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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.22 by root, Wed Nov 20 15:05:46 2013 UTC vs.
Revision 1.51 by root, Sun Apr 24 13:15:19 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
		103	#define F_ALLOW_CYCLES 0x00000008UL
62	#define F_DEDUP_STRINGS 0x00000008UL //TODO	104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_UTF8_STRINGS 0x00000020UL
		106	#define F_VALIDATE_UTF8 0x00000040UL
63		107
64	#define INIT_SIZE 32 // initial scalar size to be allocated	108	#define INIT_SIZE 32 // initial scalar size to be allocated
65		109
66	#define SB do {	110	#define SB do {
67	#define SE } while (0)	111	#define SE } while (0)
…		…
79	# define CBOR_SLOW 0	123	# define CBOR_SLOW 0
80	# define CBOR_STASH cbor_stash	124	# define CBOR_STASH cbor_stash
81	#endif	125	#endif
82		126
83	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::	127	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;	128	static SV types_true, types_false, types_error, sv_cbor, *default_filter;
85		129
86	typedef struct {	130	typedef struct {
87	U32 flags;	131	U32 flags;
88	U32 max_depth;	132	U32 max_depth;
89	STRLEN max_size;	133	STRLEN max_size;
		134	SV *filter;
		135
		136	// for the incremental parser
		137	STRLEN incr_pos; // the current offset into the text
		138	STRLEN incr_need; // minimum bytes needed to decode
		139	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90	} CBOR;	140	} CBOR;
91		141
92	ecb_inline void	142	ecb_inline void
93	cbor_init (CBOR *cbor)	143	cbor_init (CBOR *cbor)
94	{	144	{
95	Zero (cbor, 1, CBOR);	145	Zero (cbor, 1, CBOR);
96	cbor->max_depth = 512;	146	cbor->max_depth = 512;
		147	}
		148
		149	ecb_inline void
		150	cbor_free (CBOR *cbor)
		151	{
		152	SvREFCNT_dec (cbor->filter);
		153	SvREFCNT_dec (cbor->incr_count);
97	}	154	}
98		155
99	/////////////////////////////////////////////////////////////////////////////	156	/////////////////////////////////////////////////////////////////////////////
100	// utility functions	157	// utility functions
101		158
…		…
131	{	188	{
132	return idx > 23	189	return idx > 23
133	? idx > 0xffU	190	? idx > 0xffU
134	? idx > 0xffffU	191	? idx > 0xffffU
135	? idx > 0xffffffffU	192	? idx > 0xffffffffU
		193	? 11
136	? 7	194	: 7
137	: 6
138	: 5	195	: 5
139	: 4	196	: 4
140	: 3;	197	: 3;
141	}	198	}
142		199
…		…
179	static void	236	static void
180	encode_uint (enc_t *enc, int major, UV len)	237	encode_uint (enc_t *enc, int major, UV len)
181	{	238	{
182	need (enc, 9);	239	need (enc, 9);
183		240
184	if (len < 24)	241	if (ecb_expect_true (len < LENGTH_EXT1))
185	*enc->cur++ = major \| len;	242	*enc->cur++ = major \| len;
186	else if (len <= 0xff)	243	else if (ecb_expect_true (len <= 0xffU))
187	{	244	{
188	*enc->cur++ = major \| 24;	245	*enc->cur++ = major \| LENGTH_EXT1;
189	*enc->cur++ = len;	246	*enc->cur++ = len;
190	}	247	}
191	else if (len <= 0xffff)	248	else if (len <= 0xffffU)
192	{	249	{
193	*enc->cur++ = major \| 25;	250	*enc->cur++ = major \| LENGTH_EXT2;
194	*enc->cur++ = len >> 8;	251	*enc->cur++ = len >> 8;
195	*enc->cur++ = len;	252	*enc->cur++ = len;
196	}	253	}
197	else if (len <= 0xffffffff)	254	else if (len <= 0xffffffffU)
198	{	255	{
199	*enc->cur++ = major \| 26;	256	*enc->cur++ = major \| LENGTH_EXT4;
200	*enc->cur++ = len >> 24;	257	*enc->cur++ = len >> 24;
201	*enc->cur++ = len >> 16;	258	*enc->cur++ = len >> 16;
202	*enc->cur++ = len >> 8;	259	*enc->cur++ = len >> 8;
203	*enc->cur++ = len;	260	*enc->cur++ = len;
204	}	261	}
205	else	262	else
206	{	263	{
207	*enc->cur++ = major \| 27;	264	*enc->cur++ = major \| LENGTH_EXT8;
208	*enc->cur++ = len >> 56;	265	*enc->cur++ = len >> 56;
209	*enc->cur++ = len >> 48;	266	*enc->cur++ = len >> 48;
210	*enc->cur++ = len >> 40;	267	*enc->cur++ = len >> 40;
211	*enc->cur++ = len >> 32;	268	*enc->cur++ = len >> 32;
212	*enc->cur++ = len >> 24;	269	*enc->cur++ = len >> 24;
…		…
217	}	274	}
218		275
219	ecb_inline void	276	ecb_inline void
220	encode_tag (enc_t *enc, UV tag)	277	encode_tag (enc_t *enc, UV tag)
221	{	278	{
222	encode_uint (enc, 0xc0, tag);	279	encode_uint (enc, MAJOR_TAG, tag);
		280	}
		281
		282	ecb_inline void
		283	encode_str (enc_t enc, int utf8, char str, STRLEN len)
		284	{
		285	if (ecb_expect_false (enc->cbor.flags & F_UTF8_STRINGS))
		286	if (!utf8)
		287	{
		288	SV *sv = sv_newmortal ();
		289	char *s; STRLEN l;
		290
		291	sv_setpvn (sv, str, len);
		292
		293	s = SvPVutf8 (sv, l);
		294	encode_str (enc, 1, s, l);
		295	return;
		296	}
		297
		298	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		299	need (enc, len);
		300	memcpy (enc->cur, str, len);
		301	enc->cur += len;
223	}	302	}
224		303
225	static void	304	static void
226	encode_str (enc_t enc, int utf8, char str, STRLEN len)	305	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
227	{	306	{
228	if (ecb_expect_false (enc->cbor.flags & F_DEDUP_STRINGS))	307	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229	{	308	{
230	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	309	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231		310
232	if (SvOK (*svp))	311	if (SvOK (*svp))
233	{	312	{
234	// already registered, use stringref	313	// already registered, use stringref
235	encode_tag (enc, CBOR_TAG_STRINGREF);	314	encode_tag (enc, CBOR_TAG_STRINGREF);
236	encode_uint (enc, 0x00, SvUV (*svp));	315	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237	return;	316	return;
238	}	317	}
239	else if (len >= minimum_string_length (enc->stringref_idx))	318	else if (len >= minimum_string_length (enc->stringref_idx))
240	{	319	{
241	// register only	320	// register only
242	sv_setuv (*svp, enc->stringref_idx);	321	sv_setuv (*svp, enc->stringref_idx);
243	++enc->stringref_idx;	322	++enc->stringref_idx;
244	}	323	}
245	}	324	}
246		325
247	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	326	encode_str (enc, utf8, str, len);
248	need (enc, len);
249	memcpy (enc->cur, str, len);
250	enc->cur += len;
251	}	327	}
252		328
253	static void encode_sv (enc_t enc, SV sv);	329	static void encode_sv (enc_t enc, SV sv);
254		330
255	static void	331	static void
…		…
260	if (enc->depth >= enc->cbor.max_depth)	336	if (enc->depth >= enc->cbor.max_depth)
261	croak (ERR_NESTING_EXCEEDED);	337	croak (ERR_NESTING_EXCEEDED);
262		338
263	++enc->depth;	339	++enc->depth;
264		340
265	encode_uint (enc, 0x80, len + 1);	341	encode_uint (enc, MAJOR_ARRAY, len + 1);
266		342
		343	if (SvMAGICAL (av))
267	for (i = 0; i <= len; ++i)	344	for (i = 0; i <= len; ++i)
268	{	345	{
269	SV **svp = av_fetch (av, i, 0);	346	SV **svp = av_fetch (av, i, 0);
270	encode_sv (enc, svp ? *svp : &PL_sv_undef);	347	encode_sv (enc, svp ? *svp : &PL_sv_undef);
271	}	348	}
		349	else
		350	for (i = 0; i <= len; ++i)
		351	{
		352	SV *sv = AvARRAY (av)[i];
		353	encode_sv (enc, sv ? sv : &PL_sv_undef);
		354	}
272		355
273	--enc->depth;	356	--enc->depth;
274	}
275
276	ecb_inline void
277	encode_he (enc_t enc, HE he)
278	{
279	}	357	}
280		358
281	static void	359	static void
282	encode_hv (enc_t enc, HV hv)	360	encode_hv (enc_t enc, HV hv)
283	{	361	{
…		…
290		368
291	int pairs = hv_iterinit (hv);	369	int pairs = hv_iterinit (hv);
292	int mg = SvMAGICAL (hv);	370	int mg = SvMAGICAL (hv);
293		371
294	if (mg)	372	if (mg)
295	encode_ch (enc, 0xa0 \| 31);	373	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
296	else	374	else
297	encode_uint (enc, 0xa0, pairs);	375	encode_uint (enc, MAJOR_MAP, pairs);
298		376
299	while ((he = hv_iternext (hv)))	377	while ((he = hv_iternext (hv)))
300	{	378	{
301	if (HeKLEN (he) == HEf_SVKEY)	379	if (HeKLEN (he) == HEf_SVKEY)
302	encode_sv (enc, HeSVKEY (he));	380	encode_sv (enc, HeSVKEY (he));
303	else	381	else
304	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	382	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305		383
306	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	384	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307	}	385	}
308		386
309	if (mg)	387	if (mg)
310	encode_ch (enc, 0xe0 \| 31);	388	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
311		389
312	--enc->depth;	390	--enc->depth;
313	}	391	}
314		392
315	// encode objects, arrays and special \0=false and \1=true values.	393	// encode objects, arrays and special \0=false and \1=true values.
316	static void	394	static void
317	encode_rv (enc_t enc, SV sv)	395	encode_rv (enc_t enc, SV sv)
318	{	396	{
319	SvGETMAGIC (sv);	397	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		398
343	svtype svt = SvTYPE (sv);	399	svtype svt = SvTYPE (sv);
344		400
345	if (ecb_expect_false (SvOBJECT (sv)))	401	if (ecb_expect_false (SvOBJECT (sv)))
346	{	402	{
…		…
353	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	409	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
354	? cbor_tagged_stash	410	? cbor_tagged_stash
355	: gv_stashpv ("CBOR::XS::Tagged" , 1);	411	: gv_stashpv ("CBOR::XS::Tagged" , 1);
356		412
357	HV *stash = SvSTASH (sv);	413	HV *stash = SvSTASH (sv);
358	GV *method;
359		414
360	if (stash == boolean_stash)	415	if (stash == boolean_stash)
361	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	416	{
		417	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		418	return;
		419	}
362	else if (stash == error_stash)	420	else if (stash == error_stash)
363	encode_ch (enc, 0xe0 \| 23);	421	{
		422	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		423	return;
		424	}
364	else if (stash == tagged_stash)	425	else if (stash == tagged_stash)
365	{	426	{
366	if (svt != SVt_PVAV)	427	if (svt != SVt_PVAV)
367	croak ("encountered CBOR::XS::Tagged object that isn't an array");	428	croak ("encountered CBOR::XS::Tagged object that isn't an array");
368		429
369	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	430	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
370	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	431	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		432
		433	return;
		434	}
		435	}
		436
		437	if (ecb_expect_false (SvREFCNT (sv) > 1)
		438	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		439	{
		440	if (!enc->shareable)
		441	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		442
		443	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		444
		445	if (SvOK (*svp))
371	}	446	{
		447	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		448	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		449	return;
		450	}
		451	else
		452	{
		453	sv_setuv (*svp, enc->shareable_idx);
		454	++enc->shareable_idx;
		455	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		456	}
		457	}
		458
		459	if (ecb_expect_false (SvOBJECT (sv)))
		460	{
		461	HV *stash = SvSTASH (sv);
		462	GV *method;
		463
372	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	464	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373	{	465	{
374	dSP;	466	dSP;
375		467
376	ENTER; SAVETMPS; PUSHMARK (SP);	468	ENTER; SAVETMPS;
		469	PUSHMARK (SP);
377	// we re-bless the reference to get overload and other niceties right	470	// we re-bless the reference to get overload and other niceties right
378	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	471	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379		472
380	PUTBACK;	473	PUTBACK;
381	// G_SCALAR ensures that return value is 1	474	// G_SCALAR ensures that return value is 1
…		…
394	}	487	}
395	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	488	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396	{	489	{
397	dSP;	490	dSP;
398		491
399	ENTER; SAVETMPS; PUSHMARK (SP);	492	ENTER; SAVETMPS;
		493	SAVESTACK_POS ();
		494	PUSHMARK (SP);
400	EXTEND (SP, 2);	495	EXTEND (SP, 2);
401	// we re-bless the reference to get overload and other niceties right	496	// we re-bless the reference to get overload and other niceties right
402	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	497	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403	PUSHs (sv_cbor);	498	PUSHs (sv_cbor);
404		499
…		…
409	// catch this surprisingly common error	504	// catch this surprisingly common error
410	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	505	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));	506	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412		507
413	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	508	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414	encode_uint (enc, 0x80, count + 1);	509	encode_uint (enc, MAJOR_ARRAY, count + 1);
415	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	510	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416		511
417	while (count)	512	while (count)
418	encode_sv (enc, SP[1 - count--]);	513	encode_sv (enc, SP[1 - count--]);
419		514
420	PUTBACK;	515	PUTBACK;
…		…
441	{	536	{
442	double nv = SvNVX (sv);	537	double nv = SvNVX (sv);
443		538
444	need (enc, 9);	539	need (enc, 9);
445		540
446	if (ecb_expect_false (nv == (U32)nv))	541	if (ecb_expect_false (nv == (NV)(U32)nv))
447	encode_uint (enc, 0x00, (U32)nv);	542	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448	//TODO: maybe I32?	543	//TODO: maybe I32?
449	else if (ecb_expect_false (nv == (float)nv))	544	else if (ecb_expect_false (nv == (float)nv))
450	{	545	{
451	uint32_t fp = ecb_float_to_binary32 (nv);	546	uint32_t fp = ecb_float_to_binary32 (nv);
452		547
453	*enc->cur++ = 0xe0 \| 26;	548	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
454		549
455	if (!ecb_big_endian ())	550	if (!ecb_big_endian ())
456	fp = ecb_bswap32 (fp);	551	fp = ecb_bswap32 (fp);
457		552
458	memcpy (enc->cur, &fp, 4);	553	memcpy (enc->cur, &fp, 4);
…		…
460	}	555	}
461	else	556	else
462	{	557	{
463	uint64_t fp = ecb_double_to_binary64 (nv);	558	uint64_t fp = ecb_double_to_binary64 (nv);
464		559
465	*enc->cur++ = 0xe0 \| 27;	560	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
466		561
467	if (!ecb_big_endian ())	562	if (!ecb_big_endian ())
468	fp = ecb_bswap64 (fp);	563	fp = ecb_bswap64 (fp);
469		564
470	memcpy (enc->cur, &fp, 8);	565	memcpy (enc->cur, &fp, 8);
…		…
479		574
480	if (SvPOKp (sv))	575	if (SvPOKp (sv))
481	{	576	{
482	STRLEN len;	577	STRLEN len;
483	char *str = SvPV (sv, len);	578	char *str = SvPV (sv, len);
484	encode_str (enc, SvUTF8 (sv), str, len);	579	encode_strref (enc, SvUTF8 (sv), str, len);
485	}	580	}
486	else if (SvNOKp (sv))	581	else if (SvNOKp (sv))
487	encode_nv (enc, sv);	582	encode_nv (enc, sv);
488	else if (SvIOKp (sv))	583	else if (SvIOKp (sv))
489	{	584	{
490	if (SvIsUV (sv))	585	if (SvIsUV (sv))
491	encode_uint (enc, 0x00, SvUVX (sv));	586	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
492	else if (SvIVX (sv) >= 0)	587	else if (SvIVX (sv) >= 0)
493	encode_uint (enc, 0x00, SvIVX (sv));	588	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
494	else	589	else
495	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	590	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
496	}	591	}
497	else if (SvROK (sv))	592	else if (SvROK (sv))
498	encode_rv (enc, SvRV (sv));	593	encode_rv (enc, SvRV (sv));
499	else if (!SvOK (sv))	594	else if (!SvOK (sv))
500	encode_ch (enc, 0xe0 \| 22);	595	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
501	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	596	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502	encode_ch (enc, 0xe0 \| 23);	597	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
503	else	598	else
504	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	599	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	600	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506	}	601	}
507		602
508	static SV *	603	static SV *
509	encode_cbor (SV scalar, CBOR cbor)	604	encode_cbor (SV scalar, CBOR cbor)
510	{	605	{
511	enc_t enc = { };	606	enc_t enc = { 0 };
512		607
513	enc.cbor = *cbor;	608	enc.cbor = *cbor;
514	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	609	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515	enc.cur = SvPVX (enc.sv);	610	enc.cur = SvPVX (enc.sv);
516	enc.end = SvEND (enc.sv);	611	enc.end = SvEND (enc.sv);
517		612
518	SvPOK_only (enc.sv);	613	SvPOK_only (enc.sv);
519		614
520	if (cbor->flags & F_DEDUP_STRINGS)	615	if (cbor->flags & F_PACK_STRINGS)
521	{	616	{
522	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	617	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	618	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
524	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	619	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
525	}	620	}
…		…
547	CBOR cbor;	642	CBOR cbor;
548	U32 depth; // recursion depth	643	U32 depth; // recursion depth
549	U32 maxdepth; // recursion depth limit	644	U32 maxdepth; // recursion depth limit
550	AV *shareable;	645	AV *shareable;
551	AV *stringref;	646	AV *stringref;
		647	SV *decode_tagged;
552	} dec_t;	648	} dec_t;
553		649
554	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	650	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
555		651
556	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	652	#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	655	#define DEC_DEC_DEPTH --dec->depth
560		656
561	static UV	657	static UV
562	decode_uint (dec_t *dec)	658	decode_uint (dec_t *dec)
563	{	659	{
564	switch (*dec->cur & 31)	660	U8 m = *dec->cur & MINOR_MASK;
565	{	661	++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		662
571	case 24:	663	if (ecb_expect_true (m < LENGTH_EXT1))
		664	return m;
		665	else if (ecb_expect_true (m == LENGTH_EXT1))
		666	{
572	WANT (2);	667	WANT (1);
573	dec->cur += 2;	668	dec->cur += 1;
574	return dec->cur[-1];	669	return dec->cur[-1];
575		670	}
576	case 25:	671	else if (ecb_expect_true (m == LENGTH_EXT2))
		672	{
577	WANT (3);	673	WANT (2);
578	dec->cur += 3;	674	dec->cur += 2;
579	return (((UV)dec->cur[-2]) << 8)	675	return (((UV)dec->cur[-2]) << 8)
580	\| ((UV)dec->cur[-1]);	676	\| ((UV)dec->cur[-1]);
581		677	}
582	case 26:	678	else if (ecb_expect_true (m == LENGTH_EXT4))
		679	{
583	WANT (5);	680	WANT (4);
584	dec->cur += 5;	681	dec->cur += 4;
585	return (((UV)dec->cur[-4]) << 24)	682	return (((UV)dec->cur[-4]) << 24)
586	\| (((UV)dec->cur[-3]) << 16)	683	\| (((UV)dec->cur[-3]) << 16)
587	\| (((UV)dec->cur[-2]) << 8)	684	\| (((UV)dec->cur[-2]) << 8)
588	\| ((UV)dec->cur[-1]);	685	\| ((UV)dec->cur[-1]);
589		686	}
590	case 27:	687	else if (ecb_expect_true (m == LENGTH_EXT8))
		688	{
591	WANT (9);	689	WANT (8);
592	dec->cur += 9;	690	dec->cur += 8;
		691
		692	return
		693	#if UVSIZE < 8
		694	0
		695	#else
593	return (((UV)dec->cur[-8]) << 56)	696	(((UV)dec->cur[-8]) << 56)
594	\| (((UV)dec->cur[-7]) << 48)	697	\| (((UV)dec->cur[-7]) << 48)
595	\| (((UV)dec->cur[-6]) << 40)	698	\| (((UV)dec->cur[-6]) << 40)
596	\| (((UV)dec->cur[-5]) << 32)	699	\| (((UV)dec->cur[-5]) << 32)
		700	#endif
597	\| (((UV)dec->cur[-4]) << 24)	701	\| (((UV)dec->cur[-4]) << 24)
598	\| (((UV)dec->cur[-3]) << 16)	702	\| (((UV)dec->cur[-3]) << 16)
599	\| (((UV)dec->cur[-2]) << 8)	703	\| (((UV)dec->cur[-2]) << 8)
600	\| ((UV)dec->cur[-1]);	704	\| ((UV)dec->cur[-1]);
601		705	}
602	default:	706	else
603	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	707	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604	}
605		708
606	fail:	709	fail:
607	return 0;	710	return 0;
608	}	711	}
609		712
…		…
614	{	717	{
615	AV *av = newAV ();	718	AV *av = newAV ();
616		719
617	DEC_INC_DEPTH;	720	DEC_INC_DEPTH;
618		721
619	if ((*dec->cur & 31) == 31)	722	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
620	{	723	{
621	++dec->cur;	724	++dec->cur;
622		725
623	for (;;)	726	for (;;)
624	{	727	{
625	WANT (1);	728	WANT (1);
626		729
627	if (*dec->cur == (0xe0 \| 31))	730	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
628	{	731	{
629	++dec->cur;	732	++dec->cur;
630	break;	733	break;
631	}	734	}
632		735
…		…
635	}	738	}
636	else	739	else
637	{	740	{
638	int i, len = decode_uint (dec);	741	int i, len = decode_uint (dec);
639		742
		743	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);	744	av_fill (av, len - 1);
641		745
642	for (i = 0; i < len; ++i)	746	for (i = 0; i < len; ++i)
643	AvARRAY (av)[i] = decode_sv (dec);	747	AvARRAY (av)[i] = decode_sv (dec);
644	}	748	}
…		…
656	decode_he (dec_t dec, HV hv)	760	decode_he (dec_t dec, HV hv)
657	{	761	{
658	// for speed reasons, we specialcase single-string	762	// for speed reasons, we specialcase single-string
659	// byte or utf-8 strings as keys, but only when !stringref	763	// byte or utf-8 strings as keys, but only when !stringref
660		764
661	if (expect_true (!dec->stringref))	765	if (ecb_expect_true (!dec->stringref))
662	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	766	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663	{	767	{
664	I32 len = decode_uint (dec);	768	I32 len = decode_uint (dec);
665	char key = (char )dec->cur;	769	char key = (char )dec->cur;
666		770
		771	WANT (len);
667	dec->cur += len;	772	dec->cur += len;
668		773
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);	774	hv_store (hv, key, len, decode_sv (dec), 0);
673		775
674	return;	776	return;
675	}	777	}
676	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	778	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677	{	779	{
678	I32 len = decode_uint (dec);	780	I32 len = decode_uint (dec);
679	char key = (char )dec->cur;	781	char key = (char )dec->cur;
680		782
		783	WANT (len);
681	dec->cur += len;	784	dec->cur += len;
682		785
683	if (ecb_expect_false (dec->stringref))	786	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	787	if (!is_utf8_string (key, len))
		788	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685		789
686	hv_store (hv, key, -len, decode_sv (dec), 0);	790	hv_store (hv, key, -len, decode_sv (dec), 0);
687		791
688	return;	792	return;
689	}	793	}
…		…
691	SV *k = decode_sv (dec);	795	SV *k = decode_sv (dec);
692	SV *v = decode_sv (dec);	796	SV *v = decode_sv (dec);
693		797
694	hv_store_ent (hv, k, v, 0);	798	hv_store_ent (hv, k, v, 0);
695	SvREFCNT_dec (k);	799	SvREFCNT_dec (k);
		800
		801	fail:
		802	;
696	}	803	}
697		804
698	static SV *	805	static SV *
699	decode_hv (dec_t *dec)	806	decode_hv (dec_t *dec)
700	{	807	{
701	HV *hv = newHV ();	808	HV *hv = newHV ();
702		809
703	DEC_INC_DEPTH;	810	DEC_INC_DEPTH;
704		811
705	if ((*dec->cur & 31) == 31)	812	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
706	{	813	{
707	++dec->cur;	814	++dec->cur;
708		815
709	for (;;)	816	for (;;)
710	{	817	{
711	WANT (1);	818	WANT (1);
712		819
713	if (*dec->cur == (0xe0 \| 31))	820	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
714	{	821	{
715	++dec->cur;	822	++dec->cur;
716	break;	823	break;
717	}	824	}
718		825
…		…
739	static SV *	846	static SV *
740	decode_str (dec_t *dec, int utf8)	847	decode_str (dec_t *dec, int utf8)
741	{	848	{
742	SV *sv = 0;	849	SV *sv = 0;
743		850
744	if ((*dec->cur & 31) == 31)	851	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
745	{	852	{
		853	// indefinite length strings
746	++dec->cur;	854	++dec->cur;
747		855
		856	U8 major = *dec->cur & MAJOR_MISC;
		857
748	sv = newSVpvn ("", 0);	858	sv = newSVpvn ("", 0);
749		859
750	// not very fast, and certainly not robust against illegal input
751	for (;;)	860	for (;;)
752	{	861	{
753	WANT (1);	862	WANT (1);
754		863
755	if (*dec->cur == (0xe0 \| 31))	864	if ((*dec->cur - major) > LENGTH_EXT8)
		865	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
756	{	866	{
757	++dec->cur;	867	++dec->cur;
758	break;	868	break;
759	}	869	}
		870	else
		871	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760		872
761	sv_catsv (sv, decode_sv (dec));	873	STRLEN len = decode_uint (dec);
		874
		875	WANT (len);
		876	sv_catpvn (sv, dec->cur, len);
		877	dec->cur += len;
762	}	878	}
763	}	879	}
764	else	880	else
765	{	881	{
766	STRLEN len = decode_uint (dec);	882	STRLEN len = decode_uint (dec);
767		883
768	WANT (len);	884	WANT (len);
769	sv = newSVpvn (dec->cur, len);	885	sv = newSVpvn (dec->cur, len);
770	dec->cur += len;	886	dec->cur += len;
		887
		888	if (ecb_expect_false (dec->stringref)
		889	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
		890	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
771	}	891	}
772		892
773	if (utf8)	893	if (utf8)
		894	{
		895	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		896	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		897	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		898
774	SvUTF8_on (sv);	899	SvUTF8_on (sv);
775		900	}
776	if (ecb_expect_false (dec->stringref)
777	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
778	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
779		901
780	return sv;	902	return sv;
781		903
782	fail:	904	fail:
783	SvREFCNT_dec (sv);	905	SvREFCNT_dec (sv);
…		…
815	}	937	}
816	break;	938	break;
817		939
818	case CBOR_TAG_STRINGREF:	940	case CBOR_TAG_STRINGREF:
819	{	941	{
820	if ((*dec->cur >> 5) != 0)	942	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	943	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822		944
823	UV idx = decode_uint (dec);	945	UV idx = decode_uint (dec);
824		946
825	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	947	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
832	case CBOR_TAG_VALUE_SHAREABLE:	954	case CBOR_TAG_VALUE_SHAREABLE:
833	{	955	{
834	if (ecb_expect_false (!dec->shareable))	956	if (ecb_expect_false (!dec->shareable))
835	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	957	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
836		958
		959	if (dec->cbor.flags & F_ALLOW_CYCLES)
		960	{
837	sv = newSV (0);	961	sv = newSV (0);
838	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	962	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839		963
840	SV *osv = decode_sv (dec);	964	SV *osv = decode_sv (dec);
841	sv_setsv (sv, osv);	965	sv_setsv (sv, osv);
842	SvREFCNT_dec_NN (osv);	966	SvREFCNT_dec_NN (osv);
		967	}
		968	else
		969	{
		970	av_push (dec->shareable, &PL_sv_undef);
		971	int idx = AvFILLp (dec->shareable);
		972	sv = decode_sv (dec);
		973	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		974	}
843	}	975	}
844	break;	976	break;
845		977
846	case CBOR_TAG_VALUE_SHAREDREF:	978	case CBOR_TAG_VALUE_SHAREDREF:
847	{	979	{
848	if ((*dec->cur >> 5) != 0)	980	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	981	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850		982
851	UV idx = decode_uint (dec);	983	UV idx = decode_uint (dec);
852		984
853	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	985	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
854	ERR ("corrupted CBOR data (sharedref index out of bounds)");	986	ERR ("corrupted CBOR data (sharedref index out of bounds)");
855		987
856	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	988	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		989
		990	if (sv == &PL_sv_undef)
		991	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857	}	992	}
858	break;	993	break;
859		994
860	case CBOR_TAG_PERL_OBJECT:	995	case CBOR_TAG_PERL_OBJECT:
861	{	996	{
…		…
876	if (!method)	1011	if (!method)
877	ERR ("cannot decode perl-object (package does not have a THAW method)");	1012	ERR ("cannot decode perl-object (package does not have a THAW method)");
878		1013
879	dSP;	1014	dSP;
880		1015
881	ENTER; SAVETMPS; PUSHMARK (SP);	1016	ENTER; SAVETMPS;
		1017	PUSHMARK (SP);
882	EXTEND (SP, len + 1);	1018	EXTEND (SP, len + 1);
883	// we re-bless the reference to get overload and other niceties right	1019	// we re-bless the reference to get overload and other niceties right
884	PUSHs (*av_fetch (av, 0, 1));	1020	PUSHs (*av_fetch (av, 0, 1));
885	PUSHs (sv_cbor);	1021	PUSHs (sv_cbor);
886		1022
…		…
910		1046
911	default:	1047	default:
912	{	1048	{
913	sv = decode_sv (dec);	1049	sv = decode_sv (dec);
914		1050
		1051	dSP;
		1052	ENTER; SAVETMPS;
		1053	SAVESTACK_POS ();
		1054	PUSHMARK (SP);
		1055	EXTEND (SP, 2);
		1056	PUSHs (newSVuv (tag));
		1057	PUSHs (sv);
		1058
		1059	PUTBACK;
		1060	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
		1061	SPAGAIN;
		1062
		1063	if (SvTRUE (ERRSV))
		1064	{
		1065	FREETMPS; LEAVE;
		1066	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
		1067	}
		1068
		1069	if (count)
		1070	{
		1071	SvREFCNT_dec (sv);
		1072	sv = SvREFCNT_inc (POPs);
		1073	}
		1074	else
		1075	{
915	AV *av = newAV ();	1076	AV *av = newAV ();
916	av_push (av, newSVuv (tag));	1077	av_push (av, newSVuv (tag));
917	av_push (av, sv);	1078	av_push (av, sv);
918		1079
919	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1080	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
920	? cbor_tagged_stash	1081	? cbor_tagged_stash
921	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1082	: gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1083	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1084	}
		1085
		1086	PUTBACK;
		1087
		1088	FREETMPS; LEAVE;
924	}	1089	}
925	break;	1090	break;
926	}	1091	}
927		1092
928	return sv;	1093	return sv;
…		…
935	static SV *	1100	static SV *
936	decode_sv (dec_t *dec)	1101	decode_sv (dec_t *dec)
937	{	1102	{
938	WANT (1);	1103	WANT (1);
939		1104
940	switch (*dec->cur >> 5)	1105	switch (*dec->cur >> MAJOR_SHIFT)
941	{	1106	{
942	case 0: // unsigned int	1107	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943	return newSVuv (decode_uint (dec));	1108	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944	case 1: // negative int	1109	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945	return newSViv (-1 - (IV)decode_uint (dec));	1110	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946	case 2: // octet string	1111	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947	return decode_str (dec, 0);	1112	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948	case 3: // utf-8 string	1113	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949	return decode_str (dec, 1);	1114
950	case 4: // array	1115	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)	1116	switch (*dec->cur++ & MINOR_MASK)
958	{	1117	{
959	case 20:	1118	case SIMPLE_FALSE:
960	#if CBOR_SLOW	1119	#if CBOR_SLOW
961	types_false = get_bool ("Types::Serialiser::false");	1120	types_false = get_bool ("Types::Serialiser::false");
962	#endif	1121	#endif
963	return newSVsv (types_false);	1122	return newSVsv (types_false);
964	case 21:	1123	case SIMPLE_TRUE:
965	#if CBOR_SLOW	1124	#if CBOR_SLOW
966	types_true = get_bool ("Types::Serialiser::true");	1125	types_true = get_bool ("Types::Serialiser::true");
967	#endif	1126	#endif
968	return newSVsv (types_true);	1127	return newSVsv (types_true);
969	case 22:	1128	case SIMPLE_NULL:
970	return newSVsv (&PL_sv_undef);	1129	return newSVsv (&PL_sv_undef);
971	case 23:	1130	case SIMPLE_UNDEF:
972	#if CBOR_SLOW	1131	#if CBOR_SLOW
973	types_error = get_bool ("Types::Serialiser::error");	1132	types_error = get_bool ("Types::Serialiser::error");
974	#endif	1133	#endif
975	return newSVsv (types_error);	1134	return newSVsv (types_error);
976		1135
977	case 25:	1136	case MISC_FLOAT16:
978	{	1137	{
979	WANT (2);	1138	WANT (2);
980		1139
981	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1140	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
982	dec->cur += 2;	1141	dec->cur += 2;
983		1142
984	return newSVnv (ecb_binary16_to_float (fp));	1143	return newSVnv (ecb_binary16_to_float (fp));
985	}	1144	}
986		1145
987	case 26:	1146	case MISC_FLOAT32:
988	{	1147	{
989	uint32_t fp;	1148	uint32_t fp;
990	WANT (4);	1149	WANT (4);
991	memcpy (&fp, dec->cur, 4);	1150	memcpy (&fp, dec->cur, 4);
992	dec->cur += 4;	1151	dec->cur += 4;
…		…
995	fp = ecb_bswap32 (fp);	1154	fp = ecb_bswap32 (fp);
996		1155
997	return newSVnv (ecb_binary32_to_float (fp));	1156	return newSVnv (ecb_binary32_to_float (fp));
998	}	1157	}
999		1158
1000	case 27:	1159	case MISC_FLOAT64:
1001	{	1160	{
1002	uint64_t fp;	1161	uint64_t fp;
1003	WANT (8);	1162	WANT (8);
1004	memcpy (&fp, dec->cur, 8);	1163	memcpy (&fp, dec->cur, 8);
1005	dec->cur += 8;	1164	dec->cur += 8;
…		…
1008	fp = ecb_bswap64 (fp);	1167	fp = ecb_bswap64 (fp);
1009		1168
1010	return newSVnv (ecb_binary64_to_double (fp));	1169	return newSVnv (ecb_binary64_to_double (fp));
1011	}	1170	}
1012		1171
1013	// 0..19 unassigned	1172	// 0..19 unassigned simple
1014	// 24 reserved + unassigned (reserved values are not encodable)	1173	// 24 reserved + unassigned simple (reserved values are not encodable)
		1174	// 28-30 unassigned misc
		1175	// 31 break code
1015	default:	1176	default:
1016	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1177	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017	}	1178	}
1018		1179
1019	break;	1180	break;
1020	}	1181	}
1021		1182
…		…
1024	}	1185	}
1025		1186
1026	static SV *	1187	static SV *
1027	decode_cbor (SV string, CBOR cbor, char **offset_return)	1188	decode_cbor (SV string, CBOR cbor, char **offset_return)
1028	{	1189	{
1029	dec_t dec = { };	1190	dec_t dec = { 0 };
1030	SV *sv;	1191	SV *sv;
1031	STRLEN len;	1192	STRLEN len;
1032	char *data = SvPVbyte (string, len);	1193	char *data = SvPVbyte (string, len);
1033		1194
1034	if (len > cbor->max_size && cbor->max_size)	1195	if (len > cbor->max_size && cbor->max_size)
…		…
1048	if (dec.cur != dec.end && !dec.err)	1209	if (dec.cur != dec.end && !dec.err)
1049	dec.err = "garbage after CBOR object";	1210	dec.err = "garbage after CBOR object";
1050		1211
1051	if (dec.err)	1212	if (dec.err)
1052	{	1213	{
		1214	if (dec.shareable)
		1215	{
		1216	// need to break cyclic links, which whould all be in shareable
		1217	int i;
		1218	SV **svp;
		1219
		1220	for (i = av_len (dec.shareable) + 1; i--; )
		1221	if ((svp = av_fetch (dec.shareable, i, 0)))
		1222	sv_setsv (*svp, &PL_sv_undef);
		1223	}
		1224
1053	SvREFCNT_dec (sv);	1225	SvREFCNT_dec (sv);
1054	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1226	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1055	}	1227	}
1056		1228
1057	sv = sv_2mortal (sv);	1229	sv = sv_2mortal (sv);
1058		1230
1059	return sv;	1231	return sv;
1060	}	1232	}
1061		1233
		1234	/////////////////////////////////////////////////////////////////////////////
		1235	// incremental parser
		1236
		1237	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1238
		1239	// returns 0 for notyet, 1 for success or error
		1240	static int
		1241	incr_parse (CBOR self, SV cborstr)
		1242	{
		1243	STRLEN cur;
		1244	SvPV (cborstr, cur);
		1245
		1246	while (ecb_expect_true (self->incr_need <= cur))
		1247	{
		1248	// table of integer count bytes
		1249	static I8 incr_len[MINOR_MASK + 1] = {
		1250	0, 0, 0, 0, 0, 0, 0, 0,
		1251	0, 0, 0, 0, 0, 0, 0, 0,
		1252	0, 0, 0, 0, 0, 0, 0, 0,
		1253	1, 2, 4, 8,-1,-1,-1,-2
		1254	};
		1255
		1256	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1257	U8 m = *p & MINOR_MASK;
		1258	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1259	I8 ilen = incr_len[m];
		1260
		1261	self->incr_need = self->incr_pos + 1;
		1262
		1263	if (ecb_expect_false (ilen < 0))
		1264	{
		1265	if (m != MINOR_INDEF)
		1266	return 1; // error
		1267
		1268	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1269	{
		1270	if (count >= 0)
		1271	return 1; // error
		1272
		1273	count = 1;
		1274	}
		1275	else
		1276	{
		1277	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1278	count = -1;
		1279	}
		1280	}
		1281	else
		1282	{
		1283	self->incr_need += ilen;
		1284	if (ecb_expect_false (self->incr_need > cur))
		1285	return 0;
		1286
		1287	int major = *p >> MAJOR_SHIFT;
		1288
		1289	switch (major)
		1290	{
		1291	case MAJOR_TAG >> MAJOR_SHIFT:
		1292	++count; // tags merely prefix another value
		1293	break;
		1294
		1295	case MAJOR_BYTES >> MAJOR_SHIFT:
		1296	case MAJOR_TEXT >> MAJOR_SHIFT:
		1297	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1298	case MAJOR_MAP >> MAJOR_SHIFT:
		1299	{
		1300	UV len;
		1301
		1302	if (ecb_expect_false (ilen))
		1303	{
		1304	len = 0;
		1305
		1306	do {
		1307	len = (len << 8) \| *++p;
		1308	} while (--ilen);
		1309	}
		1310	else
		1311	len = m;
		1312
		1313	switch (major)
		1314	{
		1315	case MAJOR_BYTES >> MAJOR_SHIFT:
		1316	case MAJOR_TEXT >> MAJOR_SHIFT:
		1317	self->incr_need += len;
		1318	if (ecb_expect_false (self->incr_need > cur))
		1319	return 0;
		1320
		1321	break;
		1322
		1323	case MAJOR_MAP >> MAJOR_SHIFT:
		1324	len <<= 1;
		1325	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1326	if (len)
		1327	{
		1328	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1329	count = len + 1;
		1330	}
		1331	break;
		1332	}
		1333	}
		1334	}
		1335	}
		1336
		1337	self->incr_pos = self->incr_need;
		1338
		1339	if (count > 0)
		1340	{
		1341	while (!--count)
		1342	{
		1343	if (!AvFILLp (self->incr_count))
		1344	return 1; // done
		1345
		1346	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1347	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1348	}
		1349
		1350	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1351	}
		1352	}
		1353
		1354	return 0;
		1355	}
		1356
		1357
1062	/////////////////////////////////////////////////////////////////////////////	1358	/////////////////////////////////////////////////////////////////////////////
1063	// XS interface functions	1359	// XS interface functions
1064		1360
1065	MODULE = CBOR::XS PACKAGE = CBOR::XS	1361	MODULE = CBOR::XS PACKAGE = CBOR::XS
1066		1362
…		…
1073	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);	1369	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1074		1370
1075	types_true = get_bool ("Types::Serialiser::true" );	1371	types_true = get_bool ("Types::Serialiser::true" );
1076	types_false = get_bool ("Types::Serialiser::false");	1372	types_false = get_bool ("Types::Serialiser::false");
1077	types_error = get_bool ("Types::Serialiser::error");	1373	types_error = get_bool ("Types::Serialiser::error");
		1374
		1375	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1078		1376
1079	sv_cbor = newSVpv ("CBOR", 0);	1377	sv_cbor = newSVpv ("CBOR", 0);
1080	SvREADONLY_on (sv_cbor);	1378	SvREADONLY_on (sv_cbor);
1081	}	1379	}
1082		1380
…		…
1104	void shrink (CBOR *self, int enable = 1)	1402	void shrink (CBOR *self, int enable = 1)
1105	ALIAS:	1403	ALIAS:
1106	shrink = F_SHRINK	1404	shrink = F_SHRINK
1107	allow_unknown = F_ALLOW_UNKNOWN	1405	allow_unknown = F_ALLOW_UNKNOWN
1108	allow_sharing = F_ALLOW_SHARING	1406	allow_sharing = F_ALLOW_SHARING
		1407	allow_cycles = F_ALLOW_CYCLES
1109	dedup_strings = F_DEDUP_STRINGS	1408	pack_strings = F_PACK_STRINGS
		1409	utf8_strings = F_UTF8_STRINGS
		1410	validate_utf8 = F_VALIDATE_UTF8
1110	PPCODE:	1411	PPCODE:
1111	{	1412	{
1112	if (enable)	1413	if (enable)
1113	self->flags \|= ix;	1414	self->flags \|= ix;
1114	else	1415	else
…		…
1120	void get_shrink (CBOR *self)	1421	void get_shrink (CBOR *self)
1121	ALIAS:	1422	ALIAS:
1122	get_shrink = F_SHRINK	1423	get_shrink = F_SHRINK
1123	get_allow_unknown = F_ALLOW_UNKNOWN	1424	get_allow_unknown = F_ALLOW_UNKNOWN
1124	get_allow_sharing = F_ALLOW_SHARING	1425	get_allow_sharing = F_ALLOW_SHARING
		1426	get_allow_cycles = F_ALLOW_CYCLES
1125	get_dedup_strings = F_DEDUP_STRINGS	1427	get_pack_strings = F_PACK_STRINGS
		1428	get_validate_utf8 = F_VALIDATE_UTF8
1126	PPCODE:	1429	PPCODE:
1127	XPUSHs (boolSV (self->flags & ix));	1430	XPUSHs (boolSV (self->flags & ix));
1128		1431
1129	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1432	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130	PPCODE:	1433	PPCODE:
…		…
1143	XPUSHs (ST (0));	1446	XPUSHs (ST (0));
1144		1447
1145	int get_max_size (CBOR *self)	1448	int get_max_size (CBOR *self)
1146	CODE:	1449	CODE:
1147	RETVAL = self->max_size;	1450	RETVAL = self->max_size;
		1451	OUTPUT:
		1452	RETVAL
		1453
		1454	void filter (CBOR self, SV filter = 0)
		1455	PPCODE:
		1456	SvREFCNT_dec (self->filter);
		1457	self->filter = filter ? newSVsv (filter) : filter;
		1458	XPUSHs (ST (0));
		1459
		1460	SV get_filter (CBOR self)
		1461	CODE:
		1462	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148	OUTPUT:	1463	OUTPUT:
1149	RETVAL	1464	RETVAL
1150		1465
1151	void encode (CBOR self, SV scalar)	1466	void encode (CBOR self, SV scalar)
1152	PPCODE:	1467	PPCODE:
…		…
1167	EXTEND (SP, 2);	1482	EXTEND (SP, 2);
1168	PUSHs (sv);	1483	PUSHs (sv);
1169	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1484	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170	}	1485	}
1171		1486
		1487	void incr_parse (CBOR self, SV cborstr)
		1488	ALIAS:
		1489	incr_parse_multiple = 1
		1490	PPCODE:
		1491	{
		1492	if (SvUTF8 (cborstr))
		1493	sv_utf8_downgrade (cborstr, 0);
		1494
		1495	if (!self->incr_count)
		1496	{
		1497	self->incr_count = newAV ();
		1498	self->incr_pos = 0;
		1499	self->incr_need = 1;
		1500
		1501	av_push (self->incr_count, newSViv (1));
		1502	}
		1503
		1504	do
		1505	{
		1506	if (!incr_parse (self, cborstr))
		1507	{
		1508	if (self->incr_need > self->max_size && self->max_size)
		1509	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1510	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1511
		1512	break;
		1513	}
		1514
		1515	SV *sv;
		1516	char *offset;
		1517
		1518	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1519	XPUSHs (sv);
		1520
		1521	sv_chop (cborstr, offset);
		1522
		1523	av_clear (self->incr_count);
		1524	av_push (self->incr_count, newSViv (1));
		1525
		1526	self->incr_pos = 0;
		1527	self->incr_need = self->incr_pos + 1;
		1528	}
		1529	while (ix);
		1530	}
		1531
		1532	void incr_reset (CBOR *self)
		1533	CODE:
		1534	{
		1535	SvREFCNT_dec (self->incr_count);
		1536	self->incr_count = 0;
		1537	}
		1538
		1539	void DESTROY (CBOR *self)
		1540	PPCODE:
		1541	cbor_free (self);
		1542
1172	PROTOTYPES: ENABLE	1543	PROTOTYPES: ENABLE
1173		1544
1174	void encode_cbor (SV *scalar)	1545	void encode_cbor (SV *scalar)
		1546	ALIAS:
		1547	encode_cbor = 0
		1548	encode_cbor_sharing = F_ALLOW_SHARING
1175	PPCODE:	1549	PPCODE:
1176	{	1550	{
1177	CBOR cbor;	1551	CBOR cbor;
1178	cbor_init (&cbor);	1552	cbor_init (&cbor);
		1553	cbor.flags \|= ix;
1179	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1554	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180	XPUSHs (scalar);	1555	XPUSHs (scalar);
1181	}	1556	}
1182		1557
1183	void decode_cbor (SV *cborstr)	1558	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.22 by root, Wed Nov 20 15:05:46 2013 UTC vs. Revision 1.51 by root, Sun Apr 24 13:15:19 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.22 by root, Wed Nov 20 15:05:46 2013 UTC vs.
Revision 1.51 by root, Sun Apr 24 13:15:19 2016 UTC