOR::XS	1332	MODULE = CBOR::XS PACKAGE = CBOR::XS
1099		1333
…		…
1139	void shrink (CBOR *self, int enable = 1)	1373	void shrink (CBOR *self, int enable = 1)
1140	ALIAS:	1374	ALIAS:
1141	shrink = F_SHRINK	1375	shrink = F_SHRINK
1142	allow_unknown = F_ALLOW_UNKNOWN	1376	allow_unknown = F_ALLOW_UNKNOWN
1143	allow_sharing = F_ALLOW_SHARING	1377	allow_sharing = F_ALLOW_SHARING
1144	allow_stringref = F_ALLOW_STRINGREF	1378	allow_cycles = F_ALLOW_CYCLES
		1379	pack_strings = F_PACK_STRINGS
		1380	validate_utf8 = F_VALIDATE_UTF8
1145	PPCODE:	1381	PPCODE:
1146	{	1382	{
1147	if (enable)	1383	if (enable)
1148	self->flags \|= ix;	1384	self->flags \|= ix;
1149	else	1385	else
…		…
1155	void get_shrink (CBOR *self)	1391	void get_shrink (CBOR *self)
1156	ALIAS:	1392	ALIAS:
1157	get_shrink = F_SHRINK	1393	get_shrink = F_SHRINK
1158	get_allow_unknown = F_ALLOW_UNKNOWN	1394	get_allow_unknown = F_ALLOW_UNKNOWN
1159	get_allow_sharing = F_ALLOW_SHARING	1395	get_allow_sharing = F_ALLOW_SHARING
1160	get_allow_stringref = F_ALLOW_STRINGREF	1396	get_allow_cycles = F_ALLOW_CYCLES
		1397	get_pack_strings = F_PACK_STRINGS
		1398	get_validate_utf8 = F_VALIDATE_UTF8
1161	PPCODE:	1399	PPCODE:
1162	XPUSHs (boolSV (self->flags & ix));	1400	XPUSHs (boolSV (self->flags & ix));
1163		1401
1164	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1402	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1165	PPCODE:	1403	PPCODE:
…		…
1214	EXTEND (SP, 2);	1452	EXTEND (SP, 2);
1215	PUSHs (sv);	1453	PUSHs (sv);
1216	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1454	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1217	}	1455	}
1218		1456
		1457	void incr_parse (CBOR self, SV cborstr)
		1458	ALIAS:
		1459	incr_parse_multiple = 1
		1460	PPCODE:
		1461	{
		1462	if (SvUTF8 (cborstr))
		1463	sv_utf8_downgrade (cborstr, 0);
		1464
		1465	if (!self->incr_count)
		1466	{
		1467	self->incr_count = newAV ();
		1468	self->incr_pos = 0;
		1469	self->incr_need = 1;
		1470
		1471	av_push (self->incr_count, newSViv (1));
		1472	}
		1473
		1474	do
		1475	{
		1476	if (!incr_parse (self, cborstr))
		1477	{
		1478	if (self->incr_need > self->max_size && self->max_size)
		1479	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1480	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1481
		1482	break;
		1483	}
		1484
		1485	SV *sv;
		1486	char *offset;
		1487
		1488	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1489	XPUSHs (sv);
		1490
		1491	sv_chop (cborstr, offset);
		1492
		1493	av_clear (self->incr_count);
		1494	av_push (self->incr_count, newSViv (1));
		1495
		1496	self->incr_pos = 0;
		1497	self->incr_need = self->incr_pos + 1;
		1498	}
		1499	while (ix);
		1500	}
		1501
		1502	void incr_reset (CBOR *self)
		1503	CODE:
		1504	{
		1505	SvREFCNT_dec (self->incr_count);
		1506	self->incr_count = 0;
		1507	}
		1508
1219	void DESTROY (CBOR *self)	1509	void DESTROY (CBOR *self)
1220	PPCODE:	1510	PPCODE:
1221	cbor_free (self);	1511	cbor_free (self);
1222		1512
1223	PROTOTYPES: ENABLE	1513	PROTOTYPES: ENABLE
1224		1514
1225	void encode_cbor (SV *scalar)	1515	void encode_cbor (SV *scalar)
		1516	ALIAS:
		1517	encode_cbor = 0
		1518	encode_cbor_sharing = F_ALLOW_SHARING
1226	PPCODE:	1519	PPCODE:
1227	{	1520	{
1228	CBOR cbor;	1521	CBOR cbor;
1229	cbor_init (&cbor);	1522	cbor_init (&cbor);
		1523	cbor.flags \|= ix;
1230	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1524	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1231	XPUSHs (scalar);	1525	XPUSHs (scalar);
1232	}	1526	}
1233		1527
1234	void decode_cbor (SV *cborstr)	1528	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.28 by root, Sat Nov 23 18:30:59 2013 UTC vs. Revision 1.45 by root, Sun Dec 14 05:48:39 2014 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.28 by root, Sat Nov 23 18:30:59 2013 UTC vs.
Revision 1.45 by root, Sun Dec 14 05:48:39 2014 UTC