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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.33 by root, Sat Nov 30 15:23:59 2013 UTC vs.
Revision 1.58 by root, Fri Nov 25 13:27:29 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	102	#define F_ALLOW_SHARING 0x00000004UL
		103	#define F_ALLOW_CYCLES 0x00000008UL
65	#define F_PACK_STRINGS 0x00000008UL	104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_TEXT_KEYS 0x00000020UL
		106	#define F_TEXT_STRINGS 0x00000040UL
		107	#define F_VALIDATE_UTF8 0x00000080UL
66		108
67	#define INIT_SIZE 32 // initial scalar size to be allocated	109	#define INIT_SIZE 32 // initial scalar size to be allocated
68		110
69	#define SB do {	111	#define SB do {
70	#define SE } while (0)	112	#define SE } while (0)
…		…
89	typedef struct {	131	typedef struct {
90	U32 flags;	132	U32 flags;
91	U32 max_depth;	133	U32 max_depth;
92	STRLEN max_size;	134	STRLEN max_size;
93	SV *filter;	135	SV *filter;
		136
		137	// for the incremental parser
		138	STRLEN incr_pos; // the current offset into the text
		139	STRLEN incr_need; // minimum bytes needed to decode
		140	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
94	} CBOR;	141	} CBOR;
95		142
96	ecb_inline void	143	ecb_inline void
97	cbor_init (CBOR *cbor)	144	cbor_init (CBOR *cbor)
98	{	145	{
…		…
102		149
103	ecb_inline void	150	ecb_inline void
104	cbor_free (CBOR *cbor)	151	cbor_free (CBOR *cbor)
105	{	152	{
106	SvREFCNT_dec (cbor->filter);	153	SvREFCNT_dec (cbor->filter);
		154	SvREFCNT_dec (cbor->incr_count);
107	}	155	}
108		156
109	/////////////////////////////////////////////////////////////////////////////	157	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	158	// utility functions
111		159
…		…
189	static void	237	static void
190	encode_uint (enc_t *enc, int major, UV len)	238	encode_uint (enc_t *enc, int major, UV len)
191	{	239	{
192	need (enc, 9);	240	need (enc, 9);
193		241
194	if (ecb_expect_true (len < 24))	242	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major \| len;	243	*enc->cur++ = major \| len;
196	else if (ecb_expect_true (len <= 0xff))	244	else if (ecb_expect_true (len <= 0xffU))
197	{	245	{
198	*enc->cur++ = major \| 24;	246	*enc->cur++ = major \| LENGTH_EXT1;
199	*enc->cur++ = len;	247	*enc->cur++ = len;
200	}	248	}
201	else if (len <= 0xffff)	249	else if (len <= 0xffffU)
202	{	250	{
203	*enc->cur++ = major \| 25;	251	*enc->cur++ = major \| LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	252	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	253	*enc->cur++ = len;
206	}	254	}
207	else if (len <= 0xffffffff)	255	else if (len <= 0xffffffffU)
208	{	256	{
209	*enc->cur++ = major \| 26;	257	*enc->cur++ = major \| LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	258	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	259	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	260	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	261	*enc->cur++ = len;
214	}	262	}
215	else	263	else
216	{	264	{
217	*enc->cur++ = major \| 27;	265	*enc->cur++ = major \| LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	266	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	267	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	268	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	269	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	270	*enc->cur++ = len >> 24;
…		…
227	}	275	}
228		276
229	ecb_inline void	277	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	278	encode_tag (enc_t *enc, UV tag)
231	{	279	{
232	encode_uint (enc, 0xc0, tag);	280	encode_uint (enc, MAJOR_TAG, tag);
		281	}
		282
		283	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		284	ecb_noinline static void
		285	encode_str_utf8 (enc_t enc, int utf8, char str, STRLEN len)
		286	{
		287	STRLEN ulen = len;
		288	U8 p, pend = (U8 *)str + len;
		289
		290	for (p = (U8 *)str; p < pend; ++p)
		291	ulen += *p >> 7; // count set high bits
		292
		293	encode_uint (enc, MAJOR_TEXT, ulen);
		294
		295	need (enc, ulen);
		296	for (p = (U8 *)str; p < pend; ++p)
		297	if (*p < 0x80)
		298	enc->cur++ = p;
		299	else
		300	{
		301	enc->cur++ = 0xc0 + (p >> 6);
		302	enc->cur++ = 0x80 + (p & 63);
		303	}
233	}	304	}
234		305
235	ecb_inline void	306	ecb_inline void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	307	encode_str (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
237	{	308	{
238	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	309	if (ecb_expect_false (upgrade_utf8))
		310	if (!utf8)
		311	{
		312	encode_str_utf8 (enc, utf8, str, len);
		313	return;
		314	}
		315
		316	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
239	need (enc, len);	317	need (enc, len);
240	memcpy (enc->cur, str, len);	318	memcpy (enc->cur, str, len);
241	enc->cur += len;	319	enc->cur += len;
242	}	320	}
243		321
244	static void	322	ecb_inline void
245	encode_strref (enc_t enc, int utf8, char str, STRLEN len)	323	encode_strref (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
246	{	324	{
247	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))	325	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
248	{	326	{
249	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	327	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
250		328
251	if (SvOK (*svp))	329	if (SvOK (*svp))
252	{	330	{
253	// already registered, use stringref	331	// already registered, use stringref
254	encode_tag (enc, CBOR_TAG_STRINGREF);	332	encode_tag (enc, CBOR_TAG_STRINGREF);
255	encode_uint (enc, 0x00, SvUV (*svp));	333	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
256	return;	334	return;
257	}	335	}
258	else if (len >= minimum_string_length (enc->stringref_idx))	336	else if (len >= minimum_string_length (enc->stringref_idx))
259	{	337	{
260	// register only	338	// register only
261	sv_setuv (*svp, enc->stringref_idx);	339	sv_setuv (*svp, enc->stringref_idx);
262	++enc->stringref_idx;	340	++enc->stringref_idx;
263	}	341	}
264	}	342	}
265		343
266	encode_str (enc, utf8, str, len);	344	encode_str (enc, upgrade_utf8, utf8, str, len);
267	}	345	}
268		346
269	static void encode_sv (enc_t enc, SV sv);	347	static void encode_sv (enc_t enc, SV sv);
270		348
271	static void	349	static void
…		…
276	if (enc->depth >= enc->cbor.max_depth)	354	if (enc->depth >= enc->cbor.max_depth)
277	croak (ERR_NESTING_EXCEEDED);	355	croak (ERR_NESTING_EXCEEDED);
278		356
279	++enc->depth;	357	++enc->depth;
280		358
281	encode_uint (enc, 0x80, len + 1);	359	encode_uint (enc, MAJOR_ARRAY, len + 1);
282		360
		361	if (SvMAGICAL (av))
283	for (i = 0; i <= len; ++i)	362	for (i = 0; i <= len; ++i)
284	{	363	{
285	SV **svp = av_fetch (av, i, 0);	364	SV **svp = av_fetch (av, i, 0);
286	encode_sv (enc, svp ? *svp : &PL_sv_undef);	365	encode_sv (enc, svp ? *svp : &PL_sv_undef);
287	}	366	}
		367	else
		368	for (i = 0; i <= len; ++i)
		369	{
		370	SV *sv = AvARRAY (av)[i];
		371	encode_sv (enc, sv ? sv : &PL_sv_undef);
		372	}
288		373
289	--enc->depth;	374	--enc->depth;
290	}	375	}
291		376
292	static void	377	static void
…		…
301		386
302	int pairs = hv_iterinit (hv);	387	int pairs = hv_iterinit (hv);
303	int mg = SvMAGICAL (hv);	388	int mg = SvMAGICAL (hv);
304		389
305	if (mg)	390	if (mg)
306	encode_ch (enc, 0xa0 \| 31);	391	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
307	else	392	else
308	encode_uint (enc, 0xa0, pairs);	393	encode_uint (enc, MAJOR_MAP, pairs);
309		394
310	while ((he = hv_iternext (hv)))	395	while ((he = hv_iternext (hv)))
311	{	396	{
312	if (HeKLEN (he) == HEf_SVKEY)	397	if (HeKLEN (he) == HEf_SVKEY)
313	encode_sv (enc, HeSVKEY (he));	398	encode_sv (enc, HeSVKEY (he));
314	else	399	else
315	encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));	400	encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS \| F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
316		401
317	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	402	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
318	}	403	}
319		404
320	if (mg)	405	if (mg)
321	encode_ch (enc, 0xe0 \| 31);	406	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
322		407
323	--enc->depth;	408	--enc->depth;
324	}	409	}
325		410
326	// encode objects, arrays and special \0=false and \1=true values.	411	// encode objects, arrays and special \0=false and \1=true values.
…		…
342	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	427	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
343	? cbor_tagged_stash	428	? cbor_tagged_stash
344	: gv_stashpv ("CBOR::XS::Tagged" , 1);	429	: gv_stashpv ("CBOR::XS::Tagged" , 1);
345		430
346	HV *stash = SvSTASH (sv);	431	HV *stash = SvSTASH (sv);
347	GV *method;
348		432
349	if (stash == boolean_stash)	433	if (stash == boolean_stash)
350	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	434	{
		435	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
		436	return;
		437	}
351	else if (stash == error_stash)	438	else if (stash == error_stash)
352	encode_ch (enc, 0xe0 \| 23);	439	{
		440	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		441	return;
		442	}
353	else if (stash == tagged_stash)	443	else if (stash == tagged_stash)
354	{	444	{
355	if (svt != SVt_PVAV)	445	if (svt != SVt_PVAV)
356	croak ("encountered CBOR::XS::Tagged object that isn't an array");	446	croak ("encountered CBOR::XS::Tagged object that isn't an array");
357		447
358	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	448	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
359	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	449	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
		450
		451	return;
		452	}
		453	}
		454
		455	if (ecb_expect_false (SvREFCNT (sv) > 1)
		456	&& ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
		457	{
		458	if (!enc->shareable)
		459	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
		460
		461	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
		462
		463	if (SvOK (*svp))
360	}	464	{
		465	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
		466	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
		467	return;
		468	}
		469	else
		470	{
		471	sv_setuv (*svp, enc->shareable_idx);
		472	++enc->shareable_idx;
		473	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
		474	}
		475	}
		476
		477	if (ecb_expect_false (SvOBJECT (sv)))
		478	{
		479	HV *stash = SvSTASH (sv);
		480	GV *method;
		481
361	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	482	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
362	{	483	{
363	dSP;	484	dSP;
364		485
365	ENTER; SAVETMPS; PUSHMARK (SP);	486	ENTER; SAVETMPS;
		487	PUSHMARK (SP);
366	// we re-bless the reference to get overload and other niceties right	488	// we re-bless the reference to get overload and other niceties right
367	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	489	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
368		490
369	PUTBACK;	491	PUTBACK;
370	// G_SCALAR ensures that return value is 1	492	// G_SCALAR ensures that return value is 1
…		…
383	}	505	}
384	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	506	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
385	{	507	{
386	dSP;	508	dSP;
387		509
388	ENTER; SAVETMPS; PUSHMARK (SP);	510	ENTER; SAVETMPS;
		511	SAVESTACK_POS ();
		512	PUSHMARK (SP);
389	EXTEND (SP, 2);	513	EXTEND (SP, 2);
390	// we re-bless the reference to get overload and other niceties right	514	// we re-bless the reference to get overload and other niceties right
391	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	515	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
392	PUSHs (sv_cbor);	516	PUSHs (sv_cbor);
393		517
…		…
398	// catch this surprisingly common error	522	// catch this surprisingly common error
399	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	523	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
400	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));	524	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
401		525
402	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	526	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
403	encode_uint (enc, 0x80, count + 1);	527	encode_uint (enc, MAJOR_ARRAY, count + 1);
404	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	528	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
405		529
406	while (count)	530	while (count)
407	encode_sv (enc, SP[1 - count--]);	531	encode_sv (enc, SP[1 - count--]);
408		532
409	PUTBACK;	533	PUTBACK;
…		…
418	encode_hv (enc, (HV *)sv);	542	encode_hv (enc, (HV *)sv);
419	else if (svt == SVt_PVAV)	543	else if (svt == SVt_PVAV)
420	encode_av (enc, (AV *)sv);	544	encode_av (enc, (AV *)sv);
421	else	545	else
422	{	546	{
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))
432	{
433	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
434	encode_uint (enc, 0x00, 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	encode_tag (enc, CBOR_TAG_INDIRECTION);	547	encode_tag (enc, CBOR_TAG_INDIRECTION);
446	encode_sv (enc, sv);	548	encode_sv (enc, sv);
447	}	549	}
448	}	550	}
449		551
…		…
452	{	554	{
453	double nv = SvNVX (sv);	555	double nv = SvNVX (sv);
454		556
455	need (enc, 9);	557	need (enc, 9);
456		558
457	if (ecb_expect_false (nv == (U32)nv))	559	if (ecb_expect_false (nv == (NV)(U32)nv))
458	encode_uint (enc, 0x00, (U32)nv);	560	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
459	//TODO: maybe I32?	561	//TODO: maybe I32?
460	else if (ecb_expect_false (nv == (float)nv))	562	else if (ecb_expect_false (nv == (float)nv))
461	{	563	{
462	uint32_t fp = ecb_float_to_binary32 (nv);	564	uint32_t fp = ecb_float_to_binary32 (nv);
463		565
464	*enc->cur++ = 0xe0 \| 26;	566	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
465		567
466	if (!ecb_big_endian ())	568	if (!ecb_big_endian ())
467	fp = ecb_bswap32 (fp);	569	fp = ecb_bswap32 (fp);
468		570
469	memcpy (enc->cur, &fp, 4);	571	memcpy (enc->cur, &fp, 4);
…		…
471	}	573	}
472	else	574	else
473	{	575	{
474	uint64_t fp = ecb_double_to_binary64 (nv);	576	uint64_t fp = ecb_double_to_binary64 (nv);
475		577
476	*enc->cur++ = 0xe0 \| 27;	578	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
477		579
478	if (!ecb_big_endian ())	580	if (!ecb_big_endian ())
479	fp = ecb_bswap64 (fp);	581	fp = ecb_bswap64 (fp);
480		582
481	memcpy (enc->cur, &fp, 8);	583	memcpy (enc->cur, &fp, 8);
…		…
490		592
491	if (SvPOKp (sv))	593	if (SvPOKp (sv))
492	{	594	{
493	STRLEN len;	595	STRLEN len;
494	char *str = SvPV (sv, len);	596	char *str = SvPV (sv, len);
495	encode_strref (enc, SvUTF8 (sv), str, len);	597	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
496	}	598	}
497	else if (SvNOKp (sv))	599	else if (SvNOKp (sv))
498	encode_nv (enc, sv);	600	encode_nv (enc, sv);
499	else if (SvIOKp (sv))	601	else if (SvIOKp (sv))
500	{	602	{
501	if (SvIsUV (sv))	603	if (SvIsUV (sv))
502	encode_uint (enc, 0x00, SvUVX (sv));	604	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
503	else if (SvIVX (sv) >= 0)	605	else if (SvIVX (sv) >= 0)
504	encode_uint (enc, 0x00, SvIVX (sv));	606	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
505	else	607	else
506	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	608	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
507	}	609	}
508	else if (SvROK (sv))	610	else if (SvROK (sv))
509	encode_rv (enc, SvRV (sv));	611	encode_rv (enc, SvRV (sv));
510	else if (!SvOK (sv))	612	else if (!SvOK (sv))
511	encode_ch (enc, 0xe0 \| 22);	613	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
512	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	614	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
513	encode_ch (enc, 0xe0 \| 23);	615	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
514	else	616	else
515	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	617	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
516	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	618	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
517	}	619	}
518		620
519	static SV *	621	static SV *
520	encode_cbor (SV scalar, CBOR cbor)	622	encode_cbor (SV scalar, CBOR cbor)
521	{	623	{
522	enc_t enc = { };	624	enc_t enc = { 0 };
523		625
524	enc.cbor = *cbor;	626	enc.cbor = *cbor;
525	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	627	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
526	enc.cur = SvPVX (enc.sv);	628	enc.cur = SvPVX (enc.sv);
527	enc.end = SvEND (enc.sv);	629	enc.end = SvEND (enc.sv);
…		…
563	SV *decode_tagged;	665	SV *decode_tagged;
564	} dec_t;	666	} dec_t;
565		667
566	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	668	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
567		669
568	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	670	#define WANT(len) if (ecb_expect_false ((UV)(dec->end - dec->cur) < (UV)len)) ERR ("unexpected end of CBOR data")
569		671
570	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	672	#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
571	#define DEC_DEC_DEPTH --dec->depth	673	#define DEC_DEC_DEPTH --dec->depth
572		674
573	static UV	675	static UV
574	decode_uint (dec_t *dec)	676	decode_uint (dec_t *dec)
575	{	677	{
576	switch (*dec->cur & 31)	678	U8 m = *dec->cur & MINOR_MASK;
577	{	679	++dec->cur;
578	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
579	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
580	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
581	return *dec->cur++ & 31;
582		680
583	case 24:	681	if (ecb_expect_true (m < LENGTH_EXT1))
		682	return m;
		683	else if (ecb_expect_true (m == LENGTH_EXT1))
		684	{
584	WANT (2);	685	WANT (1);
585	dec->cur += 2;	686	dec->cur += 1;
586	return dec->cur[-1];	687	return dec->cur[-1];
587		688	}
588	case 25:	689	else if (ecb_expect_true (m == LENGTH_EXT2))
		690	{
589	WANT (3);	691	WANT (2);
590	dec->cur += 3;	692	dec->cur += 2;
591	return (((UV)dec->cur[-2]) << 8)	693	return (((UV)dec->cur[-2]) << 8)
592	\| ((UV)dec->cur[-1]);	694	\| ((UV)dec->cur[-1]);
593		695	}
594	case 26:	696	else if (ecb_expect_true (m == LENGTH_EXT4))
		697	{
595	WANT (5);	698	WANT (4);
596	dec->cur += 5;	699	dec->cur += 4;
597	return (((UV)dec->cur[-4]) << 24)	700	return (((UV)dec->cur[-4]) << 24)
598	\| (((UV)dec->cur[-3]) << 16)	701	\| (((UV)dec->cur[-3]) << 16)
599	\| (((UV)dec->cur[-2]) << 8)	702	\| (((UV)dec->cur[-2]) << 8)
600	\| ((UV)dec->cur[-1]);	703	\| ((UV)dec->cur[-1]);
601		704	}
602	case 27:	705	else if (ecb_expect_true (m == LENGTH_EXT8))
		706	{
603	WANT (9);	707	WANT (8);
604	dec->cur += 9;	708	dec->cur += 8;
		709
		710	return
		711	#if UVSIZE < 8
		712	0
		713	#else
605	return (((UV)dec->cur[-8]) << 56)	714	(((UV)dec->cur[-8]) << 56)
606	\| (((UV)dec->cur[-7]) << 48)	715	\| (((UV)dec->cur[-7]) << 48)
607	\| (((UV)dec->cur[-6]) << 40)	716	\| (((UV)dec->cur[-6]) << 40)
608	\| (((UV)dec->cur[-5]) << 32)	717	\| (((UV)dec->cur[-5]) << 32)
		718	#endif
609	\| (((UV)dec->cur[-4]) << 24)	719	\| (((UV)dec->cur[-4]) << 24)
610	\| (((UV)dec->cur[-3]) << 16)	720	\| (((UV)dec->cur[-3]) << 16)
611	\| (((UV)dec->cur[-2]) << 8)	721	\| (((UV)dec->cur[-2]) << 8)
612	\| ((UV)dec->cur[-1]);	722	\| ((UV)dec->cur[-1]);
613		723	}
614	default:	724	else
615	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	725	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
616	}
617		726
618	fail:	727	fail:
619	return 0;	728	return 0;
620	}	729	}
621		730
…		…
626	{	735	{
627	AV *av = newAV ();	736	AV *av = newAV ();
628		737
629	DEC_INC_DEPTH;	738	DEC_INC_DEPTH;
630		739
631	if ((*dec->cur & 31) == 31)	740	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
632	{	741	{
633	++dec->cur;	742	++dec->cur;
634		743
635	for (;;)	744	for (;;)
636	{	745	{
637	WANT (1);	746	WANT (1);
638		747
639	if (*dec->cur == (0xe0 \| 31))	748	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
640	{	749	{
641	++dec->cur;	750	++dec->cur;
642	break;	751	break;
643	}	752	}
644		753
645	av_push (av, decode_sv (dec));	754	av_push (av, decode_sv (dec));
646	}	755	}
647	}	756	}
648	else	757	else
649	{	758	{
650	int i, len = decode_uint (dec);	759	UV i, len = decode_uint (dec);
651		760
		761	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
652	av_fill (av, len - 1);	762	av_fill (av, len - 1);
653		763
654	for (i = 0; i < len; ++i)	764	for (i = 0; i < len; ++i)
655	AvARRAY (av)[i] = decode_sv (dec);	765	AvARRAY (av)[i] = decode_sv (dec);
656	}	766	}
…		…
669	{	779	{
670	// for speed reasons, we specialcase single-string	780	// for speed reasons, we specialcase single-string
671	// byte or utf-8 strings as keys, but only when !stringref	781	// byte or utf-8 strings as keys, but only when !stringref
672		782
673	if (ecb_expect_true (!dec->stringref))	783	if (ecb_expect_true (!dec->stringref))
674	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	784	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
675	{	785	{
676	I32 len = decode_uint (dec);	786	STRLEN len = decode_uint (dec);
677	char key = (char )dec->cur;	787	char key = (char )dec->cur;
678		788
		789	WANT (len);
679	dec->cur += len;	790	dec->cur += len;
680		791
681	if (ecb_expect_false (dec->stringref))
682	av_push (dec->stringref, newSVpvn (key, len));
683
684	hv_store (hv, key, len, decode_sv (dec), 0);	792	hv_store (hv, key, len, decode_sv (dec), 0);
685		793
686	return;	794	return;
687	}	795	}
688	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	796	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
689	{	797	{
690	I32 len = decode_uint (dec);	798	STRLEN len = decode_uint (dec);
691	char key = (char )dec->cur;	799	char key = (char )dec->cur;
692		800
		801	WANT (len);
693	dec->cur += len;	802	dec->cur += len;
694		803
695	if (ecb_expect_false (dec->stringref))	804	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
696	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	805	if (!is_utf8_string (key, len))
		806	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
697		807
698	hv_store (hv, key, -len, decode_sv (dec), 0);	808	hv_store (hv, key, -len, decode_sv (dec), 0);
699		809
700	return;	810	return;
701	}	811	}
…		…
703	SV *k = decode_sv (dec);	813	SV *k = decode_sv (dec);
704	SV *v = decode_sv (dec);	814	SV *v = decode_sv (dec);
705		815
706	hv_store_ent (hv, k, v, 0);	816	hv_store_ent (hv, k, v, 0);
707	SvREFCNT_dec (k);	817	SvREFCNT_dec (k);
		818
		819	fail:
		820	;
708	}	821	}
709		822
710	static SV *	823	static SV *
711	decode_hv (dec_t *dec)	824	decode_hv (dec_t *dec)
712	{	825	{
713	HV *hv = newHV ();	826	HV *hv = newHV ();
714		827
715	DEC_INC_DEPTH;	828	DEC_INC_DEPTH;
716		829
717	if ((*dec->cur & 31) == 31)	830	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
718	{	831	{
719	++dec->cur;	832	++dec->cur;
720		833
721	for (;;)	834	for (;;)
722	{	835	{
723	WANT (1);	836	WANT (1);
724		837
725	if (*dec->cur == (0xe0 \| 31))	838	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
726	{	839	{
727	++dec->cur;	840	++dec->cur;
728	break;	841	break;
729	}	842	}
730		843
731	decode_he (dec, hv);	844	decode_he (dec, hv);
732	}	845	}
733	}	846	}
734	else	847	else
735	{	848	{
736	int pairs = decode_uint (dec);	849	UV pairs = decode_uint (dec);
		850
		851	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
737		852
738	while (pairs--)	853	while (pairs--)
739	decode_he (dec, hv);	854	decode_he (dec, hv);
740	}	855	}
741		856
…		…
751	static SV *	866	static SV *
752	decode_str (dec_t *dec, int utf8)	867	decode_str (dec_t *dec, int utf8)
753	{	868	{
754	SV *sv = 0;	869	SV *sv = 0;
755		870
756	if ((*dec->cur & 31) == 31)	871	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
757	{	872	{
758	// indefinite length strings	873	// indefinite length strings
759	++dec->cur;	874	++dec->cur;
760		875
761	unsigned char major = *dec->cur & 0xe0;	876	U8 major = *dec->cur & MAJOR_MISC;
762		877
763	sv = newSVpvn ("", 0);	878	sv = newSVpvn ("", 0);
764		879
765	for (;;)	880	for (;;)
766	{	881	{
767	WANT (1);	882	WANT (1);
768		883
769	if ((*dec->cur ^ major) >= 31)	884	if ((*dec->cur - major) > LENGTH_EXT8)
770	if (*dec->cur == (0xe0 \| 31))	885	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
771	{	886	{
772	++dec->cur;	887	++dec->cur;
773	break;	888	break;
774	}	889	}
775	else	890	else
…		…
794	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	909	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
795	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	910	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
796	}	911	}
797		912
798	if (utf8)	913	if (utf8)
		914	{
		915	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		916	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		917	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		918
799	SvUTF8_on (sv);	919	SvUTF8_on (sv);
		920	}
800		921
801	return sv;	922	return sv;
802		923
803	fail:	924	fail:
804	SvREFCNT_dec (sv);	925	SvREFCNT_dec (sv);
…		…
823	sv = newRV_noinc (decode_sv (dec));	944	sv = newRV_noinc (decode_sv (dec));
824	break;	945	break;
825		946
826	case CBOR_TAG_STRINGREF_NAMESPACE:	947	case CBOR_TAG_STRINGREF_NAMESPACE:
827	{	948	{
		949	// do nmot use SAVETMPS/FREETMPS, as these will
		950	// erase mortalised caches, e.g. "shareable"
828	ENTER; SAVETMPS;	951	ENTER;
829		952
830	SAVESPTR (dec->stringref);	953	SAVESPTR (dec->stringref);
831	dec->stringref = (AV )sv_2mortal ((SV )newAV ());	954	dec->stringref = (AV )sv_2mortal ((SV )newAV ());
832		955
833	sv = decode_sv (dec);	956	sv = decode_sv (dec);
834		957
835	FREETMPS; LEAVE;	958	LEAVE;
836	}	959	}
837	break;	960	break;
838		961
839	case CBOR_TAG_STRINGREF:	962	case CBOR_TAG_STRINGREF:
840	{	963	{
841	if ((*dec->cur >> 5) != 0)	964	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
842	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	965	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
843		966
844	UV idx = decode_uint (dec);	967	UV idx = decode_uint (dec);
845		968
846	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	969	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
853	case CBOR_TAG_VALUE_SHAREABLE:	976	case CBOR_TAG_VALUE_SHAREABLE:
854	{	977	{
855	if (ecb_expect_false (!dec->shareable))	978	if (ecb_expect_false (!dec->shareable))
856	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	979	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
857		980
		981	if (dec->cbor.flags & F_ALLOW_CYCLES)
		982	{
858	sv = newSV (0);	983	sv = newSV (0);
859	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	984	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
860		985
861	SV *osv = decode_sv (dec);	986	SV *osv = decode_sv (dec);
862	sv_setsv (sv, osv);	987	sv_setsv (sv, osv);
863	SvREFCNT_dec_NN (osv);	988	SvREFCNT_dec_NN (osv);
		989	}
		990	else
		991	{
		992	av_push (dec->shareable, &PL_sv_undef);
		993	int idx = AvFILLp (dec->shareable);
		994	sv = decode_sv (dec);
		995	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		996	}
864	}	997	}
865	break;	998	break;
866		999
867	case CBOR_TAG_VALUE_SHAREDREF:	1000	case CBOR_TAG_VALUE_SHAREDREF:
868	{	1001	{
869	if ((*dec->cur >> 5) != 0)	1002	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
870	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1003	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
871		1004
872	UV idx = decode_uint (dec);	1005	UV idx = decode_uint (dec);
873		1006
874	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	1007	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
875	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1008	ERR ("corrupted CBOR data (sharedref index out of bounds)");
876		1009
877	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1010	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1011
		1012	if (sv == &PL_sv_undef)
		1013	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
878	}	1014	}
879	break;	1015	break;
880		1016
881	case CBOR_TAG_PERL_OBJECT:	1017	case CBOR_TAG_PERL_OBJECT:
882	{	1018	{
…		…
897	if (!method)	1033	if (!method)
898	ERR ("cannot decode perl-object (package does not have a THAW method)");	1034	ERR ("cannot decode perl-object (package does not have a THAW method)");
899		1035
900	dSP;	1036	dSP;
901		1037
902	ENTER; SAVETMPS; PUSHMARK (SP);	1038	ENTER; SAVETMPS;
		1039	PUSHMARK (SP);
903	EXTEND (SP, len + 1);	1040	EXTEND (SP, len + 1);
904	// we re-bless the reference to get overload and other niceties right	1041	// we re-bless the reference to get overload and other niceties right
905	PUSHs (*av_fetch (av, 0, 1));	1042	PUSHs (*av_fetch (av, 0, 1));
906	PUSHs (sv_cbor);	1043	PUSHs (sv_cbor);
907		1044
…		…
929	}	1066	}
930	break;	1067	break;
931		1068
932	default:	1069	default:
933	{	1070	{
		1071	SV *tag_sv = newSVuv (tag);
		1072
934	sv = decode_sv (dec);	1073	sv = decode_sv (dec);
935		1074
936	dSP;	1075	dSP;
937	ENTER; SAVETMPS; PUSHMARK (SP);	1076	ENTER; SAVETMPS;
		1077	SAVESTACK_POS ();
		1078	PUSHMARK (SP);
938	EXTEND (SP, 2);	1079	EXTEND (SP, 2);
939	PUSHs (newSVuv (tag));	1080	PUSHs (tag_sv);
940	PUSHs (sv);	1081	PUSHs (sv);
941		1082
942	PUTBACK;	1083	PUTBACK;
943	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);	1084	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
944	SPAGAIN;	1085	SPAGAIN;
945		1086
946	if (SvTRUE (ERRSV))	1087	if (SvTRUE (ERRSV))
947	{	1088	{
		1089	SvREFCNT_dec (tag_sv);
948	FREETMPS; LEAVE;	1090	FREETMPS; LEAVE;
949	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1091	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
950	}	1092	}
951		1093
952	if (count)	1094	if (count)
953	{	1095	{
		1096	SvREFCNT_dec (tag_sv);
954	SvREFCNT_dec (sv);	1097	SvREFCNT_dec (sv);
955	sv = SvREFCNT_inc (POPs);	1098	sv = SvREFCNT_inc (POPs);
956	}	1099	}
957	else	1100	else
958	{	1101	{
959	AV *av = newAV ();	1102	AV *av = newAV ();
960	av_push (av, newSVuv (tag));	1103	av_push (av, tag_sv);
961	av_push (av, sv);	1104	av_push (av, sv);
962		1105
963	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1106	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
964	? cbor_tagged_stash	1107	? cbor_tagged_stash
965	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1108	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
983	static SV *	1126	static SV *
984	decode_sv (dec_t *dec)	1127	decode_sv (dec_t *dec)
985	{	1128	{
986	WANT (1);	1129	WANT (1);
987		1130
988	switch (*dec->cur >> 5)	1131	switch (*dec->cur >> MAJOR_SHIFT)
989	{	1132	{
990	case 0: // unsigned int	1133	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
991	return newSVuv (decode_uint (dec));	1134	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
992	case 1: // negative int	1135	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
993	return newSViv (-1 - (IV)decode_uint (dec));	1136	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
994	case 2: // octet string	1137	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
995	return decode_str (dec, 0);	1138	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
996	case 3: // utf-8 string	1139	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
997	return decode_str (dec, 1);	1140
998	case 4: // array	1141	case MAJOR_MISC >> MAJOR_SHIFT:
999	return decode_av (dec);
1000	case 5: // map
1001	return decode_hv (dec);
1002	case 6: // tag
1003	return decode_tagged (dec);
1004	case 7: // misc
1005	switch (*dec->cur++ & 31)	1142	switch (*dec->cur++ & MINOR_MASK)
1006	{	1143	{
1007	case 20:	1144	case SIMPLE_FALSE:
1008	#if CBOR_SLOW	1145	#if CBOR_SLOW
1009	types_false = get_bool ("Types::Serialiser::false");	1146	types_false = get_bool ("Types::Serialiser::false");
1010	#endif	1147	#endif
1011	return newSVsv (types_false);	1148	return newSVsv (types_false);
1012	case 21:	1149	case SIMPLE_TRUE:
1013	#if CBOR_SLOW	1150	#if CBOR_SLOW
1014	types_true = get_bool ("Types::Serialiser::true");	1151	types_true = get_bool ("Types::Serialiser::true");
1015	#endif	1152	#endif
1016	return newSVsv (types_true);	1153	return newSVsv (types_true);
1017	case 22:	1154	case SIMPLE_NULL:
1018	return newSVsv (&PL_sv_undef);	1155	return newSVsv (&PL_sv_undef);
1019	case 23:	1156	case SIMPLE_UNDEF:
1020	#if CBOR_SLOW	1157	#if CBOR_SLOW
1021	types_error = get_bool ("Types::Serialiser::error");	1158	types_error = get_bool ("Types::Serialiser::error");
1022	#endif	1159	#endif
1023	return newSVsv (types_error);	1160	return newSVsv (types_error);
1024		1161
1025	case 25:	1162	case MISC_FLOAT16:
1026	{	1163	{
1027	WANT (2);	1164	WANT (2);
1028		1165
1029	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1166	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1030	dec->cur += 2;	1167	dec->cur += 2;
1031		1168
1032	return newSVnv (ecb_binary16_to_float (fp));	1169	return newSVnv (ecb_binary16_to_float (fp));
1033	}	1170	}
1034		1171
1035	case 26:	1172	case MISC_FLOAT32:
1036	{	1173	{
1037	uint32_t fp;	1174	uint32_t fp;
1038	WANT (4);	1175	WANT (4);
1039	memcpy (&fp, dec->cur, 4);	1176	memcpy (&fp, dec->cur, 4);
1040	dec->cur += 4;	1177	dec->cur += 4;
…		…
1043	fp = ecb_bswap32 (fp);	1180	fp = ecb_bswap32 (fp);
1044		1181
1045	return newSVnv (ecb_binary32_to_float (fp));	1182	return newSVnv (ecb_binary32_to_float (fp));
1046	}	1183	}
1047		1184
1048	case 27:	1185	case MISC_FLOAT64:
1049	{	1186	{
1050	uint64_t fp;	1187	uint64_t fp;
1051	WANT (8);	1188	WANT (8);
1052	memcpy (&fp, dec->cur, 8);	1189	memcpy (&fp, dec->cur, 8);
1053	dec->cur += 8;	1190	dec->cur += 8;
…		…
1056	fp = ecb_bswap64 (fp);	1193	fp = ecb_bswap64 (fp);
1057		1194
1058	return newSVnv (ecb_binary64_to_double (fp));	1195	return newSVnv (ecb_binary64_to_double (fp));
1059	}	1196	}
1060		1197
1061	// 0..19 unassigned	1198	// 0..19 unassigned simple
1062	// 24 reserved + unassigned (reserved values are not encodable)	1199	// 24 reserved + unassigned simple (reserved values are not encodable)
		1200	// 28-30 unassigned misc
		1201	// 31 break code
1063	default:	1202	default:
1064	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1203	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1065	}	1204	}
1066		1205
1067	break;	1206	break;
1068	}	1207	}
1069		1208
…		…
1072	}	1211	}
1073		1212
1074	static SV *	1213	static SV *
1075	decode_cbor (SV string, CBOR cbor, char **offset_return)	1214	decode_cbor (SV string, CBOR cbor, char **offset_return)
1076	{	1215	{
1077	dec_t dec = { };	1216	dec_t dec = { 0 };
1078	SV *sv;	1217	SV *sv;
1079	STRLEN len;	1218	STRLEN len;
1080	char *data = SvPVbyte (string, len);	1219	char *data = SvPVbyte (string, len);
1081		1220
1082	if (len > cbor->max_size && cbor->max_size)	1221	if (len > cbor->max_size && cbor->max_size)
…		…
1096	if (dec.cur != dec.end && !dec.err)	1235	if (dec.cur != dec.end && !dec.err)
1097	dec.err = "garbage after CBOR object";	1236	dec.err = "garbage after CBOR object";
1098		1237
1099	if (dec.err)	1238	if (dec.err)
1100	{	1239	{
		1240	if (dec.shareable)
		1241	{
		1242	// need to break cyclic links, which would all be in shareable
		1243	int i;
		1244	SV **svp;
		1245
		1246	for (i = av_len (dec.shareable) + 1; i--; )
		1247	if ((svp = av_fetch (dec.shareable, i, 0)))
		1248	sv_setsv (*svp, &PL_sv_undef);
		1249	}
		1250
1101	SvREFCNT_dec (sv);	1251	SvREFCNT_dec (sv);
1102	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1252	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1103	}	1253	}
1104		1254
1105	sv = sv_2mortal (sv);	1255	sv = sv_2mortal (sv);
1106		1256
1107	return sv;	1257	return sv;
1108	}	1258	}
1109		1259
		1260	/////////////////////////////////////////////////////////////////////////////
		1261	// incremental parser
		1262
		1263	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1264
		1265	// returns 0 for notyet, 1 for success or error
		1266	static int
		1267	incr_parse (CBOR self, SV cborstr)
		1268	{
		1269	STRLEN cur;
		1270	SvPV (cborstr, cur);
		1271
		1272	while (ecb_expect_true (self->incr_need <= cur))
		1273	{
		1274	// table of integer count bytes
		1275	static I8 incr_len[MINOR_MASK + 1] = {
		1276	0, 0, 0, 0, 0, 0, 0, 0,
		1277	0, 0, 0, 0, 0, 0, 0, 0,
		1278	0, 0, 0, 0, 0, 0, 0, 0,
		1279	1, 2, 4, 8,-1,-1,-1,-2
		1280	};
		1281
		1282	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1283	U8 m = *p & MINOR_MASK;
		1284	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1285	I8 ilen = incr_len[m];
		1286
		1287	self->incr_need = self->incr_pos + 1;
		1288
		1289	if (ecb_expect_false (ilen < 0))
		1290	{
		1291	if (m != MINOR_INDEF)
		1292	return 1; // error
		1293
		1294	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1295	{
		1296	if (count >= 0)
		1297	return 1; // error
		1298
		1299	count = 1;
		1300	}
		1301	else
		1302	{
		1303	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1304	count = -1;
		1305	}
		1306	}
		1307	else
		1308	{
		1309	self->incr_need += ilen;
		1310	if (ecb_expect_false (self->incr_need > cur))
		1311	return 0;
		1312
		1313	int major = *p >> MAJOR_SHIFT;
		1314
		1315	switch (major)
		1316	{
		1317	case MAJOR_TAG >> MAJOR_SHIFT:
		1318	++count; // tags merely prefix another value
		1319	break;
		1320
		1321	case MAJOR_BYTES >> MAJOR_SHIFT:
		1322	case MAJOR_TEXT >> MAJOR_SHIFT:
		1323	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1324	case MAJOR_MAP >> MAJOR_SHIFT:
		1325	{
		1326	UV len;
		1327
		1328	if (ecb_expect_false (ilen))
		1329	{
		1330	len = 0;
		1331
		1332	do {
		1333	len = (len << 8) \| *++p;
		1334	} while (--ilen);
		1335	}
		1336	else
		1337	len = m;
		1338
		1339	switch (major)
		1340	{
		1341	case MAJOR_BYTES >> MAJOR_SHIFT:
		1342	case MAJOR_TEXT >> MAJOR_SHIFT:
		1343	self->incr_need += len;
		1344	if (ecb_expect_false (self->incr_need > cur))
		1345	return 0;
		1346
		1347	break;
		1348
		1349	case MAJOR_MAP >> MAJOR_SHIFT:
		1350	len <<= 1;
		1351	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1352	if (len)
		1353	{
		1354	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1355	count = len + 1;
		1356	}
		1357	break;
		1358	}
		1359	}
		1360	}
		1361	}
		1362
		1363	self->incr_pos = self->incr_need;
		1364
		1365	if (count > 0)
		1366	{
		1367	while (!--count)
		1368	{
		1369	if (!AvFILLp (self->incr_count))
		1370	return 1; // done
		1371
		1372	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1373	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1374	}
		1375
		1376	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1377	}
		1378	}
		1379
		1380	return 0;
		1381	}
		1382
		1383
1110	/////////////////////////////////////////////////////////////////////////////	1384	/////////////////////////////////////////////////////////////////////////////
1111	// XS interface functions	1385	// XS interface functions
1112		1386
1113	MODULE = CBOR::XS PACKAGE = CBOR::XS	1387	MODULE = CBOR::XS PACKAGE = CBOR::XS
1114		1388
…		…
1126		1400
1127	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1401	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1128		1402
1129	sv_cbor = newSVpv ("CBOR", 0);	1403	sv_cbor = newSVpv ("CBOR", 0);
1130	SvREADONLY_on (sv_cbor);	1404	SvREADONLY_on (sv_cbor);
		1405
		1406	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1131	}	1407	}
1132		1408
1133	PROTOTYPES: DISABLE	1409	PROTOTYPES: DISABLE
1134		1410
1135	void CLONE (...)	1411	void CLONE (...)
…		…
1154	void shrink (CBOR *self, int enable = 1)	1430	void shrink (CBOR *self, int enable = 1)
1155	ALIAS:	1431	ALIAS:
1156	shrink = F_SHRINK	1432	shrink = F_SHRINK
1157	allow_unknown = F_ALLOW_UNKNOWN	1433	allow_unknown = F_ALLOW_UNKNOWN
1158	allow_sharing = F_ALLOW_SHARING	1434	allow_sharing = F_ALLOW_SHARING
		1435	allow_cycles = F_ALLOW_CYCLES
1159	pack_strings = F_PACK_STRINGS	1436	pack_strings = F_PACK_STRINGS
		1437	text_keys = F_TEXT_KEYS
		1438	text_strings = F_TEXT_STRINGS
		1439	validate_utf8 = F_VALIDATE_UTF8
1160	PPCODE:	1440	PPCODE:
1161	{	1441	{
1162	if (enable)	1442	if (enable)
1163	self->flags \|= ix;	1443	self->flags \|= ix;
1164	else	1444	else
…		…
1170	void get_shrink (CBOR *self)	1450	void get_shrink (CBOR *self)
1171	ALIAS:	1451	ALIAS:
1172	get_shrink = F_SHRINK	1452	get_shrink = F_SHRINK
1173	get_allow_unknown = F_ALLOW_UNKNOWN	1453	get_allow_unknown = F_ALLOW_UNKNOWN
1174	get_allow_sharing = F_ALLOW_SHARING	1454	get_allow_sharing = F_ALLOW_SHARING
		1455	get_allow_cycles = F_ALLOW_CYCLES
1175	get_pack_strings = F_PACK_STRINGS	1456	get_pack_strings = F_PACK_STRINGS
		1457	get_text_keys = F_TEXT_KEYS
		1458	get_text_strings = F_TEXT_STRINGS
		1459	get_validate_utf8 = F_VALIDATE_UTF8
1176	PPCODE:	1460	PPCODE:
1177	XPUSHs (boolSV (self->flags & ix));	1461	XPUSHs (boolSV (self->flags & ix));
1178		1462
1179	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1463	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1180	PPCODE:	1464	PPCODE:
…		…
1229	EXTEND (SP, 2);	1513	EXTEND (SP, 2);
1230	PUSHs (sv);	1514	PUSHs (sv);
1231	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1515	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1232	}	1516	}
1233		1517
		1518	void incr_parse (CBOR self, SV cborstr)
		1519	ALIAS:
		1520	incr_parse_multiple = 1
		1521	PPCODE:
		1522	{
		1523	if (SvUTF8 (cborstr))
		1524	sv_utf8_downgrade (cborstr, 0);
		1525
		1526	if (!self->incr_count)
		1527	{
		1528	self->incr_count = newAV ();
		1529	self->incr_pos = 0;
		1530	self->incr_need = 1;
		1531
		1532	av_push (self->incr_count, newSViv (1));
		1533	}
		1534
		1535	do
		1536	{
		1537	if (!incr_parse (self, cborstr))
		1538	{
		1539	if (self->incr_need > self->max_size && self->max_size)
		1540	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1541	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1542
		1543	break;
		1544	}
		1545
		1546	SV *sv;
		1547	char *offset;
		1548
		1549	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1550	XPUSHs (sv);
		1551
		1552	sv_chop (cborstr, offset);
		1553
		1554	av_clear (self->incr_count);
		1555	av_push (self->incr_count, newSViv (1));
		1556
		1557	self->incr_pos = 0;
		1558	self->incr_need = self->incr_pos + 1;
		1559	}
		1560	while (ix);
		1561	}
		1562
		1563	void incr_reset (CBOR *self)
		1564	CODE:
		1565	{
		1566	SvREFCNT_dec (self->incr_count);
		1567	self->incr_count = 0;
		1568	}
		1569
1234	void DESTROY (CBOR *self)	1570	void DESTROY (CBOR *self)
1235	PPCODE:	1571	PPCODE:
1236	cbor_free (self);	1572	cbor_free (self);
1237		1573
1238	PROTOTYPES: ENABLE	1574	PROTOTYPES: ENABLE
1239		1575
1240	void encode_cbor (SV *scalar)	1576	void encode_cbor (SV *scalar)
		1577	ALIAS:
		1578	encode_cbor = 0
		1579	encode_cbor_sharing = F_ALLOW_SHARING
1241	PPCODE:	1580	PPCODE:
1242	{	1581	{
1243	CBOR cbor;	1582	CBOR cbor;
1244	cbor_init (&cbor);	1583	cbor_init (&cbor);
		1584	cbor.flags \|= ix;
1245	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1585	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1246	XPUSHs (scalar);	1586	XPUSHs (scalar);
1247	}	1587	}
1248		1588
1249	void decode_cbor (SV *cborstr)	1589	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.33 by root, Sat Nov 30 15:23:59 2013 UTC vs. Revision 1.58 by root, Fri Nov 25 13:27:29 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.33 by root, Sat Nov 30 15:23:59 2013 UTC vs.
Revision 1.58 by root, Fri Nov 25 13:27:29 2016 UTC