prism 0.13.0
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.
- checksums.yaml +7 -0
- data/CHANGELOG.md +172 -0
- data/CODE_OF_CONDUCT.md +76 -0
- data/CONTRIBUTING.md +62 -0
- data/LICENSE.md +7 -0
- data/Makefile +84 -0
- data/README.md +89 -0
- data/config.yml +2481 -0
- data/docs/build_system.md +74 -0
- data/docs/building.md +22 -0
- data/docs/configuration.md +60 -0
- data/docs/design.md +53 -0
- data/docs/encoding.md +117 -0
- data/docs/fuzzing.md +93 -0
- data/docs/heredocs.md +36 -0
- data/docs/mapping.md +117 -0
- data/docs/ripper.md +36 -0
- data/docs/ruby_api.md +25 -0
- data/docs/serialization.md +181 -0
- data/docs/testing.md +55 -0
- data/ext/prism/api_node.c +4725 -0
- data/ext/prism/api_pack.c +256 -0
- data/ext/prism/extconf.rb +136 -0
- data/ext/prism/extension.c +626 -0
- data/ext/prism/extension.h +18 -0
- data/include/prism/ast.h +1932 -0
- data/include/prism/defines.h +45 -0
- data/include/prism/diagnostic.h +231 -0
- data/include/prism/enc/pm_encoding.h +95 -0
- data/include/prism/node.h +41 -0
- data/include/prism/pack.h +141 -0
- data/include/prism/parser.h +418 -0
- data/include/prism/regexp.h +19 -0
- data/include/prism/unescape.h +48 -0
- data/include/prism/util/pm_buffer.h +51 -0
- data/include/prism/util/pm_char.h +91 -0
- data/include/prism/util/pm_constant_pool.h +78 -0
- data/include/prism/util/pm_list.h +67 -0
- data/include/prism/util/pm_memchr.h +14 -0
- data/include/prism/util/pm_newline_list.h +61 -0
- data/include/prism/util/pm_state_stack.h +24 -0
- data/include/prism/util/pm_string.h +61 -0
- data/include/prism/util/pm_string_list.h +25 -0
- data/include/prism/util/pm_strpbrk.h +29 -0
- data/include/prism/version.h +4 -0
- data/include/prism.h +82 -0
- data/lib/prism/compiler.rb +465 -0
- data/lib/prism/debug.rb +157 -0
- data/lib/prism/desugar_compiler.rb +206 -0
- data/lib/prism/dispatcher.rb +2051 -0
- data/lib/prism/dsl.rb +750 -0
- data/lib/prism/ffi.rb +251 -0
- data/lib/prism/lex_compat.rb +838 -0
- data/lib/prism/mutation_compiler.rb +718 -0
- data/lib/prism/node.rb +14540 -0
- data/lib/prism/node_ext.rb +55 -0
- data/lib/prism/node_inspector.rb +68 -0
- data/lib/prism/pack.rb +185 -0
- data/lib/prism/parse_result/comments.rb +172 -0
- data/lib/prism/parse_result/newlines.rb +60 -0
- data/lib/prism/parse_result.rb +266 -0
- data/lib/prism/pattern.rb +239 -0
- data/lib/prism/ripper_compat.rb +174 -0
- data/lib/prism/serialize.rb +662 -0
- data/lib/prism/visitor.rb +470 -0
- data/lib/prism.rb +64 -0
- data/prism.gemspec +113 -0
- data/src/diagnostic.c +287 -0
- data/src/enc/pm_big5.c +52 -0
- data/src/enc/pm_euc_jp.c +58 -0
- data/src/enc/pm_gbk.c +61 -0
- data/src/enc/pm_shift_jis.c +56 -0
- data/src/enc/pm_tables.c +507 -0
- data/src/enc/pm_unicode.c +2324 -0
- data/src/enc/pm_windows_31j.c +56 -0
- data/src/node.c +2633 -0
- data/src/pack.c +493 -0
- data/src/prettyprint.c +2136 -0
- data/src/prism.c +14587 -0
- data/src/regexp.c +580 -0
- data/src/serialize.c +1899 -0
- data/src/token_type.c +349 -0
- data/src/unescape.c +637 -0
- data/src/util/pm_buffer.c +103 -0
- data/src/util/pm_char.c +272 -0
- data/src/util/pm_constant_pool.c +252 -0
- data/src/util/pm_list.c +41 -0
- data/src/util/pm_memchr.c +33 -0
- data/src/util/pm_newline_list.c +134 -0
- data/src/util/pm_state_stack.c +19 -0
- data/src/util/pm_string.c +200 -0
- data/src/util/pm_string_list.c +29 -0
- data/src/util/pm_strncasecmp.c +17 -0
- data/src/util/pm_strpbrk.c +66 -0
- metadata +138 -0
@@ -0,0 +1,626 @@
|
|
1
|
+
#include "prism/extension.h"
|
2
|
+
|
3
|
+
// NOTE: this file should contain only bindings.
|
4
|
+
// All non-trivial logic should be in librubyparser so it can be shared its the various callers.
|
5
|
+
|
6
|
+
VALUE rb_cPrism;
|
7
|
+
VALUE rb_cPrismNode;
|
8
|
+
VALUE rb_cPrismSource;
|
9
|
+
VALUE rb_cPrismToken;
|
10
|
+
VALUE rb_cPrismLocation;
|
11
|
+
|
12
|
+
VALUE rb_cPrismComment;
|
13
|
+
VALUE rb_cPrismParseError;
|
14
|
+
VALUE rb_cPrismParseWarning;
|
15
|
+
VALUE rb_cPrismParseResult;
|
16
|
+
|
17
|
+
/******************************************************************************/
|
18
|
+
/* IO of Ruby code */
|
19
|
+
/******************************************************************************/
|
20
|
+
|
21
|
+
// Check if the given VALUE is a string. If it's nil, then return NULL. If it's
|
22
|
+
// not a string, then raise a type error. Otherwise return the VALUE as a C
|
23
|
+
// string.
|
24
|
+
static const char *
|
25
|
+
check_string(VALUE value) {
|
26
|
+
// If the value is nil, then we don't need to do anything.
|
27
|
+
if (NIL_P(value)) {
|
28
|
+
return NULL;
|
29
|
+
}
|
30
|
+
|
31
|
+
// Check if the value is a string. If it's not, then raise a type error.
|
32
|
+
if (!RB_TYPE_P(value, T_STRING)) {
|
33
|
+
rb_raise(rb_eTypeError, "wrong argument type %"PRIsVALUE" (expected String)", rb_obj_class(value));
|
34
|
+
}
|
35
|
+
|
36
|
+
// Otherwise, return the value as a C string.
|
37
|
+
return RSTRING_PTR(value);
|
38
|
+
}
|
39
|
+
|
40
|
+
// Load the contents and size of the given string into the given pm_string_t.
|
41
|
+
static void
|
42
|
+
input_load_string(pm_string_t *input, VALUE string) {
|
43
|
+
// Check if the string is a string. If it's not, then raise a type error.
|
44
|
+
if (!RB_TYPE_P(string, T_STRING)) {
|
45
|
+
rb_raise(rb_eTypeError, "wrong argument type %"PRIsVALUE" (expected String)", rb_obj_class(string));
|
46
|
+
}
|
47
|
+
|
48
|
+
pm_string_constant_init(input, RSTRING_PTR(string), RSTRING_LEN(string));
|
49
|
+
}
|
50
|
+
|
51
|
+
/******************************************************************************/
|
52
|
+
/* Serializing the AST */
|
53
|
+
/******************************************************************************/
|
54
|
+
|
55
|
+
// Dump the AST corresponding to the given input to a string.
|
56
|
+
static VALUE
|
57
|
+
dump_input(pm_string_t *input, const char *filepath) {
|
58
|
+
pm_buffer_t buffer;
|
59
|
+
if (!pm_buffer_init(&buffer)) {
|
60
|
+
rb_raise(rb_eNoMemError, "failed to allocate memory");
|
61
|
+
}
|
62
|
+
|
63
|
+
pm_parser_t parser;
|
64
|
+
pm_parser_init(&parser, pm_string_source(input), pm_string_length(input), filepath);
|
65
|
+
|
66
|
+
pm_node_t *node = pm_parse(&parser);
|
67
|
+
pm_serialize(&parser, node, &buffer);
|
68
|
+
|
69
|
+
VALUE result = rb_str_new(pm_buffer_value(&buffer), pm_buffer_length(&buffer));
|
70
|
+
pm_node_destroy(&parser, node);
|
71
|
+
pm_buffer_free(&buffer);
|
72
|
+
pm_parser_free(&parser);
|
73
|
+
|
74
|
+
return result;
|
75
|
+
}
|
76
|
+
|
77
|
+
// Dump the AST corresponding to the given string to a string.
|
78
|
+
static VALUE
|
79
|
+
dump(int argc, VALUE *argv, VALUE self) {
|
80
|
+
VALUE string;
|
81
|
+
VALUE filepath;
|
82
|
+
rb_scan_args(argc, argv, "11", &string, &filepath);
|
83
|
+
|
84
|
+
pm_string_t input;
|
85
|
+
input_load_string(&input, string);
|
86
|
+
|
87
|
+
#ifdef PRISM_DEBUG_MODE_BUILD
|
88
|
+
size_t length = pm_string_length(&input);
|
89
|
+
char* dup = malloc(length);
|
90
|
+
memcpy(dup, pm_string_source(&input), length);
|
91
|
+
pm_string_constant_init(&input, dup, length);
|
92
|
+
#endif
|
93
|
+
|
94
|
+
VALUE value = dump_input(&input, check_string(filepath));
|
95
|
+
|
96
|
+
#ifdef PRISM_DEBUG_MODE_BUILD
|
97
|
+
free(dup);
|
98
|
+
#endif
|
99
|
+
|
100
|
+
return value;
|
101
|
+
}
|
102
|
+
|
103
|
+
// Dump the AST corresponding to the given file to a string.
|
104
|
+
static VALUE
|
105
|
+
dump_file(VALUE self, VALUE filepath) {
|
106
|
+
pm_string_t input;
|
107
|
+
|
108
|
+
const char *checked = check_string(filepath);
|
109
|
+
if (!pm_string_mapped_init(&input, checked)) return Qnil;
|
110
|
+
|
111
|
+
VALUE value = dump_input(&input, checked);
|
112
|
+
pm_string_free(&input);
|
113
|
+
|
114
|
+
return value;
|
115
|
+
}
|
116
|
+
|
117
|
+
/******************************************************************************/
|
118
|
+
/* Extracting values for the parse result */
|
119
|
+
/******************************************************************************/
|
120
|
+
|
121
|
+
// Extract the comments out of the parser into an array.
|
122
|
+
static VALUE
|
123
|
+
parser_comments(pm_parser_t *parser, VALUE source) {
|
124
|
+
VALUE comments = rb_ary_new();
|
125
|
+
|
126
|
+
for (pm_comment_t *comment = (pm_comment_t *) parser->comment_list.head; comment != NULL; comment = (pm_comment_t *) comment->node.next) {
|
127
|
+
VALUE location_argv[] = {
|
128
|
+
source,
|
129
|
+
LONG2FIX(comment->start - parser->start),
|
130
|
+
LONG2FIX(comment->end - comment->start)
|
131
|
+
};
|
132
|
+
|
133
|
+
VALUE type;
|
134
|
+
switch (comment->type) {
|
135
|
+
case PM_COMMENT_INLINE:
|
136
|
+
type = ID2SYM(rb_intern("inline"));
|
137
|
+
break;
|
138
|
+
case PM_COMMENT_EMBDOC:
|
139
|
+
type = ID2SYM(rb_intern("embdoc"));
|
140
|
+
break;
|
141
|
+
case PM_COMMENT___END__:
|
142
|
+
type = ID2SYM(rb_intern("__END__"));
|
143
|
+
break;
|
144
|
+
default:
|
145
|
+
type = ID2SYM(rb_intern("inline"));
|
146
|
+
break;
|
147
|
+
}
|
148
|
+
|
149
|
+
VALUE comment_argv[] = { type, rb_class_new_instance(3, location_argv, rb_cPrismLocation) };
|
150
|
+
rb_ary_push(comments, rb_class_new_instance(2, comment_argv, rb_cPrismComment));
|
151
|
+
}
|
152
|
+
|
153
|
+
return comments;
|
154
|
+
}
|
155
|
+
|
156
|
+
// Extract the errors out of the parser into an array.
|
157
|
+
static VALUE
|
158
|
+
parser_errors(pm_parser_t *parser, rb_encoding *encoding, VALUE source) {
|
159
|
+
VALUE errors = rb_ary_new();
|
160
|
+
pm_diagnostic_t *error;
|
161
|
+
|
162
|
+
for (error = (pm_diagnostic_t *) parser->error_list.head; error != NULL; error = (pm_diagnostic_t *) error->node.next) {
|
163
|
+
VALUE location_argv[] = {
|
164
|
+
source,
|
165
|
+
LONG2FIX(error->start - parser->start),
|
166
|
+
LONG2FIX(error->end - error->start)
|
167
|
+
};
|
168
|
+
|
169
|
+
VALUE error_argv[] = {
|
170
|
+
rb_enc_str_new_cstr(error->message, encoding),
|
171
|
+
rb_class_new_instance(3, location_argv, rb_cPrismLocation)
|
172
|
+
};
|
173
|
+
|
174
|
+
rb_ary_push(errors, rb_class_new_instance(2, error_argv, rb_cPrismParseError));
|
175
|
+
}
|
176
|
+
|
177
|
+
return errors;
|
178
|
+
}
|
179
|
+
|
180
|
+
// Extract the warnings out of the parser into an array.
|
181
|
+
static VALUE
|
182
|
+
parser_warnings(pm_parser_t *parser, rb_encoding *encoding, VALUE source) {
|
183
|
+
VALUE warnings = rb_ary_new();
|
184
|
+
pm_diagnostic_t *warning;
|
185
|
+
|
186
|
+
for (warning = (pm_diagnostic_t *) parser->warning_list.head; warning != NULL; warning = (pm_diagnostic_t *) warning->node.next) {
|
187
|
+
VALUE location_argv[] = {
|
188
|
+
source,
|
189
|
+
LONG2FIX(warning->start - parser->start),
|
190
|
+
LONG2FIX(warning->end - warning->start)
|
191
|
+
};
|
192
|
+
|
193
|
+
VALUE warning_argv[] = {
|
194
|
+
rb_enc_str_new_cstr(warning->message, encoding),
|
195
|
+
rb_class_new_instance(3, location_argv, rb_cPrismLocation)
|
196
|
+
};
|
197
|
+
|
198
|
+
rb_ary_push(warnings, rb_class_new_instance(2, warning_argv, rb_cPrismParseWarning));
|
199
|
+
}
|
200
|
+
|
201
|
+
return warnings;
|
202
|
+
}
|
203
|
+
|
204
|
+
/******************************************************************************/
|
205
|
+
/* Lexing Ruby code */
|
206
|
+
/******************************************************************************/
|
207
|
+
|
208
|
+
// This struct gets stored in the parser and passed in to the lex callback any
|
209
|
+
// time a new token is found. We use it to store the necessary information to
|
210
|
+
// initialize a Token instance.
|
211
|
+
typedef struct {
|
212
|
+
VALUE source;
|
213
|
+
VALUE tokens;
|
214
|
+
rb_encoding *encoding;
|
215
|
+
} parse_lex_data_t;
|
216
|
+
|
217
|
+
// This is passed as a callback to the parser. It gets called every time a new
|
218
|
+
// token is found. Once found, we initialize a new instance of Token and push it
|
219
|
+
// onto the tokens array.
|
220
|
+
static void
|
221
|
+
parse_lex_token(void *data, pm_parser_t *parser, pm_token_t *token) {
|
222
|
+
parse_lex_data_t *parse_lex_data = (parse_lex_data_t *) parser->lex_callback->data;
|
223
|
+
|
224
|
+
VALUE yields = rb_ary_new_capa(2);
|
225
|
+
rb_ary_push(yields, pm_token_new(parser, token, parse_lex_data->encoding, parse_lex_data->source));
|
226
|
+
rb_ary_push(yields, INT2FIX(parser->lex_state));
|
227
|
+
|
228
|
+
rb_ary_push(parse_lex_data->tokens, yields);
|
229
|
+
}
|
230
|
+
|
231
|
+
// This is called whenever the encoding changes based on the magic comment at
|
232
|
+
// the top of the file. We use it to update the encoding that we are using to
|
233
|
+
// create tokens.
|
234
|
+
static void
|
235
|
+
parse_lex_encoding_changed_callback(pm_parser_t *parser) {
|
236
|
+
parse_lex_data_t *parse_lex_data = (parse_lex_data_t *) parser->lex_callback->data;
|
237
|
+
parse_lex_data->encoding = rb_enc_find(parser->encoding.name);
|
238
|
+
|
239
|
+
// Since the encoding changed, we need to go back and change the encoding of
|
240
|
+
// the tokens that were already lexed. This is only going to end up being
|
241
|
+
// one or two tokens, since the encoding can only change at the top of the
|
242
|
+
// file.
|
243
|
+
VALUE tokens = parse_lex_data->tokens;
|
244
|
+
for (long index = 0; index < RARRAY_LEN(tokens); index++) {
|
245
|
+
VALUE yields = rb_ary_entry(tokens, index);
|
246
|
+
VALUE token = rb_ary_entry(yields, 0);
|
247
|
+
|
248
|
+
VALUE value = rb_ivar_get(token, rb_intern("@value"));
|
249
|
+
rb_enc_associate(value, parse_lex_data->encoding);
|
250
|
+
ENC_CODERANGE_CLEAR(value);
|
251
|
+
}
|
252
|
+
}
|
253
|
+
|
254
|
+
// Parse the given input and return a ParseResult containing just the tokens or
|
255
|
+
// the nodes and tokens.
|
256
|
+
static VALUE
|
257
|
+
parse_lex_input(pm_string_t *input, const char *filepath, bool return_nodes) {
|
258
|
+
pm_parser_t parser;
|
259
|
+
pm_parser_init(&parser, pm_string_source(input), pm_string_length(input), filepath);
|
260
|
+
pm_parser_register_encoding_changed_callback(&parser, parse_lex_encoding_changed_callback);
|
261
|
+
|
262
|
+
VALUE offsets = rb_ary_new();
|
263
|
+
VALUE source_argv[] = { rb_str_new((const char *) pm_string_source(input), pm_string_length(input)), offsets };
|
264
|
+
VALUE source = rb_class_new_instance(2, source_argv, rb_cPrismSource);
|
265
|
+
|
266
|
+
parse_lex_data_t parse_lex_data = {
|
267
|
+
.source = source,
|
268
|
+
.tokens = rb_ary_new(),
|
269
|
+
.encoding = rb_utf8_encoding()
|
270
|
+
};
|
271
|
+
|
272
|
+
parse_lex_data_t *data = &parse_lex_data;
|
273
|
+
pm_lex_callback_t lex_callback = (pm_lex_callback_t) {
|
274
|
+
.data = (void *) data,
|
275
|
+
.callback = parse_lex_token,
|
276
|
+
};
|
277
|
+
|
278
|
+
parser.lex_callback = &lex_callback;
|
279
|
+
pm_node_t *node = pm_parse(&parser);
|
280
|
+
|
281
|
+
// Here we need to update the source range to have the correct newline
|
282
|
+
// offsets. We do it here because we've already created the object and given
|
283
|
+
// it over to all of the tokens.
|
284
|
+
for (size_t index = 0; index < parser.newline_list.size; index++) {
|
285
|
+
rb_ary_push(offsets, INT2FIX(parser.newline_list.offsets[index]));
|
286
|
+
}
|
287
|
+
|
288
|
+
VALUE value;
|
289
|
+
if (return_nodes) {
|
290
|
+
value = rb_ary_new_capa(2);
|
291
|
+
rb_ary_push(value, pm_ast_new(&parser, node, parse_lex_data.encoding));
|
292
|
+
rb_ary_push(value, parse_lex_data.tokens);
|
293
|
+
} else {
|
294
|
+
value = parse_lex_data.tokens;
|
295
|
+
}
|
296
|
+
|
297
|
+
VALUE result_argv[] = {
|
298
|
+
value,
|
299
|
+
parser_comments(&parser, source),
|
300
|
+
parser_errors(&parser, parse_lex_data.encoding, source),
|
301
|
+
parser_warnings(&parser, parse_lex_data.encoding, source),
|
302
|
+
source
|
303
|
+
};
|
304
|
+
|
305
|
+
pm_node_destroy(&parser, node);
|
306
|
+
pm_parser_free(&parser);
|
307
|
+
return rb_class_new_instance(5, result_argv, rb_cPrismParseResult);
|
308
|
+
}
|
309
|
+
|
310
|
+
// Return an array of tokens corresponding to the given string.
|
311
|
+
static VALUE
|
312
|
+
lex(int argc, VALUE *argv, VALUE self) {
|
313
|
+
VALUE string;
|
314
|
+
VALUE filepath;
|
315
|
+
rb_scan_args(argc, argv, "11", &string, &filepath);
|
316
|
+
|
317
|
+
pm_string_t input;
|
318
|
+
input_load_string(&input, string);
|
319
|
+
|
320
|
+
return parse_lex_input(&input, check_string(filepath), false);
|
321
|
+
}
|
322
|
+
|
323
|
+
// Return an array of tokens corresponding to the given file.
|
324
|
+
static VALUE
|
325
|
+
lex_file(VALUE self, VALUE filepath) {
|
326
|
+
pm_string_t input;
|
327
|
+
|
328
|
+
const char *checked = check_string(filepath);
|
329
|
+
if (!pm_string_mapped_init(&input, checked)) return Qnil;
|
330
|
+
|
331
|
+
VALUE value = parse_lex_input(&input, checked, false);
|
332
|
+
pm_string_free(&input);
|
333
|
+
|
334
|
+
return value;
|
335
|
+
}
|
336
|
+
|
337
|
+
/******************************************************************************/
|
338
|
+
/* Parsing Ruby code */
|
339
|
+
/******************************************************************************/
|
340
|
+
|
341
|
+
// Parse the given input and return a ParseResult instance.
|
342
|
+
static VALUE
|
343
|
+
parse_input(pm_string_t *input, const char *filepath) {
|
344
|
+
pm_parser_t parser;
|
345
|
+
pm_parser_init(&parser, pm_string_source(input), pm_string_length(input), filepath);
|
346
|
+
|
347
|
+
pm_node_t *node = pm_parse(&parser);
|
348
|
+
rb_encoding *encoding = rb_enc_find(parser.encoding.name);
|
349
|
+
|
350
|
+
VALUE source = pm_source_new(&parser, encoding);
|
351
|
+
VALUE result_argv[] = {
|
352
|
+
pm_ast_new(&parser, node, encoding),
|
353
|
+
parser_comments(&parser, source),
|
354
|
+
parser_errors(&parser, encoding, source),
|
355
|
+
parser_warnings(&parser, encoding, source),
|
356
|
+
source
|
357
|
+
};
|
358
|
+
|
359
|
+
VALUE result = rb_class_new_instance(5, result_argv, rb_cPrismParseResult);
|
360
|
+
|
361
|
+
pm_node_destroy(&parser, node);
|
362
|
+
pm_parser_free(&parser);
|
363
|
+
|
364
|
+
return result;
|
365
|
+
}
|
366
|
+
|
367
|
+
// Parse the given string and return a ParseResult instance.
|
368
|
+
static VALUE
|
369
|
+
parse(int argc, VALUE *argv, VALUE self) {
|
370
|
+
VALUE string;
|
371
|
+
VALUE filepath;
|
372
|
+
rb_scan_args(argc, argv, "11", &string, &filepath);
|
373
|
+
|
374
|
+
pm_string_t input;
|
375
|
+
input_load_string(&input, string);
|
376
|
+
|
377
|
+
#ifdef PRISM_DEBUG_MODE_BUILD
|
378
|
+
size_t length = pm_string_length(&input);
|
379
|
+
char* dup = malloc(length);
|
380
|
+
memcpy(dup, pm_string_source(&input), length);
|
381
|
+
pm_string_constant_init(&input, dup, length);
|
382
|
+
#endif
|
383
|
+
|
384
|
+
VALUE value = parse_input(&input, check_string(filepath));
|
385
|
+
|
386
|
+
#ifdef PRISM_DEBUG_MODE_BUILD
|
387
|
+
free(dup);
|
388
|
+
#endif
|
389
|
+
|
390
|
+
return value;
|
391
|
+
}
|
392
|
+
|
393
|
+
// Parse the given file and return a ParseResult instance.
|
394
|
+
static VALUE
|
395
|
+
parse_file(VALUE self, VALUE filepath) {
|
396
|
+
pm_string_t input;
|
397
|
+
|
398
|
+
const char *checked = check_string(filepath);
|
399
|
+
if (!pm_string_mapped_init(&input, checked)) return Qnil;
|
400
|
+
|
401
|
+
VALUE value = parse_input(&input, checked);
|
402
|
+
pm_string_free(&input);
|
403
|
+
|
404
|
+
return value;
|
405
|
+
}
|
406
|
+
|
407
|
+
// Parse the given string and return a ParseResult instance.
|
408
|
+
static VALUE
|
409
|
+
parse_lex(int argc, VALUE *argv, VALUE self) {
|
410
|
+
VALUE string;
|
411
|
+
VALUE filepath;
|
412
|
+
rb_scan_args(argc, argv, "11", &string, &filepath);
|
413
|
+
|
414
|
+
pm_string_t input;
|
415
|
+
input_load_string(&input, string);
|
416
|
+
|
417
|
+
VALUE value = parse_lex_input(&input, check_string(filepath), true);
|
418
|
+
pm_string_free(&input);
|
419
|
+
|
420
|
+
return value;
|
421
|
+
}
|
422
|
+
|
423
|
+
// Parse and lex the given file and return a ParseResult instance.
|
424
|
+
static VALUE
|
425
|
+
parse_lex_file(VALUE self, VALUE filepath) {
|
426
|
+
pm_string_t input;
|
427
|
+
|
428
|
+
const char *checked = check_string(filepath);
|
429
|
+
if (!pm_string_mapped_init(&input, checked)) return Qnil;
|
430
|
+
|
431
|
+
VALUE value = parse_lex_input(&input, checked, true);
|
432
|
+
pm_string_free(&input);
|
433
|
+
|
434
|
+
return value;
|
435
|
+
}
|
436
|
+
|
437
|
+
/******************************************************************************/
|
438
|
+
/* Utility functions exposed to make testing easier */
|
439
|
+
/******************************************************************************/
|
440
|
+
|
441
|
+
// Returns an array of strings corresponding to the named capture groups in the
|
442
|
+
// given source string. If prism was unable to parse the regular expression, this
|
443
|
+
// function returns nil.
|
444
|
+
static VALUE
|
445
|
+
named_captures(VALUE self, VALUE source) {
|
446
|
+
pm_string_list_t string_list;
|
447
|
+
pm_string_list_init(&string_list);
|
448
|
+
|
449
|
+
if (!pm_regexp_named_capture_group_names((const uint8_t *) RSTRING_PTR(source), RSTRING_LEN(source), &string_list, false, &pm_encoding_utf_8)) {
|
450
|
+
pm_string_list_free(&string_list);
|
451
|
+
return Qnil;
|
452
|
+
}
|
453
|
+
|
454
|
+
VALUE names = rb_ary_new();
|
455
|
+
for (size_t index = 0; index < string_list.length; index++) {
|
456
|
+
const pm_string_t *string = &string_list.strings[index];
|
457
|
+
rb_ary_push(names, rb_str_new((const char *) pm_string_source(string), pm_string_length(string)));
|
458
|
+
}
|
459
|
+
|
460
|
+
pm_string_list_free(&string_list);
|
461
|
+
return names;
|
462
|
+
}
|
463
|
+
|
464
|
+
// Accepts a source string and a type of unescaping and returns the unescaped
|
465
|
+
// version.
|
466
|
+
static VALUE
|
467
|
+
unescape(VALUE source, pm_unescape_type_t unescape_type) {
|
468
|
+
pm_string_t result;
|
469
|
+
|
470
|
+
if (pm_unescape_string((const uint8_t *) RSTRING_PTR(source), RSTRING_LEN(source), unescape_type, &result)) {
|
471
|
+
VALUE str = rb_str_new((const char *) pm_string_source(&result), pm_string_length(&result));
|
472
|
+
pm_string_free(&result);
|
473
|
+
return str;
|
474
|
+
} else {
|
475
|
+
pm_string_free(&result);
|
476
|
+
return Qnil;
|
477
|
+
}
|
478
|
+
}
|
479
|
+
|
480
|
+
// Do not unescape anything in the given string. This is here to provide a
|
481
|
+
// consistent API.
|
482
|
+
static VALUE
|
483
|
+
unescape_none(VALUE self, VALUE source) {
|
484
|
+
return unescape(source, PM_UNESCAPE_NONE);
|
485
|
+
}
|
486
|
+
|
487
|
+
// Minimally unescape the given string. This means effectively unescaping just
|
488
|
+
// the quotes of a string. Returns the unescaped string.
|
489
|
+
static VALUE
|
490
|
+
unescape_minimal(VALUE self, VALUE source) {
|
491
|
+
return unescape(source, PM_UNESCAPE_MINIMAL);
|
492
|
+
}
|
493
|
+
|
494
|
+
// Escape the given string minimally plus whitespace. Returns the unescaped string.
|
495
|
+
static VALUE
|
496
|
+
unescape_whitespace(VALUE self, VALUE source) {
|
497
|
+
return unescape(source, PM_UNESCAPE_WHITESPACE);
|
498
|
+
}
|
499
|
+
|
500
|
+
// Unescape everything in the given string. Return the unescaped string.
|
501
|
+
static VALUE
|
502
|
+
unescape_all(VALUE self, VALUE source) {
|
503
|
+
return unescape(source, PM_UNESCAPE_ALL);
|
504
|
+
}
|
505
|
+
|
506
|
+
// Return a hash of information about the given source string's memory usage.
|
507
|
+
static VALUE
|
508
|
+
memsize(VALUE self, VALUE string) {
|
509
|
+
pm_parser_t parser;
|
510
|
+
size_t length = RSTRING_LEN(string);
|
511
|
+
pm_parser_init(&parser, (const uint8_t *) RSTRING_PTR(string), length, NULL);
|
512
|
+
|
513
|
+
pm_node_t *node = pm_parse(&parser);
|
514
|
+
pm_memsize_t memsize;
|
515
|
+
pm_node_memsize(node, &memsize);
|
516
|
+
|
517
|
+
pm_node_destroy(&parser, node);
|
518
|
+
pm_parser_free(&parser);
|
519
|
+
|
520
|
+
VALUE result = rb_hash_new();
|
521
|
+
rb_hash_aset(result, ID2SYM(rb_intern("length")), INT2FIX(length));
|
522
|
+
rb_hash_aset(result, ID2SYM(rb_intern("memsize")), INT2FIX(memsize.memsize));
|
523
|
+
rb_hash_aset(result, ID2SYM(rb_intern("node_count")), INT2FIX(memsize.node_count));
|
524
|
+
return result;
|
525
|
+
}
|
526
|
+
|
527
|
+
// Parse the file, but do nothing with the result. This is used to profile the
|
528
|
+
// parser for memory and speed.
|
529
|
+
static VALUE
|
530
|
+
profile_file(VALUE self, VALUE filepath) {
|
531
|
+
pm_string_t input;
|
532
|
+
|
533
|
+
const char *checked = check_string(filepath);
|
534
|
+
if (!pm_string_mapped_init(&input, checked)) return Qnil;
|
535
|
+
|
536
|
+
pm_parser_t parser;
|
537
|
+
pm_parser_init(&parser, pm_string_source(&input), pm_string_length(&input), checked);
|
538
|
+
|
539
|
+
pm_node_t *node = pm_parse(&parser);
|
540
|
+
pm_node_destroy(&parser, node);
|
541
|
+
pm_parser_free(&parser);
|
542
|
+
|
543
|
+
pm_string_free(&input);
|
544
|
+
|
545
|
+
return Qnil;
|
546
|
+
}
|
547
|
+
|
548
|
+
// Parse the file and serialize the result. This is mostly used to test this
|
549
|
+
// path since it is used by client libraries.
|
550
|
+
static VALUE
|
551
|
+
parse_serialize_file_metadata(VALUE self, VALUE filepath, VALUE metadata) {
|
552
|
+
pm_string_t input;
|
553
|
+
pm_buffer_t buffer;
|
554
|
+
pm_buffer_init(&buffer);
|
555
|
+
|
556
|
+
const char *checked = check_string(filepath);
|
557
|
+
if (!pm_string_mapped_init(&input, checked)) return Qnil;
|
558
|
+
|
559
|
+
pm_parse_serialize(pm_string_source(&input), pm_string_length(&input), &buffer, check_string(metadata));
|
560
|
+
VALUE result = rb_str_new(pm_buffer_value(&buffer), pm_buffer_length(&buffer));
|
561
|
+
|
562
|
+
pm_string_free(&input);
|
563
|
+
pm_buffer_free(&buffer);
|
564
|
+
return result;
|
565
|
+
}
|
566
|
+
|
567
|
+
/******************************************************************************/
|
568
|
+
/* Initialization of the extension */
|
569
|
+
/******************************************************************************/
|
570
|
+
|
571
|
+
RUBY_FUNC_EXPORTED void
|
572
|
+
Init_prism(void) {
|
573
|
+
// Make sure that the prism library version matches the expected version.
|
574
|
+
// Otherwise something was compiled incorrectly.
|
575
|
+
if (strcmp(pm_version(), EXPECTED_PRISM_VERSION) != 0) {
|
576
|
+
rb_raise(
|
577
|
+
rb_eRuntimeError,
|
578
|
+
"The prism library version (%s) does not match the expected version (%s)",
|
579
|
+
pm_version(),
|
580
|
+
EXPECTED_PRISM_VERSION
|
581
|
+
);
|
582
|
+
}
|
583
|
+
|
584
|
+
// Grab up references to all of the constants that we're going to need to
|
585
|
+
// reference throughout this extension.
|
586
|
+
rb_cPrism = rb_define_module("Prism");
|
587
|
+
rb_cPrismNode = rb_define_class_under(rb_cPrism, "Node", rb_cObject);
|
588
|
+
rb_cPrismSource = rb_define_class_under(rb_cPrism, "Source", rb_cObject);
|
589
|
+
rb_cPrismToken = rb_define_class_under(rb_cPrism, "Token", rb_cObject);
|
590
|
+
rb_cPrismLocation = rb_define_class_under(rb_cPrism, "Location", rb_cObject);
|
591
|
+
rb_cPrismComment = rb_define_class_under(rb_cPrism, "Comment", rb_cObject);
|
592
|
+
rb_cPrismParseError = rb_define_class_under(rb_cPrism, "ParseError", rb_cObject);
|
593
|
+
rb_cPrismParseWarning = rb_define_class_under(rb_cPrism, "ParseWarning", rb_cObject);
|
594
|
+
rb_cPrismParseResult = rb_define_class_under(rb_cPrism, "ParseResult", rb_cObject);
|
595
|
+
|
596
|
+
// Define the version string here so that we can use the constants defined
|
597
|
+
// in prism.h.
|
598
|
+
rb_define_const(rb_cPrism, "VERSION", rb_str_new2(EXPECTED_PRISM_VERSION));
|
599
|
+
rb_define_const(rb_cPrism, "BACKEND", ID2SYM(rb_intern("CExtension")));
|
600
|
+
|
601
|
+
// First, the functions that have to do with lexing and parsing.
|
602
|
+
rb_define_singleton_method(rb_cPrism, "dump", dump, -1);
|
603
|
+
rb_define_singleton_method(rb_cPrism, "dump_file", dump_file, 1);
|
604
|
+
rb_define_singleton_method(rb_cPrism, "lex", lex, -1);
|
605
|
+
rb_define_singleton_method(rb_cPrism, "lex_file", lex_file, 1);
|
606
|
+
rb_define_singleton_method(rb_cPrism, "parse", parse, -1);
|
607
|
+
rb_define_singleton_method(rb_cPrism, "parse_file", parse_file, 1);
|
608
|
+
rb_define_singleton_method(rb_cPrism, "parse_lex", parse_lex, -1);
|
609
|
+
rb_define_singleton_method(rb_cPrism, "parse_lex_file", parse_lex_file, 1);
|
610
|
+
|
611
|
+
// Next, the functions that will be called by the parser to perform various
|
612
|
+
// internal tasks. We expose these to make them easier to test.
|
613
|
+
VALUE rb_cPrismDebug = rb_define_module_under(rb_cPrism, "Debug");
|
614
|
+
rb_define_singleton_method(rb_cPrismDebug, "named_captures", named_captures, 1);
|
615
|
+
rb_define_singleton_method(rb_cPrismDebug, "unescape_none", unescape_none, 1);
|
616
|
+
rb_define_singleton_method(rb_cPrismDebug, "unescape_minimal", unescape_minimal, 1);
|
617
|
+
rb_define_singleton_method(rb_cPrismDebug, "unescape_whitespace", unescape_whitespace, 1);
|
618
|
+
rb_define_singleton_method(rb_cPrismDebug, "unescape_all", unescape_all, 1);
|
619
|
+
rb_define_singleton_method(rb_cPrismDebug, "memsize", memsize, 1);
|
620
|
+
rb_define_singleton_method(rb_cPrismDebug, "profile_file", profile_file, 1);
|
621
|
+
rb_define_singleton_method(rb_cPrismDebug, "parse_serialize_file_metadata", parse_serialize_file_metadata, 2);
|
622
|
+
|
623
|
+
// Next, initialize the other APIs.
|
624
|
+
Init_prism_api_node();
|
625
|
+
Init_prism_pack();
|
626
|
+
}
|
@@ -0,0 +1,18 @@
|
|
1
|
+
#ifndef PRISM_EXT_NODE_H
|
2
|
+
#define PRISM_EXT_NODE_H
|
3
|
+
|
4
|
+
#define EXPECTED_PRISM_VERSION "0.13.0"
|
5
|
+
|
6
|
+
#include <ruby.h>
|
7
|
+
#include <ruby/encoding.h>
|
8
|
+
#include "prism.h"
|
9
|
+
|
10
|
+
VALUE pm_source_new(pm_parser_t *parser, rb_encoding *encoding);
|
11
|
+
VALUE pm_token_new(pm_parser_t *parser, pm_token_t *token, rb_encoding *encoding, VALUE source);
|
12
|
+
VALUE pm_ast_new(pm_parser_t *parser, pm_node_t *node, rb_encoding *encoding);
|
13
|
+
|
14
|
+
void Init_prism_api_node(void);
|
15
|
+
void Init_prism_pack(void);
|
16
|
+
PRISM_EXPORTED_FUNCTION void Init_prism(void);
|
17
|
+
|
18
|
+
#endif
|