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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs.
Revision 1.56 by root, Fri Nov 25 11:33:03 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))
…		…
19	# define HvNAMELEN(hv) HvNAMELEN_get (hv)	20	# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20	#endif	21	#endif
21	#ifndef HvNAMEUTF8	22	#ifndef HvNAMEUTF8
22	# define HvNAMEUTF8(hv) 0	23	# define HvNAMEUTF8(hv) 0
23	#endif	24	#endif
		25	#ifndef SvREFCNT_dec_NN
		26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
		27	#endif
		28
		29	// known major and minor types
		30	enum cbor_type
		31	{
		32	MAJOR_SHIFT = 5,
		33	MINOR_MASK = 0x1f,
		34
		35	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		36	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		37	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		38	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		39	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		40	MAJOR_MAP = 5 << MAJOR_SHIFT,
		41	MAJOR_TAG = 6 << MAJOR_SHIFT,
		42	MAJOR_MISC = 7 << MAJOR_SHIFT,
		43
		44	// INT/STRING/ARRAY/MAP subtypes
		45	LENGTH_EXT1 = 24,
		46	LENGTH_EXT2 = 25,
		47	LENGTH_EXT4 = 26,
		48	LENGTH_EXT8 = 27,
		49
		50	// SIMPLE types (effectively MISC subtypes)
		51	SIMPLE_FALSE = 20,
		52	SIMPLE_TRUE = 21,
		53	SIMPLE_NULL = 22,
		54	SIMPLE_UNDEF = 23,
		55
		56	// MISC subtype (unused)
		57	MISC_EXT1 = 24,
		58	MISC_FLOAT16 = 25,
		59	MISC_FLOAT32 = 26,
		60	MISC_FLOAT64 = 27,
		61
		62	// BYTES/TEXT/ARRAY/MAP
		63	MINOR_INDEF = 31,
		64	};
24		65
25	// known tags	66	// known tags
26	enum cbor_tag	67	enum cbor_tag
27	{	68	{
28	// inofficial extensions (pending iana registration)	69	// extensions
		70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29	CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30	CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31	CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32	CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33	CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34	CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36		77
37	// rfc7049	78	// rfc7049
38	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
41	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
42	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
44		85
45	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
46	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
47	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
48	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49		90
50	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55		96
56	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
57	};	98	};
58		99
59	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
60	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
61	#define F_ALLOW_SHARING 0x00000004UL //TODO	102	#define F_ALLOW_SHARING 0x00000004UL
62	#define F_ALLOW_STRINGREF 0x00000008UL //TODO	103	#define F_ALLOW_CYCLES 0x00000008UL
		104	#define F_PACK_STRINGS 0x00000010UL
		105	#define F_TEXT_KEYS 0x00000020UL
		106	#define F_TEXT_STRINGS 0x00000040UL
		107	#define F_VALIDATE_UTF8 0x00000080UL
63		108
64	#define INIT_SIZE 32 // initial scalar size to be allocated	109	#define INIT_SIZE 32 // initial scalar size to be allocated
65		110
66	#define SB do {	111	#define SB do {
67	#define SE } while (0)	112	#define SE } while (0)
…		…
86	typedef struct {	131	typedef struct {
87	U32 flags;	132	U32 flags;
88	U32 max_depth;	133	U32 max_depth;
89	STRLEN max_size;	134	STRLEN max_size;
90	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.
91	} CBOR;	141	} CBOR;
92		142
93	ecb_inline void	143	ecb_inline void
94	cbor_init (CBOR *cbor)	144	cbor_init (CBOR *cbor)
95	{	145	{
…		…
99		149
100	ecb_inline void	150	ecb_inline void
101	cbor_free (CBOR *cbor)	151	cbor_free (CBOR *cbor)
102	{	152	{
103	SvREFCNT_dec (cbor->filter);	153	SvREFCNT_dec (cbor->filter);
		154	SvREFCNT_dec (cbor->incr_count);
104	}	155	}
105		156
106	/////////////////////////////////////////////////////////////////////////////	157	/////////////////////////////////////////////////////////////////////////////
107	// utility functions	158	// utility functions
108		159
…		…
138	{	189	{
139	return idx > 23	190	return idx > 23
140	? idx > 0xffU	191	? idx > 0xffU
141	? idx > 0xffffU	192	? idx > 0xffffU
142	? idx > 0xffffffffU	193	? idx > 0xffffffffU
		194	? 11
143	? 7	195	: 7
144	: 6
145	: 5	196	: 5
146	: 4	197	: 4
147	: 3;	198	: 3;
148	}	199	}
149		200
…		…
186	static void	237	static void
187	encode_uint (enc_t *enc, int major, UV len)	238	encode_uint (enc_t *enc, int major, UV len)
188	{	239	{
189	need (enc, 9);	240	need (enc, 9);
190		241
191	if (len < 24)	242	if (ecb_expect_true (len < LENGTH_EXT1))
192	*enc->cur++ = major \| len;	243	*enc->cur++ = major \| len;
193	else if (len <= 0xff)	244	else if (ecb_expect_true (len <= 0xffU))
194	{	245	{
195	*enc->cur++ = major \| 24;	246	*enc->cur++ = major \| LENGTH_EXT1;
196	*enc->cur++ = len;	247	*enc->cur++ = len;
197	}	248	}
198	else if (len <= 0xffff)	249	else if (len <= 0xffffU)
199	{	250	{
200	*enc->cur++ = major \| 25;	251	*enc->cur++ = major \| LENGTH_EXT2;
201	*enc->cur++ = len >> 8;	252	*enc->cur++ = len >> 8;
202	*enc->cur++ = len;	253	*enc->cur++ = len;
203	}	254	}
204	else if (len <= 0xffffffff)	255	else if (len <= 0xffffffffU)
205	{	256	{
206	*enc->cur++ = major \| 26;	257	*enc->cur++ = major \| LENGTH_EXT4;
207	*enc->cur++ = len >> 24;	258	*enc->cur++ = len >> 24;
208	*enc->cur++ = len >> 16;	259	*enc->cur++ = len >> 16;
209	*enc->cur++ = len >> 8;	260	*enc->cur++ = len >> 8;
210	*enc->cur++ = len;	261	*enc->cur++ = len;
211	}	262	}
212	else	263	else
213	{	264	{
214	*enc->cur++ = major \| 27;	265	*enc->cur++ = major \| LENGTH_EXT8;
215	*enc->cur++ = len >> 56;	266	*enc->cur++ = len >> 56;
216	*enc->cur++ = len >> 48;	267	*enc->cur++ = len >> 48;
217	*enc->cur++ = len >> 40;	268	*enc->cur++ = len >> 40;
218	*enc->cur++ = len >> 32;	269	*enc->cur++ = len >> 32;
219	*enc->cur++ = len >> 24;	270	*enc->cur++ = len >> 24;
…		…
224	}	275	}
225		276
226	ecb_inline void	277	ecb_inline void
227	encode_tag (enc_t *enc, UV tag)	278	encode_tag (enc_t *enc, UV tag)
228	{	279	{
229	encode_uint (enc, 0xc0, tag);	280	encode_uint (enc, MAJOR_TAG, tag);
230	}	281	}
231		282
232	static void	283	// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
		284	ecb_noinline static void
233	encode_str (enc_t enc, int utf8, char str, STRLEN len)	285	encode_str_utf8 (enc_t enc, int utf8, char str, STRLEN len)
234	{	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	}
		304	}
		305
		306	ecb_inline void
		307	encode_str (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
		308	{
		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);
		317	need (enc, len);
		318	memcpy (enc->cur, str, len);
		319	enc->cur += len;
		320	}
		321
		322	ecb_inline void
		323	encode_strref (enc_t enc, int upgrade_utf8, int utf8, char str, STRLEN len)
		324	{
235	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF))	325	if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
236	{	326	{
237	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);	327	SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
238		328
239	if (SvOK (*svp))	329	if (SvOK (*svp))
240	{	330	{
241	// already registered, use stringref	331	// already registered, use stringref
242	encode_tag (enc, CBOR_TAG_STRINGREF);	332	encode_tag (enc, CBOR_TAG_STRINGREF);
243	encode_uint (enc, 0x00, SvUV (*svp));	333	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
244	return;	334	return;
245	}	335	}
246	else if (len >= minimum_string_length (enc->stringref_idx))	336	else if (len >= minimum_string_length (enc->stringref_idx))
247	{	337	{
248	// register only	338	// register only
249	sv_setuv (*svp, enc->stringref_idx);	339	sv_setuv (*svp, enc->stringref_idx);
250	++enc->stringref_idx;	340	++enc->stringref_idx;
251	}	341	}
252	}	342	}
253		343
254	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	344	encode_str (enc, upgrade_utf8, utf8, str, len);
255	need (enc, len);
256	memcpy (enc->cur, str, len);
257	enc->cur += len;
258	}	345	}
259		346
260	static void encode_sv (enc_t enc, SV sv);	347	static void encode_sv (enc_t enc, SV sv);
261		348
262	static void	349	static void
…		…
267	if (enc->depth >= enc->cbor.max_depth)	354	if (enc->depth >= enc->cbor.max_depth)
268	croak (ERR_NESTING_EXCEEDED);	355	croak (ERR_NESTING_EXCEEDED);
269		356
270	++enc->depth;	357	++enc->depth;
271		358
272	encode_uint (enc, 0x80, len + 1);	359	encode_uint (enc, MAJOR_ARRAY, len + 1);
273		360
		361	if (SvMAGICAL (av))
274	for (i = 0; i <= len; ++i)	362	for (i = 0; i <= len; ++i)
275	{	363	{
276	SV **svp = av_fetch (av, i, 0);	364	SV **svp = av_fetch (av, i, 0);
277	encode_sv (enc, svp ? *svp : &PL_sv_undef);	365	encode_sv (enc, svp ? *svp : &PL_sv_undef);
278	}	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	}
279		373
280	--enc->depth;	374	--enc->depth;
281	}	375	}
282		376
283	static void	377	static void
…		…
292		386
293	int pairs = hv_iterinit (hv);	387	int pairs = hv_iterinit (hv);
294	int mg = SvMAGICAL (hv);	388	int mg = SvMAGICAL (hv);
295		389
296	if (mg)	390	if (mg)
297	encode_ch (enc, 0xa0 \| 31);	391	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
298	else	392	else
299	encode_uint (enc, 0xa0, pairs);	393	encode_uint (enc, MAJOR_MAP, pairs);
300		394
301	while ((he = hv_iternext (hv)))	395	while ((he = hv_iternext (hv)))
302	{	396	{
303	if (HeKLEN (he) == HEf_SVKEY)	397	if (HeKLEN (he) == HEf_SVKEY)
304	encode_sv (enc, HeSVKEY (he));	398	encode_sv (enc, HeSVKEY (he));
305	else	399	else
306	encode_str (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));
307		401
308	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));
309	}	403	}
310		404
311	if (mg)	405	if (mg)
312	encode_ch (enc, 0xe0 \| 31);	406	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
313		407
314	--enc->depth;	408	--enc->depth;
315	}	409	}
316		410
317	// encode objects, arrays and special \0=false and \1=true values.	411	// encode objects, arrays and special \0=false and \1=true values.
318	static void	412	static void
319	encode_rv (enc_t enc, SV sv)	413	encode_rv (enc_t enc, SV sv)
320	{	414	{
321	SvGETMAGIC (sv);	415	SvGETMAGIC (sv);
322
323	if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
324	&& ecb_expect_false (SvREFCNT (sv) > 1))
325	{
326	if (!enc->shareable)
327	enc->shareable = (HV )sv_2mortal ((SV )newHV ());
328
329	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
330
331	if (SvOK (*svp))
332	{
333	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
334	encode_uint (enc, 0x00, SvUV (*svp));
335	return;
336	}
337	else
338	{
339	sv_setuv (*svp, enc->shareable_idx);
340	++enc->shareable_idx;
341	encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
342	}
343	}
344		416
345	svtype svt = SvTYPE (sv);	417	svtype svt = SvTYPE (sv);
346		418
347	if (ecb_expect_false (SvOBJECT (sv)))	419	if (ecb_expect_false (SvOBJECT (sv)))
348	{	420	{
…		…
355	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	427	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
356	? cbor_tagged_stash	428	? cbor_tagged_stash
357	: gv_stashpv ("CBOR::XS::Tagged" , 1);	429	: gv_stashpv ("CBOR::XS::Tagged" , 1);
358		430
359	HV *stash = SvSTASH (sv);	431	HV *stash = SvSTASH (sv);
360	GV *method;
361		432
362	if (stash == boolean_stash)	433	if (stash == boolean_stash)
363	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	}
364	else if (stash == error_stash)	438	else if (stash == error_stash)
365	encode_ch (enc, 0xe0 \| 23);	439	{
		440	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
		441	return;
		442	}
366	else if (stash == tagged_stash)	443	else if (stash == tagged_stash)
367	{	444	{
368	if (svt != SVt_PVAV)	445	if (svt != SVt_PVAV)
369	croak ("encountered CBOR::XS::Tagged object that isn't an array");	446	croak ("encountered CBOR::XS::Tagged object that isn't an array");
370		447
371	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	448	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
372	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))
373	}	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
374	else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	482	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
375	{	483	{
376	dSP;	484	dSP;
377		485
378	ENTER; SAVETMPS; PUSHMARK (SP);	486	ENTER; SAVETMPS;
		487	PUSHMARK (SP);
379	// 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
380	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	489	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
381		490
382	PUTBACK;	491	PUTBACK;
383	// G_SCALAR ensures that return value is 1	492	// G_SCALAR ensures that return value is 1
…		…
396	}	505	}
397	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	506	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
398	{	507	{
399	dSP;	508	dSP;
400		509
401	ENTER; SAVETMPS; PUSHMARK (SP);	510	ENTER; SAVETMPS;
		511	SAVESTACK_POS ();
		512	PUSHMARK (SP);
402	EXTEND (SP, 2);	513	EXTEND (SP, 2);
403	// 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
404	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	515	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
405	PUSHs (sv_cbor);	516	PUSHs (sv_cbor);
406		517
…		…
411	// catch this surprisingly common error	522	// catch this surprisingly common error
412	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	523	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
413	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));
414		525
415	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	526	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
416	encode_uint (enc, 0x80, count + 1);	527	encode_uint (enc, MAJOR_ARRAY, count + 1);
417	encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	528	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
418		529
419	while (count)	530	while (count)
420	encode_sv (enc, SP[1 - count--]);	531	encode_sv (enc, SP[1 - count--]);
421		532
422	PUTBACK;	533	PUTBACK;
…		…
443	{	554	{
444	double nv = SvNVX (sv);	555	double nv = SvNVX (sv);
445		556
446	need (enc, 9);	557	need (enc, 9);
447		558
448	if (ecb_expect_false (nv == (U32)nv))	559	if (ecb_expect_false (nv == (NV)(U32)nv))
449	encode_uint (enc, 0x00, (U32)nv);	560	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
450	//TODO: maybe I32?	561	//TODO: maybe I32?
451	else if (ecb_expect_false (nv == (float)nv))	562	else if (ecb_expect_false (nv == (float)nv))
452	{	563	{
453	uint32_t fp = ecb_float_to_binary32 (nv);	564	uint32_t fp = ecb_float_to_binary32 (nv);
454		565
455	*enc->cur++ = 0xe0 \| 26;	566	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
456		567
457	if (!ecb_big_endian ())	568	if (!ecb_big_endian ())
458	fp = ecb_bswap32 (fp);	569	fp = ecb_bswap32 (fp);
459		570
460	memcpy (enc->cur, &fp, 4);	571	memcpy (enc->cur, &fp, 4);
…		…
462	}	573	}
463	else	574	else
464	{	575	{
465	uint64_t fp = ecb_double_to_binary64 (nv);	576	uint64_t fp = ecb_double_to_binary64 (nv);
466		577
467	*enc->cur++ = 0xe0 \| 27;	578	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
468		579
469	if (!ecb_big_endian ())	580	if (!ecb_big_endian ())
470	fp = ecb_bswap64 (fp);	581	fp = ecb_bswap64 (fp);
471		582
472	memcpy (enc->cur, &fp, 8);	583	memcpy (enc->cur, &fp, 8);
…		…
481		592
482	if (SvPOKp (sv))	593	if (SvPOKp (sv))
483	{	594	{
484	STRLEN len;	595	STRLEN len;
485	char *str = SvPV (sv, len);	596	char *str = SvPV (sv, len);
486	encode_str (enc, SvUTF8 (sv), str, len);	597	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
487	}	598	}
488	else if (SvNOKp (sv))	599	else if (SvNOKp (sv))
489	encode_nv (enc, sv);	600	encode_nv (enc, sv);
490	else if (SvIOKp (sv))	601	else if (SvIOKp (sv))
491	{	602	{
492	if (SvIsUV (sv))	603	if (SvIsUV (sv))
493	encode_uint (enc, 0x00, SvUVX (sv));	604	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
494	else if (SvIVX (sv) >= 0)	605	else if (SvIVX (sv) >= 0)
495	encode_uint (enc, 0x00, SvIVX (sv));	606	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
496	else	607	else
497	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	608	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
498	}	609	}
499	else if (SvROK (sv))	610	else if (SvROK (sv))
500	encode_rv (enc, SvRV (sv));	611	encode_rv (enc, SvRV (sv));
501	else if (!SvOK (sv))	612	else if (!SvOK (sv))
502	encode_ch (enc, 0xe0 \| 22);	613	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
503	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	614	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
504	encode_ch (enc, 0xe0 \| 23);	615	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
505	else	616	else
506	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",
507	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	618	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
508	}	619	}
509		620
510	static SV *	621	static SV *
511	encode_cbor (SV scalar, CBOR cbor)	622	encode_cbor (SV scalar, CBOR cbor)
512	{	623	{
513	enc_t enc = { };	624	enc_t enc = { 0 };
514		625
515	enc.cbor = *cbor;	626	enc.cbor = *cbor;
516	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	627	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
517	enc.cur = SvPVX (enc.sv);	628	enc.cur = SvPVX (enc.sv);
518	enc.end = SvEND (enc.sv);	629	enc.end = SvEND (enc.sv);
519		630
520	SvPOK_only (enc.sv);	631	SvPOK_only (enc.sv);
521		632
522	if (cbor->flags & F_ALLOW_STRINGREF)	633	if (cbor->flags & F_PACK_STRINGS)
523	{	634	{
524	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);	635	encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
525	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());	636	enc.stringref[0]= (HV )sv_2mortal ((SV )newHV ());
526	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());	637	enc.stringref[1]= (HV )sv_2mortal ((SV )newHV ());
527	}	638	}
…		…
556		667
557	#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
558		669
559	#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 (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
560		671
561	#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)
562	#define DEC_DEC_DEPTH --dec->depth	673	#define DEC_DEC_DEPTH --dec->depth
563		674
564	static UV	675	static UV
565	decode_uint (dec_t *dec)	676	decode_uint (dec_t *dec)
566	{	677	{
567	switch (*dec->cur & 31)	678	U8 m = *dec->cur & MINOR_MASK;
568	{	679	++dec->cur;
569	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
570	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
571	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
572	return *dec->cur++ & 31;
573		680
574	case 24:	681	if (ecb_expect_true (m < LENGTH_EXT1))
		682	return m;
		683	else if (ecb_expect_true (m == LENGTH_EXT1))
		684	{
575	WANT (2);	685	WANT (1);
576	dec->cur += 2;	686	dec->cur += 1;
577	return dec->cur[-1];	687	return dec->cur[-1];
578		688	}
579	case 25:	689	else if (ecb_expect_true (m == LENGTH_EXT2))
		690	{
580	WANT (3);	691	WANT (2);
581	dec->cur += 3;	692	dec->cur += 2;
582	return (((UV)dec->cur[-2]) << 8)	693	return (((UV)dec->cur[-2]) << 8)
583	\| ((UV)dec->cur[-1]);	694	\| ((UV)dec->cur[-1]);
584		695	}
585	case 26:	696	else if (ecb_expect_true (m == LENGTH_EXT4))
		697	{
586	WANT (5);	698	WANT (4);
587	dec->cur += 5;	699	dec->cur += 4;
588	return (((UV)dec->cur[-4]) << 24)	700	return (((UV)dec->cur[-4]) << 24)
589	\| (((UV)dec->cur[-3]) << 16)	701	\| (((UV)dec->cur[-3]) << 16)
590	\| (((UV)dec->cur[-2]) << 8)	702	\| (((UV)dec->cur[-2]) << 8)
591	\| ((UV)dec->cur[-1]);	703	\| ((UV)dec->cur[-1]);
592		704	}
593	case 27:	705	else if (ecb_expect_true (m == LENGTH_EXT8))
		706	{
594	WANT (9);	707	WANT (8);
595	dec->cur += 9;	708	dec->cur += 8;
		709
		710	return
		711	#if UVSIZE < 8
		712	0
		713	#else
596	return (((UV)dec->cur[-8]) << 56)	714	(((UV)dec->cur[-8]) << 56)
597	\| (((UV)dec->cur[-7]) << 48)	715	\| (((UV)dec->cur[-7]) << 48)
598	\| (((UV)dec->cur[-6]) << 40)	716	\| (((UV)dec->cur[-6]) << 40)
599	\| (((UV)dec->cur[-5]) << 32)	717	\| (((UV)dec->cur[-5]) << 32)
		718	#endif
600	\| (((UV)dec->cur[-4]) << 24)	719	\| (((UV)dec->cur[-4]) << 24)
601	\| (((UV)dec->cur[-3]) << 16)	720	\| (((UV)dec->cur[-3]) << 16)
602	\| (((UV)dec->cur[-2]) << 8)	721	\| (((UV)dec->cur[-2]) << 8)
603	\| ((UV)dec->cur[-1]);	722	\| ((UV)dec->cur[-1]);
604		723	}
605	default:	724	else
606	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	725	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
607	}
608		726
609	fail:	727	fail:
610	return 0;	728	return 0;
611	}	729	}
612		730
…		…
617	{	735	{
618	AV *av = newAV ();	736	AV *av = newAV ();
619		737
620	DEC_INC_DEPTH;	738	DEC_INC_DEPTH;
621		739
622	if ((*dec->cur & 31) == 31)	740	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
623	{	741	{
624	++dec->cur;	742	++dec->cur;
625		743
626	for (;;)	744	for (;;)
627	{	745	{
628	WANT (1);	746	WANT (1);
629		747
630	if (*dec->cur == (0xe0 \| 31))	748	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
631	{	749	{
632	++dec->cur;	750	++dec->cur;
633	break;	751	break;
634	}	752	}
635		753
…		…
638	}	756	}
639	else	757	else
640	{	758	{
641	int i, len = decode_uint (dec);	759	int i, len = decode_uint (dec);
642		760
		761	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
643	av_fill (av, len - 1);	762	av_fill (av, len - 1);
644		763
645	for (i = 0; i < len; ++i)	764	for (i = 0; i < len; ++i)
646	AvARRAY (av)[i] = decode_sv (dec);	765	AvARRAY (av)[i] = decode_sv (dec);
647	}	766	}
…		…
660	{	779	{
661	// for speed reasons, we specialcase single-string	780	// for speed reasons, we specialcase single-string
662	// byte or utf-8 strings as keys, but only when !stringref	781	// byte or utf-8 strings as keys, but only when !stringref
663		782
664	if (ecb_expect_true (!dec->stringref))	783	if (ecb_expect_true (!dec->stringref))
665	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	784	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
666	{	785	{
667	I32 len = decode_uint (dec);	786	STRLEN len = decode_uint (dec);
668	char key = (char )dec->cur;	787	char key = (char )dec->cur;
669		788
		789	WANT (len);
670	dec->cur += len;	790	dec->cur += len;
671		791
672	if (ecb_expect_false (dec->stringref))
673	av_push (dec->stringref, newSVpvn (key, len));
674
675	hv_store (hv, key, len, decode_sv (dec), 0);	792	hv_store (hv, key, len, decode_sv (dec), 0);
676		793
677	return;	794	return;
678	}	795	}
679	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	796	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
680	{	797	{
681	I32 len = decode_uint (dec);	798	STRLEN len = decode_uint (dec);
682	char key = (char )dec->cur;	799	char key = (char )dec->cur;
683		800
		801	WANT (len);
684	dec->cur += len;	802	dec->cur += len;
685		803
686	if (ecb_expect_false (dec->stringref))	804	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
687	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)");
688		807
689	hv_store (hv, key, -len, decode_sv (dec), 0);	808	hv_store (hv, key, -len, decode_sv (dec), 0);
690		809
691	return;	810	return;
692	}	811	}
…		…
694	SV *k = decode_sv (dec);	813	SV *k = decode_sv (dec);
695	SV *v = decode_sv (dec);	814	SV *v = decode_sv (dec);
696		815
697	hv_store_ent (hv, k, v, 0);	816	hv_store_ent (hv, k, v, 0);
698	SvREFCNT_dec (k);	817	SvREFCNT_dec (k);
		818
		819	fail:
		820	;
699	}	821	}
700		822
701	static SV *	823	static SV *
702	decode_hv (dec_t *dec)	824	decode_hv (dec_t *dec)
703	{	825	{
704	HV *hv = newHV ();	826	HV *hv = newHV ();
705		827
706	DEC_INC_DEPTH;	828	DEC_INC_DEPTH;
707		829
708	if ((*dec->cur & 31) == 31)	830	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
709	{	831	{
710	++dec->cur;	832	++dec->cur;
711		833
712	for (;;)	834	for (;;)
713	{	835	{
714	WANT (1);	836	WANT (1);
715		837
716	if (*dec->cur == (0xe0 \| 31))	838	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
717	{	839	{
718	++dec->cur;	840	++dec->cur;
719	break;	841	break;
720	}	842	}
721		843
…		…
742	static SV *	864	static SV *
743	decode_str (dec_t *dec, int utf8)	865	decode_str (dec_t *dec, int utf8)
744	{	866	{
745	SV *sv = 0;	867	SV *sv = 0;
746		868
747	if ((*dec->cur & 31) == 31)	869	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
748	{	870	{
		871	// indefinite length strings
749	++dec->cur;	872	++dec->cur;
750		873
		874	U8 major = *dec->cur & MAJOR_MISC;
		875
751	sv = newSVpvn ("", 0);	876	sv = newSVpvn ("", 0);
752		877
753	// not very fast, and certainly not robust against illegal input
754	for (;;)	878	for (;;)
755	{	879	{
756	WANT (1);	880	WANT (1);
757		881
758	if (*dec->cur == (0xe0 \| 31))	882	if ((*dec->cur - major) > LENGTH_EXT8)
		883	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
759	{	884	{
760	++dec->cur;	885	++dec->cur;
761	break;	886	break;
762	}	887	}
		888	else
		889	ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
763		890
764	sv_catsv (sv, decode_sv (dec));	891	STRLEN len = decode_uint (dec);
		892
		893	WANT (len);
		894	sv_catpvn (sv, dec->cur, len);
		895	dec->cur += len;
765	}	896	}
766	}	897	}
767	else	898	else
768	{	899	{
769	STRLEN len = decode_uint (dec);	900	STRLEN len = decode_uint (dec);
…		…
776	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	907	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
777	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	908	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
778	}	909	}
779		910
780	if (utf8)	911	if (utf8)
		912	{
		913	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		914	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		915	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		916
781	SvUTF8_on (sv);	917	SvUTF8_on (sv);
		918	}
782		919
783	return sv;	920	return sv;
784		921
785	fail:	922	fail:
786	SvREFCNT_dec (sv);	923	SvREFCNT_dec (sv);
…		…
805	sv = newRV_noinc (decode_sv (dec));	942	sv = newRV_noinc (decode_sv (dec));
806	break;	943	break;
807		944
808	case CBOR_TAG_STRINGREF_NAMESPACE:	945	case CBOR_TAG_STRINGREF_NAMESPACE:
809	{	946	{
		947	// do nmot use SAVETMPS/FREETMPS, as these will
		948	// erase mortalised caches, e.g. "shareable"
810	ENTER; SAVETMPS;	949	ENTER;
811		950
812	SAVESPTR (dec->stringref);	951	SAVESPTR (dec->stringref);
813	dec->stringref = (AV )sv_2mortal ((SV )newAV ());	952	dec->stringref = (AV )sv_2mortal ((SV )newAV ());
814		953
815	sv = decode_sv (dec);	954	sv = decode_sv (dec);
816		955
817	FREETMPS; LEAVE;	956	LEAVE;
818	}	957	}
819	break;	958	break;
820		959
821	case CBOR_TAG_STRINGREF:	960	case CBOR_TAG_STRINGREF:
822	{	961	{
823	if ((*dec->cur >> 5) != 0)	962	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
824	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	963	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
825		964
826	UV idx = decode_uint (dec);	965	UV idx = decode_uint (dec);
827		966
828	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	967	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
835	case CBOR_TAG_VALUE_SHAREABLE:	974	case CBOR_TAG_VALUE_SHAREABLE:
836	{	975	{
837	if (ecb_expect_false (!dec->shareable))	976	if (ecb_expect_false (!dec->shareable))
838	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	977	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
839		978
		979	if (dec->cbor.flags & F_ALLOW_CYCLES)
		980	{
840	sv = newSV (0);	981	sv = newSV (0);
841	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	982	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
842		983
843	SV *osv = decode_sv (dec);	984	SV *osv = decode_sv (dec);
844	sv_setsv (sv, osv);	985	sv_setsv (sv, osv);
845	SvREFCNT_dec_NN (osv);	986	SvREFCNT_dec_NN (osv);
		987	}
		988	else
		989	{
		990	av_push (dec->shareable, &PL_sv_undef);
		991	int idx = AvFILLp (dec->shareable);
		992	sv = decode_sv (dec);
		993	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		994	}
846	}	995	}
847	break;	996	break;
848		997
849	case CBOR_TAG_VALUE_SHAREDREF:	998	case CBOR_TAG_VALUE_SHAREDREF:
850	{	999	{
851	if ((*dec->cur >> 5) != 0)	1000	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
852	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1001	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
853		1002
854	UV idx = decode_uint (dec);	1003	UV idx = decode_uint (dec);
855		1004
856	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	1005	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
857	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1006	ERR ("corrupted CBOR data (sharedref index out of bounds)");
858		1007
859	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1008	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1009
		1010	if (sv == &PL_sv_undef)
		1011	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
860	}	1012	}
861	break;	1013	break;
862		1014
863	case CBOR_TAG_PERL_OBJECT:	1015	case CBOR_TAG_PERL_OBJECT:
864	{	1016	{
…		…
879	if (!method)	1031	if (!method)
880	ERR ("cannot decode perl-object (package does not have a THAW method)");	1032	ERR ("cannot decode perl-object (package does not have a THAW method)");
881		1033
882	dSP;	1034	dSP;
883		1035
884	ENTER; SAVETMPS; PUSHMARK (SP);	1036	ENTER; SAVETMPS;
		1037	PUSHMARK (SP);
885	EXTEND (SP, len + 1);	1038	EXTEND (SP, len + 1);
886	// we re-bless the reference to get overload and other niceties right	1039	// we re-bless the reference to get overload and other niceties right
887	PUSHs (*av_fetch (av, 0, 1));	1040	PUSHs (*av_fetch (av, 0, 1));
888	PUSHs (sv_cbor);	1041	PUSHs (sv_cbor);
889		1042
…		…
914	default:	1067	default:
915	{	1068	{
916	sv = decode_sv (dec);	1069	sv = decode_sv (dec);
917		1070
918	dSP;	1071	dSP;
919	ENTER; SAVETMPS; PUSHMARK (SP);	1072	ENTER; SAVETMPS;
		1073	SAVESTACK_POS ();
		1074	PUSHMARK (SP);
920	EXTEND (SP, 2);	1075	EXTEND (SP, 2);
921	PUSHs (newSVuv (tag));	1076	PUSHs (newSVuv (tag));
922	PUSHs (sv);	1077	PUSHs (sv);
923		1078
924	PUTBACK;	1079	PUTBACK;
…		…
965	static SV *	1120	static SV *
966	decode_sv (dec_t *dec)	1121	decode_sv (dec_t *dec)
967	{	1122	{
968	WANT (1);	1123	WANT (1);
969		1124
970	switch (*dec->cur >> 5)	1125	switch (*dec->cur >> MAJOR_SHIFT)
971	{	1126	{
972	case 0: // unsigned int	1127	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
973	return newSVuv (decode_uint (dec));	1128	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
974	case 1: // negative int	1129	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
975	return newSViv (-1 - (IV)decode_uint (dec));	1130	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
976	case 2: // octet string	1131	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
977	return decode_str (dec, 0);	1132	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
978	case 3: // utf-8 string	1133	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
979	return decode_str (dec, 1);	1134
980	case 4: // array	1135	case MAJOR_MISC >> MAJOR_SHIFT:
981	return decode_av (dec);
982	case 5: // map
983	return decode_hv (dec);
984	case 6: // tag
985	return decode_tagged (dec);
986	case 7: // misc
987	switch (*dec->cur++ & 31)	1136	switch (*dec->cur++ & MINOR_MASK)
988	{	1137	{
989	case 20:	1138	case SIMPLE_FALSE:
990	#if CBOR_SLOW	1139	#if CBOR_SLOW
991	types_false = get_bool ("Types::Serialiser::false");	1140	types_false = get_bool ("Types::Serialiser::false");
992	#endif	1141	#endif
993	return newSVsv (types_false);	1142	return newSVsv (types_false);
994	case 21:	1143	case SIMPLE_TRUE:
995	#if CBOR_SLOW	1144	#if CBOR_SLOW
996	types_true = get_bool ("Types::Serialiser::true");	1145	types_true = get_bool ("Types::Serialiser::true");
997	#endif	1146	#endif
998	return newSVsv (types_true);	1147	return newSVsv (types_true);
999	case 22:	1148	case SIMPLE_NULL:
1000	return newSVsv (&PL_sv_undef);	1149	return newSVsv (&PL_sv_undef);
1001	case 23:	1150	case SIMPLE_UNDEF:
1002	#if CBOR_SLOW	1151	#if CBOR_SLOW
1003	types_error = get_bool ("Types::Serialiser::error");	1152	types_error = get_bool ("Types::Serialiser::error");
1004	#endif	1153	#endif
1005	return newSVsv (types_error);	1154	return newSVsv (types_error);
1006		1155
1007	case 25:	1156	case MISC_FLOAT16:
1008	{	1157	{
1009	WANT (2);	1158	WANT (2);
1010		1159
1011	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1160	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1012	dec->cur += 2;	1161	dec->cur += 2;
1013		1162
1014	return newSVnv (ecb_binary16_to_float (fp));	1163	return newSVnv (ecb_binary16_to_float (fp));
1015	}	1164	}
1016		1165
1017	case 26:	1166	case MISC_FLOAT32:
1018	{	1167	{
1019	uint32_t fp;	1168	uint32_t fp;
1020	WANT (4);	1169	WANT (4);
1021	memcpy (&fp, dec->cur, 4);	1170	memcpy (&fp, dec->cur, 4);
1022	dec->cur += 4;	1171	dec->cur += 4;
…		…
1025	fp = ecb_bswap32 (fp);	1174	fp = ecb_bswap32 (fp);
1026		1175
1027	return newSVnv (ecb_binary32_to_float (fp));	1176	return newSVnv (ecb_binary32_to_float (fp));
1028	}	1177	}
1029		1178
1030	case 27:	1179	case MISC_FLOAT64:
1031	{	1180	{
1032	uint64_t fp;	1181	uint64_t fp;
1033	WANT (8);	1182	WANT (8);
1034	memcpy (&fp, dec->cur, 8);	1183	memcpy (&fp, dec->cur, 8);
1035	dec->cur += 8;	1184	dec->cur += 8;
…		…
1038	fp = ecb_bswap64 (fp);	1187	fp = ecb_bswap64 (fp);
1039		1188
1040	return newSVnv (ecb_binary64_to_double (fp));	1189	return newSVnv (ecb_binary64_to_double (fp));
1041	}	1190	}
1042		1191
1043	// 0..19 unassigned	1192	// 0..19 unassigned simple
1044	// 24 reserved + unassigned (reserved values are not encodable)	1193	// 24 reserved + unassigned simple (reserved values are not encodable)
		1194	// 28-30 unassigned misc
		1195	// 31 break code
1045	default:	1196	default:
1046	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1197	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1047	}	1198	}
1048		1199
1049	break;	1200	break;
1050	}	1201	}
1051		1202
…		…
1054	}	1205	}
1055		1206
1056	static SV *	1207	static SV *
1057	decode_cbor (SV string, CBOR cbor, char **offset_return)	1208	decode_cbor (SV string, CBOR cbor, char **offset_return)
1058	{	1209	{
1059	dec_t dec = { };	1210	dec_t dec = { 0 };
1060	SV *sv;	1211	SV *sv;
1061	STRLEN len;	1212	STRLEN len;
1062	char *data = SvPVbyte (string, len);	1213	char *data = SvPVbyte (string, len);
1063		1214
1064	if (len > cbor->max_size && cbor->max_size)	1215	if (len > cbor->max_size && cbor->max_size)
…		…
1078	if (dec.cur != dec.end && !dec.err)	1229	if (dec.cur != dec.end && !dec.err)
1079	dec.err = "garbage after CBOR object";	1230	dec.err = "garbage after CBOR object";
1080		1231
1081	if (dec.err)	1232	if (dec.err)
1082	{	1233	{
		1234	if (dec.shareable)
		1235	{
		1236	// need to break cyclic links, which whould all be in shareable
		1237	int i;
		1238	SV **svp;
		1239
		1240	for (i = av_len (dec.shareable) + 1; i--; )
		1241	if ((svp = av_fetch (dec.shareable, i, 0)))
		1242	sv_setsv (*svp, &PL_sv_undef);
		1243	}
		1244
1083	SvREFCNT_dec (sv);	1245	SvREFCNT_dec (sv);
1084	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1246	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1085	}	1247	}
1086		1248
1087	sv = sv_2mortal (sv);	1249	sv = sv_2mortal (sv);
1088		1250
1089	return sv;	1251	return sv;
1090	}	1252	}
1091		1253
		1254	/////////////////////////////////////////////////////////////////////////////
		1255	// incremental parser
		1256
		1257	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1258
		1259	// returns 0 for notyet, 1 for success or error
		1260	static int
		1261	incr_parse (CBOR self, SV cborstr)
		1262	{
		1263	STRLEN cur;
		1264	SvPV (cborstr, cur);
		1265
		1266	while (ecb_expect_true (self->incr_need <= cur))
		1267	{
		1268	// table of integer count bytes
		1269	static I8 incr_len[MINOR_MASK + 1] = {
		1270	0, 0, 0, 0, 0, 0, 0, 0,
		1271	0, 0, 0, 0, 0, 0, 0, 0,
		1272	0, 0, 0, 0, 0, 0, 0, 0,
		1273	1, 2, 4, 8,-1,-1,-1,-2
		1274	};
		1275
		1276	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1277	U8 m = *p & MINOR_MASK;
		1278	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1279	I8 ilen = incr_len[m];
		1280
		1281	self->incr_need = self->incr_pos + 1;
		1282
		1283	if (ecb_expect_false (ilen < 0))
		1284	{
		1285	if (m != MINOR_INDEF)
		1286	return 1; // error
		1287
		1288	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1289	{
		1290	if (count >= 0)
		1291	return 1; // error
		1292
		1293	count = 1;
		1294	}
		1295	else
		1296	{
		1297	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1298	count = -1;
		1299	}
		1300	}
		1301	else
		1302	{
		1303	self->incr_need += ilen;
		1304	if (ecb_expect_false (self->incr_need > cur))
		1305	return 0;
		1306
		1307	int major = *p >> MAJOR_SHIFT;
		1308
		1309	switch (major)
		1310	{
		1311	case MAJOR_TAG >> MAJOR_SHIFT:
		1312	++count; // tags merely prefix another value
		1313	break;
		1314
		1315	case MAJOR_BYTES >> MAJOR_SHIFT:
		1316	case MAJOR_TEXT >> MAJOR_SHIFT:
		1317	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1318	case MAJOR_MAP >> MAJOR_SHIFT:
		1319	{
		1320	UV len;
		1321
		1322	if (ecb_expect_false (ilen))
		1323	{
		1324	len = 0;
		1325
		1326	do {
		1327	len = (len << 8) \| *++p;
		1328	} while (--ilen);
		1329	}
		1330	else
		1331	len = m;
		1332
		1333	switch (major)
		1334	{
		1335	case MAJOR_BYTES >> MAJOR_SHIFT:
		1336	case MAJOR_TEXT >> MAJOR_SHIFT:
		1337	self->incr_need += len;
		1338	if (ecb_expect_false (self->incr_need > cur))
		1339	return 0;
		1340
		1341	break;
		1342
		1343	case MAJOR_MAP >> MAJOR_SHIFT:
		1344	len <<= 1;
		1345	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1346	if (len)
		1347	{
		1348	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1349	count = len + 1;
		1350	}
		1351	break;
		1352	}
		1353	}
		1354	}
		1355	}
		1356
		1357	self->incr_pos = self->incr_need;
		1358
		1359	if (count > 0)
		1360	{
		1361	while (!--count)
		1362	{
		1363	if (!AvFILLp (self->incr_count))
		1364	return 1; // done
		1365
		1366	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1367	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1368	}
		1369
		1370	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1371	}
		1372	}
		1373
		1374	return 0;
		1375	}
		1376
		1377
1092	/////////////////////////////////////////////////////////////////////////////	1378	/////////////////////////////////////////////////////////////////////////////
1093	// XS interface functions	1379	// XS interface functions
1094		1380
1095	MODULE = CBOR::XS PACKAGE = CBOR::XS	1381	MODULE = CBOR::XS PACKAGE = CBOR::XS
1096		1382
…		…
1108		1394
1109	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1395	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1110		1396
1111	sv_cbor = newSVpv ("CBOR", 0);	1397	sv_cbor = newSVpv ("CBOR", 0);
1112	SvREADONLY_on (sv_cbor);	1398	SvREADONLY_on (sv_cbor);
		1399
		1400	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1113	}	1401	}
1114		1402
1115	PROTOTYPES: DISABLE	1403	PROTOTYPES: DISABLE
1116		1404
1117	void CLONE (...)	1405	void CLONE (...)
…		…
1136	void shrink (CBOR *self, int enable = 1)	1424	void shrink (CBOR *self, int enable = 1)
1137	ALIAS:	1425	ALIAS:
1138	shrink = F_SHRINK	1426	shrink = F_SHRINK
1139	allow_unknown = F_ALLOW_UNKNOWN	1427	allow_unknown = F_ALLOW_UNKNOWN
1140	allow_sharing = F_ALLOW_SHARING	1428	allow_sharing = F_ALLOW_SHARING
1141	allow_stringref = F_ALLOW_STRINGREF	1429	allow_cycles = F_ALLOW_CYCLES
		1430	pack_strings = F_PACK_STRINGS
		1431	text_keys = F_TEXT_KEYS
		1432	text_strings = F_TEXT_STRINGS
		1433	validate_utf8 = F_VALIDATE_UTF8
1142	PPCODE:	1434	PPCODE:
1143	{	1435	{
1144	if (enable)	1436	if (enable)
1145	self->flags \|= ix;	1437	self->flags \|= ix;
1146	else	1438	else
…		…
1152	void get_shrink (CBOR *self)	1444	void get_shrink (CBOR *self)
1153	ALIAS:	1445	ALIAS:
1154	get_shrink = F_SHRINK	1446	get_shrink = F_SHRINK
1155	get_allow_unknown = F_ALLOW_UNKNOWN	1447	get_allow_unknown = F_ALLOW_UNKNOWN
1156	get_allow_sharing = F_ALLOW_SHARING	1448	get_allow_sharing = F_ALLOW_SHARING
1157	get_allow_stringref = F_ALLOW_STRINGREF	1449	get_allow_cycles = F_ALLOW_CYCLES
		1450	get_pack_strings = F_PACK_STRINGS
		1451	get_text_keys = F_TEXT_KEYS
		1452	get_text_strings = F_TEXT_STRINGS
		1453	get_validate_utf8 = F_VALIDATE_UTF8
1158	PPCODE:	1454	PPCODE:
1159	XPUSHs (boolSV (self->flags & ix));	1455	XPUSHs (boolSV (self->flags & ix));
1160		1456
1161	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1457	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1162	PPCODE:	1458	PPCODE:
…		…
1211	EXTEND (SP, 2);	1507	EXTEND (SP, 2);
1212	PUSHs (sv);	1508	PUSHs (sv);
1213	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1509	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1214	}	1510	}
1215		1511
		1512	void incr_parse (CBOR self, SV cborstr)
		1513	ALIAS:
		1514	incr_parse_multiple = 1
		1515	PPCODE:
		1516	{
		1517	if (SvUTF8 (cborstr))
		1518	sv_utf8_downgrade (cborstr, 0);
		1519
		1520	if (!self->incr_count)
		1521	{
		1522	self->incr_count = newAV ();
		1523	self->incr_pos = 0;
		1524	self->incr_need = 1;
		1525
		1526	av_push (self->incr_count, newSViv (1));
		1527	}
		1528
		1529	do
		1530	{
		1531	if (!incr_parse (self, cborstr))
		1532	{
		1533	if (self->incr_need > self->max_size && self->max_size)
		1534	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1535	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1536
		1537	break;
		1538	}
		1539
		1540	SV *sv;
		1541	char *offset;
		1542
		1543	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1544	XPUSHs (sv);
		1545
		1546	sv_chop (cborstr, offset);
		1547
		1548	av_clear (self->incr_count);
		1549	av_push (self->incr_count, newSViv (1));
		1550
		1551	self->incr_pos = 0;
		1552	self->incr_need = self->incr_pos + 1;
		1553	}
		1554	while (ix);
		1555	}
		1556
		1557	void incr_reset (CBOR *self)
		1558	CODE:
		1559	{
		1560	SvREFCNT_dec (self->incr_count);
		1561	self->incr_count = 0;
		1562	}
		1563
1216	void DESTROY (CBOR *self)	1564	void DESTROY (CBOR *self)
1217	PPCODE:	1565	PPCODE:
1218	cbor_free (self);	1566	cbor_free (self);
1219		1567
1220	PROTOTYPES: ENABLE	1568	PROTOTYPES: ENABLE
1221		1569
1222	void encode_cbor (SV *scalar)	1570	void encode_cbor (SV *scalar)
		1571	ALIAS:
		1572	encode_cbor = 0
		1573	encode_cbor_sharing = F_ALLOW_SHARING
1223	PPCODE:	1574	PPCODE:
1224	{	1575	{
1225	CBOR cbor;	1576	CBOR cbor;
1226	cbor_init (&cbor);	1577	cbor_init (&cbor);
		1578	cbor.flags \|= ix;
1227	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1579	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1228	XPUSHs (scalar);	1580	XPUSHs (scalar);
1229	}	1581	}
1230		1582
1231	void decode_cbor (SV *cborstr)	1583	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs. Revision 1.56 by root, Fri Nov 25 11:33:03 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs.
Revision 1.56 by root, Fri Nov 25 11:33:03 2016 UTC