… | |
… | |
12 | #if defined(__BORLANDC__) || defined(_MSC_VER) |
12 | #if defined(__BORLANDC__) || defined(_MSC_VER) |
13 | # define snprintf _snprintf // C compilers have this in stdio.h |
13 | # define snprintf _snprintf // C compilers have this in stdio.h |
14 | #endif |
14 | #endif |
15 | |
15 | |
16 | // some old perls do not have this, try to make it work, no |
16 | // some old perls do not have this, try to make it work, no |
17 | // guarentees, though. if it breaks, you get to keep the pieces. |
17 | // guarantees, though. if it breaks, you get to keep the pieces. |
18 | #ifndef UTF8_MAXBYTES |
18 | #ifndef UTF8_MAXBYTES |
19 | # define UTF8_MAXBYTES 13 |
19 | # define UTF8_MAXBYTES 13 |
20 | #endif |
20 | #endif |
21 | |
21 | |
|
|
22 | // three extra for rounding, sign, and end of string |
22 | #define IVUV_MAXCHARS (sizeof (UV) * CHAR_BIT * 28 / 93 + 2) |
23 | #define IVUV_MAXCHARS (sizeof (UV) * CHAR_BIT * 28 / 93 + 3) |
23 | |
24 | |
24 | #define F_ASCII 0x00000001UL |
25 | #define F_ASCII 0x00000001UL |
25 | #define F_LATIN1 0x00000002UL |
26 | #define F_LATIN1 0x00000002UL |
26 | #define F_UTF8 0x00000004UL |
27 | #define F_UTF8 0x00000004UL |
27 | #define F_INDENT 0x00000008UL |
28 | #define F_INDENT 0x00000008UL |
… | |
… | |
31 | #define F_ALLOW_NONREF 0x00000100UL |
32 | #define F_ALLOW_NONREF 0x00000100UL |
32 | #define F_SHRINK 0x00000200UL |
33 | #define F_SHRINK 0x00000200UL |
33 | #define F_ALLOW_BLESSED 0x00000400UL |
34 | #define F_ALLOW_BLESSED 0x00000400UL |
34 | #define F_CONV_BLESSED 0x00000800UL |
35 | #define F_CONV_BLESSED 0x00000800UL |
35 | #define F_RELAXED 0x00001000UL |
36 | #define F_RELAXED 0x00001000UL |
36 | |
37 | #define F_ALLOW_UNKNOWN 0x00002000UL |
37 | #define F_MAXDEPTH 0xf8000000UL |
|
|
38 | #define S_MAXDEPTH 27 |
|
|
39 | #define F_MAXSIZE 0x01f00000UL |
|
|
40 | #define S_MAXSIZE 20 |
|
|
41 | #define F_HOOK 0x00080000UL // some hooks exist, so slow-path processing |
38 | #define F_HOOK 0x00080000UL // some hooks exist, so slow-path processing |
42 | |
39 | |
43 | #define DEC_DEPTH(flags) (1UL << ((flags & F_MAXDEPTH) >> S_MAXDEPTH)) |
|
|
44 | #define DEC_SIZE(flags) (1UL << ((flags & F_MAXSIZE ) >> S_MAXSIZE )) |
|
|
45 | |
|
|
46 | #define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER |
40 | #define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER |
47 | #define F_DEFAULT (9UL << S_MAXDEPTH) |
|
|
48 | |
41 | |
49 | #define INIT_SIZE 32 // initial scalar size to be allocated |
42 | #define INIT_SIZE 32 // initial scalar size to be allocated |
50 | #define INDENT_STEP 3 // spaces per indentation level |
43 | #define INDENT_STEP 3 // spaces per indentation level |
51 | |
44 | |
52 | #define SHORT_STRING_LEN 16384 // special-case strings of up to this size |
45 | #define SHORT_STRING_LEN 16384 // special-case strings of up to this size |
… | |
… | |
67 | |
60 | |
68 | #define IN_RANGE_INC(type,val,beg,end) \ |
61 | #define IN_RANGE_INC(type,val,beg,end) \ |
69 | ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \ |
62 | ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \ |
70 | <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg))) |
63 | <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg))) |
71 | |
64 | |
|
|
65 | #define ERR_NESTING_EXCEEDED "json text or perl structure exceeds maximum nesting level (max_depth set too low?)" |
|
|
66 | |
72 | #ifdef USE_ITHREADS |
67 | #ifdef USE_ITHREADS |
73 | # define JSON_SLOW 1 |
68 | # define JSON_SLOW 1 |
74 | # define JSON_STASH (json_stash ? json_stash : gv_stashpv ("JSON::XS", 1)) |
69 | # define JSON_STASH (json_stash ? json_stash : gv_stashpv ("JSON::XS", 1)) |
75 | #else |
70 | #else |
76 | # define JSON_SLOW 0 |
71 | # define JSON_SLOW 0 |
… | |
… | |
82 | |
77 | |
83 | enum { |
78 | enum { |
84 | INCR_M_WS = 0, // initial whitespace skipping, must be 0 |
79 | INCR_M_WS = 0, // initial whitespace skipping, must be 0 |
85 | INCR_M_STR, // inside string |
80 | INCR_M_STR, // inside string |
86 | INCR_M_BS, // inside backslash |
81 | INCR_M_BS, // inside backslash |
|
|
82 | INCR_M_C0, // inside comment in initial whitespace sequence |
|
|
83 | INCR_M_C1, // inside comment in other places |
87 | INCR_M_JSON // outside anything, count nesting |
84 | INCR_M_JSON // outside anything, count nesting |
88 | }; |
85 | }; |
89 | |
86 | |
90 | #define INCR_DONE(json) (!(json)->incr_nest && (json)->incr_mode == INCR_M_JSON) |
87 | #define INCR_DONE(json) ((json)->incr_nest <= 0 && (json)->incr_mode == INCR_M_JSON) |
91 | |
88 | |
92 | typedef struct { |
89 | typedef struct { |
93 | U32 flags; |
90 | U32 flags; |
|
|
91 | U32 max_depth; |
|
|
92 | STRLEN max_size; |
|
|
93 | |
94 | SV *cb_object; |
94 | SV *cb_object; |
95 | HV *cb_sk_object; |
95 | HV *cb_sk_object; |
96 | |
96 | |
97 | // for the incremental parser |
97 | // for the incremental parser |
98 | SV *incr_text; // the source text so far |
98 | SV *incr_text; // the source text so far |
99 | STRLEN incr_pos; // the current offset into the text |
99 | STRLEN incr_pos; // the current offset into the text |
100 | int incr_nest; // {[]}-nesting level |
100 | int incr_nest; // {[]}-nesting level |
101 | int incr_mode; |
101 | unsigned char incr_mode; |
102 | } JSON; |
102 | } JSON; |
|
|
103 | |
|
|
104 | INLINE void |
|
|
105 | json_init (JSON *json) |
|
|
106 | { |
|
|
107 | Zero (json, 1, JSON); |
|
|
108 | json->max_depth = 512; |
|
|
109 | } |
103 | |
110 | |
104 | ///////////////////////////////////////////////////////////////////////////// |
111 | ///////////////////////////////////////////////////////////////////////////// |
105 | // utility functions |
112 | // utility functions |
106 | |
113 | |
|
|
114 | INLINE SV * |
|
|
115 | get_bool (const char *name) |
|
|
116 | { |
|
|
117 | SV *sv = get_sv (name, 1); |
|
|
118 | |
|
|
119 | SvREADONLY_on (sv); |
|
|
120 | SvREADONLY_on (SvRV (sv)); |
|
|
121 | |
|
|
122 | return sv; |
|
|
123 | } |
|
|
124 | |
107 | INLINE void |
125 | INLINE void |
108 | shrink (SV *sv) |
126 | shrink (SV *sv) |
109 | { |
127 | { |
110 | sv_utf8_downgrade (sv, 1); |
128 | sv_utf8_downgrade (sv, 1); |
|
|
129 | |
111 | if (SvLEN (sv) > SvCUR (sv) + 1) |
130 | if (SvLEN (sv) > SvCUR (sv) + 1) |
112 | { |
131 | { |
113 | #ifdef SvPV_shrink_to_cur |
132 | #ifdef SvPV_shrink_to_cur |
114 | SvPV_shrink_to_cur (sv); |
133 | SvPV_shrink_to_cur (sv); |
115 | #elif defined (SvPV_renew) |
134 | #elif defined (SvPV_renew) |
… | |
… | |
159 | *s++ = 0x80 | ( ch & 0x3f); |
178 | *s++ = 0x80 | ( ch & 0x3f); |
160 | |
179 | |
161 | return s; |
180 | return s; |
162 | } |
181 | } |
163 | |
182 | |
|
|
183 | // convert offset pointer to character index, sv must be string |
|
|
184 | static STRLEN |
|
|
185 | ptr_to_index (SV *sv, char *offset) |
|
|
186 | { |
|
|
187 | return SvUTF8 (sv) |
|
|
188 | ? utf8_distance (offset, SvPVX (sv)) |
|
|
189 | : offset - SvPVX (sv); |
|
|
190 | } |
|
|
191 | |
|
|
192 | ///////////////////////////////////////////////////////////////////////////// |
|
|
193 | // fp hell |
|
|
194 | |
|
|
195 | // scan a group of digits, and a trailing exponent |
|
|
196 | static void |
|
|
197 | json_atof_scan1 (const char *s, NV *accum, int *expo, int postdp) |
|
|
198 | { |
|
|
199 | UV uaccum = 0; |
|
|
200 | int eaccum = 0; |
|
|
201 | |
|
|
202 | for (;;) |
|
|
203 | { |
|
|
204 | U8 dig = (U8)*s - '0'; |
|
|
205 | |
|
|
206 | if (expect_false (dig >= 10)) |
|
|
207 | { |
|
|
208 | if (dig == (U8)((U8)'.' - (U8)'0')) |
|
|
209 | { |
|
|
210 | ++s; |
|
|
211 | json_atof_scan1 (s, accum, expo, 1); |
|
|
212 | } |
|
|
213 | else if ((dig | ' ') == 'e' - '0') |
|
|
214 | { |
|
|
215 | int exp2 = 0; |
|
|
216 | int neg = 0; |
|
|
217 | |
|
|
218 | ++s; |
|
|
219 | |
|
|
220 | if (*s == '-') |
|
|
221 | { |
|
|
222 | ++s; |
|
|
223 | neg = 1; |
|
|
224 | } |
|
|
225 | else if (*s == '+') |
|
|
226 | ++s; |
|
|
227 | |
|
|
228 | while ((dig = (U8)*s - '0') < 10) |
|
|
229 | exp2 = exp2 * 10 + *s++ - '0'; |
|
|
230 | |
|
|
231 | *expo += neg ? -exp2 : exp2; |
|
|
232 | } |
|
|
233 | |
|
|
234 | break; |
|
|
235 | } |
|
|
236 | |
|
|
237 | ++s; |
|
|
238 | |
|
|
239 | uaccum = uaccum * 10 + dig; |
|
|
240 | ++eaccum; |
|
|
241 | |
|
|
242 | // if we have too many digits, then recurse for more |
|
|
243 | // we actually do this for rather few digits |
|
|
244 | if (uaccum >= (UV_MAX - 9) / 10) |
|
|
245 | { |
|
|
246 | if (postdp) *expo -= eaccum; |
|
|
247 | json_atof_scan1 (s, accum, expo, postdp); |
|
|
248 | if (postdp) *expo += eaccum; |
|
|
249 | |
|
|
250 | break; |
|
|
251 | } |
|
|
252 | } |
|
|
253 | |
|
|
254 | // this relies greatly on the quality of the pow () |
|
|
255 | // implementation of the platform, but a good |
|
|
256 | // implementation is hard to beat. |
|
|
257 | if (postdp) *expo -= eaccum; |
|
|
258 | *accum += uaccum * Perl_pow (10., *expo); |
|
|
259 | *expo += eaccum; |
|
|
260 | } |
|
|
261 | |
|
|
262 | static NV |
|
|
263 | json_atof (const char *s) |
|
|
264 | { |
|
|
265 | NV accum = 0.; |
|
|
266 | int expo = 0; |
|
|
267 | int neg = 0; |
|
|
268 | |
|
|
269 | if (*s == '-') |
|
|
270 | { |
|
|
271 | ++s; |
|
|
272 | neg = 1; |
|
|
273 | } |
|
|
274 | |
|
|
275 | json_atof_scan1 (s, &accum, &expo, 0); |
|
|
276 | |
|
|
277 | return neg ? -accum : accum; |
|
|
278 | } |
164 | ///////////////////////////////////////////////////////////////////////////// |
279 | ///////////////////////////////////////////////////////////////////////////// |
165 | // encoder |
280 | // encoder |
166 | |
281 | |
167 | // structure used for encoding JSON |
282 | // structure used for encoding JSON |
168 | typedef struct |
283 | typedef struct |
… | |
… | |
170 | char *cur; // SvPVX (sv) + current output position |
285 | char *cur; // SvPVX (sv) + current output position |
171 | char *end; // SvEND (sv) |
286 | char *end; // SvEND (sv) |
172 | SV *sv; // result scalar |
287 | SV *sv; // result scalar |
173 | JSON json; |
288 | JSON json; |
174 | U32 indent; // indentation level |
289 | U32 indent; // indentation level |
175 | U32 maxdepth; // max. indentation/recursion level |
|
|
176 | UV limit; // escape character values >= this value when encoding |
290 | UV limit; // escape character values >= this value when encoding |
177 | } enc_t; |
291 | } enc_t; |
178 | |
292 | |
179 | INLINE void |
293 | INLINE void |
180 | need (enc_t *enc, STRLEN len) |
294 | need (enc_t *enc, STRLEN len) |
181 | { |
295 | { |
182 | if (expect_false (enc->cur + len >= enc->end)) |
296 | if (expect_false (enc->cur + len >= enc->end)) |
183 | { |
297 | { |
184 | STRLEN cur = enc->cur - SvPVX (enc->sv); |
298 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
185 | SvGROW (enc->sv, cur + len + 1); |
299 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
186 | enc->cur = SvPVX (enc->sv) + cur; |
300 | enc->cur = SvPVX (enc->sv) + cur; |
187 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
301 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
188 | } |
302 | } |
189 | } |
303 | } |
190 | |
304 | |
… | |
… | |
265 | (int)((uch - 0x10000) % 0x400 + 0xDC00)); |
379 | (int)((uch - 0x10000) % 0x400 + 0xDC00)); |
266 | enc->cur += 12; |
380 | enc->cur += 12; |
267 | } |
381 | } |
268 | else |
382 | else |
269 | { |
383 | { |
270 | static char hexdigit [16] = "0123456789abcdef"; |
|
|
271 | need (enc, len += 5); |
384 | need (enc, len += 5); |
272 | *enc->cur++ = '\\'; |
385 | *enc->cur++ = '\\'; |
273 | *enc->cur++ = 'u'; |
386 | *enc->cur++ = 'u'; |
274 | *enc->cur++ = hexdigit [ uch >> 12 ]; |
387 | *enc->cur++ = PL_hexdigit [ uch >> 12 ]; |
275 | *enc->cur++ = hexdigit [(uch >> 8) & 15]; |
388 | *enc->cur++ = PL_hexdigit [(uch >> 8) & 15]; |
276 | *enc->cur++ = hexdigit [(uch >> 4) & 15]; |
389 | *enc->cur++ = PL_hexdigit [(uch >> 4) & 15]; |
277 | *enc->cur++ = hexdigit [(uch >> 0) & 15]; |
390 | *enc->cur++ = PL_hexdigit [(uch >> 0) & 15]; |
278 | } |
391 | } |
279 | |
392 | |
280 | str += clen; |
393 | str += clen; |
281 | } |
394 | } |
282 | else if (enc->json.flags & F_LATIN1) |
395 | else if (enc->json.flags & F_LATIN1) |
… | |
… | |
353 | static void |
466 | static void |
354 | encode_av (enc_t *enc, AV *av) |
467 | encode_av (enc_t *enc, AV *av) |
355 | { |
468 | { |
356 | int i, len = av_len (av); |
469 | int i, len = av_len (av); |
357 | |
470 | |
358 | if (enc->indent >= enc->maxdepth) |
471 | if (enc->indent >= enc->json.max_depth) |
359 | croak ("data structure too deep (hit recursion limit)"); |
472 | croak (ERR_NESTING_EXCEEDED); |
360 | |
473 | |
361 | encode_ch (enc, '['); |
474 | encode_ch (enc, '['); |
362 | |
475 | |
363 | if (len >= 0) |
476 | if (len >= 0) |
364 | { |
477 | { |
… | |
… | |
439 | static void |
552 | static void |
440 | encode_hv (enc_t *enc, HV *hv) |
553 | encode_hv (enc_t *enc, HV *hv) |
441 | { |
554 | { |
442 | HE *he; |
555 | HE *he; |
443 | |
556 | |
444 | if (enc->indent >= enc->maxdepth) |
557 | if (enc->indent >= enc->json.max_depth) |
445 | croak ("data structure too deep (hit recursion limit)"); |
558 | croak (ERR_NESTING_EXCEEDED); |
446 | |
559 | |
447 | encode_ch (enc, '{'); |
560 | encode_ch (enc, '{'); |
448 | |
561 | |
449 | // for canonical output we have to sort by keys first |
562 | // for canonical output we have to sort by keys first |
450 | // actually, this is mostly due to the stupid so-called |
563 | // actually, this is mostly due to the stupid so-called |
451 | // security workaround added somewhere in 5.8.x. |
564 | // security workaround added somewhere in 5.8.x |
452 | // that randomises hash orderings |
565 | // that randomises hash orderings |
453 | if (enc->json.flags & F_CANONICAL) |
566 | if (enc->json.flags & F_CANONICAL && !SvRMAGICAL (hv)) |
454 | { |
567 | { |
455 | int count = hv_iterinit (hv); |
568 | int count = hv_iterinit (hv); |
456 | |
569 | |
457 | if (SvMAGICAL (hv)) |
570 | if (SvMAGICAL (hv)) |
458 | { |
571 | { |
… | |
… | |
631 | |
744 | |
632 | if (len == 1 && *pv == '1') |
745 | if (len == 1 && *pv == '1') |
633 | encode_str (enc, "true", 4, 0); |
746 | encode_str (enc, "true", 4, 0); |
634 | else if (len == 1 && *pv == '0') |
747 | else if (len == 1 && *pv == '0') |
635 | encode_str (enc, "false", 5, 0); |
748 | encode_str (enc, "false", 5, 0); |
|
|
749 | else if (enc->json.flags & F_ALLOW_UNKNOWN) |
|
|
750 | encode_str (enc, "null", 4, 0); |
636 | else |
751 | else |
637 | croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1", |
752 | croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1", |
638 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
753 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
639 | } |
754 | } |
|
|
755 | else if (enc->json.flags & F_ALLOW_UNKNOWN) |
|
|
756 | encode_str (enc, "null", 4, 0); |
640 | else |
757 | else |
641 | croak ("encountered %s, but JSON can only represent references to arrays or hashes", |
758 | croak ("encountered %s, but JSON can only represent references to arrays or hashes", |
642 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
759 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
643 | } |
760 | } |
644 | |
761 | |
… | |
… | |
706 | : snprintf (enc->cur, IVUV_MAXCHARS, "%"IVdf, (IV)SvIVX (sv)); |
823 | : snprintf (enc->cur, IVUV_MAXCHARS, "%"IVdf, (IV)SvIVX (sv)); |
707 | } |
824 | } |
708 | } |
825 | } |
709 | else if (SvROK (sv)) |
826 | else if (SvROK (sv)) |
710 | encode_rv (enc, SvRV (sv)); |
827 | encode_rv (enc, SvRV (sv)); |
711 | else if (!SvOK (sv)) |
828 | else if (!SvOK (sv) || enc->json.flags & F_ALLOW_UNKNOWN) |
712 | encode_str (enc, "null", 4, 0); |
829 | encode_str (enc, "null", 4, 0); |
713 | else |
830 | else |
714 | croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this", |
831 | croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this", |
715 | SvPV_nolen (sv), SvFLAGS (sv)); |
832 | SvPV_nolen (sv), SvFLAGS (sv)); |
716 | } |
833 | } |
… | |
… | |
726 | enc.json = *json; |
843 | enc.json = *json; |
727 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
844 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
728 | enc.cur = SvPVX (enc.sv); |
845 | enc.cur = SvPVX (enc.sv); |
729 | enc.end = SvEND (enc.sv); |
846 | enc.end = SvEND (enc.sv); |
730 | enc.indent = 0; |
847 | enc.indent = 0; |
731 | enc.maxdepth = DEC_DEPTH (enc.json.flags); |
|
|
732 | enc.limit = enc.json.flags & F_ASCII ? 0x000080UL |
848 | enc.limit = enc.json.flags & F_ASCII ? 0x000080UL |
733 | : enc.json.flags & F_LATIN1 ? 0x000100UL |
849 | : enc.json.flags & F_LATIN1 ? 0x000100UL |
734 | : 0x110000UL; |
850 | : 0x110000UL; |
735 | |
851 | |
736 | SvPOK_only (enc.sv); |
852 | SvPOK_only (enc.sv); |
737 | encode_sv (&enc, scalar); |
853 | encode_sv (&enc, scalar); |
|
|
854 | encode_nl (&enc); |
738 | |
855 | |
739 | SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); |
856 | SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); |
740 | *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings |
857 | *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings |
741 | |
858 | |
742 | if (!(enc.json.flags & (F_ASCII | F_LATIN1 | F_UTF8))) |
859 | if (!(enc.json.flags & (F_ASCII | F_LATIN1 | F_UTF8))) |
… | |
… | |
803 | if (*dec->cur != ch) \ |
920 | if (*dec->cur != ch) \ |
804 | ERR (# ch " expected"); \ |
921 | ERR (# ch " expected"); \ |
805 | ++dec->cur; \ |
922 | ++dec->cur; \ |
806 | SE |
923 | SE |
807 | |
924 | |
808 | #define DEC_INC_DEPTH if (++dec->depth > dec->maxdepth) ERR ("json datastructure exceeds maximum nesting level (set a higher max_depth)") |
925 | #define DEC_INC_DEPTH if (++dec->depth > dec->json.max_depth) ERR (ERR_NESTING_EXCEEDED) |
809 | #define DEC_DEC_DEPTH --dec->depth |
926 | #define DEC_DEC_DEPTH --dec->depth |
810 | |
927 | |
811 | static SV *decode_sv (dec_t *dec); |
928 | static SV *decode_sv (dec_t *dec); |
812 | |
929 | |
813 | static signed char decode_hexdigit[256]; |
930 | static signed char decode_hexdigit[256]; |
… | |
… | |
922 | else if (expect_true (ch >= 0x20 && ch < 0x80)) |
1039 | else if (expect_true (ch >= 0x20 && ch < 0x80)) |
923 | *cur++ = ch; |
1040 | *cur++ = ch; |
924 | else if (ch >= 0x80) |
1041 | else if (ch >= 0x80) |
925 | { |
1042 | { |
926 | STRLEN clen; |
1043 | STRLEN clen; |
927 | UV uch; |
|
|
928 | |
1044 | |
929 | --dec_cur; |
1045 | --dec_cur; |
930 | |
1046 | |
931 | uch = decode_utf8 (dec_cur, dec->end - dec_cur, &clen); |
1047 | decode_utf8 (dec_cur, dec->end - dec_cur, &clen); |
932 | if (clen == (STRLEN)-1) |
1048 | if (clen == (STRLEN)-1) |
933 | ERR ("malformed UTF-8 character in JSON string"); |
1049 | ERR ("malformed UTF-8 character in JSON string"); |
934 | |
1050 | |
935 | do |
1051 | do |
936 | *cur++ = *dec_cur++; |
1052 | *cur++ = *dec_cur++; |
… | |
… | |
953 | { |
1069 | { |
954 | STRLEN len = cur - buf; |
1070 | STRLEN len = cur - buf; |
955 | |
1071 | |
956 | if (sv) |
1072 | if (sv) |
957 | { |
1073 | { |
958 | SvGROW (sv, SvCUR (sv) + len + 1); |
1074 | STRLEN cur = SvCUR (sv); |
|
|
1075 | |
|
|
1076 | if (SvLEN (sv) <= cur + len) |
|
|
1077 | SvGROW (sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
|
|
1078 | |
959 | memcpy (SvPVX (sv) + SvCUR (sv), buf, len); |
1079 | memcpy (SvPVX (sv) + SvCUR (sv), buf, len); |
960 | SvCUR_set (sv, SvCUR (sv) + len); |
1080 | SvCUR_set (sv, SvCUR (sv) + len); |
961 | } |
1081 | } |
962 | else |
1082 | else |
963 | sv = newSVpvn (buf, len); |
1083 | sv = newSVpvn (buf, len); |
… | |
… | |
1054 | |
1174 | |
1055 | // special case the rather common 1..5-digit-int case |
1175 | // special case the rather common 1..5-digit-int case |
1056 | if (*start == '-') |
1176 | if (*start == '-') |
1057 | switch (len) |
1177 | switch (len) |
1058 | { |
1178 | { |
1059 | case 2: return newSViv (-( start [1] - '0' * 1)); |
1179 | case 2: return newSViv (-(IV)( start [1] - '0' * 1)); |
1060 | case 3: return newSViv (-( start [1] * 10 + start [2] - '0' * 11)); |
1180 | case 3: return newSViv (-(IV)( start [1] * 10 + start [2] - '0' * 11)); |
1061 | case 4: return newSViv (-( start [1] * 100 + start [2] * 10 + start [3] - '0' * 111)); |
1181 | case 4: return newSViv (-(IV)( start [1] * 100 + start [2] * 10 + start [3] - '0' * 111)); |
1062 | case 5: return newSViv (-( start [1] * 1000 + start [2] * 100 + start [3] * 10 + start [4] - '0' * 1111)); |
1182 | case 5: return newSViv (-(IV)( start [1] * 1000 + start [2] * 100 + start [3] * 10 + start [4] - '0' * 1111)); |
1063 | case 6: return newSViv (-(start [1] * 10000 + start [2] * 1000 + start [3] * 100 + start [4] * 10 + start [5] - '0' * 11111)); |
1183 | case 6: return newSViv (-(IV)(start [1] * 10000 + start [2] * 1000 + start [3] * 100 + start [4] * 10 + start [5] - '0' * 11111)); |
1064 | } |
1184 | } |
1065 | else |
1185 | else |
1066 | switch (len) |
1186 | switch (len) |
1067 | { |
1187 | { |
1068 | case 1: return newSViv ( start [0] - '0' * 1); |
1188 | case 1: return newSViv ( start [0] - '0' * 1); |
1069 | case 2: return newSViv ( start [0] * 10 + start [1] - '0' * 11); |
1189 | case 2: return newSViv ( start [0] * 10 + start [1] - '0' * 11); |
1070 | case 3: return newSViv ( start [0] * 100 + start [1] * 10 + start [2] - '0' * 111); |
1190 | case 3: return newSViv ( start [0] * 100 + start [1] * 10 + start [2] - '0' * 111); |
1071 | case 4: return newSViv ( start [0] * 1000 + start [1] * 100 + start [2] * 10 + start [3] - '0' * 1111); |
1191 | case 4: return newSViv ( start [0] * 1000 + start [1] * 100 + start [2] * 10 + start [3] - '0' * 1111); |
1072 | case 5: return newSViv ( start [0] * 10000 + start [1] * 1000 + start [2] * 100 + start [3] * 10 + start [4] - '0' * 11111); |
1192 | case 5: return newSViv ( start [0] * 10000 + start [1] * 1000 + start [2] * 100 + start [3] * 10 + start [4] - '0' * 11111); |
1073 | } |
1193 | } |
1074 | |
1194 | |
1075 | { |
1195 | { |
1076 | UV uv; |
1196 | UV uv; |
1077 | int numtype = grok_number (start, len, &uv); |
1197 | int numtype = grok_number (start, len, &uv); |
… | |
… | |
1086 | } |
1206 | } |
1087 | |
1207 | |
1088 | len -= *start == '-' ? 1 : 0; |
1208 | len -= *start == '-' ? 1 : 0; |
1089 | |
1209 | |
1090 | // does not fit into IV or UV, try NV |
1210 | // does not fit into IV or UV, try NV |
1091 | if ((sizeof (NV) == sizeof (double) && DBL_DIG >= len) |
1211 | if (len <= NV_DIG) |
1092 | #if defined (LDBL_DIG) |
|
|
1093 | || (sizeof (NV) == sizeof (long double) && LDBL_DIG >= len) |
|
|
1094 | #endif |
|
|
1095 | ) |
|
|
1096 | // fits into NV without loss of precision |
1212 | // fits into NV without loss of precision |
1097 | return newSVnv (Atof (start)); |
1213 | return newSVnv (json_atof (start)); |
1098 | |
1214 | |
1099 | // everything else fails, convert it to a string |
1215 | // everything else fails, convert it to a string |
1100 | return newSVpvn (start, dec->cur - start); |
1216 | return newSVpvn (start, dec->cur - start); |
1101 | } |
1217 | } |
1102 | |
1218 | |
1103 | // loss of precision here |
1219 | // loss of precision here |
1104 | return newSVnv (Atof (start)); |
1220 | return newSVnv (json_atof (start)); |
1105 | |
1221 | |
1106 | fail: |
1222 | fail: |
1107 | return 0; |
1223 | return 0; |
1108 | } |
1224 | } |
1109 | |
1225 | |
… | |
… | |
1343 | case 't': |
1459 | case 't': |
1344 | if (dec->end - dec->cur >= 4 && !memcmp (dec->cur, "true", 4)) |
1460 | if (dec->end - dec->cur >= 4 && !memcmp (dec->cur, "true", 4)) |
1345 | { |
1461 | { |
1346 | dec->cur += 4; |
1462 | dec->cur += 4; |
1347 | #if JSON_SLOW |
1463 | #if JSON_SLOW |
1348 | json_true = get_sv ("JSON::XS::true", 1); SvREADONLY_on (json_true); |
1464 | json_true = get_bool ("JSON::XS::true"); |
1349 | #endif |
1465 | #endif |
1350 | return SvREFCNT_inc (json_true); |
1466 | return newSVsv (json_true); |
1351 | } |
1467 | } |
1352 | else |
1468 | else |
1353 | ERR ("'true' expected"); |
1469 | ERR ("'true' expected"); |
1354 | |
1470 | |
1355 | break; |
1471 | break; |
… | |
… | |
1357 | case 'f': |
1473 | case 'f': |
1358 | if (dec->end - dec->cur >= 5 && !memcmp (dec->cur, "false", 5)) |
1474 | if (dec->end - dec->cur >= 5 && !memcmp (dec->cur, "false", 5)) |
1359 | { |
1475 | { |
1360 | dec->cur += 5; |
1476 | dec->cur += 5; |
1361 | #if JSON_SLOW |
1477 | #if JSON_SLOW |
1362 | json_false = get_sv ("JSON::XS::false", 1); SvREADONLY_on (json_false); |
1478 | json_false = get_bool ("JSON::XS::false"); |
1363 | #endif |
1479 | #endif |
1364 | return SvREFCNT_inc (json_false); |
1480 | return newSVsv (json_false); |
1365 | } |
1481 | } |
1366 | else |
1482 | else |
1367 | ERR ("'false' expected"); |
1483 | ERR ("'false' expected"); |
1368 | |
1484 | |
1369 | break; |
1485 | break; |
… | |
… | |
1387 | fail: |
1503 | fail: |
1388 | return 0; |
1504 | return 0; |
1389 | } |
1505 | } |
1390 | |
1506 | |
1391 | static SV * |
1507 | static SV * |
1392 | decode_json (SV *string, JSON *json, STRLEN *offset_return) |
1508 | decode_json (SV *string, JSON *json, char **offset_return) |
1393 | { |
1509 | { |
1394 | dec_t dec; |
1510 | dec_t dec; |
1395 | STRLEN offset; |
|
|
1396 | SV *sv; |
1511 | SV *sv; |
1397 | |
1512 | |
|
|
1513 | /* work around bugs in 5.10 where manipulating magic values |
|
|
1514 | * will perl ignore the magic in subsequent accesses |
|
|
1515 | */ |
1398 | SvGETMAGIC (string); |
1516 | /*SvGETMAGIC (string);*/ |
|
|
1517 | if (SvMAGICAL (string)) |
|
|
1518 | string = sv_2mortal (newSVsv (string)); |
|
|
1519 | |
1399 | SvUPGRADE (string, SVt_PV); |
1520 | SvUPGRADE (string, SVt_PV); |
1400 | |
1521 | |
1401 | if (json->flags & F_MAXSIZE && SvCUR (string) > DEC_SIZE (json->flags)) |
1522 | /* work around a bug in perl 5.10, which causes SvCUR to fail an |
|
|
1523 | * assertion with -DDEBUGGING, although SvCUR is documented to |
|
|
1524 | * return the xpv_cur field which certainly exists after upgrading. |
|
|
1525 | * according to nicholas clark, calling SvPOK fixes this. |
|
|
1526 | * But it doesn't fix it, so try another workaround, call SvPV_nolen |
|
|
1527 | * and hope for the best. |
|
|
1528 | * Damnit, SvPV_nolen still trips over yet another assertion. This |
|
|
1529 | * assertion business is seriously broken, try yet another workaround |
|
|
1530 | * for the broken -DDEBUGGING. |
|
|
1531 | */ |
|
|
1532 | { |
|
|
1533 | #ifdef DEBUGGING |
|
|
1534 | STRLEN offset = SvOK (string) ? sv_len (string) : 0; |
|
|
1535 | #else |
|
|
1536 | STRLEN offset = SvCUR (string); |
|
|
1537 | #endif |
|
|
1538 | |
|
|
1539 | if (offset > json->max_size && json->max_size) |
1402 | croak ("attempted decode of JSON text of %lu bytes size, but max_size is set to %lu", |
1540 | croak ("attempted decode of JSON text of %lu bytes size, but max_size is set to %lu", |
1403 | (unsigned long)SvCUR (string), (unsigned long)DEC_SIZE (json->flags)); |
1541 | (unsigned long)SvCUR (string), (unsigned long)json->max_size); |
|
|
1542 | } |
1404 | |
1543 | |
1405 | if (json->flags & F_UTF8) |
1544 | if (json->flags & F_UTF8) |
1406 | sv_utf8_downgrade (string, 0); |
1545 | sv_utf8_downgrade (string, 0); |
1407 | else |
1546 | else |
1408 | sv_utf8_upgrade (string); |
1547 | sv_utf8_upgrade (string); |
1409 | |
1548 | |
1410 | SvGROW (string, SvCUR (string) + 1); // should basically be a NOP |
1549 | SvGROW (string, SvCUR (string) + 1); // should basically be a NOP |
1411 | |
1550 | |
1412 | dec.json = *json; |
1551 | dec.json = *json; |
1413 | dec.cur = SvPVX (string); |
1552 | dec.cur = SvPVX (string); |
1414 | dec.end = SvEND (string); |
1553 | dec.end = SvEND (string); |
1415 | dec.err = 0; |
1554 | dec.err = 0; |
1416 | dec.depth = 0; |
1555 | dec.depth = 0; |
1417 | dec.maxdepth = DEC_DEPTH (dec.json.flags); |
|
|
1418 | |
1556 | |
1419 | if (dec.json.cb_object || dec.json.cb_sk_object) |
1557 | if (dec.json.cb_object || dec.json.cb_sk_object) |
1420 | dec.json.flags |= F_HOOK; |
1558 | dec.json.flags |= F_HOOK; |
1421 | |
1559 | |
1422 | *dec.end = 0; // this should basically be a nop, too, but make sure it's there |
1560 | *dec.end = 0; // this should basically be a nop, too, but make sure it's there |
1423 | |
1561 | |
1424 | decode_ws (&dec); |
1562 | decode_ws (&dec); |
1425 | sv = decode_sv (&dec); |
1563 | sv = decode_sv (&dec); |
|
|
1564 | |
|
|
1565 | if (offset_return) |
|
|
1566 | *offset_return = dec.cur; |
1426 | |
1567 | |
1427 | if (!(offset_return || !sv)) |
1568 | if (!(offset_return || !sv)) |
1428 | { |
1569 | { |
1429 | // check for trailing garbage |
1570 | // check for trailing garbage |
1430 | decode_ws (&dec); |
1571 | decode_ws (&dec); |
… | |
… | |
1433 | { |
1574 | { |
1434 | dec.err = "garbage after JSON object"; |
1575 | dec.err = "garbage after JSON object"; |
1435 | SvREFCNT_dec (sv); |
1576 | SvREFCNT_dec (sv); |
1436 | sv = 0; |
1577 | sv = 0; |
1437 | } |
1578 | } |
1438 | } |
|
|
1439 | |
|
|
1440 | if (offset_return || !sv) |
|
|
1441 | { |
|
|
1442 | offset = dec.json.flags & F_UTF8 |
|
|
1443 | ? dec.cur - SvPVX (string) |
|
|
1444 | : utf8_distance (dec.cur, SvPVX (string)); |
|
|
1445 | |
|
|
1446 | if (offset_return) |
|
|
1447 | *offset_return = offset; |
|
|
1448 | } |
1579 | } |
1449 | |
1580 | |
1450 | if (!sv) |
1581 | if (!sv) |
1451 | { |
1582 | { |
1452 | SV *uni = sv_newmortal (); |
1583 | SV *uni = sv_newmortal (); |
… | |
… | |
1458 | SAVEVPTR (PL_curcop); |
1589 | SAVEVPTR (PL_curcop); |
1459 | PL_curcop = &cop; |
1590 | PL_curcop = &cop; |
1460 | pv_uni_display (uni, dec.cur, dec.end - dec.cur, 20, UNI_DISPLAY_QQ); |
1591 | pv_uni_display (uni, dec.cur, dec.end - dec.cur, 20, UNI_DISPLAY_QQ); |
1461 | LEAVE; |
1592 | LEAVE; |
1462 | |
1593 | |
1463 | croak ("%s, at character offset %d [\"%s\"]", |
1594 | croak ("%s, at character offset %d (before \"%s\")", |
1464 | dec.err, |
1595 | dec.err, |
1465 | (int)offset, |
1596 | ptr_to_index (string, dec.cur), |
1466 | dec.cur != dec.end ? SvPV_nolen (uni) : "(end of string)"); |
1597 | dec.cur != dec.end ? SvPV_nolen (uni) : "(end of string)"); |
1467 | } |
1598 | } |
1468 | |
1599 | |
1469 | sv = sv_2mortal (sv); |
1600 | sv = sv_2mortal (sv); |
1470 | |
1601 | |
… | |
… | |
1479 | |
1610 | |
1480 | static void |
1611 | static void |
1481 | incr_parse (JSON *self) |
1612 | incr_parse (JSON *self) |
1482 | { |
1613 | { |
1483 | const char *p = SvPVX (self->incr_text) + self->incr_pos; |
1614 | const char *p = SvPVX (self->incr_text) + self->incr_pos; |
|
|
1615 | |
|
|
1616 | // the state machine here is a bit convoluted and could be simplified a lot |
|
|
1617 | // but this would make it slower, so... |
1484 | |
1618 | |
1485 | for (;;) |
1619 | for (;;) |
1486 | { |
1620 | { |
1487 | //printf ("loop pod %d *p<%c><%s>, mode %d nest %d\n", p - SvPVX (self->incr_text), *p, p, self->incr_mode, self->incr_nest);//D |
1621 | //printf ("loop pod %d *p<%c><%s>, mode %d nest %d\n", p - SvPVX (self->incr_text), *p, p, self->incr_mode, self->incr_nest);//D |
1488 | switch (self->incr_mode) |
1622 | switch (self->incr_mode) |
1489 | { |
1623 | { |
1490 | // only used for intiial whitespace skipping |
1624 | // only used for initial whitespace skipping |
1491 | case INCR_M_WS: |
1625 | case INCR_M_WS: |
1492 | for (;;) |
1626 | for (;;) |
1493 | { |
1627 | { |
1494 | if (*p > 0x20) |
1628 | if (*p > 0x20) |
1495 | { |
1629 | { |
|
|
1630 | if (*p == '#') |
|
|
1631 | { |
|
|
1632 | self->incr_mode = INCR_M_C0; |
|
|
1633 | goto incr_m_c; |
|
|
1634 | } |
|
|
1635 | else |
|
|
1636 | { |
1496 | self->incr_mode = INCR_M_JSON; |
1637 | self->incr_mode = INCR_M_JSON; |
1497 | goto incr_m_json; |
1638 | goto incr_m_json; |
|
|
1639 | } |
1498 | } |
1640 | } |
1499 | else if (!*p) |
1641 | else if (!*p) |
1500 | goto interrupt; |
1642 | goto interrupt; |
1501 | |
1643 | |
1502 | ++p; |
1644 | ++p; |
… | |
… | |
1508 | goto interrupt; |
1650 | goto interrupt; |
1509 | |
1651 | |
1510 | ++p; |
1652 | ++p; |
1511 | self->incr_mode = INCR_M_STR; |
1653 | self->incr_mode = INCR_M_STR; |
1512 | goto incr_m_str; |
1654 | goto incr_m_str; |
|
|
1655 | |
|
|
1656 | // inside #-style comments |
|
|
1657 | case INCR_M_C0: |
|
|
1658 | case INCR_M_C1: |
|
|
1659 | incr_m_c: |
|
|
1660 | for (;;) |
|
|
1661 | { |
|
|
1662 | if (*p == '\n') |
|
|
1663 | { |
|
|
1664 | self->incr_mode = self->incr_mode == INCR_M_C0 ? INCR_M_WS : INCR_M_JSON; |
|
|
1665 | break; |
|
|
1666 | } |
|
|
1667 | else if (!*p) |
|
|
1668 | goto interrupt; |
|
|
1669 | |
|
|
1670 | ++p; |
|
|
1671 | } |
|
|
1672 | |
|
|
1673 | break; |
1513 | |
1674 | |
1514 | // inside a string |
1675 | // inside a string |
1515 | case INCR_M_STR: |
1676 | case INCR_M_STR: |
1516 | incr_m_str: |
1677 | incr_m_str: |
1517 | for (;;) |
1678 | for (;;) |
… | |
… | |
1568 | self->incr_mode = INCR_M_STR; |
1729 | self->incr_mode = INCR_M_STR; |
1569 | goto incr_m_str; |
1730 | goto incr_m_str; |
1570 | |
1731 | |
1571 | case '[': |
1732 | case '[': |
1572 | case '{': |
1733 | case '{': |
1573 | ++self->incr_nest; |
1734 | if (++self->incr_nest > self->max_depth) |
|
|
1735 | croak (ERR_NESTING_EXCEEDED); |
1574 | break; |
1736 | break; |
1575 | |
1737 | |
1576 | case ']': |
1738 | case ']': |
1577 | case '}': |
1739 | case '}': |
1578 | if (!--self->incr_nest) |
1740 | if (--self->incr_nest <= 0) |
1579 | goto interrupt; |
1741 | goto interrupt; |
|
|
1742 | break; |
|
|
1743 | |
|
|
1744 | case '#': |
|
|
1745 | self->incr_mode = INCR_M_C1; |
|
|
1746 | goto incr_m_c; |
1580 | } |
1747 | } |
1581 | } |
1748 | } |
1582 | } |
1749 | } |
1583 | |
1750 | |
1584 | modechange: |
1751 | modechange: |
1585 | ; |
1752 | ; |
1586 | } |
1753 | } |
1587 | |
1754 | |
1588 | interrupt: |
1755 | interrupt: |
1589 | self->incr_pos = p - SvPVX (self->incr_text); |
1756 | self->incr_pos = p - SvPVX (self->incr_text); |
|
|
1757 | //printf ("interrupt<%.*s>\n", self->incr_pos, SvPVX(self->incr_text));//D |
1590 | //printf ("return pos %d mode %d nest %d\n", self->incr_pos, self->incr_mode, self->incr_nest);//D |
1758 | //printf ("return pos %d mode %d nest %d\n", self->incr_pos, self->incr_mode, self->incr_nest);//D |
1591 | } |
1759 | } |
1592 | |
1760 | |
1593 | ///////////////////////////////////////////////////////////////////////////// |
1761 | ///////////////////////////////////////////////////////////////////////////// |
1594 | // XS interface functions |
1762 | // XS interface functions |
… | |
… | |
1607 | : -1; |
1775 | : -1; |
1608 | |
1776 | |
1609 | json_stash = gv_stashpv ("JSON::XS" , 1); |
1777 | json_stash = gv_stashpv ("JSON::XS" , 1); |
1610 | json_boolean_stash = gv_stashpv ("JSON::XS::Boolean", 1); |
1778 | json_boolean_stash = gv_stashpv ("JSON::XS::Boolean", 1); |
1611 | |
1779 | |
1612 | json_true = get_sv ("JSON::XS::true" , 1); SvREADONLY_on (json_true ); |
1780 | json_true = get_bool ("JSON::XS::true"); |
1613 | json_false = get_sv ("JSON::XS::false", 1); SvREADONLY_on (json_false); |
1781 | json_false = get_bool ("JSON::XS::false"); |
|
|
1782 | |
|
|
1783 | CvNODEBUG_on (get_cv ("JSON::XS::incr_text", 0)); /* the debugger completely breaks lvalue subs */ |
1614 | } |
1784 | } |
1615 | |
1785 | |
1616 | PROTOTYPES: DISABLE |
1786 | PROTOTYPES: DISABLE |
1617 | |
1787 | |
1618 | void CLONE (...) |
1788 | void CLONE (...) |
… | |
… | |
1623 | void new (char *klass) |
1793 | void new (char *klass) |
1624 | PPCODE: |
1794 | PPCODE: |
1625 | { |
1795 | { |
1626 | SV *pv = NEWSV (0, sizeof (JSON)); |
1796 | SV *pv = NEWSV (0, sizeof (JSON)); |
1627 | SvPOK_only (pv); |
1797 | SvPOK_only (pv); |
1628 | Zero (SvPVX (pv), 1, JSON); |
1798 | json_init ((JSON *)SvPVX (pv)); |
1629 | ((JSON *)SvPVX (pv))->flags = F_DEFAULT; |
|
|
1630 | XPUSHs (sv_2mortal (sv_bless ( |
1799 | XPUSHs (sv_2mortal (sv_bless ( |
1631 | newRV_noinc (pv), |
1800 | newRV_noinc (pv), |
1632 | strEQ (klass, "JSON::XS") ? JSON_STASH : gv_stashpv (klass, 1) |
1801 | strEQ (klass, "JSON::XS") ? JSON_STASH : gv_stashpv (klass, 1) |
1633 | ))); |
1802 | ))); |
1634 | } |
1803 | } |
… | |
… | |
1646 | allow_nonref = F_ALLOW_NONREF |
1815 | allow_nonref = F_ALLOW_NONREF |
1647 | shrink = F_SHRINK |
1816 | shrink = F_SHRINK |
1648 | allow_blessed = F_ALLOW_BLESSED |
1817 | allow_blessed = F_ALLOW_BLESSED |
1649 | convert_blessed = F_CONV_BLESSED |
1818 | convert_blessed = F_CONV_BLESSED |
1650 | relaxed = F_RELAXED |
1819 | relaxed = F_RELAXED |
|
|
1820 | allow_unknown = F_ALLOW_UNKNOWN |
1651 | PPCODE: |
1821 | PPCODE: |
1652 | { |
1822 | { |
1653 | if (enable) |
1823 | if (enable) |
1654 | self->flags |= ix; |
1824 | self->flags |= ix; |
1655 | else |
1825 | else |
… | |
… | |
1670 | get_allow_nonref = F_ALLOW_NONREF |
1840 | get_allow_nonref = F_ALLOW_NONREF |
1671 | get_shrink = F_SHRINK |
1841 | get_shrink = F_SHRINK |
1672 | get_allow_blessed = F_ALLOW_BLESSED |
1842 | get_allow_blessed = F_ALLOW_BLESSED |
1673 | get_convert_blessed = F_CONV_BLESSED |
1843 | get_convert_blessed = F_CONV_BLESSED |
1674 | get_relaxed = F_RELAXED |
1844 | get_relaxed = F_RELAXED |
|
|
1845 | get_allow_unknown = F_ALLOW_UNKNOWN |
1675 | PPCODE: |
1846 | PPCODE: |
1676 | XPUSHs (boolSV (self->flags & ix)); |
1847 | XPUSHs (boolSV (self->flags & ix)); |
1677 | |
1848 | |
1678 | void max_depth (JSON *self, UV max_depth = 0x80000000UL) |
1849 | void max_depth (JSON *self, U32 max_depth = 0x80000000UL) |
1679 | PPCODE: |
1850 | PPCODE: |
1680 | { |
1851 | self->max_depth = max_depth; |
1681 | UV log2 = 0; |
|
|
1682 | |
|
|
1683 | if (max_depth > 0x80000000UL) max_depth = 0x80000000UL; |
|
|
1684 | |
|
|
1685 | while ((1UL << log2) < max_depth) |
|
|
1686 | ++log2; |
|
|
1687 | |
|
|
1688 | self->flags = self->flags & ~F_MAXDEPTH | (log2 << S_MAXDEPTH); |
|
|
1689 | |
|
|
1690 | XPUSHs (ST (0)); |
1852 | XPUSHs (ST (0)); |
1691 | } |
|
|
1692 | |
1853 | |
1693 | U32 get_max_depth (JSON *self) |
1854 | U32 get_max_depth (JSON *self) |
1694 | CODE: |
1855 | CODE: |
1695 | RETVAL = DEC_DEPTH (self->flags); |
1856 | RETVAL = self->max_depth; |
1696 | OUTPUT: |
1857 | OUTPUT: |
1697 | RETVAL |
1858 | RETVAL |
1698 | |
1859 | |
1699 | void max_size (JSON *self, UV max_size = 0) |
1860 | void max_size (JSON *self, U32 max_size = 0) |
1700 | PPCODE: |
1861 | PPCODE: |
1701 | { |
1862 | self->max_size = max_size; |
1702 | UV log2 = 0; |
|
|
1703 | |
|
|
1704 | if (max_size > 0x80000000UL) max_size = 0x80000000UL; |
|
|
1705 | if (max_size == 1) max_size = 2; |
|
|
1706 | |
|
|
1707 | while ((1UL << log2) < max_size) |
|
|
1708 | ++log2; |
|
|
1709 | |
|
|
1710 | self->flags = self->flags & ~F_MAXSIZE | (log2 << S_MAXSIZE); |
|
|
1711 | |
|
|
1712 | XPUSHs (ST (0)); |
1863 | XPUSHs (ST (0)); |
1713 | } |
|
|
1714 | |
1864 | |
1715 | int get_max_size (JSON *self) |
1865 | int get_max_size (JSON *self) |
1716 | CODE: |
1866 | CODE: |
1717 | RETVAL = DEC_SIZE (self->flags); |
1867 | RETVAL = self->max_size; |
1718 | OUTPUT: |
1868 | OUTPUT: |
1719 | RETVAL |
1869 | RETVAL |
1720 | |
1870 | |
1721 | void filter_json_object (JSON *self, SV *cb = &PL_sv_undef) |
1871 | void filter_json_object (JSON *self, SV *cb = &PL_sv_undef) |
1722 | PPCODE: |
1872 | PPCODE: |
… | |
… | |
1758 | XPUSHs (decode_json (jsonstr, self, 0)); |
1908 | XPUSHs (decode_json (jsonstr, self, 0)); |
1759 | |
1909 | |
1760 | void decode_prefix (JSON *self, SV *jsonstr) |
1910 | void decode_prefix (JSON *self, SV *jsonstr) |
1761 | PPCODE: |
1911 | PPCODE: |
1762 | { |
1912 | { |
1763 | STRLEN offset; |
1913 | char *offset; |
1764 | EXTEND (SP, 2); |
1914 | EXTEND (SP, 2); |
1765 | PUSHs (decode_json (jsonstr, self, &offset)); |
1915 | PUSHs (decode_json (jsonstr, self, &offset)); |
1766 | PUSHs (sv_2mortal (newSVuv (offset))); |
1916 | PUSHs (sv_2mortal (newSVuv (ptr_to_index (jsonstr, offset)))); |
1767 | } |
1917 | } |
1768 | |
1918 | |
1769 | void incr_parse (JSON *self, SV *jsonstr = 0) |
1919 | void incr_parse (JSON *self, SV *jsonstr = 0) |
1770 | PPCODE: |
1920 | PPCODE: |
1771 | { |
1921 | { |
… | |
… | |
1773 | self->incr_text = newSVpvn ("", 0); |
1923 | self->incr_text = newSVpvn ("", 0); |
1774 | |
1924 | |
1775 | // append data, if any |
1925 | // append data, if any |
1776 | if (jsonstr) |
1926 | if (jsonstr) |
1777 | { |
1927 | { |
1778 | if (SvUTF8 (jsonstr) && !SvUTF8 (self->incr_text)) |
1928 | if (SvUTF8 (jsonstr)) |
1779 | { |
1929 | { |
|
|
1930 | if (!SvUTF8 (self->incr_text)) |
|
|
1931 | { |
1780 | /* utf-8-ness differs, need to upgrade */ |
1932 | /* utf-8-ness differs, need to upgrade */ |
1781 | sv_utf8_upgrade (self->incr_text); |
1933 | sv_utf8_upgrade (self->incr_text); |
1782 | |
1934 | |
1783 | if (self->incr_pos) |
1935 | if (self->incr_pos) |
1784 | self->incr_pos = utf8_hop ((U8 *)SvPVX (self->incr_text), self->incr_pos) |
1936 | self->incr_pos = utf8_hop ((U8 *)SvPVX (self->incr_text), self->incr_pos) |
1785 | - (U8 *)SvPVX (self->incr_text); |
1937 | - (U8 *)SvPVX (self->incr_text); |
|
|
1938 | } |
1786 | } |
1939 | } |
|
|
1940 | else if (SvUTF8 (self->incr_text)) |
|
|
1941 | sv_utf8_upgrade (jsonstr); |
1787 | |
1942 | |
1788 | { |
1943 | { |
1789 | STRLEN len; |
1944 | STRLEN len; |
1790 | const char *str = SvPV (jsonstr, len); |
1945 | const char *str = SvPV (jsonstr, len); |
1791 | SvGROW (self->incr_text, SvCUR (self->incr_text) + len + 1); |
1946 | STRLEN cur = SvCUR (self->incr_text); |
|
|
1947 | |
|
|
1948 | if (SvLEN (self->incr_text) <= cur + len) |
|
|
1949 | SvGROW (self->incr_text, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
|
|
1950 | |
1792 | Move (str, SvEND (self->incr_text), len, char); |
1951 | Move (str, SvEND (self->incr_text), len, char); |
1793 | SvCUR_set (self->incr_text, SvCUR (self->incr_text) + len); |
1952 | SvCUR_set (self->incr_text, SvCUR (self->incr_text) + len); |
1794 | *SvEND (self->incr_text) = 0; // this should basically be a nop, too, but make sure it's there |
1953 | *SvEND (self->incr_text) = 0; // this should basically be a nop, too, but make sure it's there |
1795 | } |
1954 | } |
1796 | } |
1955 | } |
1797 | |
1956 | |
1798 | if (GIMME_V != G_VOID) |
1957 | if (GIMME_V != G_VOID) |
1799 | do |
1958 | do |
1800 | { |
1959 | { |
1801 | STRLEN offset; |
1960 | char *offset; |
1802 | |
|
|
1803 | incr_parse (self); |
|
|
1804 | |
1961 | |
1805 | if (!INCR_DONE (self)) |
1962 | if (!INCR_DONE (self)) |
|
|
1963 | { |
|
|
1964 | incr_parse (self); |
|
|
1965 | |
|
|
1966 | if (self->incr_pos > self->max_size && self->max_size) |
|
|
1967 | croak ("attempted decode of JSON text of %lu bytes size, but max_size is set to %lu", |
|
|
1968 | (unsigned long)self->incr_pos, (unsigned long)self->max_size); |
|
|
1969 | |
|
|
1970 | if (!INCR_DONE (self)) |
1806 | break; |
1971 | break; |
|
|
1972 | } |
1807 | |
1973 | |
1808 | XPUSHs (decode_json (self->incr_text, self, &offset)); |
1974 | XPUSHs (decode_json (self->incr_text, self, &offset)); |
1809 | |
1975 | |
1810 | sv_chop (self->incr_text, SvPV_nolen (self->incr_text) + offset); |
|
|
1811 | self->incr_pos -= offset; |
1976 | self->incr_pos -= offset - SvPVX (self->incr_text); |
1812 | self->incr_nest = 0; |
1977 | self->incr_nest = 0; |
1813 | self->incr_mode = 0; |
1978 | self->incr_mode = 0; |
|
|
1979 | |
|
|
1980 | sv_chop (self->incr_text, offset); |
1814 | } |
1981 | } |
1815 | while (GIMME_V == G_ARRAY); |
1982 | while (GIMME_V == G_ARRAY); |
1816 | } |
1983 | } |
1817 | |
1984 | |
1818 | SV *incr_text (JSON *self) |
1985 | SV *incr_text (JSON *self) |
1819 | ATTRS: lvalue |
1986 | ATTRS: lvalue |
1820 | CODE: |
1987 | CODE: |
1821 | { |
1988 | { |
1822 | if (self->incr_pos) |
1989 | if (self->incr_pos) |
1823 | croak ("incr_text can only be called after a successful incr_parse call in scalar context %d", self->incr_pos);//D |
1990 | croak ("incr_text can not be called when the incremental parser already started parsing"); |
1824 | |
1991 | |
1825 | RETVAL = self->incr_text ? SvREFCNT_inc (self->incr_text) : &PL_sv_undef; |
1992 | RETVAL = self->incr_text ? SvREFCNT_inc (self->incr_text) : &PL_sv_undef; |
1826 | } |
1993 | } |
1827 | OUTPUT: |
1994 | OUTPUT: |
1828 | RETVAL |
1995 | RETVAL |
|
|
1996 | |
|
|
1997 | void incr_skip (JSON *self) |
|
|
1998 | CODE: |
|
|
1999 | { |
|
|
2000 | if (self->incr_pos) |
|
|
2001 | { |
|
|
2002 | sv_chop (self->incr_text, SvPV_nolen (self->incr_text) + self->incr_pos); |
|
|
2003 | self->incr_pos = 0; |
|
|
2004 | self->incr_nest = 0; |
|
|
2005 | self->incr_mode = 0; |
|
|
2006 | } |
|
|
2007 | } |
|
|
2008 | |
|
|
2009 | void incr_reset (JSON *self) |
|
|
2010 | CODE: |
|
|
2011 | { |
|
|
2012 | SvREFCNT_dec (self->incr_text); |
|
|
2013 | self->incr_text = 0; |
|
|
2014 | self->incr_pos = 0; |
|
|
2015 | self->incr_nest = 0; |
|
|
2016 | self->incr_mode = 0; |
|
|
2017 | } |
1829 | |
2018 | |
1830 | void DESTROY (JSON *self) |
2019 | void DESTROY (JSON *self) |
1831 | CODE: |
2020 | CODE: |
1832 | SvREFCNT_dec (self->cb_sk_object); |
2021 | SvREFCNT_dec (self->cb_sk_object); |
1833 | SvREFCNT_dec (self->cb_object); |
2022 | SvREFCNT_dec (self->cb_object); |
… | |
… | |
1839 | ALIAS: |
2028 | ALIAS: |
1840 | to_json_ = 0 |
2029 | to_json_ = 0 |
1841 | encode_json = F_UTF8 |
2030 | encode_json = F_UTF8 |
1842 | PPCODE: |
2031 | PPCODE: |
1843 | { |
2032 | { |
1844 | JSON json = { F_DEFAULT | ix }; |
2033 | JSON json; |
|
|
2034 | json_init (&json); |
|
|
2035 | json.flags |= ix; |
1845 | XPUSHs (encode_json (scalar, &json)); |
2036 | XPUSHs (encode_json (scalar, &json)); |
1846 | } |
2037 | } |
1847 | |
2038 | |
1848 | void decode_json (SV *jsonstr) |
2039 | void decode_json (SV *jsonstr) |
1849 | ALIAS: |
2040 | ALIAS: |
1850 | from_json_ = 0 |
2041 | from_json_ = 0 |
1851 | decode_json = F_UTF8 |
2042 | decode_json = F_UTF8 |
1852 | PPCODE: |
2043 | PPCODE: |
1853 | { |
2044 | { |
1854 | JSON json = { F_DEFAULT | ix }; |
2045 | JSON json; |
|
|
2046 | json_init (&json); |
|
|
2047 | json.flags |= ix; |
1855 | XPUSHs (decode_json (jsonstr, &json, 0)); |
2048 | XPUSHs (decode_json (jsonstr, &json, 0)); |
1856 | } |
2049 | } |
1857 | |
2050 | |
1858 | |
|
|