| paul@0 | 1 | #!/usr/bin/env python |
| paul@0 | 2 | |
| paul@0 | 3 | """ |
| paul@0 | 4 | Common functions. |
| paul@0 | 5 | |
| paul@0 | 6 | Copyright (C) 2007, 2008, 2009, 2010, 2011, 2012, 2013, |
| paul@508 | 7 | 2014, 2015, 2016, 2017 Paul Boddie <paul@boddie.org.uk> |
| paul@0 | 8 | |
| paul@0 | 9 | This program is free software; you can redistribute it and/or modify it under |
| paul@0 | 10 | the terms of the GNU General Public License as published by the Free Software |
| paul@0 | 11 | Foundation; either version 3 of the License, or (at your option) any later |
| paul@0 | 12 | version. |
| paul@0 | 13 | |
| paul@0 | 14 | This program is distributed in the hope that it will be useful, but WITHOUT |
| paul@0 | 15 | ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS |
| paul@0 | 16 | FOR A PARTICULAR PURPOSE. See the GNU General Public License for more |
| paul@0 | 17 | details. |
| paul@0 | 18 | |
| paul@0 | 19 | You should have received a copy of the GNU General Public License along with |
| paul@0 | 20 | this program. If not, see <http://www.gnu.org/licenses/>. |
| paul@0 | 21 | """ |
| paul@0 | 22 | |
| paul@512 | 23 | from compiler.transformer import Transformer |
| paul@430 | 24 | from errors import InspectError |
| paul@0 | 25 | from os import listdir, makedirs, remove |
| paul@609 | 26 | from os.path import exists, getmtime, isdir, join, split |
| paul@11 | 27 | from results import ConstantValueRef, LiteralSequenceRef, NameRef |
| paul@405 | 28 | import compiler.ast |
| paul@0 | 29 | |
| paul@0 | 30 | class CommonOutput: |
| paul@0 | 31 | |
| paul@0 | 32 | "Common output functionality." |
| paul@0 | 33 | |
| paul@617 | 34 | def check_output(self, options=None): |
| paul@0 | 35 | |
| paul@0 | 36 | "Check the existing output and remove it if irrelevant." |
| paul@0 | 37 | |
| paul@0 | 38 | if not exists(self.output): |
| paul@0 | 39 | makedirs(self.output) |
| paul@0 | 40 | |
| paul@0 | 41 | details = self.importer.get_cache_details() |
| paul@0 | 42 | recorded_details = self.get_output_details() |
| paul@0 | 43 | |
| paul@617 | 44 | # Combine cache details with any options. |
| paul@617 | 45 | |
| paul@617 | 46 | full_details = options and (details + " " + options) or details |
| paul@617 | 47 | |
| paul@617 | 48 | if recorded_details != full_details: |
| paul@0 | 49 | self.remove_output() |
| paul@0 | 50 | |
| paul@617 | 51 | writefile(self.get_output_details_filename(), full_details) |
| paul@0 | 52 | |
| paul@0 | 53 | def get_output_details_filename(self): |
| paul@0 | 54 | |
| paul@0 | 55 | "Return the output details filename." |
| paul@0 | 56 | |
| paul@0 | 57 | return join(self.output, "$details") |
| paul@0 | 58 | |
| paul@0 | 59 | def get_output_details(self): |
| paul@0 | 60 | |
| paul@0 | 61 | "Return details of the existing output." |
| paul@0 | 62 | |
| paul@0 | 63 | details_filename = self.get_output_details_filename() |
| paul@0 | 64 | |
| paul@0 | 65 | if not exists(details_filename): |
| paul@0 | 66 | return None |
| paul@0 | 67 | else: |
| paul@0 | 68 | return readfile(details_filename) |
| paul@0 | 69 | |
| paul@0 | 70 | def remove_output(self, dirname=None): |
| paul@0 | 71 | |
| paul@0 | 72 | "Remove the output." |
| paul@0 | 73 | |
| paul@0 | 74 | dirname = dirname or self.output |
| paul@0 | 75 | |
| paul@0 | 76 | for filename in listdir(dirname): |
| paul@0 | 77 | path = join(dirname, filename) |
| paul@0 | 78 | if isdir(path): |
| paul@0 | 79 | self.remove_output(path) |
| paul@0 | 80 | else: |
| paul@0 | 81 | remove(path) |
| paul@0 | 82 | |
| paul@609 | 83 | def copy(source, target, only_if_newer=True): |
| paul@609 | 84 | |
| paul@609 | 85 | "Copy a text file from 'source' to 'target'." |
| paul@609 | 86 | |
| paul@609 | 87 | if isdir(target): |
| paul@609 | 88 | target = join(target, split(source)[-1]) |
| paul@609 | 89 | |
| paul@609 | 90 | if only_if_newer and not is_newer(source, target): |
| paul@609 | 91 | return |
| paul@609 | 92 | |
| paul@609 | 93 | infile = open(source) |
| paul@609 | 94 | outfile = open(target, "w") |
| paul@609 | 95 | |
| paul@609 | 96 | try: |
| paul@609 | 97 | outfile.write(infile.read()) |
| paul@609 | 98 | finally: |
| paul@609 | 99 | outfile.close() |
| paul@609 | 100 | infile.close() |
| paul@609 | 101 | |
| paul@609 | 102 | def is_newer(source, target): |
| paul@609 | 103 | |
| paul@609 | 104 | "Return whether 'source' is newer than 'target'." |
| paul@609 | 105 | |
| paul@609 | 106 | if exists(target): |
| paul@609 | 107 | target_mtime = getmtime(target) |
| paul@609 | 108 | source_mtime = getmtime(source) |
| paul@609 | 109 | return source_mtime > target_mtime |
| paul@609 | 110 | |
| paul@609 | 111 | return True |
| paul@609 | 112 | |
| paul@0 | 113 | class CommonModule: |
| paul@0 | 114 | |
| paul@0 | 115 | "A common module representation." |
| paul@0 | 116 | |
| paul@0 | 117 | def __init__(self, name, importer): |
| paul@0 | 118 | |
| paul@0 | 119 | """ |
| paul@0 | 120 | Initialise this module with the given 'name' and an 'importer' which is |
| paul@0 | 121 | used to provide access to other modules when required. |
| paul@0 | 122 | """ |
| paul@0 | 123 | |
| paul@0 | 124 | self.name = name |
| paul@0 | 125 | self.importer = importer |
| paul@0 | 126 | self.filename = None |
| paul@0 | 127 | |
| paul@0 | 128 | # Inspection-related attributes. |
| paul@0 | 129 | |
| paul@0 | 130 | self.astnode = None |
| paul@405 | 131 | self.encoding = None |
| paul@0 | 132 | self.temp = {} |
| paul@0 | 133 | self.lambdas = {} |
| paul@0 | 134 | |
| paul@0 | 135 | # Constants, literals and values. |
| paul@0 | 136 | |
| paul@0 | 137 | self.constants = {} |
| paul@0 | 138 | self.constant_values = {} |
| paul@0 | 139 | self.literals = {} |
| paul@0 | 140 | self.literal_types = {} |
| paul@0 | 141 | |
| paul@0 | 142 | # Nested namespaces. |
| paul@0 | 143 | |
| paul@0 | 144 | self.namespace_path = [] |
| paul@0 | 145 | self.in_function = False |
| paul@0 | 146 | |
| paul@124 | 147 | # Retain the assignment value expression and track invocations. |
| paul@124 | 148 | |
| paul@124 | 149 | self.in_assignment = None |
| paul@553 | 150 | self.in_invocation = None |
| paul@124 | 151 | |
| paul@124 | 152 | # Attribute chain state management. |
| paul@0 | 153 | |
| paul@0 | 154 | self.attrs = [] |
| paul@124 | 155 | self.chain_assignment = [] |
| paul@124 | 156 | self.chain_invocation = [] |
| paul@0 | 157 | |
| paul@0 | 158 | def __repr__(self): |
| paul@0 | 159 | return "CommonModule(%r, %r)" % (self.name, self.importer) |
| paul@0 | 160 | |
| paul@0 | 161 | def parse_file(self, filename): |
| paul@0 | 162 | |
| paul@0 | 163 | "Parse the file with the given 'filename', initialising attributes." |
| paul@0 | 164 | |
| paul@0 | 165 | self.filename = filename |
| paul@405 | 166 | |
| paul@405 | 167 | # Use the Transformer directly to obtain encoding information. |
| paul@405 | 168 | |
| paul@405 | 169 | t = Transformer() |
| paul@405 | 170 | f = open(filename) |
| paul@405 | 171 | |
| paul@405 | 172 | try: |
| paul@405 | 173 | self.astnode = t.parsesuite(f.read() + "\n") |
| paul@405 | 174 | self.encoding = t.encoding |
| paul@405 | 175 | finally: |
| paul@405 | 176 | f.close() |
| paul@0 | 177 | |
| paul@0 | 178 | # Module-relative naming. |
| paul@0 | 179 | |
| paul@0 | 180 | def get_global_path(self, name): |
| paul@0 | 181 | return "%s.%s" % (self.name, name) |
| paul@0 | 182 | |
| paul@0 | 183 | def get_namespace_path(self): |
| paul@0 | 184 | return ".".join([self.name] + self.namespace_path) |
| paul@0 | 185 | |
| paul@0 | 186 | def get_object_path(self, name): |
| paul@0 | 187 | return ".".join([self.name] + self.namespace_path + [name]) |
| paul@0 | 188 | |
| paul@0 | 189 | def get_parent_path(self): |
| paul@0 | 190 | return ".".join([self.name] + self.namespace_path[:-1]) |
| paul@0 | 191 | |
| paul@0 | 192 | # Namespace management. |
| paul@0 | 193 | |
| paul@0 | 194 | def enter_namespace(self, name): |
| paul@0 | 195 | |
| paul@0 | 196 | "Enter the namespace having the given 'name'." |
| paul@0 | 197 | |
| paul@0 | 198 | self.namespace_path.append(name) |
| paul@0 | 199 | |
| paul@0 | 200 | def exit_namespace(self): |
| paul@0 | 201 | |
| paul@0 | 202 | "Exit the current namespace." |
| paul@0 | 203 | |
| paul@0 | 204 | self.namespace_path.pop() |
| paul@0 | 205 | |
| paul@0 | 206 | # Constant reference naming. |
| paul@0 | 207 | |
| paul@406 | 208 | def get_constant_name(self, value, value_type, encoding=None): |
| paul@0 | 209 | |
| paul@397 | 210 | """ |
| paul@397 | 211 | Add a new constant to the current namespace for 'value' with |
| paul@397 | 212 | 'value_type'. |
| paul@397 | 213 | """ |
| paul@0 | 214 | |
| paul@0 | 215 | path = self.get_namespace_path() |
| paul@0 | 216 | init_item(self.constants, path, dict) |
| paul@406 | 217 | return "$c%d" % add_counter_item(self.constants[path], (value, value_type, encoding)) |
| paul@0 | 218 | |
| paul@0 | 219 | # Literal reference naming. |
| paul@0 | 220 | |
| paul@0 | 221 | def get_literal_name(self): |
| paul@0 | 222 | |
| paul@0 | 223 | "Add a new literal to the current namespace." |
| paul@0 | 224 | |
| paul@0 | 225 | path = self.get_namespace_path() |
| paul@0 | 226 | init_item(self.literals, path, lambda: 0) |
| paul@0 | 227 | return "$C%d" % self.literals[path] |
| paul@0 | 228 | |
| paul@0 | 229 | def next_literal(self): |
| paul@0 | 230 | self.literals[self.get_namespace_path()] += 1 |
| paul@0 | 231 | |
| paul@0 | 232 | # Temporary variable naming. |
| paul@0 | 233 | |
| paul@0 | 234 | def get_temporary_name(self): |
| paul@0 | 235 | path = self.get_namespace_path() |
| paul@0 | 236 | init_item(self.temp, path, lambda: 0) |
| paul@0 | 237 | return "$t%d" % self.temp[path] |
| paul@0 | 238 | |
| paul@0 | 239 | def next_temporary(self): |
| paul@0 | 240 | self.temp[self.get_namespace_path()] += 1 |
| paul@0 | 241 | |
| paul@0 | 242 | # Arbitrary function naming. |
| paul@0 | 243 | |
| paul@0 | 244 | def get_lambda_name(self): |
| paul@0 | 245 | path = self.get_namespace_path() |
| paul@0 | 246 | init_item(self.lambdas, path, lambda: 0) |
| paul@0 | 247 | name = "$l%d" % self.lambdas[path] |
| paul@0 | 248 | self.lambdas[path] += 1 |
| paul@0 | 249 | return name |
| paul@0 | 250 | |
| paul@0 | 251 | def reset_lambdas(self): |
| paul@0 | 252 | self.lambdas = {} |
| paul@0 | 253 | |
| paul@0 | 254 | # Constant and literal recording. |
| paul@0 | 255 | |
| paul@537 | 256 | def get_constant_value(self, value, literals=None): |
| paul@394 | 257 | |
| paul@406 | 258 | """ |
| paul@406 | 259 | Encode the 'value' if appropriate, returning a value, a typename and any |
| paul@406 | 260 | encoding. |
| paul@406 | 261 | """ |
| paul@394 | 262 | |
| paul@394 | 263 | if isinstance(value, unicode): |
| paul@406 | 264 | return value.encode("utf-8"), "unicode", self.encoding |
| paul@405 | 265 | |
| paul@405 | 266 | # Attempt to convert plain strings to text. |
| paul@405 | 267 | |
| paul@405 | 268 | elif isinstance(value, str) and self.encoding: |
| paul@513 | 269 | try: |
| paul@537 | 270 | return get_string_details(literals, self.encoding) |
| paul@513 | 271 | except UnicodeDecodeError: |
| paul@513 | 272 | pass |
| paul@405 | 273 | |
| paul@406 | 274 | return value, value.__class__.__name__, None |
| paul@394 | 275 | |
| paul@406 | 276 | def get_constant_reference(self, ref, value, encoding=None): |
| paul@0 | 277 | |
| paul@406 | 278 | """ |
| paul@406 | 279 | Return a constant reference for the given 'ref' type and 'value', with |
| paul@406 | 280 | the optional 'encoding' applying to text values. |
| paul@406 | 281 | """ |
| paul@0 | 282 | |
| paul@406 | 283 | constant_name = self.get_constant_name(value, ref.get_origin(), encoding) |
| paul@0 | 284 | |
| paul@0 | 285 | # Return a reference for the constant. |
| paul@0 | 286 | |
| paul@0 | 287 | objpath = self.get_object_path(constant_name) |
| paul@338 | 288 | name_ref = ConstantValueRef(constant_name, ref.instance_of(objpath), value) |
| paul@0 | 289 | |
| paul@0 | 290 | # Record the value and type for the constant. |
| paul@0 | 291 | |
| paul@406 | 292 | self._reserve_constant(objpath, name_ref.value, name_ref.get_origin(), encoding) |
| paul@0 | 293 | return name_ref |
| paul@0 | 294 | |
| paul@406 | 295 | def reserve_constant(self, objpath, value, origin, encoding=None): |
| paul@251 | 296 | |
| paul@251 | 297 | """ |
| paul@251 | 298 | Reserve a constant within 'objpath' with the given 'value' and having a |
| paul@406 | 299 | type with the given 'origin', with the optional 'encoding' applying to |
| paul@406 | 300 | text values. |
| paul@251 | 301 | """ |
| paul@251 | 302 | |
| paul@397 | 303 | constant_name = self.get_constant_name(value, origin) |
| paul@251 | 304 | objpath = self.get_object_path(constant_name) |
| paul@406 | 305 | self._reserve_constant(objpath, value, origin, encoding) |
| paul@251 | 306 | |
| paul@406 | 307 | def _reserve_constant(self, objpath, value, origin, encoding): |
| paul@251 | 308 | |
| paul@406 | 309 | """ |
| paul@406 | 310 | Store a constant for 'objpath' with the given 'value' and 'origin', with |
| paul@406 | 311 | the optional 'encoding' applying to text values. |
| paul@406 | 312 | """ |
| paul@251 | 313 | |
| paul@406 | 314 | self.constant_values[objpath] = value, origin, encoding |
| paul@251 | 315 | |
| paul@0 | 316 | def get_literal_reference(self, name, ref, items, cls): |
| paul@0 | 317 | |
| paul@11 | 318 | """ |
| paul@11 | 319 | Return a literal reference for the given type 'name', literal 'ref', |
| paul@11 | 320 | node 'items' and employing the given 'cls' as the class of the returned |
| paul@11 | 321 | reference object. |
| paul@11 | 322 | """ |
| paul@11 | 323 | |
| paul@0 | 324 | # Construct an invocation using the items as arguments. |
| paul@0 | 325 | |
| paul@0 | 326 | typename = "$L%s" % name |
| paul@0 | 327 | |
| paul@0 | 328 | invocation = compiler.ast.CallFunc( |
| paul@0 | 329 | compiler.ast.Name(typename), |
| paul@0 | 330 | items |
| paul@0 | 331 | ) |
| paul@0 | 332 | |
| paul@0 | 333 | # Get a name for the actual literal. |
| paul@0 | 334 | |
| paul@0 | 335 | instname = self.get_literal_name() |
| paul@0 | 336 | self.next_literal() |
| paul@0 | 337 | |
| paul@0 | 338 | # Record the type for the literal. |
| paul@0 | 339 | |
| paul@0 | 340 | objpath = self.get_object_path(instname) |
| paul@0 | 341 | self.literal_types[objpath] = ref.get_origin() |
| paul@0 | 342 | |
| paul@0 | 343 | # Return a wrapper for the invocation exposing the items. |
| paul@0 | 344 | |
| paul@0 | 345 | return cls( |
| paul@0 | 346 | instname, |
| paul@0 | 347 | ref.instance_of(), |
| paul@0 | 348 | self.process_structure_node(invocation), |
| paul@0 | 349 | invocation.args |
| paul@0 | 350 | ) |
| paul@0 | 351 | |
| paul@0 | 352 | # Node handling. |
| paul@0 | 353 | |
| paul@0 | 354 | def process_structure(self, node): |
| paul@0 | 355 | |
| paul@0 | 356 | """ |
| paul@0 | 357 | Within the given 'node', process the program structure. |
| paul@0 | 358 | |
| paul@0 | 359 | During inspection, this will process global declarations, adjusting the |
| paul@0 | 360 | module namespace, and import statements, building a module dependency |
| paul@0 | 361 | hierarchy. |
| paul@0 | 362 | |
| paul@0 | 363 | During translation, this will consult deduced program information and |
| paul@0 | 364 | output translated code. |
| paul@0 | 365 | """ |
| paul@0 | 366 | |
| paul@0 | 367 | l = [] |
| paul@0 | 368 | for n in node.getChildNodes(): |
| paul@0 | 369 | l.append(self.process_structure_node(n)) |
| paul@0 | 370 | return l |
| paul@0 | 371 | |
| paul@0 | 372 | def process_augassign_node(self, n): |
| paul@0 | 373 | |
| paul@0 | 374 | "Process the given augmented assignment node 'n'." |
| paul@0 | 375 | |
| paul@0 | 376 | op = operator_functions[n.op] |
| paul@0 | 377 | |
| paul@0 | 378 | if isinstance(n.node, compiler.ast.Getattr): |
| paul@0 | 379 | target = compiler.ast.AssAttr(n.node.expr, n.node.attrname, "OP_ASSIGN") |
| paul@0 | 380 | elif isinstance(n.node, compiler.ast.Name): |
| paul@0 | 381 | target = compiler.ast.AssName(n.node.name, "OP_ASSIGN") |
| paul@0 | 382 | else: |
| paul@0 | 383 | target = n.node |
| paul@0 | 384 | |
| paul@0 | 385 | assignment = compiler.ast.Assign( |
| paul@0 | 386 | [target], |
| paul@0 | 387 | compiler.ast.CallFunc( |
| paul@0 | 388 | compiler.ast.Name("$op%s" % op), |
| paul@0 | 389 | [n.node, n.expr])) |
| paul@0 | 390 | |
| paul@0 | 391 | return self.process_structure_node(assignment) |
| paul@0 | 392 | |
| paul@320 | 393 | def process_assignment_for_object(self, original_name, source): |
| paul@0 | 394 | |
| paul@0 | 395 | """ |
| paul@0 | 396 | Return an assignment operation making 'original_name' refer to the given |
| paul@196 | 397 | 'source'. |
| paul@0 | 398 | """ |
| paul@0 | 399 | |
| paul@0 | 400 | assignment = compiler.ast.Assign( |
| paul@0 | 401 | [compiler.ast.AssName(original_name, "OP_ASSIGN")], |
| paul@196 | 402 | source |
| paul@0 | 403 | ) |
| paul@0 | 404 | |
| paul@0 | 405 | return self.process_structure_node(assignment) |
| paul@0 | 406 | |
| paul@0 | 407 | def process_assignment_node_items(self, n, expr): |
| paul@0 | 408 | |
| paul@0 | 409 | """ |
| paul@0 | 410 | Process the given assignment node 'n' whose children are to be assigned |
| paul@0 | 411 | items of 'expr'. |
| paul@0 | 412 | """ |
| paul@0 | 413 | |
| paul@0 | 414 | name_ref = self.process_structure_node(expr) |
| paul@0 | 415 | |
| paul@509 | 416 | # Either unpack the items and present them directly to each assignment |
| paul@509 | 417 | # node. |
| paul@509 | 418 | |
| paul@509 | 419 | if isinstance(name_ref, LiteralSequenceRef) and \ |
| paul@509 | 420 | self.process_literal_sequence_items(n, name_ref): |
| paul@0 | 421 | |
| paul@509 | 422 | pass |
| paul@509 | 423 | |
| paul@509 | 424 | # Or have the assignment nodes access each item via the sequence API. |
| paul@509 | 425 | |
| paul@509 | 426 | else: |
| paul@509 | 427 | self.process_assignment_node_items_by_position(n, expr, name_ref) |
| paul@0 | 428 | |
| paul@0 | 429 | def process_assignment_node_items_by_position(self, n, expr, name_ref): |
| paul@0 | 430 | |
| paul@0 | 431 | """ |
| paul@0 | 432 | Process the given sequence assignment node 'n', converting the node to |
| paul@0 | 433 | the separate assignment of each target using positional access on a |
| paul@0 | 434 | temporary variable representing the sequence. Use 'expr' as the assigned |
| paul@0 | 435 | value and 'name_ref' as the reference providing any existing temporary |
| paul@0 | 436 | variable. |
| paul@0 | 437 | """ |
| paul@0 | 438 | |
| paul@0 | 439 | assignments = [] |
| paul@0 | 440 | |
| paul@508 | 441 | # Employ existing names to access the sequence. |
| paul@508 | 442 | # Literal sequences do not provide names of accessible objects. |
| paul@508 | 443 | |
| paul@508 | 444 | if isinstance(name_ref, NameRef) and not isinstance(name_ref, LiteralSequenceRef): |
| paul@0 | 445 | temp = name_ref.name |
| paul@508 | 446 | |
| paul@508 | 447 | # For other expressions, create a temporary name to reference the items. |
| paul@508 | 448 | |
| paul@0 | 449 | else: |
| paul@0 | 450 | temp = self.get_temporary_name() |
| paul@0 | 451 | self.next_temporary() |
| paul@0 | 452 | |
| paul@0 | 453 | assignments.append( |
| paul@0 | 454 | compiler.ast.Assign([compiler.ast.AssName(temp, "OP_ASSIGN")], expr) |
| paul@0 | 455 | ) |
| paul@0 | 456 | |
| paul@508 | 457 | # Assign the items to the target nodes. |
| paul@508 | 458 | |
| paul@0 | 459 | for i, node in enumerate(n.nodes): |
| paul@0 | 460 | assignments.append( |
| paul@0 | 461 | compiler.ast.Assign([node], compiler.ast.Subscript( |
| paul@395 | 462 | compiler.ast.Name(temp), "OP_APPLY", [compiler.ast.Const(i, str(i))])) |
| paul@0 | 463 | ) |
| paul@0 | 464 | |
| paul@0 | 465 | return self.process_structure_node(compiler.ast.Stmt(assignments)) |
| paul@0 | 466 | |
| paul@0 | 467 | def process_literal_sequence_items(self, n, name_ref): |
| paul@0 | 468 | |
| paul@0 | 469 | """ |
| paul@0 | 470 | Process the given assignment node 'n', obtaining from the given |
| paul@0 | 471 | 'name_ref' the items to be assigned to the assignment targets. |
| paul@509 | 472 | |
| paul@509 | 473 | Return whether this method was able to process the assignment node as |
| paul@509 | 474 | a sequence of direct assignments. |
| paul@0 | 475 | """ |
| paul@0 | 476 | |
| paul@0 | 477 | if len(n.nodes) == len(name_ref.items): |
| paul@509 | 478 | assigned_names, count = get_names_from_nodes(n.nodes) |
| paul@509 | 479 | accessed_names, _count = get_names_from_nodes(name_ref.items) |
| paul@509 | 480 | |
| paul@509 | 481 | # Only assign directly between items if all assigned names are |
| paul@509 | 482 | # plain names (not attribute assignments), and if the assigned names |
| paul@509 | 483 | # do not appear in the accessed names. |
| paul@509 | 484 | |
| paul@509 | 485 | if len(assigned_names) == count and \ |
| paul@509 | 486 | not assigned_names.intersection(accessed_names): |
| paul@509 | 487 | |
| paul@509 | 488 | for node, item in zip(n.nodes, name_ref.items): |
| paul@509 | 489 | self.process_assignment_node(node, item) |
| paul@509 | 490 | |
| paul@509 | 491 | return True |
| paul@509 | 492 | |
| paul@509 | 493 | # Otherwise, use the position-based mechanism to obtain values. |
| paul@509 | 494 | |
| paul@509 | 495 | else: |
| paul@509 | 496 | return False |
| paul@0 | 497 | else: |
| paul@0 | 498 | raise InspectError("In %s, item assignment needing %d items is given %d items." % ( |
| paul@0 | 499 | self.get_namespace_path(), len(n.nodes), len(name_ref.items))) |
| paul@0 | 500 | |
| paul@0 | 501 | def process_compare_node(self, n): |
| paul@0 | 502 | |
| paul@0 | 503 | """ |
| paul@0 | 504 | Process the given comparison node 'n', converting an operator sequence |
| paul@0 | 505 | from... |
| paul@0 | 506 | |
| paul@0 | 507 | <expr1> <op1> <expr2> <op2> <expr3> |
| paul@0 | 508 | |
| paul@0 | 509 | ...to... |
| paul@0 | 510 | |
| paul@0 | 511 | <op1>(<expr1>, <expr2>) and <op2>(<expr2>, <expr3>) |
| paul@0 | 512 | """ |
| paul@0 | 513 | |
| paul@0 | 514 | invocations = [] |
| paul@0 | 515 | last = n.expr |
| paul@0 | 516 | |
| paul@0 | 517 | for op, op_node in n.ops: |
| paul@0 | 518 | op = operator_functions.get(op) |
| paul@0 | 519 | |
| paul@0 | 520 | invocations.append(compiler.ast.CallFunc( |
| paul@0 | 521 | compiler.ast.Name("$op%s" % op), |
| paul@0 | 522 | [last, op_node])) |
| paul@0 | 523 | |
| paul@0 | 524 | last = op_node |
| paul@0 | 525 | |
| paul@0 | 526 | if len(invocations) > 1: |
| paul@0 | 527 | result = compiler.ast.And(invocations) |
| paul@0 | 528 | else: |
| paul@0 | 529 | result = invocations[0] |
| paul@0 | 530 | |
| paul@0 | 531 | return self.process_structure_node(result) |
| paul@0 | 532 | |
| paul@0 | 533 | def process_dict_node(self, node): |
| paul@0 | 534 | |
| paul@0 | 535 | """ |
| paul@0 | 536 | Process the given dictionary 'node', returning a list of (key, value) |
| paul@0 | 537 | tuples. |
| paul@0 | 538 | """ |
| paul@0 | 539 | |
| paul@0 | 540 | l = [] |
| paul@0 | 541 | for key, value in node.items: |
| paul@0 | 542 | l.append(( |
| paul@0 | 543 | self.process_structure_node(key), |
| paul@0 | 544 | self.process_structure_node(value))) |
| paul@0 | 545 | return l |
| paul@0 | 546 | |
| paul@0 | 547 | def process_for_node(self, n): |
| paul@0 | 548 | |
| paul@0 | 549 | """ |
| paul@0 | 550 | Generate attribute accesses for {n.list}.__iter__ and the next method on |
| paul@0 | 551 | the iterator, producing a replacement node for the original. |
| paul@0 | 552 | """ |
| paul@0 | 553 | |
| paul@705 | 554 | t0 = self.get_temporary_name() |
| paul@705 | 555 | self.next_temporary() |
| paul@705 | 556 | t1 = self.get_temporary_name() |
| paul@705 | 557 | self.next_temporary() |
| paul@704 | 558 | i0 = self.get_temporary_name() |
| paul@704 | 559 | self.next_temporary() |
| paul@704 | 560 | |
| paul@0 | 561 | node = compiler.ast.Stmt([ |
| paul@0 | 562 | |
| paul@705 | 563 | # <t0> = {n.list} |
| paul@705 | 564 | # <t1> = <t0>.__iter__() |
| paul@705 | 565 | |
| paul@705 | 566 | compiler.ast.Assign( |
| paul@705 | 567 | [compiler.ast.AssName(t0, "OP_ASSIGN")], |
| paul@705 | 568 | n.list), |
| paul@705 | 569 | |
| paul@705 | 570 | compiler.ast.Assign( |
| paul@705 | 571 | [compiler.ast.AssName(t1, "OP_ASSIGN")], |
| paul@705 | 572 | compiler.ast.CallFunc( |
| paul@705 | 573 | compiler.ast.Getattr(compiler.ast.Name(t0), "__iter__"), |
| paul@705 | 574 | [])), |
| paul@0 | 575 | |
| paul@0 | 576 | # try: |
| paul@0 | 577 | # while True: |
| paul@737 | 578 | # <var>... = <t1>.next() |
| paul@0 | 579 | # ... |
| paul@0 | 580 | # except StopIteration: |
| paul@0 | 581 | # pass |
| paul@0 | 582 | |
| paul@0 | 583 | compiler.ast.TryExcept( |
| paul@0 | 584 | compiler.ast.While( |
| paul@0 | 585 | compiler.ast.Name("True"), |
| paul@0 | 586 | compiler.ast.Stmt([ |
| paul@0 | 587 | compiler.ast.Assign( |
| paul@0 | 588 | [n.assign], |
| paul@0 | 589 | compiler.ast.CallFunc( |
| paul@737 | 590 | compiler.ast.Getattr(compiler.ast.Name(t1), "next"), |
| paul@0 | 591 | [] |
| paul@0 | 592 | )), |
| paul@0 | 593 | n.body]), |
| paul@0 | 594 | None), |
| paul@0 | 595 | [(compiler.ast.Name("StopIteration"), None, compiler.ast.Stmt([compiler.ast.Pass()]))], |
| paul@0 | 596 | None) |
| paul@0 | 597 | ]) |
| paul@0 | 598 | |
| paul@0 | 599 | self.process_structure_node(node) |
| paul@0 | 600 | |
| paul@0 | 601 | def process_literal_sequence_node(self, n, name, ref, cls): |
| paul@0 | 602 | |
| paul@0 | 603 | """ |
| paul@0 | 604 | Process the given literal sequence node 'n' as a function invocation, |
| paul@0 | 605 | with 'name' indicating the type of the sequence, and 'ref' being a |
| paul@0 | 606 | reference to the type. The 'cls' is used to instantiate a suitable name |
| paul@0 | 607 | reference. |
| paul@0 | 608 | """ |
| paul@0 | 609 | |
| paul@0 | 610 | if name == "dict": |
| paul@0 | 611 | items = [] |
| paul@0 | 612 | for key, value in n.items: |
| paul@0 | 613 | items.append(compiler.ast.Tuple([key, value])) |
| paul@0 | 614 | else: # name in ("list", "tuple"): |
| paul@0 | 615 | items = n.nodes |
| paul@0 | 616 | |
| paul@0 | 617 | return self.get_literal_reference(name, ref, items, cls) |
| paul@0 | 618 | |
| paul@0 | 619 | def process_operator_node(self, n): |
| paul@0 | 620 | |
| paul@0 | 621 | """ |
| paul@0 | 622 | Process the given operator node 'n' as an operator function invocation. |
| paul@0 | 623 | """ |
| paul@0 | 624 | |
| paul@0 | 625 | op = operator_functions[n.__class__.__name__] |
| paul@0 | 626 | invocation = compiler.ast.CallFunc( |
| paul@0 | 627 | compiler.ast.Name("$op%s" % op), |
| paul@0 | 628 | list(n.getChildNodes()) |
| paul@0 | 629 | ) |
| paul@0 | 630 | return self.process_structure_node(invocation) |
| paul@0 | 631 | |
| paul@173 | 632 | def process_print_node(self, n): |
| paul@173 | 633 | |
| paul@173 | 634 | """ |
| paul@173 | 635 | Process the given print node 'n' as an invocation on a stream of the |
| paul@173 | 636 | form... |
| paul@173 | 637 | |
| paul@173 | 638 | $print(dest, args, nl) |
| paul@173 | 639 | |
| paul@173 | 640 | The special function name will be translated elsewhere. |
| paul@173 | 641 | """ |
| paul@173 | 642 | |
| paul@173 | 643 | nl = isinstance(n, compiler.ast.Printnl) |
| paul@173 | 644 | invocation = compiler.ast.CallFunc( |
| paul@173 | 645 | compiler.ast.Name("$print"), |
| paul@173 | 646 | [n.dest or compiler.ast.Name("None"), |
| paul@173 | 647 | compiler.ast.List(list(n.nodes)), |
| paul@359 | 648 | nl and compiler.ast.Name("True") or compiler.ast.Name("False")] |
| paul@173 | 649 | ) |
| paul@173 | 650 | return self.process_structure_node(invocation) |
| paul@173 | 651 | |
| paul@0 | 652 | def process_slice_node(self, n, expr=None): |
| paul@0 | 653 | |
| paul@0 | 654 | """ |
| paul@0 | 655 | Process the given slice node 'n' as an operator function invocation. |
| paul@0 | 656 | """ |
| paul@0 | 657 | |
| paul@548 | 658 | if n.flags == "OP_ASSIGN": op = "setslice" |
| paul@548 | 659 | elif n.flags == "OP_DELETE": op = "delslice" |
| paul@548 | 660 | else: op = "getslice" |
| paul@548 | 661 | |
| paul@0 | 662 | invocation = compiler.ast.CallFunc( |
| paul@0 | 663 | compiler.ast.Name("$op%s" % op), |
| paul@0 | 664 | [n.expr, n.lower or compiler.ast.Name("None"), n.upper or compiler.ast.Name("None")] + |
| paul@0 | 665 | (expr and [expr] or []) |
| paul@0 | 666 | ) |
| paul@548 | 667 | |
| paul@548 | 668 | # Fix parse tree structure. |
| paul@548 | 669 | |
| paul@548 | 670 | if op == "delslice": |
| paul@548 | 671 | invocation = compiler.ast.Discard(invocation) |
| paul@548 | 672 | |
| paul@0 | 673 | return self.process_structure_node(invocation) |
| paul@0 | 674 | |
| paul@0 | 675 | def process_sliceobj_node(self, n): |
| paul@0 | 676 | |
| paul@0 | 677 | """ |
| paul@0 | 678 | Process the given slice object node 'n' as a slice constructor. |
| paul@0 | 679 | """ |
| paul@0 | 680 | |
| paul@0 | 681 | op = "slice" |
| paul@0 | 682 | invocation = compiler.ast.CallFunc( |
| paul@0 | 683 | compiler.ast.Name("$op%s" % op), |
| paul@0 | 684 | n.nodes |
| paul@0 | 685 | ) |
| paul@0 | 686 | return self.process_structure_node(invocation) |
| paul@0 | 687 | |
| paul@0 | 688 | def process_subscript_node(self, n, expr=None): |
| paul@0 | 689 | |
| paul@0 | 690 | """ |
| paul@0 | 691 | Process the given subscript node 'n' as an operator function invocation. |
| paul@0 | 692 | """ |
| paul@0 | 693 | |
| paul@548 | 694 | if n.flags == "OP_ASSIGN": op = "setitem" |
| paul@548 | 695 | elif n.flags == "OP_DELETE": op = "delitem" |
| paul@548 | 696 | else: op = "getitem" |
| paul@548 | 697 | |
| paul@0 | 698 | invocation = compiler.ast.CallFunc( |
| paul@0 | 699 | compiler.ast.Name("$op%s" % op), |
| paul@0 | 700 | [n.expr] + list(n.subs) + (expr and [expr] or []) |
| paul@0 | 701 | ) |
| paul@548 | 702 | |
| paul@548 | 703 | # Fix parse tree structure. |
| paul@548 | 704 | |
| paul@548 | 705 | if op == "delitem": |
| paul@548 | 706 | invocation = compiler.ast.Discard(invocation) |
| paul@548 | 707 | |
| paul@0 | 708 | return self.process_structure_node(invocation) |
| paul@0 | 709 | |
| paul@0 | 710 | def process_attribute_chain(self, n): |
| paul@0 | 711 | |
| paul@0 | 712 | """ |
| paul@0 | 713 | Process the given attribute access node 'n'. Return a reference |
| paul@0 | 714 | describing the expression. |
| paul@0 | 715 | """ |
| paul@0 | 716 | |
| paul@0 | 717 | # AssAttr/Getattr are nested with the outermost access being the last |
| paul@0 | 718 | # access in any chain. |
| paul@0 | 719 | |
| paul@0 | 720 | self.attrs.insert(0, n.attrname) |
| paul@0 | 721 | attrs = self.attrs |
| paul@0 | 722 | |
| paul@0 | 723 | # Break attribute chains where non-access nodes are found. |
| paul@0 | 724 | |
| paul@0 | 725 | if not self.have_access_expression(n): |
| paul@110 | 726 | self.reset_attribute_chain() |
| paul@0 | 727 | |
| paul@0 | 728 | # Descend into the expression, extending backwards any existing chain, |
| paul@0 | 729 | # or building another for the expression. |
| paul@0 | 730 | |
| paul@0 | 731 | name_ref = self.process_structure_node(n.expr) |
| paul@0 | 732 | |
| paul@0 | 733 | # Restore chain information applying to this node. |
| paul@0 | 734 | |
| paul@110 | 735 | if not self.have_access_expression(n): |
| paul@110 | 736 | self.restore_attribute_chain(attrs) |
| paul@0 | 737 | |
| paul@0 | 738 | # Return immediately if the expression was another access and thus a |
| paul@0 | 739 | # continuation backwards along the chain. The above processing will |
| paul@0 | 740 | # have followed the chain all the way to its conclusion. |
| paul@0 | 741 | |
| paul@0 | 742 | if self.have_access_expression(n): |
| paul@0 | 743 | del self.attrs[0] |
| paul@0 | 744 | |
| paul@0 | 745 | return name_ref |
| paul@0 | 746 | |
| paul@124 | 747 | # Attribute chain handling. |
| paul@124 | 748 | |
| paul@110 | 749 | def reset_attribute_chain(self): |
| paul@110 | 750 | |
| paul@110 | 751 | "Reset the attribute chain for a subexpression of an attribute access." |
| paul@110 | 752 | |
| paul@110 | 753 | self.attrs = [] |
| paul@124 | 754 | self.chain_assignment.append(self.in_assignment) |
| paul@124 | 755 | self.chain_invocation.append(self.in_invocation) |
| paul@124 | 756 | self.in_assignment = None |
| paul@553 | 757 | self.in_invocation = None |
| paul@110 | 758 | |
| paul@110 | 759 | def restore_attribute_chain(self, attrs): |
| paul@110 | 760 | |
| paul@110 | 761 | "Restore the attribute chain for an attribute access." |
| paul@110 | 762 | |
| paul@110 | 763 | self.attrs = attrs |
| paul@124 | 764 | self.in_assignment = self.chain_assignment.pop() |
| paul@124 | 765 | self.in_invocation = self.chain_invocation.pop() |
| paul@110 | 766 | |
| paul@0 | 767 | def have_access_expression(self, node): |
| paul@0 | 768 | |
| paul@0 | 769 | "Return whether the expression associated with 'node' is Getattr." |
| paul@0 | 770 | |
| paul@0 | 771 | return isinstance(node.expr, compiler.ast.Getattr) |
| paul@0 | 772 | |
| paul@678 | 773 | def get_name_for_tracking(self, name, name_ref=None, is_global=False): |
| paul@0 | 774 | |
| paul@0 | 775 | """ |
| paul@0 | 776 | Return the name to be used for attribute usage observations involving |
| paul@603 | 777 | the given 'name' in the current namespace. |
| paul@603 | 778 | |
| paul@603 | 779 | If the name is being used outside a function, and if 'name_ref' is |
| paul@678 | 780 | given and indicates a global or if 'is_global' is specified as a true |
| paul@678 | 781 | value, a path featuring the name in the global namespace is returned. |
| paul@678 | 782 | Otherwise, a path computed using the current namespace and the given |
| paul@678 | 783 | name is returned. |
| paul@0 | 784 | |
| paul@0 | 785 | The intention of this method is to provide a suitably-qualified name |
| paul@0 | 786 | that can be tracked across namespaces. Where globals are being |
| paul@0 | 787 | referenced in class namespaces, they should be referenced using their |
| paul@0 | 788 | path within the module, not using a path within each class. |
| paul@0 | 789 | |
| paul@0 | 790 | It may not be possible to identify a global within a function at the |
| paul@0 | 791 | time of inspection (since a global may appear later in a file). |
| paul@0 | 792 | Consequently, globals are identified by their local name rather than |
| paul@0 | 793 | their module-qualified path. |
| paul@0 | 794 | """ |
| paul@0 | 795 | |
| paul@0 | 796 | # For functions, use the appropriate local names. |
| paul@0 | 797 | |
| paul@0 | 798 | if self.in_function: |
| paul@0 | 799 | return name |
| paul@0 | 800 | |
| paul@603 | 801 | # For global names outside functions, use a global name. |
| paul@597 | 802 | |
| paul@678 | 803 | elif is_global or name_ref and name_ref.is_global_name(): |
| paul@603 | 804 | return self.get_global_path(name) |
| paul@0 | 805 | |
| paul@152 | 806 | # Otherwise, establish a name in the current namespace. |
| paul@0 | 807 | |
| paul@0 | 808 | else: |
| paul@0 | 809 | return self.get_object_path(name) |
| paul@0 | 810 | |
| paul@0 | 811 | def get_path_for_access(self): |
| paul@0 | 812 | |
| paul@0 | 813 | "Outside functions, register accesses at the module level." |
| paul@0 | 814 | |
| paul@0 | 815 | if not self.in_function: |
| paul@0 | 816 | return self.name |
| paul@0 | 817 | else: |
| paul@0 | 818 | return self.get_namespace_path() |
| paul@0 | 819 | |
| paul@0 | 820 | def get_module_name(self, node): |
| paul@0 | 821 | |
| paul@0 | 822 | """ |
| paul@0 | 823 | Using the given From 'node' in this module, calculate any relative import |
| paul@0 | 824 | information, returning a tuple containing a module to import along with any |
| paul@0 | 825 | names to import based on the node's name information. |
| paul@0 | 826 | |
| paul@0 | 827 | Where the returned module is given as None, whole module imports should |
| paul@0 | 828 | be performed for the returned modules using the returned names. |
| paul@0 | 829 | """ |
| paul@0 | 830 | |
| paul@0 | 831 | # Absolute import. |
| paul@0 | 832 | |
| paul@0 | 833 | if node.level == 0: |
| paul@0 | 834 | return node.modname, node.names |
| paul@0 | 835 | |
| paul@0 | 836 | # Relative to an ancestor of this module. |
| paul@0 | 837 | |
| paul@0 | 838 | else: |
| paul@0 | 839 | path = self.name.split(".") |
| paul@0 | 840 | level = node.level |
| paul@0 | 841 | |
| paul@0 | 842 | # Relative imports treat package roots as submodules. |
| paul@0 | 843 | |
| paul@0 | 844 | if split(self.filename)[-1] == "__init__.py": |
| paul@0 | 845 | level -= 1 |
| paul@0 | 846 | |
| paul@0 | 847 | if level > len(path): |
| paul@0 | 848 | raise InspectError("Relative import %r involves too many levels up from module %r" % ( |
| paul@0 | 849 | ("%s%s" % ("." * node.level, node.modname or "")), self.name)) |
| paul@0 | 850 | |
| paul@0 | 851 | basename = ".".join(path[:len(path)-level]) |
| paul@0 | 852 | |
| paul@0 | 853 | # Name imports from a module. |
| paul@0 | 854 | |
| paul@0 | 855 | if node.modname: |
| paul@0 | 856 | return "%s.%s" % (basename, node.modname), node.names |
| paul@0 | 857 | |
| paul@0 | 858 | # Relative whole module imports. |
| paul@0 | 859 | |
| paul@0 | 860 | else: |
| paul@0 | 861 | return basename, node.names |
| paul@0 | 862 | |
| paul@0 | 863 | def get_argnames(args): |
| paul@0 | 864 | |
| paul@0 | 865 | """ |
| paul@0 | 866 | Return a list of all names provided by 'args'. Since tuples may be |
| paul@0 | 867 | employed, the arguments are traversed depth-first. |
| paul@0 | 868 | """ |
| paul@0 | 869 | |
| paul@0 | 870 | l = [] |
| paul@0 | 871 | for arg in args: |
| paul@0 | 872 | if isinstance(arg, tuple): |
| paul@0 | 873 | l += get_argnames(arg) |
| paul@0 | 874 | else: |
| paul@0 | 875 | l.append(arg) |
| paul@0 | 876 | return l |
| paul@0 | 877 | |
| paul@509 | 878 | def get_names_from_nodes(nodes): |
| paul@509 | 879 | |
| paul@509 | 880 | """ |
| paul@509 | 881 | Return the names employed in the given 'nodes' along with the number of |
| paul@509 | 882 | nodes excluding sequences. |
| paul@509 | 883 | """ |
| paul@509 | 884 | |
| paul@509 | 885 | names = set() |
| paul@509 | 886 | count = 0 |
| paul@509 | 887 | |
| paul@509 | 888 | for node in nodes: |
| paul@509 | 889 | |
| paul@509 | 890 | # Add names and count them. |
| paul@509 | 891 | |
| paul@509 | 892 | if isinstance(node, (compiler.ast.AssName, compiler.ast.Name)): |
| paul@509 | 893 | names.add(node.name) |
| paul@509 | 894 | count += 1 |
| paul@509 | 895 | |
| paul@509 | 896 | # Add names from sequences and incorporate their counts. |
| paul@509 | 897 | |
| paul@509 | 898 | elif isinstance(node, (compiler.ast.AssList, compiler.ast.AssTuple, |
| paul@509 | 899 | compiler.ast.List, compiler.ast.Set, |
| paul@509 | 900 | compiler.ast.Tuple)): |
| paul@509 | 901 | _names, _count = get_names_from_nodes(node.nodes) |
| paul@509 | 902 | names.update(_names) |
| paul@509 | 903 | count += _count |
| paul@509 | 904 | |
| paul@509 | 905 | # Count non-name, non-sequence nodes. |
| paul@509 | 906 | |
| paul@509 | 907 | else: |
| paul@509 | 908 | count += 1 |
| paul@509 | 909 | |
| paul@509 | 910 | return names, count |
| paul@509 | 911 | |
| paul@491 | 912 | # Result classes. |
| paul@491 | 913 | |
| paul@491 | 914 | class InstructionSequence: |
| paul@491 | 915 | |
| paul@491 | 916 | "A generic sequence of instructions." |
| paul@491 | 917 | |
| paul@491 | 918 | def __init__(self, instructions): |
| paul@491 | 919 | self.instructions = instructions |
| paul@491 | 920 | |
| paul@491 | 921 | def get_value_instruction(self): |
| paul@491 | 922 | return self.instructions[-1] |
| paul@491 | 923 | |
| paul@491 | 924 | def get_init_instructions(self): |
| paul@491 | 925 | return self.instructions[:-1] |
| paul@491 | 926 | |
| paul@0 | 927 | # Dictionary utilities. |
| paul@0 | 928 | |
| paul@0 | 929 | def init_item(d, key, fn): |
| paul@0 | 930 | |
| paul@0 | 931 | """ |
| paul@0 | 932 | Add to 'd' an entry for 'key' using the callable 'fn' to make an initial |
| paul@0 | 933 | value where no entry already exists. |
| paul@0 | 934 | """ |
| paul@0 | 935 | |
| paul@0 | 936 | if not d.has_key(key): |
| paul@0 | 937 | d[key] = fn() |
| paul@0 | 938 | return d[key] |
| paul@0 | 939 | |
| paul@0 | 940 | def dict_for_keys(d, keys): |
| paul@0 | 941 | |
| paul@0 | 942 | "Return a new dictionary containing entries from 'd' for the given 'keys'." |
| paul@0 | 943 | |
| paul@0 | 944 | nd = {} |
| paul@0 | 945 | for key in keys: |
| paul@0 | 946 | if d.has_key(key): |
| paul@0 | 947 | nd[key] = d[key] |
| paul@0 | 948 | return nd |
| paul@0 | 949 | |
| paul@0 | 950 | def make_key(s): |
| paul@0 | 951 | |
| paul@0 | 952 | "Make sequence 's' into a tuple-based key, first sorting its contents." |
| paul@0 | 953 | |
| paul@0 | 954 | l = list(s) |
| paul@0 | 955 | l.sort() |
| paul@0 | 956 | return tuple(l) |
| paul@0 | 957 | |
| paul@0 | 958 | def add_counter_item(d, key): |
| paul@0 | 959 | |
| paul@0 | 960 | """ |
| paul@0 | 961 | Make a mapping in 'd' for 'key' to the number of keys added before it, thus |
| paul@0 | 962 | maintaining a mapping of keys to their order of insertion. |
| paul@0 | 963 | """ |
| paul@0 | 964 | |
| paul@0 | 965 | if not d.has_key(key): |
| paul@0 | 966 | d[key] = len(d.keys()) |
| paul@0 | 967 | return d[key] |
| paul@0 | 968 | |
| paul@0 | 969 | def remove_items(d1, d2): |
| paul@0 | 970 | |
| paul@0 | 971 | "Remove from 'd1' all items from 'd2'." |
| paul@0 | 972 | |
| paul@0 | 973 | for key in d2.keys(): |
| paul@0 | 974 | if d1.has_key(key): |
| paul@0 | 975 | del d1[key] |
| paul@0 | 976 | |
| paul@0 | 977 | # Set utilities. |
| paul@0 | 978 | |
| paul@0 | 979 | def first(s): |
| paul@0 | 980 | return list(s)[0] |
| paul@0 | 981 | |
| paul@0 | 982 | def same(s1, s2): |
| paul@0 | 983 | return set(s1) == set(s2) |
| paul@0 | 984 | |
| paul@724 | 985 | def order_dependencies(all_depends): |
| paul@724 | 986 | |
| paul@724 | 987 | """ |
| paul@724 | 988 | Produce a dependency ordering for the 'all_depends' mapping. This mapping |
| paul@724 | 989 | has the form "A depends on B, C...". The result will order A, B, C, and so |
| paul@724 | 990 | on. |
| paul@724 | 991 | """ |
| paul@724 | 992 | |
| paul@726 | 993 | usage = init_reverse_dependencies(all_depends) |
| paul@726 | 994 | |
| paul@726 | 995 | # Produce an ordering by obtaining exposed items (required by items already |
| paul@726 | 996 | # processed) and putting them at the start of the list. |
| paul@726 | 997 | |
| paul@726 | 998 | ordered = [] |
| paul@726 | 999 | |
| paul@726 | 1000 | while usage: |
| paul@726 | 1001 | have_next = False |
| paul@726 | 1002 | |
| paul@726 | 1003 | for key, n in usage.items(): |
| paul@726 | 1004 | |
| paul@726 | 1005 | # Add items needed by no other items to the ordering. |
| paul@726 | 1006 | |
| paul@726 | 1007 | if not n: |
| paul@726 | 1008 | remove_dependency(key, all_depends, usage, ordered) |
| paul@726 | 1009 | have_next = True |
| paul@726 | 1010 | |
| paul@726 | 1011 | if not have_next: |
| paul@726 | 1012 | raise ValueError, usage |
| paul@726 | 1013 | |
| paul@726 | 1014 | return ordered |
| paul@726 | 1015 | |
| paul@726 | 1016 | def order_dependencies_partial(all_depends): |
| paul@726 | 1017 | |
| paul@726 | 1018 | """ |
| paul@726 | 1019 | Produce a dependency ordering for the 'all_depends' mapping. This mapping |
| paul@726 | 1020 | has the form "A depends on B, C...". The result will order A, B, C, and so |
| paul@726 | 1021 | on. Where cycles exist, they will be broken and a partial ordering returned. |
| paul@726 | 1022 | """ |
| paul@726 | 1023 | |
| paul@726 | 1024 | usage = init_reverse_dependencies(all_depends) |
| paul@726 | 1025 | |
| paul@726 | 1026 | # Duplicate the dependencies for subsequent modification. |
| paul@726 | 1027 | |
| paul@726 | 1028 | new_depends = {} |
| paul@726 | 1029 | for key, values in all_depends.items(): |
| paul@726 | 1030 | new_depends[key] = set(values) |
| paul@726 | 1031 | |
| paul@726 | 1032 | all_depends = new_depends |
| paul@726 | 1033 | |
| paul@726 | 1034 | # Produce an ordering by obtaining exposed items (required by items already |
| paul@726 | 1035 | # processed) and putting them at the start of the list. |
| paul@726 | 1036 | |
| paul@726 | 1037 | ordered = [] |
| paul@726 | 1038 | |
| paul@726 | 1039 | while usage: |
| paul@726 | 1040 | least = None |
| paul@726 | 1041 | least_key = None |
| paul@726 | 1042 | |
| paul@726 | 1043 | for key, n in usage.items(): |
| paul@726 | 1044 | |
| paul@726 | 1045 | # Add items needed by no other items to the ordering. |
| paul@726 | 1046 | |
| paul@726 | 1047 | if not n: |
| paul@726 | 1048 | remove_dependency(key, all_depends, usage, ordered) |
| paul@726 | 1049 | least = 0 |
| paul@726 | 1050 | |
| paul@726 | 1051 | # When breaking cycles, note the least used items. |
| paul@726 | 1052 | |
| paul@726 | 1053 | elif least is None or len(n) < least: |
| paul@726 | 1054 | least_key = key |
| paul@726 | 1055 | least = len(n) |
| paul@726 | 1056 | |
| paul@726 | 1057 | if least: |
| paul@726 | 1058 | transfer_dependencies(least_key, all_depends, usage, ordered) |
| paul@726 | 1059 | |
| paul@726 | 1060 | return ordered |
| paul@726 | 1061 | |
| paul@726 | 1062 | def init_reverse_dependencies(all_depends): |
| paul@726 | 1063 | |
| paul@726 | 1064 | """ |
| paul@726 | 1065 | From 'all_depends', providing a mapping of the form "A depends on B, C...", |
| paul@726 | 1066 | record the reverse dependencies, making a mapping of the form |
| paul@726 | 1067 | "B is needed by A", "C is needed by A", and so on. |
| paul@726 | 1068 | """ |
| paul@724 | 1069 | |
| paul@724 | 1070 | usage = {} |
| paul@724 | 1071 | |
| paul@724 | 1072 | # Record path-based dependencies. |
| paul@724 | 1073 | |
| paul@724 | 1074 | for key in all_depends.keys(): |
| paul@724 | 1075 | usage[key] = set() |
| paul@724 | 1076 | |
| paul@724 | 1077 | for key, depends in all_depends.items(): |
| paul@724 | 1078 | for depend in depends: |
| paul@724 | 1079 | init_item(usage, depend, set) |
| paul@724 | 1080 | usage[depend].add(key) |
| paul@724 | 1081 | |
| paul@726 | 1082 | return usage |
| paul@726 | 1083 | |
| paul@726 | 1084 | def transfer_dependencies(key, all_depends, usage, ordered): |
| paul@726 | 1085 | |
| paul@726 | 1086 | """ |
| paul@726 | 1087 | Transfer items needed by 'key' to those items needing 'key', found using |
| paul@726 | 1088 | 'all_depends', and updating 'usage'. Insert 'key' into the 'ordered' |
| paul@726 | 1089 | collection of dependencies. |
| paul@724 | 1090 | |
| paul@726 | 1091 | If "A is needed by X" and "B is needed by A", then transferring items needed |
| paul@726 | 1092 | by A will cause "B is needed by X" to be recorded as a consequence. |
| paul@726 | 1093 | |
| paul@726 | 1094 | Transferring items also needs to occur in the reverse mapping, so that |
| paul@726 | 1095 | "A needs B" and "X needs A", then the consequence must be recorded as |
| paul@726 | 1096 | "X needs B". |
| paul@726 | 1097 | """ |
| paul@726 | 1098 | |
| paul@726 | 1099 | ordered.insert(0, key) |
| paul@724 | 1100 | |
| paul@726 | 1101 | needing = usage[key] # A is needed by X |
| paul@726 | 1102 | needed = all_depends.get(key) # A needs B |
| paul@726 | 1103 | |
| paul@726 | 1104 | if needing: |
| paul@726 | 1105 | for depend in needing: |
| paul@726 | 1106 | l = all_depends.get(depend) |
| paul@726 | 1107 | if not l: |
| paul@726 | 1108 | continue |
| paul@724 | 1109 | |
| paul@726 | 1110 | l.remove(key) # X needs (A) |
| paul@726 | 1111 | |
| paul@726 | 1112 | if needed: |
| paul@726 | 1113 | l.update(needed) # X needs B... |
| paul@726 | 1114 | |
| paul@726 | 1115 | # Prevent self references. |
| paul@726 | 1116 | |
| paul@726 | 1117 | if depend in needed: |
| paul@726 | 1118 | l.remove(depend) |
| paul@724 | 1119 | |
| paul@726 | 1120 | if needed: |
| paul@726 | 1121 | for depend in needed: |
| paul@726 | 1122 | l = usage.get(depend) |
| paul@726 | 1123 | if not l: |
| paul@726 | 1124 | continue |
| paul@726 | 1125 | |
| paul@726 | 1126 | l.remove(key) # B is needed by (A) |
| paul@726 | 1127 | l.update(needing) # B is needed by X... |
| paul@724 | 1128 | |
| paul@726 | 1129 | # Prevent self references. |
| paul@726 | 1130 | |
| paul@726 | 1131 | if depend in needing: |
| paul@726 | 1132 | l.remove(depend) |
| paul@726 | 1133 | |
| paul@726 | 1134 | if needed: |
| paul@726 | 1135 | del all_depends[key] |
| paul@726 | 1136 | del usage[key] |
| paul@726 | 1137 | |
| paul@726 | 1138 | def remove_dependency(key, all_depends, usage, ordered): |
| paul@724 | 1139 | |
| paul@726 | 1140 | """ |
| paul@726 | 1141 | Remove 'key', found in 'all_depends', from 'usage', inserting it into the |
| paul@726 | 1142 | 'ordered' collection of dependencies. |
| paul@726 | 1143 | |
| paul@726 | 1144 | Given that 'usage' for a given key A would indicate that "A needs <nothing>" |
| paul@726 | 1145 | upon removing A from 'usage', the outcome is that all keys needing A will |
| paul@726 | 1146 | have A removed from their 'usage' records. |
| paul@726 | 1147 | |
| paul@726 | 1148 | So, if "B needs A", removing A will cause "B needs <nothing>" to be recorded |
| paul@726 | 1149 | as a consequence. |
| paul@726 | 1150 | """ |
| paul@724 | 1151 | |
| paul@726 | 1152 | ordered.insert(0, key) |
| paul@726 | 1153 | |
| paul@726 | 1154 | depends = all_depends.get(key) |
| paul@726 | 1155 | |
| paul@726 | 1156 | # Reduce usage of the referenced items. |
| paul@724 | 1157 | |
| paul@726 | 1158 | if depends: |
| paul@726 | 1159 | for depend in depends: |
| paul@726 | 1160 | usage[depend].remove(key) |
| paul@726 | 1161 | |
| paul@726 | 1162 | del usage[key] |
| paul@724 | 1163 | |
| paul@0 | 1164 | # General input/output. |
| paul@0 | 1165 | |
| paul@0 | 1166 | def readfile(filename): |
| paul@0 | 1167 | |
| paul@0 | 1168 | "Return the contents of 'filename'." |
| paul@0 | 1169 | |
| paul@0 | 1170 | f = open(filename) |
| paul@0 | 1171 | try: |
| paul@0 | 1172 | return f.read() |
| paul@0 | 1173 | finally: |
| paul@0 | 1174 | f.close() |
| paul@0 | 1175 | |
| paul@0 | 1176 | def writefile(filename, s): |
| paul@0 | 1177 | |
| paul@0 | 1178 | "Write to 'filename' the string 's'." |
| paul@0 | 1179 | |
| paul@0 | 1180 | f = open(filename, "w") |
| paul@0 | 1181 | try: |
| paul@0 | 1182 | f.write(s) |
| paul@0 | 1183 | finally: |
| paul@0 | 1184 | f.close() |
| paul@0 | 1185 | |
| paul@0 | 1186 | # General encoding. |
| paul@0 | 1187 | |
| paul@0 | 1188 | def sorted_output(x): |
| paul@0 | 1189 | |
| paul@0 | 1190 | "Sort sequence 'x' and return a string with commas separating the values." |
| paul@0 | 1191 | |
| paul@0 | 1192 | x = map(str, x) |
| paul@0 | 1193 | x.sort() |
| paul@0 | 1194 | return ", ".join(x) |
| paul@0 | 1195 | |
| paul@537 | 1196 | def get_string_details(literals, encoding): |
| paul@512 | 1197 | |
| paul@512 | 1198 | """ |
| paul@537 | 1199 | Determine whether 'literals' represent Unicode strings or byte strings, |
| paul@537 | 1200 | using 'encoding' to reproduce byte sequences. |
| paul@537 | 1201 | |
| paul@537 | 1202 | Each literal is the full program representation including prefix and quotes |
| paul@537 | 1203 | recoded by the parser to UTF-8. Thus, any literal found to represent a byte |
| paul@537 | 1204 | string needs to be translated back to its original encoding. |
| paul@537 | 1205 | |
| paul@537 | 1206 | Return a single encoded literal value, a type name, and the original |
| paul@537 | 1207 | encoding as a tuple. |
| paul@537 | 1208 | """ |
| paul@537 | 1209 | |
| paul@537 | 1210 | typename = "unicode" |
| paul@537 | 1211 | |
| paul@537 | 1212 | l = [] |
| paul@537 | 1213 | |
| paul@537 | 1214 | for s in literals: |
| paul@537 | 1215 | out, _typename = get_literal_details(s) |
| paul@537 | 1216 | if _typename == "str": |
| paul@537 | 1217 | typename = "str" |
| paul@537 | 1218 | l.append(out) |
| paul@537 | 1219 | |
| paul@537 | 1220 | out = "".join(l) |
| paul@537 | 1221 | |
| paul@537 | 1222 | # For Unicode values, convert to the UTF-8 program representation. |
| paul@537 | 1223 | |
| paul@537 | 1224 | if typename == "unicode": |
| paul@537 | 1225 | return out.encode("utf-8"), typename, encoding |
| paul@537 | 1226 | |
| paul@537 | 1227 | # For byte string values, convert back to the original encoding. |
| paul@537 | 1228 | |
| paul@537 | 1229 | else: |
| paul@537 | 1230 | return out.encode(encoding), typename, encoding |
| paul@537 | 1231 | |
| paul@537 | 1232 | def get_literal_details(s): |
| paul@537 | 1233 | |
| paul@537 | 1234 | """ |
| paul@537 | 1235 | Determine whether 's' represents a Unicode string or a byte string, where |
| paul@537 | 1236 | 's' contains the full program representation of a literal including prefix |
| paul@537 | 1237 | and quotes, recoded by the parser to UTF-8. |
| paul@512 | 1238 | |
| paul@512 | 1239 | Find and convert Unicode values starting with <backslash>u or <backslash>U, |
| paul@512 | 1240 | and byte or Unicode values starting with <backslash><octal digit> or |
| paul@512 | 1241 | <backslash>x. |
| paul@512 | 1242 | |
| paul@512 | 1243 | Literals prefixed with "u" cause <backslash><octal digit> and <backslash>x |
| paul@512 | 1244 | to be considered as Unicode values. Otherwise, they produce byte values and |
| paul@512 | 1245 | cause unprefixed strings to be considered as byte strings. |
| paul@512 | 1246 | |
| paul@512 | 1247 | Literals prefixed with "r" do not have their backslash-encoded values |
| paul@512 | 1248 | converted unless also prefixed with "u", in which case only the above value |
| paul@512 | 1249 | formats are converted, not any of the other special sequences for things |
| paul@512 | 1250 | like newlines. |
| paul@512 | 1251 | |
| paul@537 | 1252 | Return the literal value as a Unicode object together with the appropriate |
| paul@537 | 1253 | type name in a tuple. |
| paul@512 | 1254 | """ |
| paul@512 | 1255 | |
| paul@512 | 1256 | l = [] |
| paul@512 | 1257 | |
| paul@512 | 1258 | # Identify the quote character and use it to identify the prefix. |
| paul@512 | 1259 | |
| paul@512 | 1260 | quote_type = s[-1] |
| paul@512 | 1261 | prefix_end = s.find(quote_type) |
| paul@512 | 1262 | prefix = s[:prefix_end].lower() |
| paul@512 | 1263 | |
| paul@512 | 1264 | if prefix not in ("", "b", "br", "r", "u", "ur"): |
| paul@512 | 1265 | raise ValueError, "String literal does not have a supported prefix: %s" % s |
| paul@512 | 1266 | |
| paul@513 | 1267 | if "b" in prefix: |
| paul@513 | 1268 | typename = "str" |
| paul@513 | 1269 | else: |
| paul@513 | 1270 | typename = "unicode" |
| paul@513 | 1271 | |
| paul@512 | 1272 | # Identify triple quotes or single quotes. |
| paul@512 | 1273 | |
| paul@512 | 1274 | if len(s) >= 6 and s[-2] == quote_type and s[-3] == quote_type: |
| paul@512 | 1275 | quote = s[prefix_end:prefix_end+3] |
| paul@512 | 1276 | current = prefix_end + 3 |
| paul@512 | 1277 | end = len(s) - 3 |
| paul@512 | 1278 | else: |
| paul@512 | 1279 | quote = s[prefix_end] |
| paul@512 | 1280 | current = prefix_end + 1 |
| paul@512 | 1281 | end = len(s) - 1 |
| paul@512 | 1282 | |
| paul@512 | 1283 | # Conversions of some quoted values. |
| paul@512 | 1284 | |
| paul@512 | 1285 | searches = { |
| paul@512 | 1286 | "u" : (6, 16), |
| paul@512 | 1287 | "U" : (10, 16), |
| paul@512 | 1288 | "x" : (4, 16), |
| paul@512 | 1289 | } |
| paul@512 | 1290 | |
| paul@512 | 1291 | octal_digits = map(str, range(0, 8)) |
| paul@512 | 1292 | |
| paul@512 | 1293 | # Translations of some quoted values. |
| paul@512 | 1294 | |
| paul@512 | 1295 | escaped = { |
| paul@512 | 1296 | "\\" : "\\", "'" : "'", '"' : '"', |
| paul@512 | 1297 | "a" : "\a", "b" : "\b", "f" : "\f", |
| paul@512 | 1298 | "n" : "\n", "r" : "\r", "t" : "\t", |
| paul@512 | 1299 | } |
| paul@512 | 1300 | |
| paul@512 | 1301 | while current < end: |
| paul@512 | 1302 | |
| paul@512 | 1303 | # Look for quoted values. |
| paul@512 | 1304 | |
| paul@512 | 1305 | index = s.find("\\", current) |
| paul@512 | 1306 | if index == -1 or index + 1 == end: |
| paul@512 | 1307 | l.append(s[current:end]) |
| paul@512 | 1308 | break |
| paul@512 | 1309 | |
| paul@512 | 1310 | # Add the preceding text. |
| paul@512 | 1311 | |
| paul@512 | 1312 | l.append(s[current:index]) |
| paul@512 | 1313 | |
| paul@512 | 1314 | # Handle quoted text. |
| paul@512 | 1315 | |
| paul@512 | 1316 | term = s[index+1] |
| paul@512 | 1317 | |
| paul@512 | 1318 | # Add Unicode values. Where a string is u-prefixed, even \o and \x |
| paul@512 | 1319 | # produce Unicode values. |
| paul@512 | 1320 | |
| paul@513 | 1321 | if typename == "unicode" and ( |
| paul@513 | 1322 | term in ("u", "U") or |
| paul@513 | 1323 | "u" in prefix and (term == "x" or term in octal_digits)): |
| paul@512 | 1324 | |
| paul@512 | 1325 | needed, base = searches.get(term, (4, 8)) |
| paul@512 | 1326 | value = convert_quoted_value(s, index, needed, end, base, unichr) |
| paul@512 | 1327 | l.append(value) |
| paul@512 | 1328 | current = index + needed |
| paul@512 | 1329 | |
| paul@512 | 1330 | # Add raw byte values, changing the string type. |
| paul@512 | 1331 | |
| paul@512 | 1332 | elif "r" not in prefix and ( |
| paul@512 | 1333 | term == "x" or term in octal_digits): |
| paul@512 | 1334 | |
| paul@512 | 1335 | needed, base = searches.get(term, (4, 8)) |
| paul@512 | 1336 | value = convert_quoted_value(s, index, needed, end, base, chr) |
| paul@512 | 1337 | l.append(value) |
| paul@512 | 1338 | typename = "str" |
| paul@512 | 1339 | current = index + needed |
| paul@512 | 1340 | |
| paul@512 | 1341 | # Add other escaped values. |
| paul@512 | 1342 | |
| paul@512 | 1343 | elif "r" not in prefix and escaped.has_key(term): |
| paul@512 | 1344 | l.append(escaped[term]) |
| paul@512 | 1345 | current = index + 2 |
| paul@512 | 1346 | |
| paul@512 | 1347 | # Add other text as found. |
| paul@512 | 1348 | |
| paul@512 | 1349 | else: |
| paul@512 | 1350 | l.append(s[index:index+2]) |
| paul@512 | 1351 | current = index + 2 |
| paul@512 | 1352 | |
| paul@537 | 1353 | # Collect the components into a single Unicode object. Since the literal |
| paul@537 | 1354 | # text was already in UTF-8 form, interpret plain strings as UTF-8 |
| paul@537 | 1355 | # sequences. |
| paul@512 | 1356 | |
| paul@537 | 1357 | out = [] |
| paul@512 | 1358 | |
| paul@537 | 1359 | for value in l: |
| paul@537 | 1360 | if isinstance(value, unicode): |
| paul@537 | 1361 | out.append(value) |
| paul@537 | 1362 | else: |
| paul@537 | 1363 | out.append(unicode(value, "utf-8")) |
| paul@512 | 1364 | |
| paul@537 | 1365 | return "".join(out), typename |
| paul@512 | 1366 | |
| paul@512 | 1367 | def convert_quoted_value(s, index, needed, end, base, fn): |
| paul@512 | 1368 | |
| paul@512 | 1369 | """ |
| paul@512 | 1370 | Interpret a quoted value in 's' at 'index' with the given 'needed' number of |
| paul@512 | 1371 | positions, and with the given 'end' indicating the first position after the |
| paul@512 | 1372 | end of the actual string content. |
| paul@512 | 1373 | |
| paul@512 | 1374 | Use 'base' as the numerical base when interpreting the value, and use 'fn' |
| paul@512 | 1375 | to convert the value to an appropriate type. |
| paul@512 | 1376 | """ |
| paul@512 | 1377 | |
| paul@512 | 1378 | s = s[index:min(index+needed, end)] |
| paul@512 | 1379 | |
| paul@512 | 1380 | # Not a complete occurrence. |
| paul@512 | 1381 | |
| paul@512 | 1382 | if len(s) < needed: |
| paul@512 | 1383 | return s |
| paul@512 | 1384 | |
| paul@512 | 1385 | # Test for a well-formed value. |
| paul@512 | 1386 | |
| paul@512 | 1387 | try: |
| paul@512 | 1388 | first = base == 8 and 1 or 2 |
| paul@512 | 1389 | value = int(s[first:needed], base) |
| paul@512 | 1390 | except ValueError: |
| paul@512 | 1391 | return s |
| paul@512 | 1392 | else: |
| paul@512 | 1393 | return fn(value) |
| paul@512 | 1394 | |
| paul@0 | 1395 | # Attribute chain decoding. |
| paul@0 | 1396 | |
| paul@0 | 1397 | def get_attrnames(attrnames): |
| paul@11 | 1398 | |
| paul@11 | 1399 | """ |
| paul@11 | 1400 | Split the qualified attribute chain 'attrnames' into its components, |
| paul@11 | 1401 | handling special attributes starting with "#" that indicate type |
| paul@11 | 1402 | conformance. |
| paul@11 | 1403 | """ |
| paul@11 | 1404 | |
| paul@0 | 1405 | if attrnames.startswith("#"): |
| paul@0 | 1406 | return [attrnames] |
| paul@0 | 1407 | else: |
| paul@0 | 1408 | return attrnames.split(".") |
| paul@0 | 1409 | |
| paul@0 | 1410 | def get_attrname_from_location(location): |
| paul@11 | 1411 | |
| paul@11 | 1412 | """ |
| paul@11 | 1413 | Extract the first attribute from the attribute names employed in a |
| paul@11 | 1414 | 'location'. |
| paul@11 | 1415 | """ |
| paul@11 | 1416 | |
| paul@0 | 1417 | path, name, attrnames, access = location |
| paul@91 | 1418 | if not attrnames: |
| paul@91 | 1419 | return attrnames |
| paul@0 | 1420 | return get_attrnames(attrnames)[0] |
| paul@0 | 1421 | |
| paul@85 | 1422 | def get_name_path(path, name): |
| paul@85 | 1423 | |
| paul@85 | 1424 | "Return a suitable qualified name from the given 'path' and 'name'." |
| paul@85 | 1425 | |
| paul@85 | 1426 | if "." in name: |
| paul@85 | 1427 | return name |
| paul@85 | 1428 | else: |
| paul@85 | 1429 | return "%s.%s" % (path, name) |
| paul@85 | 1430 | |
| paul@90 | 1431 | # Usage-related functions. |
| paul@89 | 1432 | |
| paul@89 | 1433 | def get_types_for_usage(attrnames, objects): |
| paul@89 | 1434 | |
| paul@89 | 1435 | """ |
| paul@89 | 1436 | Identify the types that can support the given 'attrnames', using the |
| paul@89 | 1437 | given 'objects' as the catalogue of type details. |
| paul@89 | 1438 | """ |
| paul@89 | 1439 | |
| paul@89 | 1440 | types = [] |
| paul@89 | 1441 | for name, _attrnames in objects.items(): |
| paul@89 | 1442 | if set(attrnames).issubset(_attrnames): |
| paul@89 | 1443 | types.append(name) |
| paul@89 | 1444 | return types |
| paul@89 | 1445 | |
| paul@90 | 1446 | def get_invoked_attributes(usage): |
| paul@90 | 1447 | |
| paul@90 | 1448 | "Obtain invoked attribute from the given 'usage'." |
| paul@90 | 1449 | |
| paul@90 | 1450 | invoked = [] |
| paul@90 | 1451 | if usage: |
| paul@107 | 1452 | for attrname, invocation, assignment in usage: |
| paul@90 | 1453 | if invocation: |
| paul@90 | 1454 | invoked.append(attrname) |
| paul@90 | 1455 | return invoked |
| paul@90 | 1456 | |
| paul@107 | 1457 | def get_assigned_attributes(usage): |
| paul@107 | 1458 | |
| paul@107 | 1459 | "Obtain assigned attribute from the given 'usage'." |
| paul@107 | 1460 | |
| paul@107 | 1461 | assigned = [] |
| paul@107 | 1462 | if usage: |
| paul@107 | 1463 | for attrname, invocation, assignment in usage: |
| paul@107 | 1464 | if assignment: |
| paul@107 | 1465 | assigned.append(attrname) |
| paul@107 | 1466 | return assigned |
| paul@107 | 1467 | |
| paul@366 | 1468 | # Type and module functions. |
| paul@538 | 1469 | # NOTE: This makes assumptions about the __builtins__ structure. |
| paul@366 | 1470 | |
| paul@366 | 1471 | def get_builtin_module(name): |
| paul@366 | 1472 | |
| paul@366 | 1473 | "Return the module name containing the given type 'name'." |
| paul@366 | 1474 | |
| paul@394 | 1475 | if name == "string": |
| paul@538 | 1476 | modname = "str" |
| paul@394 | 1477 | elif name == "utf8string": |
| paul@538 | 1478 | modname = "unicode" |
| paul@394 | 1479 | elif name == "NoneType": |
| paul@538 | 1480 | modname = "none" |
| paul@394 | 1481 | else: |
| paul@538 | 1482 | modname = name |
| paul@538 | 1483 | |
| paul@538 | 1484 | return "__builtins__.%s" % modname |
| paul@366 | 1485 | |
| paul@366 | 1486 | def get_builtin_type(name): |
| paul@366 | 1487 | |
| paul@366 | 1488 | "Return the type name provided by the given Python value 'name'." |
| paul@366 | 1489 | |
| paul@394 | 1490 | if name == "str": |
| paul@394 | 1491 | return "string" |
| paul@394 | 1492 | elif name == "unicode": |
| paul@394 | 1493 | return "utf8string" |
| paul@394 | 1494 | else: |
| paul@394 | 1495 | return name |
| paul@366 | 1496 | |
| paul@538 | 1497 | def get_builtin_class(name): |
| paul@538 | 1498 | |
| paul@538 | 1499 | "Return the full name of the built-in class having the given 'name'." |
| paul@538 | 1500 | |
| paul@538 | 1501 | typename = get_builtin_type(name) |
| paul@538 | 1502 | module = get_builtin_module(typename) |
| paul@538 | 1503 | return "%s.%s" % (module, typename) |
| paul@538 | 1504 | |
| paul@0 | 1505 | # Useful data. |
| paul@0 | 1506 | |
| paul@11 | 1507 | predefined_constants = "False", "None", "NotImplemented", "True" |
| paul@0 | 1508 | |
| paul@0 | 1509 | operator_functions = { |
| paul@0 | 1510 | |
| paul@0 | 1511 | # Fundamental operations. |
| paul@0 | 1512 | |
| paul@0 | 1513 | "is" : "is_", |
| paul@0 | 1514 | "is not" : "is_not", |
| paul@0 | 1515 | |
| paul@0 | 1516 | # Binary operations. |
| paul@0 | 1517 | |
| paul@0 | 1518 | "in" : "in_", |
| paul@0 | 1519 | "not in" : "not_in", |
| paul@0 | 1520 | "Add" : "add", |
| paul@0 | 1521 | "Bitand" : "and_", |
| paul@0 | 1522 | "Bitor" : "or_", |
| paul@0 | 1523 | "Bitxor" : "xor", |
| paul@0 | 1524 | "Div" : "div", |
| paul@0 | 1525 | "FloorDiv" : "floordiv", |
| paul@0 | 1526 | "LeftShift" : "lshift", |
| paul@0 | 1527 | "Mod" : "mod", |
| paul@0 | 1528 | "Mul" : "mul", |
| paul@0 | 1529 | "Power" : "pow", |
| paul@0 | 1530 | "RightShift" : "rshift", |
| paul@0 | 1531 | "Sub" : "sub", |
| paul@0 | 1532 | |
| paul@0 | 1533 | # Unary operations. |
| paul@0 | 1534 | |
| paul@0 | 1535 | "Invert" : "invert", |
| paul@0 | 1536 | "UnaryAdd" : "pos", |
| paul@0 | 1537 | "UnarySub" : "neg", |
| paul@0 | 1538 | |
| paul@0 | 1539 | # Augmented assignment. |
| paul@0 | 1540 | |
| paul@0 | 1541 | "+=" : "iadd", |
| paul@0 | 1542 | "-=" : "isub", |
| paul@0 | 1543 | "*=" : "imul", |
| paul@0 | 1544 | "/=" : "idiv", |
| paul@0 | 1545 | "//=" : "ifloordiv", |
| paul@0 | 1546 | "%=" : "imod", |
| paul@0 | 1547 | "**=" : "ipow", |
| paul@0 | 1548 | "<<=" : "ilshift", |
| paul@0 | 1549 | ">>=" : "irshift", |
| paul@0 | 1550 | "&=" : "iand", |
| paul@0 | 1551 | "^=" : "ixor", |
| paul@0 | 1552 | "|=" : "ior", |
| paul@0 | 1553 | |
| paul@0 | 1554 | # Comparisons. |
| paul@0 | 1555 | |
| paul@0 | 1556 | "==" : "eq", |
| paul@0 | 1557 | "!=" : "ne", |
| paul@0 | 1558 | "<" : "lt", |
| paul@0 | 1559 | "<=" : "le", |
| paul@0 | 1560 | ">=" : "ge", |
| paul@0 | 1561 | ">" : "gt", |
| paul@0 | 1562 | } |
| paul@0 | 1563 | |
| paul@0 | 1564 | # vim: tabstop=4 expandtab shiftwidth=4 |
