[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 1.48 by root, Sun Sep 20 23:25:53 2015 UTC

…		…
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
11		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	20	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	21	#endif
21	#ifndef HvNAMEUTF8	22	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	23	# define HvNAMEUTF8(hv) 0
23	#endif	24	#endif
		25	#ifndef SvREFCNT_dec_NN
		26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		27	#endif
		28
		29	// known major and minor types
		30	enum cbor_type
		31	{
		32	MAJOR_SHIFT = 5,
		33	MINOR_MASK = 0x1f,
		34
		35	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		36	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		37	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		38	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		39	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		40	MAJOR_MAP = 5 << MAJOR_SHIFT,
		41	MAJOR_TAG = 6 << MAJOR_SHIFT,
		42	MAJOR_MISC = 7 << MAJOR_SHIFT,
		43
		44	// INT/STRING/ARRAY/MAP subtypes
		45	LENGTH_EXT1 = 24,
		46	LENGTH_EXT2 = 25,
		47	LENGTH_EXT4 = 26,
		48	LENGTH_EXT8 = 27,
		49
		50	// SIMPLE types (effectively MISC subtypes)
		51	SIMPLE_FALSE = 20,
		52	SIMPLE_TRUE = 21,
		53	SIMPLE_NULL = 22,
		54	SIMPLE_UNDEF = 23,
		55
		56	// MISC subtype (unused)
		57	MISC_EXT1 = 24,
		58	MISC_FLOAT16 = 25,
		59	MISC_FLOAT32 = 26,
		60	MISC_FLOAT64 = 27,
		61
		62	// BYTES/TEXT/ARRAY/MAP
		63	MINOR_INDEF = 31,
		64	};
24		65
25	// known tags	66	// known tags
26	enum cbor_tag	67	enum cbor_tag
27	{	68	{
28	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		77
37	// rfc7049	78	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
44		85
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		90
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		96
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
57	};	98	};
58		99
59	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
62	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	103	#define F_ALLOW_CYCLES 0x00000008UL
		104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_VALIDATE_UTF8 0x00000020UL
63		106
64	#define INIT_SIZE 32 // initial scalar size to be allocated	107	#define INIT_SIZE 32 // initial scalar size to be allocated
65		108
66	#define SB do {	109	#define SB do {
67	#define SE } while (0)	110	#define SE } while (0)
…		…
79	# define CBOR_SLOW 0	122	# define CBOR_SLOW 0
80	# define CBOR_STASH cbor_stash	123	# define CBOR_STASH cbor_stash
81	#endif	124	#endif
82		125
83	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::	126	static HV cbor_stash, types_boolean_stash, types_error_stash, cbor_tagged_stash; // CBOR::XS::
84	static SV types_true, types_false, types_error, sv_cbor;	127	static SV types_true, types_false, types_error, sv_cbor, *default_filter;
85		128
86	typedef struct {	129	typedef struct {
87	U32 flags;	130	U32 flags;
88	U32 max_depth;	131	U32 max_depth;
89	STRLEN max_size;	132	STRLEN max_size;
		133	SV *filter;
		134
		135	// for the incremental parser
		136	STRLEN incr_pos; // the current offset into the text
		137	STRLEN incr_need; // minimum bytes needed to decode
		138	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90	} CBOR;	139	} CBOR;
91		140
92	ecb_inline void	141	ecb_inline void
93	cbor_init (CBOR *cbor)	142	cbor_init (CBOR *cbor)
94	{	143	{
95	Zero (cbor, 1, CBOR);	144	Zero (cbor, 1, CBOR);
96	cbor->max_depth = 512;	145	cbor->max_depth = 512;
		146	}
		147
		148	ecb_inline void
		149	cbor_free (CBOR *cbor)
		150	{
		151	SvREFCNT_dec (cbor->filter);
		152	SvREFCNT_dec (cbor->incr_count);
97	}	153	}
98		154
99	/////////////////////////////////////////////////////////////////////////////	155	/////////////////////////////////////////////////////////////////////////////
100	// utility functions	156	// utility functions
101		157
…		…
131	{	187	{
132	return idx > 23	188	return idx > 23
133	? idx > 0xffU	189	? idx > 0xffU
134	? idx > 0xffffU	190	? idx > 0xffffU
135	? idx > 0xffffffffU	191	? idx > 0xffffffffU
		192	? 11
136	? 7	193	: 7
137	: 6
138	: 5	194	: 5
139	: 4	195	: 4
140	: 3;	196	: 3;
141	}	197	}
142		198
…		…
179	static void	235	static void
180	encode_uint (enc_t *enc, int major, UV len)	236	encode_uint (enc_t *enc, int major, UV len)
181	{	237	{
182	need (enc, 9);	238	need (enc, 9);
183		239
184	if (len < 24)	240	if (ecb_expect_true (len < LENGTH_EXT1))
185	*enc->cur++ = major \| len;	241	*enc->cur++ = major \| len;
186	else if (len <= 0xff)	242	else if (ecb_expect_true (len <= 0xffU))
187	{	243	{
188	*enc->cur++ = major \| 24;	244	*enc->cur++ = major \| LENGTH_EXT1;
189	*enc->cur++ = len;	245	*enc->cur++ = len;
190	}	246	}
191	else if (len <= 0xffff)	247	else if (len <= 0xffffU)
192	{	248	{
193	*enc->cur++ = major \| 25;	249	*enc->cur++ = major \| LENGTH_EXT2;
194	*enc->cur++ = len >> 8;	250	*enc->cur++ = len >> 8;
195	*enc->cur++ = len;	251	*enc->cur++ = len;
196	}	252	}
197	else if (len <= 0xffffffff)	253	else if (len <= 0xffffffffU)
198	{	254	{
199	*enc->cur++ = major \| 26;	255	*enc->cur++ = major \| LENGTH_EXT4;
200	*enc->cur++ = len >> 24;	256	*enc->cur++ = len >> 24;
201	*enc->cur++ = len >> 16;	257	*enc->cur++ = len >> 16;
202	*enc->cur++ = len >> 8;	258	*enc->cur++ = len >> 8;
203	*enc->cur++ = len;	259	*enc->cur++ = len;
204	}	260	}
205	else	261	else
206	{	262	{
207	*enc->cur++ = major \| 27;	263	*enc->cur++ = major \| LENGTH_EXT8;
208	*enc->cur++ = len >> 56;	264	*enc->cur++ = len >> 56;
209	*enc->cur++ = len >> 48;	265	*enc->cur++ = len >> 48;
210	*enc->cur++ = len >> 40;	266	*enc->cur++ = len >> 40;
211	*enc->cur++ = len >> 32;	267	*enc->cur++ = len >> 32;
212	*enc->cur++ = len >> 24;	268	*enc->cur++ = len >> 24;
…		…
217	}	273	}
218		274
219	ecb_inline void	275	ecb_inline void
220	encode_tag (enc_t *enc, UV tag)	276	encode_tag (enc_t *enc, UV tag)
221	{	277	{
222	encode_uint (enc, 0xc0, tag);	278	encode_uint (enc, MAJOR_TAG, tag);
		279	}
		280
		281	ecb_inline void
		282	encode_str (enc_t enc, int utf8, char str, STRLEN len)
		283	{
		284	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		285	need (enc, len);
		286	memcpy (enc->cur, str, len);
		287	enc->cur += len;
223	}	288	}
224		289
225	static void	290	static void
226	encode_str (enc_t enc, int utf8, char str, STRLEN len)	291	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
227	{	292	{
228	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	293	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229	{	294	{
230	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	295	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231		296
232	if (SvOK (*svp))	297	if (SvOK (*svp))
233	{	298	{
234	// already registered, use stringref	299	// already registered, use stringref
235	encode_tag (enc, CBOR_TAG_STRINGREF);	300	encode_tag (enc, CBOR_TAG_STRINGREF);
236	encode_uint (enc, 0x00, SvUV (*svp));	301	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237	return;	302	return;
238	}	303	}
239	else if (len >= minimum_string_length (enc->stringref_idx))	304	else if (len >= minimum_string_length (enc->stringref_idx))
240	{	305	{
241	// register only	306	// register only
242	sv_setuv (*svp, enc->stringref_idx);	307	sv_setuv (*svp, enc->stringref_idx);
243	++enc->stringref_idx;	308	++enc->stringref_idx;
244	}	309	}
245	}	310	}
246		311
247	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	312	encode_str (enc, utf8, str, len);
248	need (enc, len);
249	memcpy (enc->cur, str, len);
250	enc->cur += len;
251	}	313	}
252		314
253	static void encode_sv (enc_t enc, SV sv);	315	static void encode_sv (enc_t enc, SV sv);
254		316
255	static void	317	static void
…		…
260	if (enc->depth >= enc->cbor.max_depth)	322	if (enc->depth >= enc->cbor.max_depth)
261	croak (ERR_NESTING_EXCEEDED);	323	croak (ERR_NESTING_EXCEEDED);
262		324
263	++enc->depth;	325	++enc->depth;
264		326
265	encode_uint (enc, 0x80, len + 1);	327	encode_uint (enc, MAJOR_ARRAY, len + 1);
266		328
		329	if (SvMAGICAL (av))
267	for (i = 0; i <= len; ++i)	330	for (i = 0; i <= len; ++i)
268	{	331	{
269	SV **svp = av_fetch (av, i, 0);	332	SV **svp = av_fetch (av, i, 0);
270	encode_sv (enc, svp ? *svp : &PL_sv_undef);	333	encode_sv (enc, svp ? *svp : &PL_sv_undef);
271	}	334	}
		335	else
		336	for (i = 0; i <= len; ++i)
		337	{
		338	SV *sv = AvARRAY (av)[i];
		339	encode_sv (enc, sv ? sv : &PL_sv_undef);
		340	}
272		341
273	--enc->depth;	342	--enc->depth;
274	}	343	}
275		344
276	static void	345	static void
…		…
285		354
286	int pairs = hv_iterinit (hv);	355	int pairs = hv_iterinit (hv);
287	int mg = SvMAGICAL (hv);	356	int mg = SvMAGICAL (hv);
288		357
289	if (mg)	358	if (mg)
290	encode_ch (enc, 0xa0 \| 31);	359	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
291	else	360	else
292	encode_uint (enc, 0xa0, pairs);	361	encode_uint (enc, MAJOR_MAP, pairs);
293		362
294	while ((he = hv_iternext (hv)))	363	while ((he = hv_iternext (hv)))
295	{	364	{
296	if (HeKLEN (he) == HEf_SVKEY)	365	if (HeKLEN (he) == HEf_SVKEY)
297	encode_sv (enc, HeSVKEY (he));	366	encode_sv (enc, HeSVKEY (he));
298	else	367	else
299	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	368	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300		369
301	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	370	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302	}	371	}
303		372
304	if (mg)	373	if (mg)
305	encode_ch (enc, 0xe0 \| 31);	374	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
306		375
307	--enc->depth;	376	--enc->depth;
308	}	377	}
309		378
310	// encode objects, arrays and special \0=false and \1=true values.	379	// encode objects, arrays and special \0=false and \1=true values.
311	static void	380	static void
312	encode_rv (enc_t enc, SV sv)	381	encode_rv (enc_t enc, SV sv)
313	{	382	{
314	SvGETMAGIC (sv);	383	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		384
338	svtype svt = SvTYPE (sv);	385	svtype svt = SvTYPE (sv);
339		386
340	if (ecb_expect_false (SvOBJECT (sv)))	387	if (ecb_expect_false (SvOBJECT (sv)))
341	{	388	{
…		…
348	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	395	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
349	? cbor_tagged_stash	396	? cbor_tagged_stash
350	: gv_stashpv ("CBOR::XS::Tagged" , 1);	397	: gv_stashpv ("CBOR::XS::Tagged" , 1);
351		398
352	HV *stash = SvSTASH (sv);	399	HV *stash = SvSTASH (sv);
353	GV *method;
354		400
355	if (stash == boolean_stash)	401	if (stash == boolean_stash)
356	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	402	{
		403	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		404	return;
		405	}
357	else if (stash == error_stash)	406	else if (stash == error_stash)
358	encode_ch (enc, 0xe0 \| 23);	407	{
		408	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		409	return;
		410	}
359	else if (stash == tagged_stash)	411	else if (stash == tagged_stash)
360	{	412	{
361	if (svt != SVt_PVAV)	413	if (svt != SVt_PVAV)
362	croak ("encountered CBOR::XS::Tagged object that isn't an array");	414	croak ("encountered CBOR::XS::Tagged object that isn't an array");
363		415
364	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	416	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
365	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	417	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		418
		419	return;
		420	}
		421	}
		422
		423	if (ecb_expect_false (SvREFCNT (sv) > 1)
		424	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		425	{
		426	if (!enc->shareable)
		427	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		428
		429	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		430
		431	if (SvOK (*svp))
366	}	432	{
		433	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		434	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		435	return;
		436	}
		437	else
		438	{
		439	sv_setuv (*svp, enc->shareable_idx);
		440	++enc->shareable_idx;
		441	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		442	}
		443	}
		444
		445	if (ecb_expect_false (SvOBJECT (sv)))
		446	{
		447	HV *stash = SvSTASH (sv);
		448	GV *method;
		449
367	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	450	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
368	{	451	{
369	dSP;	452	dSP;
370		453
371	ENTER; SAVETMPS; PUSHMARK (SP);	454	ENTER; SAVETMPS; PUSHMARK (SP);
372	// we re-bless the reference to get overload and other niceties right	455	// we re-bless the reference to get overload and other niceties right
…		…
404	// catch this surprisingly common error	487	// catch this surprisingly common error
405	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	488	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));	489	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
407		490
408	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	491	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
409	encode_uint (enc, 0x80, count + 1);	492	encode_uint (enc, MAJOR_ARRAY, count + 1);
410	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	493	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
411		494
412	while (count)	495	while (count)
413	encode_sv (enc, SP[1 - count--]);	496	encode_sv (enc, SP[1 - count--]);
414		497
415	PUTBACK;	498	PUTBACK;
…		…
436	{	519	{
437	double nv = SvNVX (sv);	520	double nv = SvNVX (sv);
438		521
439	need (enc, 9);	522	need (enc, 9);
440		523
441	if (ecb_expect_false (nv == (U32)nv))	524	if (ecb_expect_false (nv == (NV)(U32)nv))
442	encode_uint (enc, 0x00, (U32)nv);	525	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
443	//TODO: maybe I32?	526	//TODO: maybe I32?
444	else if (ecb_expect_false (nv == (float)nv))	527	else if (ecb_expect_false (nv == (float)nv))
445	{	528	{
446	uint32_t fp = ecb_float_to_binary32 (nv);	529	uint32_t fp = ecb_float_to_binary32 (nv);
447		530
448	*enc->cur++ = 0xe0 \| 26;	531	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
449		532
450	if (!ecb_big_endian ())	533	if (!ecb_big_endian ())
451	fp = ecb_bswap32 (fp);	534	fp = ecb_bswap32 (fp);
452		535
453	memcpy (enc->cur, &fp, 4);	536	memcpy (enc->cur, &fp, 4);
…		…
455	}	538	}
456	else	539	else
457	{	540	{
458	uint64_t fp = ecb_double_to_binary64 (nv);	541	uint64_t fp = ecb_double_to_binary64 (nv);
459		542
460	*enc->cur++ = 0xe0 \| 27;	543	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
461		544
462	if (!ecb_big_endian ())	545	if (!ecb_big_endian ())
463	fp = ecb_bswap64 (fp);	546	fp = ecb_bswap64 (fp);
464		547
465	memcpy (enc->cur, &fp, 8);	548	memcpy (enc->cur, &fp, 8);
…		…
474		557
475	if (SvPOKp (sv))	558	if (SvPOKp (sv))
476	{	559	{
477	STRLEN len;	560	STRLEN len;
478	char *str = SvPV (sv, len);	561	char *str = SvPV (sv, len);
479	encode_str (enc, SvUTF8 (sv), str, len);	562	encode_strref (enc, SvUTF8 (sv), str, len);
480	}	563	}
481	else if (SvNOKp (sv))	564	else if (SvNOKp (sv))
482	encode_nv (enc, sv);	565	encode_nv (enc, sv);
483	else if (SvIOKp (sv))	566	else if (SvIOKp (sv))
484	{	567	{
485	if (SvIsUV (sv))	568	if (SvIsUV (sv))
486	encode_uint (enc, 0x00, SvUVX (sv));	569	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
487	else if (SvIVX (sv) >= 0)	570	else if (SvIVX (sv) >= 0)
488	encode_uint (enc, 0x00, SvIVX (sv));	571	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
489	else	572	else
490	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	573	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
491	}	574	}
492	else if (SvROK (sv))	575	else if (SvROK (sv))
493	encode_rv (enc, SvRV (sv));	576	encode_rv (enc, SvRV (sv));
494	else if (!SvOK (sv))	577	else if (!SvOK (sv))
495	encode_ch (enc, 0xe0 \| 22);	578	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
496	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	579	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
497	encode_ch (enc, 0xe0 \| 23);	580	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
498	else	581	else
499	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	582	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
500	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	583	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
501	}	584	}
502		585
503	static SV *	586	static SV *
504	encode_cbor (SV scalar, CBOR cbor)	587	encode_cbor (SV scalar, CBOR cbor)
505	{	588	{
506	enc_t enc = { };	589	enc_t enc = { 0 };
507		590
508	enc.cbor = *cbor;	591	enc.cbor = *cbor;
509	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	592	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
510	enc.cur = SvPVX (enc.sv);	593	enc.cur = SvPVX (enc.sv);
511	enc.end = SvEND (enc.sv);	594	enc.end = SvEND (enc.sv);
512		595
513	SvPOK_only (enc.sv);	596	SvPOK_only (enc.sv);
514		597
515	if (cbor->flags & F_ALLOW_STRINGREF)	598	if (cbor->flags & F_PACK_STRINGS)
516	{	599	{
517	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	600	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
518	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	601	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
519	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	602	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
520	}	603	}
…		…
542	CBOR cbor;	625	CBOR cbor;
543	U32 depth; // recursion depth	626	U32 depth; // recursion depth
544	U32 maxdepth; // recursion depth limit	627	U32 maxdepth; // recursion depth limit
545	AV *shareable;	628	AV *shareable;
546	AV *stringref;	629	AV *stringref;
		630	SV *decode_tagged;
547	} dec_t;	631	} dec_t;
548		632
549	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	633	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
550		634
551	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	635	#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	638	#define DEC_DEC_DEPTH --dec->depth
555		639
556	static UV	640	static UV
557	decode_uint (dec_t *dec)	641	decode_uint (dec_t *dec)
558	{	642	{
559	switch (*dec->cur & 31)	643	U8 m = *dec->cur & MINOR_MASK;
560	{	644	++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		645
566	case 24:	646	if (ecb_expect_true (m < LENGTH_EXT1))
		647	return m;
		648	else if (ecb_expect_true (m == LENGTH_EXT1))
		649	{
567	WANT (2);	650	WANT (1);
568	dec->cur += 2;	651	dec->cur += 1;
569	return dec->cur[-1];	652	return dec->cur[-1];
570		653	}
571	case 25:	654	else if (ecb_expect_true (m == LENGTH_EXT2))
		655	{
572	WANT (3);	656	WANT (2);
573	dec->cur += 3;	657	dec->cur += 2;
574	return (((UV)dec->cur[-2]) << 8)	658	return (((UV)dec->cur[-2]) << 8)
575	\| ((UV)dec->cur[-1]);	659	\| ((UV)dec->cur[-1]);
576		660	}
577	case 26:	661	else if (ecb_expect_true (m == LENGTH_EXT4))
		662	{
578	WANT (5);	663	WANT (4);
579	dec->cur += 5;	664	dec->cur += 4;
580	return (((UV)dec->cur[-4]) << 24)	665	return (((UV)dec->cur[-4]) << 24)
581	\| (((UV)dec->cur[-3]) << 16)	666	\| (((UV)dec->cur[-3]) << 16)
582	\| (((UV)dec->cur[-2]) << 8)	667	\| (((UV)dec->cur[-2]) << 8)
583	\| ((UV)dec->cur[-1]);	668	\| ((UV)dec->cur[-1]);
584		669	}
585	case 27:	670	else if (ecb_expect_true (m == LENGTH_EXT8))
		671	{
586	WANT (9);	672	WANT (8);
587	dec->cur += 9;	673	dec->cur += 8;
		674
		675	return
		676	#if UVSIZE < 8
		677	0
		678	#else
588	return (((UV)dec->cur[-8]) << 56)	679	(((UV)dec->cur[-8]) << 56)
589	\| (((UV)dec->cur[-7]) << 48)	680	\| (((UV)dec->cur[-7]) << 48)
590	\| (((UV)dec->cur[-6]) << 40)	681	\| (((UV)dec->cur[-6]) << 40)
591	\| (((UV)dec->cur[-5]) << 32)	682	\| (((UV)dec->cur[-5]) << 32)
		683	#endif
592	\| (((UV)dec->cur[-4]) << 24)	684	\| (((UV)dec->cur[-4]) << 24)
593	\| (((UV)dec->cur[-3]) << 16)	685	\| (((UV)dec->cur[-3]) << 16)
594	\| (((UV)dec->cur[-2]) << 8)	686	\| (((UV)dec->cur[-2]) << 8)
595	\| ((UV)dec->cur[-1]);	687	\| ((UV)dec->cur[-1]);
596		688	}
597	default:	689	else
598	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	690	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
599	}
600		691
601	fail:	692	fail:
602	return 0;	693	return 0;
603	}	694	}
604		695
…		…
609	{	700	{
610	AV *av = newAV ();	701	AV *av = newAV ();
611		702
612	DEC_INC_DEPTH;	703	DEC_INC_DEPTH;
613		704
614	if ((*dec->cur & 31) == 31)	705	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
615	{	706	{
616	++dec->cur;	707	++dec->cur;
617		708
618	for (;;)	709	for (;;)
619	{	710	{
620	WANT (1);	711	WANT (1);
621		712
622	if (*dec->cur == (0xe0 \| 31))	713	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
623	{	714	{
624	++dec->cur;	715	++dec->cur;
625	break;	716	break;
626	}	717	}
627		718
…		…
630	}	721	}
631	else	722	else
632	{	723	{
633	int i, len = decode_uint (dec);	724	int i, len = decode_uint (dec);
634		725
		726	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);	727	av_fill (av, len - 1);
636		728
637	for (i = 0; i < len; ++i)	729	for (i = 0; i < len; ++i)
638	AvARRAY (av)[i] = decode_sv (dec);	730	AvARRAY (av)[i] = decode_sv (dec);
639	}	731	}
…		…
652	{	744	{
653	// for speed reasons, we specialcase single-string	745	// for speed reasons, we specialcase single-string
654	// byte or utf-8 strings as keys, but only when !stringref	746	// byte or utf-8 strings as keys, but only when !stringref
655		747
656	if (ecb_expect_true (!dec->stringref))	748	if (ecb_expect_true (!dec->stringref))
657	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	749	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
658	{	750	{
659	I32 len = decode_uint (dec);	751	I32 len = decode_uint (dec);
660	char key = (char )dec->cur;	752	char key = (char )dec->cur;
661		753
662	dec->cur += len;	754	dec->cur += len;
663		755
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);	756	hv_store (hv, key, len, decode_sv (dec), 0);
668		757
669	return;	758	return;
670	}	759	}
671	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	760	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
672	{	761	{
673	I32 len = decode_uint (dec);	762	I32 len = decode_uint (dec);
674	char key = (char )dec->cur;	763	char key = (char )dec->cur;
675		764
676	dec->cur += len;	765	dec->cur += len;
677		766
678	if (ecb_expect_false (dec->stringref))	767	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
679	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	768	if (!is_utf8_string (key, len))
		769	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
680		770
681	hv_store (hv, key, -len, decode_sv (dec), 0);	771	hv_store (hv, key, -len, decode_sv (dec), 0);
682		772
683	return;	773	return;
684	}	774	}
…		…
686	SV *k = decode_sv (dec);	776	SV *k = decode_sv (dec);
687	SV *v = decode_sv (dec);	777	SV *v = decode_sv (dec);
688		778
689	hv_store_ent (hv, k, v, 0);	779	hv_store_ent (hv, k, v, 0);
690	SvREFCNT_dec (k);	780	SvREFCNT_dec (k);
		781
		782	fail:
		783	;
691	}	784	}
692		785
693	static SV *	786	static SV *
694	decode_hv (dec_t *dec)	787	decode_hv (dec_t *dec)
695	{	788	{
696	HV *hv = newHV ();	789	HV *hv = newHV ();
697		790
698	DEC_INC_DEPTH;	791	DEC_INC_DEPTH;
699		792
700	if ((*dec->cur & 31) == 31)	793	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
701	{	794	{
702	++dec->cur;	795	++dec->cur;
703		796
704	for (;;)	797	for (;;)
705	{	798	{
706	WANT (1);	799	WANT (1);
707		800
708	if (*dec->cur == (0xe0 \| 31))	801	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
709	{	802	{
710	++dec->cur;	803	++dec->cur;
711	break;	804	break;
712	}	805	}
713		806
…		…
734	static SV *	827	static SV *
735	decode_str (dec_t *dec, int utf8)	828	decode_str (dec_t *dec, int utf8)
736	{	829	{
737	SV *sv = 0;	830	SV *sv = 0;
738		831
739	if ((*dec->cur & 31) == 31)	832	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
740	{	833	{
		834	// indefinite length strings
741	++dec->cur;	835	++dec->cur;
742		836
		837	U8 major = *dec->cur & MAJOR_MISC;
		838
743	sv = newSVpvn ("", 0);	839	sv = newSVpvn ("", 0);
744		840
745	// not very fast, and certainly not robust against illegal input
746	for (;;)	841	for (;;)
747	{	842	{
748	WANT (1);	843	WANT (1);
749		844
750	if (*dec->cur == (0xe0 \| 31))	845	if ((*dec->cur - major) > LENGTH_EXT8)
		846	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
751	{	847	{
752	++dec->cur;	848	++dec->cur;
753	break;	849	break;
754	}	850	}
		851	else
		852	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
755		853
756	sv_catsv (sv, decode_sv (dec));	854	STRLEN len = decode_uint (dec);
		855
		856	WANT (len);
		857	sv_catpvn (sv, dec->cur, len);
		858	dec->cur += len;
757	}	859	}
758	}	860	}
759	else	861	else
760	{	862	{
761	STRLEN len = decode_uint (dec);	863	STRLEN len = decode_uint (dec);
…		…
768	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	870	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
769	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	871	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770	}	872	}
771		873
772	if (utf8)	874	if (utf8)
		875	{
		876	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		877	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		878	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		879
773	SvUTF8_on (sv);	880	SvUTF8_on (sv);
		881	}
774		882
775	return sv;	883	return sv;
776		884
777	fail:	885	fail:
778	SvREFCNT_dec (sv);	886	SvREFCNT_dec (sv);
…		…
810	}	918	}
811	break;	919	break;
812		920
813	case CBOR_TAG_STRINGREF:	921	case CBOR_TAG_STRINGREF:
814	{	922	{
815	if ((*dec->cur >> 5) != 0)	923	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
816	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	924	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
817		925
818	UV idx = decode_uint (dec);	926	UV idx = decode_uint (dec);
819		927
820	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	928	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
827	case CBOR_TAG_VALUE_SHAREABLE:	935	case CBOR_TAG_VALUE_SHAREABLE:
828	{	936	{
829	if (ecb_expect_false (!dec->shareable))	937	if (ecb_expect_false (!dec->shareable))
830	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	938	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
831		939
		940	if (dec->cbor.flags & F_ALLOW_CYCLES)
		941	{
832	sv = newSV (0);	942	sv = newSV (0);
833	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	943	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
834		944
835	SV *osv = decode_sv (dec);	945	SV *osv = decode_sv (dec);
836	sv_setsv (sv, osv);	946	sv_setsv (sv, osv);
837	SvREFCNT_dec_NN (osv);	947	SvREFCNT_dec_NN (osv);
		948	}
		949	else
		950	{
		951	av_push (dec->shareable, &PL_sv_undef);
		952	int idx = AvFILLp (dec->shareable);
		953	sv = decode_sv (dec);
		954	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		955	}
838	}	956	}
839	break;	957	break;
840		958
841	case CBOR_TAG_VALUE_SHAREDREF:	959	case CBOR_TAG_VALUE_SHAREDREF:
842	{	960	{
843	if ((*dec->cur >> 5) != 0)	961	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
844	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	962	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
845		963
846	UV idx = decode_uint (dec);	964	UV idx = decode_uint (dec);
847		965
848	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	966	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
849	ERR ("corrupted CBOR data (sharedref index out of bounds)");	967	ERR ("corrupted CBOR data (sharedref index out of bounds)");
850		968
851	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	969	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		970
		971	if (sv == &PL_sv_undef)
		972	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
852	}	973	}
853	break;	974	break;
854		975
855	case CBOR_TAG_PERL_OBJECT:	976	case CBOR_TAG_PERL_OBJECT:
856	{	977	{
…		…
905		1026
906	default:	1027	default:
907	{	1028	{
908	sv = decode_sv (dec);	1029	sv = decode_sv (dec);
909		1030
		1031	dSP;
		1032	ENTER; SAVETMPS; PUSHMARK (SP);
		1033	EXTEND (SP, 2);
		1034	PUSHs (newSVuv (tag));
		1035	PUSHs (sv);
		1036
		1037	PUTBACK;
		1038	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
		1039	SPAGAIN;
		1040
		1041	if (SvTRUE (ERRSV))
		1042	{
		1043	FREETMPS; LEAVE;
		1044	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
		1045	}
		1046
		1047	if (count)
		1048	{
		1049	SvREFCNT_dec (sv);
		1050	sv = SvREFCNT_inc (POPs);
		1051	}
		1052	else
		1053	{
910	AV *av = newAV ();	1054	AV *av = newAV ();
911	av_push (av, newSVuv (tag));	1055	av_push (av, newSVuv (tag));
912	av_push (av, sv);	1056	av_push (av, sv);
913		1057
914	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1058	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
915	? cbor_tagged_stash	1059	? cbor_tagged_stash
916	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1060	: gv_stashpv ("CBOR::XS::Tagged" , 1);
917
918	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);	1061	sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
		1062	}
		1063
		1064	PUTBACK;
		1065
		1066	FREETMPS; LEAVE;
919	}	1067	}
920	break;	1068	break;
921	}	1069	}
922		1070
923	return sv;	1071	return sv;
…		…
930	static SV *	1078	static SV *
931	decode_sv (dec_t *dec)	1079	decode_sv (dec_t *dec)
932	{	1080	{
933	WANT (1);	1081	WANT (1);
934		1082
935	switch (*dec->cur >> 5)	1083	switch (*dec->cur >> MAJOR_SHIFT)
936	{	1084	{
937	case 0: // unsigned int	1085	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
938	return newSVuv (decode_uint (dec));	1086	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
939	case 1: // negative int	1087	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
940	return newSViv (-1 - (IV)decode_uint (dec));	1088	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
941	case 2: // octet string	1089	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
942	return decode_str (dec, 0);	1090	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
943	case 3: // utf-8 string	1091	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
944	return decode_str (dec, 1);	1092
945	case 4: // array	1093	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)	1094	switch (*dec->cur++ & MINOR_MASK)
953	{	1095	{
954	case 20:	1096	case SIMPLE_FALSE:
955	#if CBOR_SLOW	1097	#if CBOR_SLOW
956	types_false = get_bool ("Types::Serialiser::false");	1098	types_false = get_bool ("Types::Serialiser::false");
957	#endif	1099	#endif
958	return newSVsv (types_false);	1100	return newSVsv (types_false);
959	case 21:	1101	case SIMPLE_TRUE:
960	#if CBOR_SLOW	1102	#if CBOR_SLOW
961	types_true = get_bool ("Types::Serialiser::true");	1103	types_true = get_bool ("Types::Serialiser::true");
962	#endif	1104	#endif
963	return newSVsv (types_true);	1105	return newSVsv (types_true);
964	case 22:	1106	case SIMPLE_NULL:
965	return newSVsv (&PL_sv_undef);	1107	return newSVsv (&PL_sv_undef);
966	case 23:	1108	case SIMPLE_UNDEF:
967	#if CBOR_SLOW	1109	#if CBOR_SLOW
968	types_error = get_bool ("Types::Serialiser::error");	1110	types_error = get_bool ("Types::Serialiser::error");
969	#endif	1111	#endif
970	return newSVsv (types_error);	1112	return newSVsv (types_error);
971		1113
972	case 25:	1114	case MISC_FLOAT16:
973	{	1115	{
974	WANT (2);	1116	WANT (2);
975		1117
976	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1118	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
977	dec->cur += 2;	1119	dec->cur += 2;
978		1120
979	return newSVnv (ecb_binary16_to_float (fp));	1121	return newSVnv (ecb_binary16_to_float (fp));
980	}	1122	}
981		1123
982	case 26:	1124	case MISC_FLOAT32:
983	{	1125	{
984	uint32_t fp;	1126	uint32_t fp;
985	WANT (4);	1127	WANT (4);
986	memcpy (&fp, dec->cur, 4);	1128	memcpy (&fp, dec->cur, 4);
987	dec->cur += 4;	1129	dec->cur += 4;
…		…
990	fp = ecb_bswap32 (fp);	1132	fp = ecb_bswap32 (fp);
991		1133
992	return newSVnv (ecb_binary32_to_float (fp));	1134	return newSVnv (ecb_binary32_to_float (fp));
993	}	1135	}
994		1136
995	case 27:	1137	case MISC_FLOAT64:
996	{	1138	{
997	uint64_t fp;	1139	uint64_t fp;
998	WANT (8);	1140	WANT (8);
999	memcpy (&fp, dec->cur, 8);	1141	memcpy (&fp, dec->cur, 8);
1000	dec->cur += 8;	1142	dec->cur += 8;
…		…
1003	fp = ecb_bswap64 (fp);	1145	fp = ecb_bswap64 (fp);
1004		1146
1005	return newSVnv (ecb_binary64_to_double (fp));	1147	return newSVnv (ecb_binary64_to_double (fp));
1006	}	1148	}
1007		1149
1008	// 0..19 unassigned	1150	// 0..19 unassigned simple
1009	// 24 reserved + unassigned (reserved values are not encodable)	1151	// 24 reserved + unassigned simple (reserved values are not encodable)
		1152	// 28-30 unassigned misc
		1153	// 31 break code
1010	default:	1154	default:
1011	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1155	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1012	}	1156	}
1013		1157
1014	break;	1158	break;
1015	}	1159	}
1016		1160
…		…
1019	}	1163	}
1020		1164
1021	static SV *	1165	static SV *
1022	decode_cbor (SV string, CBOR cbor, char **offset_return)	1166	decode_cbor (SV string, CBOR cbor, char **offset_return)
1023	{	1167	{
1024	dec_t dec = { };	1168	dec_t dec = { 0 };
1025	SV *sv;	1169	SV *sv;
1026	STRLEN len;	1170	STRLEN len;
1027	char *data = SvPVbyte (string, len);	1171	char *data = SvPVbyte (string, len);
1028		1172
1029	if (len > cbor->max_size && cbor->max_size)	1173	if (len > cbor->max_size && cbor->max_size)
…		…
1043	if (dec.cur != dec.end && !dec.err)	1187	if (dec.cur != dec.end && !dec.err)
1044	dec.err = "garbage after CBOR object";	1188	dec.err = "garbage after CBOR object";
1045		1189
1046	if (dec.err)	1190	if (dec.err)
1047	{	1191	{
		1192	if (dec.shareable)
		1193	{
		1194	// need to break cyclic links, which whould all be in shareable
		1195	int i;
		1196	SV **svp;
		1197
		1198	for (i = av_len (dec.shareable) + 1; i--; )
		1199	if ((svp = av_fetch (dec.shareable, i, 0)))
		1200	sv_setsv (*svp, &PL_sv_undef);
		1201	}
		1202
1048	SvREFCNT_dec (sv);	1203	SvREFCNT_dec (sv);
1049	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1204	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1050	}	1205	}
1051		1206
1052	sv = sv_2mortal (sv);	1207	sv = sv_2mortal (sv);
1053		1208
1054	return sv;	1209	return sv;
1055	}	1210	}
1056		1211
		1212	/////////////////////////////////////////////////////////////////////////////
		1213	// incremental parser
		1214
		1215	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1216
		1217	// returns 0 for notyet, 1 for success or error
		1218	static int
		1219	incr_parse (CBOR self, SV cborstr)
		1220	{
		1221	STRLEN cur;
		1222	SvPV (cborstr, cur);
		1223
		1224	while (ecb_expect_true (self->incr_need <= cur))
		1225	{
		1226	// table of integer count bytes
		1227	static I8 incr_len[MINOR_MASK + 1] = {
		1228	0, 0, 0, 0, 0, 0, 0, 0,
		1229	0, 0, 0, 0, 0, 0, 0, 0,
		1230	0, 0, 0, 0, 0, 0, 0, 0,
		1231	1, 2, 4, 8,-1,-1,-1,-2
		1232	};
		1233
		1234	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1235	U8 m = *p & MINOR_MASK;
		1236	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1237	I8 ilen = incr_len[m];
		1238
		1239	self->incr_need = self->incr_pos + 1;
		1240
		1241	if (ecb_expect_false (ilen < 0))
		1242	{
		1243	if (m != MINOR_INDEF)
		1244	return 1; // error
		1245
		1246	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1247	{
		1248	if (count >= 0)
		1249	return 1; // error
		1250
		1251	count = 1;
		1252	}
		1253	else
		1254	{
		1255	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1256	count = -1;
		1257	}
		1258	}
		1259	else
		1260	{
		1261	self->incr_need += ilen;
		1262	if (ecb_expect_false (self->incr_need > cur))
		1263	return 0;
		1264
		1265	int major = *p >> MAJOR_SHIFT;
		1266
		1267	switch (major)
		1268	{
		1269	case MAJOR_TAG >> MAJOR_SHIFT:
		1270	++count; // tags merely prefix another value
		1271	break;
		1272
		1273	case MAJOR_BYTES >> MAJOR_SHIFT:
		1274	case MAJOR_TEXT >> MAJOR_SHIFT:
		1275	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1276	case MAJOR_MAP >> MAJOR_SHIFT:
		1277	{
		1278	UV len;
		1279
		1280	if (ecb_expect_false (ilen))
		1281	{
		1282	len = 0;
		1283
		1284	do {
		1285	len = (len << 8) \| *++p;
		1286	} while (--ilen);
		1287	}
		1288	else
		1289	len = m;
		1290
		1291	switch (major)
		1292	{
		1293	case MAJOR_BYTES >> MAJOR_SHIFT:
		1294	case MAJOR_TEXT >> MAJOR_SHIFT:
		1295	self->incr_need += len;
		1296	if (ecb_expect_false (self->incr_need > cur))
		1297	return 0;
		1298
		1299	break;
		1300
		1301	case MAJOR_MAP >> MAJOR_SHIFT:
		1302	len <<= 1;
		1303	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1304	if (len)
		1305	{
		1306	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1307	count = len + 1;
		1308	}
		1309	break;
		1310	}
		1311	}
		1312	}
		1313	}
		1314
		1315	self->incr_pos = self->incr_need;
		1316
		1317	if (count > 0)
		1318	{
		1319	while (!--count)
		1320	{
		1321	if (!AvFILLp (self->incr_count))
		1322	return 1; // done
		1323
		1324	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1325	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1326	}
		1327
		1328	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1329	}
		1330	}
		1331
		1332	return 0;
		1333	}
		1334
		1335
1057	/////////////////////////////////////////////////////////////////////////////	1336	/////////////////////////////////////////////////////////////////////////////
1058	// XS interface functions	1337	// XS interface functions
1059		1338
1060	MODULE = CBOR::XS PACKAGE = CBOR::XS	1339	MODULE = CBOR::XS PACKAGE = CBOR::XS
1061		1340
…		…
1068	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);	1347	types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1069		1348
1070	types_true = get_bool ("Types::Serialiser::true" );	1349	types_true = get_bool ("Types::Serialiser::true" );
1071	types_false = get_bool ("Types::Serialiser::false");	1350	types_false = get_bool ("Types::Serialiser::false");
1072	types_error = get_bool ("Types::Serialiser::error");	1351	types_error = get_bool ("Types::Serialiser::error");
		1352
		1353	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1073		1354
1074	sv_cbor = newSVpv ("CBOR", 0);	1355	sv_cbor = newSVpv ("CBOR", 0);
1075	SvREADONLY_on (sv_cbor);	1356	SvREADONLY_on (sv_cbor);
1076	}	1357	}
1077		1358
…		…
1099	void shrink (CBOR *self, int enable = 1)	1380	void shrink (CBOR *self, int enable = 1)
1100	ALIAS:	1381	ALIAS:
1101	shrink = F_SHRINK	1382	shrink = F_SHRINK
1102	allow_unknown = F_ALLOW_UNKNOWN	1383	allow_unknown = F_ALLOW_UNKNOWN
1103	allow_sharing = F_ALLOW_SHARING	1384	allow_sharing = F_ALLOW_SHARING
1104	allow_stringref = F_ALLOW_STRINGREF	1385	allow_cycles = F_ALLOW_CYCLES
		1386	pack_strings = F_PACK_STRINGS
		1387	validate_utf8 = F_VALIDATE_UTF8
1105	PPCODE:	1388	PPCODE:
1106	{	1389	{
1107	if (enable)	1390	if (enable)
1108	self->flags \|= ix;	1391	self->flags \|= ix;
1109	else	1392	else
…		…
1115	void get_shrink (CBOR *self)	1398	void get_shrink (CBOR *self)
1116	ALIAS:	1399	ALIAS:
1117	get_shrink = F_SHRINK	1400	get_shrink = F_SHRINK
1118	get_allow_unknown = F_ALLOW_UNKNOWN	1401	get_allow_unknown = F_ALLOW_UNKNOWN
1119	get_allow_sharing = F_ALLOW_SHARING	1402	get_allow_sharing = F_ALLOW_SHARING
1120	get_allow_stringref = F_ALLOW_STRINGREF	1403	get_allow_cycles = F_ALLOW_CYCLES
		1404	get_pack_strings = F_PACK_STRINGS
		1405	get_validate_utf8 = F_VALIDATE_UTF8
1121	PPCODE:	1406	PPCODE:
1122	XPUSHs (boolSV (self->flags & ix));	1407	XPUSHs (boolSV (self->flags & ix));
1123		1408
1124	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1409	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1125	PPCODE:	1410	PPCODE:
…		…
1138	XPUSHs (ST (0));	1423	XPUSHs (ST (0));
1139		1424
1140	int get_max_size (CBOR *self)	1425	int get_max_size (CBOR *self)
1141	CODE:	1426	CODE:
1142	RETVAL = self->max_size;	1427	RETVAL = self->max_size;
		1428	OUTPUT:
		1429	RETVAL
		1430
		1431	void filter (CBOR self, SV filter = 0)
		1432	PPCODE:
		1433	SvREFCNT_dec (self->filter);
		1434	self->filter = filter ? newSVsv (filter) : filter;
		1435	XPUSHs (ST (0));
		1436
		1437	SV get_filter (CBOR self)
		1438	CODE:
		1439	RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1143	OUTPUT:	1440	OUTPUT:
1144	RETVAL	1441	RETVAL
1145		1442
1146	void encode (CBOR self, SV scalar)	1443	void encode (CBOR self, SV scalar)
1147	PPCODE:	1444	PPCODE:
…		…
1162	EXTEND (SP, 2);	1459	EXTEND (SP, 2);
1163	PUSHs (sv);	1460	PUSHs (sv);
1164	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1461	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1165	}	1462	}
1166		1463
		1464	void incr_parse (CBOR self, SV cborstr)
		1465	ALIAS:
		1466	incr_parse_multiple = 1
		1467	PPCODE:
		1468	{
		1469	if (SvUTF8 (cborstr))
		1470	sv_utf8_downgrade (cborstr, 0);
		1471
		1472	if (!self->incr_count)
		1473	{
		1474	self->incr_count = newAV ();
		1475	self->incr_pos = 0;
		1476	self->incr_need = 1;
		1477
		1478	av_push (self->incr_count, newSViv (1));
		1479	}
		1480
		1481	do
		1482	{
		1483	if (!incr_parse (self, cborstr))
		1484	{
		1485	if (self->incr_need > self->max_size && self->max_size)
		1486	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1487	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1488
		1489	break;
		1490	}
		1491
		1492	SV *sv;
		1493	char *offset;
		1494
		1495	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1496	XPUSHs (sv);
		1497
		1498	sv_chop (cborstr, offset);
		1499
		1500	av_clear (self->incr_count);
		1501	av_push (self->incr_count, newSViv (1));
		1502
		1503	self->incr_pos = 0;
		1504	self->incr_need = self->incr_pos + 1;
		1505	}
		1506	while (ix);
		1507	}
		1508
		1509	void incr_reset (CBOR *self)
		1510	CODE:
		1511	{
		1512	SvREFCNT_dec (self->incr_count);
		1513	self->incr_count = 0;
		1514	}
		1515
		1516	void DESTROY (CBOR *self)
		1517	PPCODE:
		1518	cbor_free (self);
		1519
1167	PROTOTYPES: ENABLE	1520	PROTOTYPES: ENABLE
1168		1521
1169	void encode_cbor (SV *scalar)	1522	void encode_cbor (SV *scalar)
		1523	ALIAS:
		1524	encode_cbor = 0
		1525	encode_cbor_sharing = F_ALLOW_SHARING
1170	PPCODE:	1526	PPCODE:
1171	{	1527	{
1172	CBOR cbor;	1528	CBOR cbor;
1173	cbor_init (&cbor);	1529	cbor_init (&cbor);
		1530	cbor.flags \|= ix;
1174	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1531	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1175	XPUSHs (scalar);	1532	XPUSHs (scalar);
1176	}	1533	}
1177		1534
1178	void decode_cbor (SV *cborstr)	1535	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 1.48 by root, Sun Sep 20 23:25:53 2015 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 1.48 by root, Sun Sep 20 23:25:53 2015 UTC