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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.28 by root, Sat Nov 23 18:30:59 2013 UTC vs.
Revision 1.51 by root, Sun Apr 24 13:15:19 2016 UTC

…		…
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
11		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
23	#endif	24	#endif
24	#ifndef SvREFCNT_dec_NN	25	#ifndef SvREFCNT_dec_NN
25	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)	26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26	#endif	27	#endif
27		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	};
		65
28	// known tags	66	// known tags
29	enum cbor_tag	67	enum cbor_tag
30	{	68	{
31	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
32	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
33	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
34	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
35	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
36	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
37	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
38	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
39		77
40	// rfc7049	78	// rfc7049
41	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
42	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
43	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
44	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
45	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
46	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
47		85
48	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
49	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
50	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
51	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
52		90
53	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
54	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
55	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
56	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
57	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
58		96
59	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
60	};	98	};
61		99
62	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
65	#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
66		107
67	#define INIT_SIZE 32 // initial scalar size to be allocated	108	#define INIT_SIZE 32 // initial scalar size to be allocated
68		109
69	#define SB do {	110	#define SB do {
70	#define SE } while (0)	111	#define SE } while (0)
…		…
89	typedef struct {	130	typedef struct {
90	U32 flags;	131	U32 flags;
91	U32 max_depth;	132	U32 max_depth;
92	STRLEN max_size;	133	STRLEN max_size;
93	SV *filter;	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.
94	} CBOR;	140	} CBOR;
95		141
96	ecb_inline void	142	ecb_inline void
97	cbor_init (CBOR *cbor)	143	cbor_init (CBOR *cbor)
98	{	144	{
…		…
102		148
103	ecb_inline void	149	ecb_inline void
104	cbor_free (CBOR *cbor)	150	cbor_free (CBOR *cbor)
105	{	151	{
106	SvREFCNT_dec (cbor->filter);	152	SvREFCNT_dec (cbor->filter);
		153	SvREFCNT_dec (cbor->incr_count);
107	}	154	}
108		155
109	/////////////////////////////////////////////////////////////////////////////	156	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	157	// utility functions
111		158
…		…
141	{	188	{
142	return idx > 23	189	return idx > 23
143	? idx > 0xffU	190	? idx > 0xffU
144	? idx > 0xffffU	191	? idx > 0xffffU
145	? idx > 0xffffffffU	192	? idx > 0xffffffffU
		193	? 11
146	? 7	194	: 7
147	: 6
148	: 5	195	: 5
149	: 4	196	: 4
150	: 3;	197	: 3;
151	}	198	}
152		199
…		…
189	static void	236	static void
190	encode_uint (enc_t *enc, int major, UV len)	237	encode_uint (enc_t *enc, int major, UV len)
191	{	238	{
192	need (enc, 9);	239	need (enc, 9);
193		240
194	if (len < 24)	241	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major \| len;	242	*enc->cur++ = major \| len;
196	else if (len <= 0xff)	243	else if (ecb_expect_true (len <= 0xffU))
197	{	244	{
198	*enc->cur++ = major \| 24;	245	*enc->cur++ = major \| LENGTH_EXT1;
199	*enc->cur++ = len;	246	*enc->cur++ = len;
200	}	247	}
201	else if (len <= 0xffff)	248	else if (len <= 0xffffU)
202	{	249	{
203	*enc->cur++ = major \| 25;	250	*enc->cur++ = major \| LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	251	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	252	*enc->cur++ = len;
206	}	253	}
207	else if (len <= 0xffffffff)	254	else if (len <= 0xffffffffU)
208	{	255	{
209	*enc->cur++ = major \| 26;	256	*enc->cur++ = major \| LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	257	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	258	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	259	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	260	*enc->cur++ = len;
214	}	261	}
215	else	262	else
216	{	263	{
217	*enc->cur++ = major \| 27;	264	*enc->cur++ = major \| LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	265	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	266	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	267	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	268	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	269	*enc->cur++ = len >> 24;
…		…
227	}	274	}
228		275
229	ecb_inline void	276	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	277	encode_tag (enc_t *enc, UV tag)
231	{	278	{
232	encode_uint (enc, 0xc0, tag);	279	encode_uint (enc, MAJOR_TAG, tag);
		280	}
		281
		282	ecb_inline void
		283	encode_str (enc_t enc, int utf8, char str, STRLEN len)
		284	{
		285	if (ecb_expect_false (enc->cbor.flags & F_UTF8_STRINGS))
		286	if (!utf8)
		287	{
		288	SV *sv = sv_newmortal ();
		289	char *s; STRLEN l;
		290
		291	sv_setpvn (sv, str, len);
		292
		293	s = SvPVutf8 (sv, l);
		294	encode_str (enc, 1, s, l);
		295	return;
		296	}
		297
		298	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		299	need (enc, len);
		300	memcpy (enc->cur, str, len);
		301	enc->cur += len;
233	}	302	}
234		303
235	static void	304	static void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	305	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
237	{	306	{
238	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	307	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
239	{	308	{
240	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	309	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
241		310
242	if (SvOK (*svp))	311	if (SvOK (*svp))
243	{	312	{
244	// already registered, use stringref	313	// already registered, use stringref
245	encode_tag (enc, CBOR_TAG_STRINGREF);	314	encode_tag (enc, CBOR_TAG_STRINGREF);
246	encode_uint (enc, 0x00, SvUV (*svp));	315	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
247	return;	316	return;
248	}	317	}
249	else if (len >= minimum_string_length (enc->stringref_idx))	318	else if (len >= minimum_string_length (enc->stringref_idx))
250	{	319	{
251	// register only	320	// register only
252	sv_setuv (*svp, enc->stringref_idx);	321	sv_setuv (*svp, enc->stringref_idx);
253	++enc->stringref_idx;	322	++enc->stringref_idx;
254	}	323	}
255	}	324	}
256		325
257	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	326	encode_str (enc, utf8, str, len);
258	need (enc, len);
259	memcpy (enc->cur, str, len);
260	enc->cur += len;
261	}	327	}
262		328
263	static void encode_sv (enc_t enc, SV sv);	329	static void encode_sv (enc_t enc, SV sv);
264		330
265	static void	331	static void
…		…
270	if (enc->depth >= enc->cbor.max_depth)	336	if (enc->depth >= enc->cbor.max_depth)
271	croak (ERR_NESTING_EXCEEDED);	337	croak (ERR_NESTING_EXCEEDED);
272		338
273	++enc->depth;	339	++enc->depth;
274		340
275	encode_uint (enc, 0x80, len + 1);	341	encode_uint (enc, MAJOR_ARRAY, len + 1);
276		342
		343	if (SvMAGICAL (av))
277	for (i = 0; i <= len; ++i)	344	for (i = 0; i <= len; ++i)
278	{	345	{
279	SV **svp = av_fetch (av, i, 0);	346	SV **svp = av_fetch (av, i, 0);
280	encode_sv (enc, svp ? *svp : &PL_sv_undef);	347	encode_sv (enc, svp ? *svp : &PL_sv_undef);
281	}	348	}
		349	else
		350	for (i = 0; i <= len; ++i)
		351	{
		352	SV *sv = AvARRAY (av)[i];
		353	encode_sv (enc, sv ? sv : &PL_sv_undef);
		354	}
282		355
283	--enc->depth;	356	--enc->depth;
284	}	357	}
285		358
286	static void	359	static void
…		…
295		368
296	int pairs = hv_iterinit (hv);	369	int pairs = hv_iterinit (hv);
297	int mg = SvMAGICAL (hv);	370	int mg = SvMAGICAL (hv);
298		371
299	if (mg)	372	if (mg)
300	encode_ch (enc, 0xa0 \| 31);	373	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
301	else	374	else
302	encode_uint (enc, 0xa0, pairs);	375	encode_uint (enc, MAJOR_MAP, pairs);
303		376
304	while ((he = hv_iternext (hv)))	377	while ((he = hv_iternext (hv)))
305	{	378	{
306	if (HeKLEN (he) == HEf_SVKEY)	379	if (HeKLEN (he) == HEf_SVKEY)
307	encode_sv (enc, HeSVKEY (he));	380	encode_sv (enc, HeSVKEY (he));
308	else	381	else
309	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	382	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
310		383
311	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	384	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
312	}	385	}
313		386
314	if (mg)	387	if (mg)
315	encode_ch (enc, 0xe0 \| 31);	388	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
316		389
317	--enc->depth;	390	--enc->depth;
318	}	391	}
319		392
320	// encode objects, arrays and special \0=false and \1=true values.	393	// encode objects, arrays and special \0=false and \1=true values.
321	static void	394	static void
322	encode_rv (enc_t enc, SV sv)	395	encode_rv (enc_t enc, SV sv)
323	{	396	{
324	SvGETMAGIC (sv);	397	SvGETMAGIC (sv);
325
326	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
327	&& ecb_expect_false (SvREFCNT (sv) > 1))
328	{
329	if (!enc->shareable)
330	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
331
332	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
333
334	if (SvOK (*svp))
335	{
336	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
337	encode_uint (enc, 0x00, SvUV (*svp));
338	return;
339	}
340	else
341	{
342	sv_setuv (*svp, enc->shareable_idx);
343	++enc->shareable_idx;
344	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
345	}
346	}
347		398
348	svtype svt = SvTYPE (sv);	399	svtype svt = SvTYPE (sv);
349		400
350	if (ecb_expect_false (SvOBJECT (sv)))	401	if (ecb_expect_false (SvOBJECT (sv)))
351	{	402	{
…		…
358	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	409	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
359	? cbor_tagged_stash	410	? cbor_tagged_stash
360	: gv_stashpv ("CBOR::XS::Tagged" , 1);	411	: gv_stashpv ("CBOR::XS::Tagged" , 1);
361		412
362	HV *stash = SvSTASH (sv);	413	HV *stash = SvSTASH (sv);
363	GV *method;
364		414
365	if (stash == boolean_stash)	415	if (stash == boolean_stash)
366	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	416	{
		417	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		418	return;
		419	}
367	else if (stash == error_stash)	420	else if (stash == error_stash)
368	encode_ch (enc, 0xe0 \| 23);	421	{
		422	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		423	return;
		424	}
369	else if (stash == tagged_stash)	425	else if (stash == tagged_stash)
370	{	426	{
371	if (svt != SVt_PVAV)	427	if (svt != SVt_PVAV)
372	croak ("encountered CBOR::XS::Tagged object that isn't an array");	428	croak ("encountered CBOR::XS::Tagged object that isn't an array");
373		429
374	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	430	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
375	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	431	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		432
		433	return;
		434	}
		435	}
		436
		437	if (ecb_expect_false (SvREFCNT (sv) > 1)
		438	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		439	{
		440	if (!enc->shareable)
		441	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		442
		443	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		444
		445	if (SvOK (*svp))
376	}	446	{
		447	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		448	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		449	return;
		450	}
		451	else
		452	{
		453	sv_setuv (*svp, enc->shareable_idx);
		454	++enc->shareable_idx;
		455	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		456	}
		457	}
		458
		459	if (ecb_expect_false (SvOBJECT (sv)))
		460	{
		461	HV *stash = SvSTASH (sv);
		462	GV *method;
		463
377	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	464	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
378	{	465	{
379	dSP;	466	dSP;
380		467
381	ENTER; SAVETMPS; PUSHMARK (SP);	468	ENTER; SAVETMPS;
		469	PUSHMARK (SP);
382	// we re-bless the reference to get overload and other niceties right	470	// we re-bless the reference to get overload and other niceties right
383	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	471	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
384		472
385	PUTBACK;	473	PUTBACK;
386	// G_SCALAR ensures that return value is 1	474	// G_SCALAR ensures that return value is 1
…		…
399	}	487	}
400	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	488	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
401	{	489	{
402	dSP;	490	dSP;
403		491
404	ENTER; SAVETMPS; PUSHMARK (SP);	492	ENTER; SAVETMPS;
		493	SAVESTACK_POS ();
		494	PUSHMARK (SP);
405	EXTEND (SP, 2);	495	EXTEND (SP, 2);
406	// we re-bless the reference to get overload and other niceties right	496	// we re-bless the reference to get overload and other niceties right
407	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	497	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
408	PUSHs (sv_cbor);	498	PUSHs (sv_cbor);
409		499
…		…
414	// catch this surprisingly common error	504	// catch this surprisingly common error
415	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	505	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
416	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	506	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
417		507
418	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	508	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
419	encode_uint (enc, 0x80, count + 1);	509	encode_uint (enc, MAJOR_ARRAY, count + 1);
420	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	510	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
421		511
422	while (count)	512	while (count)
423	encode_sv (enc, SP[1 - count--]);	513	encode_sv (enc, SP[1 - count--]);
424		514
425	PUTBACK;	515	PUTBACK;
…		…
446	{	536	{
447	double nv = SvNVX (sv);	537	double nv = SvNVX (sv);
448		538
449	need (enc, 9);	539	need (enc, 9);
450		540
451	if (ecb_expect_false (nv == (U32)nv))	541	if (ecb_expect_false (nv == (NV)(U32)nv))
452	encode_uint (enc, 0x00, (U32)nv);	542	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
453	//TODO: maybe I32?	543	//TODO: maybe I32?
454	else if (ecb_expect_false (nv == (float)nv))	544	else if (ecb_expect_false (nv == (float)nv))
455	{	545	{
456	uint32_t fp = ecb_float_to_binary32 (nv);	546	uint32_t fp = ecb_float_to_binary32 (nv);
457		547
458	*enc->cur++ = 0xe0 \| 26;	548	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
459		549
460	if (!ecb_big_endian ())	550	if (!ecb_big_endian ())
461	fp = ecb_bswap32 (fp);	551	fp = ecb_bswap32 (fp);
462		552
463	memcpy (enc->cur, &fp, 4);	553	memcpy (enc->cur, &fp, 4);
…		…
465	}	555	}
466	else	556	else
467	{	557	{
468	uint64_t fp = ecb_double_to_binary64 (nv);	558	uint64_t fp = ecb_double_to_binary64 (nv);
469		559
470	*enc->cur++ = 0xe0 \| 27;	560	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
471		561
472	if (!ecb_big_endian ())	562	if (!ecb_big_endian ())
473	fp = ecb_bswap64 (fp);	563	fp = ecb_bswap64 (fp);
474		564
475	memcpy (enc->cur, &fp, 8);	565	memcpy (enc->cur, &fp, 8);
…		…
484		574
485	if (SvPOKp (sv))	575	if (SvPOKp (sv))
486	{	576	{
487	STRLEN len;	577	STRLEN len;
488	char *str = SvPV (sv, len);	578	char *str = SvPV (sv, len);
489	encode_str (enc, SvUTF8 (sv), str, len);	579	encode_strref (enc, SvUTF8 (sv), str, len);
490	}	580	}
491	else if (SvNOKp (sv))	581	else if (SvNOKp (sv))
492	encode_nv (enc, sv);	582	encode_nv (enc, sv);
493	else if (SvIOKp (sv))	583	else if (SvIOKp (sv))
494	{	584	{
495	if (SvIsUV (sv))	585	if (SvIsUV (sv))
496	encode_uint (enc, 0x00, SvUVX (sv));	586	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
497	else if (SvIVX (sv) >= 0)	587	else if (SvIVX (sv) >= 0)
498	encode_uint (enc, 0x00, SvIVX (sv));	588	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
499	else	589	else
500	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	590	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
501	}	591	}
502	else if (SvROK (sv))	592	else if (SvROK (sv))
503	encode_rv (enc, SvRV (sv));	593	encode_rv (enc, SvRV (sv));
504	else if (!SvOK (sv))	594	else if (!SvOK (sv))
505	encode_ch (enc, 0xe0 \| 22);	595	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
506	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	596	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
507	encode_ch (enc, 0xe0 \| 23);	597	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
508	else	598	else
509	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	599	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
510	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	600	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
511	}	601	}
512		602
513	static SV *	603	static SV *
514	encode_cbor (SV scalar, CBOR cbor)	604	encode_cbor (SV scalar, CBOR cbor)
515	{	605	{
516	enc_t enc = { };	606	enc_t enc = { 0 };
517		607
518	enc.cbor = *cbor;	608	enc.cbor = *cbor;
519	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	609	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
520	enc.cur = SvPVX (enc.sv);	610	enc.cur = SvPVX (enc.sv);
521	enc.end = SvEND (enc.sv);	611	enc.end = SvEND (enc.sv);
522		612
523	SvPOK_only (enc.sv);	613	SvPOK_only (enc.sv);
524		614
525	if (cbor->flags & F_ALLOW_STRINGREF)	615	if (cbor->flags & F_PACK_STRINGS)
526	{	616	{
527	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	617	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
528	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	618	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
529	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	619	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
530	}	620	}
…		…
565	#define DEC_DEC_DEPTH --dec->depth	655	#define DEC_DEC_DEPTH --dec->depth
566		656
567	static UV	657	static UV
568	decode_uint (dec_t *dec)	658	decode_uint (dec_t *dec)
569	{	659	{
570	switch (*dec->cur & 31)	660	U8 m = *dec->cur & MINOR_MASK;
571	{	661	++dec->cur;
572	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
573	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
574	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
575	return *dec->cur++ & 31;
576		662
577	case 24:	663	if (ecb_expect_true (m < LENGTH_EXT1))
		664	return m;
		665	else if (ecb_expect_true (m == LENGTH_EXT1))
		666	{
578	WANT (2);	667	WANT (1);
579	dec->cur += 2;	668	dec->cur += 1;
580	return dec->cur[-1];	669	return dec->cur[-1];
581		670	}
582	case 25:	671	else if (ecb_expect_true (m == LENGTH_EXT2))
		672	{
583	WANT (3);	673	WANT (2);
584	dec->cur += 3;	674	dec->cur += 2;
585	return (((UV)dec->cur[-2]) << 8)	675	return (((UV)dec->cur[-2]) << 8)
586	\| ((UV)dec->cur[-1]);	676	\| ((UV)dec->cur[-1]);
587		677	}
588	case 26:	678	else if (ecb_expect_true (m == LENGTH_EXT4))
		679	{
589	WANT (5);	680	WANT (4);
590	dec->cur += 5;	681	dec->cur += 4;
591	return (((UV)dec->cur[-4]) << 24)	682	return (((UV)dec->cur[-4]) << 24)
592	\| (((UV)dec->cur[-3]) << 16)	683	\| (((UV)dec->cur[-3]) << 16)
593	\| (((UV)dec->cur[-2]) << 8)	684	\| (((UV)dec->cur[-2]) << 8)
594	\| ((UV)dec->cur[-1]);	685	\| ((UV)dec->cur[-1]);
595		686	}
596	case 27:	687	else if (ecb_expect_true (m == LENGTH_EXT8))
		688	{
597	WANT (9);	689	WANT (8);
598	dec->cur += 9;	690	dec->cur += 8;
		691
		692	return
		693	#if UVSIZE < 8
		694	0
		695	#else
599	return (((UV)dec->cur[-8]) << 56)	696	(((UV)dec->cur[-8]) << 56)
600	\| (((UV)dec->cur[-7]) << 48)	697	\| (((UV)dec->cur[-7]) << 48)
601	\| (((UV)dec->cur[-6]) << 40)	698	\| (((UV)dec->cur[-6]) << 40)
602	\| (((UV)dec->cur[-5]) << 32)	699	\| (((UV)dec->cur[-5]) << 32)
		700	#endif
603	\| (((UV)dec->cur[-4]) << 24)	701	\| (((UV)dec->cur[-4]) << 24)
604	\| (((UV)dec->cur[-3]) << 16)	702	\| (((UV)dec->cur[-3]) << 16)
605	\| (((UV)dec->cur[-2]) << 8)	703	\| (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	704	\| ((UV)dec->cur[-1]);
607		705	}
608	default:	706	else
609	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	707	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
610	}
611		708
612	fail:	709	fail:
613	return 0;	710	return 0;
614	}	711	}
615		712
…		…
620	{	717	{
621	AV *av = newAV ();	718	AV *av = newAV ();
622		719
623	DEC_INC_DEPTH;	720	DEC_INC_DEPTH;
624		721
625	if ((*dec->cur & 31) == 31)	722	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
626	{	723	{
627	++dec->cur;	724	++dec->cur;
628		725
629	for (;;)	726	for (;;)
630	{	727	{
631	WANT (1);	728	WANT (1);
632		729
633	if (*dec->cur == (0xe0 \| 31))	730	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
634	{	731	{
635	++dec->cur;	732	++dec->cur;
636	break;	733	break;
637	}	734	}
638		735
…		…
641	}	738	}
642	else	739	else
643	{	740	{
644	int i, len = decode_uint (dec);	741	int i, len = decode_uint (dec);
645		742
		743	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
646	av_fill (av, len - 1);	744	av_fill (av, len - 1);
647		745
648	for (i = 0; i < len; ++i)	746	for (i = 0; i < len; ++i)
649	AvARRAY (av)[i] = decode_sv (dec);	747	AvARRAY (av)[i] = decode_sv (dec);
650	}	748	}
…		…
663	{	761	{
664	// for speed reasons, we specialcase single-string	762	// for speed reasons, we specialcase single-string
665	// byte or utf-8 strings as keys, but only when !stringref	763	// byte or utf-8 strings as keys, but only when !stringref
666		764
667	if (ecb_expect_true (!dec->stringref))	765	if (ecb_expect_true (!dec->stringref))
668	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	766	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
669	{	767	{
670	I32 len = decode_uint (dec);	768	I32 len = decode_uint (dec);
671	char key = (char )dec->cur;	769	char key = (char )dec->cur;
672		770
		771	WANT (len);
673	dec->cur += len;	772	dec->cur += len;
674		773
675	if (ecb_expect_false (dec->stringref))
676	av_push (dec->stringref, newSVpvn (key, len));
677
678	hv_store (hv, key, len, decode_sv (dec), 0);	774	hv_store (hv, key, len, decode_sv (dec), 0);
679		775
680	return;	776	return;
681	}	777	}
682	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	778	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
683	{	779	{
684	I32 len = decode_uint (dec);	780	I32 len = decode_uint (dec);
685	char key = (char )dec->cur;	781	char key = (char )dec->cur;
686		782
		783	WANT (len);
687	dec->cur += len;	784	dec->cur += len;
688		785
689	if (ecb_expect_false (dec->stringref))	786	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
690	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	787	if (!is_utf8_string (key, len))
		788	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
691		789
692	hv_store (hv, key, -len, decode_sv (dec), 0);	790	hv_store (hv, key, -len, decode_sv (dec), 0);
693		791
694	return;	792	return;
695	}	793	}
…		…
697	SV *k = decode_sv (dec);	795	SV *k = decode_sv (dec);
698	SV *v = decode_sv (dec);	796	SV *v = decode_sv (dec);
699		797
700	hv_store_ent (hv, k, v, 0);	798	hv_store_ent (hv, k, v, 0);
701	SvREFCNT_dec (k);	799	SvREFCNT_dec (k);
		800
		801	fail:
		802	;
702	}	803	}
703		804
704	static SV *	805	static SV *
705	decode_hv (dec_t *dec)	806	decode_hv (dec_t *dec)
706	{	807	{
707	HV *hv = newHV ();	808	HV *hv = newHV ();
708		809
709	DEC_INC_DEPTH;	810	DEC_INC_DEPTH;
710		811
711	if ((*dec->cur & 31) == 31)	812	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
712	{	813	{
713	++dec->cur;	814	++dec->cur;
714		815
715	for (;;)	816	for (;;)
716	{	817	{
717	WANT (1);	818	WANT (1);
718		819
719	if (*dec->cur == (0xe0 \| 31))	820	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
720	{	821	{
721	++dec->cur;	822	++dec->cur;
722	break;	823	break;
723	}	824	}
724		825
…		…
745	static SV *	846	static SV *
746	decode_str (dec_t *dec, int utf8)	847	decode_str (dec_t *dec, int utf8)
747	{	848	{
748	SV *sv = 0;	849	SV *sv = 0;
749		850
750	if ((*dec->cur & 31) == 31)	851	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
751	{	852	{
		853	// indefinite length strings
752	++dec->cur;	854	++dec->cur;
753		855
		856	U8 major = *dec->cur & MAJOR_MISC;
		857
754	sv = newSVpvn ("", 0);	858	sv = newSVpvn ("", 0);
755		859
756	// not very fast, and certainly not robust against illegal input
757	for (;;)	860	for (;;)
758	{	861	{
759	WANT (1);	862	WANT (1);
760		863
761	if (*dec->cur == (0xe0 \| 31))	864	if ((*dec->cur - major) > LENGTH_EXT8)
		865	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
762	{	866	{
763	++dec->cur;	867	++dec->cur;
764	break;	868	break;
765	}	869	}
		870	else
		871	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
766		872
767	sv_catsv (sv, decode_sv (dec));	873	STRLEN len = decode_uint (dec);
		874
		875	WANT (len);
		876	sv_catpvn (sv, dec->cur, len);
		877	dec->cur += len;
768	}	878	}
769	}	879	}
770	else	880	else
771	{	881	{
772	STRLEN len = decode_uint (dec);	882	STRLEN len = decode_uint (dec);
…		…
779	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	889	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
780	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	890	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
781	}	891	}
782		892
783	if (utf8)	893	if (utf8)
		894	{
		895	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		896	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		897	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		898
784	SvUTF8_on (sv);	899	SvUTF8_on (sv);
		900	}
785		901
786	return sv;	902	return sv;
787		903
788	fail:	904	fail:
789	SvREFCNT_dec (sv);	905	SvREFCNT_dec (sv);
…		…
821	}	937	}
822	break;	938	break;
823		939
824	case CBOR_TAG_STRINGREF:	940	case CBOR_TAG_STRINGREF:
825	{	941	{
826	if ((*dec->cur >> 5) != 0)	942	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
827	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	943	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
828		944
829	UV idx = decode_uint (dec);	945	UV idx = decode_uint (dec);
830		946
831	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	947	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
838	case CBOR_TAG_VALUE_SHAREABLE:	954	case CBOR_TAG_VALUE_SHAREABLE:
839	{	955	{
840	if (ecb_expect_false (!dec->shareable))	956	if (ecb_expect_false (!dec->shareable))
841	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	957	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
842		958
		959	if (dec->cbor.flags & F_ALLOW_CYCLES)
		960	{
843	sv = newSV (0);	961	sv = newSV (0);
844	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	962	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
845		963
846	SV *osv = decode_sv (dec);	964	SV *osv = decode_sv (dec);
847	sv_setsv (sv, osv);	965	sv_setsv (sv, osv);
848	SvREFCNT_dec_NN (osv);	966	SvREFCNT_dec_NN (osv);
		967	}
		968	else
		969	{
		970	av_push (dec->shareable, &PL_sv_undef);
		971	int idx = AvFILLp (dec->shareable);
		972	sv = decode_sv (dec);
		973	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		974	}
849	}	975	}
850	break;	976	break;
851		977
852	case CBOR_TAG_VALUE_SHAREDREF:	978	case CBOR_TAG_VALUE_SHAREDREF:
853	{	979	{
854	if ((*dec->cur >> 5) != 0)	980	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
855	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	981	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
856		982
857	UV idx = decode_uint (dec);	983	UV idx = decode_uint (dec);
858		984
859	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	985	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
860	ERR ("corrupted CBOR data (sharedref index out of bounds)");	986	ERR ("corrupted CBOR data (sharedref index out of bounds)");
861		987
862	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	988	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		989
		990	if (sv == &PL_sv_undef)
		991	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
863	}	992	}
864	break;	993	break;
865		994
866	case CBOR_TAG_PERL_OBJECT:	995	case CBOR_TAG_PERL_OBJECT:
867	{	996	{
…		…
882	if (!method)	1011	if (!method)
883	ERR ("cannot decode perl-object (package does not have a THAW method)");	1012	ERR ("cannot decode perl-object (package does not have a THAW method)");
884		1013
885	dSP;	1014	dSP;
886		1015
887	ENTER; SAVETMPS; PUSHMARK (SP);	1016	ENTER; SAVETMPS;
		1017	PUSHMARK (SP);
888	EXTEND (SP, len + 1);	1018	EXTEND (SP, len + 1);
889	// we re-bless the reference to get overload and other niceties right	1019	// we re-bless the reference to get overload and other niceties right
890	PUSHs (*av_fetch (av, 0, 1));	1020	PUSHs (*av_fetch (av, 0, 1));
891	PUSHs (sv_cbor);	1021	PUSHs (sv_cbor);
892		1022
…		…
917	default:	1047	default:
918	{	1048	{
919	sv = decode_sv (dec);	1049	sv = decode_sv (dec);
920		1050
921	dSP;	1051	dSP;
922	ENTER; SAVETMPS; PUSHMARK (SP);	1052	ENTER; SAVETMPS;
		1053	SAVESTACK_POS ();
		1054	PUSHMARK (SP);
923	EXTEND (SP, 2);	1055	EXTEND (SP, 2);
924	PUSHs (newSVuv (tag));	1056	PUSHs (newSVuv (tag));
925	PUSHs (sv);	1057	PUSHs (sv);
926		1058
927	PUTBACK;	1059	PUTBACK;
…		…
968	static SV *	1100	static SV *
969	decode_sv (dec_t *dec)	1101	decode_sv (dec_t *dec)
970	{	1102	{
971	WANT (1);	1103	WANT (1);
972		1104
973	switch (*dec->cur >> 5)	1105	switch (*dec->cur >> MAJOR_SHIFT)
974	{	1106	{
975	case 0: // unsigned int	1107	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
976	return newSVuv (decode_uint (dec));	1108	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
977	case 1: // negative int	1109	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
978	return newSViv (-1 - (IV)decode_uint (dec));	1110	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
979	case 2: // octet string	1111	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
980	return decode_str (dec, 0);	1112	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
981	case 3: // utf-8 string	1113	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
982	return decode_str (dec, 1);	1114
983	case 4: // array	1115	case MAJOR_MISC >> MAJOR_SHIFT:
984	return decode_av (dec);
985	case 5: // map
986	return decode_hv (dec);
987	case 6: // tag
988	return decode_tagged (dec);
989	case 7: // misc
990	switch (*dec->cur++ & 31)	1116	switch (*dec->cur++ & MINOR_MASK)
991	{	1117	{
992	case 20:	1118	case SIMPLE_FALSE:
993	#if CBOR_SLOW	1119	#if CBOR_SLOW
994	types_false = get_bool ("Types::Serialiser::false");	1120	types_false = get_bool ("Types::Serialiser::false");
995	#endif	1121	#endif
996	return newSVsv (types_false);	1122	return newSVsv (types_false);
997	case 21:	1123	case SIMPLE_TRUE:
998	#if CBOR_SLOW	1124	#if CBOR_SLOW
999	types_true = get_bool ("Types::Serialiser::true");	1125	types_true = get_bool ("Types::Serialiser::true");
1000	#endif	1126	#endif
1001	return newSVsv (types_true);	1127	return newSVsv (types_true);
1002	case 22:	1128	case SIMPLE_NULL:
1003	return newSVsv (&PL_sv_undef);	1129	return newSVsv (&PL_sv_undef);
1004	case 23:	1130	case SIMPLE_UNDEF:
1005	#if CBOR_SLOW	1131	#if CBOR_SLOW
1006	types_error = get_bool ("Types::Serialiser::error");	1132	types_error = get_bool ("Types::Serialiser::error");
1007	#endif	1133	#endif
1008	return newSVsv (types_error);	1134	return newSVsv (types_error);
1009		1135
1010	case 25:	1136	case MISC_FLOAT16:
1011	{	1137	{
1012	WANT (2);	1138	WANT (2);
1013		1139
1014	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1140	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1015	dec->cur += 2;	1141	dec->cur += 2;
1016		1142
1017	return newSVnv (ecb_binary16_to_float (fp));	1143	return newSVnv (ecb_binary16_to_float (fp));
1018	}	1144	}
1019		1145
1020	case 26:	1146	case MISC_FLOAT32:
1021	{	1147	{
1022	uint32_t fp;	1148	uint32_t fp;
1023	WANT (4);	1149	WANT (4);
1024	memcpy (&fp, dec->cur, 4);	1150	memcpy (&fp, dec->cur, 4);
1025	dec->cur += 4;	1151	dec->cur += 4;
…		…
1028	fp = ecb_bswap32 (fp);	1154	fp = ecb_bswap32 (fp);
1029		1155
1030	return newSVnv (ecb_binary32_to_float (fp));	1156	return newSVnv (ecb_binary32_to_float (fp));
1031	}	1157	}
1032		1158
1033	case 27:	1159	case MISC_FLOAT64:
1034	{	1160	{
1035	uint64_t fp;	1161	uint64_t fp;
1036	WANT (8);	1162	WANT (8);
1037	memcpy (&fp, dec->cur, 8);	1163	memcpy (&fp, dec->cur, 8);
1038	dec->cur += 8;	1164	dec->cur += 8;
…		…
1041	fp = ecb_bswap64 (fp);	1167	fp = ecb_bswap64 (fp);
1042		1168
1043	return newSVnv (ecb_binary64_to_double (fp));	1169	return newSVnv (ecb_binary64_to_double (fp));
1044	}	1170	}
1045		1171
1046	// 0..19 unassigned	1172	// 0..19 unassigned simple
1047	// 24 reserved + unassigned (reserved values are not encodable)	1173	// 24 reserved + unassigned simple (reserved values are not encodable)
		1174	// 28-30 unassigned misc
		1175	// 31 break code
1048	default:	1176	default:
1049	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1177	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1050	}	1178	}
1051		1179
1052	break;	1180	break;
1053	}	1181	}
1054		1182
…		…
1057	}	1185	}
1058		1186
1059	static SV *	1187	static SV *
1060	decode_cbor (SV string, CBOR cbor, char **offset_return)	1188	decode_cbor (SV string, CBOR cbor, char **offset_return)
1061	{	1189	{
1062	dec_t dec = { };	1190	dec_t dec = { 0 };
1063	SV *sv;	1191	SV *sv;
1064	STRLEN len;	1192	STRLEN len;
1065	char *data = SvPVbyte (string, len);	1193	char *data = SvPVbyte (string, len);
1066		1194
1067	if (len > cbor->max_size && cbor->max_size)	1195	if (len > cbor->max_size && cbor->max_size)
…		…
1081	if (dec.cur != dec.end && !dec.err)	1209	if (dec.cur != dec.end && !dec.err)
1082	dec.err = "garbage after CBOR object";	1210	dec.err = "garbage after CBOR object";
1083		1211
1084	if (dec.err)	1212	if (dec.err)
1085	{	1213	{
		1214	if (dec.shareable)
		1215	{
		1216	// need to break cyclic links, which whould all be in shareable
		1217	int i;
		1218	SV **svp;
		1219
		1220	for (i = av_len (dec.shareable) + 1; i--; )
		1221	if ((svp = av_fetch (dec.shareable, i, 0)))
		1222	sv_setsv (*svp, &PL_sv_undef);
		1223	}
		1224
1086	SvREFCNT_dec (sv);	1225	SvREFCNT_dec (sv);
1087	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1226	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1088	}	1227	}
1089		1228
1090	sv = sv_2mortal (sv);	1229	sv = sv_2mortal (sv);
1091		1230
1092	return sv;	1231	return sv;
1093	}	1232	}
1094		1233
		1234	/////////////////////////////////////////////////////////////////////////////
		1235	// incremental parser
		1236
		1237	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1238
		1239	// returns 0 for notyet, 1 for success or error
		1240	static int
		1241	incr_parse (CBOR self, SV cborstr)
		1242	{
		1243	STRLEN cur;
		1244	SvPV (cborstr, cur);
		1245
		1246	while (ecb_expect_true (self->incr_need <= cur))
		1247	{
		1248	// table of integer count bytes
		1249	static I8 incr_len[MINOR_MASK + 1] = {
		1250	0, 0, 0, 0, 0, 0, 0, 0,
		1251	0, 0, 0, 0, 0, 0, 0, 0,
		1252	0, 0, 0, 0, 0, 0, 0, 0,
		1253	1, 2, 4, 8,-1,-1,-1,-2
		1254	};
		1255
		1256	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1257	U8 m = *p & MINOR_MASK;
		1258	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1259	I8 ilen = incr_len[m];
		1260
		1261	self->incr_need = self->incr_pos + 1;
		1262
		1263	if (ecb_expect_false (ilen < 0))
		1264	{
		1265	if (m != MINOR_INDEF)
		1266	return 1; // error
		1267
		1268	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1269	{
		1270	if (count >= 0)
		1271	return 1; // error
		1272
		1273	count = 1;
		1274	}
		1275	else
		1276	{
		1277	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1278	count = -1;
		1279	}
		1280	}
		1281	else
		1282	{
		1283	self->incr_need += ilen;
		1284	if (ecb_expect_false (self->incr_need > cur))
		1285	return 0;
		1286
		1287	int major = *p >> MAJOR_SHIFT;
		1288
		1289	switch (major)
		1290	{
		1291	case MAJOR_TAG >> MAJOR_SHIFT:
		1292	++count; // tags merely prefix another value
		1293	break;
		1294
		1295	case MAJOR_BYTES >> MAJOR_SHIFT:
		1296	case MAJOR_TEXT >> MAJOR_SHIFT:
		1297	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1298	case MAJOR_MAP >> MAJOR_SHIFT:
		1299	{
		1300	UV len;
		1301
		1302	if (ecb_expect_false (ilen))
		1303	{
		1304	len = 0;
		1305
		1306	do {
		1307	len = (len << 8) \| *++p;
		1308	} while (--ilen);
		1309	}
		1310	else
		1311	len = m;
		1312
		1313	switch (major)
		1314	{
		1315	case MAJOR_BYTES >> MAJOR_SHIFT:
		1316	case MAJOR_TEXT >> MAJOR_SHIFT:
		1317	self->incr_need += len;
		1318	if (ecb_expect_false (self->incr_need > cur))
		1319	return 0;
		1320
		1321	break;
		1322
		1323	case MAJOR_MAP >> MAJOR_SHIFT:
		1324	len <<= 1;
		1325	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1326	if (len)
		1327	{
		1328	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1329	count = len + 1;
		1330	}
		1331	break;
		1332	}
		1333	}
		1334	}
		1335	}
		1336
		1337	self->incr_pos = self->incr_need;
		1338
		1339	if (count > 0)
		1340	{
		1341	while (!--count)
		1342	{
		1343	if (!AvFILLp (self->incr_count))
		1344	return 1; // done
		1345
		1346	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1347	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1348	}
		1349
		1350	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1351	}
		1352	}
		1353
		1354	return 0;
		1355	}
		1356
		1357
1095	/////////////////////////////////////////////////////////////////////////////	1358	/////////////////////////////////////////////////////////////////////////////
1096	// XS interface functions	1359	// XS interface functions
1097		1360
1098	MODULE = CBOR::XS PACKAGE = CBOR::XS	1361	MODULE = CBOR::XS PACKAGE = CBOR::XS
1099		1362
…		…
1139	void shrink (CBOR *self, int enable = 1)	1402	void shrink (CBOR *self, int enable = 1)
1140	ALIAS:	1403	ALIAS:
1141	shrink = F_SHRINK	1404	shrink = F_SHRINK
1142	allow_unknown = F_ALLOW_UNKNOWN	1405	allow_unknown = F_ALLOW_UNKNOWN
1143	allow_sharing = F_ALLOW_SHARING	1406	allow_sharing = F_ALLOW_SHARING
1144	allow_stringref = F_ALLOW_STRINGREF	1407	allow_cycles = F_ALLOW_CYCLES
		1408	pack_strings = F_PACK_STRINGS
		1409	utf8_strings = F_UTF8_STRINGS
		1410	validate_utf8 = F_VALIDATE_UTF8
1145	PPCODE:	1411	PPCODE:
1146	{	1412	{
1147	if (enable)	1413	if (enable)
1148	self->flags \|= ix;	1414	self->flags \|= ix;
1149	else	1415	else
…		…
1155	void get_shrink (CBOR *self)	1421	void get_shrink (CBOR *self)
1156	ALIAS:	1422	ALIAS:
1157	get_shrink = F_SHRINK	1423	get_shrink = F_SHRINK
1158	get_allow_unknown = F_ALLOW_UNKNOWN	1424	get_allow_unknown = F_ALLOW_UNKNOWN
1159	get_allow_sharing = F_ALLOW_SHARING	1425	get_allow_sharing = F_ALLOW_SHARING
1160	get_allow_stringref = F_ALLOW_STRINGREF	1426	get_allow_cycles = F_ALLOW_CYCLES
		1427	get_pack_strings = F_PACK_STRINGS
		1428	get_validate_utf8 = F_VALIDATE_UTF8
1161	PPCODE:	1429	PPCODE:
1162	XPUSHs (boolSV (self->flags & ix));	1430	XPUSHs (boolSV (self->flags & ix));
1163		1431
1164	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1432	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1165	PPCODE:	1433	PPCODE:
…		…
1214	EXTEND (SP, 2);	1482	EXTEND (SP, 2);
1215	PUSHs (sv);	1483	PUSHs (sv);
1216	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1484	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1217	}	1485	}
1218		1486
		1487	void incr_parse (CBOR self, SV cborstr)
		1488	ALIAS:
		1489	incr_parse_multiple = 1
		1490	PPCODE:
		1491	{
		1492	if (SvUTF8 (cborstr))
		1493	sv_utf8_downgrade (cborstr, 0);
		1494
		1495	if (!self->incr_count)
		1496	{
		1497	self->incr_count = newAV ();
		1498	self->incr_pos = 0;
		1499	self->incr_need = 1;
		1500
		1501	av_push (self->incr_count, newSViv (1));
		1502	}
		1503
		1504	do
		1505	{
		1506	if (!incr_parse (self, cborstr))
		1507	{
		1508	if (self->incr_need > self->max_size && self->max_size)
		1509	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1510	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1511
		1512	break;
		1513	}
		1514
		1515	SV *sv;
		1516	char *offset;
		1517
		1518	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1519	XPUSHs (sv);
		1520
		1521	sv_chop (cborstr, offset);
		1522
		1523	av_clear (self->incr_count);
		1524	av_push (self->incr_count, newSViv (1));
		1525
		1526	self->incr_pos = 0;
		1527	self->incr_need = self->incr_pos + 1;
		1528	}
		1529	while (ix);
		1530	}
		1531
		1532	void incr_reset (CBOR *self)
		1533	CODE:
		1534	{
		1535	SvREFCNT_dec (self->incr_count);
		1536	self->incr_count = 0;
		1537	}
		1538
1219	void DESTROY (CBOR *self)	1539	void DESTROY (CBOR *self)
1220	PPCODE:	1540	PPCODE:
1221	cbor_free (self);	1541	cbor_free (self);
1222		1542
1223	PROTOTYPES: ENABLE	1543	PROTOTYPES: ENABLE
1224		1544
1225	void encode_cbor (SV *scalar)	1545	void encode_cbor (SV *scalar)
		1546	ALIAS:
		1547	encode_cbor = 0
		1548	encode_cbor_sharing = F_ALLOW_SHARING
1226	PPCODE:	1549	PPCODE:
1227	{	1550	{
1228	CBOR cbor;	1551	CBOR cbor;
1229	cbor_init (&cbor);	1552	cbor_init (&cbor);
		1553	cbor.flags \|= ix;
1230	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1554	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1231	XPUSHs (scalar);	1555	XPUSHs (scalar);
1232	}	1556	}
1233		1557
1234	void decode_cbor (SV *cborstr)	1558	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.28 by root, Sat Nov 23 18:30:59 2013 UTC vs. Revision 1.51 by root, Sun Apr 24 13:15:19 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.28 by root, Sat Nov 23 18:30:59 2013 UTC vs.
Revision 1.51 by root, Sun Apr 24 13:15:19 2016 UTC