ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/CBOR-XS/XS.xs
(Generate patch)

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.18 by root, Wed Nov 20 01:09:46 2013 UTC vs.
Revision 1.34 by root, Sat Nov 30 16:19:59 2013 UTC

19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 19# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 20#endif
21#ifndef HvNAMEUTF8 21#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 22# define HvNAMEUTF8(hv) 0
23#endif 23#endif
24#ifndef SvREFCNT_dec_NN
25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26#endif
24 27
25// known tags 28// known tags
26enum cbor_tag 29enum cbor_tag
27{ 30{
28 // inofficial extensions (pending iana registration) 31 // extensions
32 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 33 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 34 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHARABLE = 26, // http://cbor.schmorp.de/value-sharing 35 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 36 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 37 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34 CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection 38 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 39
37 // rfc7049 40 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 41 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 42 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 57 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 58
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 59 CBOR_TAG_MAGIC = 55799 // self-describe cbor
57}; 60};
58 61
59#define F_SHRINK 0x00000001UL 62#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 63#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 64#define F_ALLOW_SHARING 0x00000004UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 65#define F_PACK_STRINGS 0x00000008UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO
64 66
65#define INIT_SIZE 32 // initial scalar size to be allocated 67#define INIT_SIZE 32 // initial scalar size to be allocated
66 68
67#define SB do { 69#define SB do {
68#define SE } while (0) 70#define SE } while (0)
80# define CBOR_SLOW 0 82# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 83# define CBOR_STASH cbor_stash
82#endif 84#endif
83 85
84static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 86static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 87static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 88
87typedef struct { 89typedef struct {
88 U32 flags; 90 U32 flags;
89 U32 max_depth; 91 U32 max_depth;
90 STRLEN max_size; 92 STRLEN max_size;
93 SV *filter;
91} CBOR; 94} CBOR;
92 95
93ecb_inline void 96ecb_inline void
94cbor_init (CBOR *cbor) 97cbor_init (CBOR *cbor)
95{ 98{
96 Zero (cbor, 1, CBOR); 99 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 100 cbor->max_depth = 512;
101}
102
103ecb_inline void
104cbor_free (CBOR *cbor)
105{
106 SvREFCNT_dec (cbor->filter);
98} 107}
99 108
100///////////////////////////////////////////////////////////////////////////// 109/////////////////////////////////////////////////////////////////////////////
101// utility functions 110// utility functions
102 111
124 SvPV_renew (sv, SvCUR (sv) + 1); 133 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 134#endif
126 } 135 }
127} 136}
128 137
138// minimum length of a string to be registered for stringref
139ecb_inline int
140minimum_string_length (UV idx)
141{
142 return idx > 23
143 ? idx > 0xffU
144 ? idx > 0xffffU
145 ? idx > 0xffffffffU
146 ? 11
147 : 7
148 : 5
149 : 4
150 : 3;
151}
152
129///////////////////////////////////////////////////////////////////////////// 153/////////////////////////////////////////////////////////////////////////////
130// encoder 154// encoder
131 155
132// structure used for encoding CBOR 156// structure used for encoding CBOR
133typedef struct 157typedef struct
135 char *cur; // SvPVX (sv) + current output position 159 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 160 char *end; // SvEND (sv)
137 SV *sv; // result scalar 161 SV *sv; // result scalar
138 CBOR cbor; 162 CBOR cbor;
139 U32 depth; // recursion level 163 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 164 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
165 UV stringref_idx;
141 HV *sharable; // ptr => index, or 0 166 HV *shareable; // ptr => index, or 0
142 HV *sharable_idx; 167 UV shareable_idx;
143} enc_t; 168} enc_t;
144 169
145ecb_inline void 170ecb_inline void
146need (enc_t *enc, STRLEN len) 171need (enc_t *enc, STRLEN len)
147{ 172{
164static void 189static void
165encode_uint (enc_t *enc, int major, UV len) 190encode_uint (enc_t *enc, int major, UV len)
166{ 191{
167 need (enc, 9); 192 need (enc, 9);
168 193
169 if (len < 24) 194 if (ecb_expect_true (len < 24))
170 *enc->cur++ = major | len; 195 *enc->cur++ = major | len;
171 else if (len <= 0xff) 196 else if (ecb_expect_true (len <= 0xff))
172 { 197 {
173 *enc->cur++ = major | 24; 198 *enc->cur++ = major | 24;
174 *enc->cur++ = len; 199 *enc->cur++ = len;
175 } 200 }
176 else if (len <= 0xffff) 201 else if (len <= 0xffff)
199 *enc->cur++ = len >> 8; 224 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 225 *enc->cur++ = len;
201 } 226 }
202} 227}
203 228
204static void 229ecb_inline void
230encode_tag (enc_t *enc, UV tag)
231{
232 encode_uint (enc, 0xc0, tag);
233}
234
235ecb_inline void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 236encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 237{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 238 encode_uint (enc, utf8 ? 0x60 : 0x40, len);
208 need (enc, len); 239 need (enc, len);
209 memcpy (enc->cur, str, len); 240 memcpy (enc->cur, str, len);
210 enc->cur += len; 241 enc->cur += len;
211} 242}
212 243
213ecb_inline void
214encode_tag (enc_t *enc, UV tag)
215{
216 encode_uint (enc, 0xc0, tag);
217}
218
219static int 244static void
220encode_sharable2 (enc_t *enc, SV *sv) 245encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
221{ 246{
222 if (!enc->sharable) 247 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
223 enc->sharable = (HV *)sv_2mortal ((SV *)newHV ()); 248 {
249 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
224 250
225 SV **svp = hv_fetch (enc->sharable, &sv, sizeof (sv), 1);
226
227 if (SvOK (*svp)) 251 if (SvOK (*svp))
228 { 252 {
253 // already registered, use stringref
229 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); 254 encode_tag (enc, CBOR_TAG_STRINGREF);
230 encode_uint (enc, 0x00, SvUV (*svp)); 255 encode_uint (enc, 0x00, SvUV (*svp));
231
232 return 1; 256 return;
233 } 257 }
234 else 258 else if (len >= minimum_string_length (enc->stringref_idx))
235 { 259 {
260 // register only
236 sv_setuv (*svp, enc->sharable_idx++); 261 sv_setuv (*svp, enc->stringref_idx);
237 encode_tag (enc, CBOR_TAG_VALUE_SHARABLE); 262 ++enc->stringref_idx;
238 263 }
239 return 0;
240 } 264 }
241}
242 265
243ecb_inline int 266 encode_str (enc, utf8, str, len);
244encode_sharable (enc_t *enc, SV *sv)
245{
246 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
247 && ecb_expect_false (SvREFCNT (sv) > 1))
248 return encode_sharable2 (enc, sv);
249
250 return 0;
251} 267}
252 268
253static void encode_sv (enc_t *enc, SV *sv); 269static void encode_sv (enc_t *enc, SV *sv);
254 270
255static void 271static void
294 while ((he = hv_iternext (hv))) 310 while ((he = hv_iternext (hv)))
295 { 311 {
296 if (HeKLEN (he) == HEf_SVKEY) 312 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 313 encode_sv (enc, HeSVKEY (he));
298 else 314 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 315 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 316
301 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 317 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302 } 318 }
303 319
304 if (mg) 320 if (mg)
309 325
310// encode objects, arrays and special \0=false and \1=true values. 326// encode objects, arrays and special \0=false and \1=true values.
311static void 327static void
312encode_rv (enc_t *enc, SV *sv) 328encode_rv (enc_t *enc, SV *sv)
313{ 329{
314 svtype svt;
315
316 SvGETMAGIC (sv); 330 SvGETMAGIC (sv);
331
317 svt = SvTYPE (sv); 332 svtype svt = SvTYPE (sv);
318
319 if (encode_sharable (enc, sv))
320 return;
321 333
322 if (ecb_expect_false (SvOBJECT (sv))) 334 if (ecb_expect_false (SvOBJECT (sv)))
323 { 335 {
324 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 336 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
325 ? types_boolean_stash 337 ? types_boolean_stash
330 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 342 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
331 ? cbor_tagged_stash 343 ? cbor_tagged_stash
332 : gv_stashpv ("CBOR::XS::Tagged" , 1); 344 : gv_stashpv ("CBOR::XS::Tagged" , 1);
333 345
334 HV *stash = SvSTASH (sv); 346 HV *stash = SvSTASH (sv);
335 GV *method;
336 347
337 if (stash == boolean_stash) 348 if (stash == boolean_stash)
349 {
338 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 350 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20);
351 return;
352 }
339 else if (stash == error_stash) 353 else if (stash == error_stash)
354 {
340 encode_ch (enc, 0xe0 | 23); 355 encode_ch (enc, 0xe0 | 23);
356 return;
357 }
341 else if (stash == tagged_stash) 358 else if (stash == tagged_stash)
342 { 359 {
343 if (svt != SVt_PVAV) 360 if (svt != SVt_PVAV)
344 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 361 croak ("encountered CBOR::XS::Tagged object that isn't an array");
345 362
346 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 363 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1)));
347 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 364 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
365
366 return;
367 }
368 }
369
370 if (ecb_expect_false (SvREFCNT (sv) > 1)
371 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
372 {
373 if (!enc->shareable)
374 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
375
376 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
377
378 if (SvOK (*svp))
348 } 379 {
380 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
381 encode_uint (enc, 0x00, SvUV (*svp));
382 return;
383 }
384 else
385 {
386 sv_setuv (*svp, enc->shareable_idx);
387 ++enc->shareable_idx;
388 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
389 }
390 }
391
392 if (ecb_expect_false (SvOBJECT (sv)))
393 {
394 HV *stash = SvSTASH (sv);
395 GV *method;
396
349 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 397 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
350 { 398 {
351 dSP; 399 dSP;
352 400
353 ENTER; SAVETMPS; PUSHMARK (SP); 401 ENTER; SAVETMPS; PUSHMARK (SP);
354 // we re-bless the reference to get overload and other niceties right 402 // we re-bless the reference to get overload and other niceties right
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 435 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
388 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 436 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
389 437
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 438 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 439 encode_uint (enc, 0x80, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 440 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 441
394 while (count) 442 while (count)
395 encode_sv (enc, SP[1 - count--]); 443 encode_sv (enc, SP[1 - count--]);
396 444
397 PUTBACK; 445 PUTBACK;
452static void 500static void
453encode_sv (enc_t *enc, SV *sv) 501encode_sv (enc_t *enc, SV *sv)
454{ 502{
455 SvGETMAGIC (sv); 503 SvGETMAGIC (sv);
456 504
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 505 if (SvPOKp (sv))
461 { 506 {
462 STRLEN len; 507 STRLEN len;
463 char *str = SvPV (sv, len); 508 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 509 encode_strref (enc, SvUTF8 (sv), str, len);
465 } 510 }
466 else if (SvNOKp (sv)) 511 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 512 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 513 else if (SvIOKp (sv))
469 { 514 {
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 539 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 540 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 541 enc.end = SvEND (enc.sv);
497 542
498 SvPOK_only (enc.sv); 543 SvPOK_only (enc.sv);
544
545 if (cbor->flags & F_PACK_STRINGS)
546 {
547 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
548 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
549 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
550 }
551
499 encode_sv (&enc, scalar); 552 encode_sv (&enc, scalar);
500 553
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 554 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 555 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
503 556
517 U8 *end; // end of input string 570 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 571 const char *err; // parse error, if != 0
519 CBOR cbor; 572 CBOR cbor;
520 U32 depth; // recursion depth 573 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 574 U32 maxdepth; // recursion depth limit
522 AV *sharable; 575 AV *shareable;
576 AV *stringref;
577 SV *decode_tagged;
523} dec_t; 578} dec_t;
524 579
525#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 580#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
526 581
527#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 582#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
559 | ((UV)dec->cur[-1]); 614 | ((UV)dec->cur[-1]);
560 615
561 case 27: 616 case 27:
562 WANT (9); 617 WANT (9);
563 dec->cur += 9; 618 dec->cur += 9;
619
620 return
621#if UVSIZE < 8
622 0
623#else
564 return (((UV)dec->cur[-8]) << 56) 624 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 625 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 626 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 627 | (((UV)dec->cur[-5]) << 32)
628#endif
568 | (((UV)dec->cur[-4]) << 24) 629 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 630 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 631 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 632 | ((UV)dec->cur[-1]);
572 633
625 686
626static void 687static void
627decode_he (dec_t *dec, HV *hv) 688decode_he (dec_t *dec, HV *hv)
628{ 689{
629 // for speed reasons, we specialcase single-string 690 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 691 // byte or utf-8 strings as keys, but only when !stringref
631 692
693 if (ecb_expect_true (!dec->stringref))
632 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 694 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27)
633 { 695 {
634 I32 len = decode_uint (dec); 696 I32 len = decode_uint (dec);
635 char *key = (char *)dec->cur; 697 char *key = (char *)dec->cur;
636 698
637 dec->cur += len; 699 dec->cur += len;
638 700
701 if (ecb_expect_false (dec->stringref))
702 av_push (dec->stringref, newSVpvn (key, len));
703
639 hv_store (hv, key, len, decode_sv (dec), 0); 704 hv_store (hv, key, len, decode_sv (dec), 0);
705
706 return;
640 } 707 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 708 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27)
642 { 709 {
643 I32 len = decode_uint (dec); 710 I32 len = decode_uint (dec);
644 char *key = (char *)dec->cur; 711 char *key = (char *)dec->cur;
645 712
646 dec->cur += len; 713 dec->cur += len;
647 714
715 if (ecb_expect_false (dec->stringref))
716 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));
717
648 hv_store (hv, key, -len, decode_sv (dec), 0); 718 hv_store (hv, key, -len, decode_sv (dec), 0);
719
720 return;
649 } 721 }
650 else 722
651 {
652 SV *k = decode_sv (dec); 723 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 724 SV *v = decode_sv (dec);
654 725
655 hv_store_ent (hv, k, v, 0); 726 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 727 SvREFCNT_dec (k);
657 }
658} 728}
659 729
660static SV * 730static SV *
661decode_hv (dec_t *dec) 731decode_hv (dec_t *dec)
662{ 732{
703{ 773{
704 SV *sv = 0; 774 SV *sv = 0;
705 775
706 if ((*dec->cur & 31) == 31) 776 if ((*dec->cur & 31) == 31)
707 { 777 {
778 // indefinite length strings
708 ++dec->cur; 779 ++dec->cur;
709 780
781 unsigned char major = *dec->cur & 0xe0;
782
710 sv = newSVpvn ("", 0); 783 sv = newSVpvn ("", 0);
711 784
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 785 for (;;)
714 { 786 {
715 WANT (1); 787 WANT (1);
716 788
789 if ((*dec->cur ^ major) >= 31)
717 if (*dec->cur == (0xe0 | 31)) 790 if (*dec->cur == (0xe0 | 31))
718 { 791 {
719 ++dec->cur; 792 ++dec->cur;
720 break; 793 break;
721 } 794 }
795 else
796 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 797
723 sv_catsv (sv, decode_sv (dec)); 798 STRLEN len = decode_uint (dec);
799
800 WANT (len);
801 sv_catpvn (sv, dec->cur, len);
802 dec->cur += len;
724 } 803 }
725 } 804 }
726 else 805 else
727 { 806 {
728 STRLEN len = decode_uint (dec); 807 STRLEN len = decode_uint (dec);
729 808
730 WANT (len); 809 WANT (len);
731 sv = newSVpvn (dec->cur, len); 810 sv = newSVpvn (dec->cur, len);
732 dec->cur += len; 811 dec->cur += len;
812
813 if (ecb_expect_false (dec->stringref)
814 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
815 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
733 } 816 }
734 817
735 if (utf8) 818 if (utf8)
736 SvUTF8_on (sv); 819 SvUTF8_on (sv);
737 820
743} 826}
744 827
745static SV * 828static SV *
746decode_tagged (dec_t *dec) 829decode_tagged (dec_t *dec)
747{ 830{
831 SV *sv = 0;
748 UV tag = decode_uint (dec); 832 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 833
834 WANT (1);
750 835
751 switch (tag) 836 switch (tag)
752 { 837 {
753 case CBOR_TAG_MAGIC: 838 case CBOR_TAG_MAGIC:
754 return sv; 839 sv = decode_sv (dec);
840 break;
755 841
756 case CBOR_TAG_INDIRECTION: 842 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 843 sv = newRV_noinc (decode_sv (dec));
844 break;
758 845
846 case CBOR_TAG_STRINGREF_NAMESPACE:
847 {
848 ENTER; SAVETMPS;
849
850 SAVESPTR (dec->stringref);
851 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
852
853 sv = decode_sv (dec);
854
855 FREETMPS; LEAVE;
856 }
857 break;
858
859 case CBOR_TAG_STRINGREF:
860 {
861 if ((*dec->cur >> 5) != 0)
862 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
863
864 UV idx = decode_uint (dec);
865
866 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
867 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
868
869 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
870 }
871 break;
872
759 case CBOR_TAG_VALUE_SHARABLE: 873 case CBOR_TAG_VALUE_SHAREABLE:
874 {
760 if (ecb_expect_false (!dec->sharable)) 875 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 876 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 877
878 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 879 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 880
765 return sv; 881 SV *osv = decode_sv (dec);
882 sv_setsv (sv, osv);
883 SvREFCNT_dec_NN (osv);
884 }
885 break;
766 886
767 case CBOR_TAG_VALUE_SHAREDREF: 887 case CBOR_TAG_VALUE_SHAREDREF:
768 { 888 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 889 if ((*dec->cur >> 5) != 0)
770 UV idx = SvUV (sv); 890 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 891
892 UV idx = decode_uint (dec);
893
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 894 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 895 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 896
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 897 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
778 } 898 }
899 break;
779 900
780 case CBOR_TAG_PERL_OBJECT: 901 case CBOR_TAG_PERL_OBJECT:
781 { 902 {
903 sv = decode_sv (dec);
904
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 905 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 906 ERR ("corrupted CBOR data (non-array perl object)");
784 907
785 AV *av = (AV *)SvRV (sv); 908 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 909 int len = av_len (av) + 1;
821 sv = SvREFCNT_inc (POPs); 944 sv = SvREFCNT_inc (POPs);
822 945
823 PUTBACK; 946 PUTBACK;
824 947
825 FREETMPS; LEAVE; 948 FREETMPS; LEAVE;
826
827 return sv;
828 } 949 }
950 break;
829 951
830 default: 952 default:
831 { 953 {
954 sv = decode_sv (dec);
955
956 dSP;
957 ENTER; SAVETMPS; PUSHMARK (SP);
958 EXTEND (SP, 2);
959 PUSHs (newSVuv (tag));
960 PUSHs (sv);
961
962 PUTBACK;
963 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
964 SPAGAIN;
965
966 if (SvTRUE (ERRSV))
967 {
968 FREETMPS; LEAVE;
969 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
970 }
971
972 if (count)
973 {
974 SvREFCNT_dec (sv);
975 sv = SvREFCNT_inc (POPs);
976 }
977 else
978 {
832 AV *av = newAV (); 979 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 980 av_push (av, newSVuv (tag));
834 av_push (av, sv); 981 av_push (av, sv);
835 982
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 983 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 984 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 985 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 986 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 987 }
988
989 PUTBACK;
990
991 FREETMPS; LEAVE;
992 }
993 break;
842 } 994 }
995
996 return sv;
843 997
844fail: 998fail:
845 SvREFCNT_dec (sv); 999 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1000 return &PL_sv_undef;
847} 1001}
988 1142
989 types_true = get_bool ("Types::Serialiser::true" ); 1143 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1144 types_false = get_bool ("Types::Serialiser::false");
991 types_error = get_bool ("Types::Serialiser::error"); 1145 types_error = get_bool ("Types::Serialiser::error");
992 1146
1147 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1148
993 sv_cbor = newSVpv ("CBOR", 0); 1149 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1150 SvREADONLY_on (sv_cbor);
995} 1151}
996 1152
997PROTOTYPES: DISABLE 1153PROTOTYPES: DISABLE
1018void shrink (CBOR *self, int enable = 1) 1174void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1175 ALIAS:
1020 shrink = F_SHRINK 1176 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1177 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1178 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS
1024 dedup_strings = F_DEDUP_STRINGS 1179 pack_strings = F_PACK_STRINGS
1025 PPCODE: 1180 PPCODE:
1026{ 1181{
1027 if (enable) 1182 if (enable)
1028 self->flags |= ix; 1183 self->flags |= ix;
1029 else 1184 else
1035void get_shrink (CBOR *self) 1190void get_shrink (CBOR *self)
1036 ALIAS: 1191 ALIAS:
1037 get_shrink = F_SHRINK 1192 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1193 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1194 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS
1041 get_dedup_strings = F_DEDUP_STRINGS 1195 get_pack_strings = F_PACK_STRINGS
1042 PPCODE: 1196 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1197 XPUSHs (boolSV (self->flags & ix));
1044 1198
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1199void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1200 PPCODE:
1059 XPUSHs (ST (0)); 1213 XPUSHs (ST (0));
1060 1214
1061int get_max_size (CBOR *self) 1215int get_max_size (CBOR *self)
1062 CODE: 1216 CODE:
1063 RETVAL = self->max_size; 1217 RETVAL = self->max_size;
1218 OUTPUT:
1219 RETVAL
1220
1221void filter (CBOR *self, SV *filter = 0)
1222 PPCODE:
1223 SvREFCNT_dec (self->filter);
1224 self->filter = filter ? newSVsv (filter) : filter;
1225 XPUSHs (ST (0));
1226
1227SV *get_filter (CBOR *self)
1228 CODE:
1229 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1064 OUTPUT: 1230 OUTPUT:
1065 RETVAL 1231 RETVAL
1066 1232
1067void encode (CBOR *self, SV *scalar) 1233void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1234 PPCODE:
1083 EXTEND (SP, 2); 1249 EXTEND (SP, 2);
1084 PUSHs (sv); 1250 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1251 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1252}
1087 1253
1254void DESTROY (CBOR *self)
1255 PPCODE:
1256 cbor_free (self);
1257
1088PROTOTYPES: ENABLE 1258PROTOTYPES: ENABLE
1089 1259
1090void encode_cbor (SV *scalar) 1260void encode_cbor (SV *scalar)
1091 PPCODE: 1261 PPCODE:
1092{ 1262{

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines