[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 1.50 by root, Thu Feb 25 02:29:22 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_VALIDATE_UTF8 0x00000020UL
66		106
67	#define INIT_SIZE 32 // initial scalar size to be allocated	107	#define INIT_SIZE 32 // initial scalar size to be allocated
68		108
69	#define SB do {	109	#define SB do {
70	#define SE } while (0)	110	#define SE } while (0)
…		…
89	typedef struct {	129	typedef struct {
90	U32 flags;	130	U32 flags;
91	U32 max_depth;	131	U32 max_depth;
92	STRLEN max_size;	132	STRLEN max_size;
93	SV *filter;	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.
94	} CBOR;	139	} CBOR;
95		140
96	ecb_inline void	141	ecb_inline void
97	cbor_init (CBOR *cbor)	142	cbor_init (CBOR *cbor)
98	{	143	{
…		…
102		147
103	ecb_inline void	148	ecb_inline void
104	cbor_free (CBOR *cbor)	149	cbor_free (CBOR *cbor)
105	{	150	{
106	SvREFCNT_dec (cbor->filter);	151	SvREFCNT_dec (cbor->filter);
		152	SvREFCNT_dec (cbor->incr_count);
107	}	153	}
108		154
109	/////////////////////////////////////////////////////////////////////////////	155	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	156	// utility functions
111		157
…		…
141	{	187	{
142	return idx > 23	188	return idx > 23
143	? idx > 0xffU	189	? idx > 0xffU
144	? idx > 0xffffU	190	? idx > 0xffffU
145	? idx > 0xffffffffU	191	? idx > 0xffffffffU
		192	? 11
146	? 7	193	: 7
147	: 6
148	: 5	194	: 5
149	: 4	195	: 4
150	: 3;	196	: 3;
151	}	197	}
152		198
…		…
189	static void	235	static void
190	encode_uint (enc_t *enc, int major, UV len)	236	encode_uint (enc_t *enc, int major, UV len)
191	{	237	{
192	need (enc, 9);	238	need (enc, 9);
193		239
194	if (len < 24)	240	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major \| len;	241	*enc->cur++ = major \| len;
196	else if (len <= 0xff)	242	else if (ecb_expect_true (len <= 0xffU))
197	{	243	{
198	*enc->cur++ = major \| 24;	244	*enc->cur++ = major \| LENGTH_EXT1;
199	*enc->cur++ = len;	245	*enc->cur++ = len;
200	}	246	}
201	else if (len <= 0xffff)	247	else if (len <= 0xffffU)
202	{	248	{
203	*enc->cur++ = major \| 25;	249	*enc->cur++ = major \| LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	250	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	251	*enc->cur++ = len;
206	}	252	}
207	else if (len <= 0xffffffff)	253	else if (len <= 0xffffffffU)
208	{	254	{
209	*enc->cur++ = major \| 26;	255	*enc->cur++ = major \| LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	256	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	257	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	258	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	259	*enc->cur++ = len;
214	}	260	}
215	else	261	else
216	{	262	{
217	*enc->cur++ = major \| 27;	263	*enc->cur++ = major \| LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	264	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	265	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	266	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	267	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	268	*enc->cur++ = len >> 24;
…		…
227	}	273	}
228		274
229	ecb_inline void	275	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	276	encode_tag (enc_t *enc, UV tag)
231	{	277	{
232	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;
233	}	288	}
234		289
235	static void	290	static void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	291	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
237	{	292	{
238	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	293	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
239	{	294	{
240	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	295	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
241		296
242	if (SvOK (*svp))	297	if (SvOK (*svp))
243	{	298	{
244	// already registered, use stringref	299	// already registered, use stringref
245	encode_tag (enc, CBOR_TAG_STRINGREF);	300	encode_tag (enc, CBOR_TAG_STRINGREF);
246	encode_uint (enc, 0x00, SvUV (*svp));	301	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
247	return;	302	return;
248	}	303	}
249	else if (len >= minimum_string_length (enc->stringref_idx))	304	else if (len >= minimum_string_length (enc->stringref_idx))
250	{	305	{
251	// register only	306	// register only
252	sv_setuv (*svp, enc->stringref_idx);	307	sv_setuv (*svp, enc->stringref_idx);
253	++enc->stringref_idx;	308	++enc->stringref_idx;
254	}	309	}
255	}	310	}
256		311
257	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	312	encode_str (enc, utf8, str, len);
258	need (enc, len);
259	memcpy (enc->cur, str, len);
260	enc->cur += len;
261	}	313	}
262		314
263	static void encode_sv (enc_t enc, SV sv);	315	static void encode_sv (enc_t enc, SV sv);
264		316
265	static void	317	static void
…		…
270	if (enc->depth >= enc->cbor.max_depth)	322	if (enc->depth >= enc->cbor.max_depth)
271	croak (ERR_NESTING_EXCEEDED);	323	croak (ERR_NESTING_EXCEEDED);
272		324
273	++enc->depth;	325	++enc->depth;
274		326
275	encode_uint (enc, 0x80, len + 1);	327	encode_uint (enc, MAJOR_ARRAY, len + 1);
276		328
		329	if (SvMAGICAL (av))
277	for (i = 0; i <= len; ++i)	330	for (i = 0; i <= len; ++i)
278	{	331	{
279	SV **svp = av_fetch (av, i, 0);	332	SV **svp = av_fetch (av, i, 0);
280	encode_sv (enc, svp ? *svp : &PL_sv_undef);	333	encode_sv (enc, svp ? *svp : &PL_sv_undef);
281	}	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	}
282		341
283	--enc->depth;	342	--enc->depth;
284	}	343	}
285		344
286	static void	345	static void
…		…
295		354
296	int pairs = hv_iterinit (hv);	355	int pairs = hv_iterinit (hv);
297	int mg = SvMAGICAL (hv);	356	int mg = SvMAGICAL (hv);
298		357
299	if (mg)	358	if (mg)
300	encode_ch (enc, 0xa0 \| 31);	359	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
301	else	360	else
302	encode_uint (enc, 0xa0, pairs);	361	encode_uint (enc, MAJOR_MAP, pairs);
303		362
304	while ((he = hv_iternext (hv)))	363	while ((he = hv_iternext (hv)))
305	{	364	{
306	if (HeKLEN (he) == HEf_SVKEY)	365	if (HeKLEN (he) == HEf_SVKEY)
307	encode_sv (enc, HeSVKEY (he));	366	encode_sv (enc, HeSVKEY (he));
308	else	367	else
309	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	368	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
310		369
311	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));
312	}	371	}
313		372
314	if (mg)	373	if (mg)
315	encode_ch (enc, 0xe0 \| 31);	374	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
316		375
317	--enc->depth;	376	--enc->depth;
318	}	377	}
319		378
320	// encode objects, arrays and special \0=false and \1=true values.	379	// encode objects, arrays and special \0=false and \1=true values.
321	static void	380	static void
322	encode_rv (enc_t enc, SV sv)	381	encode_rv (enc_t enc, SV sv)
323	{	382	{
324	SvGETMAGIC (sv);	383	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		384
348	svtype svt = SvTYPE (sv);	385	svtype svt = SvTYPE (sv);
349		386
350	if (ecb_expect_false (SvOBJECT (sv)))	387	if (ecb_expect_false (SvOBJECT (sv)))
351	{	388	{
…		…
358	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	395	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
359	? cbor_tagged_stash	396	? cbor_tagged_stash
360	: gv_stashpv ("CBOR::XS::Tagged" , 1);	397	: gv_stashpv ("CBOR::XS::Tagged" , 1);
361		398
362	HV *stash = SvSTASH (sv);	399	HV *stash = SvSTASH (sv);
363	GV *method;
364		400
365	if (stash == boolean_stash)	401	if (stash == boolean_stash)
366	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	}
367	else if (stash == error_stash)	406	else if (stash == error_stash)
368	encode_ch (enc, 0xe0 \| 23);	407	{
		408	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		409	return;
		410	}
369	else if (stash == tagged_stash)	411	else if (stash == tagged_stash)
370	{	412	{
371	if (svt != SVt_PVAV)	413	if (svt != SVt_PVAV)
372	croak ("encountered CBOR::XS::Tagged object that isn't an array");	414	croak ("encountered CBOR::XS::Tagged object that isn't an array");
373		415
374	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	416	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
375	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))
376	}	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
377	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	450	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
378	{	451	{
379	dSP;	452	dSP;
380		453
381	ENTER; SAVETMPS; PUSHMARK (SP);	454	ENTER; SAVETMPS;
		455	PUSHMARK (SP);
382	// we re-bless the reference to get overload and other niceties right	456	// we re-bless the reference to get overload and other niceties right
383	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	457	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
384		458
385	PUTBACK;	459	PUTBACK;
386	// G_SCALAR ensures that return value is 1	460	// G_SCALAR ensures that return value is 1
…		…
399	}	473	}
400	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	474	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
401	{	475	{
402	dSP;	476	dSP;
403		477
404	ENTER; SAVETMPS; PUSHMARK (SP);	478	ENTER; SAVETMPS;
		479	SAVESTACK_POS ();
		480	PUSHMARK (SP);
405	EXTEND (SP, 2);	481	EXTEND (SP, 2);
406	// we re-bless the reference to get overload and other niceties right	482	// we re-bless the reference to get overload and other niceties right
407	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
408	PUSHs (sv_cbor);	484	PUSHs (sv_cbor);
409		485
…		…
414	// catch this surprisingly common error	490	// catch this surprisingly common error
415	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	491	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));	492	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
417		493
418	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	494	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
419	encode_uint (enc, 0x80, count + 1);	495	encode_uint (enc, MAJOR_ARRAY, count + 1);
420	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	496	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
421		497
422	while (count)	498	while (count)
423	encode_sv (enc, SP[1 - count--]);	499	encode_sv (enc, SP[1 - count--]);
424		500
425	PUTBACK;	501	PUTBACK;
…		…
446	{	522	{
447	double nv = SvNVX (sv);	523	double nv = SvNVX (sv);
448		524
449	need (enc, 9);	525	need (enc, 9);
450		526
451	if (ecb_expect_false (nv == (U32)nv))	527	if (ecb_expect_false (nv == (NV)(U32)nv))
452	encode_uint (enc, 0x00, (U32)nv);	528	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
453	//TODO: maybe I32?	529	//TODO: maybe I32?
454	else if (ecb_expect_false (nv == (float)nv))	530	else if (ecb_expect_false (nv == (float)nv))
455	{	531	{
456	uint32_t fp = ecb_float_to_binary32 (nv);	532	uint32_t fp = ecb_float_to_binary32 (nv);
457		533
458	*enc->cur++ = 0xe0 \| 26;	534	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
459		535
460	if (!ecb_big_endian ())	536	if (!ecb_big_endian ())
461	fp = ecb_bswap32 (fp);	537	fp = ecb_bswap32 (fp);
462		538
463	memcpy (enc->cur, &fp, 4);	539	memcpy (enc->cur, &fp, 4);
…		…
465	}	541	}
466	else	542	else
467	{	543	{
468	uint64_t fp = ecb_double_to_binary64 (nv);	544	uint64_t fp = ecb_double_to_binary64 (nv);
469		545
470	*enc->cur++ = 0xe0 \| 27;	546	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
471		547
472	if (!ecb_big_endian ())	548	if (!ecb_big_endian ())
473	fp = ecb_bswap64 (fp);	549	fp = ecb_bswap64 (fp);
474		550
475	memcpy (enc->cur, &fp, 8);	551	memcpy (enc->cur, &fp, 8);
…		…
484		560
485	if (SvPOKp (sv))	561	if (SvPOKp (sv))
486	{	562	{
487	STRLEN len;	563	STRLEN len;
488	char *str = SvPV (sv, len);	564	char *str = SvPV (sv, len);
489	encode_str (enc, SvUTF8 (sv), str, len);	565	encode_strref (enc, SvUTF8 (sv), str, len);
490	}	566	}
491	else if (SvNOKp (sv))	567	else if (SvNOKp (sv))
492	encode_nv (enc, sv);	568	encode_nv (enc, sv);
493	else if (SvIOKp (sv))	569	else if (SvIOKp (sv))
494	{	570	{
495	if (SvIsUV (sv))	571	if (SvIsUV (sv))
496	encode_uint (enc, 0x00, SvUVX (sv));	572	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
497	else if (SvIVX (sv) >= 0)	573	else if (SvIVX (sv) >= 0)
498	encode_uint (enc, 0x00, SvIVX (sv));	574	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
499	else	575	else
500	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	576	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
501	}	577	}
502	else if (SvROK (sv))	578	else if (SvROK (sv))
503	encode_rv (enc, SvRV (sv));	579	encode_rv (enc, SvRV (sv));
504	else if (!SvOK (sv))	580	else if (!SvOK (sv))
505	encode_ch (enc, 0xe0 \| 22);	581	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
506	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	582	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
507	encode_ch (enc, 0xe0 \| 23);	583	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
508	else	584	else
509	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	585	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
510	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	586	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
511	}	587	}
512		588
513	static SV *	589	static SV *
514	encode_cbor (SV scalar, CBOR cbor)	590	encode_cbor (SV scalar, CBOR cbor)
515	{	591	{
516	enc_t enc = { };	592	enc_t enc = { 0 };
517		593
518	enc.cbor = *cbor;	594	enc.cbor = *cbor;
519	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	595	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
520	enc.cur = SvPVX (enc.sv);	596	enc.cur = SvPVX (enc.sv);
521	enc.end = SvEND (enc.sv);	597	enc.end = SvEND (enc.sv);
522		598
523	SvPOK_only (enc.sv);	599	SvPOK_only (enc.sv);
524		600
525	if (cbor->flags & F_ALLOW_STRINGREF)	601	if (cbor->flags & F_PACK_STRINGS)
526	{	602	{
527	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	603	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
528	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	604	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
529	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	605	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
530	}	606	}
…		…
565	#define DEC_DEC_DEPTH --dec->depth	641	#define DEC_DEC_DEPTH --dec->depth
566		642
567	static UV	643	static UV
568	decode_uint (dec_t *dec)	644	decode_uint (dec_t *dec)
569	{	645	{
570	switch (*dec->cur & 31)	646	U8 m = *dec->cur & MINOR_MASK;
571	{	647	++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		648
577	case 24:	649	if (ecb_expect_true (m < LENGTH_EXT1))
		650	return m;
		651	else if (ecb_expect_true (m == LENGTH_EXT1))
		652	{
578	WANT (2);	653	WANT (1);
579	dec->cur += 2;	654	dec->cur += 1;
580	return dec->cur[-1];	655	return dec->cur[-1];
581		656	}
582	case 25:	657	else if (ecb_expect_true (m == LENGTH_EXT2))
		658	{
583	WANT (3);	659	WANT (2);
584	dec->cur += 3;	660	dec->cur += 2;
585	return (((UV)dec->cur[-2]) << 8)	661	return (((UV)dec->cur[-2]) << 8)
586	\| ((UV)dec->cur[-1]);	662	\| ((UV)dec->cur[-1]);
587		663	}
588	case 26:	664	else if (ecb_expect_true (m == LENGTH_EXT4))
		665	{
589	WANT (5);	666	WANT (4);
590	dec->cur += 5;	667	dec->cur += 4;
591	return (((UV)dec->cur[-4]) << 24)	668	return (((UV)dec->cur[-4]) << 24)
592	\| (((UV)dec->cur[-3]) << 16)	669	\| (((UV)dec->cur[-3]) << 16)
593	\| (((UV)dec->cur[-2]) << 8)	670	\| (((UV)dec->cur[-2]) << 8)
594	\| ((UV)dec->cur[-1]);	671	\| ((UV)dec->cur[-1]);
595		672	}
596	case 27:	673	else if (ecb_expect_true (m == LENGTH_EXT8))
		674	{
597	WANT (9);	675	WANT (8);
598	dec->cur += 9;	676	dec->cur += 8;
		677
		678	return
		679	#if UVSIZE < 8
		680	0
		681	#else
599	return (((UV)dec->cur[-8]) << 56)	682	(((UV)dec->cur[-8]) << 56)
600	\| (((UV)dec->cur[-7]) << 48)	683	\| (((UV)dec->cur[-7]) << 48)
601	\| (((UV)dec->cur[-6]) << 40)	684	\| (((UV)dec->cur[-6]) << 40)
602	\| (((UV)dec->cur[-5]) << 32)	685	\| (((UV)dec->cur[-5]) << 32)
		686	#endif
603	\| (((UV)dec->cur[-4]) << 24)	687	\| (((UV)dec->cur[-4]) << 24)
604	\| (((UV)dec->cur[-3]) << 16)	688	\| (((UV)dec->cur[-3]) << 16)
605	\| (((UV)dec->cur[-2]) << 8)	689	\| (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	690	\| ((UV)dec->cur[-1]);
607		691	}
608	default:	692	else
609	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	693	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
610	}
611		694
612	fail:	695	fail:
613	return 0;	696	return 0;
614	}	697	}
615		698
…		…
620	{	703	{
621	AV *av = newAV ();	704	AV *av = newAV ();
622		705
623	DEC_INC_DEPTH;	706	DEC_INC_DEPTH;
624		707
625	if ((*dec->cur & 31) == 31)	708	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
626	{	709	{
627	++dec->cur;	710	++dec->cur;
628		711
629	for (;;)	712	for (;;)
630	{	713	{
631	WANT (1);	714	WANT (1);
632		715
633	if (*dec->cur == (0xe0 \| 31))	716	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
634	{	717	{
635	++dec->cur;	718	++dec->cur;
636	break;	719	break;
637	}	720	}
638		721
…		…
641	}	724	}
642	else	725	else
643	{	726	{
644	int i, len = decode_uint (dec);	727	int i, len = decode_uint (dec);
645		728
		729	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);	730	av_fill (av, len - 1);
647		731
648	for (i = 0; i < len; ++i)	732	for (i = 0; i < len; ++i)
649	AvARRAY (av)[i] = decode_sv (dec);	733	AvARRAY (av)[i] = decode_sv (dec);
650	}	734	}
…		…
663	{	747	{
664	// for speed reasons, we specialcase single-string	748	// for speed reasons, we specialcase single-string
665	// byte or utf-8 strings as keys, but only when !stringref	749	// byte or utf-8 strings as keys, but only when !stringref
666		750
667	if (ecb_expect_true (!dec->stringref))	751	if (ecb_expect_true (!dec->stringref))
668	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	752	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
669	{	753	{
670	I32 len = decode_uint (dec);	754	I32 len = decode_uint (dec);
671	char key = (char )dec->cur;	755	char key = (char )dec->cur;
672		756
		757	WANT (len);
673	dec->cur += len;	758	dec->cur += len;
674		759
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);	760	hv_store (hv, key, len, decode_sv (dec), 0);
679		761
680	return;	762	return;
681	}	763	}
682	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	764	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
683	{	765	{
684	I32 len = decode_uint (dec);	766	I32 len = decode_uint (dec);
685	char key = (char )dec->cur;	767	char key = (char )dec->cur;
686		768
		769	WANT (len);
687	dec->cur += len;	770	dec->cur += len;
688		771
689	if (ecb_expect_false (dec->stringref))	772	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
690	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	773	if (!is_utf8_string (key, len))
		774	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
691		775
692	hv_store (hv, key, -len, decode_sv (dec), 0);	776	hv_store (hv, key, -len, decode_sv (dec), 0);
693		777
694	return;	778	return;
695	}	779	}
…		…
697	SV *k = decode_sv (dec);	781	SV *k = decode_sv (dec);
698	SV *v = decode_sv (dec);	782	SV *v = decode_sv (dec);
699		783
700	hv_store_ent (hv, k, v, 0);	784	hv_store_ent (hv, k, v, 0);
701	SvREFCNT_dec (k);	785	SvREFCNT_dec (k);
		786
		787	fail:
		788	;
702	}	789	}
703		790
704	static SV *	791	static SV *
705	decode_hv (dec_t *dec)	792	decode_hv (dec_t *dec)
706	{	793	{
707	HV *hv = newHV ();	794	HV *hv = newHV ();
708		795
709	DEC_INC_DEPTH;	796	DEC_INC_DEPTH;
710		797
711	if ((*dec->cur & 31) == 31)	798	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
712	{	799	{
713	++dec->cur;	800	++dec->cur;
714		801
715	for (;;)	802	for (;;)
716	{	803	{
717	WANT (1);	804	WANT (1);
718		805
719	if (*dec->cur == (0xe0 \| 31))	806	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
720	{	807	{
721	++dec->cur;	808	++dec->cur;
722	break;	809	break;
723	}	810	}
724		811
…		…
745	static SV *	832	static SV *
746	decode_str (dec_t *dec, int utf8)	833	decode_str (dec_t *dec, int utf8)
747	{	834	{
748	SV *sv = 0;	835	SV *sv = 0;
749		836
750	if ((*dec->cur & 31) == 31)	837	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
751	{	838	{
		839	// indefinite length strings
752	++dec->cur;	840	++dec->cur;
753		841
		842	U8 major = *dec->cur & MAJOR_MISC;
		843
754	sv = newSVpvn ("", 0);	844	sv = newSVpvn ("", 0);
755		845
756	// not very fast, and certainly not robust against illegal input
757	for (;;)	846	for (;;)
758	{	847	{
759	WANT (1);	848	WANT (1);
760		849
761	if (*dec->cur == (0xe0 \| 31))	850	if ((*dec->cur - major) > LENGTH_EXT8)
		851	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
762	{	852	{
763	++dec->cur;	853	++dec->cur;
764	break;	854	break;
765	}	855	}
		856	else
		857	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
766		858
767	sv_catsv (sv, decode_sv (dec));	859	STRLEN len = decode_uint (dec);
		860
		861	WANT (len);
		862	sv_catpvn (sv, dec->cur, len);
		863	dec->cur += len;
768	}	864	}
769	}	865	}
770	else	866	else
771	{	867	{
772	STRLEN len = decode_uint (dec);	868	STRLEN len = decode_uint (dec);
…		…
779	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	875	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
780	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	876	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
781	}	877	}
782		878
783	if (utf8)	879	if (utf8)
		880	{
		881	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		882	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		883	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		884
784	SvUTF8_on (sv);	885	SvUTF8_on (sv);
		886	}
785		887
786	return sv;	888	return sv;
787		889
788	fail:	890	fail:
789	SvREFCNT_dec (sv);	891	SvREFCNT_dec (sv);
…		…
821	}	923	}
822	break;	924	break;
823		925
824	case CBOR_TAG_STRINGREF:	926	case CBOR_TAG_STRINGREF:
825	{	927	{
826	if ((*dec->cur >> 5) != 0)	928	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
827	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	929	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
828		930
829	UV idx = decode_uint (dec);	931	UV idx = decode_uint (dec);
830		932
831	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	933	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
838	case CBOR_TAG_VALUE_SHAREABLE:	940	case CBOR_TAG_VALUE_SHAREABLE:
839	{	941	{
840	if (ecb_expect_false (!dec->shareable))	942	if (ecb_expect_false (!dec->shareable))
841	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	943	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
842		944
		945	if (dec->cbor.flags & F_ALLOW_CYCLES)
		946	{
843	sv = newSV (0);	947	sv = newSV (0);
844	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	948	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
845		949
846	SV *osv = decode_sv (dec);	950	SV *osv = decode_sv (dec);
847	sv_setsv (sv, osv);	951	sv_setsv (sv, osv);
848	SvREFCNT_dec_NN (osv);	952	SvREFCNT_dec_NN (osv);
		953	}
		954	else
		955	{
		956	av_push (dec->shareable, &PL_sv_undef);
		957	int idx = AvFILLp (dec->shareable);
		958	sv = decode_sv (dec);
		959	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		960	}
849	}	961	}
850	break;	962	break;
851		963
852	case CBOR_TAG_VALUE_SHAREDREF:	964	case CBOR_TAG_VALUE_SHAREDREF:
853	{	965	{
854	if ((*dec->cur >> 5) != 0)	966	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
855	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	967	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
856		968
857	UV idx = decode_uint (dec);	969	UV idx = decode_uint (dec);
858		970
859	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	971	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
860	ERR ("corrupted CBOR data (sharedref index out of bounds)");	972	ERR ("corrupted CBOR data (sharedref index out of bounds)");
861		973
862	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	974	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		975
		976	if (sv == &PL_sv_undef)
		977	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
863	}	978	}
864	break;	979	break;
865		980
866	case CBOR_TAG_PERL_OBJECT:	981	case CBOR_TAG_PERL_OBJECT:
867	{	982	{
…		…
882	if (!method)	997	if (!method)
883	ERR ("cannot decode perl-object (package does not have a THAW method)");	998	ERR ("cannot decode perl-object (package does not have a THAW method)");
884		999
885	dSP;	1000	dSP;
886		1001
887	ENTER; SAVETMPS; PUSHMARK (SP);	1002	ENTER; SAVETMPS;
		1003	PUSHMARK (SP);
888	EXTEND (SP, len + 1);	1004	EXTEND (SP, len + 1);
889	// we re-bless the reference to get overload and other niceties right	1005	// we re-bless the reference to get overload and other niceties right
890	PUSHs (*av_fetch (av, 0, 1));	1006	PUSHs (*av_fetch (av, 0, 1));
891	PUSHs (sv_cbor);	1007	PUSHs (sv_cbor);
892		1008
…		…
917	default:	1033	default:
918	{	1034	{
919	sv = decode_sv (dec);	1035	sv = decode_sv (dec);
920		1036
921	dSP;	1037	dSP;
922	ENTER; SAVETMPS; PUSHMARK (SP);	1038	ENTER; SAVETMPS;
		1039	SAVESTACK_POS ();
		1040	PUSHMARK (SP);
923	EXTEND (SP, 2);	1041	EXTEND (SP, 2);
924	PUSHs (newSVuv (tag));	1042	PUSHs (newSVuv (tag));
925	PUSHs (sv);	1043	PUSHs (sv);
926		1044
927	PUTBACK;	1045	PUTBACK;
…		…
968	static SV *	1086	static SV *
969	decode_sv (dec_t *dec)	1087	decode_sv (dec_t *dec)
970	{	1088	{
971	WANT (1);	1089	WANT (1);
972		1090
973	switch (*dec->cur >> 5)	1091	switch (*dec->cur >> MAJOR_SHIFT)
974	{	1092	{
975	case 0: // unsigned int	1093	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
976	return newSVuv (decode_uint (dec));	1094	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
977	case 1: // negative int	1095	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
978	return newSViv (-1 - (IV)decode_uint (dec));	1096	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
979	case 2: // octet string	1097	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
980	return decode_str (dec, 0);	1098	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
981	case 3: // utf-8 string	1099	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
982	return decode_str (dec, 1);	1100
983	case 4: // array	1101	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)	1102	switch (*dec->cur++ & MINOR_MASK)
991	{	1103	{
992	case 20:	1104	case SIMPLE_FALSE:
993	#if CBOR_SLOW	1105	#if CBOR_SLOW
994	types_false = get_bool ("Types::Serialiser::false");	1106	types_false = get_bool ("Types::Serialiser::false");
995	#endif	1107	#endif
996	return newSVsv (types_false);	1108	return newSVsv (types_false);
997	case 21:	1109	case SIMPLE_TRUE:
998	#if CBOR_SLOW	1110	#if CBOR_SLOW
999	types_true = get_bool ("Types::Serialiser::true");	1111	types_true = get_bool ("Types::Serialiser::true");
1000	#endif	1112	#endif
1001	return newSVsv (types_true);	1113	return newSVsv (types_true);
1002	case 22:	1114	case SIMPLE_NULL:
1003	return newSVsv (&PL_sv_undef);	1115	return newSVsv (&PL_sv_undef);
1004	case 23:	1116	case SIMPLE_UNDEF:
1005	#if CBOR_SLOW	1117	#if CBOR_SLOW
1006	types_error = get_bool ("Types::Serialiser::error");	1118	types_error = get_bool ("Types::Serialiser::error");
1007	#endif	1119	#endif
1008	return newSVsv (types_error);	1120	return newSVsv (types_error);
1009		1121
1010	case 25:	1122	case MISC_FLOAT16:
1011	{	1123	{
1012	WANT (2);	1124	WANT (2);
1013		1125
1014	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1126	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1015	dec->cur += 2;	1127	dec->cur += 2;
1016		1128
1017	return newSVnv (ecb_binary16_to_float (fp));	1129	return newSVnv (ecb_binary16_to_float (fp));
1018	}	1130	}
1019		1131
1020	case 26:	1132	case MISC_FLOAT32:
1021	{	1133	{
1022	uint32_t fp;	1134	uint32_t fp;
1023	WANT (4);	1135	WANT (4);
1024	memcpy (&fp, dec->cur, 4);	1136	memcpy (&fp, dec->cur, 4);
1025	dec->cur += 4;	1137	dec->cur += 4;
…		…
1028	fp = ecb_bswap32 (fp);	1140	fp = ecb_bswap32 (fp);
1029		1141
1030	return newSVnv (ecb_binary32_to_float (fp));	1142	return newSVnv (ecb_binary32_to_float (fp));
1031	}	1143	}
1032		1144
1033	case 27:	1145	case MISC_FLOAT64:
1034	{	1146	{
1035	uint64_t fp;	1147	uint64_t fp;
1036	WANT (8);	1148	WANT (8);
1037	memcpy (&fp, dec->cur, 8);	1149	memcpy (&fp, dec->cur, 8);
1038	dec->cur += 8;	1150	dec->cur += 8;
…		…
1041	fp = ecb_bswap64 (fp);	1153	fp = ecb_bswap64 (fp);
1042		1154
1043	return newSVnv (ecb_binary64_to_double (fp));	1155	return newSVnv (ecb_binary64_to_double (fp));
1044	}	1156	}
1045		1157
1046	// 0..19 unassigned	1158	// 0..19 unassigned simple
1047	// 24 reserved + unassigned (reserved values are not encodable)	1159	// 24 reserved + unassigned simple (reserved values are not encodable)
		1160	// 28-30 unassigned misc
		1161	// 31 break code
1048	default:	1162	default:
1049	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1163	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1050	}	1164	}
1051		1165
1052	break;	1166	break;
1053	}	1167	}
1054		1168
…		…
1057	}	1171	}
1058		1172
1059	static SV *	1173	static SV *
1060	decode_cbor (SV string, CBOR cbor, char **offset_return)	1174	decode_cbor (SV string, CBOR cbor, char **offset_return)
1061	{	1175	{
1062	dec_t dec = { };	1176	dec_t dec = { 0 };
1063	SV *sv;	1177	SV *sv;
1064	STRLEN len;	1178	STRLEN len;
1065	char *data = SvPVbyte (string, len);	1179	char *data = SvPVbyte (string, len);
1066		1180
1067	if (len > cbor->max_size && cbor->max_size)	1181	if (len > cbor->max_size && cbor->max_size)
…		…
1081	if (dec.cur != dec.end && !dec.err)	1195	if (dec.cur != dec.end && !dec.err)
1082	dec.err = "garbage after CBOR object";	1196	dec.err = "garbage after CBOR object";
1083		1197
1084	if (dec.err)	1198	if (dec.err)
1085	{	1199	{
		1200	if (dec.shareable)
		1201	{
		1202	// need to break cyclic links, which whould all be in shareable
		1203	int i;
		1204	SV **svp;
		1205
		1206	for (i = av_len (dec.shareable) + 1; i--; )
		1207	if ((svp = av_fetch (dec.shareable, i, 0)))
		1208	sv_setsv (*svp, &PL_sv_undef);
		1209	}
		1210
1086	SvREFCNT_dec (sv);	1211	SvREFCNT_dec (sv);
1087	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1212	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1088	}	1213	}
1089		1214
1090	sv = sv_2mortal (sv);	1215	sv = sv_2mortal (sv);
1091		1216
1092	return sv;	1217	return sv;
1093	}	1218	}
1094		1219
		1220	/////////////////////////////////////////////////////////////////////////////
		1221	// incremental parser
		1222
		1223	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1224
		1225	// returns 0 for notyet, 1 for success or error
		1226	static int
		1227	incr_parse (CBOR self, SV cborstr)
		1228	{
		1229	STRLEN cur;
		1230	SvPV (cborstr, cur);
		1231
		1232	while (ecb_expect_true (self->incr_need <= cur))
		1233	{
		1234	// table of integer count bytes
		1235	static I8 incr_len[MINOR_MASK + 1] = {
		1236	0, 0, 0, 0, 0, 0, 0, 0,
		1237	0, 0, 0, 0, 0, 0, 0, 0,
		1238	0, 0, 0, 0, 0, 0, 0, 0,
		1239	1, 2, 4, 8,-1,-1,-1,-2
		1240	};
		1241
		1242	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1243	U8 m = *p & MINOR_MASK;
		1244	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1245	I8 ilen = incr_len[m];
		1246
		1247	self->incr_need = self->incr_pos + 1;
		1248
		1249	if (ecb_expect_false (ilen < 0))
		1250	{
		1251	if (m != MINOR_INDEF)
		1252	return 1; // error
		1253
		1254	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1255	{
		1256	if (count >= 0)
		1257	return 1; // error
		1258
		1259	count = 1;
		1260	}
		1261	else
		1262	{
		1263	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1264	count = -1;
		1265	}
		1266	}
		1267	else
		1268	{
		1269	self->incr_need += ilen;
		1270	if (ecb_expect_false (self->incr_need > cur))
		1271	return 0;
		1272
		1273	int major = *p >> MAJOR_SHIFT;
		1274
		1275	switch (major)
		1276	{
		1277	case MAJOR_TAG >> MAJOR_SHIFT:
		1278	++count; // tags merely prefix another value
		1279	break;
		1280
		1281	case MAJOR_BYTES >> MAJOR_SHIFT:
		1282	case MAJOR_TEXT >> MAJOR_SHIFT:
		1283	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1284	case MAJOR_MAP >> MAJOR_SHIFT:
		1285	{
		1286	UV len;
		1287
		1288	if (ecb_expect_false (ilen))
		1289	{
		1290	len = 0;
		1291
		1292	do {
		1293	len = (len << 8) \| *++p;
		1294	} while (--ilen);
		1295	}
		1296	else
		1297	len = m;
		1298
		1299	switch (major)
		1300	{
		1301	case MAJOR_BYTES >> MAJOR_SHIFT:
		1302	case MAJOR_TEXT >> MAJOR_SHIFT:
		1303	self->incr_need += len;
		1304	if (ecb_expect_false (self->incr_need > cur))
		1305	return 0;
		1306
		1307	break;
		1308
		1309	case MAJOR_MAP >> MAJOR_SHIFT:
		1310	len <<= 1;
		1311	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1312	if (len)
		1313	{
		1314	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1315	count = len + 1;
		1316	}
		1317	break;
		1318	}
		1319	}
		1320	}
		1321	}
		1322
		1323	self->incr_pos = self->incr_need;
		1324
		1325	if (count > 0)
		1326	{
		1327	while (!--count)
		1328	{
		1329	if (!AvFILLp (self->incr_count))
		1330	return 1; // done
		1331
		1332	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1333	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1334	}
		1335
		1336	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1337	}
		1338	}
		1339
		1340	return 0;
		1341	}
		1342
		1343
1095	/////////////////////////////////////////////////////////////////////////////	1344	/////////////////////////////////////////////////////////////////////////////
1096	// XS interface functions	1345	// XS interface functions
1097		1346
1098	MODULE = CBOR::XS PACKAGE = CBOR::XS	1347	MODULE = CBOR::XS PACKAGE = CBOR::XS
1099		1348
…		…
1139	void shrink (CBOR *self, int enable = 1)	1388	void shrink (CBOR *self, int enable = 1)
1140	ALIAS:	1389	ALIAS:
1141	shrink = F_SHRINK	1390	shrink = F_SHRINK
1142	allow_unknown = F_ALLOW_UNKNOWN	1391	allow_unknown = F_ALLOW_UNKNOWN
1143	allow_sharing = F_ALLOW_SHARING	1392	allow_sharing = F_ALLOW_SHARING
1144	allow_stringref = F_ALLOW_STRINGREF	1393	allow_cycles = F_ALLOW_CYCLES
		1394	pack_strings = F_PACK_STRINGS
		1395	validate_utf8 = F_VALIDATE_UTF8
1145	PPCODE:	1396	PPCODE:
1146	{	1397	{
1147	if (enable)	1398	if (enable)
1148	self->flags \|= ix;	1399	self->flags \|= ix;
1149	else	1400	else
…		…
1155	void get_shrink (CBOR *self)	1406	void get_shrink (CBOR *self)
1156	ALIAS:	1407	ALIAS:
1157	get_shrink = F_SHRINK	1408	get_shrink = F_SHRINK
1158	get_allow_unknown = F_ALLOW_UNKNOWN	1409	get_allow_unknown = F_ALLOW_UNKNOWN
1159	get_allow_sharing = F_ALLOW_SHARING	1410	get_allow_sharing = F_ALLOW_SHARING
1160	get_allow_stringref = F_ALLOW_STRINGREF	1411	get_allow_cycles = F_ALLOW_CYCLES
		1412	get_pack_strings = F_PACK_STRINGS
		1413	get_validate_utf8 = F_VALIDATE_UTF8
1161	PPCODE:	1414	PPCODE:
1162	XPUSHs (boolSV (self->flags & ix));	1415	XPUSHs (boolSV (self->flags & ix));
1163		1416
1164	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1417	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1165	PPCODE:	1418	PPCODE:
…		…
1214	EXTEND (SP, 2);	1467	EXTEND (SP, 2);
1215	PUSHs (sv);	1468	PUSHs (sv);
1216	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1469	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1217	}	1470	}
1218		1471
		1472	void incr_parse (CBOR self, SV cborstr)
		1473	ALIAS:
		1474	incr_parse_multiple = 1
		1475	PPCODE:
		1476	{
		1477	if (SvUTF8 (cborstr))
		1478	sv_utf8_downgrade (cborstr, 0);
		1479
		1480	if (!self->incr_count)
		1481	{
		1482	self->incr_count = newAV ();
		1483	self->incr_pos = 0;
		1484	self->incr_need = 1;
		1485
		1486	av_push (self->incr_count, newSViv (1));
		1487	}
		1488
		1489	do
		1490	{
		1491	if (!incr_parse (self, cborstr))
		1492	{
		1493	if (self->incr_need > self->max_size && self->max_size)
		1494	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1495	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1496
		1497	break;
		1498	}
		1499
		1500	SV *sv;
		1501	char *offset;
		1502
		1503	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1504	XPUSHs (sv);
		1505
		1506	sv_chop (cborstr, offset);
		1507
		1508	av_clear (self->incr_count);
		1509	av_push (self->incr_count, newSViv (1));
		1510
		1511	self->incr_pos = 0;
		1512	self->incr_need = self->incr_pos + 1;
		1513	}
		1514	while (ix);
		1515	}
		1516
		1517	void incr_reset (CBOR *self)
		1518	CODE:
		1519	{
		1520	SvREFCNT_dec (self->incr_count);
		1521	self->incr_count = 0;
		1522	}
		1523
1219	void DESTROY (CBOR *self)	1524	void DESTROY (CBOR *self)
1220	PPCODE:	1525	PPCODE:
1221	cbor_free (self);	1526	cbor_free (self);
1222		1527
1223	PROTOTYPES: ENABLE	1528	PROTOTYPES: ENABLE
1224		1529
1225	void encode_cbor (SV *scalar)	1530	void encode_cbor (SV *scalar)
		1531	ALIAS:
		1532	encode_cbor = 0
		1533	encode_cbor_sharing = F_ALLOW_SHARING
1226	PPCODE:	1534	PPCODE:
1227	{	1535	{
1228	CBOR cbor;	1536	CBOR cbor;
1229	cbor_init (&cbor);	1537	cbor_init (&cbor);
		1538	cbor.flags \|= ix;
1230	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1539	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1231	XPUSHs (scalar);	1540	XPUSHs (scalar);
1232	}	1541	}
1233		1542
1234	void decode_cbor (SV *cborstr)	1543	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 1.50 by root, Thu Feb 25 02:29:22 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 1.50 by root, Thu Feb 25 02:29:22 2016 UTC