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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.29 by root, Thu Nov 28 09:13:12 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC

…		…
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
11		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
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	// extensions	69	// extensions
32	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref	70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
33	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
34	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
35	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
36	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
37	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // 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
…		…
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	// exceptional path for bytze strings that need to be utf8-encoded
		289	STRLEN ulen = len;
		290	U8 p, pend = (U8 *)str + len;
		291
		292	for (p = (U8 *)str; p < pend; ++p)
		293	ulen += *p >> 7; // count set high bits
		294
		295	encode_uint (enc, MAJOR_TEXT, ulen);
		296
		297	need (enc, ulen);
		298	for (p = (U8 *)str; p < pend; ++p)
		299	if (*p < 0x80)
		300	enc->cur++ = p;
		301	else
		302	{
		303	enc->cur++ = 0xc0 + (p >> 6);
		304	enc->cur++ = 0x80 + (p & 63);
		305	}
		306
		307	return;
		308	}
		309
		310	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
		311	need (enc, len);
		312	memcpy (enc->cur, str, len);
		313	enc->cur += len;
233	}	314	}
234		315
235	static void	316	static void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	317	encode_strref (enc_t enc, int utf8, char str, STRLEN len)
237	{	318	{
238	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	319	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
239	{	320	{
240	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	321	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
241		322
242	if (SvOK (*svp))	323	if (SvOK (*svp))
243	{	324	{
244	// already registered, use stringref	325	// already registered, use stringref
245	encode_tag (enc, CBOR_TAG_STRINGREF);	326	encode_tag (enc, CBOR_TAG_STRINGREF);
246	encode_uint (enc, 0x00, SvUV (*svp));	327	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
247	return;	328	return;
248	}	329	}
249	else if (len >= minimum_string_length (enc->stringref_idx))	330	else if (len >= minimum_string_length (enc->stringref_idx))
250	{	331	{
251	// register only	332	// register only
252	sv_setuv (*svp, enc->stringref_idx);	333	sv_setuv (*svp, enc->stringref_idx);
253	++enc->stringref_idx;	334	++enc->stringref_idx;
254	}	335	}
255	}	336	}
256		337
257	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	338	encode_str (enc, utf8, str, len);
258	need (enc, len);
259	memcpy (enc->cur, str, len);
260	enc->cur += len;
261	}	339	}
262		340
263	static void encode_sv (enc_t enc, SV sv);	341	static void encode_sv (enc_t enc, SV sv);
264		342
265	static void	343	static void
…		…
270	if (enc->depth >= enc->cbor.max_depth)	348	if (enc->depth >= enc->cbor.max_depth)
271	croak (ERR_NESTING_EXCEEDED);	349	croak (ERR_NESTING_EXCEEDED);
272		350
273	++enc->depth;	351	++enc->depth;
274		352
275	encode_uint (enc, 0x80, len + 1);	353	encode_uint (enc, MAJOR_ARRAY, len + 1);
276		354
		355	if (SvMAGICAL (av))
277	for (i = 0; i <= len; ++i)	356	for (i = 0; i <= len; ++i)
278	{	357	{
279	SV **svp = av_fetch (av, i, 0);	358	SV **svp = av_fetch (av, i, 0);
280	encode_sv (enc, svp ? *svp : &PL_sv_undef);	359	encode_sv (enc, svp ? *svp : &PL_sv_undef);
281	}	360	}
		361	else
		362	for (i = 0; i <= len; ++i)
		363	{
		364	SV *sv = AvARRAY (av)[i];
		365	encode_sv (enc, sv ? sv : &PL_sv_undef);
		366	}
282		367
283	--enc->depth;	368	--enc->depth;
284	}	369	}
285		370
286	static void	371	static void
…		…
295		380
296	int pairs = hv_iterinit (hv);	381	int pairs = hv_iterinit (hv);
297	int mg = SvMAGICAL (hv);	382	int mg = SvMAGICAL (hv);
298		383
299	if (mg)	384	if (mg)
300	encode_ch (enc, 0xa0 \| 31);	385	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
301	else	386	else
302	encode_uint (enc, 0xa0, pairs);	387	encode_uint (enc, MAJOR_MAP, pairs);
303		388
304	while ((he = hv_iternext (hv)))	389	while ((he = hv_iternext (hv)))
305	{	390	{
306	if (HeKLEN (he) == HEf_SVKEY)	391	if (HeKLEN (he) == HEf_SVKEY)
307	encode_sv (enc, HeSVKEY (he));	392	encode_sv (enc, HeSVKEY (he));
308	else	393	else
309	encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	394	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
310		395
311	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	396	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
312	}	397	}
313		398
314	if (mg)	399	if (mg)
315	encode_ch (enc, 0xe0 \| 31);	400	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
316		401
317	--enc->depth;	402	--enc->depth;
318	}	403	}
319		404
320	// encode objects, arrays and special \0=false and \1=true values.	405	// encode objects, arrays and special \0=false and \1=true values.
321	static void	406	static void
322	encode_rv (enc_t enc, SV sv)	407	encode_rv (enc_t enc, SV sv)
323	{	408	{
324	SvGETMAGIC (sv);	409	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		410
348	svtype svt = SvTYPE (sv);	411	svtype svt = SvTYPE (sv);
349		412
350	if (ecb_expect_false (SvOBJECT (sv)))	413	if (ecb_expect_false (SvOBJECT (sv)))
351	{	414	{
…		…
358	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	421	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
359	? cbor_tagged_stash	422	? cbor_tagged_stash
360	: gv_stashpv ("CBOR::XS::Tagged" , 1);	423	: gv_stashpv ("CBOR::XS::Tagged" , 1);
361		424
362	HV *stash = SvSTASH (sv);	425	HV *stash = SvSTASH (sv);
363	GV *method;
364		426
365	if (stash == boolean_stash)	427	if (stash == boolean_stash)
366	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	428	{
		429	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		430	return;
		431	}
367	else if (stash == error_stash)	432	else if (stash == error_stash)
368	encode_ch (enc, 0xe0 \| 23);	433	{
		434	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		435	return;
		436	}
369	else if (stash == tagged_stash)	437	else if (stash == tagged_stash)
370	{	438	{
371	if (svt != SVt_PVAV)	439	if (svt != SVt_PVAV)
372	croak ("encountered CBOR::XS::Tagged object that isn't an array");	440	croak ("encountered CBOR::XS::Tagged object that isn't an array");
373		441
374	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	442	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
375	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	443	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		444
		445	return;
		446	}
		447	}
		448
		449	if (ecb_expect_false (SvREFCNT (sv) > 1)
		450	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		451	{
		452	if (!enc->shareable)
		453	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		454
		455	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		456
		457	if (SvOK (*svp))
376	}	458	{
		459	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		460	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		461	return;
		462	}
		463	else
		464	{
		465	sv_setuv (*svp, enc->shareable_idx);
		466	++enc->shareable_idx;
		467	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		468	}
		469	}
		470
		471	if (ecb_expect_false (SvOBJECT (sv)))
		472	{
		473	HV *stash = SvSTASH (sv);
		474	GV *method;
		475
377	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	476	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
378	{	477	{
379	dSP;	478	dSP;
380		479
381	ENTER; SAVETMPS; PUSHMARK (SP);	480	ENTER; SAVETMPS;
		481	PUSHMARK (SP);
382	// 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
383	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
384		484
385	PUTBACK;	485	PUTBACK;
386	// G_SCALAR ensures that return value is 1	486	// G_SCALAR ensures that return value is 1
…		…
399	}	499	}
400	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	500	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
401	{	501	{
402	dSP;	502	dSP;
403		503
404	ENTER; SAVETMPS; PUSHMARK (SP);	504	ENTER; SAVETMPS;
		505	SAVESTACK_POS ();
		506	PUSHMARK (SP);
405	EXTEND (SP, 2);	507	EXTEND (SP, 2);
406	// we re-bless the reference to get overload and other niceties right	508	// we re-bless the reference to get overload and other niceties right
407	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	509	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
408	PUSHs (sv_cbor);	510	PUSHs (sv_cbor);
409		511
…		…
414	// catch this surprisingly common error	516	// catch this surprisingly common error
415	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	517	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));	518	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
417		519
418	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	520	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
419	encode_uint (enc, 0x80, count + 1);	521	encode_uint (enc, MAJOR_ARRAY, count + 1);
420	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	522	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
421		523
422	while (count)	524	while (count)
423	encode_sv (enc, SP[1 - count--]);	525	encode_sv (enc, SP[1 - count--]);
424		526
425	PUTBACK;	527	PUTBACK;
…		…
446	{	548	{
447	double nv = SvNVX (sv);	549	double nv = SvNVX (sv);
448		550
449	need (enc, 9);	551	need (enc, 9);
450		552
451	if (ecb_expect_false (nv == (U32)nv))	553	if (ecb_expect_false (nv == (NV)(U32)nv))
452	encode_uint (enc, 0x00, (U32)nv);	554	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
453	//TODO: maybe I32?	555	//TODO: maybe I32?
454	else if (ecb_expect_false (nv == (float)nv))	556	else if (ecb_expect_false (nv == (float)nv))
455	{	557	{
456	uint32_t fp = ecb_float_to_binary32 (nv);	558	uint32_t fp = ecb_float_to_binary32 (nv);
457		559
458	*enc->cur++ = 0xe0 \| 26;	560	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
459		561
460	if (!ecb_big_endian ())	562	if (!ecb_big_endian ())
461	fp = ecb_bswap32 (fp);	563	fp = ecb_bswap32 (fp);
462		564
463	memcpy (enc->cur, &fp, 4);	565	memcpy (enc->cur, &fp, 4);
…		…
465	}	567	}
466	else	568	else
467	{	569	{
468	uint64_t fp = ecb_double_to_binary64 (nv);	570	uint64_t fp = ecb_double_to_binary64 (nv);
469		571
470	*enc->cur++ = 0xe0 \| 27;	572	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
471		573
472	if (!ecb_big_endian ())	574	if (!ecb_big_endian ())
473	fp = ecb_bswap64 (fp);	575	fp = ecb_bswap64 (fp);
474		576
475	memcpy (enc->cur, &fp, 8);	577	memcpy (enc->cur, &fp, 8);
…		…
484		586
485	if (SvPOKp (sv))	587	if (SvPOKp (sv))
486	{	588	{
487	STRLEN len;	589	STRLEN len;
488	char *str = SvPV (sv, len);	590	char *str = SvPV (sv, len);
489	encode_str (enc, SvUTF8 (sv), str, len);	591	encode_strref (enc, SvUTF8 (sv), str, len);
490	}	592	}
491	else if (SvNOKp (sv))	593	else if (SvNOKp (sv))
492	encode_nv (enc, sv);	594	encode_nv (enc, sv);
493	else if (SvIOKp (sv))	595	else if (SvIOKp (sv))
494	{	596	{
495	if (SvIsUV (sv))	597	if (SvIsUV (sv))
496	encode_uint (enc, 0x00, SvUVX (sv));	598	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
497	else if (SvIVX (sv) >= 0)	599	else if (SvIVX (sv) >= 0)
498	encode_uint (enc, 0x00, SvIVX (sv));	600	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
499	else	601	else
500	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	602	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
501	}	603	}
502	else if (SvROK (sv))	604	else if (SvROK (sv))
503	encode_rv (enc, SvRV (sv));	605	encode_rv (enc, SvRV (sv));
504	else if (!SvOK (sv))	606	else if (!SvOK (sv))
505	encode_ch (enc, 0xe0 \| 22);	607	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
506	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	608	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
507	encode_ch (enc, 0xe0 \| 23);	609	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
508	else	610	else
509	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	611	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
510	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	612	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
511	}	613	}
512		614
513	static SV *	615	static SV *
514	encode_cbor (SV scalar, CBOR cbor)	616	encode_cbor (SV scalar, CBOR cbor)
515	{	617	{
516	enc_t enc = { };	618	enc_t enc = { 0 };
517		619
518	enc.cbor = *cbor;	620	enc.cbor = *cbor;
519	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	621	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
520	enc.cur = SvPVX (enc.sv);	622	enc.cur = SvPVX (enc.sv);
521	enc.end = SvEND (enc.sv);	623	enc.end = SvEND (enc.sv);
522		624
523	SvPOK_only (enc.sv);	625	SvPOK_only (enc.sv);
524		626
525	if (cbor->flags & F_ALLOW_STRINGREF)	627	if (cbor->flags & F_PACK_STRINGS)
526	{	628	{
527	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	629	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
528	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	630	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
529	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	631	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
530	}	632	}
…		…
565	#define DEC_DEC_DEPTH --dec->depth	667	#define DEC_DEC_DEPTH --dec->depth
566		668
567	static UV	669	static UV
568	decode_uint (dec_t *dec)	670	decode_uint (dec_t *dec)
569	{	671	{
570	switch (*dec->cur & 31)	672	U8 m = *dec->cur & MINOR_MASK;
571	{	673	++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		674
577	case 24:	675	if (ecb_expect_true (m < LENGTH_EXT1))
		676	return m;
		677	else if (ecb_expect_true (m == LENGTH_EXT1))
		678	{
578	WANT (2);	679	WANT (1);
579	dec->cur += 2;	680	dec->cur += 1;
580	return dec->cur[-1];	681	return dec->cur[-1];
581		682	}
582	case 25:	683	else if (ecb_expect_true (m == LENGTH_EXT2))
		684	{
583	WANT (3);	685	WANT (2);
584	dec->cur += 3;	686	dec->cur += 2;
585	return (((UV)dec->cur[-2]) << 8)	687	return (((UV)dec->cur[-2]) << 8)
586	\| ((UV)dec->cur[-1]);	688	\| ((UV)dec->cur[-1]);
587		689	}
588	case 26:	690	else if (ecb_expect_true (m == LENGTH_EXT4))
		691	{
589	WANT (5);	692	WANT (4);
590	dec->cur += 5;	693	dec->cur += 4;
591	return (((UV)dec->cur[-4]) << 24)	694	return (((UV)dec->cur[-4]) << 24)
592	\| (((UV)dec->cur[-3]) << 16)	695	\| (((UV)dec->cur[-3]) << 16)
593	\| (((UV)dec->cur[-2]) << 8)	696	\| (((UV)dec->cur[-2]) << 8)
594	\| ((UV)dec->cur[-1]);	697	\| ((UV)dec->cur[-1]);
595		698	}
596	case 27:	699	else if (ecb_expect_true (m == LENGTH_EXT8))
		700	{
597	WANT (9);	701	WANT (8);
598	dec->cur += 9;	702	dec->cur += 8;
		703
		704	return
		705	#if UVSIZE < 8
		706	0
		707	#else
599	return (((UV)dec->cur[-8]) << 56)	708	(((UV)dec->cur[-8]) << 56)
600	\| (((UV)dec->cur[-7]) << 48)	709	\| (((UV)dec->cur[-7]) << 48)
601	\| (((UV)dec->cur[-6]) << 40)	710	\| (((UV)dec->cur[-6]) << 40)
602	\| (((UV)dec->cur[-5]) << 32)	711	\| (((UV)dec->cur[-5]) << 32)
		712	#endif
603	\| (((UV)dec->cur[-4]) << 24)	713	\| (((UV)dec->cur[-4]) << 24)
604	\| (((UV)dec->cur[-3]) << 16)	714	\| (((UV)dec->cur[-3]) << 16)
605	\| (((UV)dec->cur[-2]) << 8)	715	\| (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	716	\| ((UV)dec->cur[-1]);
607		717	}
608	default:	718	else
609	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	719	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
610	}
611		720
612	fail:	721	fail:
613	return 0;	722	return 0;
614	}	723	}
615		724
…		…
620	{	729	{
621	AV *av = newAV ();	730	AV *av = newAV ();
622		731
623	DEC_INC_DEPTH;	732	DEC_INC_DEPTH;
624		733
625	if ((*dec->cur & 31) == 31)	734	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
626	{	735	{
627	++dec->cur;	736	++dec->cur;
628		737
629	for (;;)	738	for (;;)
630	{	739	{
631	WANT (1);	740	WANT (1);
632		741
633	if (*dec->cur == (0xe0 \| 31))	742	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
634	{	743	{
635	++dec->cur;	744	++dec->cur;
636	break;	745	break;
637	}	746	}
638		747
…		…
641	}	750	}
642	else	751	else
643	{	752	{
644	int i, len = decode_uint (dec);	753	int i, len = decode_uint (dec);
645		754
		755	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);	756	av_fill (av, len - 1);
647		757
648	for (i = 0; i < len; ++i)	758	for (i = 0; i < len; ++i)
649	AvARRAY (av)[i] = decode_sv (dec);	759	AvARRAY (av)[i] = decode_sv (dec);
650	}	760	}
…		…
663	{	773	{
664	// for speed reasons, we specialcase single-string	774	// for speed reasons, we specialcase single-string
665	// byte or utf-8 strings as keys, but only when !stringref	775	// byte or utf-8 strings as keys, but only when !stringref
666		776
667	if (ecb_expect_true (!dec->stringref))	777	if (ecb_expect_true (!dec->stringref))
668	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	778	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
669	{	779	{
670	I32 len = decode_uint (dec);	780	I32 len = decode_uint (dec);
671	char key = (char )dec->cur;	781	char key = (char )dec->cur;
672		782
		783	WANT (len);
673	dec->cur += len;	784	dec->cur += len;
674		785
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);	786	hv_store (hv, key, len, decode_sv (dec), 0);
679		787
680	return;	788	return;
681	}	789	}
682	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	790	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
683	{	791	{
684	I32 len = decode_uint (dec);	792	I32 len = decode_uint (dec);
685	char key = (char )dec->cur;	793	char key = (char )dec->cur;
686		794
		795	WANT (len);
687	dec->cur += len;	796	dec->cur += len;
688		797
689	if (ecb_expect_false (dec->stringref))	798	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
690	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	799	if (!is_utf8_string (key, len))
		800	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
691		801
692	hv_store (hv, key, -len, decode_sv (dec), 0);	802	hv_store (hv, key, -len, decode_sv (dec), 0);
693		803
694	return;	804	return;
695	}	805	}
…		…
697	SV *k = decode_sv (dec);	807	SV *k = decode_sv (dec);
698	SV *v = decode_sv (dec);	808	SV *v = decode_sv (dec);
699		809
700	hv_store_ent (hv, k, v, 0);	810	hv_store_ent (hv, k, v, 0);
701	SvREFCNT_dec (k);	811	SvREFCNT_dec (k);
		812
		813	fail:
		814	;
702	}	815	}
703		816
704	static SV *	817	static SV *
705	decode_hv (dec_t *dec)	818	decode_hv (dec_t *dec)
706	{	819	{
707	HV *hv = newHV ();	820	HV *hv = newHV ();
708		821
709	DEC_INC_DEPTH;	822	DEC_INC_DEPTH;
710		823
711	if ((*dec->cur & 31) == 31)	824	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
712	{	825	{
713	++dec->cur;	826	++dec->cur;
714		827
715	for (;;)	828	for (;;)
716	{	829	{
717	WANT (1);	830	WANT (1);
718		831
719	if (*dec->cur == (0xe0 \| 31))	832	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
720	{	833	{
721	++dec->cur;	834	++dec->cur;
722	break;	835	break;
723	}	836	}
724		837
…		…
745	static SV *	858	static SV *
746	decode_str (dec_t *dec, int utf8)	859	decode_str (dec_t *dec, int utf8)
747	{	860	{
748	SV *sv = 0;	861	SV *sv = 0;
749		862
750	if ((*dec->cur & 31) == 31)	863	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
751	{	864	{
		865	// indefinite length strings
752	++dec->cur;	866	++dec->cur;
753		867
		868	U8 major = *dec->cur & MAJOR_MISC;
		869
754	sv = newSVpvn ("", 0);	870	sv = newSVpvn ("", 0);
755		871
756	// not very fast, and certainly not robust against illegal input
757	for (;;)	872	for (;;)
758	{	873	{
759	WANT (1);	874	WANT (1);
760		875
761	if (*dec->cur == (0xe0 \| 31))	876	if ((*dec->cur - major) > LENGTH_EXT8)
		877	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
762	{	878	{
763	++dec->cur;	879	++dec->cur;
764	break;	880	break;
765	}	881	}
		882	else
		883	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
766		884
767	sv_catsv (sv, decode_sv (dec));	885	STRLEN len = decode_uint (dec);
		886
		887	WANT (len);
		888	sv_catpvn (sv, dec->cur, len);
		889	dec->cur += len;
768	}	890	}
769	}	891	}
770	else	892	else
771	{	893	{
772	STRLEN len = decode_uint (dec);	894	STRLEN len = decode_uint (dec);
…		…
779	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	901	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
780	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	902	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
781	}	903	}
782		904
783	if (utf8)	905	if (utf8)
		906	{
		907	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		908	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		909	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		910
784	SvUTF8_on (sv);	911	SvUTF8_on (sv);
		912	}
785		913
786	return sv;	914	return sv;
787		915
788	fail:	916	fail:
789	SvREFCNT_dec (sv);	917	SvREFCNT_dec (sv);
…		…
821	}	949	}
822	break;	950	break;
823		951
824	case CBOR_TAG_STRINGREF:	952	case CBOR_TAG_STRINGREF:
825	{	953	{
826	if ((*dec->cur >> 5) != 0)	954	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
827	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	955	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
828		956
829	UV idx = decode_uint (dec);	957	UV idx = decode_uint (dec);
830		958
831	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	959	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
838	case CBOR_TAG_VALUE_SHAREABLE:	966	case CBOR_TAG_VALUE_SHAREABLE:
839	{	967	{
840	if (ecb_expect_false (!dec->shareable))	968	if (ecb_expect_false (!dec->shareable))
841	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	969	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
842		970
		971	if (dec->cbor.flags & F_ALLOW_CYCLES)
		972	{
843	sv = newSV (0);	973	sv = newSV (0);
844	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	974	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
845		975
846	SV *osv = decode_sv (dec);	976	SV *osv = decode_sv (dec);
847	sv_setsv (sv, osv);	977	sv_setsv (sv, osv);
848	SvREFCNT_dec_NN (osv);	978	SvREFCNT_dec_NN (osv);
		979	}
		980	else
		981	{
		982	av_push (dec->shareable, &PL_sv_undef);
		983	int idx = AvFILLp (dec->shareable);
		984	sv = decode_sv (dec);
		985	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		986	}
849	}	987	}
850	break;	988	break;
851		989
852	case CBOR_TAG_VALUE_SHAREDREF:	990	case CBOR_TAG_VALUE_SHAREDREF:
853	{	991	{
854	if ((*dec->cur >> 5) != 0)	992	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
855	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	993	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
856		994
857	UV idx = decode_uint (dec);	995	UV idx = decode_uint (dec);
858		996
859	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	997	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
860	ERR ("corrupted CBOR data (sharedref index out of bounds)");	998	ERR ("corrupted CBOR data (sharedref index out of bounds)");
861		999
862	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1000	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1001
		1002	if (sv == &PL_sv_undef)
		1003	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
863	}	1004	}
864	break;	1005	break;
865		1006
866	case CBOR_TAG_PERL_OBJECT:	1007	case CBOR_TAG_PERL_OBJECT:
867	{	1008	{
…		…
882	if (!method)	1023	if (!method)
883	ERR ("cannot decode perl-object (package does not have a THAW method)");	1024	ERR ("cannot decode perl-object (package does not have a THAW method)");
884		1025
885	dSP;	1026	dSP;
886		1027
887	ENTER; SAVETMPS; PUSHMARK (SP);	1028	ENTER; SAVETMPS;
		1029	PUSHMARK (SP);
888	EXTEND (SP, len + 1);	1030	EXTEND (SP, len + 1);
889	// we re-bless the reference to get overload and other niceties right	1031	// we re-bless the reference to get overload and other niceties right
890	PUSHs (*av_fetch (av, 0, 1));	1032	PUSHs (*av_fetch (av, 0, 1));
891	PUSHs (sv_cbor);	1033	PUSHs (sv_cbor);
892		1034
…		…
917	default:	1059	default:
918	{	1060	{
919	sv = decode_sv (dec);	1061	sv = decode_sv (dec);
920		1062
921	dSP;	1063	dSP;
922	ENTER; SAVETMPS; PUSHMARK (SP);	1064	ENTER; SAVETMPS;
		1065	SAVESTACK_POS ();
		1066	PUSHMARK (SP);
923	EXTEND (SP, 2);	1067	EXTEND (SP, 2);
924	PUSHs (newSVuv (tag));	1068	PUSHs (newSVuv (tag));
925	PUSHs (sv);	1069	PUSHs (sv);
926		1070
927	PUTBACK;	1071	PUTBACK;
…		…
968	static SV *	1112	static SV *
969	decode_sv (dec_t *dec)	1113	decode_sv (dec_t *dec)
970	{	1114	{
971	WANT (1);	1115	WANT (1);
972		1116
973	switch (*dec->cur >> 5)	1117	switch (*dec->cur >> MAJOR_SHIFT)
974	{	1118	{
975	case 0: // unsigned int	1119	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
976	return newSVuv (decode_uint (dec));	1120	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
977	case 1: // negative int	1121	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
978	return newSViv (-1 - (IV)decode_uint (dec));	1122	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
979	case 2: // octet string	1123	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
980	return decode_str (dec, 0);	1124	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
981	case 3: // utf-8 string	1125	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
982	return decode_str (dec, 1);	1126
983	case 4: // array	1127	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)	1128	switch (*dec->cur++ & MINOR_MASK)
991	{	1129	{
992	case 20:	1130	case SIMPLE_FALSE:
993	#if CBOR_SLOW	1131	#if CBOR_SLOW
994	types_false = get_bool ("Types::Serialiser::false");	1132	types_false = get_bool ("Types::Serialiser::false");
995	#endif	1133	#endif
996	return newSVsv (types_false);	1134	return newSVsv (types_false);
997	case 21:	1135	case SIMPLE_TRUE:
998	#if CBOR_SLOW	1136	#if CBOR_SLOW
999	types_true = get_bool ("Types::Serialiser::true");	1137	types_true = get_bool ("Types::Serialiser::true");
1000	#endif	1138	#endif
1001	return newSVsv (types_true);	1139	return newSVsv (types_true);
1002	case 22:	1140	case SIMPLE_NULL:
1003	return newSVsv (&PL_sv_undef);	1141	return newSVsv (&PL_sv_undef);
1004	case 23:	1142	case SIMPLE_UNDEF:
1005	#if CBOR_SLOW	1143	#if CBOR_SLOW
1006	types_error = get_bool ("Types::Serialiser::error");	1144	types_error = get_bool ("Types::Serialiser::error");
1007	#endif	1145	#endif
1008	return newSVsv (types_error);	1146	return newSVsv (types_error);
1009		1147
1010	case 25:	1148	case MISC_FLOAT16:
1011	{	1149	{
1012	WANT (2);	1150	WANT (2);
1013		1151
1014	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1152	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1015	dec->cur += 2;	1153	dec->cur += 2;
1016		1154
1017	return newSVnv (ecb_binary16_to_float (fp));	1155	return newSVnv (ecb_binary16_to_float (fp));
1018	}	1156	}
1019		1157
1020	case 26:	1158	case MISC_FLOAT32:
1021	{	1159	{
1022	uint32_t fp;	1160	uint32_t fp;
1023	WANT (4);	1161	WANT (4);
1024	memcpy (&fp, dec->cur, 4);	1162	memcpy (&fp, dec->cur, 4);
1025	dec->cur += 4;	1163	dec->cur += 4;
…		…
1028	fp = ecb_bswap32 (fp);	1166	fp = ecb_bswap32 (fp);
1029		1167
1030	return newSVnv (ecb_binary32_to_float (fp));	1168	return newSVnv (ecb_binary32_to_float (fp));
1031	}	1169	}
1032		1170
1033	case 27:	1171	case MISC_FLOAT64:
1034	{	1172	{
1035	uint64_t fp;	1173	uint64_t fp;
1036	WANT (8);	1174	WANT (8);
1037	memcpy (&fp, dec->cur, 8);	1175	memcpy (&fp, dec->cur, 8);
1038	dec->cur += 8;	1176	dec->cur += 8;
…		…
1041	fp = ecb_bswap64 (fp);	1179	fp = ecb_bswap64 (fp);
1042		1180
1043	return newSVnv (ecb_binary64_to_double (fp));	1181	return newSVnv (ecb_binary64_to_double (fp));
1044	}	1182	}
1045		1183
1046	// 0..19 unassigned	1184	// 0..19 unassigned simple
1047	// 24 reserved + unassigned (reserved values are not encodable)	1185	// 24 reserved + unassigned simple (reserved values are not encodable)
		1186	// 28-30 unassigned misc
		1187	// 31 break code
1048	default:	1188	default:
1049	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1189	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1050	}	1190	}
1051		1191
1052	break;	1192	break;
1053	}	1193	}
1054		1194
…		…
1057	}	1197	}
1058		1198
1059	static SV *	1199	static SV *
1060	decode_cbor (SV string, CBOR cbor, char **offset_return)	1200	decode_cbor (SV string, CBOR cbor, char **offset_return)
1061	{	1201	{
1062	dec_t dec = { };	1202	dec_t dec = { 0 };
1063	SV *sv;	1203	SV *sv;
1064	STRLEN len;	1204	STRLEN len;
1065	char *data = SvPVbyte (string, len);	1205	char *data = SvPVbyte (string, len);
1066		1206
1067	if (len > cbor->max_size && cbor->max_size)	1207	if (len > cbor->max_size && cbor->max_size)
…		…
1081	if (dec.cur != dec.end && !dec.err)	1221	if (dec.cur != dec.end && !dec.err)
1082	dec.err = "garbage after CBOR object";	1222	dec.err = "garbage after CBOR object";
1083		1223
1084	if (dec.err)	1224	if (dec.err)
1085	{	1225	{
		1226	if (dec.shareable)
		1227	{
		1228	// need to break cyclic links, which whould all be in shareable
		1229	int i;
		1230	SV **svp;
		1231
		1232	for (i = av_len (dec.shareable) + 1; i--; )
		1233	if ((svp = av_fetch (dec.shareable, i, 0)))
		1234	sv_setsv (*svp, &PL_sv_undef);
		1235	}
		1236
1086	SvREFCNT_dec (sv);	1237	SvREFCNT_dec (sv);
1087	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1238	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1088	}	1239	}
1089		1240
1090	sv = sv_2mortal (sv);	1241	sv = sv_2mortal (sv);
1091		1242
1092	return sv;	1243	return sv;
1093	}	1244	}
1094		1245
		1246	/////////////////////////////////////////////////////////////////////////////
		1247	// incremental parser
		1248
		1249	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1250
		1251	// returns 0 for notyet, 1 for success or error
		1252	static int
		1253	incr_parse (CBOR self, SV cborstr)
		1254	{
		1255	STRLEN cur;
		1256	SvPV (cborstr, cur);
		1257
		1258	while (ecb_expect_true (self->incr_need <= cur))
		1259	{
		1260	// table of integer count bytes
		1261	static I8 incr_len[MINOR_MASK + 1] = {
		1262	0, 0, 0, 0, 0, 0, 0, 0,
		1263	0, 0, 0, 0, 0, 0, 0, 0,
		1264	0, 0, 0, 0, 0, 0, 0, 0,
		1265	1, 2, 4, 8,-1,-1,-1,-2
		1266	};
		1267
		1268	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1269	U8 m = *p & MINOR_MASK;
		1270	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1271	I8 ilen = incr_len[m];
		1272
		1273	self->incr_need = self->incr_pos + 1;
		1274
		1275	if (ecb_expect_false (ilen < 0))
		1276	{
		1277	if (m != MINOR_INDEF)
		1278	return 1; // error
		1279
		1280	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1281	{
		1282	if (count >= 0)
		1283	return 1; // error
		1284
		1285	count = 1;
		1286	}
		1287	else
		1288	{
		1289	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1290	count = -1;
		1291	}
		1292	}
		1293	else
		1294	{
		1295	self->incr_need += ilen;
		1296	if (ecb_expect_false (self->incr_need > cur))
		1297	return 0;
		1298
		1299	int major = *p >> MAJOR_SHIFT;
		1300
		1301	switch (major)
		1302	{
		1303	case MAJOR_TAG >> MAJOR_SHIFT:
		1304	++count; // tags merely prefix another value
		1305	break;
		1306
		1307	case MAJOR_BYTES >> MAJOR_SHIFT:
		1308	case MAJOR_TEXT >> MAJOR_SHIFT:
		1309	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1310	case MAJOR_MAP >> MAJOR_SHIFT:
		1311	{
		1312	UV len;
		1313
		1314	if (ecb_expect_false (ilen))
		1315	{
		1316	len = 0;
		1317
		1318	do {
		1319	len = (len << 8) \| *++p;
		1320	} while (--ilen);
		1321	}
		1322	else
		1323	len = m;
		1324
		1325	switch (major)
		1326	{
		1327	case MAJOR_BYTES >> MAJOR_SHIFT:
		1328	case MAJOR_TEXT >> MAJOR_SHIFT:
		1329	self->incr_need += len;
		1330	if (ecb_expect_false (self->incr_need > cur))
		1331	return 0;
		1332
		1333	break;
		1334
		1335	case MAJOR_MAP >> MAJOR_SHIFT:
		1336	len <<= 1;
		1337	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1338	if (len)
		1339	{
		1340	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1341	count = len + 1;
		1342	}
		1343	break;
		1344	}
		1345	}
		1346	}
		1347	}
		1348
		1349	self->incr_pos = self->incr_need;
		1350
		1351	if (count > 0)
		1352	{
		1353	while (!--count)
		1354	{
		1355	if (!AvFILLp (self->incr_count))
		1356	return 1; // done
		1357
		1358	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1359	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1360	}
		1361
		1362	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1363	}
		1364	}
		1365
		1366	return 0;
		1367	}
		1368
		1369
1095	/////////////////////////////////////////////////////////////////////////////	1370	/////////////////////////////////////////////////////////////////////////////
1096	// XS interface functions	1371	// XS interface functions
1097		1372
1098	MODULE = CBOR::XS PACKAGE = CBOR::XS	1373	MODULE = CBOR::XS PACKAGE = CBOR::XS
1099		1374
…		…
1139	void shrink (CBOR *self, int enable = 1)	1414	void shrink (CBOR *self, int enable = 1)
1140	ALIAS:	1415	ALIAS:
1141	shrink = F_SHRINK	1416	shrink = F_SHRINK
1142	allow_unknown = F_ALLOW_UNKNOWN	1417	allow_unknown = F_ALLOW_UNKNOWN
1143	allow_sharing = F_ALLOW_SHARING	1418	allow_sharing = F_ALLOW_SHARING
1144	allow_stringref = F_ALLOW_STRINGREF	1419	allow_cycles = F_ALLOW_CYCLES
		1420	pack_strings = F_PACK_STRINGS
		1421	utf8_strings = F_UTF8_STRINGS
		1422	validate_utf8 = F_VALIDATE_UTF8
1145	PPCODE:	1423	PPCODE:
1146	{	1424	{
1147	if (enable)	1425	if (enable)
1148	self->flags \|= ix;	1426	self->flags \|= ix;
1149	else	1427	else
…		…
1155	void get_shrink (CBOR *self)	1433	void get_shrink (CBOR *self)
1156	ALIAS:	1434	ALIAS:
1157	get_shrink = F_SHRINK	1435	get_shrink = F_SHRINK
1158	get_allow_unknown = F_ALLOW_UNKNOWN	1436	get_allow_unknown = F_ALLOW_UNKNOWN
1159	get_allow_sharing = F_ALLOW_SHARING	1437	get_allow_sharing = F_ALLOW_SHARING
1160	get_allow_stringref = F_ALLOW_STRINGREF	1438	get_allow_cycles = F_ALLOW_CYCLES
		1439	get_pack_strings = F_PACK_STRINGS
		1440	get_validate_utf8 = F_VALIDATE_UTF8
1161	PPCODE:	1441	PPCODE:
1162	XPUSHs (boolSV (self->flags & ix));	1442	XPUSHs (boolSV (self->flags & ix));
1163		1443
1164	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1444	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1165	PPCODE:	1445	PPCODE:
…		…
1214	EXTEND (SP, 2);	1494	EXTEND (SP, 2);
1215	PUSHs (sv);	1495	PUSHs (sv);
1216	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1496	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1217	}	1497	}
1218		1498
		1499	void incr_parse (CBOR self, SV cborstr)
		1500	ALIAS:
		1501	incr_parse_multiple = 1
		1502	PPCODE:
		1503	{
		1504	if (SvUTF8 (cborstr))
		1505	sv_utf8_downgrade (cborstr, 0);
		1506
		1507	if (!self->incr_count)
		1508	{
		1509	self->incr_count = newAV ();
		1510	self->incr_pos = 0;
		1511	self->incr_need = 1;
		1512
		1513	av_push (self->incr_count, newSViv (1));
		1514	}
		1515
		1516	do
		1517	{
		1518	if (!incr_parse (self, cborstr))
		1519	{
		1520	if (self->incr_need > self->max_size && self->max_size)
		1521	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1522	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1523
		1524	break;
		1525	}
		1526
		1527	SV *sv;
		1528	char *offset;
		1529
		1530	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1531	XPUSHs (sv);
		1532
		1533	sv_chop (cborstr, offset);
		1534
		1535	av_clear (self->incr_count);
		1536	av_push (self->incr_count, newSViv (1));
		1537
		1538	self->incr_pos = 0;
		1539	self->incr_need = self->incr_pos + 1;
		1540	}
		1541	while (ix);
		1542	}
		1543
		1544	void incr_reset (CBOR *self)
		1545	CODE:
		1546	{
		1547	SvREFCNT_dec (self->incr_count);
		1548	self->incr_count = 0;
		1549	}
		1550
1219	void DESTROY (CBOR *self)	1551	void DESTROY (CBOR *self)
1220	PPCODE:	1552	PPCODE:
1221	cbor_free (self);	1553	cbor_free (self);
1222		1554
1223	PROTOTYPES: ENABLE	1555	PROTOTYPES: ENABLE
1224		1556
1225	void encode_cbor (SV *scalar)	1557	void encode_cbor (SV *scalar)
		1558	ALIAS:
		1559	encode_cbor = 0
		1560	encode_cbor_sharing = F_ALLOW_SHARING
1226	PPCODE:	1561	PPCODE:
1227	{	1562	{
1228	CBOR cbor;	1563	CBOR cbor;
1229	cbor_init (&cbor);	1564	cbor_init (&cbor);
		1565	cbor.flags \|= ix;
1230	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1566	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1231	XPUSHs (scalar);	1567	XPUSHs (scalar);
1232	}	1568	}
1233		1569
1234	void decode_cbor (SV *cborstr)	1570	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.29 by root, Thu Nov 28 09:13:12 2013 UTC vs. Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.29 by root, Thu Nov 28 09:13:12 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC