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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs.
Revision 1.54 by root, Sun Apr 24 20:08:41 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.
…		…
345		430
346	HV *stash = SvSTASH (sv);	431	HV *stash = SvSTASH (sv);
347		432
348	if (stash == boolean_stash)	433	if (stash == boolean_stash)
349	{	434	{
350	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	435	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
351	return;	436	return;
352	}	437	}
353	else if (stash == error_stash)	438	else if (stash == error_stash)
354	{	439	{
355	encode_ch (enc, 0xe0 \| 23);	440	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
356	return;	441	return;
357	}	442	}
358	else if (stash == tagged_stash)	443	else if (stash == tagged_stash)
359	{	444	{
360	if (svt != SVt_PVAV)	445	if (svt != SVt_PVAV)
361	croak ("encountered CBOR::XS::Tagged object that isn't an array");	446	croak ("encountered CBOR::XS::Tagged object that isn't an array");
362		447
363	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	448	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
364	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	449	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
365		450
366	return;	451	return;
367	}	452	}
368	}	453	}
…		…
376	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);	461	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
377		462
378	if (SvOK (*svp))	463	if (SvOK (*svp))
379	{	464	{
380	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);	465	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381	encode_uint (enc, 0x00, SvUV (*svp));	466	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
382	return;	467	return;
383	}	468	}
384	else	469	else
385	{	470	{
386	sv_setuv (*svp, enc->shareable_idx);	471	sv_setuv (*svp, enc->shareable_idx);
…		…
396		481
397	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	482	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
398	{	483	{
399	dSP;	484	dSP;
400		485
401	ENTER; SAVETMPS; PUSHMARK (SP);	486	ENTER; SAVETMPS;
		487	PUSHMARK (SP);
402	// 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
403	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	489	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
404		490
405	PUTBACK;	491	PUTBACK;
406	// G_SCALAR ensures that return value is 1	492	// G_SCALAR ensures that return value is 1
…		…
419	}	505	}
420	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	506	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
421	{	507	{
422	dSP;	508	dSP;
423		509
424	ENTER; SAVETMPS; PUSHMARK (SP);	510	ENTER; SAVETMPS;
		511	SAVESTACK_POS ();
		512	PUSHMARK (SP);
425	EXTEND (SP, 2);	513	EXTEND (SP, 2);
426	// 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
427	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	515	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
428	PUSHs (sv_cbor);	516	PUSHs (sv_cbor);
429		517
…		…
434	// catch this surprisingly common error	522	// catch this surprisingly common error
435	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	523	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
436	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));
437		525
438	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	526	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
439	encode_uint (enc, 0x80, count + 1);	527	encode_uint (enc, MAJOR_ARRAY, count + 1);
440	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	528	encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
441		529
442	while (count)	530	while (count)
443	encode_sv (enc, SP[1 - count--]);	531	encode_sv (enc, SP[1 - count--]);
444		532
445	PUTBACK;	533	PUTBACK;
…		…
466	{	554	{
467	double nv = SvNVX (sv);	555	double nv = SvNVX (sv);
468		556
469	need (enc, 9);	557	need (enc, 9);
470		558
471	if (ecb_expect_false (nv == (U32)nv))	559	if (ecb_expect_false (nv == (NV)(U32)nv))
472	encode_uint (enc, 0x00, (U32)nv);	560	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
473	//TODO: maybe I32?	561	//TODO: maybe I32?
474	else if (ecb_expect_false (nv == (float)nv))	562	else if (ecb_expect_false (nv == (float)nv))
475	{	563	{
476	uint32_t fp = ecb_float_to_binary32 (nv);	564	uint32_t fp = ecb_float_to_binary32 (nv);
477		565
478	*enc->cur++ = 0xe0 \| 26;	566	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
479		567
480	if (!ecb_big_endian ())	568	if (!ecb_big_endian ())
481	fp = ecb_bswap32 (fp);	569	fp = ecb_bswap32 (fp);
482		570
483	memcpy (enc->cur, &fp, 4);	571	memcpy (enc->cur, &fp, 4);
…		…
485	}	573	}
486	else	574	else
487	{	575	{
488	uint64_t fp = ecb_double_to_binary64 (nv);	576	uint64_t fp = ecb_double_to_binary64 (nv);
489		577
490	*enc->cur++ = 0xe0 \| 27;	578	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
491		579
492	if (!ecb_big_endian ())	580	if (!ecb_big_endian ())
493	fp = ecb_bswap64 (fp);	581	fp = ecb_bswap64 (fp);
494		582
495	memcpy (enc->cur, &fp, 8);	583	memcpy (enc->cur, &fp, 8);
…		…
504		592
505	if (SvPOKp (sv))	593	if (SvPOKp (sv))
506	{	594	{
507	STRLEN len;	595	STRLEN len;
508	char *str = SvPV (sv, len);	596	char *str = SvPV (sv, len);
509	encode_strref (enc, SvUTF8 (sv), str, len);	597	encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
510	}	598	}
511	else if (SvNOKp (sv))	599	else if (SvNOKp (sv))
512	encode_nv (enc, sv);	600	encode_nv (enc, sv);
513	else if (SvIOKp (sv))	601	else if (SvIOKp (sv))
514	{	602	{
515	if (SvIsUV (sv))	603	if (SvIsUV (sv))
516	encode_uint (enc, 0x00, SvUVX (sv));	604	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
517	else if (SvIVX (sv) >= 0)	605	else if (SvIVX (sv) >= 0)
518	encode_uint (enc, 0x00, SvIVX (sv));	606	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
519	else	607	else
520	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	608	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
521	}	609	}
522	else if (SvROK (sv))	610	else if (SvROK (sv))
523	encode_rv (enc, SvRV (sv));	611	encode_rv (enc, SvRV (sv));
524	else if (!SvOK (sv))	612	else if (!SvOK (sv))
525	encode_ch (enc, 0xe0 \| 22);	613	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
526	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	614	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
527	encode_ch (enc, 0xe0 \| 23);	615	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
528	else	616	else
529	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",
530	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	618	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
531	}	619	}
532		620
533	static SV *	621	static SV *
534	encode_cbor (SV scalar, CBOR cbor)	622	encode_cbor (SV scalar, CBOR cbor)
535	{	623	{
536	enc_t enc = { };	624	enc_t enc = { 0 };
537		625
538	enc.cbor = *cbor;	626	enc.cbor = *cbor;
539	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	627	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
540	enc.cur = SvPVX (enc.sv);	628	enc.cur = SvPVX (enc.sv);
541	enc.end = SvEND (enc.sv);	629	enc.end = SvEND (enc.sv);
…		…
585	#define DEC_DEC_DEPTH --dec->depth	673	#define DEC_DEC_DEPTH --dec->depth
586		674
587	static UV	675	static UV
588	decode_uint (dec_t *dec)	676	decode_uint (dec_t *dec)
589	{	677	{
590	switch (*dec->cur & 31)	678	U8 m = *dec->cur & MINOR_MASK;
591	{	679	++dec->cur;
592	case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
593	case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
594	case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
595	return *dec->cur++ & 31;
596		680
597	case 24:	681	if (ecb_expect_true (m < LENGTH_EXT1))
		682	return m;
		683	else if (ecb_expect_true (m == LENGTH_EXT1))
		684	{
598	WANT (2);	685	WANT (1);
599	dec->cur += 2;	686	dec->cur += 1;
600	return dec->cur[-1];	687	return dec->cur[-1];
601		688	}
602	case 25:	689	else if (ecb_expect_true (m == LENGTH_EXT2))
		690	{
603	WANT (3);	691	WANT (2);
604	dec->cur += 3;	692	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	693	return (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	694	\| ((UV)dec->cur[-1]);
607		695	}
608	case 26:	696	else if (ecb_expect_true (m == LENGTH_EXT4))
		697	{
609	WANT (5);	698	WANT (4);
610	dec->cur += 5;	699	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	700	return (((UV)dec->cur[-4]) << 24)
612	\| (((UV)dec->cur[-3]) << 16)	701	\| (((UV)dec->cur[-3]) << 16)
613	\| (((UV)dec->cur[-2]) << 8)	702	\| (((UV)dec->cur[-2]) << 8)
614	\| ((UV)dec->cur[-1]);	703	\| ((UV)dec->cur[-1]);
615		704	}
616	case 27:	705	else if (ecb_expect_true (m == LENGTH_EXT8))
		706	{
617	WANT (9);	707	WANT (8);
618	dec->cur += 9;	708	dec->cur += 8;
619		709
620	return	710	return
621	#if UVSIZE < 8	711	#if UVSIZE < 8
622	0	712	0
623	#else	713	#else
624	(((UV)dec->cur[-8]) << 56)	714	(((UV)dec->cur[-8]) << 56)
625	\| (((UV)dec->cur[-7]) << 48)	715	\| (((UV)dec->cur[-7]) << 48)
626	\| (((UV)dec->cur[-6]) << 40)	716	\| (((UV)dec->cur[-6]) << 40)
627	\| (((UV)dec->cur[-5]) << 32)	717	\| (((UV)dec->cur[-5]) << 32)
628	#endif	718	#endif
629	\| (((UV)dec->cur[-4]) << 24)	719	\| (((UV)dec->cur[-4]) << 24)
630	\| (((UV)dec->cur[-3]) << 16)	720	\| (((UV)dec->cur[-3]) << 16)
631	\| (((UV)dec->cur[-2]) << 8)	721	\| (((UV)dec->cur[-2]) << 8)
632	\| ((UV)dec->cur[-1]);	722	\| ((UV)dec->cur[-1]);
633		723	}
634	default:	724	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	725	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		726
638	fail:	727	fail:
639	return 0;	728	return 0;
640	}	729	}
641		730
…		…
646	{	735	{
647	AV *av = newAV ();	736	AV *av = newAV ();
648		737
649	DEC_INC_DEPTH;	738	DEC_INC_DEPTH;
650		739
651	if ((*dec->cur & 31) == 31)	740	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
652	{	741	{
653	++dec->cur;	742	++dec->cur;
654		743
655	for (;;)	744	for (;;)
656	{	745	{
657	WANT (1);	746	WANT (1);
658		747
659	if (*dec->cur == (0xe0 \| 31))	748	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
660	{	749	{
661	++dec->cur;	750	++dec->cur;
662	break;	751	break;
663	}	752	}
664		753
…		…
667	}	756	}
668	else	757	else
669	{	758	{
670	int i, len = decode_uint (dec);	759	int i, len = decode_uint (dec);
671		760
		761	WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
672	av_fill (av, len - 1);	762	av_fill (av, len - 1);
673		763
674	for (i = 0; i < len; ++i)	764	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	765	AvARRAY (av)[i] = decode_sv (dec);
676	}	766	}
…		…
689	{	779	{
690	// for speed reasons, we specialcase single-string	780	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	781	// byte or utf-8 strings as keys, but only when !stringref
692		782
693	if (ecb_expect_true (!dec->stringref))	783	if (ecb_expect_true (!dec->stringref))
694	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	784	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	785	{
696	I32 len = decode_uint (dec);	786	I32 len = decode_uint (dec);
697	char key = (char )dec->cur;	787	char key = (char )dec->cur;
698		788
		789	WANT (len);
699	dec->cur += len;	790	dec->cur += len;
700		791
701	if (ecb_expect_false (dec->stringref))
702	av_push (dec->stringref, newSVpvn (key, len));
703
704	hv_store (hv, key, len, decode_sv (dec), 0);	792	hv_store (hv, key, len, decode_sv (dec), 0);
705		793
706	return;	794	return;
707	}	795	}
708	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	796	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	797	{
710	I32 len = decode_uint (dec);	798	I32 len = decode_uint (dec);
711	char key = (char )dec->cur;	799	char key = (char )dec->cur;
712		800
		801	WANT (len);
713	dec->cur += len;	802	dec->cur += len;
714		803
715	if (ecb_expect_false (dec->stringref))	804	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	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)");
717		807
718	hv_store (hv, key, -len, decode_sv (dec), 0);	808	hv_store (hv, key, -len, decode_sv (dec), 0);
719		809
720	return;	810	return;
721	}	811	}
…		…
723	SV *k = decode_sv (dec);	813	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	814	SV *v = decode_sv (dec);
725		815
726	hv_store_ent (hv, k, v, 0);	816	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	817	SvREFCNT_dec (k);
		818
		819	fail:
		820	;
728	}	821	}
729		822
730	static SV *	823	static SV *
731	decode_hv (dec_t *dec)	824	decode_hv (dec_t *dec)
732	{	825	{
733	HV *hv = newHV ();	826	HV *hv = newHV ();
734		827
735	DEC_INC_DEPTH;	828	DEC_INC_DEPTH;
736		829
737	if ((*dec->cur & 31) == 31)	830	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
738	{	831	{
739	++dec->cur;	832	++dec->cur;
740		833
741	for (;;)	834	for (;;)
742	{	835	{
743	WANT (1);	836	WANT (1);
744		837
745	if (*dec->cur == (0xe0 \| 31))	838	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
746	{	839	{
747	++dec->cur;	840	++dec->cur;
748	break;	841	break;
749	}	842	}
750		843
…		…
771	static SV *	864	static SV *
772	decode_str (dec_t *dec, int utf8)	865	decode_str (dec_t *dec, int utf8)
773	{	866	{
774	SV *sv = 0;	867	SV *sv = 0;
775		868
776	if ((*dec->cur & 31) == 31)	869	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
777	{	870	{
778	// indefinite length strings	871	// indefinite length strings
779	++dec->cur;	872	++dec->cur;
780		873
781	unsigned char major = *dec->cur & 0xe0;	874	U8 major = *dec->cur & MAJOR_MISC;
782		875
783	sv = newSVpvn ("", 0);	876	sv = newSVpvn ("", 0);
784		877
785	for (;;)	878	for (;;)
786	{	879	{
787	WANT (1);	880	WANT (1);
788		881
789	if ((*dec->cur ^ major) >= 31)	882	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 \| 31))	883	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
791	{	884	{
792	++dec->cur;	885	++dec->cur;
793	break;	886	break;
794	}	887	}
795	else	888	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	907	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	908	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	909	}
817		910
818	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
819	SvUTF8_on (sv);	917	SvUTF8_on (sv);
		918	}
820		919
821	return sv;	920	return sv;
822		921
823	fail:	922	fail:
824	SvREFCNT_dec (sv);	923	SvREFCNT_dec (sv);
…		…
856	}	955	}
857	break;	956	break;
858		957
859	case CBOR_TAG_STRINGREF:	958	case CBOR_TAG_STRINGREF:
860	{	959	{
861	if ((*dec->cur >> 5) != 0)	960	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	961	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		962
864	UV idx = decode_uint (dec);	963	UV idx = decode_uint (dec);
865		964
866	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	965	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	972	case CBOR_TAG_VALUE_SHAREABLE:
874	{	973	{
875	if (ecb_expect_false (!dec->shareable))	974	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	975	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
877		976
		977	if (dec->cbor.flags & F_ALLOW_CYCLES)
		978	{
878	sv = newSV (0);	979	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	980	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		981
881	SV *osv = decode_sv (dec);	982	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	983	sv_setsv (sv, osv);
883	SvREFCNT_dec_NN (osv);	984	SvREFCNT_dec_NN (osv);
		985	}
		986	else
		987	{
		988	av_push (dec->shareable, &PL_sv_undef);
		989	int idx = AvFILLp (dec->shareable);
		990	sv = decode_sv (dec);
		991	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		992	}
884	}	993	}
885	break;	994	break;
886		995
887	case CBOR_TAG_VALUE_SHAREDREF:	996	case CBOR_TAG_VALUE_SHAREDREF:
888	{	997	{
889	if ((*dec->cur >> 5) != 0)	998	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	999	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		1000
892	UV idx = decode_uint (dec);	1001	UV idx = decode_uint (dec);
893		1002
894	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	1003	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1004	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		1005
897	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1006	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1007
		1008	if (sv == &PL_sv_undef)
		1009	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
898	}	1010	}
899	break;	1011	break;
900		1012
901	case CBOR_TAG_PERL_OBJECT:	1013	case CBOR_TAG_PERL_OBJECT:
902	{	1014	{
…		…
917	if (!method)	1029	if (!method)
918	ERR ("cannot decode perl-object (package does not have a THAW method)");	1030	ERR ("cannot decode perl-object (package does not have a THAW method)");
919		1031
920	dSP;	1032	dSP;
921		1033
922	ENTER; SAVETMPS; PUSHMARK (SP);	1034	ENTER; SAVETMPS;
		1035	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1036	EXTEND (SP, len + 1);
924	// we re-bless the reference to get overload and other niceties right	1037	// we re-bless the reference to get overload and other niceties right
925	PUSHs (*av_fetch (av, 0, 1));	1038	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1039	PUSHs (sv_cbor);
927		1040
…		…
952	default:	1065	default:
953	{	1066	{
954	sv = decode_sv (dec);	1067	sv = decode_sv (dec);
955		1068
956	dSP;	1069	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1070	ENTER; SAVETMPS;
		1071	SAVESTACK_POS ();
		1072	PUSHMARK (SP);
958	EXTEND (SP, 2);	1073	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1074	PUSHs (newSVuv (tag));
960	PUSHs (sv);	1075	PUSHs (sv);
961		1076
962	PUTBACK;	1077	PUTBACK;
…		…
1003	static SV *	1118	static SV *
1004	decode_sv (dec_t *dec)	1119	decode_sv (dec_t *dec)
1005	{	1120	{
1006	WANT (1);	1121	WANT (1);
1007		1122
1008	switch (*dec->cur >> 5)	1123	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1124	{
1010	case 0: // unsigned int	1125	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1126	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1127	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1128	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1129	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1130	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1131	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1132
1018	case 4: // array	1133	case MAJOR_MISC >> MAJOR_SHIFT:
1019	return decode_av (dec);
1020	case 5: // map
1021	return decode_hv (dec);
1022	case 6: // tag
1023	return decode_tagged (dec);
1024	case 7: // misc
1025	switch (*dec->cur++ & 31)	1134	switch (*dec->cur++ & MINOR_MASK)
1026	{	1135	{
1027	case 20:	1136	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1137	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1138	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1139	#endif
1031	return newSVsv (types_false);	1140	return newSVsv (types_false);
1032	case 21:	1141	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1142	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1143	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1144	#endif
1036	return newSVsv (types_true);	1145	return newSVsv (types_true);
1037	case 22:	1146	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1147	return newSVsv (&PL_sv_undef);
1039	case 23:	1148	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1149	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1150	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1151	#endif
1043	return newSVsv (types_error);	1152	return newSVsv (types_error);
1044		1153
1045	case 25:	1154	case MISC_FLOAT16:
1046	{	1155	{
1047	WANT (2);	1156	WANT (2);
1048		1157
1049	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1158	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1050	dec->cur += 2;	1159	dec->cur += 2;
1051		1160
1052	return newSVnv (ecb_binary16_to_float (fp));	1161	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1162	}
1054		1163
1055	case 26:	1164	case MISC_FLOAT32:
1056	{	1165	{
1057	uint32_t fp;	1166	uint32_t fp;
1058	WANT (4);	1167	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1168	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1169	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1172	fp = ecb_bswap32 (fp);
1064		1173
1065	return newSVnv (ecb_binary32_to_float (fp));	1174	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1175	}
1067		1176
1068	case 27:	1177	case MISC_FLOAT64:
1069	{	1178	{
1070	uint64_t fp;	1179	uint64_t fp;
1071	WANT (8);	1180	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1181	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1182	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1185	fp = ecb_bswap64 (fp);
1077		1186
1078	return newSVnv (ecb_binary64_to_double (fp));	1187	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1188	}
1080		1189
1081	// 0..19 unassigned	1190	// 0..19 unassigned simple
1082	// 24 reserved + unassigned (reserved values are not encodable)	1191	// 24 reserved + unassigned simple (reserved values are not encodable)
		1192	// 28-30 unassigned misc
		1193	// 31 break code
1083	default:	1194	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1195	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1196	}
1086		1197
1087	break;	1198	break;
1088	}	1199	}
1089		1200
…		…
1092	}	1203	}
1093		1204
1094	static SV *	1205	static SV *
1095	decode_cbor (SV string, CBOR cbor, char **offset_return)	1206	decode_cbor (SV string, CBOR cbor, char **offset_return)
1096	{	1207	{
1097	dec_t dec = { };	1208	dec_t dec = { 0 };
1098	SV *sv;	1209	SV *sv;
1099	STRLEN len;	1210	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1211	char *data = SvPVbyte (string, len);
1101		1212
1102	if (len > cbor->max_size && cbor->max_size)	1213	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1227	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1228	dec.err = "garbage after CBOR object";
1118		1229
1119	if (dec.err)	1230	if (dec.err)
1120	{	1231	{
		1232	if (dec.shareable)
		1233	{
		1234	// need to break cyclic links, which whould all be in shareable
		1235	int i;
		1236	SV **svp;
		1237
		1238	for (i = av_len (dec.shareable) + 1; i--; )
		1239	if ((svp = av_fetch (dec.shareable, i, 0)))
		1240	sv_setsv (*svp, &PL_sv_undef);
		1241	}
		1242
1121	SvREFCNT_dec (sv);	1243	SvREFCNT_dec (sv);
1122	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1244	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1123	}	1245	}
1124		1246
1125	sv = sv_2mortal (sv);	1247	sv = sv_2mortal (sv);
1126		1248
1127	return sv;	1249	return sv;
1128	}	1250	}
1129		1251
		1252	/////////////////////////////////////////////////////////////////////////////
		1253	// incremental parser
		1254
		1255	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1256
		1257	// returns 0 for notyet, 1 for success or error
		1258	static int
		1259	incr_parse (CBOR self, SV cborstr)
		1260	{
		1261	STRLEN cur;
		1262	SvPV (cborstr, cur);
		1263
		1264	while (ecb_expect_true (self->incr_need <= cur))
		1265	{
		1266	// table of integer count bytes
		1267	static I8 incr_len[MINOR_MASK + 1] = {
		1268	0, 0, 0, 0, 0, 0, 0, 0,
		1269	0, 0, 0, 0, 0, 0, 0, 0,
		1270	0, 0, 0, 0, 0, 0, 0, 0,
		1271	1, 2, 4, 8,-1,-1,-1,-2
		1272	};
		1273
		1274	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1275	U8 m = *p & MINOR_MASK;
		1276	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1277	I8 ilen = incr_len[m];
		1278
		1279	self->incr_need = self->incr_pos + 1;
		1280
		1281	if (ecb_expect_false (ilen < 0))
		1282	{
		1283	if (m != MINOR_INDEF)
		1284	return 1; // error
		1285
		1286	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1287	{
		1288	if (count >= 0)
		1289	return 1; // error
		1290
		1291	count = 1;
		1292	}
		1293	else
		1294	{
		1295	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1296	count = -1;
		1297	}
		1298	}
		1299	else
		1300	{
		1301	self->incr_need += ilen;
		1302	if (ecb_expect_false (self->incr_need > cur))
		1303	return 0;
		1304
		1305	int major = *p >> MAJOR_SHIFT;
		1306
		1307	switch (major)
		1308	{
		1309	case MAJOR_TAG >> MAJOR_SHIFT:
		1310	++count; // tags merely prefix another value
		1311	break;
		1312
		1313	case MAJOR_BYTES >> MAJOR_SHIFT:
		1314	case MAJOR_TEXT >> MAJOR_SHIFT:
		1315	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1316	case MAJOR_MAP >> MAJOR_SHIFT:
		1317	{
		1318	UV len;
		1319
		1320	if (ecb_expect_false (ilen))
		1321	{
		1322	len = 0;
		1323
		1324	do {
		1325	len = (len << 8) \| *++p;
		1326	} while (--ilen);
		1327	}
		1328	else
		1329	len = m;
		1330
		1331	switch (major)
		1332	{
		1333	case MAJOR_BYTES >> MAJOR_SHIFT:
		1334	case MAJOR_TEXT >> MAJOR_SHIFT:
		1335	self->incr_need += len;
		1336	if (ecb_expect_false (self->incr_need > cur))
		1337	return 0;
		1338
		1339	break;
		1340
		1341	case MAJOR_MAP >> MAJOR_SHIFT:
		1342	len <<= 1;
		1343	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1344	if (len)
		1345	{
		1346	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1347	count = len + 1;
		1348	}
		1349	break;
		1350	}
		1351	}
		1352	}
		1353	}
		1354
		1355	self->incr_pos = self->incr_need;
		1356
		1357	if (count > 0)
		1358	{
		1359	while (!--count)
		1360	{
		1361	if (!AvFILLp (self->incr_count))
		1362	return 1; // done
		1363
		1364	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1365	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1366	}
		1367
		1368	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1369	}
		1370	}
		1371
		1372	return 0;
		1373	}
		1374
		1375
1130	/////////////////////////////////////////////////////////////////////////////	1376	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1377	// XS interface functions
1132		1378
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1379	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1380
…		…
1174	void shrink (CBOR *self, int enable = 1)	1420	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1421	ALIAS:
1176	shrink = F_SHRINK	1422	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1423	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1424	allow_sharing = F_ALLOW_SHARING
		1425	allow_cycles = F_ALLOW_CYCLES
1179	pack_strings = F_PACK_STRINGS	1426	pack_strings = F_PACK_STRINGS
		1427	text_keys = F_TEXT_KEYS
		1428	text_strings = F_TEXT_STRINGS
		1429	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1430	PPCODE:
1181	{	1431	{
1182	if (enable)	1432	if (enable)
1183	self->flags \|= ix;	1433	self->flags \|= ix;
1184	else	1434	else
…		…
1190	void get_shrink (CBOR *self)	1440	void get_shrink (CBOR *self)
1191	ALIAS:	1441	ALIAS:
1192	get_shrink = F_SHRINK	1442	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1443	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1444	get_allow_sharing = F_ALLOW_SHARING
		1445	get_allow_cycles = F_ALLOW_CYCLES
1195	get_pack_strings = F_PACK_STRINGS	1446	get_pack_strings = F_PACK_STRINGS
		1447	get_text_keys = F_TEXT_KEYS
		1448	get_text_strings = F_TEXT_STRINGS
		1449	get_validate_utf8 = F_VALIDATE_UTF8
1196	PPCODE:	1450	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1451	XPUSHs (boolSV (self->flags & ix));
1198		1452
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1453	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1454	PPCODE:
…		…
1249	EXTEND (SP, 2);	1503	EXTEND (SP, 2);
1250	PUSHs (sv);	1504	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1505	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1506	}
1253		1507
		1508	void incr_parse (CBOR self, SV cborstr)
		1509	ALIAS:
		1510	incr_parse_multiple = 1
		1511	PPCODE:
		1512	{
		1513	if (SvUTF8 (cborstr))
		1514	sv_utf8_downgrade (cborstr, 0);
		1515
		1516	if (!self->incr_count)
		1517	{
		1518	self->incr_count = newAV ();
		1519	self->incr_pos = 0;
		1520	self->incr_need = 1;
		1521
		1522	av_push (self->incr_count, newSViv (1));
		1523	}
		1524
		1525	do
		1526	{
		1527	if (!incr_parse (self, cborstr))
		1528	{
		1529	if (self->incr_need > self->max_size && self->max_size)
		1530	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1531	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1532
		1533	break;
		1534	}
		1535
		1536	SV *sv;
		1537	char *offset;
		1538
		1539	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1540	XPUSHs (sv);
		1541
		1542	sv_chop (cborstr, offset);
		1543
		1544	av_clear (self->incr_count);
		1545	av_push (self->incr_count, newSViv (1));
		1546
		1547	self->incr_pos = 0;
		1548	self->incr_need = self->incr_pos + 1;
		1549	}
		1550	while (ix);
		1551	}
		1552
		1553	void incr_reset (CBOR *self)
		1554	CODE:
		1555	{
		1556	SvREFCNT_dec (self->incr_count);
		1557	self->incr_count = 0;
		1558	}
		1559
1254	void DESTROY (CBOR *self)	1560	void DESTROY (CBOR *self)
1255	PPCODE:	1561	PPCODE:
1256	cbor_free (self);	1562	cbor_free (self);
1257		1563
1258	PROTOTYPES: ENABLE	1564	PROTOTYPES: ENABLE
1259		1565
1260	void encode_cbor (SV *scalar)	1566	void encode_cbor (SV *scalar)
		1567	ALIAS:
		1568	encode_cbor = 0
		1569	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1570	PPCODE:
1262	{	1571	{
1263	CBOR cbor;	1572	CBOR cbor;
1264	cbor_init (&cbor);	1573	cbor_init (&cbor);
		1574	cbor.flags \|= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1575	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1576	XPUSHs (scalar);
1267	}	1577	}
1268		1578
1269	void decode_cbor (SV *cborstr)	1579	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs. Revision 1.54 by root, Sun Apr 24 20:08:41 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs.
Revision 1.54 by root, Sun Apr 24 20:08:41 2016 UTC