pg 0.18.3 → 1.4.3

Sign up to get free protection for your applications and to get access to all the features.
Files changed (118) hide show
  1. checksums.yaml +5 -5
  2. checksums.yaml.gz.sig +0 -0
  3. data/.appveyor.yml +36 -0
  4. data/.gems +6 -0
  5. data/.github/workflows/binary-gems.yml +86 -0
  6. data/.github/workflows/source-gem.yml +129 -0
  7. data/.gitignore +13 -0
  8. data/.hgsigs +34 -0
  9. data/.hgtags +41 -0
  10. data/.irbrc +23 -0
  11. data/.pryrc +23 -0
  12. data/.tm_properties +21 -0
  13. data/.travis.yml +49 -0
  14. data/BSDL +2 -2
  15. data/Gemfile +14 -0
  16. data/History.rdoc +448 -4
  17. data/Manifest.txt +8 -21
  18. data/README-Windows.rdoc +4 -4
  19. data/README.ja.rdoc +1 -2
  20. data/README.rdoc +69 -17
  21. data/Rakefile +33 -135
  22. data/Rakefile.cross +70 -69
  23. data/certs/ged.pem +24 -0
  24. data/certs/larskanis-2022.pem +26 -0
  25. data/ext/errorcodes.def +109 -0
  26. data/ext/errorcodes.rb +1 -1
  27. data/ext/errorcodes.txt +35 -2
  28. data/ext/extconf.rb +120 -54
  29. data/ext/gvl_wrappers.c +8 -0
  30. data/ext/gvl_wrappers.h +44 -33
  31. data/ext/pg.c +213 -171
  32. data/ext/pg.h +92 -98
  33. data/ext/pg_binary_decoder.c +84 -15
  34. data/ext/pg_binary_encoder.c +24 -21
  35. data/ext/pg_coder.c +175 -39
  36. data/ext/pg_connection.c +1730 -1135
  37. data/ext/pg_copy_coder.c +94 -27
  38. data/ext/pg_record_coder.c +521 -0
  39. data/ext/pg_result.c +640 -221
  40. data/ext/pg_text_decoder.c +608 -40
  41. data/ext/pg_text_encoder.c +253 -99
  42. data/ext/pg_tuple.c +569 -0
  43. data/ext/pg_type_map.c +61 -21
  44. data/ext/pg_type_map_all_strings.c +19 -5
  45. data/ext/pg_type_map_by_class.c +54 -24
  46. data/ext/pg_type_map_by_column.c +79 -40
  47. data/ext/pg_type_map_by_mri_type.c +48 -19
  48. data/ext/pg_type_map_by_oid.c +55 -25
  49. data/ext/pg_type_map_in_ruby.c +51 -20
  50. data/ext/{util.c → pg_util.c} +12 -12
  51. data/ext/{util.h → pg_util.h} +2 -2
  52. data/lib/pg/basic_type_map_based_on_result.rb +47 -0
  53. data/lib/pg/basic_type_map_for_queries.rb +193 -0
  54. data/lib/pg/basic_type_map_for_results.rb +81 -0
  55. data/lib/pg/basic_type_registry.rb +301 -0
  56. data/lib/pg/binary_decoder.rb +23 -0
  57. data/lib/pg/coder.rb +24 -3
  58. data/lib/pg/connection.rb +723 -65
  59. data/lib/pg/constants.rb +2 -1
  60. data/lib/pg/exceptions.rb +9 -2
  61. data/lib/pg/result.rb +24 -7
  62. data/lib/pg/text_decoder.rb +24 -22
  63. data/lib/pg/text_encoder.rb +40 -8
  64. data/lib/pg/tuple.rb +30 -0
  65. data/lib/pg/type_map_by_column.rb +3 -2
  66. data/lib/pg/version.rb +4 -0
  67. data/lib/pg.rb +61 -36
  68. data/misc/openssl-pg-segfault.rb +31 -0
  69. data/misc/postgres/History.txt +9 -0
  70. data/misc/postgres/Manifest.txt +5 -0
  71. data/misc/postgres/README.txt +21 -0
  72. data/misc/postgres/Rakefile +21 -0
  73. data/misc/postgres/lib/postgres.rb +16 -0
  74. data/misc/ruby-pg/History.txt +9 -0
  75. data/misc/ruby-pg/Manifest.txt +5 -0
  76. data/misc/ruby-pg/README.txt +21 -0
  77. data/misc/ruby-pg/Rakefile +21 -0
  78. data/misc/ruby-pg/lib/ruby/pg.rb +16 -0
  79. data/pg.gemspec +32 -0
  80. data/rakelib/task_extension.rb +46 -0
  81. data/sample/array_insert.rb +1 -1
  82. data/sample/async_api.rb +4 -8
  83. data/sample/async_copyto.rb +1 -1
  84. data/sample/async_mixed.rb +1 -1
  85. data/sample/check_conn.rb +1 -1
  86. data/sample/copydata.rb +71 -0
  87. data/sample/copyfrom.rb +1 -1
  88. data/sample/copyto.rb +1 -1
  89. data/sample/cursor.rb +1 -1
  90. data/sample/disk_usage_report.rb +6 -15
  91. data/sample/issue-119.rb +2 -2
  92. data/sample/losample.rb +1 -1
  93. data/sample/minimal-testcase.rb +2 -2
  94. data/sample/notify_wait.rb +1 -1
  95. data/sample/pg_statistics.rb +6 -15
  96. data/sample/replication_monitor.rb +9 -18
  97. data/sample/test_binary_values.rb +1 -1
  98. data/sample/wal_shipper.rb +2 -2
  99. data/sample/warehouse_partitions.rb +8 -17
  100. data.tar.gz.sig +0 -0
  101. metadata +80 -230
  102. metadata.gz.sig +0 -0
  103. data/ChangeLog +0 -5804
  104. data/lib/pg/basic_type_mapping.rb +0 -399
  105. data/spec/data/expected_trace.out +0 -26
  106. data/spec/data/random_binary_data +0 -0
  107. data/spec/helpers.rb +0 -355
  108. data/spec/pg/basic_type_mapping_spec.rb +0 -251
  109. data/spec/pg/connection_spec.rb +0 -1538
  110. data/spec/pg/result_spec.rb +0 -449
  111. data/spec/pg/type_map_by_class_spec.rb +0 -138
  112. data/spec/pg/type_map_by_column_spec.rb +0 -222
  113. data/spec/pg/type_map_by_mri_type_spec.rb +0 -136
  114. data/spec/pg/type_map_by_oid_spec.rb +0 -149
  115. data/spec/pg/type_map_in_ruby_spec.rb +0 -164
  116. data/spec/pg/type_map_spec.rb +0 -22
  117. data/spec/pg/type_spec.rb +0 -690
  118. data/spec/pg_spec.rb +0 -50
@@ -1,6 +1,6 @@
1
1
  /*
2
2
  * pg_text_encoder.c - PG::TextEncoder module
3
- * $Id: pg_text_encoder.c,v b859963462b2 2015/03/13 17:39:35 lars $
3
+ * $Id$
4
4
  *
5
5
  */
6
6
 
@@ -28,6 +28,7 @@
28
28
  * intermediate - Pointer to a VALUE that might be set by the encoding function to some
29
29
  * value in the first call that can be retrieved later in the second call.
30
30
  * This VALUE is not yet initialized by the caller.
31
+ * enc_idx - Index of the output Encoding that strings should be converted to.
31
32
  *
32
33
  * Returns:
33
34
  * >= 0 - If out==NULL the encoder function must return the expected output buffer size.
@@ -40,15 +41,20 @@
40
41
 
41
42
 
42
43
  #include "pg.h"
43
- #include "util.h"
44
+ #include "pg_util.h"
45
+ #ifdef HAVE_INTTYPES_H
44
46
  #include <inttypes.h>
47
+ #endif
45
48
  #include <math.h>
46
49
 
47
50
  VALUE rb_mPG_TextEncoder;
48
51
  static ID s_id_encode;
49
52
  static ID s_id_to_i;
53
+ static ID s_id_to_s;
54
+ static ID s_cBigDecimal;
55
+ static VALUE s_str_F;
50
56
 
51
- static int pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate);
57
+ static int pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx);
52
58
 
53
59
  VALUE
54
60
  pg_obj_to_i( VALUE value )
@@ -74,7 +80,7 @@ pg_obj_to_i( VALUE value )
74
80
  *
75
81
  */
76
82
  static int
77
- pg_text_enc_boolean(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate)
83
+ pg_text_enc_boolean(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx)
78
84
  {
79
85
  switch( TYPE(value) ){
80
86
  case T_FALSE:
@@ -92,10 +98,10 @@ pg_text_enc_boolean(t_pg_coder *this, VALUE value, char *out, VALUE *intermediat
92
98
  if(out) *out = '1';
93
99
  return 1;
94
100
  } else {
95
- return pg_text_enc_integer(this, value, out, intermediate);
101
+ return pg_text_enc_integer(this, value, out, intermediate, enc_idx);
96
102
  }
97
103
  default:
98
- return pg_coder_enc_to_s(this, value, out, intermediate);
104
+ return pg_coder_enc_to_s(this, value, out, intermediate, enc_idx);
99
105
  }
100
106
  /* never reached */
101
107
  return 0;
@@ -111,45 +117,71 @@ pg_text_enc_boolean(t_pg_coder *this, VALUE value, char *out, VALUE *intermediat
111
117
  *
112
118
  */
113
119
  int
114
- pg_coder_enc_to_s(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate)
120
+ pg_coder_enc_to_s(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx)
115
121
  {
116
- *intermediate = rb_obj_as_string(value);
122
+ VALUE str = rb_obj_as_string(value);
123
+ if( ENCODING_GET(str) == enc_idx ){
124
+ *intermediate = str;
125
+ }else{
126
+ *intermediate = rb_str_export_to_enc(str, rb_enc_from_index(enc_idx));
127
+ }
117
128
  return -1;
118
129
  }
119
130
 
131
+ static int
132
+ count_leading_zero_bits(unsigned long long x)
133
+ {
134
+ #if defined(__GNUC__) || defined(__clang__)
135
+ return __builtin_clzll(x);
136
+ #elif defined(_MSC_VER)
137
+ DWORD r = 0;
138
+ _BitScanForward64(&r, x);
139
+ return (int)r;
140
+ #else
141
+ unsigned int a;
142
+ for(a=0; a < sizeof(unsigned long long) * 8; a++){
143
+ if( x & (1 << (sizeof(unsigned long long) * 8 - 1))) return a;
144
+ x <<= 1;
145
+ }
146
+ return a;
147
+ #endif
148
+ }
120
149
 
121
150
  /*
122
151
  * Document-class: PG::TextEncoder::Integer < PG::SimpleEncoder
123
152
  *
124
- * This is the encoder class for the PostgreSQL int types.
153
+ * This is the encoder class for the PostgreSQL integer types.
125
154
  *
126
155
  * Non-Integer values are expected to have method +to_i+ defined.
127
156
  *
128
157
  */
129
158
  static int
130
- pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate)
159
+ pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx)
131
160
  {
132
161
  if(out){
133
162
  if(TYPE(*intermediate) == T_STRING){
134
- return pg_coder_enc_to_s(this, value, out, intermediate);
163
+ return pg_coder_enc_to_s(this, value, out, intermediate, enc_idx);
135
164
  }else{
136
165
  char *start = out;
137
166
  int len;
138
167
  int neg = 0;
139
- long long ll = NUM2LL(*intermediate);
168
+ long long sll = NUM2LL(*intermediate);
169
+ unsigned long long ll;
140
170
 
141
- if (ll < 0) {
142
- /* We don't expect problems with the most negative integer not being representable
143
- * as a positive integer, because Fixnum is only up to 63 bits.
171
+ if (sll < 0) {
172
+ /* Avoid problems with the most negative integer not being representable
173
+ * as a positive integer, by using unsigned long long for encoding.
144
174
  */
145
- ll = -ll;
175
+ ll = -sll;
146
176
  neg = 1;
177
+ } else {
178
+ ll = sll;
147
179
  }
148
180
 
149
181
  /* Compute the result string backwards. */
150
182
  do {
151
- long long remainder;
152
- long long oldval = ll;
183
+ unsigned long long remainder;
184
+ unsigned long long oldval = ll;
153
185
 
154
186
  ll /= 10;
155
187
  remainder = oldval - ll * 10;
@@ -159,7 +191,7 @@ pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediat
159
191
  if (neg)
160
192
  *out++ = '-';
161
193
 
162
- len = out - start;
194
+ len = (int)(out - start);
163
195
 
164
196
  /* Reverse string. */
165
197
  out--;
@@ -176,45 +208,17 @@ pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediat
176
208
  }else{
177
209
  *intermediate = pg_obj_to_i(value);
178
210
  if(TYPE(*intermediate) == T_FIXNUM){
179
- int len;
180
211
  long long sll = NUM2LL(*intermediate);
181
- long long ll = sll < 0 ? -sll : sll;
182
- if( ll < 100000000 ){
183
- if( ll < 10000 ){
184
- if( ll < 100 ){
185
- len = ll < 10 ? 1 : 2;
186
- }else{
187
- len = ll < 1000 ? 3 : 4;
188
- }
189
- }else{
190
- if( ll < 1000000 ){
191
- len = ll < 100000 ? 5 : 6;
192
- }else{
193
- len = ll < 10000000 ? 7 : 8;
194
- }
195
- }
196
- }else{
197
- if( ll < 1000000000000LL ){
198
- if( ll < 10000000000LL ){
199
- len = ll < 1000000000LL ? 9 : 10;
200
- }else{
201
- len = ll < 100000000000LL ? 11 : 12;
202
- }
203
- }else{
204
- if( ll < 100000000000000LL ){
205
- len = ll < 10000000000000LL ? 13 : 14;
206
- }else{
207
- return pg_coder_enc_to_s(this, *intermediate, NULL, intermediate);
208
- }
209
- }
210
- }
211
- return sll < 0 ? len+1 : len;
212
+ unsigned long long ll = sll < 0 ? -sll : sll;
213
+ int len = (sizeof(unsigned long long) * 8 - count_leading_zero_bits(ll)) / 3;
214
+ return sll < 0 ? len+2 : len+1;
212
215
  }else{
213
- return pg_coder_enc_to_s(this, *intermediate, NULL, intermediate);
216
+ return pg_coder_enc_to_s(this, *intermediate, NULL, intermediate, enc_idx);
214
217
  }
215
218
  }
216
219
  }
217
220
 
221
+ #define MAX_DOUBLE_DIGITS 16
218
222
 
219
223
  /*
220
224
  * Document-class: PG::TextEncoder::Float < PG::SimpleEncoder
@@ -223,10 +227,16 @@ pg_text_enc_integer(t_pg_coder *this, VALUE value, char *out, VALUE *intermediat
223
227
  *
224
228
  */
225
229
  static int
226
- pg_text_enc_float(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
230
+ pg_text_enc_float(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
227
231
  {
228
232
  if(out){
229
233
  double dvalue = NUM2DBL(value);
234
+ int len = 0;
235
+ int neg = 0;
236
+ int exp2i, exp10i, i;
237
+ unsigned long long ll, remainder, oldval;
238
+ VALUE intermediate;
239
+
230
240
  /* Cast to the same strings as value.to_s . */
231
241
  if( isinf(dvalue) ){
232
242
  if( dvalue < 0 ){
@@ -240,12 +250,128 @@ pg_text_enc_float(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
240
250
  memcpy( out, "NaN", 3);
241
251
  return 3;
242
252
  }
243
- return sprintf( out, "%.16E", dvalue);
253
+
254
+ /*
255
+ * The following computation is roughly a conversion kind of
256
+ * sprintf( out, "%.16E", dvalue);
257
+ */
258
+
259
+ /* write the algebraic sign */
260
+ if( dvalue < 0 ) {
261
+ dvalue = -dvalue;
262
+ *out++ = '-';
263
+ neg++;
264
+ }
265
+
266
+ /* retrieve the power of 2 exponent */
267
+ frexp(dvalue, &exp2i);
268
+ /* compute the power of 10 exponent */
269
+ exp10i = (int)floor(exp2i * 0.30102999566398114); /* Math.log(2)/Math.log(10) */
270
+ /* move the decimal point, so that we get an integer of MAX_DOUBLE_DIGITS decimal digits */
271
+ ll = (unsigned long long)(dvalue * pow(10, MAX_DOUBLE_DIGITS - 1 - exp10i) + 0.5);
272
+
273
+ /* avoid leading zeros due to inaccuracy of deriving exp10i from exp2i */
274
+ /* otherwise we would print "09.0" instead of "9.0" */
275
+ if( ll < 1000000000000000 ){ /* pow(10, MAX_DOUBLE_DIGITS-1) */
276
+ exp10i--;
277
+ ll *= 10;
278
+ }
279
+
280
+ if( exp10i <= -5 || exp10i >= 15 ) {
281
+ /* Write the float in exponent format (1.23e45) */
282
+
283
+ /* write fraction digits from right to left */
284
+ for( i = MAX_DOUBLE_DIGITS; i > 1; i--){
285
+ oldval = ll;
286
+ ll /= 10;
287
+ remainder = oldval - ll * 10;
288
+ /* omit trailing zeros */
289
+ if(remainder != 0 || len ) {
290
+ out[i] = '0' + remainder;
291
+ len++;
292
+ }
293
+ }
294
+
295
+ /* write decimal point */
296
+ if( len ){
297
+ out[1] = '.';
298
+ len++;
299
+ }
300
+
301
+ /* write remaining single digit left to the decimal point */
302
+ oldval = ll;
303
+ ll /= 10;
304
+ remainder = oldval - ll * 10;
305
+ out[0] = '0' + remainder;
306
+ len++;
307
+
308
+ /* write exponent */
309
+ out[len++] = 'e';
310
+ intermediate = INT2NUM(exp10i);
311
+
312
+ return neg + len + pg_text_enc_integer(conv, Qnil, out + len, &intermediate, enc_idx);
313
+ } else {
314
+ /* write the float in non exponent format (0.001234 or 123450.0) */
315
+
316
+ /* write digits from right to left */
317
+ int lz = exp10i < 0 ? 0 : exp10i;
318
+ for( i = MAX_DOUBLE_DIGITS - (exp10i < 0 ? exp10i : 0); i >= 0; i-- ){
319
+ oldval = ll;
320
+ ll /= 10;
321
+ remainder = oldval - ll * 10;
322
+ /* write decimal point */
323
+ if( i - 1 == lz ){
324
+ out[i--] = '.';
325
+ len++;
326
+ }
327
+ /* if possible then omit trailing zeros */
328
+ if(remainder != 0 || len || i - 2 == lz) {
329
+ out[i] = '0' + remainder;
330
+ len++;
331
+ }
332
+ }
333
+ return neg + len;
334
+ }
244
335
  }else{
245
- return 23;
336
+ return 1 /*sign*/ + MAX_DOUBLE_DIGITS + 1 /*dot*/ + 1 /*e*/ + 1 /*exp sign*/ + 3 /*exp digits*/;
246
337
  }
247
338
  }
248
339
 
340
+
341
+ /*
342
+ * Document-class: PG::TextEncoder::Numeric < PG::SimpleEncoder
343
+ *
344
+ * This is the encoder class for the PostgreSQL numeric types.
345
+ *
346
+ * It converts Integer, Float and BigDecimal objects.
347
+ * All other objects are expected to respond to +to_s+.
348
+ */
349
+ static int
350
+ pg_text_enc_numeric(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx)
351
+ {
352
+ switch(TYPE(value)){
353
+ case T_FIXNUM:
354
+ case T_BIGNUM:
355
+ return pg_text_enc_integer(this, value, out, intermediate, enc_idx);
356
+ case T_FLOAT:
357
+ return pg_text_enc_float(this, value, out, intermediate, enc_idx);
358
+ default:
359
+ if(out){ /* second pass */
360
+ rb_bug("unexpected value type: %d", TYPE(value));
361
+ } else { /* first pass */
362
+ if( rb_obj_is_kind_of(value, s_cBigDecimal) ){
363
+ /* value.to_s('F') */
364
+ *intermediate = rb_funcall(value, s_id_to_s, 1, s_str_F);
365
+ return -1; /* no second pass */
366
+ } else {
367
+ return pg_coder_enc_to_s(this, value, NULL, intermediate, enc_idx);
368
+ /* no second pass */
369
+ }
370
+ }
371
+ }
372
+ }
373
+
374
+
249
375
  static const char hextab[] = {
250
376
  '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f'
251
377
  };
@@ -253,8 +379,7 @@ static const char hextab[] = {
253
379
  /*
254
380
  * Document-class: PG::TextEncoder::Bytea < PG::SimpleEncoder
255
381
  *
256
- * This is an encoder class for the PostgreSQL bytea type for server version 9.0
257
- * or newer.
382
+ * This is an encoder class for the PostgreSQL +bytea+ type.
258
383
  *
259
384
  * The binary String is converted to hexadecimal representation for transmission
260
385
  * in text format. For query bind parameters it is recommended to use
@@ -263,7 +388,7 @@ static const char hextab[] = {
263
388
  *
264
389
  */
265
390
  static int
266
- pg_text_enc_bytea(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
391
+ pg_text_enc_bytea(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
267
392
  {
268
393
  if(out){
269
394
  size_t strlen = RSTRING_LEN(*intermediate);
@@ -278,11 +403,11 @@ pg_text_enc_bytea(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
278
403
  *optr++ = hextab[c >> 4];
279
404
  *optr++ = hextab[c & 0xf];
280
405
  }
281
- return optr - out;
406
+ return (int)(optr - out);
282
407
  }else{
283
408
  *intermediate = rb_obj_as_string(value);
284
409
  /* The output starts with "\x" and each character is converted to hex. */
285
- return 2 + RSTRING_LEN(*intermediate) * 2;
410
+ return 2 + RSTRING_LENINT(*intermediate) * 2;
286
411
  }
287
412
  }
288
413
 
@@ -342,13 +467,13 @@ quote_array_buffer( void *_this, char *p_in, int strlen, char *p_out ){
342
467
  }
343
468
 
344
469
  static char *
345
- quote_string(t_pg_coder *this, VALUE value, VALUE string, char *current_out, int with_quote, t_quote_func quote_buffer, void *func_data)
470
+ quote_string(t_pg_coder *this, VALUE value, VALUE string, char *current_out, int with_quote, t_quote_func quote_buffer, void *func_data, int enc_idx)
346
471
  {
347
472
  int strlen;
348
473
  VALUE subint;
349
474
  t_pg_coder_enc_func enc_func = pg_coder_enc_func(this);
350
475
 
351
- strlen = enc_func(this, value, NULL, &subint);
476
+ strlen = enc_func(this, value, NULL, &subint, enc_idx);
352
477
 
353
478
  if( strlen == -1 ){
354
479
  /* we can directly use String value in subint */
@@ -374,20 +499,20 @@ quote_string(t_pg_coder *this, VALUE value, VALUE string, char *current_out, int
374
499
  current_out = pg_rb_str_ensure_capa( string, 2 * strlen + 2, current_out, NULL );
375
500
 
376
501
  /* Place the unescaped string at current output position. */
377
- strlen = enc_func(this, value, current_out, &subint);
502
+ strlen = enc_func(this, value, current_out, &subint, enc_idx);
378
503
 
379
504
  current_out += quote_buffer( func_data, current_out, strlen, current_out );
380
505
  }else{
381
506
  /* size of the unquoted string */
382
507
  current_out = pg_rb_str_ensure_capa( string, strlen, current_out, NULL );
383
- current_out += enc_func(this, value, current_out, &subint);
508
+ current_out += enc_func(this, value, current_out, &subint, enc_idx);
384
509
  }
385
510
  }
386
511
  return current_out;
387
512
  }
388
513
 
389
514
  static char *
390
- write_array(t_pg_composite_coder *this, VALUE value, char *current_out, VALUE string, int quote)
515
+ write_array(t_pg_composite_coder *this, VALUE value, char *current_out, VALUE string, int quote, int enc_idx)
391
516
  {
392
517
  int i;
393
518
 
@@ -405,7 +530,7 @@ write_array(t_pg_composite_coder *this, VALUE value, char *current_out, VALUE st
405
530
 
406
531
  switch(TYPE(entry)){
407
532
  case T_ARRAY:
408
- current_out = write_array(this, entry, current_out, string, quote);
533
+ current_out = write_array(this, entry, current_out, string, quote, enc_idx);
409
534
  break;
410
535
  case T_NIL:
411
536
  current_out = pg_rb_str_ensure_capa( string, 4, current_out, NULL );
@@ -415,7 +540,7 @@ write_array(t_pg_composite_coder *this, VALUE value, char *current_out, VALUE st
415
540
  *current_out++ = 'L';
416
541
  break;
417
542
  default:
418
- current_out = quote_string( this->elem, entry, string, current_out, quote, quote_array_buffer, this );
543
+ current_out = quote_string( this->elem, entry, string, current_out, quote, quote_array_buffer, this, enc_idx );
419
544
  }
420
545
  }
421
546
  current_out = pg_rb_str_ensure_capa( string, 1, current_out, NULL );
@@ -437,41 +562,42 @@ write_array(t_pg_composite_coder *this, VALUE value, char *current_out, VALUE st
437
562
  *
438
563
  */
439
564
  static int
440
- pg_text_enc_array(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
565
+ pg_text_enc_array(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
441
566
  {
442
567
  char *end_ptr;
443
568
  t_pg_composite_coder *this = (t_pg_composite_coder *)conv;
444
569
 
445
570
  if( TYPE(value) == T_ARRAY){
446
- *intermediate = rb_str_new(NULL, 0);
571
+ VALUE out_str = rb_str_new(NULL, 0);
572
+ PG_ENCODING_SET_NOCHECK(out_str, enc_idx);
447
573
 
448
- end_ptr = write_array(this, value, RSTRING_PTR(*intermediate), *intermediate, this->needs_quotation);
574
+ end_ptr = write_array(this, value, RSTRING_PTR(out_str), out_str, this->needs_quotation, enc_idx);
449
575
 
450
- rb_str_set_len( *intermediate, end_ptr - RSTRING_PTR(*intermediate) );
576
+ rb_str_set_len( out_str, end_ptr - RSTRING_PTR(out_str) );
577
+ *intermediate = out_str;
451
578
 
452
579
  return -1;
453
580
  } else {
454
- return pg_coder_enc_to_s( conv, value, out, intermediate );
581
+ return pg_coder_enc_to_s( conv, value, out, intermediate, enc_idx );
455
582
  }
456
583
  }
457
584
 
458
585
  static char *
459
586
  quote_identifier( VALUE value, VALUE out_string, char *current_out ){
460
587
  char *p_in = RSTRING_PTR(value);
461
- char *ptr1;
462
588
  size_t strlen = RSTRING_LEN(value);
589
+ char *p_inend = p_in + strlen;
463
590
  char *end_capa = current_out;
464
591
 
465
592
  PG_RB_STR_ENSURE_CAPA( out_string, strlen + 2, current_out, end_capa );
466
593
  *current_out++ = '"';
467
- for(ptr1 = p_in; ptr1 != p_in + strlen; ptr1++) {
468
- char c = *ptr1;
594
+ for(; p_in != p_inend; p_in++) {
595
+ char c = *p_in;
469
596
  if (c == '"'){
470
- strlen++;
471
- PG_RB_STR_ENSURE_CAPA( out_string, p_in - ptr1 + strlen + 1, current_out, end_capa );
597
+ PG_RB_STR_ENSURE_CAPA( out_string, p_inend - p_in + 2, current_out, end_capa );
472
598
  *current_out++ = '"';
473
599
  } else if (c == 0){
474
- break;
600
+ rb_raise(rb_eArgError, "string contains null byte");
475
601
  }
476
602
  *current_out++ = c;
477
603
  }
@@ -482,10 +608,10 @@ quote_identifier( VALUE value, VALUE out_string, char *current_out ){
482
608
  }
483
609
 
484
610
  static char *
485
- pg_text_enc_array_identifier(VALUE value, VALUE string, char *out)
611
+ pg_text_enc_array_identifier(VALUE value, VALUE string, char *out, int enc_idx)
486
612
  {
487
- int i;
488
- int nr_elems;
613
+ long i;
614
+ long nr_elems;
489
615
 
490
616
  Check_Type(value, T_ARRAY);
491
617
  nr_elems = RARRAY_LEN(value);
@@ -493,6 +619,10 @@ pg_text_enc_array_identifier(VALUE value, VALUE string, char *out)
493
619
  for( i=0; i<nr_elems; i++){
494
620
  VALUE entry = rb_ary_entry(value, i);
495
621
 
622
+ StringValue(entry);
623
+ if( ENCODING_GET(entry) != enc_idx ){
624
+ entry = rb_str_export_to_enc(entry, rb_enc_from_index(enc_idx));
625
+ }
496
626
  out = quote_identifier(entry, string, out);
497
627
  if( i < nr_elems-1 ){
498
628
  out = pg_rb_str_ensure_capa( string, 1, out, NULL );
@@ -507,27 +637,34 @@ pg_text_enc_array_identifier(VALUE value, VALUE string, char *out)
507
637
  *
508
638
  * This is the encoder class for PostgreSQL identifiers.
509
639
  *
510
- * An Array value can be used for "schema.table.column" type identifiers:
640
+ * An Array value can be used for identifiers of the kind "schema.table.column".
641
+ * This ensures that each element is properly quoted:
511
642
  * PG::TextEncoder::Identifier.new.encode(['schema', 'table', 'column'])
512
643
  * => '"schema"."table"."column"'
513
644
  *
514
645
  * This encoder can also be used per PG::Connection#quote_ident .
515
646
  */
516
647
  int
517
- pg_text_enc_identifier(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate)
648
+ pg_text_enc_identifier(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx)
518
649
  {
650
+ VALUE out_str;
519
651
  UNUSED( this );
520
652
  if( TYPE(value) == T_ARRAY){
521
- *intermediate = rb_str_new(NULL, 0);
522
- out = RSTRING_PTR(*intermediate);
523
- out = pg_text_enc_array_identifier(value, *intermediate, out);
653
+ out_str = rb_str_new(NULL, 0);
654
+ out = RSTRING_PTR(out_str);
655
+ out = pg_text_enc_array_identifier(value, out_str, out, enc_idx);
524
656
  } else {
525
657
  StringValue(value);
526
- *intermediate = rb_str_new(NULL, RSTRING_LEN(value) + 2);
527
- out = RSTRING_PTR(*intermediate);
528
- out = quote_identifier(value, *intermediate, out);
658
+ if( ENCODING_GET(value) != enc_idx ){
659
+ value = rb_str_export_to_enc(value, rb_enc_from_index(enc_idx));
660
+ }
661
+ out_str = rb_str_new(NULL, RSTRING_LEN(value) + 2);
662
+ out = RSTRING_PTR(out_str);
663
+ out = quote_identifier(value, out_str, out);
529
664
  }
530
- rb_str_set_len( *intermediate, out - RSTRING_PTR(*intermediate) );
665
+ rb_str_set_len( out_str, out - RSTRING_PTR(out_str) );
666
+ PG_ENCODING_SET_NOCHECK(out_str, enc_idx);
667
+ *intermediate = out_str;
531
668
  return -1;
532
669
  }
533
670
 
@@ -569,18 +706,26 @@ quote_literal_buffer( void *_this, char *p_in, int strlen, char *p_out ){
569
706
  *
570
707
  * This is the encoder class for PostgreSQL literals.
571
708
  *
572
- * A literal is quoted and escaped by the +'+ character.
709
+ * A literal is quoted and escaped by the <tt>'</tt> character, so that it can be inserted into SQL queries.
710
+ * It works equal to PG::Connection#escape_literal, but integrates into the type cast system of ruby-pg.
711
+ *
712
+ * Both expressions have the same result:
713
+ * conn.escape_literal(PG::TextEncoder::Array.new.encode(["v1","v2"])) # => "'{v1,v2}'"
714
+ * PG::TextEncoder::QuotedLiteral.new(elements_type: PG::TextEncoder::Array.new).encode(["v1","v2"]) # => "'{v1,v2}'"
715
+ * While escape_literal requires a intermediate ruby string allocation, QuotedLiteral encodes the values directly to the result string.
573
716
  *
574
717
  */
575
718
  static int
576
- pg_text_enc_quoted_literal(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
719
+ pg_text_enc_quoted_literal(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
577
720
  {
578
721
  t_pg_composite_coder *this = (t_pg_composite_coder *)conv;
722
+ VALUE out_str = rb_str_new(NULL, 0);
723
+ PG_ENCODING_SET_NOCHECK(out_str, enc_idx);
579
724
 
580
- *intermediate = rb_str_new(NULL, 0);
581
- out = RSTRING_PTR(*intermediate);
582
- out = quote_string(this->elem, value, *intermediate, out, this->needs_quotation, quote_literal_buffer, this);
583
- rb_str_set_len( *intermediate, out - RSTRING_PTR(*intermediate) );
725
+ out = RSTRING_PTR(out_str);
726
+ out = quote_string(this->elem, value, out_str, out, this->needs_quotation, quote_literal_buffer, this, enc_idx);
727
+ rb_str_set_len( out_str, out - RSTRING_PTR(out_str) );
728
+ *intermediate = out_str;
584
729
  return -1;
585
730
  }
586
731
 
@@ -591,7 +736,7 @@ pg_text_enc_quoted_literal(t_pg_coder *conv, VALUE value, char *out, VALUE *inte
591
736
  *
592
737
  */
593
738
  static int
594
- pg_text_enc_to_base64(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate)
739
+ pg_text_enc_to_base64(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
595
740
  {
596
741
  int strlen;
597
742
  VALUE subint;
@@ -600,13 +745,13 @@ pg_text_enc_to_base64(t_pg_coder *conv, VALUE value, char *out, VALUE *intermedi
600
745
 
601
746
  if(out){
602
747
  /* Second encoder pass, if required */
603
- strlen = enc_func(this->elem, value, out, intermediate);
748
+ strlen = enc_func(this->elem, value, out, intermediate, enc_idx);
604
749
  base64_encode( out, out, strlen );
605
750
 
606
751
  return BASE64_ENCODED_SIZE(strlen);
607
752
  } else {
608
753
  /* First encoder pass */
609
- strlen = enc_func(this->elem, value, NULL, &subint);
754
+ strlen = enc_func(this->elem, value, NULL, &subint, enc_idx);
610
755
 
611
756
  if( strlen == -1 ){
612
757
  /* Encoded string is returned in subint */
@@ -614,6 +759,7 @@ pg_text_enc_to_base64(t_pg_coder *conv, VALUE value, char *out, VALUE *intermedi
614
759
 
615
760
  strlen = RSTRING_LENINT(subint);
616
761
  out_str = rb_str_new(NULL, BASE64_ENCODED_SIZE(strlen));
762
+ PG_ENCODING_SET_NOCHECK(out_str, enc_idx);
617
763
 
618
764
  base64_encode( RSTRING_PTR(out_str), RSTRING_PTR(subint), strlen);
619
765
  *intermediate = out_str;
@@ -633,6 +779,12 @@ init_pg_text_encoder()
633
779
  {
634
780
  s_id_encode = rb_intern("encode");
635
781
  s_id_to_i = rb_intern("to_i");
782
+ s_id_to_s = rb_intern("to_s");
783
+ s_str_F = rb_str_freeze(rb_str_new_cstr("F"));
784
+ rb_global_variable(&s_str_F);
785
+ rb_require("bigdecimal");
786
+ s_cBigDecimal = rb_const_get(rb_cObject, rb_intern("BigDecimal"));
787
+
636
788
 
637
789
  /* This module encapsulates all encoder classes with text output format */
638
790
  rb_mPG_TextEncoder = rb_define_module_under( rb_mPG, "TextEncoder" );
@@ -644,6 +796,8 @@ init_pg_text_encoder()
644
796
  pg_define_coder( "Integer", pg_text_enc_integer, rb_cPG_SimpleEncoder, rb_mPG_TextEncoder );
645
797
  /* dummy = rb_define_class_under( rb_mPG_TextEncoder, "Float", rb_cPG_SimpleEncoder ); */
646
798
  pg_define_coder( "Float", pg_text_enc_float, rb_cPG_SimpleEncoder, rb_mPG_TextEncoder );
799
+ /* dummy = rb_define_class_under( rb_mPG_TextEncoder, "Numeric", rb_cPG_SimpleEncoder ); */
800
+ pg_define_coder( "Numeric", pg_text_enc_numeric, rb_cPG_SimpleEncoder, rb_mPG_TextEncoder );
647
801
  /* dummy = rb_define_class_under( rb_mPG_TextEncoder, "String", rb_cPG_SimpleEncoder ); */
648
802
  pg_define_coder( "String", pg_coder_enc_to_s, rb_cPG_SimpleEncoder, rb_mPG_TextEncoder );
649
803
  /* dummy = rb_define_class_under( rb_mPG_TextEncoder, "Bytea", rb_cPG_SimpleEncoder ); */