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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.25 by root, Fri Nov 22 05:54:07 2013 UTC vs.
Revision 1.50 by root, Thu Feb 25 02:29:22 2016 UTC

…		…
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
11		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	20	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	21	#endif
21	#ifndef HvNAMEUTF8	22	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	23	# define HvNAMEUTF8(hv) 0
23	#endif	24	#endif
		25	#ifndef SvREFCNT_dec_NN
		26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		27	#endif
		28
		29	// known major and minor types
		30	enum cbor_type
		31	{
		32	MAJOR_SHIFT = 5,
		33	MINOR_MASK = 0x1f,
		34
		35	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		36	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		37	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		38	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		39	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		40	MAJOR_MAP = 5 << MAJOR_SHIFT,
		41	MAJOR_TAG = 6 << MAJOR_SHIFT,
		42	MAJOR_MISC = 7 << MAJOR_SHIFT,
		43
		44	// INT/STRING/ARRAY/MAP subtypes
		45	LENGTH_EXT1 = 24,
		46	LENGTH_EXT2 = 25,
		47	LENGTH_EXT4 = 26,
		48	LENGTH_EXT8 = 27,
		49
		50	// SIMPLE types (effectively MISC subtypes)
		51	SIMPLE_FALSE = 20,
		52	SIMPLE_TRUE = 21,
		53	SIMPLE_NULL = 22,
		54	SIMPLE_UNDEF = 23,
		55
		56	// MISC subtype (unused)
		57	MISC_EXT1 = 24,
		58	MISC_FLOAT16 = 25,
		59	MISC_FLOAT32 = 26,
		60	MISC_FLOAT64 = 27,
		61
		62	// BYTES/TEXT/ARRAY/MAP
		63	MINOR_INDEF = 31,
		64	};
24		65
25	// known tags	66	// known tags
26	enum cbor_tag	67	enum cbor_tag
27	{	68	{
28	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		77
37	// rfc7049	78	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
44		85
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		90
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		96
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
57	};	98	};
58		99
59	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
62	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	103	#define F_ALLOW_CYCLES 0x00000008UL
		104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_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)
…		…
79	# define CBOR_SLOW 0	122	# define CBOR_SLOW 0
80	# define CBOR_STASH cbor_stash	123	# define CBOR_STASH cbor_stash
81	#endif	124	#endif
82		125
83	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::	126	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::
84	static SV types_true, types_false, types_error, sv_cbor;	127	static SV types_true, types_false, types_error, sv_cbor, *default_filter;
85		128
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;
		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.
90	} CBOR;	139	} CBOR;
91		140
92	ecb_inline void	141	ecb_inline void
93	cbor_init (CBOR *cbor)	142	cbor_init (CBOR *cbor)
94	{	143	{
95	Zero (cbor, 1, CBOR);	144	Zero (cbor, 1, CBOR);
96	cbor->max_depth = 512;	145	cbor->max_depth = 512;
		146	}
		147
		148	ecb_inline void
		149	cbor_free (CBOR *cbor)
		150	{
		151	SvREFCNT_dec (cbor->filter);
		152	SvREFCNT_dec (cbor->incr_count);
97	}	153	}
98		154
99	/////////////////////////////////////////////////////////////////////////////	155	/////////////////////////////////////////////////////////////////////////////
100	// utility functions	156	// utility functions
101		157
…		…
131	{	187	{
132	return idx > 23	188	return idx > 23
133	? idx > 0xffU	189	? idx > 0xffU
134	? idx > 0xffffU	190	? idx > 0xffffU
135	? idx > 0xffffffffU	191	? idx > 0xffffffffU
		192	? 11
136	? 7	193	: 7
137	: 6
138	: 5	194	: 5
139	: 4	195	: 4
140	: 3;	196	: 3;
141	}	197	}
142		198
…		…
179	static void	235	static void
180	encode_uint (enc_t *enc, int major, UV len)	236	encode_uint (enc_t *enc, int major, UV len)
181	{	237	{
182	need (enc, 9);	238	need (enc, 9);
183		239
184	if (len < 24)	240	if (ecb_expect_true (len < LENGTH_EXT1))
185	*enc->cur++ = major \| len;	241	*enc->cur++ = major \| len;
186	else if (len <= 0xff)	242	else if (ecb_expect_true (len <= 0xffU))
187	{	243	{
188	*enc->cur++ = major \| 24;	244	*enc->cur++ = major \| LENGTH_EXT1;
189	*enc->cur++ = len;	245	*enc->cur++ = len;
190	}	246	}
191	else if (len <= 0xffff)	247	else if (len <= 0xffffU)
192	{	248	{
193	*enc->cur++ = major \| 25;	249	*enc->cur++ = major \| LENGTH_EXT2;
194	*enc->cur++ = len >> 8;	250	*enc->cur++ = len >> 8;
195	*enc->cur++ = len;	251	*enc->cur++ = len;
196	}	252	}
197	else if (len <= 0xffffffff)	253	else if (len <= 0xffffffffU)
198	{	254	{
199	*enc->cur++ = major \| 26;	255	*enc->cur++ = major \| LENGTH_EXT4;
200	*enc->cur++ = len >> 24;	256	*enc->cur++ = len >> 24;
201	*enc->cur++ = len >> 16;	257	*enc->cur++ = len >> 16;
202	*enc->cur++ = len >> 8;	258	*enc->cur++ = len >> 8;
203	*enc->cur++ = len;	259	*enc->cur++ = len;
204	}	260	}
205	else	261	else
206	{	262	{
207	*enc->cur++ = major \| 27;	263	*enc->cur++ = major \| LENGTH_EXT8;
208	*enc->cur++ = len >> 56;	264	*enc->cur++ = len >> 56;
209	*enc->cur++ = len >> 48;	265	*enc->cur++ = len >> 48;
210	*enc->cur++ = len >> 40;	266	*enc->cur++ = len >> 40;
211	*enc->cur++ = len >> 32;	267	*enc->cur++ = len >> 32;
212	*enc->cur++ = len >> 24;	268	*enc->cur++ = len >> 24;
…		…
217	}	273	}
218		274
219	ecb_inline void	275	ecb_inline void
220	encode_tag (enc_t *enc, UV tag)	276	encode_tag (enc_t *enc, UV tag)
221	{	277	{
222	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;
223	}	288	}
224		289
225	static void	290	static void
226	encode_str (enc_t enc, int utf8, char str, STRLEN len)	291	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
227	{	292	{
228	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	293	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229	{	294	{
230	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	295	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231		296
232	if (SvOK (*svp))	297	if (SvOK (*svp))
233	{	298	{
234	// already registered, use stringref	299	// already registered, use stringref
235	encode_tag (enc, CBOR_TAG_STRINGREF);	300	encode_tag (enc, CBOR_TAG_STRINGREF);
236	encode_uint (enc, 0x00, SvUV (*svp));	301	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237	return;	302	return;
238	}	303	}
239	else if (len >= minimum_string_length (enc->stringref_idx))	304	else if (len >= minimum_string_length (enc->stringref_idx))
240	{	305	{
241	// register only	306	// register only
242	sv_setuv (*svp, enc->stringref_idx);	307	sv_setuv (*svp, enc->stringref_idx);
243	++enc->stringref_idx;	308	++enc->stringref_idx;
244	}	309	}
245	}	310	}
246		311
247	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	312	encode_str (enc, utf8, str, len);
248	need (enc, len);
249	memcpy (enc->cur, str, len);
250	enc->cur += len;
251	}	313	}
252		314
253	static void encode_sv (enc_t enc, SV sv);	315	static void encode_sv (enc_t enc, SV sv);
254		316
255	static void	317	static void
…		…
260	if (enc->depth >= enc->cbor.max_depth)	322	if (enc->depth >= enc->cbor.max_depth)
261	croak (ERR_NESTING_EXCEEDED);	323	croak (ERR_NESTING_EXCEEDED);
262		324
263	++enc->depth;	325	++enc->depth;
264		326
265	encode_uint (enc, 0x80, len + 1);	327	encode_uint (enc, MAJOR_ARRAY, len + 1);
266		328
		329	if (SvMAGICAL (av))
267	for (i = 0; i <= len; ++i)	330	for (i = 0; i <= len; ++i)
268	{	331	{
269	SV **svp = av_fetch (av, i, 0);	332	SV **svp = av_fetch (av, i, 0);
270	encode_sv (enc, svp ? *svp : &PL_sv_undef);	333	encode_sv (enc, svp ? *svp : &PL_sv_undef);
271	}	334	}
		335	else
		336	for (i = 0; i <= len; ++i)
		337	{
		338	SV *sv = AvARRAY (av)[i];
		339	encode_sv (enc, sv ? sv : &PL_sv_undef);
		340	}
272		341
273	--enc->depth;	342	--enc->depth;
274	}
275
276	ecb_inline void
277	encode_he (enc_t enc, HE he)
278	{
279	}	343	}
280		344
281	static void	345	static void
282	encode_hv (enc_t enc, HV hv)	346	encode_hv (enc_t enc, HV hv)
283	{	347	{
…		…
290		354
291	int pairs = hv_iterinit (hv);	355	int pairs = hv_iterinit (hv);
292	int mg = SvMAGICAL (hv);	356	int mg = SvMAGICAL (hv);
293		357
294	if (mg)	358	if (mg)
295	encode_ch (enc, 0xa0 \| 31);	359	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
296	else	360	else
297	encode_uint (enc, 0xa0, pairs);	361	encode_uint (enc, MAJOR_MAP, pairs);
298		362
299	while ((he = hv_iternext (hv)))	363	while ((he = hv_iternext (hv)))
300	{	364	{
301	if (HeKLEN (he) == HEf_SVKEY)	365	if (HeKLEN (he) == HEf_SVKEY)
302	encode_sv (enc, HeSVKEY (he));	366	encode_sv (enc, HeSVKEY (he));
303	else	367	else
304	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	368	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305		369
306	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	370	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307	}	371	}
308		372
309	if (mg)	373	if (mg)
310	encode_ch (enc, 0xe0 \| 31);	374	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
311		375
312	--enc->depth;	376	--enc->depth;
313	}	377	}
314		378
315	// encode objects, arrays and special \0=false and \1=true values.	379	// encode objects, arrays and special \0=false and \1=true values.
316	static void	380	static void
317	encode_rv (enc_t enc, SV sv)	381	encode_rv (enc_t enc, SV sv)
318	{	382	{
319	SvGETMAGIC (sv);	383	SvGETMAGIC (sv);
320
321	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
322	&& ecb_expect_false (SvREFCNT (sv) > 1))
323	{
324	if (!enc->shareable)
325	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
326
327	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
328
329	if (SvOK (*svp))
330	{
331	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
332	encode_uint (enc, 0x00, SvUV (*svp));
333	return;
334	}
335	else
336	{
337	sv_setuv (*svp, enc->shareable_idx);
338	++enc->shareable_idx;
339	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
340	}
341	}
342		384
343	svtype svt = SvTYPE (sv);	385	svtype svt = SvTYPE (sv);
344		386
345	if (ecb_expect_false (SvOBJECT (sv)))	387	if (ecb_expect_false (SvOBJECT (sv)))
346	{	388	{
…		…
353	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	395	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
354	? cbor_tagged_stash	396	? cbor_tagged_stash
355	: gv_stashpv ("CBOR::XS::Tagged" , 1);	397	: gv_stashpv ("CBOR::XS::Tagged" , 1);
356		398
357	HV *stash = SvSTASH (sv);	399	HV *stash = SvSTASH (sv);
358	GV *method;
359		400
360	if (stash == boolean_stash)	401	if (stash == boolean_stash)
361	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	402	{
		403	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		404	return;
		405	}
362	else if (stash == error_stash)	406	else if (stash == error_stash)
363	encode_ch (enc, 0xe0 \| 23);	407	{
		408	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		409	return;
		410	}
364	else if (stash == tagged_stash)	411	else if (stash == tagged_stash)
365	{	412	{
366	if (svt != SVt_PVAV)	413	if (svt != SVt_PVAV)
367	croak ("encountered CBOR::XS::Tagged object that isn't an array");	414	croak ("encountered CBOR::XS::Tagged object that isn't an array");
368		415
369	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	416	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
370	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	417	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		418
		419	return;
		420	}
		421	}
		422
		423	if (ecb_expect_false (SvREFCNT (sv) > 1)
		424	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		425	{
		426	if (!enc->shareable)
		427	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		428
		429	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		430
		431	if (SvOK (*svp))
371	}	432	{
		433	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		434	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		435	return;
		436	}
		437	else
		438	{
		439	sv_setuv (*svp, enc->shareable_idx);
		440	++enc->shareable_idx;
		441	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		442	}
		443	}
		444
		445	if (ecb_expect_false (SvOBJECT (sv)))
		446	{
		447	HV *stash = SvSTASH (sv);
		448	GV *method;
		449
372	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	450	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373	{	451	{
374	dSP;	452	dSP;
375		453
376	ENTER; SAVETMPS; PUSHMARK (SP);	454	ENTER; SAVETMPS;
		455	PUSHMARK (SP);
377	// we re-bless the reference to get overload and other niceties right	456	// we re-bless the reference to get overload and other niceties right
378	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	457	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379		458
380	PUTBACK;	459	PUTBACK;
381	// G_SCALAR ensures that return value is 1	460	// G_SCALAR ensures that return value is 1
…		…
394	}	473	}
395	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	474	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396	{	475	{
397	dSP;	476	dSP;
398		477
399	ENTER; SAVETMPS; PUSHMARK (SP);	478	ENTER; SAVETMPS;
		479	SAVESTACK_POS ();
		480	PUSHMARK (SP);
400	EXTEND (SP, 2);	481	EXTEND (SP, 2);
401	// we re-bless the reference to get overload and other niceties right	482	// we re-bless the reference to get overload and other niceties right
402	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403	PUSHs (sv_cbor);	484	PUSHs (sv_cbor);
404		485
…		…
409	// catch this surprisingly common error	490	// catch this surprisingly common error
410	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	491	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
411	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	492	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412		493
413	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	494	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414	encode_uint (enc, 0x80, count + 1);	495	encode_uint (enc, MAJOR_ARRAY, count + 1);
415	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	496	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416		497
417	while (count)	498	while (count)
418	encode_sv (enc, SP[1 - count--]);	499	encode_sv (enc, SP[1 - count--]);
419		500
420	PUTBACK;	501	PUTBACK;
…		…
441	{	522	{
442	double nv = SvNVX (sv);	523	double nv = SvNVX (sv);
443		524
444	need (enc, 9);	525	need (enc, 9);
445		526
446	if (ecb_expect_false (nv == (U32)nv))	527	if (ecb_expect_false (nv == (NV)(U32)nv))
447	encode_uint (enc, 0x00, (U32)nv);	528	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448	//TODO: maybe I32?	529	//TODO: maybe I32?
449	else if (ecb_expect_false (nv == (float)nv))	530	else if (ecb_expect_false (nv == (float)nv))
450	{	531	{
451	uint32_t fp = ecb_float_to_binary32 (nv);	532	uint32_t fp = ecb_float_to_binary32 (nv);
452		533
453	*enc->cur++ = 0xe0 \| 26;	534	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
454		535
455	if (!ecb_big_endian ())	536	if (!ecb_big_endian ())
456	fp = ecb_bswap32 (fp);	537	fp = ecb_bswap32 (fp);
457		538
458	memcpy (enc->cur, &fp, 4);	539	memcpy (enc->cur, &fp, 4);
…		…
460	}	541	}
461	else	542	else
462	{	543	{
463	uint64_t fp = ecb_double_to_binary64 (nv);	544	uint64_t fp = ecb_double_to_binary64 (nv);
464		545
465	*enc->cur++ = 0xe0 \| 27;	546	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
466		547
467	if (!ecb_big_endian ())	548	if (!ecb_big_endian ())
468	fp = ecb_bswap64 (fp);	549	fp = ecb_bswap64 (fp);
469		550
470	memcpy (enc->cur, &fp, 8);	551	memcpy (enc->cur, &fp, 8);
…		…
479		560
480	if (SvPOKp (sv))	561	if (SvPOKp (sv))
481	{	562	{
482	STRLEN len;	563	STRLEN len;
483	char *str = SvPV (sv, len);	564	char *str = SvPV (sv, len);
484	encode_str (enc, SvUTF8 (sv), str, len);	565	encode_strref (enc, SvUTF8 (sv), str, len);
485	}	566	}
486	else if (SvNOKp (sv))	567	else if (SvNOKp (sv))
487	encode_nv (enc, sv);	568	encode_nv (enc, sv);
488	else if (SvIOKp (sv))	569	else if (SvIOKp (sv))
489	{	570	{
490	if (SvIsUV (sv))	571	if (SvIsUV (sv))
491	encode_uint (enc, 0x00, SvUVX (sv));	572	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
492	else if (SvIVX (sv) >= 0)	573	else if (SvIVX (sv) >= 0)
493	encode_uint (enc, 0x00, SvIVX (sv));	574	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
494	else	575	else
495	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	576	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
496	}	577	}
497	else if (SvROK (sv))	578	else if (SvROK (sv))
498	encode_rv (enc, SvRV (sv));	579	encode_rv (enc, SvRV (sv));
499	else if (!SvOK (sv))	580	else if (!SvOK (sv))
500	encode_ch (enc, 0xe0 \| 22);	581	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
501	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	582	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502	encode_ch (enc, 0xe0 \| 23);	583	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
503	else	584	else
504	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	585	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	586	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506	}	587	}
507		588
508	static SV *	589	static SV *
509	encode_cbor (SV scalar, CBOR cbor)	590	encode_cbor (SV scalar, CBOR cbor)
510	{	591	{
511	enc_t enc = { };	592	enc_t enc = { 0 };
512		593
513	enc.cbor = *cbor;	594	enc.cbor = *cbor;
514	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	595	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515	enc.cur = SvPVX (enc.sv);	596	enc.cur = SvPVX (enc.sv);
516	enc.end = SvEND (enc.sv);	597	enc.end = SvEND (enc.sv);
517		598
518	SvPOK_only (enc.sv);	599	SvPOK_only (enc.sv);
519		600
520	if (cbor->flags & F_ALLOW_STRINGREF)	601	if (cbor->flags & F_PACK_STRINGS)
521	{	602	{
522	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	603	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	604	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
524	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	605	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
525	}	606	}
…		…
547	CBOR cbor;	628	CBOR cbor;
548	U32 depth; // recursion depth	629	U32 depth; // recursion depth
549	U32 maxdepth; // recursion depth limit	630	U32 maxdepth; // recursion depth limit
550	AV *shareable;	631	AV *shareable;
551	AV *stringref;	632	AV *stringref;
		633	SV *decode_tagged;
552	} dec_t;	634	} dec_t;
553		635
554	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	636	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
555		637
556	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	638	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
…		…
559	#define DEC_DEC_DEPTH --dec->depth	641	#define DEC_DEC_DEPTH --dec->depth
560		642
561	static UV	643	static UV
562	decode_uint (dec_t *dec)	644	decode_uint (dec_t *dec)
563	{	645	{
564	switch (*dec->cur & 31)	646	U8 m = *dec->cur & MINOR_MASK;
565	{	647	++dec->cur;
566	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
567	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
568	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
569	return *dec->cur++ & 31;
570		648
571	case 24:	649	if (ecb_expect_true (m < LENGTH_EXT1))
		650	return m;
		651	else if (ecb_expect_true (m == LENGTH_EXT1))
		652	{
572	WANT (2);	653	WANT (1);
573	dec->cur += 2;	654	dec->cur += 1;
574	return dec->cur[-1];	655	return dec->cur[-1];
575		656	}
576	case 25:	657	else if (ecb_expect_true (m == LENGTH_EXT2))
		658	{
577	WANT (3);	659	WANT (2);
578	dec->cur += 3;	660	dec->cur += 2;
579	return (((UV)dec->cur[-2]) << 8)	661	return (((UV)dec->cur[-2]) << 8)
580	\| ((UV)dec->cur[-1]);	662	\| ((UV)dec->cur[-1]);
581		663	}
582	case 26:	664	else if (ecb_expect_true (m == LENGTH_EXT4))
		665	{
583	WANT (5);	666	WANT (4);
584	dec->cur += 5;	667	dec->cur += 4;
585	return (((UV)dec->cur[-4]) << 24)	668	return (((UV)dec->cur[-4]) << 24)
586	\| (((UV)dec->cur[-3]) << 16)	669	\| (((UV)dec->cur[-3]) << 16)
587	\| (((UV)dec->cur[-2]) << 8)	670	\| (((UV)dec->cur[-2]) << 8)
588	\| ((UV)dec->cur[-1]);	671	\| ((UV)dec->cur[-1]);
589		672	}
590	case 27:	673	else if (ecb_expect_true (m == LENGTH_EXT8))
		674	{
591	WANT (9);	675	WANT (8);
592	dec->cur += 9;	676	dec->cur += 8;
		677
		678	return
		679	#if UVSIZE < 8
		680	0
		681	#else
593	return (((UV)dec->cur[-8]) << 56)	682	(((UV)dec->cur[-8]) << 56)
594	\| (((UV)dec->cur[-7]) << 48)	683	\| (((UV)dec->cur[-7]) << 48)
595	\| (((UV)dec->cur[-6]) << 40)	684	\| (((UV)dec->cur[-6]) << 40)
596	\| (((UV)dec->cur[-5]) << 32)	685	\| (((UV)dec->cur[-5]) << 32)
		686	#endif
597	\| (((UV)dec->cur[-4]) << 24)	687	\| (((UV)dec->cur[-4]) << 24)
598	\| (((UV)dec->cur[-3]) << 16)	688	\| (((UV)dec->cur[-3]) << 16)
599	\| (((UV)dec->cur[-2]) << 8)	689	\| (((UV)dec->cur[-2]) << 8)
600	\| ((UV)dec->cur[-1]);	690	\| ((UV)dec->cur[-1]);
601		691	}
602	default:	692	else
603	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	693	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604	}
605		694
606	fail:	695	fail:
607	return 0;	696	return 0;
608	}	697	}
609		698
…		…
614	{	703	{
615	AV *av = newAV ();	704	AV *av = newAV ();
616		705
617	DEC_INC_DEPTH;	706	DEC_INC_DEPTH;
618		707
619	if ((*dec->cur & 31) == 31)	708	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
620	{	709	{
621	++dec->cur;	710	++dec->cur;
622		711
623	for (;;)	712	for (;;)
624	{	713	{
625	WANT (1);	714	WANT (1);
626		715
627	if (*dec->cur == (0xe0 \| 31))	716	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
628	{	717	{
629	++dec->cur;	718	++dec->cur;
630	break;	719	break;
631	}	720	}
632		721
…		…
635	}	724	}
636	else	725	else
637	{	726	{
638	int i, len = decode_uint (dec);	727	int i, len = decode_uint (dec);
639		728
		729	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
640	av_fill (av, len - 1);	730	av_fill (av, len - 1);
641		731
642	for (i = 0; i < len; ++i)	732	for (i = 0; i < len; ++i)
643	AvARRAY (av)[i] = decode_sv (dec);	733	AvARRAY (av)[i] = decode_sv (dec);
644	}	734	}
…		…
657	{	747	{
658	// for speed reasons, we specialcase single-string	748	// for speed reasons, we specialcase single-string
659	// byte or utf-8 strings as keys, but only when !stringref	749	// byte or utf-8 strings as keys, but only when !stringref
660		750
661	if (ecb_expect_true (!dec->stringref))	751	if (ecb_expect_true (!dec->stringref))
662	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	752	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663	{	753	{
664	I32 len = decode_uint (dec);	754	I32 len = decode_uint (dec);
665	char key = (char )dec->cur;	755	char key = (char )dec->cur;
666		756
		757	WANT (len);
667	dec->cur += len;	758	dec->cur += len;
668		759
669	if (ecb_expect_false (dec->stringref))
670	av_push (dec->stringref, newSVpvn (key, len));
671
672	hv_store (hv, key, len, decode_sv (dec), 0);	760	hv_store (hv, key, len, decode_sv (dec), 0);
673		761
674	return;	762	return;
675	}	763	}
676	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	764	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677	{	765	{
678	I32 len = decode_uint (dec);	766	I32 len = decode_uint (dec);
679	char key = (char )dec->cur;	767	char key = (char )dec->cur;
680		768
		769	WANT (len);
681	dec->cur += len;	770	dec->cur += len;
682		771
683	if (ecb_expect_false (dec->stringref))	772	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	773	if (!is_utf8_string (key, len))
		774	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685		775
686	hv_store (hv, key, -len, decode_sv (dec), 0);	776	hv_store (hv, key, -len, decode_sv (dec), 0);
687		777
688	return;	778	return;
689	}	779	}
…		…
691	SV *k = decode_sv (dec);	781	SV *k = decode_sv (dec);
692	SV *v = decode_sv (dec);	782	SV *v = decode_sv (dec);
693		783
694	hv_store_ent (hv, k, v, 0);	784	hv_store_ent (hv, k, v, 0);
695	SvREFCNT_dec (k);	785	SvREFCNT_dec (k);
		786
		787	fail:
		788	;
696	}	789	}
697		790
698	static SV *	791	static SV *
699	decode_hv (dec_t *dec)	792	decode_hv (dec_t *dec)
700	{	793	{
701	HV *hv = newHV ();	794	HV *hv = newHV ();
702		795
703	DEC_INC_DEPTH;	796	DEC_INC_DEPTH;
704		797
705	if ((*dec->cur & 31) == 31)	798	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
706	{	799	{
707	++dec->cur;	800	++dec->cur;
708		801
709	for (;;)	802	for (;;)
710	{	803	{
711	WANT (1);	804	WANT (1);
712		805
713	if (*dec->cur == (0xe0 \| 31))	806	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
714	{	807	{
715	++dec->cur;	808	++dec->cur;
716	break;	809	break;
717	}	810	}
718		811
…		…
739	static SV *	832	static SV *
740	decode_str (dec_t *dec, int utf8)	833	decode_str (dec_t *dec, int utf8)
741	{	834	{
742	SV *sv = 0;	835	SV *sv = 0;
743		836
744	if ((*dec->cur & 31) == 31)	837	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
745	{	838	{
		839	// indefinite length strings
746	++dec->cur;	840	++dec->cur;
747		841
		842	U8 major = *dec->cur & MAJOR_MISC;
		843
748	sv = newSVpvn ("", 0);	844	sv = newSVpvn ("", 0);
749		845
750	// not very fast, and certainly not robust against illegal input
751	for (;;)	846	for (;;)
752	{	847	{
753	WANT (1);	848	WANT (1);
754		849
755	if (*dec->cur == (0xe0 \| 31))	850	if ((*dec->cur - major) > LENGTH_EXT8)
		851	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
756	{	852	{
757	++dec->cur;	853	++dec->cur;
758	break;	854	break;
759	}	855	}
		856	else
		857	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760		858
761	sv_catsv (sv, decode_sv (dec));	859	STRLEN len = decode_uint (dec);
		860
		861	WANT (len);
		862	sv_catpvn (sv, dec->cur, len);
		863	dec->cur += len;
762	}	864	}
763	}	865	}
764	else	866	else
765	{	867	{
766	STRLEN len = decode_uint (dec);	868	STRLEN len = decode_uint (dec);
…		…
773	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	875	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
774	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	876	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
775	}	877	}
776		878
777	if (utf8)	879	if (utf8)
		880	{
		881	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		882	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		883	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		884
778	SvUTF8_on (sv);	885	SvUTF8_on (sv);
		886	}
779		887
780	return sv;	888	return sv;
781		889
782	fail:	890	fail:
783	SvREFCNT_dec (sv);	891	SvREFCNT_dec (sv);
…		…
815	}	923	}
816	break;	924	break;
817		925
818	case CBOR_TAG_STRINGREF:	926	case CBOR_TAG_STRINGREF:
819	{	927	{
820	if ((*dec->cur >> 5) != 0)	928	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	929	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822		930
823	UV idx = decode_uint (dec);	931	UV idx = decode_uint (dec);
824		932
825	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	933	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
832	case CBOR_TAG_VALUE_SHAREABLE:	940	case CBOR_TAG_VALUE_SHAREABLE:
833	{	941	{
834	if (ecb_expect_false (!dec->shareable))	942	if (ecb_expect_false (!dec->shareable))
835	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	943	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
836		944
		945	if (dec->cbor.flags & F_ALLOW_CYCLES)
		946	{
837	sv = newSV (0);	947	sv = newSV (0);
838	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	948	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839		949
840	SV *osv = decode_sv (dec);	950	SV *osv = decode_sv (dec);
841	sv_setsv (sv, osv);	951	sv_setsv (sv, osv);
842	SvREFCNT_dec_NN (osv);	952	SvREFCNT_dec_NN (osv);
		953	}
		954	else
		955	{
		956	av_push (dec->shareable, &PL_sv_undef);
		957	int idx = AvFILLp (dec->shareable);
		958	sv = decode_sv (dec);
		959	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		960	}
843	}	961	}
844	break;	962	break;
845		963
846	case CBOR_TAG_VALUE_SHAREDREF:	964	case CBOR_TAG_VALUE_SHAREDREF:
847	{	965	{
848	if ((*dec->cur >> 5) != 0)	966	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	967	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850		968
851	UV idx = decode_uint (dec);	969	UV idx = decode_uint (dec);
852		970
853	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	971	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
854	ERR ("corrupted CBOR data (sharedref index out of bounds)");	972	ERR ("corrupted CBOR data (sharedref index out of bounds)");
855		973
856	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	974	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		975
		976	if (sv == &PL_sv_undef)
		977	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857	}	978	}
858	break;	979	break;
859		980
860	case CBOR_TAG_PERL_OBJECT:	981	case CBOR_TAG_PERL_OBJECT:
861	{	982	{
…		…
876	if (!method)	997	if (!method)
877	ERR ("cannot decode perl-object (package does not have a THAW method)");	998	ERR ("cannot decode perl-object (package does not have a THAW method)");
878		999
879	dSP;	1000	dSP;
880		1001
881	ENTER; SAVETMPS; PUSHMARK (SP);	1002	ENTER; SAVETMPS;
		1003	PUSHMARK (SP);
882	EXTEND (SP, len + 1);	1004	EXTEND (SP, len + 1);
883	// we re-bless the reference to get overload and other niceties right	1005	// we re-bless the reference to get overload and other niceties right
884	PUSHs (*av_fetch (av, 0, 1));	1006	PUSHs (*av_fetch (av, 0, 1));
885	PUSHs (sv_cbor);	1007	PUSHs (sv_cbor);
886		1008
…		…
910		1032
911	default:	1033	default:
912	{	1034	{
913	sv = decode_sv (dec);	1035	sv = decode_sv (dec);
914		1036
		1037	dSP;
		1038	ENTER; SAVETMPS;
		1039	SAVESTACK_POS ();
		1040	PUSHMARK (SP);
		1041	EXTEND (SP, 2);
		1042	PUSHs (newSVuv (tag));
		1043	PUSHs (sv);
		1044
		1045	PUTBACK;
		1046	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
		1047	SPAGAIN;
		1048
		1049	if (SvTRUE (ERRSV))
		1050	{
		1051	FREETMPS; LEAVE;
		1052	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
		1053	}
		1054
		1055	if (count)
		1056	{
		1057	SvREFCNT_dec (sv);
		1058	sv = SvREFCNT_inc (POPs);
		1059	}
		1060	else
		1061	{
915	AV *av = newAV ();	1062	AV *av = newAV ();
916	av_push (av, newSVuv (tag));	1063	av_push (av, newSVuv (tag));
917	av_push (av, sv);	1064	av_push (av, sv);
918		1065
919	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1066	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
920	? cbor_tagged_stash	1067	? cbor_tagged_stash
921	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1068	: gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1069	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1070	}
		1071
		1072	PUTBACK;
		1073
		1074	FREETMPS; LEAVE;
924	}	1075	}
925	break;	1076	break;
926	}	1077	}
927		1078
928	return sv;	1079	return sv;
…		…
935	static SV *	1086	static SV *
936	decode_sv (dec_t *dec)	1087	decode_sv (dec_t *dec)
937	{	1088	{
938	WANT (1);	1089	WANT (1);
939		1090
940	switch (*dec->cur >> 5)	1091	switch (*dec->cur >> MAJOR_SHIFT)
941	{	1092	{
942	case 0: // unsigned int	1093	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943	return newSVuv (decode_uint (dec));	1094	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944	case 1: // negative int	1095	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945	return newSViv (-1 - (IV)decode_uint (dec));	1096	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946	case 2: // octet string	1097	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947	return decode_str (dec, 0);	1098	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948	case 3: // utf-8 string	1099	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949	return decode_str (dec, 1);	1100
950	case 4: // array	1101	case MAJOR_MISC >> MAJOR_SHIFT:
951	return decode_av (dec);
952	case 5: // map
953	return decode_hv (dec);
954	case 6: // tag
955	return decode_tagged (dec);
956	case 7: // misc
957	switch (*dec->cur++ & 31)	1102	switch (*dec->cur++ & MINOR_MASK)
958	{	1103	{
959	case 20:	1104	case SIMPLE_FALSE:
960	#if CBOR_SLOW	1105	#if CBOR_SLOW
961	types_false = get_bool ("Types::Serialiser::false");	1106	types_false = get_bool ("Types::Serialiser::false");
962	#endif	1107	#endif
963	return newSVsv (types_false);	1108	return newSVsv (types_false);
964	case 21:	1109	case SIMPLE_TRUE:
965	#if CBOR_SLOW	1110	#if CBOR_SLOW
966	types_true = get_bool ("Types::Serialiser::true");	1111	types_true = get_bool ("Types::Serialiser::true");
967	#endif	1112	#endif
968	return newSVsv (types_true);	1113	return newSVsv (types_true);
969	case 22:	1114	case SIMPLE_NULL:
970	return newSVsv (&PL_sv_undef);	1115	return newSVsv (&PL_sv_undef);
971	case 23:	1116	case SIMPLE_UNDEF:
972	#if CBOR_SLOW	1117	#if CBOR_SLOW
973	types_error = get_bool ("Types::Serialiser::error");	1118	types_error = get_bool ("Types::Serialiser::error");
974	#endif	1119	#endif
975	return newSVsv (types_error);	1120	return newSVsv (types_error);
976		1121
977	case 25:	1122	case MISC_FLOAT16:
978	{	1123	{
979	WANT (2);	1124	WANT (2);
980		1125
981	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1126	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
982	dec->cur += 2;	1127	dec->cur += 2;
983		1128
984	return newSVnv (ecb_binary16_to_float (fp));	1129	return newSVnv (ecb_binary16_to_float (fp));
985	}	1130	}
986		1131
987	case 26:	1132	case MISC_FLOAT32:
988	{	1133	{
989	uint32_t fp;	1134	uint32_t fp;
990	WANT (4);	1135	WANT (4);
991	memcpy (&fp, dec->cur, 4);	1136	memcpy (&fp, dec->cur, 4);
992	dec->cur += 4;	1137	dec->cur += 4;
…		…
995	fp = ecb_bswap32 (fp);	1140	fp = ecb_bswap32 (fp);
996		1141
997	return newSVnv (ecb_binary32_to_float (fp));	1142	return newSVnv (ecb_binary32_to_float (fp));
998	}	1143	}
999		1144
1000	case 27:	1145	case MISC_FLOAT64:
1001	{	1146	{
1002	uint64_t fp;	1147	uint64_t fp;
1003	WANT (8);	1148	WANT (8);
1004	memcpy (&fp, dec->cur, 8);	1149	memcpy (&fp, dec->cur, 8);
1005	dec->cur += 8;	1150	dec->cur += 8;
…		…
1008	fp = ecb_bswap64 (fp);	1153	fp = ecb_bswap64 (fp);
1009		1154
1010	return newSVnv (ecb_binary64_to_double (fp));	1155	return newSVnv (ecb_binary64_to_double (fp));
1011	}	1156	}
1012		1157
1013	// 0..19 unassigned	1158	// 0..19 unassigned simple
1014	// 24 reserved + unassigned (reserved values are not encodable)	1159	// 24 reserved + unassigned simple (reserved values are not encodable)
		1160	// 28-30 unassigned misc
		1161	// 31 break code
1015	default:	1162	default:
1016	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1163	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017	}	1164	}
1018		1165
1019	break;	1166	break;
1020	}	1167	}
1021		1168
…		…
1024	}	1171	}
1025		1172
1026	static SV *	1173	static SV *
1027	decode_cbor (SV string, CBOR cbor, char **offset_return)	1174	decode_cbor (SV string, CBOR cbor, char **offset_return)
1028	{	1175	{
1029	dec_t dec = { };	1176	dec_t dec = { 0 };
1030	SV *sv;	1177	SV *sv;
1031	STRLEN len;	1178	STRLEN len;
1032	char *data = SvPVbyte (string, len);	1179	char *data = SvPVbyte (string, len);
1033		1180
1034	if (len > cbor->max_size && cbor->max_size)	1181	if (len > cbor->max_size && cbor->max_size)
…		…
1048	if (dec.cur != dec.end && !dec.err)	1195	if (dec.cur != dec.end && !dec.err)
1049	dec.err = "garbage after CBOR object";	1196	dec.err = "garbage after CBOR object";
1050		1197
1051	if (dec.err)	1198	if (dec.err)
1052	{	1199	{
		1200	if (dec.shareable)
		1201	{
		1202	// need to break cyclic links, which whould all be in shareable
		1203	int i;
		1204	SV **svp;
		1205
		1206	for (i = av_len (dec.shareable) + 1; i--; )
		1207	if ((svp = av_fetch (dec.shareable, i, 0)))
		1208	sv_setsv (*svp, &PL_sv_undef);
		1209	}
		1210
1053	SvREFCNT_dec (sv);	1211	SvREFCNT_dec (sv);
1054	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1212	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1055	}	1213	}
1056		1214
1057	sv = sv_2mortal (sv);	1215	sv = sv_2mortal (sv);
1058		1216
1059	return sv;	1217	return sv;
1060	}	1218	}
1061		1219
		1220	/////////////////////////////////////////////////////////////////////////////
		1221	// incremental parser
		1222
		1223	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1224
		1225	// returns 0 for notyet, 1 for success or error
		1226	static int
		1227	incr_parse (CBOR self, SV cborstr)
		1228	{
		1229	STRLEN cur;
		1230	SvPV (cborstr, cur);
		1231
		1232	while (ecb_expect_true (self->incr_need <= cur))
		1233	{
		1234	// table of integer count bytes
		1235	static I8 incr_len[MINOR_MASK + 1] = {
		1236	0, 0, 0, 0, 0, 0, 0, 0,
		1237	0, 0, 0, 0, 0, 0, 0, 0,
		1238	0, 0, 0, 0, 0, 0, 0, 0,
		1239	1, 2, 4, 8,-1,-1,-1,-2
		1240	};
		1241
		1242	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1243	U8 m = *p & MINOR_MASK;
		1244	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1245	I8 ilen = incr_len[m];
		1246
		1247	self->incr_need = self->incr_pos + 1;
		1248
		1249	if (ecb_expect_false (ilen < 0))
		1250	{
		1251	if (m != MINOR_INDEF)
		1252	return 1; // error
		1253
		1254	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1255	{
		1256	if (count >= 0)
		1257	return 1; // error
		1258
		1259	count = 1;
		1260	}
		1261	else
		1262	{
		1263	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1264	count = -1;
		1265	}
		1266	}
		1267	else
		1268	{
		1269	self->incr_need += ilen;
		1270	if (ecb_expect_false (self->incr_need > cur))
		1271	return 0;
		1272
		1273	int major = *p >> MAJOR_SHIFT;
		1274
		1275	switch (major)
		1276	{
		1277	case MAJOR_TAG >> MAJOR_SHIFT:
		1278	++count; // tags merely prefix another value
		1279	break;
		1280
		1281	case MAJOR_BYTES >> MAJOR_SHIFT:
		1282	case MAJOR_TEXT >> MAJOR_SHIFT:
		1283	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1284	case MAJOR_MAP >> MAJOR_SHIFT:
		1285	{
		1286	UV len;
		1287
		1288	if (ecb_expect_false (ilen))
		1289	{
		1290	len = 0;
		1291
		1292	do {
		1293	len = (len << 8) \| *++p;
		1294	} while (--ilen);
		1295	}
		1296	else
		1297	len = m;
		1298
		1299	switch (major)
		1300	{
		1301	case MAJOR_BYTES >> MAJOR_SHIFT:
		1302	case MAJOR_TEXT >> MAJOR_SHIFT:
		1303	self->incr_need += len;
		1304	if (ecb_expect_false (self->incr_need > cur))
		1305	return 0;
		1306
		1307	break;
		1308
		1309	case MAJOR_MAP >> MAJOR_SHIFT:
		1310	len <<= 1;
		1311	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1312	if (len)
		1313	{
		1314	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1315	count = len + 1;
		1316	}
		1317	break;
		1318	}
		1319	}
		1320	}
		1321	}
		1322
		1323	self->incr_pos = self->incr_need;
		1324
		1325	if (count > 0)
		1326	{
		1327	while (!--count)
		1328	{
		1329	if (!AvFILLp (self->incr_count))
		1330	return 1; // done
		1331
		1332	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1333	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1334	}
		1335
		1336	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1337	}
		1338	}
		1339
		1340	return 0;
		1341	}
		1342
		1343
1062	/////////////////////////////////////////////////////////////////////////////	1344	/////////////////////////////////////////////////////////////////////////////
1063	// XS interface functions	1345	// XS interface functions
1064		1346
1065	MODULE = CBOR::XS PACKAGE = CBOR::XS	1347	MODULE = CBOR::XS PACKAGE = CBOR::XS
1066		1348
…		…
1073	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);	1355	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1074		1356
1075	types_true = get_bool ("Types::Serialiser::true" );	1357	types_true = get_bool ("Types::Serialiser::true" );
1076	types_false = get_bool ("Types::Serialiser::false");	1358	types_false = get_bool ("Types::Serialiser::false");
1077	types_error = get_bool ("Types::Serialiser::error");	1359	types_error = get_bool ("Types::Serialiser::error");
		1360
		1361	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1078		1362
1079	sv_cbor = newSVpv ("CBOR", 0);	1363	sv_cbor = newSVpv ("CBOR", 0);
1080	SvREADONLY_on (sv_cbor);	1364	SvREADONLY_on (sv_cbor);
1081	}	1365	}
1082		1366
…		…
1104	void shrink (CBOR *self, int enable = 1)	1388	void shrink (CBOR *self, int enable = 1)
1105	ALIAS:	1389	ALIAS:
1106	shrink = F_SHRINK	1390	shrink = F_SHRINK
1107	allow_unknown = F_ALLOW_UNKNOWN	1391	allow_unknown = F_ALLOW_UNKNOWN
1108	allow_sharing = F_ALLOW_SHARING	1392	allow_sharing = F_ALLOW_SHARING
1109	allow_stringref = F_ALLOW_STRINGREF	1393	allow_cycles = F_ALLOW_CYCLES
		1394	pack_strings = F_PACK_STRINGS
		1395	validate_utf8 = F_VALIDATE_UTF8
1110	PPCODE:	1396	PPCODE:
1111	{	1397	{
1112	if (enable)	1398	if (enable)
1113	self->flags \|= ix;	1399	self->flags \|= ix;
1114	else	1400	else
…		…
1120	void get_shrink (CBOR *self)	1406	void get_shrink (CBOR *self)
1121	ALIAS:	1407	ALIAS:
1122	get_shrink = F_SHRINK	1408	get_shrink = F_SHRINK
1123	get_allow_unknown = F_ALLOW_UNKNOWN	1409	get_allow_unknown = F_ALLOW_UNKNOWN
1124	get_allow_sharing = F_ALLOW_SHARING	1410	get_allow_sharing = F_ALLOW_SHARING
1125	get_allow_stringref = F_ALLOW_STRINGREF	1411	get_allow_cycles = F_ALLOW_CYCLES
		1412	get_pack_strings = F_PACK_STRINGS
		1413	get_validate_utf8 = F_VALIDATE_UTF8
1126	PPCODE:	1414	PPCODE:
1127	XPUSHs (boolSV (self->flags & ix));	1415	XPUSHs (boolSV (self->flags & ix));
1128		1416
1129	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1417	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130	PPCODE:	1418	PPCODE:
…		…
1143	XPUSHs (ST (0));	1431	XPUSHs (ST (0));
1144		1432
1145	int get_max_size (CBOR *self)	1433	int get_max_size (CBOR *self)
1146	CODE:	1434	CODE:
1147	RETVAL = self->max_size;	1435	RETVAL = self->max_size;
		1436	OUTPUT:
		1437	RETVAL
		1438
		1439	void filter (CBOR self, SV filter = 0)
		1440	PPCODE:
		1441	SvREFCNT_dec (self->filter);
		1442	self->filter = filter ? newSVsv (filter) : filter;
		1443	XPUSHs (ST (0));
		1444
		1445	SV get_filter (CBOR self)
		1446	CODE:
		1447	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148	OUTPUT:	1448	OUTPUT:
1149	RETVAL	1449	RETVAL
1150		1450
1151	void encode (CBOR self, SV scalar)	1451	void encode (CBOR self, SV scalar)
1152	PPCODE:	1452	PPCODE:
…		…
1167	EXTEND (SP, 2);	1467	EXTEND (SP, 2);
1168	PUSHs (sv);	1468	PUSHs (sv);
1169	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1469	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170	}	1470	}
1171		1471
		1472	void incr_parse (CBOR self, SV cborstr)
		1473	ALIAS:
		1474	incr_parse_multiple = 1
		1475	PPCODE:
		1476	{
		1477	if (SvUTF8 (cborstr))
		1478	sv_utf8_downgrade (cborstr, 0);
		1479
		1480	if (!self->incr_count)
		1481	{
		1482	self->incr_count = newAV ();
		1483	self->incr_pos = 0;
		1484	self->incr_need = 1;
		1485
		1486	av_push (self->incr_count, newSViv (1));
		1487	}
		1488
		1489	do
		1490	{
		1491	if (!incr_parse (self, cborstr))
		1492	{
		1493	if (self->incr_need > self->max_size && self->max_size)
		1494	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1495	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1496
		1497	break;
		1498	}
		1499
		1500	SV *sv;
		1501	char *offset;
		1502
		1503	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1504	XPUSHs (sv);
		1505
		1506	sv_chop (cborstr, offset);
		1507
		1508	av_clear (self->incr_count);
		1509	av_push (self->incr_count, newSViv (1));
		1510
		1511	self->incr_pos = 0;
		1512	self->incr_need = self->incr_pos + 1;
		1513	}
		1514	while (ix);
		1515	}
		1516
		1517	void incr_reset (CBOR *self)
		1518	CODE:
		1519	{
		1520	SvREFCNT_dec (self->incr_count);
		1521	self->incr_count = 0;
		1522	}
		1523
		1524	void DESTROY (CBOR *self)
		1525	PPCODE:
		1526	cbor_free (self);
		1527
1172	PROTOTYPES: ENABLE	1528	PROTOTYPES: ENABLE
1173		1529
1174	void encode_cbor (SV *scalar)	1530	void encode_cbor (SV *scalar)
		1531	ALIAS:
		1532	encode_cbor = 0
		1533	encode_cbor_sharing = F_ALLOW_SHARING
1175	PPCODE:	1534	PPCODE:
1176	{	1535	{
1177	CBOR cbor;	1536	CBOR cbor;
1178	cbor_init (&cbor);	1537	cbor_init (&cbor);
		1538	cbor.flags \|= ix;
1179	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1539	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180	XPUSHs (scalar);	1540	XPUSHs (scalar);
1181	}	1541	}
1182		1542
1183	void decode_cbor (SV *cborstr)	1543	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.25 by root, Fri Nov 22 05:54:07 2013 UTC vs. Revision 1.50 by root, Thu Feb 25 02:29:22 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.25 by root, Fri Nov 22 05:54:07 2013 UTC vs.
Revision 1.50 by root, Thu Feb 25 02:29:22 2016 UTC