This file is indexed.

/usr/share/pyshared/epydoc/docbuilder.py is in python-epydoc 3.0.1+dfsg-5.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

   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
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 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
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 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
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 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
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 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
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
# epydoc -- Documentation Builder
#
# Copyright (C) 2005 Edward Loper
# Author: Edward Loper <edloper@loper.org>
# URL: <http://epydoc.sf.net>
#
# $Id: docbuilder.py 1683 2008-01-29 22:17:39Z edloper $

"""
Construct data structures that encode the API documentation for Python
objects.  These data structures are created using a series of steps:

  1. B{Building docs}: Extract basic information about the objects,
     and objects that are related to them.  This can be done by
     introspecting the objects' values (with L{epydoc.docintrospecter}; or
     by parsing their source code (with L{epydoc.docparser}.

  2. B{Merging}: Combine the information obtained from introspection &
     parsing each object into a single structure.

  3. B{Linking}: Replace any 'pointers' that were created for imported
     variables by their target (if it's available).
  
  4. B{Naming}: Chose a unique 'canonical name' for each
     object.
  
  5. B{Docstring Parsing}: Parse the docstring of each object, and
     extract any pertinant information.
  
  6. B{Inheritance}: Add information about variables that classes
     inherit from their base classes.

The documentation information for each individual object is
represented using an L{APIDoc}; and the documentation for a collection
of objects is represented using a L{DocIndex}.

The main interface to C{epydoc.docbuilder} consists of two functions:

  - L{build_doc()} -- Builds documentation for a single item, and
    returns it as an L{APIDoc} object.
  - L{build_doc_index()} -- Builds documentation for a collection of
    items, and returns it as a L{DocIndex} object.

The remaining functions are used by these two main functions to
perform individual steps in the creation of the documentation.

@group Documentation Construction: build_doc, build_doc_index,
    _get_docs_from_*, _report_valdoc_progress
@group Merging: *MERGE*, *merge*
@group Linking: link_imports
@group Naming: _name_scores, _unreachable_names, assign_canonical_names,
    _var_shadows_self, _fix_self_shadowing_var, _unreachable_name_for
@group Inheritance: inherit_docs, _inherit_info
"""
__docformat__ = 'epytext en'

######################################################################
## Contents
######################################################################
## 1. build_doc() & build_doc_index() -- the main interface.
## 2. merge_docs() -- helper, used to merge parse & introspect info
## 3. link_imports() -- helper, used to connect imported vars w/ values
## 4. assign_canonical_names() -- helper, used to set canonical names
## 5. inherit_docs() -- helper, used to inherit docs from base classes

######################################################################
## Imports
######################################################################

import sys, os, os.path, __builtin__, imp, re, inspect
from epydoc.apidoc import *
from epydoc.docintrospecter import introspect_docs
from epydoc.docparser import parse_docs, ParseError
from epydoc.docstringparser import parse_docstring
from epydoc import log
from epydoc.util import *
from epydoc.compat import * # Backwards compatibility

######################################################################
## 1. build_doc()
######################################################################

class BuildOptions:
    """
    Holds the parameters for a documentation building process.
    """
    def __init__(self, introspect=True, parse=True,
                 exclude_introspect=None, exclude_parse=None,
                 add_submodules=True):
        self.introspect = introspect
        self.parse = parse
        self.exclude_introspect = exclude_introspect
        self.exclude_parse = exclude_parse
        self.add_submodules = add_submodules

        # Test for pattern syntax and compile them into pattern objects.
        try:
            self._introspect_regexp = (exclude_introspect
                and re.compile(exclude_introspect) or None)
            self._parse_regexp = (exclude_parse
                and re.compile(exclude_parse) or None)
        except Exception, exc:
            log.error('Error in regular expression pattern: %s' % exc)
            raise

    def must_introspect(self, name):
        """
        Return C{True} if a module is to be introsepcted with the current
        settings.

        @param name: The name of the module to test
        @type name: L{DottedName} or C{str}
        """
        return self.introspect \
            and not self._matches_filter(name, self._introspect_regexp)

    def must_parse(self, name):
        """
        Return C{True} if a module is to be parsed with the current settings.

        @param name: The name of the module to test
        @type name: L{DottedName} or C{str}
        """
        return self.parse \
            and not self._matches_filter(name, self._parse_regexp)

    def _matches_filter(self, name, regexp):
        """
        Test if a module name matches a pattern.

        @param name: The name of the module to test
        @type name: L{DottedName} or C{str}
        @param regexp: The pattern object to match C{name} against.
            If C{None}, return C{False}
        @type regexp: C{pattern}
        @return: C{True} if C{name} in dotted format matches C{regexp},
            else C{False}
        @rtype: C{bool}
        """
        if regexp is None: return False

        if isinstance(name, DottedName):
            name = str(name)

        return bool(regexp.search(name))


def build_doc(item, introspect=True, parse=True, add_submodules=True,
              exclude_introspect=None, exclude_parse=None):
    """
    Build API documentation for a given item, and return it as
    an L{APIDoc} object.

    @rtype: L{APIDoc}
    @param item: The item to document, specified using any of the
        following:
          - A string, naming a python package directory
            (e.g., C{'epydoc/markup'})
          - A string, naming a python file
            (e.g., C{'epydoc/docparser.py'})
          - A string, naming a python object
            (e.g., C{'epydoc.docparser.DocParser'})
          - Any (non-string) python object
            (e.g., C{list.append})
    @param introspect: If true, then use introspection to examine the
        specified items.  Otherwise, just use parsing.
    @param parse: If true, then use parsing to examine the specified
        items.  Otherwise, just use introspection.
    """
    docindex = build_doc_index([item], introspect, parse, add_submodules,
                               exclude_introspect=exclude_introspect,
                               exclude_parse=exclude_parse)
    return docindex.root[0]

def build_doc_index(items, introspect=True, parse=True, add_submodules=True,
                    exclude_introspect=None, exclude_parse=None):
    """
    Build API documentation for the given list of items, and
    return it in the form of a L{DocIndex}.

    @rtype: L{DocIndex}
    @param items: The items to document, specified using any of the
        following:
          - A string, naming a python package directory
            (e.g., C{'epydoc/markup'})
          - A string, naming a python file
            (e.g., C{'epydoc/docparser.py'})
          - A string, naming a python object
            (e.g., C{'epydoc.docparser.DocParser'})
          - Any (non-string) python object
            (e.g., C{list.append})
    @param introspect: If true, then use introspection to examine the
        specified items.  Otherwise, just use parsing.
    @param parse: If true, then use parsing to examine the specified
        items.  Otherwise, just use introspection.
    """
    try:
        options = BuildOptions(parse=parse, introspect=introspect,
            exclude_introspect=exclude_introspect, exclude_parse=exclude_parse,
            add_submodules=add_submodules)
    except Exception, e:
        # log.error already reported by constructor.
        return None

    # Get the basic docs for each item.
    doc_pairs = _get_docs_from_items(items, options)

    # Merge the introspection & parse docs.
    if options.parse and options.introspect:
        log.start_progress('Merging parsed & introspected information')
        docs = []
        for i, (introspect_doc, parse_doc) in enumerate(doc_pairs):
            if introspect_doc is not None and parse_doc is not None:
                if introspect_doc.canonical_name not in (None, UNKNOWN):
                    name = introspect_doc.canonical_name
                else:
                    name = parse_doc.canonical_name
                log.progress(float(i)/len(doc_pairs), name)
                docs.append(merge_docs(introspect_doc, parse_doc))
            elif introspect_doc is not None:
                docs.append(introspect_doc)
            elif parse_doc is not None:
                docs.append(parse_doc)
        log.end_progress()
    elif options.introspect:
        docs = [doc_pair[0] for doc_pair in doc_pairs if doc_pair[0]]
    else:
        docs = [doc_pair[1] for doc_pair in doc_pairs if doc_pair[1]]

    if len(docs) == 0:
        log.error('Nothing left to document!')
        return None

    # Collect the docs into a single index.
    docindex = DocIndex(docs)

    # Replace any proxy valuedocs that we got from importing with
    # their targets.
    if options.parse:
        log.start_progress('Linking imported variables')
        valdocs = sorted(docindex.reachable_valdocs(
            imports=False, submodules=False, packages=False, subclasses=False))
        for i, val_doc in enumerate(valdocs):
            _report_valdoc_progress(i, val_doc, valdocs)
            link_imports(val_doc, docindex)
        log.end_progress()

    # Assign canonical names.
    log.start_progress('Indexing documentation')
    for i, val_doc in enumerate(docindex.root):
        log.progress(float(i)/len(docindex.root), val_doc.canonical_name)
        assign_canonical_names(val_doc, val_doc.canonical_name, docindex)
    log.end_progress()

    # Set overrides pointers
    log.start_progress('Checking for overridden methods')
    valdocs = sorted(docindex.reachable_valdocs(
        imports=False, submodules=False, packages=False, subclasses=False))
    for i, val_doc in enumerate(valdocs):
        if isinstance(val_doc, ClassDoc):
            percent = float(i)/len(valdocs)
            log.progress(percent, val_doc.canonical_name)
            find_overrides(val_doc)
    log.end_progress()
    
    # Parse the docstrings for each object.
    log.start_progress('Parsing docstrings')
    suppress_warnings = set(valdocs).difference(
        docindex.reachable_valdocs(
            imports=False, submodules=False, packages=False, subclasses=False,
            bases=False, overrides=True))
    for i, val_doc in enumerate(valdocs):
        _report_valdoc_progress(i, val_doc, valdocs)
        # the value's docstring
        parse_docstring(val_doc, docindex, suppress_warnings)
        # the value's variables' docstrings
        if (isinstance(val_doc, NamespaceDoc) and
            val_doc.variables not in (None, UNKNOWN)):
            for var_doc in val_doc.variables.values():
                # Now we have a chance to propagate the defining module
                # to objects for which introspection is not possible,
                # such as properties.
                if (isinstance(var_doc.value, ValueDoc)
                    and var_doc.value.defining_module is UNKNOWN):
                    var_doc.value.defining_module = val_doc.defining_module
                parse_docstring(var_doc, docindex, suppress_warnings)
    log.end_progress()

    # Take care of inheritance.
    log.start_progress('Inheriting documentation')
    for i, val_doc in enumerate(valdocs):
        if isinstance(val_doc, ClassDoc):
            percent = float(i)/len(valdocs)
            log.progress(percent, val_doc.canonical_name)
            inherit_docs(val_doc)
    log.end_progress()

    # Initialize the groups & sortedvars attributes.
    log.start_progress('Sorting & Grouping')
    for i, val_doc in enumerate(valdocs):
        if isinstance(val_doc, NamespaceDoc):
            percent = float(i)/len(valdocs)
            log.progress(percent, val_doc.canonical_name)
            val_doc.init_sorted_variables()
            val_doc.init_variable_groups()
            if isinstance(val_doc, ModuleDoc):
                val_doc.init_submodule_groups()
            val_doc.report_unused_groups()
    log.end_progress()

    return docindex

def _report_valdoc_progress(i, val_doc, val_docs):
    if (isinstance(val_doc, (ModuleDoc, ClassDoc)) and
        val_doc.canonical_name is not UNKNOWN and
        not val_doc.canonical_name[0].startswith('??')):
        log.progress(float(i)/len(val_docs), val_doc.canonical_name)

#/////////////////////////////////////////////////////////////////
# Documentation Generation
#/////////////////////////////////////////////////////////////////

def _get_docs_from_items(items, options):

    # Start the progress bar.
    log.start_progress('Building documentation')
    progress_estimator = _ProgressEstimator(items)

    # Check for duplicate item names.
    item_set = set()
    for item in items[:]:
        if item in item_set:
            log.warning("Name %r given multiple times" % item)
            items.remove(item)
        item_set.add(item)

    # Keep track of what top-level canonical names we've assigned, to
    # make sure there are no naming conflicts.  This dict maps
    # canonical names to the item names they came from (so we can print
    # useful error messages).
    canonical_names = {}

    # Collect (introspectdoc, parsedoc) pairs for each item.
    doc_pairs = []
    for item in items:
        if isinstance(item, basestring):
            if is_module_file(item):
                doc_pairs.append(_get_docs_from_module_file(
                    item, options, progress_estimator))
            elif is_package_dir(item):
                pkgfile = os.path.abspath(os.path.join(item, '__init__'))
                doc_pairs.append(_get_docs_from_module_file(
                    pkgfile, options, progress_estimator))
            elif os.path.isfile(item):
                doc_pairs.append(_get_docs_from_pyscript(
                    item, options, progress_estimator))
            elif hasattr(__builtin__, item):
                val = getattr(__builtin__, item)
                doc_pairs.append(_get_docs_from_pyobject(
                    val, options, progress_estimator))
            elif is_pyname(item):
                doc_pairs.append(_get_docs_from_pyname(
                    item, options, progress_estimator))
            elif os.path.isdir(item):
                log.error("Directory %r is not a package" % item)
                continue
            elif os.path.isfile(item):
                log.error("File %s is not a Python module" % item)
                continue
            else:
                log.error("Could not find a file or object named %s" %
                          item)
                continue
        else:
            doc_pairs.append(_get_docs_from_pyobject(
                item, options, progress_estimator))

        # Make sure there are no naming conflicts.
        name = (getattr(doc_pairs[-1][0], 'canonical_name', None) or
                getattr(doc_pairs[-1][1], 'canonical_name', None))
        if name in canonical_names:
            log.error(
                'Two of the specified items, %r and %r, have the same '
                'canonical name ("%s").  This may mean that you specified '
                'two different files that both use the same module name.  '
                'Ignoring the second item (%r)' %
                (canonical_names[name], item, name, canonical_names[name]))
            doc_pairs.pop()
        else:
            canonical_names[name] = item                

        # This will only have an effect if doc_pairs[-1] contains a
        # package's docs.  The 'not is_module_file(item)' prevents
        # us from adding subdirectories if they explicitly specify
        # a package's __init__.py file.
        if options.add_submodules and not is_module_file(item):
            doc_pairs += _get_docs_from_submodules(
                item, doc_pairs[-1], options, progress_estimator)

    log.end_progress()
    return doc_pairs

def _get_docs_from_pyobject(obj, options, progress_estimator):
    progress_estimator.complete += 1
    log.progress(progress_estimator.progress(), repr(obj))
    
    if not options.introspect:
        log.error("Cannot get docs for Python objects without "
                  "introspecting them.")
            
    introspect_doc = parse_doc = None
    introspect_error = parse_error = None
    try:
        introspect_doc = introspect_docs(value=obj)
    except ImportError, e:
        log.error(e)
        return (None, None)
    if options.parse:
        if introspect_doc.canonical_name is not None:
            prev_introspect = options.introspect
            options.introspect = False
            try:
                _, parse_docs = _get_docs_from_pyname(
                    str(introspect_doc.canonical_name), options,
                    progress_estimator, suppress_warnings=True)
            finally:
                options.introspect = prev_introspect

    # We need a name:
    if introspect_doc.canonical_name in (None, UNKNOWN):
        if hasattr(obj, '__name__'):
            introspect_doc.canonical_name = DottedName(
                DottedName.UNREACHABLE, obj.__name__)
        else:
            introspect_doc.canonical_name = DottedName(
                DottedName.UNREACHABLE)
    return (introspect_doc, parse_doc)

def _get_docs_from_pyname(name, options, progress_estimator,
                          suppress_warnings=False):
    progress_estimator.complete += 1
    if options.must_introspect(name) or options.must_parse(name):
        log.progress(progress_estimator.progress(), name)
    
    introspect_doc = parse_doc = None
    introspect_error = parse_error = None
    if options.must_introspect(name):
        try:
            introspect_doc = introspect_docs(name=name)
        except ImportError, e:
            introspect_error = str(e)
    if options.must_parse(name):
        try:
            parse_doc = parse_docs(name=name)
        except ParseError, e:
            parse_error = str(e)
        except ImportError, e:
            # If we get here, then there' probably no python source
            # available; don't bother to generate a warnining.
            pass
        
    # Report any errors we encountered.
    if not suppress_warnings:
        _report_errors(name, introspect_doc, parse_doc,
                       introspect_error, parse_error)

    # Return the docs we found.
    return (introspect_doc, parse_doc)

def _get_docs_from_pyscript(filename, options, progress_estimator):
    # [xx] I should be careful about what names I allow as filenames,
    # and maybe do some munging to prevent problems.

    introspect_doc = parse_doc = None
    introspect_error = parse_error = None
    if options.introspect:
        try:
            introspect_doc = introspect_docs(filename=filename, is_script=True)
            if introspect_doc.canonical_name is UNKNOWN:
                introspect_doc.canonical_name = munge_script_name(filename)
        except ImportError, e:
            introspect_error = str(e)
    if options.parse:
        try:
            parse_doc = parse_docs(filename=filename, is_script=True)
        except ParseError, e:
            parse_error = str(e)
        except ImportError, e:
            parse_error = str(e)
                
    # Report any errors we encountered.
    _report_errors(filename, introspect_doc, parse_doc,
                   introspect_error, parse_error)

    # Return the docs we found.
    return (introspect_doc, parse_doc)
    
def _get_docs_from_module_file(filename, options, progress_estimator,
                               parent_docs=(None,None)):
    """
    Construct and return the API documentation for the python
    module with the given filename.

    @param parent_docs: The C{ModuleDoc} of the containing package.
        If C{parent_docs} is not provided, then this method will
        check if the given filename is contained in a package; and
        if so, it will construct a stub C{ModuleDoc} for the
        containing package(s).  C{parent_docs} is a tuple, where
        the first element is the parent from introspection, and
        the second element is the parent from parsing.
    """
    # Record our progress.
    modulename = os.path.splitext(os.path.split(filename)[1])[0]
    if modulename == '__init__':
        modulename = os.path.split(os.path.split(filename)[0])[1]
    if parent_docs[0]:
        modulename = DottedName(parent_docs[0].canonical_name, modulename)
    elif parent_docs[1]:
        modulename = DottedName(parent_docs[1].canonical_name, modulename)
    if options.must_introspect(modulename) or options.must_parse(modulename):
        log.progress(progress_estimator.progress(),
                     '%s (%s)' % (modulename, filename))
    progress_estimator.complete += 1
    
    # Normalize the filename.
    filename = os.path.normpath(os.path.abspath(filename))

    # When possible, use the source version of the file.
    try:
        filename = py_src_filename(filename)
        src_file_available = True
    except ValueError:
        src_file_available = False

    # Get the introspected & parsed docs (as appropriate)
    introspect_doc = parse_doc = None
    introspect_error = parse_error = None
    if options.must_introspect(modulename):
        try:
            introspect_doc = introspect_docs(
                filename=filename, context=parent_docs[0])
            if introspect_doc.canonical_name is UNKNOWN:
                introspect_doc.canonical_name = modulename
        except ImportError, e:
            introspect_error = str(e)
    if src_file_available and options.must_parse(modulename):
        try:
            parse_doc = parse_docs(
                filename=filename, context=parent_docs[1])
        except ParseError, e:
            parse_error = str(e)
        except ImportError, e:
            parse_error = str(e)

    # Report any errors we encountered.
    _report_errors(filename, introspect_doc, parse_doc,
                   introspect_error, parse_error)

    # Return the docs we found.
    return (introspect_doc, parse_doc)

def _get_docs_from_submodules(item, pkg_docs, options, progress_estimator):
    # Extract the package's __path__.
    if isinstance(pkg_docs[0], ModuleDoc) and pkg_docs[0].is_package:
        pkg_path = pkg_docs[0].path
        package_dir = os.path.split(pkg_docs[0].filename)[0]
    elif isinstance(pkg_docs[1], ModuleDoc) and pkg_docs[1].is_package:
        pkg_path = pkg_docs[1].path
        package_dir = os.path.split(pkg_docs[1].filename)[0]
    else:
        return []

    module_filenames = {}
    subpackage_dirs = set()
    for subdir in pkg_path:
        if os.path.isdir(subdir):
            for name in os.listdir(subdir):
                filename = os.path.join(subdir, name)
                # Is it a valid module filename?
                if is_module_file(filename):
                    basename = os.path.splitext(filename)[0]
                    if os.path.split(basename)[1] != '__init__':
                        module_filenames[basename] = filename
                # Is it a valid package filename?
                if is_package_dir(filename):
                    subpackage_dirs.add(filename)

    # Update our estimate of the number of modules in this package.
    progress_estimator.revise_estimate(item, module_filenames.items(),
                                       subpackage_dirs)

    docs = [pkg_docs]
    for module_filename in module_filenames.values():
        d = _get_docs_from_module_file(
            module_filename, options, progress_estimator, pkg_docs)
        docs.append(d)
    for subpackage_dir in subpackage_dirs:
        subpackage_file = os.path.join(subpackage_dir, '__init__')
        docs.append(_get_docs_from_module_file(
            subpackage_file, options, progress_estimator, pkg_docs))
        docs += _get_docs_from_submodules(
            subpackage_dir, docs[-1], options, progress_estimator)
    return docs

def _report_errors(name, introspect_doc, parse_doc,
                   introspect_error, parse_error):
    hdr = 'In %s:\n' % name
    if introspect_doc == parse_doc == None:
        log.start_block('%sNo documentation available!' % hdr)
        if introspect_error:
            log.error('Import failed:\n%s' % introspect_error)
        if parse_error:
            log.error('Source code parsing failed:\n%s' % parse_error)
        log.end_block()
    elif introspect_error:
        log.start_block('%sImport failed (but source code parsing '
                        'was successful).' % hdr)
        log.error(introspect_error)
        log.end_block()
    elif parse_error:
        log.start_block('%sSource code parsing failed (but '
                        'introspection was successful).' % hdr)
        log.error(parse_error)
        log.end_block()


#/////////////////////////////////////////////////////////////////
# Progress Estimation (for Documentation Generation)
#/////////////////////////////////////////////////////////////////

class _ProgressEstimator:
    """
    Used to keep track of progress when generating the initial docs
    for the given items.  (It is not known in advance how many items a
    package directory will contain, since it might depend on those
    packages' __path__ values.)
    """
    def __init__(self, items):
        self.est_totals = {}
        self.complete = 0
        
        for item in items:
            if is_package_dir(item):
                self.est_totals[item] = self._est_pkg_modules(item)
            else:
                self.est_totals[item] = 1

    def progress(self):
        total = sum(self.est_totals.values())
        return float(self.complete) / total

    def revise_estimate(self, pkg_item, modules, subpackages):
        del self.est_totals[pkg_item]
        for item in modules:
            self.est_totals[item] = 1
        for item in subpackages:
            self.est_totals[item] = self._est_pkg_modules(item)

    def _est_pkg_modules(self, package_dir):
        num_items = 0
        
        if is_package_dir(package_dir):
            for name in os.listdir(package_dir):
                filename = os.path.join(package_dir, name)
                if is_module_file(filename):
                    num_items += 1
                elif is_package_dir(filename):
                    num_items += self._est_pkg_modules(filename)
                    
        return num_items
        
######################################################################
## Doc Merger
######################################################################

MERGE_PRECEDENCE = {
    'repr': 'parse',

    # The names we get from introspection match the names that users
    # can actually use -- i.e., they take magic into account.
    'canonical_name': 'introspect',

    # Only fall-back on the parser for is_imported if the introspecter
    # isn't sure.  Otherwise, we can end up thinking that vars
    # containing modules are not imported, which can cause external
    # modules to show up in the docs (sf bug #1653486)
    'is_imported': 'introspect',

    # The parser can tell if an assignment creates an alias or not.
    'is_alias': 'parse',

    # The parser is better able to determine what text file something
    # came from; e.g., it can't be fooled by 'covert' imports.
    'docformat': 'parse',

    # The parse should be able to tell definitively whether a module
    # is a package or not.
    'is_package': 'parse',

    # Extract the sort spec from the order in which values are defined
    # in the source file.
    'sort_spec': 'parse',
    
    'submodules': 'introspect',

    # The filename used by 'parse' is the source file.
    'filename': 'parse',

    # 'parse' is more likely to get the encoding right, but
    # 'introspect' will handle programatically generated docstrings.
    # Which is better?
    'docstring': 'introspect',
    }
"""Indicates whether information from introspection or parsing should be
given precedence, for specific attributes.  This dictionary maps from
attribute names to either C{'introspect'} or C{'parse'}."""

DEFAULT_MERGE_PRECEDENCE = 'introspect'
"""Indicates whether information from introspection or parsing should be
given precedence.  Should be either C{'introspect'} or C{'parse'}"""

_attribute_mergefunc_registry = {}
def register_attribute_mergefunc(attrib, mergefunc):
    """
    Register an attribute merge function.  This function will be
    called by L{merge_docs()} when it needs to merge the attribute
    values of two C{APIDoc}s.

    @param attrib: The name of the attribute whose values are merged
    by C{mergefunc}.

    @param mergefunc: The merge function, whose sinature is:

    >>> def mergefunc(introspect_val, parse_val, precedence, cyclecheck, path):
    ...     return calculate_merged_value(introspect_val, parse_val)

    Where C{introspect_val} and C{parse_val} are the two values to
    combine; C{precedence} is a string indicating which value takes
    precedence for this attribute (C{'introspect'} or C{'parse'});
    C{cyclecheck} is a value used by C{merge_docs()} to make sure that
    it only visits each pair of docs once; and C{path} is a string
    describing the path that was taken from the root to this
    attribute (used to generate log messages).

    If the merge function needs to call C{merge_docs}, then it should
    pass C{cyclecheck} and C{path} back in.  (When appropriate, a
    suffix should be added to C{path} to describe the path taken to
    the merged values.)
    """
    _attribute_mergefunc_registry[attrib] = mergefunc

def merge_docs(introspect_doc, parse_doc, cyclecheck=None, path=None):
    """
    Merge the API documentation information that was obtained from
    introspection with information that was obtained from parsing.
    C{introspect_doc} and C{parse_doc} should be two C{APIDoc} instances
    that describe the same object.  C{merge_docs} combines the
    information from these two instances, and returns the merged
    C{APIDoc}.

    If C{introspect_doc} and C{parse_doc} are compatible, then they will
    be I{merged} -- i.e., they will be coerced to a common class, and
    their state will be stored in a shared dictionary.  Once they have
    been merged, any change made to the attributes of one will affect
    the other.  The value for the each of the merged C{APIDoc}'s
    attributes is formed by combining the values of the source
    C{APIDoc}s' attributes, as follows:

      - If either of the source attributes' value is C{UNKNOWN}, then
        use the other source attribute's value.
      - Otherwise, if an attribute merge function has been registered
        for the attribute, then use that function to calculate the
        merged value from the two source attribute values.
      - Otherwise, if L{MERGE_PRECEDENCE} is defined for the
        attribute, then use the attribute value from the source that
        it indicates.
      - Otherwise, use the attribute value from the source indicated
        by L{DEFAULT_MERGE_PRECEDENCE}.

    If C{introspect_doc} and C{parse_doc} are I{not} compatible (e.g., if
    their values have incompatible types), then C{merge_docs()} will
    simply return either C{introspect_doc} or C{parse_doc}, depending on
    the value of L{DEFAULT_MERGE_PRECEDENCE}.  The two input
    C{APIDoc}s will not be merged or modified in any way.

    @param cyclecheck, path: These arguments should only be provided
        when C{merge_docs()} is called by an attribute merge
        function.  See L{register_attribute_mergefunc()} for more
        details.
    """
    assert isinstance(introspect_doc, APIDoc)
    assert isinstance(parse_doc, APIDoc)

    if cyclecheck is None:
        cyclecheck = set()
        if introspect_doc.canonical_name not in (None, UNKNOWN):
            path = '%s' % introspect_doc.canonical_name
        elif parse_doc.canonical_name not in (None, UNKNOWN):
            path = '%s' % parse_doc.canonical_name
        else:
            path = '??'

    # If we've already examined this pair, then there's nothing
    # more to do.  The reason that we check id's here is that we
    # want to avoid hashing the APIDoc objects for now, so we can
    # use APIDoc.merge_and_overwrite() later.
    if (id(introspect_doc), id(parse_doc)) in cyclecheck:
        return introspect_doc
    cyclecheck.add( (id(introspect_doc), id(parse_doc)) )

    # If these two are already merged, then we're done.  (Two
    # APIDoc's compare equal iff they are identical or have been
    # merged.)
    if introspect_doc == parse_doc:
        return introspect_doc

    # If both values are GenericValueDoc, then we don't want to merge
    # them.  E.g., we don't want to merge 2+2 with 4.  So just copy
    # the parse_doc's parse_repr to introspect_doc, & return it.
    # (In particular, do *not* call merge_and_overwrite.)
    if type(introspect_doc) == type(parse_doc) == GenericValueDoc:
        if parse_doc.parse_repr is not UNKNOWN:
            introspect_doc.parse_repr = parse_doc.parse_repr
        introspect_doc.docs_extracted_by = 'both'
        return introspect_doc

    # Perform several sanity checks here -- if we accidentally
    # merge values that shouldn't get merged, then bad things can
    # happen.
    mismatch = None
    if (introspect_doc.__class__ != parse_doc.__class__ and
        not (issubclass(introspect_doc.__class__, parse_doc.__class__) or
             issubclass(parse_doc.__class__, introspect_doc.__class__))):
        mismatch = ("value types don't match -- i=%r, p=%r." %
                    (introspect_doc.__class__, parse_doc.__class__))
    if (isinstance(introspect_doc, ValueDoc) and
        isinstance(parse_doc, ValueDoc)):
        if (introspect_doc.pyval is not UNKNOWN and
            parse_doc.pyval is not UNKNOWN and
            introspect_doc.pyval is not parse_doc.pyval):
            mismatch = "values don't match."
        elif (introspect_doc.canonical_name not in (None, UNKNOWN) and
            parse_doc.canonical_name not in (None, UNKNOWN) and
            introspect_doc.canonical_name != parse_doc.canonical_name):
            mismatch = "canonical names don't match."
    if mismatch is not None:
        log.info("Not merging the parsed & introspected values of %s, "
                 "since their %s" % (path, mismatch))
        if DEFAULT_MERGE_PRECEDENCE == 'introspect':
            return introspect_doc
        else:
            return parse_doc

    # If one apidoc's class is a superclass of the other's, then
    # specialize it to the more specific class.
    if introspect_doc.__class__ is not parse_doc.__class__:
        if issubclass(introspect_doc.__class__, parse_doc.__class__):
            parse_doc.specialize_to(introspect_doc.__class__)
        if issubclass(parse_doc.__class__, introspect_doc.__class__):
            introspect_doc.specialize_to(parse_doc.__class__)
    assert introspect_doc.__class__ is parse_doc.__class__

    # The posargs and defaults are tied together -- if we merge
    # the posargs one way, then we need to merge the defaults the
    # same way.  So check them first.  (This is a minor hack)
    if (isinstance(introspect_doc, RoutineDoc) and
        isinstance(parse_doc, RoutineDoc)):
        _merge_posargs_and_defaults(introspect_doc, parse_doc, path)
    
    # Merge the two api_doc's attributes.
    for attrib in set(introspect_doc.__dict__.keys() +
                      parse_doc.__dict__.keys()):
        # Be sure not to merge any private attributes (especially
        # __mergeset or __has_been_hashed!)
        if attrib.startswith('_'): continue
        merge_attribute(attrib, introspect_doc, parse_doc,
                             cyclecheck, path)

    # Set the dictionaries to be shared.
    return introspect_doc.merge_and_overwrite(parse_doc)

def _merge_posargs_and_defaults(introspect_doc, parse_doc, path):
    # If either is unknown, then let merge_attrib handle it.
    if introspect_doc.posargs is UNKNOWN or parse_doc.posargs is UNKNOWN:
        return 
        
    # If the introspected doc just has '...', then trust the parsed doc.
    if introspect_doc.posargs == ['...'] and parse_doc.posargs != ['...']:
        introspect_doc.posargs = parse_doc.posargs
        introspect_doc.posarg_defaults = parse_doc.posarg_defaults

    # If they are incompatible, then check the precedence.
    elif introspect_doc.posargs != parse_doc.posargs:
        log.info("Not merging the parsed & introspected arg "
                 "lists for %s, since they don't match (%s vs %s)"
                  % (path, introspect_doc.posargs, parse_doc.posargs))
        if (MERGE_PRECEDENCE.get('posargs', DEFAULT_MERGE_PRECEDENCE) ==
            'introspect'):
            parse_doc.posargs = introspect_doc.posargs
            parse_doc.posarg_defaults = introspect_doc.posarg_defaults
        else:
            introspect_doc.posargs = parse_doc.posargs
            introspect_doc.posarg_defaults = parse_doc.posarg_defaults

def merge_attribute(attrib, introspect_doc, parse_doc, cyclecheck, path):
    precedence = MERGE_PRECEDENCE.get(attrib, DEFAULT_MERGE_PRECEDENCE)
    if precedence not in ('parse', 'introspect'):
        raise ValueError('Bad precedence value %r' % precedence)
    
    if (getattr(introspect_doc, attrib) is UNKNOWN and
        getattr(parse_doc, attrib) is not UNKNOWN):
        setattr(introspect_doc, attrib, getattr(parse_doc, attrib))
    elif (getattr(introspect_doc, attrib) is not UNKNOWN and
          getattr(parse_doc, attrib) is UNKNOWN):
        setattr(parse_doc, attrib, getattr(introspect_doc, attrib))
    elif (getattr(introspect_doc, attrib) is UNKNOWN and
          getattr(parse_doc, attrib) is UNKNOWN):
        pass
    else:
        # Both APIDoc objects have values; we need to merge them.
        introspect_val = getattr(introspect_doc, attrib)
        parse_val = getattr(parse_doc, attrib)
        if attrib in _attribute_mergefunc_registry:
            handler = _attribute_mergefunc_registry[attrib]
            merged_val = handler(introspect_val, parse_val, precedence,
                                 cyclecheck, path)
        elif precedence == 'introspect':
            merged_val = introspect_val
        elif precedence == 'parse':
            merged_val = parse_val

        setattr(introspect_doc, attrib, merged_val)
        setattr(parse_doc, attrib, merged_val)

def merge_variables(varlist1, varlist2, precedence, cyclecheck, path):
    # Merge all variables that are in both sets.
    for varname, var1 in varlist1.items():
        var2 = varlist2.get(varname)
        if var2 is not None:
            var = merge_docs(var1, var2, cyclecheck, path+'.'+varname)
            varlist1[varname] = var
            varlist2[varname] = var

    # Copy any variables that are not in varlist1 over.
    for varname, var in varlist2.items():
        varlist1.setdefault(varname, var)

    return varlist1

def merge_value(value1, value2, precedence, cyclecheck, path):
    assert value1 is not None and value2 is not None
    return merge_docs(value1, value2, cyclecheck, path)

def merge_overrides(v1, v2, precedence, cyclecheck, path):
    return merge_value(v1, v2, precedence, cyclecheck, path+'.<overrides>')
def merge_fget(v1, v2, precedence, cyclecheck, path):
    return merge_value(v1, v2, precedence, cyclecheck, path+'.fget')
def merge_fset(v1, v2, precedence, cyclecheck, path):
    return merge_value(v1, v2, precedence, cyclecheck, path+'.fset')
def merge_fdel(v1, v2, precedence, cyclecheck, path):
    return merge_value(v1, v2, precedence, cyclecheck, path+'.fdel')

def merge_proxy_for(v1, v2, precedence, cyclecheck, path):
    # Anything we got from introspection shouldn't have a proxy_for
    # attribute -- it should be the actual object's documentation.
    return v1

def merge_bases(baselist1, baselist2, precedence, cyclecheck, path):
    # Be careful here -- if we get it wrong, then we could end up
    # merging two unrelated classes, which could lead to bad
    # things (e.g., a class that's its own subclass).  So only
    # merge two bases if we're quite sure they're the same class.
    # (In particular, if they have the same canonical name.)

    # If the lengths don't match up, then give up.  This is most
    # often caused by __metaclass__.
    if len(baselist1) != len(baselist2):
        log.info("Not merging the introspected & parsed base lists "
                 "for %s, since their lengths don't match (%s vs %s)" %
                 (path, len(baselist1), len(baselist2)))
        if precedence == 'introspect': return baselist1
        else: return baselist2

    # If any names disagree, then give up.
    for base1, base2 in zip(baselist1, baselist2):
        if ((base1.canonical_name not in (None, UNKNOWN) and
             base2.canonical_name not in (None, UNKNOWN)) and
            base1.canonical_name != base2.canonical_name):
            log.info("Not merging the parsed & introspected base "
                     "lists for %s, since the bases' names don't match "
                     "(%s vs %s)" % (path, base1.canonical_name,
                                     base2.canonical_name))
            if precedence == 'introspect': return baselist1
            else: return baselist2

    for i, (base1, base2) in enumerate(zip(baselist1, baselist2)):
        base = merge_docs(base1, base2, cyclecheck,
                           '%s.__bases__[%d]' % (path, i))
        baselist1[i] = baselist2[i] = base

    return baselist1

def merge_posarg_defaults(defaults1, defaults2, precedence, cyclecheck, path):
    if len(defaults1) != len(defaults2):
        if precedence == 'introspect': return defaults1
        else: return defaults2
    defaults = []
    for i, (d1, d2) in enumerate(zip(defaults1, defaults2)):
        if d1 is not None and d2 is not None:
            d_path = '%s.<default-arg-val>[%d]' % (path, i)
            defaults.append(merge_docs(d1, d2, cyclecheck, d_path))
        elif precedence == 'introspect':
            defaults.append(d1)
        else:
            defaults.append(d2)
    return defaults

def merge_docstring(docstring1, docstring2, precedence, cyclecheck, path):
    if docstring1 is None or docstring1 is UNKNOWN or precedence=='parse':
        return docstring2
    else:
        return docstring1

def merge_docs_extracted_by(v1, v2, precedence, cyclecheck, path):
    return 'both'

def merge_submodules(v1, v2, precedence, cyclecheck, path):
    n1 = sorted([m.canonical_name for m in v1])
    n2 = sorted([m.canonical_name for m in v2])
    if (n1 != n2) and (n2 != []):
        log.info('Introspector & parser disagree about submodules '
                 'for %s: (%s) vs (%s)' % (path,
                                           ', '.join([str(n) for n in n1]),
                                           ', '.join([str(n) for n in n2])))
        return v1 + [m for m in v2 if m.canonical_name not in n1]
                
    return v1

register_attribute_mergefunc('variables', merge_variables)
register_attribute_mergefunc('value', merge_value)
register_attribute_mergefunc('overrides', merge_overrides)
register_attribute_mergefunc('fget', merge_fget)
register_attribute_mergefunc('fset', merge_fset)
register_attribute_mergefunc('fdel', merge_fdel)
register_attribute_mergefunc('proxy_for', merge_proxy_for)
register_attribute_mergefunc('bases', merge_bases)
register_attribute_mergefunc('posarg_defaults', merge_posarg_defaults)
register_attribute_mergefunc('docstring', merge_docstring)
register_attribute_mergefunc('docs_extracted_by', merge_docs_extracted_by)
register_attribute_mergefunc('submodules', merge_submodules)

######################################################################
## Import Linking
######################################################################

def link_imports(val_doc, docindex):
    # Check if the ValueDoc has an unresolved proxy_for link.
    # If so, then resolve it.
    while val_doc.proxy_for not in (UNKNOWN, None):
        # Find the valuedoc that the proxy_for name points to.
        src_doc = docindex.get_valdoc(val_doc.proxy_for)

        # If we don't have any valuedoc at that address, then
        # set that address as its canonical name.
        # [XXX] Do I really want to do this?
        if src_doc is None:
            val_doc.canonical_name = val_doc.proxy_for
            return

        # If we *do* have something at that address, then
        # merge the proxy `val_doc` with it.
        elif src_doc != val_doc:
            # Copy any subclass information from val_doc->src_doc.
            if (isinstance(val_doc, ClassDoc) and
                isinstance(src_doc, ClassDoc)):
                for subclass in val_doc.subclasses:
                    if subclass not in src_doc.subclasses:
                        src_doc.subclasses.append(subclass)
            # Then overwrite val_doc with the contents of src_doc.
            src_doc.merge_and_overwrite(val_doc, ignore_hash_conflict=True)

        # If the proxy_for link points back at src_doc
        # itself, then we most likely have a variable that's
        # shadowing a submodule that it should be equal to.
        # So just get rid of the variable.
        elif src_doc == val_doc:
            parent_name = val_doc.proxy_for[:-1]
            var_name = val_doc.proxy_for[-1]
            parent = docindex.get_valdoc(parent_name)
            if parent is not None and var_name in parent.variables:
                del parent.variables[var_name]
            src_doc.proxy_for = None

######################################################################
## Canonical Name Assignment
######################################################################

_name_scores = {}
"""A dictionary mapping from each C{ValueDoc} to the score that has
been assigned to its current cannonical name.  If
L{assign_canonical_names()} finds a canonical name with a better
score, then it will replace the old name."""

_unreachable_names = {DottedName(DottedName.UNREACHABLE):1}
"""The set of names that have been used for unreachable objects.  This
is used to ensure there are no duplicate cannonical names assigned.
C{_unreachable_names} is a dictionary mapping from dotted names to
integer ids, where the next unused unreachable name derived from
dotted name C{n} is
C{DottedName('%s-%s' % (n, str(_unreachable_names[n]+1))}"""

def assign_canonical_names(val_doc, name, docindex, score=0):
    """
    Assign a canonical name to C{val_doc} (if it doesn't have one
    already), and (recursively) to each variable in C{val_doc}.
    In particular, C{val_doc} will be assigned the canonical name
    C{name} iff either:
      - C{val_doc}'s canonical name is C{UNKNOWN}; or
      - C{val_doc}'s current canonical name was assigned by this
        method; but the score of the new name (C{score}) is higher
        than the score of the current name (C{score_dict[val_doc]}).
        
    Note that canonical names will even be assigned to values
    like integers and C{None}; but these should be harmless.
    """
    # If we've already visited this node, and our new score
    # doesn't beat our old score, then there's nothing more to do.
    # Note that since score increases strictly monotonically, this
    # also prevents us from going in cycles.
    if val_doc in _name_scores and score <= _name_scores[val_doc]:
        return

    # Update val_doc's canonical name, if appropriate.
    if (val_doc not in _name_scores and
        val_doc.canonical_name is not UNKNOWN):
        # If this is the first time we've seen val_doc, and it
        # already has a name, then don't change that name.
        _name_scores[val_doc] = sys.maxint
        name = val_doc.canonical_name
        score = 0
    else:
        # Otherwise, update the name iff the new score is better
        # than the old one.
        if (val_doc not in _name_scores or
            score > _name_scores[val_doc]):
            val_doc.canonical_name = name
            _name_scores[val_doc] = score

    # Recurse to any contained values.
    if isinstance(val_doc, NamespaceDoc):
        for var_doc in val_doc.variables.values():
            # Set the variable's canonical name.
            varname = DottedName(name, var_doc.name)
            var_doc.canonical_name = varname

            # If the value is unknown, or is a generic value doc, then
            # the valuedoc doesn't get assigned a name; move on.
            if (var_doc.value is UNKNOWN
                or isinstance(var_doc.value, GenericValueDoc)):
                continue
            
            # [XX] After svn commit 1644-1647, I'm not sure if this
            # ever gets used:  This check is for cases like
            # curses.wrapper, where an imported variable shadows its
            # value's "real" location.
            if _var_shadows_self(var_doc, varname):
                _fix_self_shadowing_var(var_doc, varname, docindex)
    
            # Find the score for this new name.            
            vardoc_score = score-1
            if var_doc.is_imported is UNKNOWN: vardoc_score -= 10
            elif var_doc.is_imported: vardoc_score -= 100
            if var_doc.is_alias is UNKNOWN: vardoc_score -= 10
            elif var_doc.is_alias: vardoc_score -= 1000
            
            assign_canonical_names(var_doc.value, varname,
                                   docindex, vardoc_score)

    # Recurse to any directly reachable values.
    for val_doc_2 in val_doc.apidoc_links(variables=False):
        val_name, val_score = _unreachable_name_for(val_doc_2, docindex)
        assign_canonical_names(val_doc_2, val_name, docindex, val_score)

def _var_shadows_self(var_doc, varname):
    return (var_doc.value not in (None, UNKNOWN) and
            var_doc.value.canonical_name not in (None, UNKNOWN) and
            var_doc.value.canonical_name != varname and
            varname.dominates(var_doc.value.canonical_name))

def _fix_self_shadowing_var(var_doc, varname, docindex):
    # If possible, find another name for the shadowed value.
    cname = var_doc.value.canonical_name
    for i in range(1, len(cname)-1):
        new_name = cname[:i] + (cname[i]+"'") + cname[i+1:]
        val_doc = docindex.get_valdoc(new_name)
        if val_doc is not None:
            log.warning("%s shadows its own value -- using %s instead" %
                     (varname, new_name))
            var_doc.value = val_doc
            return

    # If we couldn't find the actual value, use an unreachable name.
    name, score = _unreachable_name_for(var_doc.value, docindex)
    log.warning('%s shadows itself -- using %s instead' % (varname, name))
    var_doc.value.canonical_name = name

def _unreachable_name_for(val_doc, docindex):
    assert isinstance(val_doc, ValueDoc)
    
    # [xx] (when) does this help?
    if (isinstance(val_doc, ModuleDoc) and
        len(val_doc.canonical_name)==1 and val_doc.package is None):
        for root_val in docindex.root:
            if root_val.canonical_name == val_doc.canonical_name:
                if root_val != val_doc: 
                    log.error("Name conflict: %r vs %r" %
                              (val_doc, root_val))
                break
        else:
            return val_doc.canonical_name, -1000

    # Assign it an 'unreachable' name:
    if (val_doc.pyval is not UNKNOWN and
          hasattr(val_doc.pyval, '__name__')):
        try:
            name = DottedName(DottedName.UNREACHABLE,
                              val_doc.pyval.__name__, strict=True)
        except DottedName.InvalidDottedName:
            name = DottedName(DottedName.UNREACHABLE)
    else:
        name = DottedName(DottedName.UNREACHABLE)

    # Uniquify the name.
    if name in _unreachable_names:
        _unreachable_names[name] += 1
        name = DottedName('%s-%s' % (name, _unreachable_names[name]-1))
    else:
        _unreachable_names[name] = 1
    
    return name, -10000

######################################################################
## Documentation Inheritance
######################################################################

def find_overrides(class_doc):
    """
    Set the C{overrides} attribute for all variables in C{class_doc}.
    This needs to be done early (before docstring parsing), so we can
    know which docstrings to suppress warnings for.
    """
    for base_class in list(class_doc.mro(warn_about_bad_bases=True)):
        if base_class == class_doc: continue
        if base_class.variables is UNKNOWN: continue
        for name, var_doc in base_class.variables.items():
            if ( not (name.startswith('__') and not name.endswith('__')) and
                 base_class == var_doc.container and
                 name in class_doc.variables and 
                 class_doc.variables[name].container==class_doc and
                 class_doc.variables[name].overrides is UNKNOWN ):
                class_doc.variables[name].overrides = var_doc
    
    
def inherit_docs(class_doc):
    for base_class in list(class_doc.mro(warn_about_bad_bases=True)):
        if base_class == class_doc: continue

        # Inherit any groups.  Place them *after* this class's groups,
        # so that any groups that are important to this class come
        # first.
        if base_class.group_specs not in (None, UNKNOWN):
            class_doc.group_specs += [gs for gs in base_class.group_specs
                                      if gs not in class_doc.group_specs]

        # Inherit any variables.
        if base_class.variables is UNKNOWN: continue
        for name, var_doc in base_class.variables.items():
            # If it's a __private variable, then don't inherit it.
            if name.startswith('__') and not name.endswith('__'):
                continue
            
            # Inhetit only from the defining class. Or else, in case of
            # multiple inheritance, we may import from a grand-ancestor
            # variables overridden by a class that follows in mro.
            if base_class != var_doc.container:
                continue
            
            # If class_doc doesn't have a variable with this name,
            # then inherit it.
            if name not in class_doc.variables:
                class_doc.variables[name] = var_doc

            # Otherwise, class_doc already contains a variable
            # that shadows var_doc.  But if class_doc's var is
            # local, then record the fact that it overrides
            # var_doc.
            elif class_doc.variables[name].container==class_doc:
                class_doc.variables[name].overrides = var_doc
                _inherit_info(class_doc.variables[name])

_INHERITED_ATTRIBS = [
    'descr', 'summary', 'metadata', 'extra_docstring_fields',
    'type_descr', 'arg_descrs', 'arg_types', 'return_descr',
    'return_type', 'exception_descrs']

_method_descriptor = type(list.append)

def _inherit_info(var_doc):
    """
    Copy any relevant documentation information from the variable that
    C{var_doc} overrides into C{var_doc} itself.
    """
    src_var = var_doc.overrides
    src_val = var_doc.overrides.value
    val_doc = var_doc.value

    # Special case: if the source value and target values are both c
    # extension methods, and the target value's signature is not
    # specified, then inherit the source value's signature.
    if (isinstance(val_doc, RoutineDoc) and
        isinstance(src_val, RoutineDoc) and
        (inspect.isbuiltin(val_doc.pyval) or
         isinstance(val_doc.pyval, _method_descriptor)) and
        (inspect.isbuiltin(src_val.pyval) or
         isinstance(src_val.pyval, _method_descriptor)) and
        val_doc.all_args() in (['...'], UNKNOWN) and
        src_val.all_args() not in (['...'], UNKNOWN)):
        for attrib in ['posargs', 'posarg_defaults', 'vararg',
                       'kwarg', 'return_type']:
            setattr(val_doc, attrib, getattr(src_val, attrib))
    
    # If the new variable has a docstring, then don't inherit
    # anything, even if the docstring is blank.
    if var_doc.docstring not in (None, UNKNOWN):
        return
    # [xx] Do I want a check like this:?
#     # If it's a method and the signature doesn't match well enough,
#     # then give up.
#     if (isinstance(src_val, RoutineDoc) and
#         isinstance(val_doc, RoutineDoc)):
#         if (src_val.posargs != val_doc.posargs[:len(src_val.posargs)] or
#             src_val.vararg != None and src_val.vararg != val_doc.vararg):
#             log.docstring_warning(
#                 "The signature of %s does not match the signature of the "
#                 "method it overrides (%s); not inheriting documentation." %
#                 (var_doc.canonical_name, src_var.canonical_name))
#             return

    # Inherit attributes!
    for attrib in _INHERITED_ATTRIBS:
        if (hasattr(var_doc, attrib) and hasattr(src_var, attrib) and
            getattr(src_var, attrib) not in (None, UNKNOWN)):
            setattr(var_doc, attrib, getattr(src_var, attrib))
        elif (src_val is not None and
              hasattr(val_doc, attrib) and hasattr(src_val, attrib) and
              getattr(src_val, attrib) not in (None, UNKNOWN) and
              getattr(val_doc, attrib) in (None, UNKNOWN, [])):
            setattr(val_doc, attrib, getattr(src_val, attrib))