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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 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_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	// exceptional path for bytze strings that need to be utf8-encoded
		289	STRLEN ulen = len;
		290	U8 p, pend = (U8 *)str + len;
		291
		292	for (p = (U8 *)str; p < pend; ++p)
		293	ulen += *p >> 7; // count set high bits
		294
		295	encode_uint (enc, MAJOR_TEXT, ulen);
		296
		297	need (enc, ulen);
		298	for (p = (U8 *)str; p < pend; ++p)
		299	if (*p < 0x80)
		300	enc->cur++ = p;
		301	else
		302	{
		303	enc->cur++ = 0xc0 + (p >> 6);
		304	enc->cur++ = 0x80 + (p & 63);
		305	}
		306
		307	return;
		308	}
		309
		310	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		311	need (enc, len);
		312	memcpy (enc->cur, str, len);
		313	enc->cur += len;
223	}	314	}
224		315
225	static void	316	static void
226	encode_str (enc_t enc, int utf8, char str, STRLEN len)	317	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
227	{	318	{
228	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	319	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229	{	320	{
230	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	321	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231		322
232	if (SvOK (*svp))	323	if (SvOK (*svp))
233	{	324	{
234	// already registered, use stringref	325	// already registered, use stringref
235	encode_tag (enc, CBOR_TAG_STRINGREF);	326	encode_tag (enc, CBOR_TAG_STRINGREF);
236	encode_uint (enc, 0x00, SvUV (*svp));	327	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237	return;	328	return;
238	}	329	}
239	else if (len >= minimum_string_length (enc->stringref_idx))	330	else if (len >= minimum_string_length (enc->stringref_idx))
240	{	331	{
241	// register only	332	// register only
242	sv_setuv (*svp, enc->stringref_idx);	333	sv_setuv (*svp, enc->stringref_idx);
243	++enc->stringref_idx;	334	++enc->stringref_idx;
244	}	335	}
245	}	336	}
246		337
247	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	338	encode_str (enc, utf8, str, len);
248	need (enc, len);
249	memcpy (enc->cur, str, len);
250	enc->cur += len;
251	}	339	}
252		340
253	static void encode_sv (enc_t enc, SV sv);	341	static void encode_sv (enc_t enc, SV sv);
254		342
255	static void	343	static void
…		…
260	if (enc->depth >= enc->cbor.max_depth)	348	if (enc->depth >= enc->cbor.max_depth)
261	croak (ERR_NESTING_EXCEEDED);	349	croak (ERR_NESTING_EXCEEDED);
262		350
263	++enc->depth;	351	++enc->depth;
264		352
265	encode_uint (enc, 0x80, len + 1);	353	encode_uint (enc, MAJOR_ARRAY, len + 1);
266		354
		355	if (SvMAGICAL (av))
267	for (i = 0; i <= len; ++i)	356	for (i = 0; i <= len; ++i)
268	{	357	{
269	SV **svp = av_fetch (av, i, 0);	358	SV **svp = av_fetch (av, i, 0);
270	encode_sv (enc, svp ? *svp : &PL_sv_undef);	359	encode_sv (enc, svp ? *svp : &PL_sv_undef);
271	}	360	}
		361	else
		362	for (i = 0; i <= len; ++i)
		363	{
		364	SV *sv = AvARRAY (av)[i];
		365	encode_sv (enc, sv ? sv : &PL_sv_undef);
		366	}
272		367
273	--enc->depth;	368	--enc->depth;
274	}
275
276	ecb_inline void
277	encode_he (enc_t enc, HE he)
278	{
279	}	369	}
280		370
281	static void	371	static void
282	encode_hv (enc_t enc, HV hv)	372	encode_hv (enc_t enc, HV hv)
283	{	373	{
…		…
290		380
291	int pairs = hv_iterinit (hv);	381	int pairs = hv_iterinit (hv);
292	int mg = SvMAGICAL (hv);	382	int mg = SvMAGICAL (hv);
293		383
294	if (mg)	384	if (mg)
295	encode_ch (enc, 0xa0 \| 31);	385	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
296	else	386	else
297	encode_uint (enc, 0xa0, pairs);	387	encode_uint (enc, MAJOR_MAP, pairs);
298		388
299	while ((he = hv_iternext (hv)))	389	while ((he = hv_iternext (hv)))
300	{	390	{
301	if (HeKLEN (he) == HEf_SVKEY)	391	if (HeKLEN (he) == HEf_SVKEY)
302	encode_sv (enc, HeSVKEY (he));	392	encode_sv (enc, HeSVKEY (he));
303	else	393	else
304	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	394	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305		395
306	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	396	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307	}	397	}
308		398
309	if (mg)	399	if (mg)
310	encode_ch (enc, 0xe0 \| 31);	400	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
311		401
312	--enc->depth;	402	--enc->depth;
313	}	403	}
314		404
315	// encode objects, arrays and special \0=false and \1=true values.	405	// encode objects, arrays and special \0=false and \1=true values.
316	static void	406	static void
317	encode_rv (enc_t enc, SV sv)	407	encode_rv (enc_t enc, SV sv)
318	{	408	{
319	SvGETMAGIC (sv);	409	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		410
343	svtype svt = SvTYPE (sv);	411	svtype svt = SvTYPE (sv);
344		412
345	if (ecb_expect_false (SvOBJECT (sv)))	413	if (ecb_expect_false (SvOBJECT (sv)))
346	{	414	{
…		…
353	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	421	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
354	? cbor_tagged_stash	422	? cbor_tagged_stash
355	: gv_stashpv ("CBOR::XS::Tagged" , 1);	423	: gv_stashpv ("CBOR::XS::Tagged" , 1);
356		424
357	HV *stash = SvSTASH (sv);	425	HV *stash = SvSTASH (sv);
358	GV *method;
359		426
360	if (stash == boolean_stash)	427	if (stash == boolean_stash)
361	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	428	{
		429	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		430	return;
		431	}
362	else if (stash == error_stash)	432	else if (stash == error_stash)
363	encode_ch (enc, 0xe0 \| 23);	433	{
		434	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		435	return;
		436	}
364	else if (stash == tagged_stash)	437	else if (stash == tagged_stash)
365	{	438	{
366	if (svt != SVt_PVAV)	439	if (svt != SVt_PVAV)
367	croak ("encountered CBOR::XS::Tagged object that isn't an array");	440	croak ("encountered CBOR::XS::Tagged object that isn't an array");
368		441
369	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	442	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
370	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	443	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		444
		445	return;
		446	}
		447	}
		448
		449	if (ecb_expect_false (SvREFCNT (sv) > 1)
		450	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		451	{
		452	if (!enc->shareable)
		453	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		454
		455	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		456
		457	if (SvOK (*svp))
371	}	458	{
		459	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		460	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		461	return;
		462	}
		463	else
		464	{
		465	sv_setuv (*svp, enc->shareable_idx);
		466	++enc->shareable_idx;
		467	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		468	}
		469	}
		470
		471	if (ecb_expect_false (SvOBJECT (sv)))
		472	{
		473	HV *stash = SvSTASH (sv);
		474	GV *method;
		475
372	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	476	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373	{	477	{
374	dSP;	478	dSP;
375		479
376	ENTER; SAVETMPS; PUSHMARK (SP);	480	ENTER; SAVETMPS;
		481	PUSHMARK (SP);
377	// 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
378	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379		484
380	PUTBACK;	485	PUTBACK;
381	// G_SCALAR ensures that return value is 1	486	// G_SCALAR ensures that return value is 1
…		…
394	}	499	}
395	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	500	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396	{	501	{
397	dSP;	502	dSP;
398		503
399	ENTER; SAVETMPS; PUSHMARK (SP);	504	ENTER; SAVETMPS;
		505	SAVESTACK_POS ();
		506	PUSHMARK (SP);
400	EXTEND (SP, 2);	507	EXTEND (SP, 2);
401	// we re-bless the reference to get overload and other niceties right	508	// we re-bless the reference to get overload and other niceties right
402	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	509	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403	PUSHs (sv_cbor);	510	PUSHs (sv_cbor);
404		511
…		…
409	// catch this surprisingly common error	516	// catch this surprisingly common error
410	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	517	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));	518	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412		519
413	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	520	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414	encode_uint (enc, 0x80, count + 1);	521	encode_uint (enc, MAJOR_ARRAY, count + 1);
415	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	522	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416		523
417	while (count)	524	while (count)
418	encode_sv (enc, SP[1 - count--]);	525	encode_sv (enc, SP[1 - count--]);
419		526
420	PUTBACK;	527	PUTBACK;
…		…
441	{	548	{
442	double nv = SvNVX (sv);	549	double nv = SvNVX (sv);
443		550
444	need (enc, 9);	551	need (enc, 9);
445		552
446	if (ecb_expect_false (nv == (U32)nv))	553	if (ecb_expect_false (nv == (NV)(U32)nv))
447	encode_uint (enc, 0x00, (U32)nv);	554	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448	//TODO: maybe I32?	555	//TODO: maybe I32?
449	else if (ecb_expect_false (nv == (float)nv))	556	else if (ecb_expect_false (nv == (float)nv))
450	{	557	{
451	uint32_t fp = ecb_float_to_binary32 (nv);	558	uint32_t fp = ecb_float_to_binary32 (nv);
452		559
453	*enc->cur++ = 0xe0 \| 26;	560	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
454		561
455	if (!ecb_big_endian ())	562	if (!ecb_big_endian ())
456	fp = ecb_bswap32 (fp);	563	fp = ecb_bswap32 (fp);
457		564
458	memcpy (enc->cur, &fp, 4);	565	memcpy (enc->cur, &fp, 4);
…		…
460	}	567	}
461	else	568	else
462	{	569	{
463	uint64_t fp = ecb_double_to_binary64 (nv);	570	uint64_t fp = ecb_double_to_binary64 (nv);
464		571
465	*enc->cur++ = 0xe0 \| 27;	572	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
466		573
467	if (!ecb_big_endian ())	574	if (!ecb_big_endian ())
468	fp = ecb_bswap64 (fp);	575	fp = ecb_bswap64 (fp);
469		576
470	memcpy (enc->cur, &fp, 8);	577	memcpy (enc->cur, &fp, 8);
…		…
479		586
480	if (SvPOKp (sv))	587	if (SvPOKp (sv))
481	{	588	{
482	STRLEN len;	589	STRLEN len;
483	char *str = SvPV (sv, len);	590	char *str = SvPV (sv, len);
484	encode_str (enc, SvUTF8 (sv), str, len);	591	encode_strref (enc, SvUTF8 (sv), str, len);
485	}	592	}
486	else if (SvNOKp (sv))	593	else if (SvNOKp (sv))
487	encode_nv (enc, sv);	594	encode_nv (enc, sv);
488	else if (SvIOKp (sv))	595	else if (SvIOKp (sv))
489	{	596	{
490	if (SvIsUV (sv))	597	if (SvIsUV (sv))
491	encode_uint (enc, 0x00, SvUVX (sv));	598	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
492	else if (SvIVX (sv) >= 0)	599	else if (SvIVX (sv) >= 0)
493	encode_uint (enc, 0x00, SvIVX (sv));	600	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
494	else	601	else
495	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	602	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
496	}	603	}
497	else if (SvROK (sv))	604	else if (SvROK (sv))
498	encode_rv (enc, SvRV (sv));	605	encode_rv (enc, SvRV (sv));
499	else if (!SvOK (sv))	606	else if (!SvOK (sv))
500	encode_ch (enc, 0xe0 \| 22);	607	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
501	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	608	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502	encode_ch (enc, 0xe0 \| 23);	609	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
503	else	610	else
504	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	611	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	612	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506	}	613	}
507		614
508	static SV *	615	static SV *
509	encode_cbor (SV scalar, CBOR cbor)	616	encode_cbor (SV scalar, CBOR cbor)
510	{	617	{
511	enc_t enc = { };	618	enc_t enc = { 0 };
512		619
513	enc.cbor = *cbor;	620	enc.cbor = *cbor;
514	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	621	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515	enc.cur = SvPVX (enc.sv);	622	enc.cur = SvPVX (enc.sv);
516	enc.end = SvEND (enc.sv);	623	enc.end = SvEND (enc.sv);
517		624
518	SvPOK_only (enc.sv);	625	SvPOK_only (enc.sv);
519		626
520	if (cbor->flags & F_ALLOW_STRINGREF)	627	if (cbor->flags & F_PACK_STRINGS)
521	{	628	{
522	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	629	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	630	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
524	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	631	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
525	}	632	}
…		…
547	CBOR cbor;	654	CBOR cbor;
548	U32 depth; // recursion depth	655	U32 depth; // recursion depth
549	U32 maxdepth; // recursion depth limit	656	U32 maxdepth; // recursion depth limit
550	AV *shareable;	657	AV *shareable;
551	AV *stringref;	658	AV *stringref;
		659	SV *decode_tagged;
552	} dec_t;	660	} dec_t;
553		661
554	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	662	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
555		663
556	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	664	#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	667	#define DEC_DEC_DEPTH --dec->depth
560		668
561	static UV	669	static UV
562	decode_uint (dec_t *dec)	670	decode_uint (dec_t *dec)
563	{	671	{
564	switch (*dec->cur & 31)	672	U8 m = *dec->cur & MINOR_MASK;
565	{	673	++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		674
571	case 24:	675	if (ecb_expect_true (m < LENGTH_EXT1))
		676	return m;
		677	else if (ecb_expect_true (m == LENGTH_EXT1))
		678	{
572	WANT (2);	679	WANT (1);
573	dec->cur += 2;	680	dec->cur += 1;
574	return dec->cur[-1];	681	return dec->cur[-1];
575		682	}
576	case 25:	683	else if (ecb_expect_true (m == LENGTH_EXT2))
		684	{
577	WANT (3);	685	WANT (2);
578	dec->cur += 3;	686	dec->cur += 2;
579	return (((UV)dec->cur[-2]) << 8)	687	return (((UV)dec->cur[-2]) << 8)
580	\| ((UV)dec->cur[-1]);	688	\| ((UV)dec->cur[-1]);
581		689	}
582	case 26:	690	else if (ecb_expect_true (m == LENGTH_EXT4))
		691	{
583	WANT (5);	692	WANT (4);
584	dec->cur += 5;	693	dec->cur += 4;
585	return (((UV)dec->cur[-4]) << 24)	694	return (((UV)dec->cur[-4]) << 24)
586	\| (((UV)dec->cur[-3]) << 16)	695	\| (((UV)dec->cur[-3]) << 16)
587	\| (((UV)dec->cur[-2]) << 8)	696	\| (((UV)dec->cur[-2]) << 8)
588	\| ((UV)dec->cur[-1]);	697	\| ((UV)dec->cur[-1]);
589		698	}
590	case 27:	699	else if (ecb_expect_true (m == LENGTH_EXT8))
		700	{
591	WANT (9);	701	WANT (8);
592	dec->cur += 9;	702	dec->cur += 8;
		703
		704	return
		705	#if UVSIZE < 8
		706	0
		707	#else
593	return (((UV)dec->cur[-8]) << 56)	708	(((UV)dec->cur[-8]) << 56)
594	\| (((UV)dec->cur[-7]) << 48)	709	\| (((UV)dec->cur[-7]) << 48)
595	\| (((UV)dec->cur[-6]) << 40)	710	\| (((UV)dec->cur[-6]) << 40)
596	\| (((UV)dec->cur[-5]) << 32)	711	\| (((UV)dec->cur[-5]) << 32)
		712	#endif
597	\| (((UV)dec->cur[-4]) << 24)	713	\| (((UV)dec->cur[-4]) << 24)
598	\| (((UV)dec->cur[-3]) << 16)	714	\| (((UV)dec->cur[-3]) << 16)
599	\| (((UV)dec->cur[-2]) << 8)	715	\| (((UV)dec->cur[-2]) << 8)
600	\| ((UV)dec->cur[-1]);	716	\| ((UV)dec->cur[-1]);
601		717	}
602	default:	718	else
603	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	719	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604	}
605		720
606	fail:	721	fail:
607	return 0;	722	return 0;
608	}	723	}
609		724
…		…
614	{	729	{
615	AV *av = newAV ();	730	AV *av = newAV ();
616		731
617	DEC_INC_DEPTH;	732	DEC_INC_DEPTH;
618		733
619	if ((*dec->cur & 31) == 31)	734	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
620	{	735	{
621	++dec->cur;	736	++dec->cur;
622		737
623	for (;;)	738	for (;;)
624	{	739	{
625	WANT (1);	740	WANT (1);
626		741
627	if (*dec->cur == (0xe0 \| 31))	742	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
628	{	743	{
629	++dec->cur;	744	++dec->cur;
630	break;	745	break;
631	}	746	}
632		747
…		…
635	}	750	}
636	else	751	else
637	{	752	{
638	int i, len = decode_uint (dec);	753	int i, len = decode_uint (dec);
639		754
		755	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);	756	av_fill (av, len - 1);
641		757
642	for (i = 0; i < len; ++i)	758	for (i = 0; i < len; ++i)
643	AvARRAY (av)[i] = decode_sv (dec);	759	AvARRAY (av)[i] = decode_sv (dec);
644	}	760	}
…		…
657	{	773	{
658	// for speed reasons, we specialcase single-string	774	// for speed reasons, we specialcase single-string
659	// byte or utf-8 strings as keys, but only when !stringref	775	// byte or utf-8 strings as keys, but only when !stringref
660		776
661	if (ecb_expect_true (!dec->stringref))	777	if (ecb_expect_true (!dec->stringref))
662	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	778	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663	{	779	{
664	I32 len = decode_uint (dec);	780	I32 len = decode_uint (dec);
665	char key = (char )dec->cur;	781	char key = (char )dec->cur;
666		782
		783	WANT (len);
667	dec->cur += len;	784	dec->cur += len;
668		785
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);	786	hv_store (hv, key, len, decode_sv (dec), 0);
673		787
674	return;	788	return;
675	}	789	}
676	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	790	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677	{	791	{
678	I32 len = decode_uint (dec);	792	I32 len = decode_uint (dec);
679	char key = (char )dec->cur;	793	char key = (char )dec->cur;
680		794
		795	WANT (len);
681	dec->cur += len;	796	dec->cur += len;
682		797
683	if (ecb_expect_false (dec->stringref))	798	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	799	if (!is_utf8_string (key, len))
		800	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685		801
686	hv_store (hv, key, -len, decode_sv (dec), 0);	802	hv_store (hv, key, -len, decode_sv (dec), 0);
687		803
688	return;	804	return;
689	}	805	}
…		…
691	SV *k = decode_sv (dec);	807	SV *k = decode_sv (dec);
692	SV *v = decode_sv (dec);	808	SV *v = decode_sv (dec);
693		809
694	hv_store_ent (hv, k, v, 0);	810	hv_store_ent (hv, k, v, 0);
695	SvREFCNT_dec (k);	811	SvREFCNT_dec (k);
		812
		813	fail:
		814	;
696	}	815	}
697		816
698	static SV *	817	static SV *
699	decode_hv (dec_t *dec)	818	decode_hv (dec_t *dec)
700	{	819	{
701	HV *hv = newHV ();	820	HV *hv = newHV ();
702		821
703	DEC_INC_DEPTH;	822	DEC_INC_DEPTH;
704		823
705	if ((*dec->cur & 31) == 31)	824	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
706	{	825	{
707	++dec->cur;	826	++dec->cur;
708		827
709	for (;;)	828	for (;;)
710	{	829	{
711	WANT (1);	830	WANT (1);
712		831
713	if (*dec->cur == (0xe0 \| 31))	832	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
714	{	833	{
715	++dec->cur;	834	++dec->cur;
716	break;	835	break;
717	}	836	}
718		837
…		…
739	static SV *	858	static SV *
740	decode_str (dec_t *dec, int utf8)	859	decode_str (dec_t *dec, int utf8)
741	{	860	{
742	SV *sv = 0;	861	SV *sv = 0;
743		862
744	if ((*dec->cur & 31) == 31)	863	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
745	{	864	{
		865	// indefinite length strings
746	++dec->cur;	866	++dec->cur;
747		867
		868	U8 major = *dec->cur & MAJOR_MISC;
		869
748	sv = newSVpvn ("", 0);	870	sv = newSVpvn ("", 0);
749		871
750	// not very fast, and certainly not robust against illegal input
751	for (;;)	872	for (;;)
752	{	873	{
753	WANT (1);	874	WANT (1);
754		875
755	if (*dec->cur == (0xe0 \| 31))	876	if ((*dec->cur - major) > LENGTH_EXT8)
		877	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
756	{	878	{
757	++dec->cur;	879	++dec->cur;
758	break;	880	break;
759	}	881	}
		882	else
		883	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760		884
761	sv_catsv (sv, decode_sv (dec));	885	STRLEN len = decode_uint (dec);
		886
		887	WANT (len);
		888	sv_catpvn (sv, dec->cur, len);
		889	dec->cur += len;
762	}	890	}
763	}	891	}
764	else	892	else
765	{	893	{
766	STRLEN len = decode_uint (dec);	894	STRLEN len = decode_uint (dec);
767		895
768	WANT (len);	896	WANT (len);
769	sv = newSVpvn (dec->cur, len);	897	sv = newSVpvn (dec->cur, len);
770	dec->cur += len;	898	dec->cur += len;
		899
		900	if (ecb_expect_false (dec->stringref)
		901	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
		902	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
771	}	903	}
772		904
773	if (utf8)	905	if (utf8)
		906	{
		907	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		908	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		909	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		910
774	SvUTF8_on (sv);	911	SvUTF8_on (sv);
775		912	}
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		913
780	return sv;	914	return sv;
781		915
782	fail:	916	fail:
783	SvREFCNT_dec (sv);	917	SvREFCNT_dec (sv);
…		…
815	}	949	}
816	break;	950	break;
817		951
818	case CBOR_TAG_STRINGREF:	952	case CBOR_TAG_STRINGREF:
819	{	953	{
820	if ((*dec->cur >> 5) != 0)	954	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	955	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822		956
823	UV idx = decode_uint (dec);	957	UV idx = decode_uint (dec);
824		958
825	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	959	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
832	case CBOR_TAG_VALUE_SHAREABLE:	966	case CBOR_TAG_VALUE_SHAREABLE:
833	{	967	{
834	if (ecb_expect_false (!dec->shareable))	968	if (ecb_expect_false (!dec->shareable))
835	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	969	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
836		970
		971	if (dec->cbor.flags & F_ALLOW_CYCLES)
		972	{
837	sv = newSV (0);	973	sv = newSV (0);
838	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	974	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839		975
840	SV *osv = decode_sv (dec);	976	SV *osv = decode_sv (dec);
841	sv_setsv (sv, osv);	977	sv_setsv (sv, osv);
842	SvREFCNT_dec_NN (osv);	978	SvREFCNT_dec_NN (osv);
		979	}
		980	else
		981	{
		982	av_push (dec->shareable, &PL_sv_undef);
		983	int idx = AvFILLp (dec->shareable);
		984	sv = decode_sv (dec);
		985	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		986	}
843	}	987	}
844	break;	988	break;
845		989
846	case CBOR_TAG_VALUE_SHAREDREF:	990	case CBOR_TAG_VALUE_SHAREDREF:
847	{	991	{
848	if ((*dec->cur >> 5) != 0)	992	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	993	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850		994
851	UV idx = decode_uint (dec);	995	UV idx = decode_uint (dec);
852		996
853	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	997	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
854	ERR ("corrupted CBOR data (sharedref index out of bounds)");	998	ERR ("corrupted CBOR data (sharedref index out of bounds)");
855		999
856	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1000	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1001
		1002	if (sv == &PL_sv_undef)
		1003	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857	}	1004	}
858	break;	1005	break;
859		1006
860	case CBOR_TAG_PERL_OBJECT:	1007	case CBOR_TAG_PERL_OBJECT:
861	{	1008	{
…		…
876	if (!method)	1023	if (!method)
877	ERR ("cannot decode perl-object (package does not have a THAW method)");	1024	ERR ("cannot decode perl-object (package does not have a THAW method)");
878		1025
879	dSP;	1026	dSP;
880		1027
881	ENTER; SAVETMPS; PUSHMARK (SP);	1028	ENTER; SAVETMPS;
		1029	PUSHMARK (SP);
882	EXTEND (SP, len + 1);	1030	EXTEND (SP, len + 1);
883	// we re-bless the reference to get overload and other niceties right	1031	// we re-bless the reference to get overload and other niceties right
884	PUSHs (*av_fetch (av, 0, 1));	1032	PUSHs (*av_fetch (av, 0, 1));
885	PUSHs (sv_cbor);	1033	PUSHs (sv_cbor);
886		1034
…		…
910		1058
911	default:	1059	default:
912	{	1060	{
913	sv = decode_sv (dec);	1061	sv = decode_sv (dec);
914		1062
		1063	dSP;
		1064	ENTER; SAVETMPS;
		1065	SAVESTACK_POS ();
		1066	PUSHMARK (SP);
		1067	EXTEND (SP, 2);
		1068	PUSHs (newSVuv (tag));
		1069	PUSHs (sv);
		1070
		1071	PUTBACK;
		1072	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
		1073	SPAGAIN;
		1074
		1075	if (SvTRUE (ERRSV))
		1076	{
		1077	FREETMPS; LEAVE;
		1078	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
		1079	}
		1080
		1081	if (count)
		1082	{
		1083	SvREFCNT_dec (sv);
		1084	sv = SvREFCNT_inc (POPs);
		1085	}
		1086	else
		1087	{
915	AV *av = newAV ();	1088	AV *av = newAV ();
916	av_push (av, newSVuv (tag));	1089	av_push (av, newSVuv (tag));
917	av_push (av, sv);	1090	av_push (av, sv);
918		1091
919	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1092	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
920	? cbor_tagged_stash	1093	? cbor_tagged_stash
921	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1094	: gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1095	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1096	}
		1097
		1098	PUTBACK;
		1099
		1100	FREETMPS; LEAVE;
924	}	1101	}
925	break;	1102	break;
926	}	1103	}
927		1104
928	return sv;	1105	return sv;
…		…
935	static SV *	1112	static SV *
936	decode_sv (dec_t *dec)	1113	decode_sv (dec_t *dec)
937	{	1114	{
938	WANT (1);	1115	WANT (1);
939		1116
940	switch (*dec->cur >> 5)	1117	switch (*dec->cur >> MAJOR_SHIFT)
941	{	1118	{
942	case 0: // unsigned int	1119	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943	return newSVuv (decode_uint (dec));	1120	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944	case 1: // negative int	1121	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945	return newSViv (-1 - (IV)decode_uint (dec));	1122	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946	case 2: // octet string	1123	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947	return decode_str (dec, 0);	1124	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948	case 3: // utf-8 string	1125	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949	return decode_str (dec, 1);	1126
950	case 4: // array	1127	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)	1128	switch (*dec->cur++ & MINOR_MASK)
958	{	1129	{
959	case 20:	1130	case SIMPLE_FALSE:
960	#if CBOR_SLOW	1131	#if CBOR_SLOW
961	types_false = get_bool ("Types::Serialiser::false");	1132	types_false = get_bool ("Types::Serialiser::false");
962	#endif	1133	#endif
963	return newSVsv (types_false);	1134	return newSVsv (types_false);
964	case 21:	1135	case SIMPLE_TRUE:
965	#if CBOR_SLOW	1136	#if CBOR_SLOW
966	types_true = get_bool ("Types::Serialiser::true");	1137	types_true = get_bool ("Types::Serialiser::true");
967	#endif	1138	#endif
968	return newSVsv (types_true);	1139	return newSVsv (types_true);
969	case 22:	1140	case SIMPLE_NULL:
970	return newSVsv (&PL_sv_undef);	1141	return newSVsv (&PL_sv_undef);
971	case 23:	1142	case SIMPLE_UNDEF:
972	#if CBOR_SLOW	1143	#if CBOR_SLOW
973	types_error = get_bool ("Types::Serialiser::error");	1144	types_error = get_bool ("Types::Serialiser::error");
974	#endif	1145	#endif
975	return newSVsv (types_error);	1146	return newSVsv (types_error);
976		1147
977	case 25:	1148	case MISC_FLOAT16:
978	{	1149	{
979	WANT (2);	1150	WANT (2);
980		1151
981	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1152	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
982	dec->cur += 2;	1153	dec->cur += 2;
983		1154
984	return newSVnv (ecb_binary16_to_float (fp));	1155	return newSVnv (ecb_binary16_to_float (fp));
985	}	1156	}
986		1157
987	case 26:	1158	case MISC_FLOAT32:
988	{	1159	{
989	uint32_t fp;	1160	uint32_t fp;
990	WANT (4);	1161	WANT (4);
991	memcpy (&fp, dec->cur, 4);	1162	memcpy (&fp, dec->cur, 4);
992	dec->cur += 4;	1163	dec->cur += 4;
…		…
995	fp = ecb_bswap32 (fp);	1166	fp = ecb_bswap32 (fp);
996		1167
997	return newSVnv (ecb_binary32_to_float (fp));	1168	return newSVnv (ecb_binary32_to_float (fp));
998	}	1169	}
999		1170
1000	case 27:	1171	case MISC_FLOAT64:
1001	{	1172	{
1002	uint64_t fp;	1173	uint64_t fp;
1003	WANT (8);	1174	WANT (8);
1004	memcpy (&fp, dec->cur, 8);	1175	memcpy (&fp, dec->cur, 8);
1005	dec->cur += 8;	1176	dec->cur += 8;
…		…
1008	fp = ecb_bswap64 (fp);	1179	fp = ecb_bswap64 (fp);
1009		1180
1010	return newSVnv (ecb_binary64_to_double (fp));	1181	return newSVnv (ecb_binary64_to_double (fp));
1011	}	1182	}
1012		1183
1013	// 0..19 unassigned	1184	// 0..19 unassigned simple
1014	// 24 reserved + unassigned (reserved values are not encodable)	1185	// 24 reserved + unassigned simple (reserved values are not encodable)
		1186	// 28-30 unassigned misc
		1187	// 31 break code
1015	default:	1188	default:
1016	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1189	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017	}	1190	}
1018		1191
1019	break;	1192	break;
1020	}	1193	}
1021		1194
…		…
1024	}	1197	}
1025		1198
1026	static SV *	1199	static SV *
1027	decode_cbor (SV string, CBOR cbor, char **offset_return)	1200	decode_cbor (SV string, CBOR cbor, char **offset_return)
1028	{	1201	{
1029	dec_t dec = { };	1202	dec_t dec = { 0 };
1030	SV *sv;	1203	SV *sv;
1031	STRLEN len;	1204	STRLEN len;
1032	char *data = SvPVbyte (string, len);	1205	char *data = SvPVbyte (string, len);
1033		1206
1034	if (len > cbor->max_size && cbor->max_size)	1207	if (len > cbor->max_size && cbor->max_size)
…		…
1048	if (dec.cur != dec.end && !dec.err)	1221	if (dec.cur != dec.end && !dec.err)
1049	dec.err = "garbage after CBOR object";	1222	dec.err = "garbage after CBOR object";
1050		1223
1051	if (dec.err)	1224	if (dec.err)
1052	{	1225	{
		1226	if (dec.shareable)
		1227	{
		1228	// need to break cyclic links, which whould all be in shareable
		1229	int i;
		1230	SV **svp;
		1231
		1232	for (i = av_len (dec.shareable) + 1; i--; )
		1233	if ((svp = av_fetch (dec.shareable, i, 0)))
		1234	sv_setsv (*svp, &PL_sv_undef);
		1235	}
		1236
1053	SvREFCNT_dec (sv);	1237	SvREFCNT_dec (sv);
1054	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1238	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1055	}	1239	}
1056		1240
1057	sv = sv_2mortal (sv);	1241	sv = sv_2mortal (sv);
1058		1242
1059	return sv;	1243	return sv;
1060	}	1244	}
1061		1245
		1246	/////////////////////////////////////////////////////////////////////////////
		1247	// incremental parser
		1248
		1249	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1250
		1251	// returns 0 for notyet, 1 for success or error
		1252	static int
		1253	incr_parse (CBOR self, SV cborstr)
		1254	{
		1255	STRLEN cur;
		1256	SvPV (cborstr, cur);
		1257
		1258	while (ecb_expect_true (self->incr_need <= cur))
		1259	{
		1260	// table of integer count bytes
		1261	static I8 incr_len[MINOR_MASK + 1] = {
		1262	0, 0, 0, 0, 0, 0, 0, 0,
		1263	0, 0, 0, 0, 0, 0, 0, 0,
		1264	0, 0, 0, 0, 0, 0, 0, 0,
		1265	1, 2, 4, 8,-1,-1,-1,-2
		1266	};
		1267
		1268	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1269	U8 m = *p & MINOR_MASK;
		1270	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1271	I8 ilen = incr_len[m];
		1272
		1273	self->incr_need = self->incr_pos + 1;
		1274
		1275	if (ecb_expect_false (ilen < 0))
		1276	{
		1277	if (m != MINOR_INDEF)
		1278	return 1; // error
		1279
		1280	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1281	{
		1282	if (count >= 0)
		1283	return 1; // error
		1284
		1285	count = 1;
		1286	}
		1287	else
		1288	{
		1289	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1290	count = -1;
		1291	}
		1292	}
		1293	else
		1294	{
		1295	self->incr_need += ilen;
		1296	if (ecb_expect_false (self->incr_need > cur))
		1297	return 0;
		1298
		1299	int major = *p >> MAJOR_SHIFT;
		1300
		1301	switch (major)
		1302	{
		1303	case MAJOR_TAG >> MAJOR_SHIFT:
		1304	++count; // tags merely prefix another value
		1305	break;
		1306
		1307	case MAJOR_BYTES >> MAJOR_SHIFT:
		1308	case MAJOR_TEXT >> MAJOR_SHIFT:
		1309	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1310	case MAJOR_MAP >> MAJOR_SHIFT:
		1311	{
		1312	UV len;
		1313
		1314	if (ecb_expect_false (ilen))
		1315	{
		1316	len = 0;
		1317
		1318	do {
		1319	len = (len << 8) \| *++p;
		1320	} while (--ilen);
		1321	}
		1322	else
		1323	len = m;
		1324
		1325	switch (major)
		1326	{
		1327	case MAJOR_BYTES >> MAJOR_SHIFT:
		1328	case MAJOR_TEXT >> MAJOR_SHIFT:
		1329	self->incr_need += len;
		1330	if (ecb_expect_false (self->incr_need > cur))
		1331	return 0;
		1332
		1333	break;
		1334
		1335	case MAJOR_MAP >> MAJOR_SHIFT:
		1336	len <<= 1;
		1337	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1338	if (len)
		1339	{
		1340	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1341	count = len + 1;
		1342	}
		1343	break;
		1344	}
		1345	}
		1346	}
		1347	}
		1348
		1349	self->incr_pos = self->incr_need;
		1350
		1351	if (count > 0)
		1352	{
		1353	while (!--count)
		1354	{
		1355	if (!AvFILLp (self->incr_count))
		1356	return 1; // done
		1357
		1358	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1359	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1360	}
		1361
		1362	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1363	}
		1364	}
		1365
		1366	return 0;
		1367	}
		1368
		1369
1062	/////////////////////////////////////////////////////////////////////////////	1370	/////////////////////////////////////////////////////////////////////////////
1063	// XS interface functions	1371	// XS interface functions
1064		1372
1065	MODULE = CBOR::XS PACKAGE = CBOR::XS	1373	MODULE = CBOR::XS PACKAGE = CBOR::XS
1066		1374
…		…
1073	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);	1381	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1074		1382
1075	types_true = get_bool ("Types::Serialiser::true" );	1383	types_true = get_bool ("Types::Serialiser::true" );
1076	types_false = get_bool ("Types::Serialiser::false");	1384	types_false = get_bool ("Types::Serialiser::false");
1077	types_error = get_bool ("Types::Serialiser::error");	1385	types_error = get_bool ("Types::Serialiser::error");
		1386
		1387	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1078		1388
1079	sv_cbor = newSVpv ("CBOR", 0);	1389	sv_cbor = newSVpv ("CBOR", 0);
1080	SvREADONLY_on (sv_cbor);	1390	SvREADONLY_on (sv_cbor);
1081	}	1391	}
1082		1392
…		…
1104	void shrink (CBOR *self, int enable = 1)	1414	void shrink (CBOR *self, int enable = 1)
1105	ALIAS:	1415	ALIAS:
1106	shrink = F_SHRINK	1416	shrink = F_SHRINK
1107	allow_unknown = F_ALLOW_UNKNOWN	1417	allow_unknown = F_ALLOW_UNKNOWN
1108	allow_sharing = F_ALLOW_SHARING	1418	allow_sharing = F_ALLOW_SHARING
1109	allow_stringref = F_ALLOW_STRINGREF	1419	allow_cycles = F_ALLOW_CYCLES
		1420	pack_strings = F_PACK_STRINGS
		1421	utf8_strings = F_UTF8_STRINGS
		1422	validate_utf8 = F_VALIDATE_UTF8
1110	PPCODE:	1423	PPCODE:
1111	{	1424	{
1112	if (enable)	1425	if (enable)
1113	self->flags \|= ix;	1426	self->flags \|= ix;
1114	else	1427	else
…		…
1120	void get_shrink (CBOR *self)	1433	void get_shrink (CBOR *self)
1121	ALIAS:	1434	ALIAS:
1122	get_shrink = F_SHRINK	1435	get_shrink = F_SHRINK
1123	get_allow_unknown = F_ALLOW_UNKNOWN	1436	get_allow_unknown = F_ALLOW_UNKNOWN
1124	get_allow_sharing = F_ALLOW_SHARING	1437	get_allow_sharing = F_ALLOW_SHARING
1125	get_allow_stringref = F_ALLOW_STRINGREF	1438	get_allow_cycles = F_ALLOW_CYCLES
		1439	get_pack_strings = F_PACK_STRINGS
		1440	get_validate_utf8 = F_VALIDATE_UTF8
1126	PPCODE:	1441	PPCODE:
1127	XPUSHs (boolSV (self->flags & ix));	1442	XPUSHs (boolSV (self->flags & ix));
1128		1443
1129	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1444	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130	PPCODE:	1445	PPCODE:
…		…
1143	XPUSHs (ST (0));	1458	XPUSHs (ST (0));
1144		1459
1145	int get_max_size (CBOR *self)	1460	int get_max_size (CBOR *self)
1146	CODE:	1461	CODE:
1147	RETVAL = self->max_size;	1462	RETVAL = self->max_size;
		1463	OUTPUT:
		1464	RETVAL
		1465
		1466	void filter (CBOR self, SV filter = 0)
		1467	PPCODE:
		1468	SvREFCNT_dec (self->filter);
		1469	self->filter = filter ? newSVsv (filter) : filter;
		1470	XPUSHs (ST (0));
		1471
		1472	SV get_filter (CBOR self)
		1473	CODE:
		1474	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148	OUTPUT:	1475	OUTPUT:
1149	RETVAL	1476	RETVAL
1150		1477
1151	void encode (CBOR self, SV scalar)	1478	void encode (CBOR self, SV scalar)
1152	PPCODE:	1479	PPCODE:
…		…
1167	EXTEND (SP, 2);	1494	EXTEND (SP, 2);
1168	PUSHs (sv);	1495	PUSHs (sv);
1169	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1496	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170	}	1497	}
1171		1498
		1499	void incr_parse (CBOR self, SV cborstr)
		1500	ALIAS:
		1501	incr_parse_multiple = 1
		1502	PPCODE:
		1503	{
		1504	if (SvUTF8 (cborstr))
		1505	sv_utf8_downgrade (cborstr, 0);
		1506
		1507	if (!self->incr_count)
		1508	{
		1509	self->incr_count = newAV ();
		1510	self->incr_pos = 0;
		1511	self->incr_need = 1;
		1512
		1513	av_push (self->incr_count, newSViv (1));
		1514	}
		1515
		1516	do
		1517	{
		1518	if (!incr_parse (self, cborstr))
		1519	{
		1520	if (self->incr_need > self->max_size && self->max_size)
		1521	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1522	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1523
		1524	break;
		1525	}
		1526
		1527	SV *sv;
		1528	char *offset;
		1529
		1530	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1531	XPUSHs (sv);
		1532
		1533	sv_chop (cborstr, offset);
		1534
		1535	av_clear (self->incr_count);
		1536	av_push (self->incr_count, newSViv (1));
		1537
		1538	self->incr_pos = 0;
		1539	self->incr_need = self->incr_pos + 1;
		1540	}
		1541	while (ix);
		1542	}
		1543
		1544	void incr_reset (CBOR *self)
		1545	CODE:
		1546	{
		1547	SvREFCNT_dec (self->incr_count);
		1548	self->incr_count = 0;
		1549	}
		1550
		1551	void DESTROY (CBOR *self)
		1552	PPCODE:
		1553	cbor_free (self);
		1554
1172	PROTOTYPES: ENABLE	1555	PROTOTYPES: ENABLE
1173		1556
1174	void encode_cbor (SV *scalar)	1557	void encode_cbor (SV *scalar)
		1558	ALIAS:
		1559	encode_cbor = 0
		1560	encode_cbor_sharing = F_ALLOW_SHARING
1175	PPCODE:	1561	PPCODE:
1176	{	1562	{
1177	CBOR cbor;	1563	CBOR cbor;
1178	cbor_init (&cbor);	1564	cbor_init (&cbor);
		1565	cbor.flags \|= ix;
1179	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1566	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180	XPUSHs (scalar);	1567	XPUSHs (scalar);
1181	}	1568	}
1182		1569
1183	void decode_cbor (SV *cborstr)	1570	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs. Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC