yarp 0.9.0 → 0.11.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (62) hide show
  1. checksums.yaml +4 -4
  2. data/CHANGELOG.md +39 -1
  3. data/CONTRIBUTING.md +7 -0
  4. data/Makefile +5 -1
  5. data/config.yml +308 -166
  6. data/docs/configuration.md +0 -1
  7. data/docs/encoding.md +5 -5
  8. data/docs/mapping.md +91 -91
  9. data/docs/serialization.md +25 -22
  10. data/ext/yarp/api_node.c +1210 -483
  11. data/ext/yarp/extension.c +22 -8
  12. data/ext/yarp/extension.h +2 -2
  13. data/include/yarp/ast.h +692 -183
  14. data/include/yarp/defines.h +2 -1
  15. data/include/yarp/diagnostic.h +200 -3
  16. data/include/yarp/enc/yp_encoding.h +10 -10
  17. data/include/yarp/node.h +0 -4
  18. data/include/yarp/parser.h +19 -19
  19. data/include/yarp/regexp.h +1 -1
  20. data/include/yarp/unescape.h +4 -4
  21. data/include/yarp/util/yp_buffer.h +3 -0
  22. data/include/yarp/util/yp_char.h +16 -16
  23. data/include/yarp/util/yp_constant_pool.h +12 -5
  24. data/include/yarp/util/yp_newline_list.h +5 -5
  25. data/include/yarp/util/yp_string.h +4 -4
  26. data/include/yarp/util/yp_string_list.h +0 -3
  27. data/include/yarp/util/yp_strpbrk.h +1 -1
  28. data/include/yarp/version.h +2 -2
  29. data/include/yarp.h +5 -4
  30. data/lib/yarp/desugar_visitor.rb +59 -122
  31. data/lib/yarp/mutation_visitor.rb +22 -12
  32. data/lib/yarp/node.rb +3081 -501
  33. data/lib/yarp/parse_result/comments.rb +172 -0
  34. data/lib/yarp/parse_result/newlines.rb +60 -0
  35. data/lib/yarp/pattern.rb +239 -0
  36. data/lib/yarp/serialize.rb +152 -129
  37. data/lib/yarp.rb +109 -49
  38. data/src/diagnostic.c +254 -2
  39. data/src/enc/yp_big5.c +15 -42
  40. data/src/enc/yp_euc_jp.c +16 -43
  41. data/src/enc/yp_gbk.c +19 -46
  42. data/src/enc/yp_shift_jis.c +16 -43
  43. data/src/enc/yp_tables.c +36 -38
  44. data/src/enc/yp_unicode.c +20 -25
  45. data/src/enc/yp_windows_31j.c +16 -43
  46. data/src/node.c +1871 -1466
  47. data/src/prettyprint.c +463 -230
  48. data/src/regexp.c +21 -21
  49. data/src/serialize.c +352 -184
  50. data/src/unescape.c +152 -122
  51. data/src/util/yp_buffer.c +7 -2
  52. data/src/util/yp_char.c +35 -40
  53. data/src/util/yp_constant_pool.c +45 -12
  54. data/src/util/yp_memchr.c +1 -1
  55. data/src/util/yp_newline_list.c +10 -5
  56. data/src/util/yp_string.c +22 -20
  57. data/src/util/yp_string_list.c +4 -7
  58. data/src/util/yp_strncasecmp.c +3 -6
  59. data/src/util/yp_strpbrk.c +8 -8
  60. data/src/yarp.c +1288 -1021
  61. data/yarp.gemspec +4 -1
  62. metadata +6 -3
data/src/enc/yp_windows_31j.c CHANGED
@@ -1,73 +1,46 @@
1
1
  #include "yarp/enc/yp_encoding.h"
2
2
 
3
- typedef uint16_t yp_windows_31j_codepoint_t;
4
-
5
- static yp_windows_31j_codepoint_t
6
- yp_windows_31j_codepoint(const char *c, ptrdiff_t n, size_t *width) {
7
- const unsigned char *uc = (const unsigned char *) c;
8
-
3
+ static size_t
4
+ yp_encoding_windows_31j_char_width(const uint8_t *b, ptrdiff_t n) {
9
5
  // These are the single byte characters.
10
- if (*uc < 0x80 || (*uc >= 0xA1 && *uc <= 0xDF)) {
11
- *width = 1;
12
- return *uc;
6
+ if (*b < 0x80 || (*b >= 0xA1 && *b <= 0xDF)) {
7
+ return 1;
13
8
  }
14
9
 
15
10
  // These are the double byte characters.
16
11
  if (
17
12
  (n > 1) &&
18
- ((uc[0] >= 0x81 && uc[0] <= 0x9F) || (uc[0] >= 0xE0 && uc[0] <= 0xFC)) &&
19
- (uc[1] >= 0x40 && uc[1] <= 0xFC)
13
+ ((b[0] >= 0x81 && b[0] <= 0x9F) || (b[0] >= 0xE0 && b[0] <= 0xFC)) &&
14
+ (b[1] >= 0x40 && b[1] <= 0xFC)
20
15
  ) {
21
- *width = 2;
22
- return (yp_windows_31j_codepoint_t) (uc[0] << 8 | uc[1]);
16
+ return 2;
23
17
  }
24
18
 
25
- *width = 0;
26
19
  return 0;
27
20
  }
28
21
 
29
22
  static size_t
30
- yp_encoding_windows_31j_char_width(const char *c, ptrdiff_t n) {
31
- size_t width;
32
- yp_windows_31j_codepoint(c, n, &width);
33
-
34
- return width;
35
- }
36
-
37
- static size_t
38
- yp_encoding_windows_31j_alpha_char(const char *c, ptrdiff_t n) {
39
- size_t width;
40
- yp_windows_31j_codepoint_t codepoint = yp_windows_31j_codepoint(c, n, &width);
41
-
42
- if (width == 1) {
43
- const char value = (const char) codepoint;
44
- return yp_encoding_ascii_alpha_char(&value, n);
23
+ yp_encoding_windows_31j_alpha_char(const uint8_t *b, ptrdiff_t n) {
24
+ if (yp_encoding_windows_31j_char_width(b, n) == 1) {
25
+ return yp_encoding_ascii_alpha_char(b, n);
45
26
  } else {
46
27
  return 0;
47
28
  }
48
29
  }
49
30
 
50
31
  static size_t
51
- yp_encoding_windows_31j_alnum_char(const char *c, ptrdiff_t n) {
52
- size_t width;
53
- yp_windows_31j_codepoint_t codepoint = yp_windows_31j_codepoint(c, n, &width);
54
-
55
- if (width == 1) {
56
- const char value = (const char) codepoint;
57
- return yp_encoding_ascii_alnum_char(&value, n);
32
+ yp_encoding_windows_31j_alnum_char(const uint8_t *b, ptrdiff_t n) {
33
+ if (yp_encoding_windows_31j_char_width(b, n) == 1) {
34
+ return yp_encoding_ascii_alnum_char(b, n);
58
35
  } else {
59
36
  return 0;
60
37
  }
61
38
  }
62
39
 
63
40
  static bool
64
- yp_encoding_windows_31j_isupper_char(const char *c, ptrdiff_t n) {
65
- size_t width;
66
- yp_windows_31j_codepoint_t codepoint = yp_windows_31j_codepoint(c, n, &width);
67
-
68
- if (width == 1) {
69
- const char value = (const char) codepoint;
70
- return yp_encoding_ascii_isupper_char(&value, n);
41
+ yp_encoding_windows_31j_isupper_char(const uint8_t *b, ptrdiff_t n) {
42
+ if (yp_encoding_windows_31j_char_width(b, n) == 1) {
43
+ return yp_encoding_ascii_isupper_char(b, n);
71
44
  } else {
72
45
  return false;
73
46
  }