gen-postmortem-metadata.py 25.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
#!/usr/bin/env python

#
# Copyright 2012 the V8 project authors. All rights reserved.
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
#     * Redistributions of source code must retain the above copyright
#       notice, this list of conditions and the following disclaimer.
#     * Redistributions in binary form must reproduce the above
#       copyright notice, this list of conditions and the following
#       disclaimer in the documentation and/or other materials provided
#       with the distribution.
#     * Neither the name of Google Inc. nor the names of its
#       contributors may be used to endorse or promote products derived
#       from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#

#
# Emits a C++ file to be compiled and linked into libv8 to support postmortem
# debugging tools.  Most importantly, this tool emits constants describing V8
# internals:
#
#    v8dbg_type_CLASS__TYPE = VALUE             Describes class type values
#    v8dbg_class_CLASS__FIELD__TYPE = OFFSET    Describes class fields
#    v8dbg_parent_CLASS__PARENT                 Describes class hierarchy
#    v8dbg_frametype_NAME = VALUE               Describes stack frame values
#    v8dbg_off_fp_NAME = OFFSET                 Frame pointer offsets
#    v8dbg_prop_NAME = OFFSET                   Object property offsets
#    v8dbg_NAME = VALUE                         Miscellaneous values
#
# These constants are declared as global integers so that they'll be present in
# the generated libv8 binary.
#

49 50 51
# for py2/py3 compatibility
from __future__ import print_function

52 53 54 55
import re
import sys

#
56 57
# Miscellaneous constants such as tags and masks used for object identification,
# enumeration values used as indexes in internal tables, etc..
58 59 60
#
consts_misc = [
    { 'name': 'FirstNonstringType',     'value': 'FIRST_NONSTRING_TYPE' },
61 62
    { 'name': 'APIObjectType',          'value': 'JS_API_OBJECT_TYPE' },
    { 'name': 'SpecialAPIObjectType',   'value': 'JS_SPECIAL_API_OBJECT_TYPE' },
63

64 65 66
    { 'name': 'FirstContextType',     'value': 'FIRST_CONTEXT_TYPE' },
    { 'name': 'LastContextType',     'value': 'LAST_CONTEXT_TYPE' },

67 68 69 70 71
    { 'name': 'IsNotStringMask',        'value': 'kIsNotStringMask' },
    { 'name': 'StringTag',              'value': 'kStringTag' },

    { 'name': 'StringEncodingMask',     'value': 'kStringEncodingMask' },
    { 'name': 'TwoByteStringTag',       'value': 'kTwoByteStringTag' },
72
    { 'name': 'OneByteStringTag',       'value': 'kOneByteStringTag' },
73 74 75 76 77 78

    { 'name': 'StringRepresentationMask',
        'value': 'kStringRepresentationMask' },
    { 'name': 'SeqStringTag',           'value': 'kSeqStringTag' },
    { 'name': 'ConsStringTag',          'value': 'kConsStringTag' },
    { 'name': 'ExternalStringTag',      'value': 'kExternalStringTag' },
79
    { 'name': 'SlicedStringTag',        'value': 'kSlicedStringTag' },
80
    { 'name': 'ThinStringTag',          'value': 'kThinStringTag' },
81 82 83 84 85 86

    { 'name': 'HeapObjectTag',          'value': 'kHeapObjectTag' },
    { 'name': 'HeapObjectTagMask',      'value': 'kHeapObjectTagMask' },
    { 'name': 'SmiTag',                 'value': 'kSmiTag' },
    { 'name': 'SmiTagMask',             'value': 'kSmiTagMask' },
    { 'name': 'SmiValueShift',          'value': 'kSmiTagSize' },
87
    { 'name': 'SmiShiftSize',           'value': 'kSmiShiftSize' },
88

89 90 91 92
    { 'name': 'OddballFalse',           'value': 'Oddball::kFalse' },
    { 'name': 'OddballTrue',            'value': 'Oddball::kTrue' },
    { 'name': 'OddballTheHole',         'value': 'Oddball::kTheHole' },
    { 'name': 'OddballNull',            'value': 'Oddball::kNull' },
93
    { 'name': 'OddballArgumentsMarker', 'value': 'Oddball::kArgumentsMarker' },
94 95 96 97 98
    { 'name': 'OddballUndefined',       'value': 'Oddball::kUndefined' },
    { 'name': 'OddballUninitialized',   'value': 'Oddball::kUninitialized' },
    { 'name': 'OddballOther',           'value': 'Oddball::kOther' },
    { 'name': 'OddballException',       'value': 'Oddball::kException' },

99 100 101 102 103 104
    { 'name': 'prop_kind_Data',
        'value': 'kData' },
    { 'name': 'prop_kind_Accessor',
        'value': 'kAccessor' },
    { 'name': 'prop_kind_mask',
        'value': 'PropertyDetails::KindField::kMask' },
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
    { 'name': 'prop_location_Descriptor',
        'value': 'kDescriptor' },
    { 'name': 'prop_location_Field',
        'value': 'kField' },
    { 'name': 'prop_location_mask',
        'value': 'PropertyDetails::LocationField::kMask' },
    { 'name': 'prop_location_shift',
        'value': 'PropertyDetails::LocationField::kShift' },
    { 'name': 'prop_attributes_NONE', 'value': 'NONE' },
    { 'name': 'prop_attributes_READ_ONLY', 'value': 'READ_ONLY' },
    { 'name': 'prop_attributes_DONT_ENUM', 'value': 'DONT_ENUM' },
    { 'name': 'prop_attributes_DONT_DELETE', 'value': 'DONT_DELETE' },
    { 'name': 'prop_attributes_mask',
        'value': 'PropertyDetails::AttributesField::kMask' },
    { 'name': 'prop_attributes_shift',
        'value': 'PropertyDetails::AttributesField::kShift' },
121 122 123 124
    { 'name': 'prop_index_mask',
        'value': 'PropertyDetails::FieldIndexField::kMask' },
    { 'name': 'prop_index_shift',
        'value': 'PropertyDetails::FieldIndexField::kShift' },
125 126 127 128 129 130 131 132 133 134 135 136
    { 'name': 'prop_representation_mask',
        'value': 'PropertyDetails::RepresentationField::kMask' },
    { 'name': 'prop_representation_shift',
        'value': 'PropertyDetails::RepresentationField::kShift' },
    { 'name': 'prop_representation_smi',
        'value': 'Representation::Kind::kSmi' },
    { 'name': 'prop_representation_double',
        'value': 'Representation::Kind::kDouble' },
    { 'name': 'prop_representation_heapobject',
        'value': 'Representation::Kind::kHeapObject' },
    { 'name': 'prop_representation_tagged',
        'value': 'Representation::Kind::kTagged' },
137

138
    { 'name': 'prop_desc_key',
139
        'value': 'DescriptorArray::kEntryKeyIndex' },
140
    { 'name': 'prop_desc_details',
141
        'value': 'DescriptorArray::kEntryDetailsIndex' },
142
    { 'name': 'prop_desc_value',
143
        'value': 'DescriptorArray::kEntryValueIndex' },
144
    { 'name': 'prop_desc_size',
145
        'value': 'DescriptorArray::kEntrySize' },
146

147
    { 'name': 'elements_fast_holey_elements',
148
        'value': 'HOLEY_ELEMENTS' },
149
    { 'name': 'elements_fast_elements',
150
        'value': 'PACKED_ELEMENTS' },
151 152 153 154
    { 'name': 'elements_dictionary_elements',
        'value': 'DICTIONARY_ELEMENTS' },

    { 'name': 'bit_field2_elements_kind_mask',
155
        'value': 'Map::ElementsKindBits::kMask' },
156
    { 'name': 'bit_field2_elements_kind_shift',
157
        'value': 'Map::ElementsKindBits::kShift' },
158 159
    { 'name': 'bit_field3_is_dictionary_map_shift',
        'value': 'Map::IsDictionaryMapBit::kShift' },
160 161 162 163
    { 'name': 'bit_field3_number_of_own_descriptors_mask',
        'value': 'Map::NumberOfOwnDescriptorsBits::kMask' },
    { 'name': 'bit_field3_number_of_own_descriptors_shift',
        'value': 'Map::NumberOfOwnDescriptorsBits::kShift' },
164

165 166
    { 'name': 'off_fp_context_or_frame_type',
        'value': 'CommonFrameConstants::kContextOrFrameTypeOffset'},
167 168
    { 'name': 'off_fp_context',
        'value': 'StandardFrameConstants::kContextOffset' },
169 170
    { 'name': 'off_fp_constant_pool',
        'value': 'StandardFrameConstants::kConstantPoolOffset' },
171 172 173 174
    { 'name': 'off_fp_function',
        'value': 'JavaScriptFrameConstants::kFunctionOffset' },
    { 'name': 'off_fp_args',
        'value': 'JavaScriptFrameConstants::kLastParameterOffset' },
175 176 177 178 179 180 181

    { 'name': 'scopeinfo_idx_nparams',
        'value': 'ScopeInfo::kParameterCount' },
    { 'name': 'scopeinfo_idx_ncontextlocals',
        'value': 'ScopeInfo::kContextLocalCount' },
    { 'name': 'scopeinfo_idx_first_vars',
        'value': 'ScopeInfo::kVariablePartIndex' },
182

183 184 185 186
    { 'name': 'jsarray_buffer_was_detached_mask',
        'value': 'JSArrayBuffer::WasDetachedBit::kMask' },
    { 'name': 'jsarray_buffer_was_detached_shift',
        'value': 'JSArrayBuffer::WasDetachedBit::kShift' },
187

188 189
    { 'name': 'context_idx_scope_info',
        'value': 'Context::SCOPE_INFO_INDEX' },
190 191 192 193 194 195 196 197
    { 'name': 'context_idx_native',
        'value': 'Context::NATIVE_CONTEXT_INDEX' },
    { 'name': 'context_idx_prev',
        'value': 'Context::PREVIOUS_INDEX' },
    { 'name': 'context_idx_ext',
        'value': 'Context::EXTENSION_INDEX' },
    { 'name': 'context_min_slots',
        'value': 'Context::MIN_CONTEXT_SLOTS' },
198 199
    { 'name': 'native_context_embedder_data_offset',
        'value': 'Internals::kNativeContextEmbedderDataOffset' },
200

201 202 203 204 205

    { 'name': 'namedictionaryshape_prefix_size',
        'value': 'NameDictionaryShape::kPrefixSize' },
    { 'name': 'namedictionaryshape_entry_size',
        'value': 'NameDictionaryShape::kEntrySize' },
206 207
    { 'name': 'globaldictionaryshape_entry_size',
        'value': 'GlobalDictionaryShape::kEntrySize' },
208 209 210 211

    { 'name': 'namedictionary_prefix_start_index',
        'value': 'NameDictionary::kPrefixStartIndex' },

212 213 214 215
    { 'name': 'numberdictionaryshape_prefix_size',
        'value': 'NumberDictionaryShape::kPrefixSize' },
    { 'name': 'numberdictionaryshape_entry_size',
        'value': 'NumberDictionaryShape::kEntrySize' },
216 217 218 219 220

    { 'name': 'simplenumberdictionaryshape_prefix_size',
        'value': 'SimpleNumberDictionaryShape::kPrefixSize' },
    { 'name': 'simplenumberdictionaryshape_entry_size',
        'value': 'SimpleNumberDictionaryShape::kEntrySize' },
221 222

    { 'name': 'type_JSError__JS_ERROR_TYPE', 'value': 'JS_ERROR_TYPE' },
223 224 225

    { 'name': 'class_SharedFunctionInfo__function_data__Object',
        'value': 'SharedFunctionInfo::kFunctionDataOffset' },
226 227 228 229
];

#
# The following useful fields are missing accessors, so we define fake ones.
230 231 232 233 234
# Please note that extra accessors should _only_ be added to expose offsets that
# can be used to access actual V8 objects' properties. They should not be added
# for exposing other values. For instance, enumeration values or class'
# constants should be exposed by adding an entry in the "consts_misc" table, not
# in this "extras_accessors" table.
235 236
#
extras_accessors = [
237
    'JSFunction, context, Context, kContextOffset',
238
    'JSFunction, shared, SharedFunctionInfo, kSharedFunctionInfoOffset',
239 240
    'HeapObject, map, Map, kMapOffset',
    'JSObject, elements, Object, kElementsOffset',
241
    'JSObject, internal_fields, uintptr_t, kHeaderSize',
242
    'FixedArray, data, uintptr_t, kHeaderSize',
243
    'FixedTypedArrayBase, external_pointer, Object, kExternalPointerOffset',
244
    'JSArrayBuffer, backing_store, Object, kBackingStoreOffset',
245 246 247
    'JSArrayBuffer, byte_length, size_t, kByteLengthOffset',
    'JSArrayBufferView, byte_length, size_t, kByteLengthOffset',
    'JSArrayBufferView, byte_offset, size_t, kByteOffsetOffset',
248
    'JSTypedArray, length, Object, kLengthOffset',
249 250
    'Map, instance_size_in_words, char, kInstanceSizeInWordsOffset',
    'Map, inobject_properties_start_or_constructor_function_index, char, kInObjectPropertiesStartOrConstructorFunctionIndexOffset',
251
    'Map, instance_type, uint16_t, kInstanceTypeOffset',
252 253
    'Map, bit_field, char, kBitFieldOffset',
    'Map, bit_field2, char, kBitField2Offset',
254
    'Map, bit_field3, int, kBitField3Offset',
255 256
    'Map, prototype, Object, kPrototypeOffset',
    'Oddball, kind_offset, int, kKindOffset',
257 258 259 260
    'HeapNumber, value, double, kValueOffset',
    'ConsString, first, String, kFirstOffset',
    'ConsString, second, String, kSecondOffset',
    'ExternalString, resource, Object, kResourceOffset',
261
    'SeqOneByteString, chars, char, kHeaderSize',
262
    'SeqTwoByteString, chars, char, kHeaderSize',
263 264 265
    'UncompiledData, start_position, int32_t, kStartPositionOffset',
    'UncompiledData, end_position, int32_t, kEndPositionOffset',
    'SharedFunctionInfo, raw_function_token_offset, int16_t, kFunctionTokenOffsetOffset',
266
    'SharedFunctionInfo, internal_formal_parameter_count, uint16_t, kFormalParameterCountOffset',
267
    'SharedFunctionInfo, flags, int, kFlagsOffset',
268
    'SharedFunctionInfo, length, uint16_t, kLengthOffset',
269
    'SlicedString, parent, String, kParentOffset',
270 271
    'Code, instruction_start, uintptr_t, kHeaderSize',
    'Code, instruction_size, int, kInstructionSizeOffset',
272
    'String, length, int32_t, kLengthOffset',
273 274 275 276 277 278 279 280 281 282
];

#
# The following is a whitelist of classes we expect to find when scanning the
# source code. This list is not exhaustive, but it's still useful to identify
# when this script gets out of sync with the source. See load_objects().
#
expected_classes = [
    'ConsString', 'FixedArray', 'HeapNumber', 'JSArray', 'JSFunction',
    'JSObject', 'JSRegExp', 'JSValue', 'Map', 'Oddball', 'Script',
283
    'SeqOneByteString', 'SharedFunctionInfo', 'ScopeInfo', 'JSPromise'
284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
];


#
# The following structures store high-level representations of the structures
# for which we're going to emit descriptive constants.
#
types = {};             # set of all type names
typeclasses = {};       # maps type names to corresponding class names
klasses = {};           # known classes, including parents
fields = [];            # field declarations

header = '''
/*
 * This file is generated by %s.  Do not edit directly.
 */

301 302 303
#include "src/v8.h"
#include "src/frames.h"
#include "src/frames-inl.h" /* for architecture-specific frame constants */
304
#include "src/contexts.h"
305
#include "src/objects.h"
306
#include "src/objects/data-handler.h"
307
#include "src/objects/js-promise.h"
308
#include "src/objects/js-regexp-string-iterator.h"
309

310 311
namespace v8 {
namespace internal {
312 313 314 315 316 317 318 319 320 321 322 323 324 325 326

extern "C" {

/* stack frame constants */
#define FRAME_CONST(value, klass)       \
    int v8dbg_frametype_##klass = StackFrame::value;

STACK_FRAME_TYPE_LIST(FRAME_CONST)

#undef FRAME_CONST

''' % sys.argv[0];

footer = '''
}
327 328 329

}
}
330 331
'''

332 333 334 335 336 337 338 339 340 341 342 343 344 345
#
# Get the base class
#
def get_base_class(klass):
        if (klass == 'Object'):
                return klass;

        if (not (klass in klasses)):
                return None;

        k = klasses[klass];

        return get_base_class(k['parent']);

346
#
347
# Loads class hierarchy and type information from "objects.h" etc.
348 349 350 351 352 353 354 355 356
#
def load_objects():
        #
        # Construct a dictionary for the classes we're sure should be present.
        #
        checktypes = {};
        for klass in expected_classes:
                checktypes[klass] = True;

357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374

        for filename in sys.argv[2:]:
                if not filename.endswith("-inl.h"):
                        load_objects_from_file(filename, checktypes)

        if (len(checktypes) > 0):
                for klass in checktypes:
                        print('error: expected class \"%s\" not found' % klass);

                sys.exit(1);


def load_objects_from_file(objfilename, checktypes):
        objfile = open(objfilename, 'r');
        in_insttype = False;

        typestr = '';

375
        #
376 377 378 379
        # Iterate the header file line-by-line to collect type and class
        # information. For types, we accumulate a string representing the entire
        # InstanceType enum definition and parse it later because it's easier to
        # do so without the embedded newlines.
380 381
        #
        for line in objfile:
382
                if (line.startswith('enum InstanceType : uint16_t {')):
383 384 385 386 387 388 389
                        in_insttype = True;
                        continue;

                if (in_insttype and line.startswith('};')):
                        in_insttype = False;
                        continue;

390
                line = re.sub('//.*', '', line.strip());
391 392 393 394 395

                if (in_insttype):
                        typestr += line;
                        continue;

396
                match = re.match('class (\w[^:]*)(: public (\w[^{]*))?\s*{\s*',
397 398 399
                    line);

                if (match):
400
                        klass = match.group(1).strip();
401
                        pklass = match.group(3);
402
                        if (pklass):
403
                                pklass = pklass.strip();
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423
                        klasses[klass] = { 'parent': pklass };

        #
        # Process the instance type declaration.
        #
        entries = typestr.split(',');
        for entry in entries:
                types[re.sub('\s*=.*', '', entry).lstrip()] = True;

        #
        # Infer class names for each type based on a systematic transformation.
        # For example, "JS_FUNCTION_TYPE" becomes "JSFunction".  We find the
        # class for each type rather than the other way around because there are
        # fewer cases where one type maps to more than one class than the other
        # way around.
        #
        for type in types:
                #
                # REGEXP behaves like REG_EXP, as in JS_REGEXP_TYPE => JSRegExp.
                #
424
                usetype = re.sub('_REGEXP_', '_REG_EXP_', type);
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451

                #
                # Remove the "_TYPE" suffix and then convert to camel case,
                # except that a "JS" prefix remains uppercase (as in
                # "JS_FUNCTION_TYPE" => "JSFunction").
                #
                if (not usetype.endswith('_TYPE')):
                        continue;

                usetype = usetype[0:len(usetype) - len('_TYPE')];
                parts = usetype.split('_');
                cctype = '';

                if (parts[0] == 'JS'):
                        cctype = 'JS';
                        start = 1;
                else:
                        cctype = '';
                        start = 0;

                for ii in range(start, len(parts)):
                        part = parts[ii];
                        cctype += part[0].upper() + part[1:].lower();

                #
                # Mapping string types is more complicated.  Both types and
                # class names for Strings specify a representation (e.g., Seq,
452
                # Cons, External, or Sliced) and an encoding (TwoByte/OneByte),
453 454 455
                # In the simplest case, both of these are explicit in both
                # names, as in:
                #
456
                #       EXTERNAL_ONE_BYTE_STRING_TYPE => ExternalOneByteString
457 458 459 460 461 462 463 464 465 466
                #
                # However, either the representation or encoding can be omitted
                # from the type name, in which case "Seq" and "TwoByte" are
                # assumed, as in:
                #
                #       STRING_TYPE => SeqTwoByteString
                #
                # Additionally, sometimes the type name has more information
                # than the class, as in:
                #
467
                #       CONS_ONE_BYTE_STRING_TYPE => ConsString
468 469 470 471 472 473 474 475 476 477
                #
                # To figure this out dynamically, we first check for a
                # representation and encoding and add them if they're not
                # present.  If that doesn't yield a valid class name, then we
                # strip out the representation.
                #
                if (cctype.endswith('String')):
                        if (cctype.find('Cons') == -1 and
                            cctype.find('External') == -1 and
                            cctype.find('Sliced') == -1):
478 479
                                if (cctype.find('OneByte') != -1):
                                        cctype = re.sub('OneByteString$',
480
                                            'SeqOneByteString', cctype);
481 482 483 484
                                else:
                                        cctype = re.sub('String$',
                                            'SeqString', cctype);

485
                        if (cctype.find('OneByte') == -1):
486 487 488 489
                                cctype = re.sub('String$', 'TwoByteString',
                                    cctype);

                        if (not (cctype in klasses)):
490
                                cctype = re.sub('OneByte', '', cctype);
491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517
                                cctype = re.sub('TwoByte', '', cctype);

                #
                # Despite all that, some types have no corresponding class.
                #
                if (cctype in klasses):
                        typeclasses[type] = cctype;
                        if (cctype in checktypes):
                                del checktypes[cctype];

#
# For a given macro call, pick apart the arguments and return an object
# describing the corresponding output constant.  See load_fields().
#
def parse_field(call):
        # Replace newlines with spaces.
        for ii in range(0, len(call)):
                if (call[ii] == '\n'):
                        call[ii] == ' ';

        idx = call.find('(');
        kind = call[0:idx];
        rest = call[idx + 1: len(call) - 1];
        args = re.split('\s*,\s*', rest);

        consts = [];

518 519
        if (kind == 'ACCESSORS' or kind == 'ACCESSORS2' or
            kind == 'ACCESSORS_GCSAFE'):
520 521
                klass = args[0];
                field = args[1];
522
                dtype = args[2].replace('<', '_').replace('>', '_')
523 524 525 526 527 528 529
                offset = args[3];

                return ({
                    'name': 'class_%s__%s__%s' % (klass, field, dtype),
                    'value': '%s::%s' % (klass, offset)
                });

530
        assert(kind == 'SMI_ACCESSORS' or kind == 'ACCESSORS_TO_SMI');
531 532 533 534 535 536 537 538 539 540
        klass = args[0];
        field = args[1];
        offset = args[2];

        return ({
            'name': 'class_%s__%s__%s' % (klass, field, 'SMI'),
            'value': '%s::%s' % (klass, offset)
        });

#
541
# Load field offset information from objects-inl.h etc.
542 543
#
def load_fields():
544 545 546 547 548 549 550 551 552 553
        for filename in sys.argv[2:]:
                if filename.endswith("-inl.h"):
                        load_fields_from_file(filename)

        for body in extras_accessors:
                fields.append(parse_field('ACCESSORS(%s)' % body));


def load_fields_from_file(filename):
        inlfile = open(filename, 'r');
554 555 556 557 558 559 560 561

        #
        # Each class's fields and the corresponding offsets are described in the
        # source by calls to macros like "ACCESSORS" (and friends).  All we do
        # here is extract these macro invocations, taking into account that they
        # may span multiple lines and may contain nested parentheses.  We also
        # call parse_field() to pick apart the invocation.
        #
562
        prefixes = [ 'ACCESSORS', 'ACCESSORS2', 'ACCESSORS_GCSAFE',
563
                     'SMI_ACCESSORS', 'ACCESSORS_TO_SMI' ];
564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608
        current = '';
        opens = 0;

        for line in inlfile:
                if (opens > 0):
                        # Continuation line
                        for ii in range(0, len(line)):
                                if (line[ii] == '('):
                                        opens += 1;
                                elif (line[ii] == ')'):
                                        opens -= 1;

                                if (opens == 0):
                                        break;

                        current += line[0:ii + 1];
                        continue;

                for prefix in prefixes:
                        if (not line.startswith(prefix + '(')):
                                continue;

                        if (len(current) > 0):
                                fields.append(parse_field(current));
                                current = '';

                        for ii in range(len(prefix), len(line)):
                                if (line[ii] == '('):
                                        opens += 1;
                                elif (line[ii] == ')'):
                                        opens -= 1;

                                if (opens == 0):
                                        break;

                        current += line[0:ii + 1];

        if (len(current) > 0):
                fields.append(parse_field(current));
                current = '';

#
# Emit a block of constants.
#
def emit_set(out, consts):
609 610 611 612 613 614 615
        # Fix up overzealous parses.  This could be done inside the
        # parsers but as there are several, it's easiest to do it here.
        ws = re.compile('\s+')
        for const in consts:
                name = ws.sub('', const['name'])
                value = ws.sub('', str(const['value']))  # Can be a number.
                out.write('int v8dbg_%s = %s;\n' % (name, value))
616 617 618 619 620 621
        out.write('\n');

#
# Emit the whole output file.
#
def emit_config():
622
        out = open(sys.argv[1], 'w');
623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647

        out.write(header);

        out.write('/* miscellaneous constants */\n');
        emit_set(out, consts_misc);

        out.write('/* class type information */\n');
        consts = [];
        keys = typeclasses.keys();
        keys.sort();
        for typename in keys:
                klass = typeclasses[typename];
                consts.append({
                    'name': 'type_%s__%s' % (klass, typename),
                    'value': typename
                });

        emit_set(out, consts);

        out.write('/* class hierarchy information */\n');
        consts = [];
        keys = klasses.keys();
        keys.sort();
        for klassname in keys:
                pklass = klasses[klassname]['parent'];
648 649 650
                bklass = get_base_class(klassname);
                if (bklass != 'Object'):
                        continue;
651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672
                if (pklass == None):
                        continue;

                consts.append({
                    'name': 'parent_%s__%s' % (klassname, pklass),
                    'value': 0
                });

        emit_set(out, consts);

        out.write('/* field information */\n');
        emit_set(out, fields);

        out.write(footer);

if (len(sys.argv) < 4):
        print('usage: %s output.cc objects.h objects-inl.h' % sys.argv[0]);
        sys.exit(2);

load_objects();
load_fields();
emit_config();