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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.26 by root, Fri Nov 22 09:40:13 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	}	369	}
275		370
276	static void	371	static void
…		…
285		380
286	int pairs = hv_iterinit (hv);	381	int pairs = hv_iterinit (hv);
287	int mg = SvMAGICAL (hv);	382	int mg = SvMAGICAL (hv);
288		383
289	if (mg)	384	if (mg)
290	encode_ch (enc, 0xa0 \| 31);	385	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
291	else	386	else
292	encode_uint (enc, 0xa0, pairs);	387	encode_uint (enc, MAJOR_MAP, pairs);
293		388
294	while ((he = hv_iternext (hv)))	389	while ((he = hv_iternext (hv)))
295	{	390	{
296	if (HeKLEN (he) == HEf_SVKEY)	391	if (HeKLEN (he) == HEf_SVKEY)
297	encode_sv (enc, HeSVKEY (he));	392	encode_sv (enc, HeSVKEY (he));
298	else	393	else
299	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	394	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300		395
301	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));
302	}	397	}
303		398
304	if (mg)	399	if (mg)
305	encode_ch (enc, 0xe0 \| 31);	400	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
306		401
307	--enc->depth;	402	--enc->depth;
308	}	403	}
309		404
310	// encode objects, arrays and special \0=false and \1=true values.	405	// encode objects, arrays and special \0=false and \1=true values.
311	static void	406	static void
312	encode_rv (enc_t enc, SV sv)	407	encode_rv (enc_t enc, SV sv)
313	{	408	{
314	SvGETMAGIC (sv);	409	SvGETMAGIC (sv);
315
316	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
317	&& ecb_expect_false (SvREFCNT (sv) > 1))
318	{
319	if (!enc->shareable)
320	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
321
322	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
323
324	if (SvOK (*svp))
325	{
326	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
327	encode_uint (enc, 0x00, SvUV (*svp));
328	return;
329	}
330	else
331	{
332	sv_setuv (*svp, enc->shareable_idx);
333	++enc->shareable_idx;
334	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
335	}
336	}
337		410
338	svtype svt = SvTYPE (sv);	411	svtype svt = SvTYPE (sv);
339		412
340	if (ecb_expect_false (SvOBJECT (sv)))	413	if (ecb_expect_false (SvOBJECT (sv)))
341	{	414	{
…		…
348	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	421	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
349	? cbor_tagged_stash	422	? cbor_tagged_stash
350	: gv_stashpv ("CBOR::XS::Tagged" , 1);	423	: gv_stashpv ("CBOR::XS::Tagged" , 1);
351		424
352	HV *stash = SvSTASH (sv);	425	HV *stash = SvSTASH (sv);
353	GV *method;
354		426
355	if (stash == boolean_stash)	427	if (stash == boolean_stash)
356	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	}
357	else if (stash == error_stash)	432	else if (stash == error_stash)
358	encode_ch (enc, 0xe0 \| 23);	433	{
		434	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		435	return;
		436	}
359	else if (stash == tagged_stash)	437	else if (stash == tagged_stash)
360	{	438	{
361	if (svt != SVt_PVAV)	439	if (svt != SVt_PVAV)
362	croak ("encountered CBOR::XS::Tagged object that isn't an array");	440	croak ("encountered CBOR::XS::Tagged object that isn't an array");
363		441
364	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	442	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
365	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))
366	}	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
367	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	476	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
368	{	477	{
369	dSP;	478	dSP;
370		479
371	ENTER; SAVETMPS; PUSHMARK (SP);	480	ENTER; SAVETMPS;
		481	PUSHMARK (SP);
372	// 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
373	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
374		484
375	PUTBACK;	485	PUTBACK;
376	// G_SCALAR ensures that return value is 1	486	// G_SCALAR ensures that return value is 1
…		…
389	}	499	}
390	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	500	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
391	{	501	{
392	dSP;	502	dSP;
393		503
394	ENTER; SAVETMPS; PUSHMARK (SP);	504	ENTER; SAVETMPS;
		505	SAVESTACK_POS ();
		506	PUSHMARK (SP);
395	EXTEND (SP, 2);	507	EXTEND (SP, 2);
396	// 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
397	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	509	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
398	PUSHs (sv_cbor);	510	PUSHs (sv_cbor);
399		511
…		…
404	// catch this surprisingly common error	516	// catch this surprisingly common error
405	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	517	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
406	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));
407		519
408	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	520	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
409	encode_uint (enc, 0x80, count + 1);	521	encode_uint (enc, MAJOR_ARRAY, count + 1);
410	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	522	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
411		523
412	while (count)	524	while (count)
413	encode_sv (enc, SP[1 - count--]);	525	encode_sv (enc, SP[1 - count--]);
414		526
415	PUTBACK;	527	PUTBACK;
…		…
436	{	548	{
437	double nv = SvNVX (sv);	549	double nv = SvNVX (sv);
438		550
439	need (enc, 9);	551	need (enc, 9);
440		552
441	if (ecb_expect_false (nv == (U32)nv))	553	if (ecb_expect_false (nv == (NV)(U32)nv))
442	encode_uint (enc, 0x00, (U32)nv);	554	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
443	//TODO: maybe I32?	555	//TODO: maybe I32?
444	else if (ecb_expect_false (nv == (float)nv))	556	else if (ecb_expect_false (nv == (float)nv))
445	{	557	{
446	uint32_t fp = ecb_float_to_binary32 (nv);	558	uint32_t fp = ecb_float_to_binary32 (nv);
447		559
448	*enc->cur++ = 0xe0 \| 26;	560	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
449		561
450	if (!ecb_big_endian ())	562	if (!ecb_big_endian ())
451	fp = ecb_bswap32 (fp);	563	fp = ecb_bswap32 (fp);
452		564
453	memcpy (enc->cur, &fp, 4);	565	memcpy (enc->cur, &fp, 4);
…		…
455	}	567	}
456	else	568	else
457	{	569	{
458	uint64_t fp = ecb_double_to_binary64 (nv);	570	uint64_t fp = ecb_double_to_binary64 (nv);
459		571
460	*enc->cur++ = 0xe0 \| 27;	572	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
461		573
462	if (!ecb_big_endian ())	574	if (!ecb_big_endian ())
463	fp = ecb_bswap64 (fp);	575	fp = ecb_bswap64 (fp);
464		576
465	memcpy (enc->cur, &fp, 8);	577	memcpy (enc->cur, &fp, 8);
…		…
474		586
475	if (SvPOKp (sv))	587	if (SvPOKp (sv))
476	{	588	{
477	STRLEN len;	589	STRLEN len;
478	char *str = SvPV (sv, len);	590	char *str = SvPV (sv, len);
479	encode_str (enc, SvUTF8 (sv), str, len);	591	encode_strref (enc, SvUTF8 (sv), str, len);
480	}	592	}
481	else if (SvNOKp (sv))	593	else if (SvNOKp (sv))
482	encode_nv (enc, sv);	594	encode_nv (enc, sv);
483	else if (SvIOKp (sv))	595	else if (SvIOKp (sv))
484	{	596	{
485	if (SvIsUV (sv))	597	if (SvIsUV (sv))
486	encode_uint (enc, 0x00, SvUVX (sv));	598	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
487	else if (SvIVX (sv) >= 0)	599	else if (SvIVX (sv) >= 0)
488	encode_uint (enc, 0x00, SvIVX (sv));	600	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
489	else	601	else
490	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	602	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
491	}	603	}
492	else if (SvROK (sv))	604	else if (SvROK (sv))
493	encode_rv (enc, SvRV (sv));	605	encode_rv (enc, SvRV (sv));
494	else if (!SvOK (sv))	606	else if (!SvOK (sv))
495	encode_ch (enc, 0xe0 \| 22);	607	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
496	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	608	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
497	encode_ch (enc, 0xe0 \| 23);	609	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
498	else	610	else
499	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",
500	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	612	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
501	}	613	}
502		614
503	static SV *	615	static SV *
504	encode_cbor (SV scalar, CBOR cbor)	616	encode_cbor (SV scalar, CBOR cbor)
505	{	617	{
506	enc_t enc = { };	618	enc_t enc = { 0 };
507		619
508	enc.cbor = *cbor;	620	enc.cbor = *cbor;
509	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	621	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
510	enc.cur = SvPVX (enc.sv);	622	enc.cur = SvPVX (enc.sv);
511	enc.end = SvEND (enc.sv);	623	enc.end = SvEND (enc.sv);
512		624
513	SvPOK_only (enc.sv);	625	SvPOK_only (enc.sv);
514		626
515	if (cbor->flags & F_ALLOW_STRINGREF)	627	if (cbor->flags & F_PACK_STRINGS)
516	{	628	{
517	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	629	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
518	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	630	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
519	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	631	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
520	}	632	}
…		…
542	CBOR cbor;	654	CBOR cbor;
543	U32 depth; // recursion depth	655	U32 depth; // recursion depth
544	U32 maxdepth; // recursion depth limit	656	U32 maxdepth; // recursion depth limit
545	AV *shareable;	657	AV *shareable;
546	AV *stringref;	658	AV *stringref;
		659	SV *decode_tagged;
547	} dec_t;	660	} dec_t;
548		661
549	#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
550		663
551	#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")
…		…
554	#define DEC_DEC_DEPTH --dec->depth	667	#define DEC_DEC_DEPTH --dec->depth
555		668
556	static UV	669	static UV
557	decode_uint (dec_t *dec)	670	decode_uint (dec_t *dec)
558	{	671	{
559	switch (*dec->cur & 31)	672	U8 m = *dec->cur & MINOR_MASK;
560	{	673	++dec->cur;
561	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
562	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
563	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
564	return *dec->cur++ & 31;
565		674
566	case 24:	675	if (ecb_expect_true (m < LENGTH_EXT1))
		676	return m;
		677	else if (ecb_expect_true (m == LENGTH_EXT1))
		678	{
567	WANT (2);	679	WANT (1);
568	dec->cur += 2;	680	dec->cur += 1;
569	return dec->cur[-1];	681	return dec->cur[-1];
570		682	}
571	case 25:	683	else if (ecb_expect_true (m == LENGTH_EXT2))
		684	{
572	WANT (3);	685	WANT (2);
573	dec->cur += 3;	686	dec->cur += 2;
574	return (((UV)dec->cur[-2]) << 8)	687	return (((UV)dec->cur[-2]) << 8)
575	\| ((UV)dec->cur[-1]);	688	\| ((UV)dec->cur[-1]);
576		689	}
577	case 26:	690	else if (ecb_expect_true (m == LENGTH_EXT4))
		691	{
578	WANT (5);	692	WANT (4);
579	dec->cur += 5;	693	dec->cur += 4;
580	return (((UV)dec->cur[-4]) << 24)	694	return (((UV)dec->cur[-4]) << 24)
581	\| (((UV)dec->cur[-3]) << 16)	695	\| (((UV)dec->cur[-3]) << 16)
582	\| (((UV)dec->cur[-2]) << 8)	696	\| (((UV)dec->cur[-2]) << 8)
583	\| ((UV)dec->cur[-1]);	697	\| ((UV)dec->cur[-1]);
584		698	}
585	case 27:	699	else if (ecb_expect_true (m == LENGTH_EXT8))
		700	{
586	WANT (9);	701	WANT (8);
587	dec->cur += 9;	702	dec->cur += 8;
		703
		704	return
		705	#if UVSIZE < 8
		706	0
		707	#else
588	return (((UV)dec->cur[-8]) << 56)	708	(((UV)dec->cur[-8]) << 56)
589	\| (((UV)dec->cur[-7]) << 48)	709	\| (((UV)dec->cur[-7]) << 48)
590	\| (((UV)dec->cur[-6]) << 40)	710	\| (((UV)dec->cur[-6]) << 40)
591	\| (((UV)dec->cur[-5]) << 32)	711	\| (((UV)dec->cur[-5]) << 32)
		712	#endif
592	\| (((UV)dec->cur[-4]) << 24)	713	\| (((UV)dec->cur[-4]) << 24)
593	\| (((UV)dec->cur[-3]) << 16)	714	\| (((UV)dec->cur[-3]) << 16)
594	\| (((UV)dec->cur[-2]) << 8)	715	\| (((UV)dec->cur[-2]) << 8)
595	\| ((UV)dec->cur[-1]);	716	\| ((UV)dec->cur[-1]);
596		717	}
597	default:	718	else
598	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	719	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
599	}
600		720
601	fail:	721	fail:
602	return 0;	722	return 0;
603	}	723	}
604		724
…		…
609	{	729	{
610	AV *av = newAV ();	730	AV *av = newAV ();
611		731
612	DEC_INC_DEPTH;	732	DEC_INC_DEPTH;
613		733
614	if ((*dec->cur & 31) == 31)	734	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
615	{	735	{
616	++dec->cur;	736	++dec->cur;
617		737
618	for (;;)	738	for (;;)
619	{	739	{
620	WANT (1);	740	WANT (1);
621		741
622	if (*dec->cur == (0xe0 \| 31))	742	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
623	{	743	{
624	++dec->cur;	744	++dec->cur;
625	break;	745	break;
626	}	746	}
627		747
…		…
630	}	750	}
631	else	751	else
632	{	752	{
633	int i, len = decode_uint (dec);	753	int i, len = decode_uint (dec);
634		754
		755	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
635	av_fill (av, len - 1);	756	av_fill (av, len - 1);
636		757
637	for (i = 0; i < len; ++i)	758	for (i = 0; i < len; ++i)
638	AvARRAY (av)[i] = decode_sv (dec);	759	AvARRAY (av)[i] = decode_sv (dec);
639	}	760	}
…		…
652	{	773	{
653	// for speed reasons, we specialcase single-string	774	// for speed reasons, we specialcase single-string
654	// byte or utf-8 strings as keys, but only when !stringref	775	// byte or utf-8 strings as keys, but only when !stringref
655		776
656	if (ecb_expect_true (!dec->stringref))	777	if (ecb_expect_true (!dec->stringref))
657	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	778	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
658	{	779	{
659	I32 len = decode_uint (dec);	780	I32 len = decode_uint (dec);
660	char key = (char )dec->cur;	781	char key = (char )dec->cur;
661		782
		783	WANT (len);
662	dec->cur += len;	784	dec->cur += len;
663		785
664	if (ecb_expect_false (dec->stringref))
665	av_push (dec->stringref, newSVpvn (key, len));
666
667	hv_store (hv, key, len, decode_sv (dec), 0);	786	hv_store (hv, key, len, decode_sv (dec), 0);
668		787
669	return;	788	return;
670	}	789	}
671	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	790	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
672	{	791	{
673	I32 len = decode_uint (dec);	792	I32 len = decode_uint (dec);
674	char key = (char )dec->cur;	793	char key = (char )dec->cur;
675		794
		795	WANT (len);
676	dec->cur += len;	796	dec->cur += len;
677		797
678	if (ecb_expect_false (dec->stringref))	798	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
679	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)");
680		801
681	hv_store (hv, key, -len, decode_sv (dec), 0);	802	hv_store (hv, key, -len, decode_sv (dec), 0);
682		803
683	return;	804	return;
684	}	805	}
…		…
686	SV *k = decode_sv (dec);	807	SV *k = decode_sv (dec);
687	SV *v = decode_sv (dec);	808	SV *v = decode_sv (dec);
688		809
689	hv_store_ent (hv, k, v, 0);	810	hv_store_ent (hv, k, v, 0);
690	SvREFCNT_dec (k);	811	SvREFCNT_dec (k);
		812
		813	fail:
		814	;
691	}	815	}
692		816
693	static SV *	817	static SV *
694	decode_hv (dec_t *dec)	818	decode_hv (dec_t *dec)
695	{	819	{
696	HV *hv = newHV ();	820	HV *hv = newHV ();
697		821
698	DEC_INC_DEPTH;	822	DEC_INC_DEPTH;
699		823
700	if ((*dec->cur & 31) == 31)	824	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
701	{	825	{
702	++dec->cur;	826	++dec->cur;
703		827
704	for (;;)	828	for (;;)
705	{	829	{
706	WANT (1);	830	WANT (1);
707		831
708	if (*dec->cur == (0xe0 \| 31))	832	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
709	{	833	{
710	++dec->cur;	834	++dec->cur;
711	break;	835	break;
712	}	836	}
713		837
…		…
734	static SV *	858	static SV *
735	decode_str (dec_t *dec, int utf8)	859	decode_str (dec_t *dec, int utf8)
736	{	860	{
737	SV *sv = 0;	861	SV *sv = 0;
738		862
739	if ((*dec->cur & 31) == 31)	863	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
740	{	864	{
		865	// indefinite length strings
741	++dec->cur;	866	++dec->cur;
742		867
		868	U8 major = *dec->cur & MAJOR_MISC;
		869
743	sv = newSVpvn ("", 0);	870	sv = newSVpvn ("", 0);
744		871
745	// not very fast, and certainly not robust against illegal input
746	for (;;)	872	for (;;)
747	{	873	{
748	WANT (1);	874	WANT (1);
749		875
750	if (*dec->cur == (0xe0 \| 31))	876	if ((*dec->cur - major) > LENGTH_EXT8)
		877	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
751	{	878	{
752	++dec->cur;	879	++dec->cur;
753	break;	880	break;
754	}	881	}
		882	else
		883	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
755		884
756	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;
757	}	890	}
758	}	891	}
759	else	892	else
760	{	893	{
761	STRLEN len = decode_uint (dec);	894	STRLEN len = decode_uint (dec);
…		…
768	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	901	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
769	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	902	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770	}	903	}
771		904
772	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
773	SvUTF8_on (sv);	911	SvUTF8_on (sv);
		912	}
774		913
775	return sv;	914	return sv;
776		915
777	fail:	916	fail:
778	SvREFCNT_dec (sv);	917	SvREFCNT_dec (sv);
…		…
810	}	949	}
811	break;	950	break;
812		951
813	case CBOR_TAG_STRINGREF:	952	case CBOR_TAG_STRINGREF:
814	{	953	{
815	if ((*dec->cur >> 5) != 0)	954	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
816	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	955	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
817		956
818	UV idx = decode_uint (dec);	957	UV idx = decode_uint (dec);
819		958
820	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	959	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
827	case CBOR_TAG_VALUE_SHAREABLE:	966	case CBOR_TAG_VALUE_SHAREABLE:
828	{	967	{
829	if (ecb_expect_false (!dec->shareable))	968	if (ecb_expect_false (!dec->shareable))
830	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	969	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
831		970
		971	if (dec->cbor.flags & F_ALLOW_CYCLES)
		972	{
832	sv = newSV (0);	973	sv = newSV (0);
833	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	974	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
834		975
835	SV *osv = decode_sv (dec);	976	SV *osv = decode_sv (dec);
836	sv_setsv (sv, osv);	977	sv_setsv (sv, osv);
837	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	}
838	}	987	}
839	break;	988	break;
840		989
841	case CBOR_TAG_VALUE_SHAREDREF:	990	case CBOR_TAG_VALUE_SHAREDREF:
842	{	991	{
843	if ((*dec->cur >> 5) != 0)	992	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
844	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	993	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
845		994
846	UV idx = decode_uint (dec);	995	UV idx = decode_uint (dec);
847		996
848	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	997	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
849	ERR ("corrupted CBOR data (sharedref index out of bounds)");	998	ERR ("corrupted CBOR data (sharedref index out of bounds)");
850		999
851	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");
852	}	1004	}
853	break;	1005	break;
854		1006
855	case CBOR_TAG_PERL_OBJECT:	1007	case CBOR_TAG_PERL_OBJECT:
856	{	1008	{
…		…
871	if (!method)	1023	if (!method)
872	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)");
873		1025
874	dSP;	1026	dSP;
875		1027
876	ENTER; SAVETMPS; PUSHMARK (SP);	1028	ENTER; SAVETMPS;
		1029	PUSHMARK (SP);
877	EXTEND (SP, len + 1);	1030	EXTEND (SP, len + 1);
878	// 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
879	PUSHs (*av_fetch (av, 0, 1));	1032	PUSHs (*av_fetch (av, 0, 1));
880	PUSHs (sv_cbor);	1033	PUSHs (sv_cbor);
881		1034
…		…
905		1058
906	default:	1059	default:
907	{	1060	{
908	sv = decode_sv (dec);	1061	sv = decode_sv (dec);
909		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	{
910	AV *av = newAV ();	1088	AV *av = newAV ();
911	av_push (av, newSVuv (tag));	1089	av_push (av, newSVuv (tag));
912	av_push (av, sv);	1090	av_push (av, sv);
913		1091
914	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1092	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
915	? cbor_tagged_stash	1093	? cbor_tagged_stash
916	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1094	: gv_stashpv ("CBOR::XS::Tagged" , 1);
917
918	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;
919	}	1101	}
920	break;	1102	break;
921	}	1103	}
922		1104
923	return sv;	1105	return sv;
…		…
930	static SV *	1112	static SV *
931	decode_sv (dec_t *dec)	1113	decode_sv (dec_t *dec)
932	{	1114	{
933	WANT (1);	1115	WANT (1);
934		1116
935	switch (*dec->cur >> 5)	1117	switch (*dec->cur >> MAJOR_SHIFT)
936	{	1118	{
937	case 0: // unsigned int	1119	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
938	return newSVuv (decode_uint (dec));	1120	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
939	case 1: // negative int	1121	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
940	return newSViv (-1 - (IV)decode_uint (dec));	1122	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
941	case 2: // octet string	1123	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
942	return decode_str (dec, 0);	1124	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
943	case 3: // utf-8 string	1125	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
944	return decode_str (dec, 1);	1126
945	case 4: // array	1127	case MAJOR_MISC >> MAJOR_SHIFT:
946	return decode_av (dec);
947	case 5: // map
948	return decode_hv (dec);
949	case 6: // tag
950	return decode_tagged (dec);
951	case 7: // misc
952	switch (*dec->cur++ & 31)	1128	switch (*dec->cur++ & MINOR_MASK)
953	{	1129	{
954	case 20:	1130	case SIMPLE_FALSE:
955	#if CBOR_SLOW	1131	#if CBOR_SLOW
956	types_false = get_bool ("Types::Serialiser::false");	1132	types_false = get_bool ("Types::Serialiser::false");
957	#endif	1133	#endif
958	return newSVsv (types_false);	1134	return newSVsv (types_false);
959	case 21:	1135	case SIMPLE_TRUE:
960	#if CBOR_SLOW	1136	#if CBOR_SLOW
961	types_true = get_bool ("Types::Serialiser::true");	1137	types_true = get_bool ("Types::Serialiser::true");
962	#endif	1138	#endif
963	return newSVsv (types_true);	1139	return newSVsv (types_true);
964	case 22:	1140	case SIMPLE_NULL:
965	return newSVsv (&PL_sv_undef);	1141	return newSVsv (&PL_sv_undef);
966	case 23:	1142	case SIMPLE_UNDEF:
967	#if CBOR_SLOW	1143	#if CBOR_SLOW
968	types_error = get_bool ("Types::Serialiser::error");	1144	types_error = get_bool ("Types::Serialiser::error");
969	#endif	1145	#endif
970	return newSVsv (types_error);	1146	return newSVsv (types_error);
971		1147
972	case 25:	1148	case MISC_FLOAT16:
973	{	1149	{
974	WANT (2);	1150	WANT (2);
975		1151
976	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1152	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
977	dec->cur += 2;	1153	dec->cur += 2;
978		1154
979	return newSVnv (ecb_binary16_to_float (fp));	1155	return newSVnv (ecb_binary16_to_float (fp));
980	}	1156	}
981		1157
982	case 26:	1158	case MISC_FLOAT32:
983	{	1159	{
984	uint32_t fp;	1160	uint32_t fp;
985	WANT (4);	1161	WANT (4);
986	memcpy (&fp, dec->cur, 4);	1162	memcpy (&fp, dec->cur, 4);
987	dec->cur += 4;	1163	dec->cur += 4;
…		…
990	fp = ecb_bswap32 (fp);	1166	fp = ecb_bswap32 (fp);
991		1167
992	return newSVnv (ecb_binary32_to_float (fp));	1168	return newSVnv (ecb_binary32_to_float (fp));
993	}	1169	}
994		1170
995	case 27:	1171	case MISC_FLOAT64:
996	{	1172	{
997	uint64_t fp;	1173	uint64_t fp;
998	WANT (8);	1174	WANT (8);
999	memcpy (&fp, dec->cur, 8);	1175	memcpy (&fp, dec->cur, 8);
1000	dec->cur += 8;	1176	dec->cur += 8;
…		…
1003	fp = ecb_bswap64 (fp);	1179	fp = ecb_bswap64 (fp);
1004		1180
1005	return newSVnv (ecb_binary64_to_double (fp));	1181	return newSVnv (ecb_binary64_to_double (fp));
1006	}	1182	}
1007		1183
1008	// 0..19 unassigned	1184	// 0..19 unassigned simple
1009	// 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
1010	default:	1188	default:
1011	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1189	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1012	}	1190	}
1013		1191
1014	break;	1192	break;
1015	}	1193	}
1016		1194
…		…
1019	}	1197	}
1020		1198
1021	static SV *	1199	static SV *
1022	decode_cbor (SV string, CBOR cbor, char **offset_return)	1200	decode_cbor (SV string, CBOR cbor, char **offset_return)
1023	{	1201	{
1024	dec_t dec = { };	1202	dec_t dec = { 0 };
1025	SV *sv;	1203	SV *sv;
1026	STRLEN len;	1204	STRLEN len;
1027	char *data = SvPVbyte (string, len);	1205	char *data = SvPVbyte (string, len);
1028		1206
1029	if (len > cbor->max_size && cbor->max_size)	1207	if (len > cbor->max_size && cbor->max_size)
…		…
1043	if (dec.cur != dec.end && !dec.err)	1221	if (dec.cur != dec.end && !dec.err)
1044	dec.err = "garbage after CBOR object";	1222	dec.err = "garbage after CBOR object";
1045		1223
1046	if (dec.err)	1224	if (dec.err)
1047	{	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
1048	SvREFCNT_dec (sv);	1237	SvREFCNT_dec (sv);
1049	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);
1050	}	1239	}
1051		1240
1052	sv = sv_2mortal (sv);	1241	sv = sv_2mortal (sv);
1053		1242
1054	return sv;	1243	return sv;
1055	}	1244	}
1056		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
1057	/////////////////////////////////////////////////////////////////////////////	1370	/////////////////////////////////////////////////////////////////////////////
1058	// XS interface functions	1371	// XS interface functions
1059		1372
1060	MODULE = CBOR::XS PACKAGE = CBOR::XS	1373	MODULE = CBOR::XS PACKAGE = CBOR::XS
1061		1374
…		…
1068	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);	1381	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1069		1382
1070	types_true = get_bool ("Types::Serialiser::true" );	1383	types_true = get_bool ("Types::Serialiser::true" );
1071	types_false = get_bool ("Types::Serialiser::false");	1384	types_false = get_bool ("Types::Serialiser::false");
1072	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);
1073		1388
1074	sv_cbor = newSVpv ("CBOR", 0);	1389	sv_cbor = newSVpv ("CBOR", 0);
1075	SvREADONLY_on (sv_cbor);	1390	SvREADONLY_on (sv_cbor);
1076	}	1391	}
1077		1392
…		…
1099	void shrink (CBOR *self, int enable = 1)	1414	void shrink (CBOR *self, int enable = 1)
1100	ALIAS:	1415	ALIAS:
1101	shrink = F_SHRINK	1416	shrink = F_SHRINK
1102	allow_unknown = F_ALLOW_UNKNOWN	1417	allow_unknown = F_ALLOW_UNKNOWN
1103	allow_sharing = F_ALLOW_SHARING	1418	allow_sharing = F_ALLOW_SHARING
1104	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
1105	PPCODE:	1423	PPCODE:
1106	{	1424	{
1107	if (enable)	1425	if (enable)
1108	self->flags \|= ix;	1426	self->flags \|= ix;
1109	else	1427	else
…		…
1115	void get_shrink (CBOR *self)	1433	void get_shrink (CBOR *self)
1116	ALIAS:	1434	ALIAS:
1117	get_shrink = F_SHRINK	1435	get_shrink = F_SHRINK
1118	get_allow_unknown = F_ALLOW_UNKNOWN	1436	get_allow_unknown = F_ALLOW_UNKNOWN
1119	get_allow_sharing = F_ALLOW_SHARING	1437	get_allow_sharing = F_ALLOW_SHARING
1120	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
1121	PPCODE:	1441	PPCODE:
1122	XPUSHs (boolSV (self->flags & ix));	1442	XPUSHs (boolSV (self->flags & ix));
1123		1443
1124	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1444	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1125	PPCODE:	1445	PPCODE:
…		…
1138	XPUSHs (ST (0));	1458	XPUSHs (ST (0));
1139		1459
1140	int get_max_size (CBOR *self)	1460	int get_max_size (CBOR *self)
1141	CODE:	1461	CODE:
1142	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);
1143	OUTPUT:	1475	OUTPUT:
1144	RETVAL	1476	RETVAL
1145		1477
1146	void encode (CBOR self, SV scalar)	1478	void encode (CBOR self, SV scalar)
1147	PPCODE:	1479	PPCODE:
…		…
1162	EXTEND (SP, 2);	1494	EXTEND (SP, 2);
1163	PUSHs (sv);	1495	PUSHs (sv);
1164	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1496	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1165	}	1497	}
1166		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
1167	PROTOTYPES: ENABLE	1555	PROTOTYPES: ENABLE
1168		1556
1169	void encode_cbor (SV *scalar)	1557	void encode_cbor (SV *scalar)
		1558	ALIAS:
		1559	encode_cbor = 0
		1560	encode_cbor_sharing = F_ALLOW_SHARING
1170	PPCODE:	1561	PPCODE:
1171	{	1562	{
1172	CBOR cbor;	1563	CBOR cbor;
1173	cbor_init (&cbor);	1564	cbor_init (&cbor);
		1565	cbor.flags \|= ix;
1174	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1566	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1175	XPUSHs (scalar);	1567	XPUSHs (scalar);
1176	}	1568	}
1177		1569
1178	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 1.26 by root, Fri Nov 22 09:40:13 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 1.26 by root, Fri Nov 22 09:40:13 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC