Lib/imputil.py

   1 """
   2 Import utilities
   3
   4 Exported classes:
   5     ImportManager   Manage the import process
   6
   7     Importer        Base class for replacing standard import functions
   8     BuiltinImporter Emulate the import mechanism for builtin and frozen modules
   9
  10     DynLoadSuffixImporter
  11 """
  12
  13 # note: avoid importing non-builtin modules
  14 import imp                      ### not available in JPython?
  15 import sys
  16 import __builtin__
  17
  18 # for the DirectoryImporter
  19 import struct
  20 import marshal
  21
  22 __all__ = ["ImportManager","Importer","BuiltinImporter"]
  23
  24 _StringType = type('')
  25 _ModuleType = type(sys)         ### doesn't work in JPython...
  26
  27 class ImportManager:
  28     "Manage the import process."
  29
  30     def install(self, namespace=vars(__builtin__)):
  31         "Install this ImportManager into the specified namespace."
  32
  33         if isinstance(namespace, _ModuleType):
  34             namespace = vars(namespace)
  35
  36         # Note: we have no notion of "chaining"
  37
  38         # Record the previous import hook, then install our own.
  39         self.previous_importer = namespace['__import__']
  40         self.namespace = namespace
  41         namespace['__import__'] = self._import_hook
  42
  43         ### fix this
  44         #namespace['reload'] = self._reload_hook
  45
  46     def uninstall(self):
  47         "Restore the previous import mechanism."
  48         self.namespace['__import__'] = self.previous_importer
  49
  50     def add_suffix(self, suffix, importFunc):
  51         assert callable(importFunc)
  52         self.fs_imp.add_suffix(suffix, importFunc)
  53
  54     ######################################################################
  55     #
  56     # PRIVATE METHODS
  57     #
  58
  59     clsFilesystemImporter = None
  60
  61     def __init__(self, fs_imp=None):
  62         # we're definitely going to be importing something in the future,
  63         # so let's just load the OS-related facilities.
  64         if not _os_stat:
  65             _os_bootstrap()
  66
  67         # This is the Importer that we use for grabbing stuff from the
  68         # filesystem. It defines one more method (import_from_dir) for our use.
  69         if fs_imp is None:
  70             cls = self.clsFilesystemImporter or _FilesystemImporter
  71             fs_imp = cls()
  72         self.fs_imp = fs_imp
  73
  74         # Initialize the set of suffixes that we recognize and import.
  75         # The default will import dynamic-load modules first, followed by
  76         # .py files (or a .py file's cached bytecode)
  77         for desc in imp.get_suffixes():
  78             if desc[2] == imp.C_EXTENSION:
  79                 self.add_suffix(desc[0],
  80                                 DynLoadSuffixImporter(desc).import_file)
  81         self.add_suffix('.py', py_suffix_importer)
  82
  83     def _import_hook(self, fqname, globals=None, locals=None, fromlist=None):
  84         """Python calls this hook to locate and import a module."""
  85
  86         parts = fqname.split('.')
  87
  88         # determine the context of this import
  89         parent = self._determine_import_context(globals)
  90
  91         # if there is a parent, then its importer should manage this import
  92         if parent:
  93             module = parent.__importer__._do_import(parent, parts, fromlist)
  94             if module:
  95                 return module
  96
  97         # has the top module already been imported?
  98         try:
  99             top_module = sys.modules[parts[0]]
 100         except KeyError:
 101
 102             # look for the topmost module
 103             top_module = self._import_top_module(parts[0])
 104             if not top_module:
 105                 # the topmost module wasn't found at all.
 106                 raise ImportError, 'No module named ' + fqname
 107
 108         # fast-path simple imports
 109         if len(parts) == 1:
 110             if not fromlist:
 111                 return top_module
 112
 113             if not top_module.__dict__.get('__ispkg__'):
 114                 # __ispkg__ isn't defined (the module was not imported by us),
 115                 # or it is zero.
 116                 #
 117                 # In the former case, there is no way that we could import
 118                 # sub-modules that occur in the fromlist (but we can't raise an
 119                 # error because it may just be names) because we don't know how
 120                 # to deal with packages that were imported by other systems.
 121                 #
 122                 # In the latter case (__ispkg__ == 0), there can't be any sub-
 123                 # modules present, so we can just return.
 124                 #
 125                 # In both cases, since len(parts) == 1, the top_module is also
 126                 # the "bottom" which is the defined return when a fromlist
 127                 # exists.
 128                 return top_module
 129
 130         importer = top_module.__dict__.get('__importer__')
 131         if importer:
 132             return importer._finish_import(top_module, parts[1:], fromlist)
 133
 134         # Grrr, some people "import os.path"
 135         if len(parts) == 2 and hasattr(top_module, parts[1]):
 136             return top_module
 137
 138         # If the importer does not exist, then we have to bail. A missing
 139         # importer means that something else imported the module, and we have
 140         # no knowledge of how to get sub-modules out of the thing.
 141         raise ImportError, 'No module named ' + fqname
 142
 143     def _determine_import_context(self, globals):
 144         """Returns the context in which a module should be imported.
 145
 146         The context could be a loaded (package) module and the imported module
 147         will be looked for within that package. The context could also be None,
 148         meaning there is no context -- the module should be looked for as a
 149         "top-level" module.
 150         """
 151
 152         if not globals or not globals.get('__importer__'):
 153             # globals does not refer to one of our modules or packages. That
 154             # implies there is no relative import context (as far as we are
 155             # concerned), and it should just pick it off the standard path.
 156             return None
 157
 158         # The globals refer to a module or package of ours. It will define
 159         # the context of the new import. Get the module/package fqname.
 160         parent_fqname = globals['__name__']
 161
 162         # if a package is performing the import, then return itself (imports
 163         # refer to pkg contents)
 164         if globals['__ispkg__']:
 165             parent = sys.modules[parent_fqname]
 166             assert globals is parent.__dict__
 167             return parent
 168
 169         i = parent_fqname.rfind('.')
 170
 171         # a module outside of a package has no particular import context
 172         if i == -1:
 173             return None
 174
 175         # if a module in a package is performing the import, then return the
 176         # package (imports refer to siblings)
 177         parent_fqname = parent_fqname[:i]
 178         parent = sys.modules[parent_fqname]
 179         assert parent.__name__ == parent_fqname
 180         return parent
 181
 182     def _import_top_module(self, name):
 183         # scan sys.path looking for a location in the filesystem that contains
 184         # the module, or an Importer object that can import the module.
 185         for item in sys.path:
 186             if isinstance(item, _StringType):
 187                 module = self.fs_imp.import_from_dir(item, name)
 188             else:
 189                 module = item.import_top(name)
 190             if module:
 191                 return module
 192         return None
 193
 194     def _reload_hook(self, module):
 195         "Python calls this hook to reload a module."
 196
 197         # reloading of a module may or may not be possible (depending on the
 198         # importer), but at least we can validate that it's ours to reload
 199         importer = module.__dict__.get('__importer__')
 200         if not importer:
 201             ### oops. now what...
 202             pass
 203
 204         # okay. it is using the imputil system, and we must delegate it, but
 205         # we don't know what to do (yet)
 206         ### we should blast the module dict and do another get_code(). need to
 207         ### flesh this out and add proper docco...
 208         raise SystemError, "reload not yet implemented"
 209
 210
 211 class Importer:
 212     "Base class for replacing standard import functions."
 213
 214     def import_top(self, name):
 215         "Import a top-level module."
 216         return self._import_one(None, name, name)
 217
 218     ######################################################################
 219     #
 220     # PRIVATE METHODS
 221     #
 222     def _finish_import(self, top, parts, fromlist):
 223         # if "a.b.c" was provided, then load the ".b.c" portion down from
 224         # below the top-level module.
 225         bottom = self._load_tail(top, parts)
 226
 227         # if the form is "import a.b.c", then return "a"
 228         if not fromlist:
 229             # no fromlist: return the top of the import tree
 230             return top
 231
 232         # the top module was imported by self.
 233         #
 234         # this means that the bottom module was also imported by self (just
 235         # now, or in the past and we fetched it from sys.modules).
 236         #
 237         # since we imported/handled the bottom module, this means that we can
 238         # also handle its fromlist (and reliably use __ispkg__).
 239
 240         # if the bottom node is a package, then (potentially) import some
 241         # modules.
 242         #
 243         # note: if it is not a package, then "fromlist" refers to names in
 244         #       the bottom module rather than modules.
 245         # note: for a mix of names and modules in the fromlist, we will
 246         #       import all modules and insert those into the namespace of
 247         #       the package module. Python will pick up all fromlist names
 248         #       from the bottom (package) module; some will be modules that
 249         #       we imported and stored in the namespace, others are expected
 250         #       to be present already.
 251         if bottom.__ispkg__:
 252             self._import_fromlist(bottom, fromlist)
 253
 254         # if the form is "from a.b import c, d" then return "b"
 255         return bottom
 256
 257     def _import_one(self, parent, modname, fqname):
 258         "Import a single module."
 259
 260         # has the module already been imported?
 261         try:
 262             return sys.modules[fqname]
 263         except KeyError:
 264             pass
 265
 266         # load the module's code, or fetch the module itself
 267         result = self.get_code(parent, modname, fqname)
 268         if result is None:
 269             return None
 270
 271         module = self._process_result(result, fqname)
 272
 273         # insert the module into its parent
 274         if parent:
 275             setattr(parent, modname, module)
 276         return module
 277
 278     def _process_result(self, (ispkg, code, values), fqname):
 279         # did get_code() return an actual module? (rather than a code object)
 280         is_module = isinstance(code, _ModuleType)
 281
 282         # use the returned module, or create a new one to exec code into
 283         if is_module:
 284             module = code
 285         else:
 286             module = imp.new_module(fqname)
 287
 288         ### record packages a bit differently??
 289         module.__importer__ = self
 290         module.__ispkg__ = ispkg
 291
 292         # insert additional values into the module (before executing the code)
 293         module.__dict__.update(values)
 294
 295         # the module is almost ready... make it visible
 296         sys.modules[fqname] = module
 297
 298         # execute the code within the module's namespace
 299         if not is_module:
 300             exec code in module.__dict__
 301
 302         # fetch from sys.modules instead of returning module directly.
 303         # also make module's __name__ agree with fqname, in case
 304         # the "exec code in module.__dict__" played games on us.
 305         module = sys.modules[fqname]
 306         module.__name__ = fqname
 307         return module
 308
 309     def _load_tail(self, m, parts):
 310         """Import the rest of the modules, down from the top-level module.
 311
 312         Returns the last module in the dotted list of modules.
 313         """
 314         for part in parts:
 315             fqname = "%s.%s" % (m.__name__, part)
 316             m = self._import_one(m, part, fqname)
 317             if not m:
 318                 raise ImportError, "No module named " + fqname
 319         return m
 320
 321     def _import_fromlist(self, package, fromlist):
 322         'Import any sub-modules in the "from" list.'
 323
 324         # if '*' is present in the fromlist, then look for the '__all__'
 325         # variable to find additional items (modules) to import.
 326         if '*' in fromlist:
 327             fromlist = list(fromlist) + \
 328                        list(package.__dict__.get('__all__', []))
 329
 330         for sub in fromlist:
 331             # if the name is already present, then don't try to import it (it
 332             # might not be a module!).
 333             if sub != '*' and not hasattr(package, sub):
 334                 subname = "%s.%s" % (package.__name__, sub)
 335                 submod = self._import_one(package, sub, subname)
 336                 if not submod:
 337                     raise ImportError, "cannot import name " + subname
 338
 339     def _do_import(self, parent, parts, fromlist):
 340         """Attempt to import the module relative to parent.
 341
 342         This method is used when the import context specifies that <self>
 343         imported the parent module.
 344         """
 345         top_name = parts[0]
 346         top_fqname = parent.__name__ + '.' + top_name
 347         top_module = self._import_one(parent, top_name, top_fqname)
 348         if not top_module:
 349             # this importer and parent could not find the module (relatively)
 350             return None
 351
 352         return self._finish_import(top_module, parts[1:], fromlist)
 353
 354     ######################################################################
 355     #
 356     # METHODS TO OVERRIDE
 357     #
 358     def get_code(self, parent, modname, fqname):
 359         """Find and retrieve the code for the given module.
 360
 361         parent specifies a parent module to define a context for importing. It
 362         may be None, indicating no particular context for the search.
 363
 364         modname specifies a single module (not dotted) within the parent.
 365
 366         fqname specifies the fully-qualified module name. This is a
 367         (potentially) dotted name from the "root" of the module namespace
 368         down to the modname.
 369         If there is no parent, then modname==fqname.
 370
 371         This method should return None, or a 3-tuple.
 372
 373         * If the module was not found, then None should be returned.
 374
 375         * The first item of the 2- or 3-tuple should be the integer 0 or 1,
 376             specifying whether the module that was found is a package or not.
 377
 378         * The second item is the code object for the module (it will be
 379             executed within the new module's namespace). This item can also
 380             be a fully-loaded module object (e.g. loaded from a shared lib).
 381
 382         * The third item is a dictionary of name/value pairs that will be
 383             inserted into new module before the code object is executed. This
 384             is provided in case the module's code expects certain values (such
 385             as where the module was found). When the second item is a module
 386             object, then these names/values will be inserted *after* the module
 387             has been loaded/initialized.
 388         """
 389         raise RuntimeError, "get_code not implemented"
 390
 391
 392 ######################################################################
 393 #
 394 # Some handy stuff for the Importers
 395 #
 396
 397 # byte-compiled file suffix character
 398 _suffix_char = __debug__ and 'c' or 'o'
 399
 400 # byte-compiled file suffix
 401 _suffix = '.py' + _suffix_char
 402
 403 def _compile(pathname, timestamp):
 404     """Compile (and cache) a Python source file.
 405
 406     The file specified by <pathname> is compiled to a code object and
 407     returned.
 408
 409     Presuming the appropriate privileges exist, the bytecodes will be
 410     saved back to the filesystem for future imports. The source file's
 411     modification timestamp must be provided as a Long value.
 412     """
 413     codestring = open(pathname, 'rU').read()
 414     if codestring and codestring[-1] != '\n':
 415         codestring = codestring + '\n'
 416     code = __builtin__.compile(codestring, pathname, 'exec')
 417
 418     # try to cache the compiled code
 419     try:
 420         f = open(pathname + _suffix_char, 'wb')
 421     except IOError:
 422         pass
 423     else:
 424         f.write('\0\0\0\0')
 425         f.write(struct.pack('<I', timestamp))
 426         marshal.dump(code, f)
 427         f.flush()
 428         f.seek(0, 0)
 429         f.write(imp.get_magic())
 430         f.close()
 431
 432     return code
 433
 434 _os_stat = _os_path_join = None
 435 def _os_bootstrap():
 436     "Set up 'os' module replacement functions for use during import bootstrap."
 437
 438     names = sys.builtin_module_names
 439
 440     join = None
 441     if 'posix' in names:
 442         sep = '/'
 443         from posix import stat
 444     elif 'nt' in names:
 445         sep = '\\'
 446         from nt import stat
 447     elif 'dos' in names:
 448         sep = '\\'
 449         from dos import stat
 450     elif 'os2' in names:
 451         sep = '\\'
 452         from os2 import stat
 453     elif 'mac' in names:
 454         from mac import stat
 455         def join(a, b):
 456             if a == '':
 457                 return b
 458             if ':' not in a:
 459                 a = ':' + a
 460             if a[-1:] != ':':
 461                 a = a + ':'
 462             return a + b
 463     else:
 464         raise ImportError, 'no os specific module found'
 465
 466     if join is None:
 467         def join(a, b, sep=sep):
 468             if a == '':
 469                 return b
 470             lastchar = a[-1:]
 471             if lastchar == '/' or lastchar == sep:
 472                 return a + b
 473             return a + sep + b
 474
 475     global _os_stat
 476     _os_stat = stat
 477
 478     global _os_path_join
 479     _os_path_join = join
 480
 481 def _os_path_isdir(pathname):
 482     "Local replacement for os.path.isdir()."
 483     try:
 484         s = _os_stat(pathname)
 485     except OSError:
 486         return None
 487     return (s.st_mode & 0170000) == 0040000
 488
 489 def _timestamp(pathname):
 490     "Return the file modification time as a Long."
 491     try:
 492         s = _os_stat(pathname)
 493     except OSError:
 494         return None
 495     return long(s.st_mtime)
 496
 497
 498 ######################################################################
 499 #
 500 # Emulate the import mechanism for builtin and frozen modules
 501 #
 502 class BuiltinImporter(Importer):
 503     def get_code(self, parent, modname, fqname):
 504         if parent:
 505             # these modules definitely do not occur within a package context
 506             return None
 507
 508         # look for the module
 509         if imp.is_builtin(modname):
 510             type = imp.C_BUILTIN
 511         elif imp.is_frozen(modname):
 512             type = imp.PY_FROZEN
 513         else:
 514             # not found
 515             return None
 516
 517         # got it. now load and return it.
 518         module = imp.load_module(modname, None, modname, ('', '', type))
 519         return 0, module, { }
 520
 521
 522 ######################################################################
 523 #
 524 # Internal importer used for importing from the filesystem
 525 #
 526 class _FilesystemImporter(Importer):
 527     def __init__(self):
 528         self.suffixes = [ ]
 529
 530     def add_suffix(self, suffix, importFunc):
 531         assert callable(importFunc)
 532         self.suffixes.append((suffix, importFunc))
 533
 534     def import_from_dir(self, dir, fqname):
 535         result = self._import_pathname(_os_path_join(dir, fqname), fqname)
 536         if result:
 537             return self._process_result(result, fqname)
 538         return None
 539
 540     def get_code(self, parent, modname, fqname):
 541         # This importer is never used with an empty parent. Its existence is
 542         # private to the ImportManager. The ImportManager uses the
 543         # import_from_dir() method to import top-level modules/packages.
 544         # This method is only used when we look for a module within a package.
 545         assert parent
 546
 547         return self._import_pathname(_os_path_join(parent.__pkgdir__, modname),
 548                                      fqname)
 549
 550     def _import_pathname(self, pathname, fqname):
 551         if _os_path_isdir(pathname):
 552             result = self._import_pathname(_os_path_join(pathname, '__init__'),
 553                                            fqname)
 554             if result:
 555                 values = result[2]
 556                 values['__pkgdir__'] = pathname
 557                 values['__path__'] = [ pathname ]
 558                 return 1, result[1], values
 559             return None
 560
 561         for suffix, importFunc in self.suffixes:
 562             filename = pathname + suffix
 563             try:
 564                 finfo = _os_stat(filename)
 565             except OSError:
 566                 pass
 567             else:
 568                 return importFunc(filename, finfo, fqname)
 569         return None
 570
 571 ######################################################################
 572 #
 573 # SUFFIX-BASED IMPORTERS
 574 #
 575
 576 def py_suffix_importer(filename, finfo, fqname):
 577     file = filename[:-3] + _suffix
 578     t_py = long(finfo[8])
 579     t_pyc = _timestamp(file)
 580
 581     code = None
 582     if t_pyc is not None and t_pyc >= t_py:
 583         f = open(file, 'rb')
 584         if f.read(4) == imp.get_magic():
 585             t = struct.unpack('<I', f.read(4))[0]
 586             if t == t_py:
 587                 code = marshal.load(f)
 588         f.close()
 589     if code is None:
 590         file = filename
 591         code = _compile(file, t_py)
 592
 593     return 0, code, { '__file__' : file }
 594
 595 class DynLoadSuffixImporter:
 596     def __init__(self, desc):
 597         self.desc = desc
 598
 599     def import_file(self, filename, finfo, fqname):
 600         fp = open(filename, self.desc[1])
 601         module = imp.load_module(fqname, fp, filename, self.desc)
 602         module.__file__ = filename
 603         return 0, module, { }
 604
 605
 606 ######################################################################
 607
 608 def _print_importers():
 609     items = sys.modules.items()
 610     items.sort()
 611     for name, module in items:
 612         if module:
 613             print name, module.__dict__.get('__importer__', '-- no importer')
 614         else:
 615             print name, '-- non-existent module'
 616
 617 def _test_revamp():
 618     ImportManager().install()
 619     sys.path.insert(0, BuiltinImporter())
 620
 621 ######################################################################
 622
 623 #
 624 # TODO
 625 #
 626 # from Finn Bock:
 627 #   type(sys) is not a module in JPython. what to use instead?
 628 #   imp.C_EXTENSION is not in JPython. same for get_suffixes and new_module
 629 #
 630 #   given foo.py of:
 631 #      import sys
 632 #      sys.modules['foo'] = sys
 633 #
 634 #   ---- standard import mechanism
 635 #   >>> import foo
 636 #   >>> foo
 637 #   <module 'sys' (built-in)>
 638 #
 639 #   ---- revamped import mechanism
 640 #   >>> import imputil
 641 #   >>> imputil._test_revamp()
 642 #   >>> import foo
 643 #   >>> foo
 644 #   <module 'foo' from 'foo.py'>
 645 #
 646 #
 647 # from MAL:
 648 #   should BuiltinImporter exist in sys.path or hard-wired in ImportManager?
 649 #   need __path__ processing
 650 #   performance
 651 #   move chaining to a subclass [gjs: it's been nuked]
 652 #   deinstall should be possible
 653 #   query mechanism needed: is a specific Importer installed?
 654 #   py/pyc/pyo piping hooks to filter/process these files
 655 #   wish list:
 656 #     distutils importer hooked to list of standard Internet repositories
 657 #     module->file location mapper to speed FS-based imports
 658 #     relative imports
 659 #     keep chaining so that it can play nice with other import hooks
 660 #
 661 # from Gordon:
 662 #   push MAL's mapper into sys.path[0] as a cache (hard-coded for apps)
 663 #
 664 # from Guido:
 665 #   need to change sys.* references for rexec environs
 666 #   need hook for MAL's walk-me-up import strategy, or Tim's absolute strategy
 667 #   watch out for sys.modules[...] is None
 668 #   flag to force absolute imports? (speeds _determine_import_context and
 669 #       checking for a relative module)
 670 #   insert names of archives into sys.path  (see quote below)
 671 #   note: reload does NOT blast module dict
 672 #   shift import mechanisms and policies around; provide for hooks, overrides
 673 #       (see quote below)
 674 #   add get_source stuff
 675 #   get_topcode and get_subcode
 676 #   CRLF handling in _compile
 677 #   race condition in _compile
 678 #   refactoring of os.py to deal with _os_bootstrap problem
 679 #   any special handling to do for importing a module with a SyntaxError?
 680 #       (e.g. clean up the traceback)
 681 #   implement "domain" for path-type functionality using pkg namespace
 682 #       (rather than FS-names like __path__)
 683 #   don't use the word "private"... maybe "internal"
 684 #
 685 #
 686 # Guido's comments on sys.path caching:
 687 #
 688 # We could cache this in a dictionary: the ImportManager can have a
 689 # cache dict mapping pathnames to importer objects, and a separate
 690 # method for coming up with an importer given a pathname that's not yet
 691 # in the cache.  The method should do a stat and/or look at the
 692 # extension to decide which importer class to use; you can register new
 693 # importer classes by registering a suffix or a Boolean function, plus a
 694 # class.  If you register a new importer class, the cache is zapped.
 695 # The cache is independent from sys.path (but maintained per
 696 # ImportManager instance) so that rearrangements of sys.path do the
 697 # right thing.  If a path is dropped from sys.path the corresponding
 698 # cache entry is simply no longer used.
 699 #
 700 # My/Guido's comments on factoring ImportManager and Importer:
 701 #
 702 # > However, we still have a tension occurring here:
 703 # >
 704 # > 1) implementing policy in ImportManager assists in single-point policy
 705 # >    changes for app/rexec situations
 706 # > 2) implementing policy in Importer assists in package-private policy
 707 # >    changes for normal, operating conditions
 708 # >
 709 # > I'll see if I can sort out a way to do this. Maybe the Importer class will
 710 # > implement the methods (which can be overridden to change policy) by
 711 # > delegating to ImportManager.
 712 #
 713 # Maybe also think about what kind of policies an Importer would be
 714 # likely to want to change.  I have a feeling that a lot of the code
 715 # there is actually not so much policy but a *necessity* to get things
 716 # working given the calling conventions for the __import__ hook: whether
 717 # to return the head or tail of a dotted name, or when to do the "finish
 718 # fromlist" stuff.
 719 #