[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 1.52 by root, Sun Apr 24 19:31:55 2016 UTC

…		…
7	#include <stdlib.h>	7	#include <stdlib.h>
8	#include <stdio.h>	8	#include <stdio.h>
9	#include <limits.h>	9	#include <limits.h>
10	#include <float.h>	10	#include <float.h>
11		11
		12	#define ECB_NO_THREADS 1
12	#include "ecb.h"	13	#include "ecb.h"
13		14
14	// compatibility with perl <5.18	15	// compatibility with perl <5.18
15	#ifndef HvNAMELEN_get	16	#ifndef HvNAMELEN_get
16	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))	17	# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
…		…
23	#endif	24	#endif
24	#ifndef SvREFCNT_dec_NN	25	#ifndef SvREFCNT_dec_NN
25	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)	26	# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26	#endif	27	#endif
27		28
		29	// known major and minor types
		30	enum cbor_type
		31	{
		32	MAJOR_SHIFT = 5,
		33	MINOR_MASK = 0x1f,
		34
		35	MAJOR_POS_INT = 0 << MAJOR_SHIFT,
		36	MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
		37	MAJOR_BYTES = 2 << MAJOR_SHIFT,
		38	MAJOR_TEXT = 3 << MAJOR_SHIFT,
		39	MAJOR_ARRAY = 4 << MAJOR_SHIFT,
		40	MAJOR_MAP = 5 << MAJOR_SHIFT,
		41	MAJOR_TAG = 6 << MAJOR_SHIFT,
		42	MAJOR_MISC = 7 << MAJOR_SHIFT,
		43
		44	// INT/STRING/ARRAY/MAP subtypes
		45	LENGTH_EXT1 = 24,
		46	LENGTH_EXT2 = 25,
		47	LENGTH_EXT4 = 26,
		48	LENGTH_EXT8 = 27,
		49
		50	// SIMPLE types (effectively MISC subtypes)
		51	SIMPLE_FALSE = 20,
		52	SIMPLE_TRUE = 21,
		53	SIMPLE_NULL = 22,
		54	SIMPLE_UNDEF = 23,
		55
		56	// MISC subtype (unused)
		57	MISC_EXT1 = 24,
		58	MISC_FLOAT16 = 25,
		59	MISC_FLOAT32 = 26,
		60	MISC_FLOAT64 = 27,
		61
		62	// BYTES/TEXT/ARRAY/MAP
		63	MINOR_INDEF = 31,
		64	};
		65
28	// known tags	66	// known tags
29	enum cbor_tag	67	enum cbor_tag
30	{	68	{
31	// extensions	69	// extensions
32	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref	70	CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
33	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object	71	CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
34	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object	72	CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
35	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing	73	CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
36	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing	74	CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
37	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref	75	CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
38	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection	76	CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
39		77
40	// rfc7049	78	// rfc7049
41	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8	79	CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
42	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any	80	CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
43	CBOR_TAG_POS_BIGNUM = 2, // byte string	81	CBOR_TAG_POS_BIGNUM = 2, // byte string
44	CBOR_TAG_NEG_BIGNUM = 3, // byte string	82	CBOR_TAG_NEG_BIGNUM = 3, // byte string
45	CBOR_TAG_DECIMAL = 4, // decimal fraction, array	83	CBOR_TAG_DECIMAL = 4, // decimal fraction, array
46	CBOR_TAG_BIGFLOAT = 5, // array	84	CBOR_TAG_BIGFLOAT = 5, // array
47		85
48	CBOR_TAG_CONV_B64U = 21, // base64url, any	86	CBOR_TAG_CONV_B64U = 21, // base64url, any
49	CBOR_TAG_CONV_B64 = 22, // base64, any	87	CBOR_TAG_CONV_B64 = 22, // base64, any
50	CBOR_TAG_CONV_HEX = 23, // base16, any	88	CBOR_TAG_CONV_HEX = 23, // base16, any
51	CBOR_TAG_CBOR = 24, // embedded cbor, byte string	89	CBOR_TAG_CBOR = 24, // embedded cbor, byte string
52		90
53	CBOR_TAG_URI = 32, // URI rfc3986, utf-8	91	CBOR_TAG_URI = 32, // URI rfc3986, utf-8
54	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8	92	CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
55	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8	93	CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
56	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8	94	CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
57	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8	95	CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
58		96
59	CBOR_TAG_MAGIC = 55799 // self-describe cbor	97	CBOR_TAG_MAGIC = 55799, // self-describe cbor
60	};	98	};
61		99
62	#define F_SHRINK 0x00000001UL	100	#define F_SHRINK 0x00000001UL
63	#define F_ALLOW_UNKNOWN 0x00000002UL	101	#define F_ALLOW_UNKNOWN 0x00000002UL
64	#define F_ALLOW_SHARING 0x00000004UL	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_UTF8_STRINGS 0x00000020UL
		106	#define F_VALIDATE_UTF8 0x00000040UL
66		107
67	#define INIT_SIZE 32 // initial scalar size to be allocated	108	#define INIT_SIZE 32 // initial scalar size to be allocated
68		109
69	#define SB do {	110	#define SB do {
70	#define SE } while (0)	111	#define SE } while (0)
…		…
89	typedef struct {	130	typedef struct {
90	U32 flags;	131	U32 flags;
91	U32 max_depth;	132	U32 max_depth;
92	STRLEN max_size;	133	STRLEN max_size;
93	SV *filter;	134	SV *filter;
		135
		136	// for the incremental parser
		137	STRLEN incr_pos; // the current offset into the text
		138	STRLEN incr_need; // minimum bytes needed to decode
		139	AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
94	} CBOR;	140	} CBOR;
95		141
96	ecb_inline void	142	ecb_inline void
97	cbor_init (CBOR *cbor)	143	cbor_init (CBOR *cbor)
98	{	144	{
…		…
102		148
103	ecb_inline void	149	ecb_inline void
104	cbor_free (CBOR *cbor)	150	cbor_free (CBOR *cbor)
105	{	151	{
106	SvREFCNT_dec (cbor->filter);	152	SvREFCNT_dec (cbor->filter);
		153	SvREFCNT_dec (cbor->incr_count);
107	}	154	}
108		155
109	/////////////////////////////////////////////////////////////////////////////	156	/////////////////////////////////////////////////////////////////////////////
110	// utility functions	157	// utility functions
111		158
…		…
189	static void	236	static void
190	encode_uint (enc_t *enc, int major, UV len)	237	encode_uint (enc_t *enc, int major, UV len)
191	{	238	{
192	need (enc, 9);	239	need (enc, 9);
193		240
194	if (ecb_expect_true (len < 24))	241	if (ecb_expect_true (len < LENGTH_EXT1))
195	*enc->cur++ = major \| len;	242	*enc->cur++ = major \| len;
196	else if (ecb_expect_true (len <= 0xff))	243	else if (ecb_expect_true (len <= 0xffU))
197	{	244	{
198	*enc->cur++ = major \| 24;	245	*enc->cur++ = major \| LENGTH_EXT1;
199	*enc->cur++ = len;	246	*enc->cur++ = len;
200	}	247	}
201	else if (len <= 0xffff)	248	else if (len <= 0xffffU)
202	{	249	{
203	*enc->cur++ = major \| 25;	250	*enc->cur++ = major \| LENGTH_EXT2;
204	*enc->cur++ = len >> 8;	251	*enc->cur++ = len >> 8;
205	*enc->cur++ = len;	252	*enc->cur++ = len;
206	}	253	}
207	else if (len <= 0xffffffff)	254	else if (len <= 0xffffffffU)
208	{	255	{
209	*enc->cur++ = major \| 26;	256	*enc->cur++ = major \| LENGTH_EXT4;
210	*enc->cur++ = len >> 24;	257	*enc->cur++ = len >> 24;
211	*enc->cur++ = len >> 16;	258	*enc->cur++ = len >> 16;
212	*enc->cur++ = len >> 8;	259	*enc->cur++ = len >> 8;
213	*enc->cur++ = len;	260	*enc->cur++ = len;
214	}	261	}
215	else	262	else
216	{	263	{
217	*enc->cur++ = major \| 27;	264	*enc->cur++ = major \| LENGTH_EXT8;
218	*enc->cur++ = len >> 56;	265	*enc->cur++ = len >> 56;
219	*enc->cur++ = len >> 48;	266	*enc->cur++ = len >> 48;
220	*enc->cur++ = len >> 40;	267	*enc->cur++ = len >> 40;
221	*enc->cur++ = len >> 32;	268	*enc->cur++ = len >> 32;
222	*enc->cur++ = len >> 24;	269	*enc->cur++ = len >> 24;
…		…
227	}	274	}
228		275
229	ecb_inline void	276	ecb_inline void
230	encode_tag (enc_t *enc, UV tag)	277	encode_tag (enc_t *enc, UV tag)
231	{	278	{
232	encode_uint (enc, 0xc0, tag);	279	encode_uint (enc, MAJOR_TAG, tag);
233	}	280	}
234		281
235	ecb_inline void	282	ecb_inline void
236	encode_str (enc_t enc, int utf8, char str, STRLEN len)	283	encode_str (enc_t enc, int utf8, char str, STRLEN len)
237	{	284	{
238	encode_uint (enc, utf8 ? 0x60 : 0x40, len);	285	if (ecb_expect_false (enc->cbor.flags & F_UTF8_STRINGS))
		286	if (!utf8)
		287	{
		288	// exceptional path for bytze strings that need to be utf8-encoded
		289	STRLEN ulen = len;
		290	U8 p, pend = (U8 *)str + len;
		291
		292	for (p = (U8 *)str; p < pend; ++p)
		293	ulen += *p >> 7; // count set high bits
		294
		295	encode_uint (enc, MAJOR_TEXT, ulen);
		296
		297	need (enc, ulen);
		298	for (p = (U8 *)str; p < pend; ++p)
		299	if (*p < 0x80)
		300	enc->cur++ = p;
		301	else
		302	{
		303	enc->cur++ = 0xc0 + (p >> 6);
		304	enc->cur++ = 0x80 + (p & 63);
		305	}
		306
		307	return;
		308	}
		309
		310	encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
239	need (enc, len);	311	need (enc, len);
240	memcpy (enc->cur, str, len);	312	memcpy (enc->cur, str, len);
241	enc->cur += len;	313	enc->cur += len;
242	}	314	}
243		315
…		…
250		322
251	if (SvOK (*svp))	323	if (SvOK (*svp))
252	{	324	{
253	// already registered, use stringref	325	// already registered, use stringref
254	encode_tag (enc, CBOR_TAG_STRINGREF);	326	encode_tag (enc, CBOR_TAG_STRINGREF);
255	encode_uint (enc, 0x00, SvUV (*svp));	327	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
256	return;	328	return;
257	}	329	}
258	else if (len >= minimum_string_length (enc->stringref_idx))	330	else if (len >= minimum_string_length (enc->stringref_idx))
259	{	331	{
260	// register only	332	// register only
…		…
276	if (enc->depth >= enc->cbor.max_depth)	348	if (enc->depth >= enc->cbor.max_depth)
277	croak (ERR_NESTING_EXCEEDED);	349	croak (ERR_NESTING_EXCEEDED);
278		350
279	++enc->depth;	351	++enc->depth;
280		352
281	encode_uint (enc, 0x80, len + 1);	353	encode_uint (enc, MAJOR_ARRAY, len + 1);
282		354
		355	if (SvMAGICAL (av))
283	for (i = 0; i <= len; ++i)	356	for (i = 0; i <= len; ++i)
284	{	357	{
285	SV **svp = av_fetch (av, i, 0);	358	SV **svp = av_fetch (av, i, 0);
286	encode_sv (enc, svp ? *svp : &PL_sv_undef);	359	encode_sv (enc, svp ? *svp : &PL_sv_undef);
287	}	360	}
		361	else
		362	for (i = 0; i <= len; ++i)
		363	{
		364	SV *sv = AvARRAY (av)[i];
		365	encode_sv (enc, sv ? sv : &PL_sv_undef);
		366	}
288		367
289	--enc->depth;	368	--enc->depth;
290	}	369	}
291		370
292	static void	371	static void
…		…
301		380
302	int pairs = hv_iterinit (hv);	381	int pairs = hv_iterinit (hv);
303	int mg = SvMAGICAL (hv);	382	int mg = SvMAGICAL (hv);
304		383
305	if (mg)	384	if (mg)
306	encode_ch (enc, 0xa0 \| 31);	385	encode_ch (enc, MAJOR_MAP \| MINOR_INDEF);
307	else	386	else
308	encode_uint (enc, 0xa0, pairs);	387	encode_uint (enc, MAJOR_MAP, pairs);
309		388
310	while ((he = hv_iternext (hv)))	389	while ((he = hv_iternext (hv)))
311	{	390	{
312	if (HeKLEN (he) == HEf_SVKEY)	391	if (HeKLEN (he) == HEf_SVKEY)
313	encode_sv (enc, HeSVKEY (he));	392	encode_sv (enc, HeSVKEY (he));
…		…
316		395
317	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));	396	encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
318	}	397	}
319		398
320	if (mg)	399	if (mg)
321	encode_ch (enc, 0xe0 \| 31);	400	encode_ch (enc, MAJOR_MISC \| MINOR_INDEF);
322		401
323	--enc->depth;	402	--enc->depth;
324	}	403	}
325		404
326	// encode objects, arrays and special \0=false and \1=true values.	405	// encode objects, arrays and special \0=false and \1=true values.
…		…
345		424
346	HV *stash = SvSTASH (sv);	425	HV *stash = SvSTASH (sv);
347		426
348	if (stash == boolean_stash)	427	if (stash == boolean_stash)
349	{	428	{
350	encode_ch (enc, SvIV (sv) ? 0xe0 \| 21 : 0xe0 \| 20);	429	encode_ch (enc, SvIV (sv) ? MAJOR_MISC \| SIMPLE_TRUE : MAJOR_MISC \| SIMPLE_FALSE);
351	return;	430	return;
352	}	431	}
353	else if (stash == error_stash)	432	else if (stash == error_stash)
354	{	433	{
355	encode_ch (enc, 0xe0 \| 23);	434	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
356	return;	435	return;
357	}	436	}
358	else if (stash == tagged_stash)	437	else if (stash == tagged_stash)
359	{	438	{
360	if (svt != SVt_PVAV)	439	if (svt != SVt_PVAV)
361	croak ("encountered CBOR::XS::Tagged object that isn't an array");	440	croak ("encountered CBOR::XS::Tagged object that isn't an array");
362		441
363	encode_uint (enc, 0xc0, SvUV (av_fetch ((AV )sv, 0, 1)));	442	encode_uint (enc, MAJOR_TAG, SvUV (av_fetch ((AV )sv, 0, 1)));
364	encode_sv (enc, av_fetch ((AV )sv, 1, 1));	443	encode_sv (enc, av_fetch ((AV )sv, 1, 1));
365		444
366	return;	445	return;
367	}	446	}
368	}	447	}
…		…
376	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);	455	SV *svp = hv_fetch (enc->shareable, (char )&sv, sizeof (sv), 1);
377		456
378	if (SvOK (*svp))	457	if (SvOK (*svp))
379	{	458	{
380	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);	459	encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381	encode_uint (enc, 0x00, SvUV (*svp));	460	encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
382	return;	461	return;
383	}	462	}
384	else	463	else
385	{	464	{
386	sv_setuv (*svp, enc->shareable_idx);	465	sv_setuv (*svp, enc->shareable_idx);
…		…
396		475
397	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))	476	if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
398	{	477	{
399	dSP;	478	dSP;
400		479
401	ENTER; SAVETMPS; PUSHMARK (SP);	480	ENTER; SAVETMPS;
		481	PUSHMARK (SP);
402	// we re-bless the reference to get overload and other niceties right	482	// we re-bless the reference to get overload and other niceties right
403	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	483	XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
404		484
405	PUTBACK;	485	PUTBACK;
406	// G_SCALAR ensures that return value is 1	486	// G_SCALAR ensures that return value is 1
…		…
419	}	499	}
420	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)	500	else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
421	{	501	{
422	dSP;	502	dSP;
423		503
424	ENTER; SAVETMPS; PUSHMARK (SP);	504	ENTER; SAVETMPS;
		505	SAVESTACK_POS ();
		506	PUSHMARK (SP);
425	EXTEND (SP, 2);	507	EXTEND (SP, 2);
426	// we re-bless the reference to get overload and other niceties right	508	// we re-bless the reference to get overload and other niceties right
427	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));	509	PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
428	PUSHs (sv_cbor);	510	PUSHs (sv_cbor);
429		511
…		…
434	// catch this surprisingly common error	516	// catch this surprisingly common error
435	if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)	517	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));	518	croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
437		519
438	encode_tag (enc, CBOR_TAG_PERL_OBJECT);	520	encode_tag (enc, CBOR_TAG_PERL_OBJECT);
439	encode_uint (enc, 0x80, count + 1);	521	encode_uint (enc, MAJOR_ARRAY, count + 1);
440	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));	522	encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
441		523
442	while (count)	524	while (count)
443	encode_sv (enc, SP[1 - count--]);	525	encode_sv (enc, SP[1 - count--]);
444		526
…		…
466	{	548	{
467	double nv = SvNVX (sv);	549	double nv = SvNVX (sv);
468		550
469	need (enc, 9);	551	need (enc, 9);
470		552
471	if (ecb_expect_false (nv == (U32)nv))	553	if (ecb_expect_false (nv == (NV)(U32)nv))
472	encode_uint (enc, 0x00, (U32)nv);	554	encode_uint (enc, MAJOR_POS_INT, (U32)nv);
473	//TODO: maybe I32?	555	//TODO: maybe I32?
474	else if (ecb_expect_false (nv == (float)nv))	556	else if (ecb_expect_false (nv == (float)nv))
475	{	557	{
476	uint32_t fp = ecb_float_to_binary32 (nv);	558	uint32_t fp = ecb_float_to_binary32 (nv);
477		559
478	*enc->cur++ = 0xe0 \| 26;	560	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT32;
479		561
480	if (!ecb_big_endian ())	562	if (!ecb_big_endian ())
481	fp = ecb_bswap32 (fp);	563	fp = ecb_bswap32 (fp);
482		564
483	memcpy (enc->cur, &fp, 4);	565	memcpy (enc->cur, &fp, 4);
…		…
485	}	567	}
486	else	568	else
487	{	569	{
488	uint64_t fp = ecb_double_to_binary64 (nv);	570	uint64_t fp = ecb_double_to_binary64 (nv);
489		571
490	*enc->cur++ = 0xe0 \| 27;	572	*enc->cur++ = MAJOR_MISC \| MISC_FLOAT64;
491		573
492	if (!ecb_big_endian ())	574	if (!ecb_big_endian ())
493	fp = ecb_bswap64 (fp);	575	fp = ecb_bswap64 (fp);
494		576
495	memcpy (enc->cur, &fp, 8);	577	memcpy (enc->cur, &fp, 8);
…		…
511	else if (SvNOKp (sv))	593	else if (SvNOKp (sv))
512	encode_nv (enc, sv);	594	encode_nv (enc, sv);
513	else if (SvIOKp (sv))	595	else if (SvIOKp (sv))
514	{	596	{
515	if (SvIsUV (sv))	597	if (SvIsUV (sv))
516	encode_uint (enc, 0x00, SvUVX (sv));	598	encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
517	else if (SvIVX (sv) >= 0)	599	else if (SvIVX (sv) >= 0)
518	encode_uint (enc, 0x00, SvIVX (sv));	600	encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
519	else	601	else
520	encode_uint (enc, 0x20, -(SvIVX (sv) + 1));	602	encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
521	}	603	}
522	else if (SvROK (sv))	604	else if (SvROK (sv))
523	encode_rv (enc, SvRV (sv));	605	encode_rv (enc, SvRV (sv));
524	else if (!SvOK (sv))	606	else if (!SvOK (sv))
525	encode_ch (enc, 0xe0 \| 22);	607	encode_ch (enc, MAJOR_MISC \| SIMPLE_NULL);
526	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)	608	else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
527	encode_ch (enc, 0xe0 \| 23);	609	encode_ch (enc, MAJOR_MISC \| SIMPLE_UNDEF);
528	else	610	else
529	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",	611	croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
530	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));	612	SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
531	}	613	}
532		614
533	static SV *	615	static SV *
534	encode_cbor (SV scalar, CBOR cbor)	616	encode_cbor (SV scalar, CBOR cbor)
535	{	617	{
536	enc_t enc = { };	618	enc_t enc = { 0 };
537		619
538	enc.cbor = *cbor;	620	enc.cbor = *cbor;
539	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));	621	enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
540	enc.cur = SvPVX (enc.sv);	622	enc.cur = SvPVX (enc.sv);
541	enc.end = SvEND (enc.sv);	623	enc.end = SvEND (enc.sv);
…		…
585	#define DEC_DEC_DEPTH --dec->depth	667	#define DEC_DEC_DEPTH --dec->depth
586		668
587	static UV	669	static UV
588	decode_uint (dec_t *dec)	670	decode_uint (dec_t *dec)
589	{	671	{
590	switch (*dec->cur & 31)	672	U8 m = *dec->cur & MINOR_MASK;
591	{	673	++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		674
597	case 24:	675	if (ecb_expect_true (m < LENGTH_EXT1))
		676	return m;
		677	else if (ecb_expect_true (m == LENGTH_EXT1))
		678	{
598	WANT (2);	679	WANT (1);
599	dec->cur += 2;	680	dec->cur += 1;
600	return dec->cur[-1];	681	return dec->cur[-1];
601		682	}
602	case 25:	683	else if (ecb_expect_true (m == LENGTH_EXT2))
		684	{
603	WANT (3);	685	WANT (2);
604	dec->cur += 3;	686	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	687	return (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	688	\| ((UV)dec->cur[-1]);
607		689	}
608	case 26:	690	else if (ecb_expect_true (m == LENGTH_EXT4))
		691	{
609	WANT (5);	692	WANT (4);
610	dec->cur += 5;	693	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	694	return (((UV)dec->cur[-4]) << 24)
612	\| (((UV)dec->cur[-3]) << 16)	695	\| (((UV)dec->cur[-3]) << 16)
613	\| (((UV)dec->cur[-2]) << 8)	696	\| (((UV)dec->cur[-2]) << 8)
614	\| ((UV)dec->cur[-1]);	697	\| ((UV)dec->cur[-1]);
615		698	}
616	case 27:	699	else if (ecb_expect_true (m == LENGTH_EXT8))
		700	{
617	WANT (9);	701	WANT (8);
618	dec->cur += 9;	702	dec->cur += 8;
619		703
620	return	704	return
621	#if UVSIZE < 8	705	#if UVSIZE < 8
622	0	706	0
623	#else	707	#else
624	(((UV)dec->cur[-8]) << 56)	708	(((UV)dec->cur[-8]) << 56)
625	\| (((UV)dec->cur[-7]) << 48)	709	\| (((UV)dec->cur[-7]) << 48)
626	\| (((UV)dec->cur[-6]) << 40)	710	\| (((UV)dec->cur[-6]) << 40)
627	\| (((UV)dec->cur[-5]) << 32)	711	\| (((UV)dec->cur[-5]) << 32)
628	#endif	712	#endif
629	\| (((UV)dec->cur[-4]) << 24)	713	\| (((UV)dec->cur[-4]) << 24)
630	\| (((UV)dec->cur[-3]) << 16)	714	\| (((UV)dec->cur[-3]) << 16)
631	\| (((UV)dec->cur[-2]) << 8)	715	\| (((UV)dec->cur[-2]) << 8)
632	\| ((UV)dec->cur[-1]);	716	\| ((UV)dec->cur[-1]);
633		717	}
634	default:	718	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	719	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		720
638	fail:	721	fail:
639	return 0;	722	return 0;
640	}	723	}
641		724
…		…
646	{	729	{
647	AV *av = newAV ();	730	AV *av = newAV ();
648		731
649	DEC_INC_DEPTH;	732	DEC_INC_DEPTH;
650		733
651	if ((*dec->cur & 31) == 31)	734	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
652	{	735	{
653	++dec->cur;	736	++dec->cur;
654		737
655	for (;;)	738	for (;;)
656	{	739	{
657	WANT (1);	740	WANT (1);
658		741
659	if (*dec->cur == (0xe0 \| 31))	742	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
660	{	743	{
661	++dec->cur;	744	++dec->cur;
662	break;	745	break;
663	}	746	}
664		747
…		…
667	}	750	}
668	else	751	else
669	{	752	{
670	int i, len = decode_uint (dec);	753	int i, len = decode_uint (dec);
671		754
		755	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);	756	av_fill (av, len - 1);
673		757
674	for (i = 0; i < len; ++i)	758	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	759	AvARRAY (av)[i] = decode_sv (dec);
676	}	760	}
…		…
689	{	773	{
690	// for speed reasons, we specialcase single-string	774	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	775	// byte or utf-8 strings as keys, but only when !stringref
692		776
693	if (ecb_expect_true (!dec->stringref))	777	if (ecb_expect_true (!dec->stringref))
694	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	778	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	779	{
696	I32 len = decode_uint (dec);	780	I32 len = decode_uint (dec);
697	char key = (char )dec->cur;	781	char key = (char )dec->cur;
698		782
		783	WANT (len);
699	dec->cur += len;	784	dec->cur += len;
700		785
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);	786	hv_store (hv, key, len, decode_sv (dec), 0);
705		787
706	return;	788	return;
707	}	789	}
708	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	790	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	791	{
710	I32 len = decode_uint (dec);	792	I32 len = decode_uint (dec);
711	char key = (char )dec->cur;	793	char key = (char )dec->cur;
712		794
		795	WANT (len);
713	dec->cur += len;	796	dec->cur += len;
714		797
715	if (ecb_expect_false (dec->stringref))	798	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	799	if (!is_utf8_string (key, len))
		800	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
717		801
718	hv_store (hv, key, -len, decode_sv (dec), 0);	802	hv_store (hv, key, -len, decode_sv (dec), 0);
719		803
720	return;	804	return;
721	}	805	}
…		…
723	SV *k = decode_sv (dec);	807	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	808	SV *v = decode_sv (dec);
725		809
726	hv_store_ent (hv, k, v, 0);	810	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	811	SvREFCNT_dec (k);
		812
		813	fail:
		814	;
728	}	815	}
729		816
730	static SV *	817	static SV *
731	decode_hv (dec_t *dec)	818	decode_hv (dec_t *dec)
732	{	819	{
733	HV *hv = newHV ();	820	HV *hv = newHV ();
734		821
735	DEC_INC_DEPTH;	822	DEC_INC_DEPTH;
736		823
737	if ((*dec->cur & 31) == 31)	824	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
738	{	825	{
739	++dec->cur;	826	++dec->cur;
740		827
741	for (;;)	828	for (;;)
742	{	829	{
743	WANT (1);	830	WANT (1);
744		831
745	if (*dec->cur == (0xe0 \| 31))	832	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
746	{	833	{
747	++dec->cur;	834	++dec->cur;
748	break;	835	break;
749	}	836	}
750		837
…		…
771	static SV *	858	static SV *
772	decode_str (dec_t *dec, int utf8)	859	decode_str (dec_t *dec, int utf8)
773	{	860	{
774	SV *sv = 0;	861	SV *sv = 0;
775		862
776	if ((*dec->cur & 31) == 31)	863	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
777	{	864	{
778	// indefinite length strings	865	// indefinite length strings
779	++dec->cur;	866	++dec->cur;
780		867
781	unsigned char major = *dec->cur & 0xe0;	868	U8 major = *dec->cur & MAJOR_MISC;
782		869
783	sv = newSVpvn ("", 0);	870	sv = newSVpvn ("", 0);
784		871
785	for (;;)	872	for (;;)
786	{	873	{
787	WANT (1);	874	WANT (1);
788		875
789	if ((*dec->cur ^ major) >= 31)	876	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 \| 31))	877	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
791	{	878	{
792	++dec->cur;	879	++dec->cur;
793	break;	880	break;
794	}	881	}
795	else	882	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	901	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	902	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	903	}
817		904
818	if (utf8)	905	if (utf8)
		906	{
		907	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		908	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		909	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		910
819	SvUTF8_on (sv);	911	SvUTF8_on (sv);
		912	}
820		913
821	return sv;	914	return sv;
822		915
823	fail:	916	fail:
824	SvREFCNT_dec (sv);	917	SvREFCNT_dec (sv);
…		…
856	}	949	}
857	break;	950	break;
858		951
859	case CBOR_TAG_STRINGREF:	952	case CBOR_TAG_STRINGREF:
860	{	953	{
861	if ((*dec->cur >> 5) != 0)	954	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	955	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		956
864	UV idx = decode_uint (dec);	957	UV idx = decode_uint (dec);
865		958
866	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	959	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	966	case CBOR_TAG_VALUE_SHAREABLE:
874	{	967	{
875	if (ecb_expect_false (!dec->shareable))	968	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	969	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
877		970
		971	if (dec->cbor.flags & F_ALLOW_CYCLES)
		972	{
878	sv = newSV (0);	973	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	974	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		975
881	SV *osv = decode_sv (dec);	976	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	977	sv_setsv (sv, osv);
883	SvREFCNT_dec_NN (osv);	978	SvREFCNT_dec_NN (osv);
		979	}
		980	else
		981	{
		982	av_push (dec->shareable, &PL_sv_undef);
		983	int idx = AvFILLp (dec->shareable);
		984	sv = decode_sv (dec);
		985	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		986	}
884	}	987	}
885	break;	988	break;
886		989
887	case CBOR_TAG_VALUE_SHAREDREF:	990	case CBOR_TAG_VALUE_SHAREDREF:
888	{	991	{
889	if ((*dec->cur >> 5) != 0)	992	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	993	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		994
892	UV idx = decode_uint (dec);	995	UV idx = decode_uint (dec);
893		996
894	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	997	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	998	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		999
897	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1000	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1001
		1002	if (sv == &PL_sv_undef)
		1003	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
898	}	1004	}
899	break;	1005	break;
900		1006
901	case CBOR_TAG_PERL_OBJECT:	1007	case CBOR_TAG_PERL_OBJECT:
902	{	1008	{
…		…
917	if (!method)	1023	if (!method)
918	ERR ("cannot decode perl-object (package does not have a THAW method)");	1024	ERR ("cannot decode perl-object (package does not have a THAW method)");
919		1025
920	dSP;	1026	dSP;
921		1027
922	ENTER; SAVETMPS; PUSHMARK (SP);	1028	ENTER; SAVETMPS;
		1029	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1030	EXTEND (SP, len + 1);
924	// we re-bless the reference to get overload and other niceties right	1031	// we re-bless the reference to get overload and other niceties right
925	PUSHs (*av_fetch (av, 0, 1));	1032	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1033	PUSHs (sv_cbor);
927		1034
…		…
952	default:	1059	default:
953	{	1060	{
954	sv = decode_sv (dec);	1061	sv = decode_sv (dec);
955		1062
956	dSP;	1063	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1064	ENTER; SAVETMPS;
		1065	SAVESTACK_POS ();
		1066	PUSHMARK (SP);
958	EXTEND (SP, 2);	1067	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1068	PUSHs (newSVuv (tag));
960	PUSHs (sv);	1069	PUSHs (sv);
961		1070
962	PUTBACK;	1071	PUTBACK;
…		…
1003	static SV *	1112	static SV *
1004	decode_sv (dec_t *dec)	1113	decode_sv (dec_t *dec)
1005	{	1114	{
1006	WANT (1);	1115	WANT (1);
1007		1116
1008	switch (*dec->cur >> 5)	1117	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1118	{
1010	case 0: // unsigned int	1119	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1120	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1121	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1122	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1123	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1124	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1125	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1126
1018	case 4: // array	1127	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)	1128	switch (*dec->cur++ & MINOR_MASK)
1026	{	1129	{
1027	case 20:	1130	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1131	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1132	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1133	#endif
1031	return newSVsv (types_false);	1134	return newSVsv (types_false);
1032	case 21:	1135	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1136	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1137	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1138	#endif
1036	return newSVsv (types_true);	1139	return newSVsv (types_true);
1037	case 22:	1140	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1141	return newSVsv (&PL_sv_undef);
1039	case 23:	1142	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1143	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1144	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1145	#endif
1043	return newSVsv (types_error);	1146	return newSVsv (types_error);
1044		1147
1045	case 25:	1148	case MISC_FLOAT16:
1046	{	1149	{
1047	WANT (2);	1150	WANT (2);
1048		1151
1049	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1152	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1050	dec->cur += 2;	1153	dec->cur += 2;
1051		1154
1052	return newSVnv (ecb_binary16_to_float (fp));	1155	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1156	}
1054		1157
1055	case 26:	1158	case MISC_FLOAT32:
1056	{	1159	{
1057	uint32_t fp;	1160	uint32_t fp;
1058	WANT (4);	1161	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1162	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1163	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1166	fp = ecb_bswap32 (fp);
1064		1167
1065	return newSVnv (ecb_binary32_to_float (fp));	1168	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1169	}
1067		1170
1068	case 27:	1171	case MISC_FLOAT64:
1069	{	1172	{
1070	uint64_t fp;	1173	uint64_t fp;
1071	WANT (8);	1174	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1175	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1176	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1179	fp = ecb_bswap64 (fp);
1077		1180
1078	return newSVnv (ecb_binary64_to_double (fp));	1181	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1182	}
1080		1183
1081	// 0..19 unassigned	1184	// 0..19 unassigned simple
1082	// 24 reserved + unassigned (reserved values are not encodable)	1185	// 24 reserved + unassigned simple (reserved values are not encodable)
		1186	// 28-30 unassigned misc
		1187	// 31 break code
1083	default:	1188	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1189	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1190	}
1086		1191
1087	break;	1192	break;
1088	}	1193	}
1089		1194
…		…
1092	}	1197	}
1093		1198
1094	static SV *	1199	static SV *
1095	decode_cbor (SV string, CBOR cbor, char **offset_return)	1200	decode_cbor (SV string, CBOR cbor, char **offset_return)
1096	{	1201	{
1097	dec_t dec = { };	1202	dec_t dec = { 0 };
1098	SV *sv;	1203	SV *sv;
1099	STRLEN len;	1204	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1205	char *data = SvPVbyte (string, len);
1101		1206
1102	if (len > cbor->max_size && cbor->max_size)	1207	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1221	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1222	dec.err = "garbage after CBOR object";
1118		1223
1119	if (dec.err)	1224	if (dec.err)
1120	{	1225	{
		1226	if (dec.shareable)
		1227	{
		1228	// need to break cyclic links, which whould all be in shareable
		1229	int i;
		1230	SV **svp;
		1231
		1232	for (i = av_len (dec.shareable) + 1; i--; )
		1233	if ((svp = av_fetch (dec.shareable, i, 0)))
		1234	sv_setsv (*svp, &PL_sv_undef);
		1235	}
		1236
1121	SvREFCNT_dec (sv);	1237	SvREFCNT_dec (sv);
1122	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1238	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1123	}	1239	}
1124		1240
1125	sv = sv_2mortal (sv);	1241	sv = sv_2mortal (sv);
1126		1242
1127	return sv;	1243	return sv;
1128	}	1244	}
1129		1245
		1246	/////////////////////////////////////////////////////////////////////////////
		1247	// incremental parser
		1248
		1249	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1250
		1251	// returns 0 for notyet, 1 for success or error
		1252	static int
		1253	incr_parse (CBOR self, SV cborstr)
		1254	{
		1255	STRLEN cur;
		1256	SvPV (cborstr, cur);
		1257
		1258	while (ecb_expect_true (self->incr_need <= cur))
		1259	{
		1260	// table of integer count bytes
		1261	static I8 incr_len[MINOR_MASK + 1] = {
		1262	0, 0, 0, 0, 0, 0, 0, 0,
		1263	0, 0, 0, 0, 0, 0, 0, 0,
		1264	0, 0, 0, 0, 0, 0, 0, 0,
		1265	1, 2, 4, 8,-1,-1,-1,-2
		1266	};
		1267
		1268	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1269	U8 m = *p & MINOR_MASK;
		1270	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1271	I8 ilen = incr_len[m];
		1272
		1273	self->incr_need = self->incr_pos + 1;
		1274
		1275	if (ecb_expect_false (ilen < 0))
		1276	{
		1277	if (m != MINOR_INDEF)
		1278	return 1; // error
		1279
		1280	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1281	{
		1282	if (count >= 0)
		1283	return 1; // error
		1284
		1285	count = 1;
		1286	}
		1287	else
		1288	{
		1289	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1290	count = -1;
		1291	}
		1292	}
		1293	else
		1294	{
		1295	self->incr_need += ilen;
		1296	if (ecb_expect_false (self->incr_need > cur))
		1297	return 0;
		1298
		1299	int major = *p >> MAJOR_SHIFT;
		1300
		1301	switch (major)
		1302	{
		1303	case MAJOR_TAG >> MAJOR_SHIFT:
		1304	++count; // tags merely prefix another value
		1305	break;
		1306
		1307	case MAJOR_BYTES >> MAJOR_SHIFT:
		1308	case MAJOR_TEXT >> MAJOR_SHIFT:
		1309	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1310	case MAJOR_MAP >> MAJOR_SHIFT:
		1311	{
		1312	UV len;
		1313
		1314	if (ecb_expect_false (ilen))
		1315	{
		1316	len = 0;
		1317
		1318	do {
		1319	len = (len << 8) \| *++p;
		1320	} while (--ilen);
		1321	}
		1322	else
		1323	len = m;
		1324
		1325	switch (major)
		1326	{
		1327	case MAJOR_BYTES >> MAJOR_SHIFT:
		1328	case MAJOR_TEXT >> MAJOR_SHIFT:
		1329	self->incr_need += len;
		1330	if (ecb_expect_false (self->incr_need > cur))
		1331	return 0;
		1332
		1333	break;
		1334
		1335	case MAJOR_MAP >> MAJOR_SHIFT:
		1336	len <<= 1;
		1337	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1338	if (len)
		1339	{
		1340	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1341	count = len + 1;
		1342	}
		1343	break;
		1344	}
		1345	}
		1346	}
		1347	}
		1348
		1349	self->incr_pos = self->incr_need;
		1350
		1351	if (count > 0)
		1352	{
		1353	while (!--count)
		1354	{
		1355	if (!AvFILLp (self->incr_count))
		1356	return 1; // done
		1357
		1358	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1359	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1360	}
		1361
		1362	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1363	}
		1364	}
		1365
		1366	return 0;
		1367	}
		1368
		1369
1130	/////////////////////////////////////////////////////////////////////////////	1370	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1371	// XS interface functions
1132		1372
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1373	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1374
…		…
1174	void shrink (CBOR *self, int enable = 1)	1414	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1415	ALIAS:
1176	shrink = F_SHRINK	1416	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1417	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1418	allow_sharing = F_ALLOW_SHARING
		1419	allow_cycles = F_ALLOW_CYCLES
1179	pack_strings = F_PACK_STRINGS	1420	pack_strings = F_PACK_STRINGS
		1421	utf8_strings = F_UTF8_STRINGS
		1422	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1423	PPCODE:
1181	{	1424	{
1182	if (enable)	1425	if (enable)
1183	self->flags \|= ix;	1426	self->flags \|= ix;
1184	else	1427	else
…		…
1190	void get_shrink (CBOR *self)	1433	void get_shrink (CBOR *self)
1191	ALIAS:	1434	ALIAS:
1192	get_shrink = F_SHRINK	1435	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1436	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1437	get_allow_sharing = F_ALLOW_SHARING
		1438	get_allow_cycles = F_ALLOW_CYCLES
1195	get_pack_strings = F_PACK_STRINGS	1439	get_pack_strings = F_PACK_STRINGS
		1440	get_validate_utf8 = F_VALIDATE_UTF8
1196	PPCODE:	1441	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1442	XPUSHs (boolSV (self->flags & ix));
1198		1443
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1444	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1445	PPCODE:
…		…
1249	EXTEND (SP, 2);	1494	EXTEND (SP, 2);
1250	PUSHs (sv);	1495	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1496	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1497	}
1253		1498
		1499	void incr_parse (CBOR self, SV cborstr)
		1500	ALIAS:
		1501	incr_parse_multiple = 1
		1502	PPCODE:
		1503	{
		1504	if (SvUTF8 (cborstr))
		1505	sv_utf8_downgrade (cborstr, 0);
		1506
		1507	if (!self->incr_count)
		1508	{
		1509	self->incr_count = newAV ();
		1510	self->incr_pos = 0;
		1511	self->incr_need = 1;
		1512
		1513	av_push (self->incr_count, newSViv (1));
		1514	}
		1515
		1516	do
		1517	{
		1518	if (!incr_parse (self, cborstr))
		1519	{
		1520	if (self->incr_need > self->max_size && self->max_size)
		1521	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1522	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1523
		1524	break;
		1525	}
		1526
		1527	SV *sv;
		1528	char *offset;
		1529
		1530	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1531	XPUSHs (sv);
		1532
		1533	sv_chop (cborstr, offset);
		1534
		1535	av_clear (self->incr_count);
		1536	av_push (self->incr_count, newSViv (1));
		1537
		1538	self->incr_pos = 0;
		1539	self->incr_need = self->incr_pos + 1;
		1540	}
		1541	while (ix);
		1542	}
		1543
		1544	void incr_reset (CBOR *self)
		1545	CODE:
		1546	{
		1547	SvREFCNT_dec (self->incr_count);
		1548	self->incr_count = 0;
		1549	}
		1550
1254	void DESTROY (CBOR *self)	1551	void DESTROY (CBOR *self)
1255	PPCODE:	1552	PPCODE:
1256	cbor_free (self);	1553	cbor_free (self);
1257		1554
1258	PROTOTYPES: ENABLE	1555	PROTOTYPES: ENABLE
1259		1556
1260	void encode_cbor (SV *scalar)	1557	void encode_cbor (SV *scalar)
		1558	ALIAS:
		1559	encode_cbor = 0
		1560	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1561	PPCODE:
1262	{	1562	{
1263	CBOR cbor;	1563	CBOR cbor;
1264	cbor_init (&cbor);	1564	cbor_init (&cbor);
		1565	cbor.flags \|= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1566	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1567	XPUSHs (scalar);
1267	}	1568	}
1268		1569
1269	void decode_cbor (SV *cborstr)	1570	void decode_cbor (SV *cborstr)

Diff Legend

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

Comparing CBOR-XS/XS.xs (file contents): Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs. Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC

Diff Legend

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC