[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 1.59 by root, Fri Nov 25 23:37:27 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);
…		…
573	U32 depth; // recursion depth	661	U32 depth; // recursion depth
574	U32 maxdepth; // recursion depth limit	662	U32 maxdepth; // recursion depth limit
575	AV *shareable;	663	AV *shareable;
576	AV *stringref;	664	AV *stringref;
577	SV *decode_tagged;	665	SV *decode_tagged;
		666	SV *err_sv; // optional sv for error, needs to be freed
578	} dec_t;	667	} dec_t;
579		668
580	#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE	669	// set dec->err to ERRSV
		670	ecb_cold static void
		671	err_errsv (dec_t *dec)
		672	{
		673	if (!dec->err)
		674	{
		675	dec->err_sv = newSVsv (ERRSV);
581		676
582	#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")	677	// chop off the trailing \n
		678	SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
		679	*SvEND (dec->err_sv) = 0;
583		680
		681	dec->err = SvPVutf8_nolen (dec->err_sv);
		682	}
		683	}
		684
		685	// the following functions are used to reduce code size and help the compiler to optimise
		686	ecb_cold static void
		687	err_set (dec_t dec, const char reason)
		688	{
		689	if (!dec->err)
		690	dec->err = reason;
		691	}
		692
		693	ecb_cold static void
		694	err_unexpected_end (dec_t *dec)
		695	{
		696	err_set (dec, "unexpected end of CBOR data");
		697	}
		698
		699	ecb_cold static void
		700	err_nesting_exceeded (dec_t *dec)
		701	{
		702	err_set (dec, ERR_NESTING_EXCEEDED);
		703	}
		704
		705	#define ERR_DO(do) SB do; goto fail; SE
		706	#define ERR(reason) ERR_DO (err_set (dec, reason))
		707	#define ERR_ERRSV ERR_DO (err_errsv (dec))
		708
		709	#define WANT(len) if (ecb_expect_false ((UV)(dec->end - dec->cur) < (UV)len)) ERR_DO (err_unexpected_end (dec))
		710
584	#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)	711	#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
585	#define DEC_DEC_DEPTH --dec->depth	712	#define DEC_DEC_DEPTH --dec->depth
586		713
587	static UV	714	static UV
588	decode_uint (dec_t *dec)	715	decode_uint (dec_t *dec)
589	{	716	{
590	switch (*dec->cur & 31)	717	U8 m = *dec->cur & MINOR_MASK;
591	{	718	++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		719
597	case 24:	720	if (ecb_expect_true (m < LENGTH_EXT1))
		721	return m;
		722	else if (ecb_expect_true (m == LENGTH_EXT1))
		723	{
598	WANT (2);	724	WANT (1);
599	dec->cur += 2;	725	dec->cur += 1;
600	return dec->cur[-1];	726	return dec->cur[-1];
601		727	}
602	case 25:	728	else if (ecb_expect_true (m == LENGTH_EXT2))
		729	{
603	WANT (3);	730	WANT (2);
604	dec->cur += 3;	731	dec->cur += 2;
605	return (((UV)dec->cur[-2]) << 8)	732	return (((UV)dec->cur[-2]) << 8)
606	\| ((UV)dec->cur[-1]);	733	\| ((UV)dec->cur[-1]);
607		734	}
608	case 26:	735	else if (ecb_expect_true (m == LENGTH_EXT4))
		736	{
609	WANT (5);	737	WANT (4);
610	dec->cur += 5;	738	dec->cur += 4;
611	return (((UV)dec->cur[-4]) << 24)	739	return (((UV)dec->cur[-4]) << 24)
612	\| (((UV)dec->cur[-3]) << 16)	740	\| (((UV)dec->cur[-3]) << 16)
613	\| (((UV)dec->cur[-2]) << 8)	741	\| (((UV)dec->cur[-2]) << 8)
614	\| ((UV)dec->cur[-1]);	742	\| ((UV)dec->cur[-1]);
615		743	}
616	case 27:	744	else if (ecb_expect_true (m == LENGTH_EXT8))
		745	{
617	WANT (9);	746	WANT (8);
618	dec->cur += 9;	747	dec->cur += 8;
619		748
620	return	749	return
621	#if UVSIZE < 8	750	#if UVSIZE < 8
622	0	751	0
623	#else	752	#else
624	(((UV)dec->cur[-8]) << 56)	753	(((UV)dec->cur[-8]) << 56)
625	\| (((UV)dec->cur[-7]) << 48)	754	\| (((UV)dec->cur[-7]) << 48)
626	\| (((UV)dec->cur[-6]) << 40)	755	\| (((UV)dec->cur[-6]) << 40)
627	\| (((UV)dec->cur[-5]) << 32)	756	\| (((UV)dec->cur[-5]) << 32)
628	#endif	757	#endif
629	\| (((UV)dec->cur[-4]) << 24)	758	\| (((UV)dec->cur[-4]) << 24)
630	\| (((UV)dec->cur[-3]) << 16)	759	\| (((UV)dec->cur[-3]) << 16)
631	\| (((UV)dec->cur[-2]) << 8)	760	\| (((UV)dec->cur[-2]) << 8)
632	\| ((UV)dec->cur[-1]);	761	\| ((UV)dec->cur[-1]);
633		762	}
634	default:	763	else
635	ERR ("corrupted CBOR data (unsupported integer minor encoding)");	764	ERR ("corrupted CBOR data (unsupported integer minor encoding)");
636	}
637		765
638	fail:	766	fail:
639	return 0;	767	return 0;
640	}	768	}
641		769
…		…
646	{	774	{
647	AV *av = newAV ();	775	AV *av = newAV ();
648		776
649	DEC_INC_DEPTH;	777	DEC_INC_DEPTH;
650		778
651	if ((*dec->cur & 31) == 31)	779	if (*dec->cur == (MAJOR_ARRAY \| MINOR_INDEF))
652	{	780	{
653	++dec->cur;	781	++dec->cur;
654		782
655	for (;;)	783	for (;;)
656	{	784	{
657	WANT (1);	785	WANT (1);
658		786
659	if (*dec->cur == (0xe0 \| 31))	787	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
660	{	788	{
661	++dec->cur;	789	++dec->cur;
662	break;	790	break;
663	}	791	}
664		792
665	av_push (av, decode_sv (dec));	793	av_push (av, decode_sv (dec));
666	}	794	}
667	}	795	}
668	else	796	else
669	{	797	{
670	int i, len = decode_uint (dec);	798	UV i, len = decode_uint (dec);
671		799
		800	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);	801	av_fill (av, len - 1);
673		802
674	for (i = 0; i < len; ++i)	803	for (i = 0; i < len; ++i)
675	AvARRAY (av)[i] = decode_sv (dec);	804	AvARRAY (av)[i] = decode_sv (dec);
676	}	805	}
…		…
689	{	818	{
690	// for speed reasons, we specialcase single-string	819	// for speed reasons, we specialcase single-string
691	// byte or utf-8 strings as keys, but only when !stringref	820	// byte or utf-8 strings as keys, but only when !stringref
692		821
693	if (ecb_expect_true (!dec->stringref))	822	if (ecb_expect_true (!dec->stringref))
694	if (dec->cur >= 0x40 && dec->cur <= 0x40 + 27)	823	if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
695	{	824	{
696	I32 len = decode_uint (dec);	825	STRLEN len = decode_uint (dec);
697	char key = (char )dec->cur;	826	char key = (char )dec->cur;
698		827
		828	WANT (len);
699	dec->cur += len;	829	dec->cur += len;
700		830
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);	831	hv_store (hv, key, len, decode_sv (dec), 0);
705		832
706	return;	833	return;
707	}	834	}
708	else if (dec->cur >= 0x60 && dec->cur <= 0x60 + 27)	835	else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
709	{	836	{
710	I32 len = decode_uint (dec);	837	STRLEN len = decode_uint (dec);
711	char key = (char )dec->cur;	838	char key = (char )dec->cur;
712		839
		840	WANT (len);
713	dec->cur += len;	841	dec->cur += len;
714		842
715	if (ecb_expect_false (dec->stringref))	843	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
716	av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));	844	if (!is_utf8_string (key, len))
		845	ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
717		846
718	hv_store (hv, key, -len, decode_sv (dec), 0);	847	hv_store (hv, key, -len, decode_sv (dec), 0);
719		848
720	return;	849	return;
721	}	850	}
722		851
723	SV *k = decode_sv (dec);	852	SV *k = decode_sv (dec);
724	SV *v = decode_sv (dec);	853	SV *v = decode_sv (dec);
725		854
		855	// we leak memory if uncaught exceptions are thrown by random magical
		856	// methods, and this is hopefully the only place where it can happen,
		857	// so if there is a chance of an exception, take the very slow path.
		858	// since catching exceptions is "undocumented/internal/forbidden" by
		859	// the new p5p powers, we need to call out to a perl function :/
		860	if (ecb_expect_false (SvAMAGIC (k)))
		861	{
		862	dSP;
		863
		864	ENTER; SAVETMPS;
		865	PUSHMARK (SP);
		866	EXTEND (SP, 3);
		867	PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
		868	PUSHs (sv_2mortal (k));
		869	PUSHs (sv_2mortal (v));
		870
		871	PUTBACK;
		872	call_pv ("CBOR::XS::_hv_store", G_VOID \| G_DISCARD \| G_EVAL);
		873	SPAGAIN;
		874
		875	FREETMPS; LEAVE;
		876
		877	if (SvTRUE (ERRSV))
		878	ERR_ERRSV;
		879
		880	return;
		881	}
		882
726	hv_store_ent (hv, k, v, 0);	883	hv_store_ent (hv, k, v, 0);
727	SvREFCNT_dec (k);	884	SvREFCNT_dec (k);
		885
		886	fail:
		887	;
728	}	888	}
729		889
730	static SV *	890	static SV *
731	decode_hv (dec_t *dec)	891	decode_hv (dec_t *dec)
732	{	892	{
733	HV *hv = newHV ();	893	HV *hv = newHV ();
734		894
735	DEC_INC_DEPTH;	895	DEC_INC_DEPTH;
736		896
737	if ((*dec->cur & 31) == 31)	897	if (*dec->cur == (MAJOR_MAP \| MINOR_INDEF))
738	{	898	{
739	++dec->cur;	899	++dec->cur;
740		900
741	for (;;)	901	for (;;)
742	{	902	{
743	WANT (1);	903	WANT (1);
744		904
745	if (*dec->cur == (0xe0 \| 31))	905	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
746	{	906	{
747	++dec->cur;	907	++dec->cur;
748	break;	908	break;
749	}	909	}
750		910
751	decode_he (dec, hv);	911	decode_he (dec, hv);
752	}	912	}
753	}	913	}
754	else	914	else
755	{	915	{
756	int pairs = decode_uint (dec);	916	UV pairs = decode_uint (dec);
		917
		918	WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
757		919
758	while (pairs--)	920	while (pairs--)
759	decode_he (dec, hv);	921	decode_he (dec, hv);
760	}	922	}
761		923
…		…
771	static SV *	933	static SV *
772	decode_str (dec_t *dec, int utf8)	934	decode_str (dec_t *dec, int utf8)
773	{	935	{
774	SV *sv = 0;	936	SV *sv = 0;
775		937
776	if ((*dec->cur & 31) == 31)	938	if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
777	{	939	{
778	// indefinite length strings	940	// indefinite length strings
779	++dec->cur;	941	++dec->cur;
780		942
781	unsigned char major = *dec->cur & 0xe0;	943	U8 major = *dec->cur & MAJOR_MISC;
782		944
783	sv = newSVpvn ("", 0);	945	sv = newSVpvn ("", 0);
784		946
785	for (;;)	947	for (;;)
786	{	948	{
787	WANT (1);	949	WANT (1);
788		950
789	if ((*dec->cur ^ major) >= 31)	951	if ((*dec->cur - major) > LENGTH_EXT8)
790	if (*dec->cur == (0xe0 \| 31))	952	if (*dec->cur == (MAJOR_MISC \| MINOR_INDEF))
791	{	953	{
792	++dec->cur;	954	++dec->cur;
793	break;	955	break;
794	}	956	}
795	else	957	else
…		…
814	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))	976	&& SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815	av_push (dec->stringref, SvREFCNT_inc_NN (sv));	977	av_push (dec->stringref, SvREFCNT_inc_NN (sv));
816	}	978	}
817		979
818	if (utf8)	980	if (utf8)
		981	{
		982	if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
		983	if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
		984	ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
		985
819	SvUTF8_on (sv);	986	SvUTF8_on (sv);
		987	}
820		988
821	return sv;	989	return sv;
822		990
823	fail:	991	fail:
824	SvREFCNT_dec (sv);	992	SvREFCNT_dec (sv);
…		…
843	sv = newRV_noinc (decode_sv (dec));	1011	sv = newRV_noinc (decode_sv (dec));
844	break;	1012	break;
845		1013
846	case CBOR_TAG_STRINGREF_NAMESPACE:	1014	case CBOR_TAG_STRINGREF_NAMESPACE:
847	{	1015	{
		1016	// do nmot use SAVETMPS/FREETMPS, as these will
		1017	// erase mortalised caches, e.g. "shareable"
848	ENTER; SAVETMPS;	1018	ENTER;
849		1019
850	SAVESPTR (dec->stringref);	1020	SAVESPTR (dec->stringref);
851	dec->stringref = (AV )sv_2mortal ((SV )newAV ());	1021	dec->stringref = (AV )sv_2mortal ((SV )newAV ());
852		1022
853	sv = decode_sv (dec);	1023	sv = decode_sv (dec);
854		1024
855	FREETMPS; LEAVE;	1025	LEAVE;
856	}	1026	}
857	break;	1027	break;
858		1028
859	case CBOR_TAG_STRINGREF:	1029	case CBOR_TAG_STRINGREF:
860	{	1030	{
861	if ((*dec->cur >> 5) != 0)	1031	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
862	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");	1032	ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863		1033
864	UV idx = decode_uint (dec);	1034	UV idx = decode_uint (dec);
865		1035
866	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))	1036	if (!dec->stringref \|\| (int)idx > AvFILLp (dec->stringref))
…		…
873	case CBOR_TAG_VALUE_SHAREABLE:	1043	case CBOR_TAG_VALUE_SHAREABLE:
874	{	1044	{
875	if (ecb_expect_false (!dec->shareable))	1045	if (ecb_expect_false (!dec->shareable))
876	dec->shareable = (AV )sv_2mortal ((SV )newAV ());	1046	dec->shareable = (AV )sv_2mortal ((SV )newAV ());
877		1047
		1048	if (dec->cbor.flags & F_ALLOW_CYCLES)
		1049	{
878	sv = newSV (0);	1050	sv = newSV (0);
879	av_push (dec->shareable, SvREFCNT_inc_NN (sv));	1051	av_push (dec->shareable, SvREFCNT_inc_NN (sv));
880		1052
881	SV *osv = decode_sv (dec);	1053	SV *osv = decode_sv (dec);
882	sv_setsv (sv, osv);	1054	sv_setsv (sv, osv);
883	SvREFCNT_dec_NN (osv);	1055	SvREFCNT_dec_NN (osv);
		1056	}
		1057	else
		1058	{
		1059	av_push (dec->shareable, &PL_sv_undef);
		1060	int idx = AvFILLp (dec->shareable);
		1061	sv = decode_sv (dec);
		1062	av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
		1063	}
884	}	1064	}
885	break;	1065	break;
886		1066
887	case CBOR_TAG_VALUE_SHAREDREF:	1067	case CBOR_TAG_VALUE_SHAREDREF:
888	{	1068	{
889	if ((*dec->cur >> 5) != 0)	1069	if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
890	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");	1070	ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
891		1071
892	UV idx = decode_uint (dec);	1072	UV idx = decode_uint (dec);
893		1073
894	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))	1074	if (!dec->shareable \|\| (int)idx > AvFILLp (dec->shareable))
895	ERR ("corrupted CBOR data (sharedref index out of bounds)");	1075	ERR ("corrupted CBOR data (sharedref index out of bounds)");
896		1076
897	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);	1077	sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
		1078
		1079	if (sv == &PL_sv_undef)
		1080	ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
898	}	1081	}
899	break;	1082	break;
900		1083
901	case CBOR_TAG_PERL_OBJECT:	1084	case CBOR_TAG_PERL_OBJECT:
902	{	1085	{
…		…
917	if (!method)	1100	if (!method)
918	ERR ("cannot decode perl-object (package does not have a THAW method)");	1101	ERR ("cannot decode perl-object (package does not have a THAW method)");
919		1102
920	dSP;	1103	dSP;
921		1104
922	ENTER; SAVETMPS; PUSHMARK (SP);	1105	ENTER; SAVETMPS;
		1106	PUSHMARK (SP);
923	EXTEND (SP, len + 1);	1107	EXTEND (SP, len + 1);
924	// we re-bless the reference to get overload and other niceties right	1108	// we re-bless the reference to get overload and other niceties right
925	PUSHs (*av_fetch (av, 0, 1));	1109	PUSHs (*av_fetch (av, 0, 1));
926	PUSHs (sv_cbor);	1110	PUSHs (sv_cbor);
927		1111
…		…
935	SPAGAIN;	1119	SPAGAIN;
936		1120
937	if (SvTRUE (ERRSV))	1121	if (SvTRUE (ERRSV))
938	{	1122	{
939	FREETMPS; LEAVE;	1123	FREETMPS; LEAVE;
940	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1124	ERR_ERRSV;
941	}	1125	}
942		1126
943	SvREFCNT_dec (sv);	1127	SvREFCNT_dec (sv);
944	sv = SvREFCNT_inc (POPs);	1128	sv = SvREFCNT_inc (POPs);
945		1129
…		…
949	}	1133	}
950	break;	1134	break;
951		1135
952	default:	1136	default:
953	{	1137	{
		1138	SV *tag_sv = newSVuv (tag);
		1139
954	sv = decode_sv (dec);	1140	sv = decode_sv (dec);
955		1141
956	dSP;	1142	dSP;
957	ENTER; SAVETMPS; PUSHMARK (SP);	1143	ENTER; SAVETMPS;
		1144	SAVESTACK_POS ();
		1145	PUSHMARK (SP);
958	EXTEND (SP, 2);	1146	EXTEND (SP, 2);
959	PUSHs (newSVuv (tag));	1147	PUSHs (tag_sv);
960	PUSHs (sv);	1148	PUSHs (sv);
961		1149
962	PUTBACK;	1150	PUTBACK;
963	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);	1151	int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY \| G_EVAL);
964	SPAGAIN;	1152	SPAGAIN;
965		1153
966	if (SvTRUE (ERRSV))	1154	if (SvTRUE (ERRSV))
967	{	1155	{
		1156	SvREFCNT_dec (tag_sv);
968	FREETMPS; LEAVE;	1157	FREETMPS; LEAVE;
969	ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));	1158	ERR_ERRSV;
970	}	1159	}
971		1160
972	if (count)	1161	if (count)
973	{	1162	{
		1163	SvREFCNT_dec (tag_sv);
974	SvREFCNT_dec (sv);	1164	SvREFCNT_dec (sv);
975	sv = SvREFCNT_inc (POPs);	1165	sv = SvREFCNT_inc (POPs);
976	}	1166	}
977	else	1167	else
978	{	1168	{
979	AV *av = newAV ();	1169	AV *av = newAV ();
980	av_push (av, newSVuv (tag));	1170	av_push (av, tag_sv);
981	av_push (av, sv);	1171	av_push (av, sv);
982		1172
983	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash	1173	HV *tagged_stash = !CBOR_SLOW \|\| cbor_tagged_stash
984	? cbor_tagged_stash	1174	? cbor_tagged_stash
985	: gv_stashpv ("CBOR::XS::Tagged" , 1);	1175	: gv_stashpv ("CBOR::XS::Tagged" , 1);
…		…
1003	static SV *	1193	static SV *
1004	decode_sv (dec_t *dec)	1194	decode_sv (dec_t *dec)
1005	{	1195	{
1006	WANT (1);	1196	WANT (1);
1007		1197
1008	switch (*dec->cur >> 5)	1198	switch (*dec->cur >> MAJOR_SHIFT)
1009	{	1199	{
1010	case 0: // unsigned int	1200	case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
1011	return newSVuv (decode_uint (dec));	1201	case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
1012	case 1: // negative int	1202	case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
1013	return newSViv (-1 - (IV)decode_uint (dec));	1203	case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
1014	case 2: // octet string	1204	case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
1015	return decode_str (dec, 0);	1205	case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
1016	case 3: // utf-8 string	1206	case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
1017	return decode_str (dec, 1);	1207
1018	case 4: // array	1208	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)	1209	switch (*dec->cur++ & MINOR_MASK)
1026	{	1210	{
1027	case 20:	1211	case SIMPLE_FALSE:
1028	#if CBOR_SLOW	1212	#if CBOR_SLOW
1029	types_false = get_bool ("Types::Serialiser::false");	1213	types_false = get_bool ("Types::Serialiser::false");
1030	#endif	1214	#endif
1031	return newSVsv (types_false);	1215	return newSVsv (types_false);
1032	case 21:	1216	case SIMPLE_TRUE:
1033	#if CBOR_SLOW	1217	#if CBOR_SLOW
1034	types_true = get_bool ("Types::Serialiser::true");	1218	types_true = get_bool ("Types::Serialiser::true");
1035	#endif	1219	#endif
1036	return newSVsv (types_true);	1220	return newSVsv (types_true);
1037	case 22:	1221	case SIMPLE_NULL:
1038	return newSVsv (&PL_sv_undef);	1222	return newSVsv (&PL_sv_undef);
1039	case 23:	1223	case SIMPLE_UNDEF:
1040	#if CBOR_SLOW	1224	#if CBOR_SLOW
1041	types_error = get_bool ("Types::Serialiser::error");	1225	types_error = get_bool ("Types::Serialiser::error");
1042	#endif	1226	#endif
1043	return newSVsv (types_error);	1227	return newSVsv (types_error);
1044		1228
1045	case 25:	1229	case MISC_FLOAT16:
1046	{	1230	{
1047	WANT (2);	1231	WANT (2);
1048		1232
1049	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];	1233	uint16_t fp = (dec->cur[0] << 8) \| dec->cur[1];
1050	dec->cur += 2;	1234	dec->cur += 2;
1051		1235
1052	return newSVnv (ecb_binary16_to_float (fp));	1236	return newSVnv (ecb_binary16_to_float (fp));
1053	}	1237	}
1054		1238
1055	case 26:	1239	case MISC_FLOAT32:
1056	{	1240	{
1057	uint32_t fp;	1241	uint32_t fp;
1058	WANT (4);	1242	WANT (4);
1059	memcpy (&fp, dec->cur, 4);	1243	memcpy (&fp, dec->cur, 4);
1060	dec->cur += 4;	1244	dec->cur += 4;
…		…
1063	fp = ecb_bswap32 (fp);	1247	fp = ecb_bswap32 (fp);
1064		1248
1065	return newSVnv (ecb_binary32_to_float (fp));	1249	return newSVnv (ecb_binary32_to_float (fp));
1066	}	1250	}
1067		1251
1068	case 27:	1252	case MISC_FLOAT64:
1069	{	1253	{
1070	uint64_t fp;	1254	uint64_t fp;
1071	WANT (8);	1255	WANT (8);
1072	memcpy (&fp, dec->cur, 8);	1256	memcpy (&fp, dec->cur, 8);
1073	dec->cur += 8;	1257	dec->cur += 8;
…		…
1076	fp = ecb_bswap64 (fp);	1260	fp = ecb_bswap64 (fp);
1077		1261
1078	return newSVnv (ecb_binary64_to_double (fp));	1262	return newSVnv (ecb_binary64_to_double (fp));
1079	}	1263	}
1080		1264
1081	// 0..19 unassigned	1265	// 0..19 unassigned simple
1082	// 24 reserved + unassigned (reserved values are not encodable)	1266	// 24 reserved + unassigned simple (reserved values are not encodable)
		1267	// 28-30 unassigned misc
		1268	// 31 break code
1083	default:	1269	default:
1084	ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");	1270	ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1085	}	1271	}
1086		1272
1087	break;	1273	break;
1088	}	1274	}
1089		1275
…		…
1092	}	1278	}
1093		1279
1094	static SV *	1280	static SV *
1095	decode_cbor (SV string, CBOR cbor, char **offset_return)	1281	decode_cbor (SV string, CBOR cbor, char **offset_return)
1096	{	1282	{
1097	dec_t dec = { };	1283	dec_t dec = { 0 };
1098	SV *sv;	1284	SV *sv;
1099	STRLEN len;	1285	STRLEN len;
1100	char *data = SvPVbyte (string, len);	1286	char *data = SvPVbyte (string, len);
1101		1287
1102	if (len > cbor->max_size && cbor->max_size)	1288	if (len > cbor->max_size && cbor->max_size)
…		…
1116	if (dec.cur != dec.end && !dec.err)	1302	if (dec.cur != dec.end && !dec.err)
1117	dec.err = "garbage after CBOR object";	1303	dec.err = "garbage after CBOR object";
1118		1304
1119	if (dec.err)	1305	if (dec.err)
1120	{	1306	{
		1307	if (dec.shareable)
		1308	{
		1309	// need to break cyclic links, which would all be in shareable
		1310	int i;
		1311	SV **svp;
		1312
		1313	for (i = av_len (dec.shareable) + 1; i--; )
		1314	if ((svp = av_fetch (dec.shareable, i, 0)))
		1315	sv_setsv (*svp, &PL_sv_undef);
		1316	}
		1317
1121	SvREFCNT_dec (sv);	1318	SvREFCNT_dec (sv);
		1319
		1320	if (dec.err_sv)
		1321	sv_2mortal (dec.err_sv);
		1322
1122	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);	1323	croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 )data, (int)(uint8_t)dec.cur);
1123	}	1324	}
1124		1325
1125	sv = sv_2mortal (sv);	1326	sv = sv_2mortal (sv);
1126		1327
1127	return sv;	1328	return sv;
1128	}	1329	}
1129		1330
		1331	/////////////////////////////////////////////////////////////////////////////
		1332	// incremental parser
		1333
		1334	#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
		1335
		1336	// returns 0 for notyet, 1 for success or error
		1337	static int
		1338	incr_parse (CBOR self, SV cborstr)
		1339	{
		1340	STRLEN cur;
		1341	SvPV (cborstr, cur);
		1342
		1343	while (ecb_expect_true (self->incr_need <= cur))
		1344	{
		1345	// table of integer count bytes
		1346	static I8 incr_len[MINOR_MASK + 1] = {
		1347	0, 0, 0, 0, 0, 0, 0, 0,
		1348	0, 0, 0, 0, 0, 0, 0, 0,
		1349	0, 0, 0, 0, 0, 0, 0, 0,
		1350	1, 2, 4, 8,-1,-1,-1,-2
		1351	};
		1352
		1353	const U8 *p = SvPVX (cborstr) + self->incr_pos;
		1354	U8 m = *p & MINOR_MASK;
		1355	IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1356	I8 ilen = incr_len[m];
		1357
		1358	self->incr_need = self->incr_pos + 1;
		1359
		1360	if (ecb_expect_false (ilen < 0))
		1361	{
		1362	if (m != MINOR_INDEF)
		1363	return 1; // error
		1364
		1365	if (*p == (MAJOR_MISC \| MINOR_INDEF))
		1366	{
		1367	if (count >= 0)
		1368	return 1; // error
		1369
		1370	count = 1;
		1371	}
		1372	else
		1373	{
		1374	av_push (self->incr_count, newSViv (-1)); //TODO: nest
		1375	count = -1;
		1376	}
		1377	}
		1378	else
		1379	{
		1380	self->incr_need += ilen;
		1381	if (ecb_expect_false (self->incr_need > cur))
		1382	return 0;
		1383
		1384	int major = *p >> MAJOR_SHIFT;
		1385
		1386	switch (major)
		1387	{
		1388	case MAJOR_TAG >> MAJOR_SHIFT:
		1389	++count; // tags merely prefix another value
		1390	break;
		1391
		1392	case MAJOR_BYTES >> MAJOR_SHIFT:
		1393	case MAJOR_TEXT >> MAJOR_SHIFT:
		1394	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1395	case MAJOR_MAP >> MAJOR_SHIFT:
		1396	{
		1397	UV len;
		1398
		1399	if (ecb_expect_false (ilen))
		1400	{
		1401	len = 0;
		1402
		1403	do {
		1404	len = (len << 8) \| *++p;
		1405	} while (--ilen);
		1406	}
		1407	else
		1408	len = m;
		1409
		1410	switch (major)
		1411	{
		1412	case MAJOR_BYTES >> MAJOR_SHIFT:
		1413	case MAJOR_TEXT >> MAJOR_SHIFT:
		1414	self->incr_need += len;
		1415	if (ecb_expect_false (self->incr_need > cur))
		1416	return 0;
		1417
		1418	break;
		1419
		1420	case MAJOR_MAP >> MAJOR_SHIFT:
		1421	len <<= 1;
		1422	case MAJOR_ARRAY >> MAJOR_SHIFT:
		1423	if (len)
		1424	{
		1425	av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
		1426	count = len + 1;
		1427	}
		1428	break;
		1429	}
		1430	}
		1431	}
		1432	}
		1433
		1434	self->incr_pos = self->incr_need;
		1435
		1436	if (count > 0)
		1437	{
		1438	while (!--count)
		1439	{
		1440	if (!AvFILLp (self->incr_count))
		1441	return 1; // done
		1442
		1443	SvREFCNT_dec_NN (av_pop (self->incr_count));
		1444	count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
		1445	}
		1446
		1447	SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
		1448	}
		1449	}
		1450
		1451	return 0;
		1452	}
		1453
		1454
1130	/////////////////////////////////////////////////////////////////////////////	1455	/////////////////////////////////////////////////////////////////////////////
1131	// XS interface functions	1456	// XS interface functions
1132		1457
1133	MODULE = CBOR::XS PACKAGE = CBOR::XS	1458	MODULE = CBOR::XS PACKAGE = CBOR::XS
1134		1459
…		…
1146		1471
1147	default_filter = newSVpv ("CBOR::XS::default_filter", 0);	1472	default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1148		1473
1149	sv_cbor = newSVpv ("CBOR", 0);	1474	sv_cbor = newSVpv ("CBOR", 0);
1150	SvREADONLY_on (sv_cbor);	1475	SvREADONLY_on (sv_cbor);
		1476
		1477	assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1151	}	1478	}
1152		1479
1153	PROTOTYPES: DISABLE	1480	PROTOTYPES: DISABLE
1154		1481
1155	void CLONE (...)	1482	void CLONE (...)
…		…
1174	void shrink (CBOR *self, int enable = 1)	1501	void shrink (CBOR *self, int enable = 1)
1175	ALIAS:	1502	ALIAS:
1176	shrink = F_SHRINK	1503	shrink = F_SHRINK
1177	allow_unknown = F_ALLOW_UNKNOWN	1504	allow_unknown = F_ALLOW_UNKNOWN
1178	allow_sharing = F_ALLOW_SHARING	1505	allow_sharing = F_ALLOW_SHARING
		1506	allow_cycles = F_ALLOW_CYCLES
1179	pack_strings = F_PACK_STRINGS	1507	pack_strings = F_PACK_STRINGS
		1508	text_keys = F_TEXT_KEYS
		1509	text_strings = F_TEXT_STRINGS
		1510	validate_utf8 = F_VALIDATE_UTF8
1180	PPCODE:	1511	PPCODE:
1181	{	1512	{
1182	if (enable)	1513	if (enable)
1183	self->flags \|= ix;	1514	self->flags \|= ix;
1184	else	1515	else
…		…
1190	void get_shrink (CBOR *self)	1521	void get_shrink (CBOR *self)
1191	ALIAS:	1522	ALIAS:
1192	get_shrink = F_SHRINK	1523	get_shrink = F_SHRINK
1193	get_allow_unknown = F_ALLOW_UNKNOWN	1524	get_allow_unknown = F_ALLOW_UNKNOWN
1194	get_allow_sharing = F_ALLOW_SHARING	1525	get_allow_sharing = F_ALLOW_SHARING
		1526	get_allow_cycles = F_ALLOW_CYCLES
1195	get_pack_strings = F_PACK_STRINGS	1527	get_pack_strings = F_PACK_STRINGS
		1528	get_text_keys = F_TEXT_KEYS
		1529	get_text_strings = F_TEXT_STRINGS
		1530	get_validate_utf8 = F_VALIDATE_UTF8
1196	PPCODE:	1531	PPCODE:
1197	XPUSHs (boolSV (self->flags & ix));	1532	XPUSHs (boolSV (self->flags & ix));
1198		1533
1199	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)	1534	void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1200	PPCODE:	1535	PPCODE:
…		…
1249	EXTEND (SP, 2);	1584	EXTEND (SP, 2);
1250	PUSHs (sv);	1585	PUSHs (sv);
1251	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));	1586	PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1252	}	1587	}
1253		1588
		1589	void incr_parse (CBOR self, SV cborstr)
		1590	ALIAS:
		1591	incr_parse_multiple = 1
		1592	PPCODE:
		1593	{
		1594	if (SvUTF8 (cborstr))
		1595	sv_utf8_downgrade (cborstr, 0);
		1596
		1597	if (!self->incr_count)
		1598	{
		1599	self->incr_count = newAV ();
		1600	self->incr_pos = 0;
		1601	self->incr_need = 1;
		1602
		1603	av_push (self->incr_count, newSViv (1));
		1604	}
		1605
		1606	do
		1607	{
		1608	if (!incr_parse (self, cborstr))
		1609	{
		1610	if (self->incr_need > self->max_size && self->max_size)
		1611	croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
		1612	(unsigned long)self->incr_need, (unsigned long)self->max_size);
		1613
		1614	break;
		1615	}
		1616
		1617	SV *sv;
		1618	char *offset;
		1619
		1620	PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
		1621	XPUSHs (sv);
		1622
		1623	sv_chop (cborstr, offset);
		1624
		1625	av_clear (self->incr_count);
		1626	av_push (self->incr_count, newSViv (1));
		1627
		1628	self->incr_pos = 0;
		1629	self->incr_need = self->incr_pos + 1;
		1630	}
		1631	while (ix);
		1632	}
		1633
		1634	void incr_reset (CBOR *self)
		1635	CODE:
		1636	{
		1637	SvREFCNT_dec (self->incr_count);
		1638	self->incr_count = 0;
		1639	}
		1640
1254	void DESTROY (CBOR *self)	1641	void DESTROY (CBOR *self)
1255	PPCODE:	1642	PPCODE:
1256	cbor_free (self);	1643	cbor_free (self);
1257		1644
1258	PROTOTYPES: ENABLE	1645	PROTOTYPES: ENABLE
1259		1646
1260	void encode_cbor (SV *scalar)	1647	void encode_cbor (SV *scalar)
		1648	ALIAS:
		1649	encode_cbor = 0
		1650	encode_cbor_sharing = F_ALLOW_SHARING
1261	PPCODE:	1651	PPCODE:
1262	{	1652	{
1263	CBOR cbor;	1653	CBOR cbor;
1264	cbor_init (&cbor);	1654	cbor_init (&cbor);
		1655	cbor.flags \|= ix;
1265	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;	1656	PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1266	XPUSHs (scalar);	1657	XPUSHs (scalar);
1267	}	1658	}
1268		1659
1269	void decode_cbor (SV *cborstr)	1660	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 1.59 by root, Fri Nov 25 23:37:27 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 1.59 by root, Fri Nov 25 23:37:27 2016 UTC