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" or do "from os.path import ..."
 135         if len(parts) == 2 and hasattr(top_module, parts[1]):
 136             if fromlist:
 137                 return getattr(top_module, parts[1])
 138             else:
 139                 return top_module
 140
 141         # If the importer does not exist, then we have to bail. A missing
 142         # importer means that something else imported the module, and we have
 143         # no knowledge of how to get sub-modules out of the thing.
 144         raise ImportError, 'No module named ' + fqname
 145
 146     def _determine_import_context(self, globals):
 147         """Returns the context in which a module should be imported.
 148
 149         The context could be a loaded (package) module and the imported module
 150         will be looked for within that package. The context could also be None,
 151         meaning there is no context -- the module should be looked for as a
 152         "top-level" module.
 153         """
 154
 155         if not globals or not globals.get('__importer__'):
 156             # globals does not refer to one of our modules or packages. That
 157             # implies there is no relative import context (as far as we are
 158             # concerned), and it should just pick it off the standard path.
 159             return None
 160
 161         # The globals refer to a module or package of ours. It will define
 162         # the context of the new import. Get the module/package fqname.
 163         parent_fqname = globals['__name__']
 164
 165         # if a package is performing the import, then return itself (imports
 166         # refer to pkg contents)
 167         if globals['__ispkg__']:
 168             parent = sys.modules[parent_fqname]
 169             assert globals is parent.__dict__
 170             return parent
 171
 172         i = parent_fqname.rfind('.')
 173
 174         # a module outside of a package has no particular import context
 175         if i == -1:
 176             return None
 177
 178         # if a module in a package is performing the import, then return the
 179         # package (imports refer to siblings)
 180         parent_fqname = parent_fqname[:i]
 181         parent = sys.modules[parent_fqname]
 182         assert parent.__name__ == parent_fqname
 183         return parent
 184
 185     def _import_top_module(self, name):
 186         # scan sys.path looking for a location in the filesystem that contains
 187         # the module, or an Importer object that can import the module.
 188         for item in sys.path:
 189             if isinstance(item, _StringType):
 190                 module = self.fs_imp.import_from_dir(item, name)
 191             else:
 192                 module = item.import_top(name)
 193             if module:
 194                 return module
 195         return None
 196
 197     def _reload_hook(self, module):
 198         "Python calls this hook to reload a module."
 199
 200         # reloading of a module may or may not be possible (depending on the
 201         # importer), but at least we can validate that it's ours to reload
 202         importer = module.__dict__.get('__importer__')
 203         if not importer:
 204             ### oops. now what...
 205             pass
 206
 207         # okay. it is using the imputil system, and we must delegate it, but
 208         # we don't know what to do (yet)
 209         ### we should blast the module dict and do another get_code(). need to
 210         ### flesh this out and add proper docco...
 211         raise SystemError, "reload not yet implemented"
 212
 213
 214 class Importer:
 215     "Base class for replacing standard import functions."
 216
 217     def import_top(self, name):
 218         "Import a top-level module."
 219         return self._import_one(None, name, name)
 220
 221     ######################################################################
 222     #
 223     # PRIVATE METHODS
 224     #
 225     def _finish_import(self, top, parts, fromlist):
 226         # if "a.b.c" was provided, then load the ".b.c" portion down from
 227         # below the top-level module.
 228         bottom = self._load_tail(top, parts)
 229
 230         # if the form is "import a.b.c", then return "a"
 231         if not fromlist:
 232             # no fromlist: return the top of the import tree
 233             return top
 234
 235         # the top module was imported by self.
 236         #
 237         # this means that the bottom module was also imported by self (just
 238         # now, or in the past and we fetched it from sys.modules).
 239         #
 240         # since we imported/handled the bottom module, this means that we can
 241         # also handle its fromlist (and reliably use __ispkg__).
 242
 243         # if the bottom node is a package, then (potentially) import some
 244         # modules.
 245         #
 246         # note: if it is not a package, then "fromlist" refers to names in
 247         #       the bottom module rather than modules.
 248         # note: for a mix of names and modules in the fromlist, we will
 249         #       import all modules and insert those into the namespace of
 250         #       the package module. Python will pick up all fromlist names
 251         #       from the bottom (package) module; some will be modules that
 252         #       we imported and stored in the namespace, others are expected
 253         #       to be present already.
 254         if bottom.__ispkg__:
 255             self._import_fromlist(bottom, fromlist)
 256
 257         # if the form is "from a.b import c, d" then return "b"
 258         return bottom
 259
 260     def _import_one(self, parent, modname, fqname):
 261         "Import a single module."
 262
 263         # has the module already been imported?
 264         try:
 265             return sys.modules[fqname]
 266         except KeyError:
 267             pass
 268
 269         # load the module's code, or fetch the module itself
 270         result = self.get_code(parent, modname, fqname)
 271         if result is None:
 272             return None
 273
 274         module = self._process_result(result, fqname)
 275
 276         # insert the module into its parent
 277         if parent:
 278             setattr(parent, modname, module)
 279         return module
 280
 281     def _process_result(self, (ispkg, code, values), fqname):
 282         # did get_code() return an actual module? (rather than a code object)
 283         is_module = isinstance(code, _ModuleType)
 284
 285         # use the returned module, or create a new one to exec code into
 286         if is_module:
 287             module = code
 288         else:
 289             module = imp.new_module(fqname)
 290
 291         ### record packages a bit differently??
 292         module.__importer__ = self
 293         module.__ispkg__ = ispkg
 294
 295         # insert additional values into the module (before executing the code)
 296         module.__dict__.update(values)
 297
 298         # the module is almost ready... make it visible
 299         sys.modules[fqname] = module
 300
 301         # execute the code within the module's namespace
 302         if not is_module:
 303             try:
 304                 exec code in module.__dict__
 305             except:
 306                 if fqname in sys.modules:
 307                     del sys.modules[fqname]
 308                 raise
 309
 310         # fetch from sys.modules instead of returning module directly.
 311         # also make module's __name__ agree with fqname, in case
 312         # the "exec code in module.__dict__" played games on us.
 313         module = sys.modules[fqname]
 314         module.__name__ = fqname
 315         return module
 316
 317     def _load_tail(self, m, parts):
 318         """Import the rest of the modules, down from the top-level module.
 319
 320         Returns the last module in the dotted list of modules.
 321         """
 322         for part in parts:
 323             fqname = "%s.%s" % (m.__name__, part)
 324             m = self._import_one(m, part, fqname)
 325             if not m:
 326                 raise ImportError, "No module named " + fqname
 327         return m
 328
 329     def _import_fromlist(self, package, fromlist):
 330         'Import any sub-modules in the "from" list.'
 331
 332         # if '*' is present in the fromlist, then look for the '__all__'
 333         # variable to find additional items (modules) to import.
 334         if '*' in fromlist:
 335             fromlist = list(fromlist) + \
 336                        list(package.__dict__.get('__all__', []))
 337
 338         for sub in fromlist:
 339             # if the name is already present, then don't try to import it (it
 340             # might not be a module!).
 341             if sub != '*' and not hasattr(package, sub):
 342                 subname = "%s.%s" % (package.__name__, sub)
 343                 submod = self._import_one(package, sub, subname)
 344                 if not submod:
 345                     raise ImportError, "cannot import name " + subname
 346
 347     def _do_import(self, parent, parts, fromlist):
 348         """Attempt to import the module relative to parent.
 349
 350         This method is used when the import context specifies that <self>
 351         imported the parent module.
 352         """
 353         top_name = parts[0]
 354         top_fqname = parent.__name__ + '.' + top_name
 355         top_module = self._import_one(parent, top_name, top_fqname)
 356         if not top_module:
 357             # this importer and parent could not find the module (relatively)
 358             return None
 359
 360         return self._finish_import(top_module, parts[1:], fromlist)
 361
 362     ######################################################################
 363     #
 364     # METHODS TO OVERRIDE
 365     #
 366     def get_code(self, parent, modname, fqname):
 367         """Find and retrieve the code for the given module.
 368
 369         parent specifies a parent module to define a context for importing. It
 370         may be None, indicating no particular context for the search.
 371
 372         modname specifies a single module (not dotted) within the parent.
 373
 374         fqname specifies the fully-qualified module name. This is a
 375         (potentially) dotted name from the "root" of the module namespace
 376         down to the modname.
 377         If there is no parent, then modname==fqname.
 378
 379         This method should return None, or a 3-tuple.
 380
 381         * If the module was not found, then None should be returned.
 382
 383         * The first item of the 2- or 3-tuple should be the integer 0 or 1,
 384             specifying whether the module that was found is a package or not.
 385
 386         * The second item is the code object for the module (it will be
 387             executed within the new module's namespace). This item can also
 388             be a fully-loaded module object (e.g. loaded from a shared lib).
 389
 390         * The third item is a dictionary of name/value pairs that will be
 391             inserted into new module before the code object is executed. This
 392             is provided in case the module's code expects certain values (such
 393             as where the module was found). When the second item is a module
 394             object, then these names/values will be inserted *after* the module
 395             has been loaded/initialized.
 396         """
 397         raise RuntimeError, "get_code not implemented"
 398
 399
 400 ######################################################################
 401 #
 402 # Some handy stuff for the Importers
 403 #
 404
 405 # byte-compiled file suffix character
 406 _suffix_char = __debug__ and 'c' or 'o'
 407
 408 # byte-compiled file suffix
 409 _suffix = '.py' + _suffix_char
 410
 411 def _compile(pathname, timestamp):
 412     """Compile (and cache) a Python source file.
 413
 414     The file specified by <pathname> is compiled to a code object and
 415     returned.
 416
 417     Presuming the appropriate privileges exist, the bytecodes will be
 418     saved back to the filesystem for future imports. The source file's
 419     modification timestamp must be provided as a Long value.
 420     """
 421     codestring = open(pathname, 'rU').read()
 422     if codestring and codestring[-1] != '\n':
 423         codestring = codestring + '\n'
 424     code = __builtin__.compile(codestring, pathname, 'exec')
 425
 426     # try to cache the compiled code
 427     try:
 428         f = open(pathname + _suffix_char, 'wb')
 429     except IOError:
 430         pass
 431     else:
 432         f.write('\0\0\0\0')
 433         f.write(struct.pack('<I', timestamp))
 434         marshal.dump(code, f)
 435         f.flush()
 436         f.seek(0, 0)
 437         f.write(imp.get_magic())
 438         f.close()
 439
 440     return code
 441
 442 _os_stat = _os_path_join = None
 443 def _os_bootstrap():
 444     "Set up 'os' module replacement functions for use during import bootstrap."
 445
 446     names = sys.builtin_module_names
 447
 448     join = None
 449     if 'posix' in names:
 450         sep = '/'
 451         from posix import stat
 452     elif 'nt' in names:
 453         sep = '\\'
 454         from nt import stat
 455     elif 'dos' in names:
 456         sep = '\\'
 457         from dos import stat
 458     elif 'os2' in names:
 459         sep = '\\'
 460         from os2 import stat
 461     elif 'mac' in names:
 462         from mac import stat
 463         def join(a, b):
 464             if a == '':
 465                 return b
 466             if ':' not in a:
 467                 a = ':' + a
 468             if a[-1:] != ':':
 469                 a = a + ':'
 470             return a + b
 471     else:
 472         raise ImportError, 'no os specific module found'
 473
 474     if join is None:
 475         def join(a, b, sep=sep):
 476             if a == '':
 477                 return b
 478             lastchar = a[-1:]
 479             if lastchar == '/' or lastchar == sep:
 480                 return a + b
 481             return a + sep + b
 482
 483     global _os_stat
 484     _os_stat = stat
 485
 486     global _os_path_join
 487     _os_path_join = join
 488
 489 def _os_path_isdir(pathname):
 490     "Local replacement for os.path.isdir()."
 491     try:
 492         s = _os_stat(pathname)
 493     except OSError:
 494         return None
 495     return (s.st_mode & 0170000) == 0040000
 496
 497 def _timestamp(pathname):
 498     "Return the file modification time as a Long."
 499     try:
 500         s = _os_stat(pathname)
 501     except OSError:
 502         return None
 503     return long(s.st_mtime)
 504
 505
 506 ######################################################################
 507 #
 508 # Emulate the import mechanism for builtin and frozen modules
 509 #
 510 class BuiltinImporter(Importer):
 511     def get_code(self, parent, modname, fqname):
 512         if parent:
 513             # these modules definitely do not occur within a package context
 514             return None
 515
 516         # look for the module
 517         if imp.is_builtin(modname):
 518             type = imp.C_BUILTIN
 519         elif imp.is_frozen(modname):
 520             type = imp.PY_FROZEN
 521         else:
 522             # not found
 523             return None
 524
 525         # got it. now load and return it.
 526         module = imp.load_module(modname, None, modname, ('', '', type))
 527         return 0, module, { }
 528
 529
 530 ######################################################################
 531 #
 532 # Internal importer used for importing from the filesystem
 533 #
 534 class _FilesystemImporter(Importer):
 535     def __init__(self):
 536         self.suffixes = [ ]
 537
 538     def add_suffix(self, suffix, importFunc):
 539         assert callable(importFunc)
 540         self.suffixes.append((suffix, importFunc))
 541
 542     def import_from_dir(self, dir, fqname):
 543         result = self._import_pathname(_os_path_join(dir, fqname), fqname)
 544         if result:
 545             return self._process_result(result, fqname)
 546         return None
 547
 548     def get_code(self, parent, modname, fqname):
 549         # This importer is never used with an empty parent. Its existence is
 550         # private to the ImportManager. The ImportManager uses the
 551         # import_from_dir() method to import top-level modules/packages.
 552         # This method is only used when we look for a module within a package.
 553         assert parent
 554
 555         return self._import_pathname(_os_path_join(parent.__pkgdir__, modname),
 556                                      fqname)
 557
 558     def _import_pathname(self, pathname, fqname):
 559         if _os_path_isdir(pathname):
 560             result = self._import_pathname(_os_path_join(pathname, '__init__'),
 561                                            fqname)
 562             if result:
 563                 values = result[2]
 564                 values['__pkgdir__'] = pathname
 565                 values['__path__'] = [ pathname ]
 566                 return 1, result[1], values
 567             return None
 568
 569         for suffix, importFunc in self.suffixes:
 570             filename = pathname + suffix
 571             try:
 572                 finfo = _os_stat(filename)
 573             except OSError:
 574                 pass
 575             else:
 576                 return importFunc(filename, finfo, fqname)
 577         return None
 578
 579 ######################################################################
 580 #
 581 # SUFFIX-BASED IMPORTERS
 582 #
 583
 584 def py_suffix_importer(filename, finfo, fqname):
 585     file = filename[:-3] + _suffix
 586     t_py = long(finfo[8])
 587     t_pyc = _timestamp(file)
 588
 589     code = None
 590     if t_pyc is not None and t_pyc >= t_py:
 591         f = open(file, 'rb')
 592         if f.read(4) == imp.get_magic():
 593             t = struct.unpack('<I', f.read(4))[0]
 594             if t == t_py:
 595                 code = marshal.load(f)
 596         f.close()
 597     if code is None:
 598         file = filename
 599         code = _compile(file, t_py)
 600
 601     return 0, code, { '__file__' : file }
 602
 603 class DynLoadSuffixImporter:
 604     def __init__(self, desc):
 605         self.desc = desc
 606
 607     def import_file(self, filename, finfo, fqname):
 608         fp = open(filename, self.desc[1])
 609         module = imp.load_module(fqname, fp, filename, self.desc)
 610         module.__file__ = filename
 611         return 0, module, { }
 612
 613
 614 ######################################################################
 615
 616 def _print_importers():
 617     items = sys.modules.items()
 618     items.sort()
 619     for name, module in items:
 620         if module:
 621             print name, module.__dict__.get('__importer__', '-- no importer')
 622         else:
 623             print name, '-- non-existent module'
 624
 625 def _test_revamp():
 626     ImportManager().install()
 627     sys.path.insert(0, BuiltinImporter())
 628
 629 ######################################################################
 630
 631 #
 632 # TODO
 633 #
 634 # from Finn Bock:
 635 #   type(sys) is not a module in JPython. what to use instead?
 636 #   imp.C_EXTENSION is not in JPython. same for get_suffixes and new_module
 637 #
 638 #   given foo.py of:
 639 #      import sys
 640 #      sys.modules['foo'] = sys
 641 #
 642 #   ---- standard import mechanism
 643 #   >>> import foo
 644 #   >>> foo
 645 #   <module 'sys' (built-in)>
 646 #
 647 #   ---- revamped import mechanism
 648 #   >>> import imputil
 649 #   >>> imputil._test_revamp()
 650 #   >>> import foo
 651 #   >>> foo
 652 #   <module 'foo' from 'foo.py'>
 653 #
 654 #
 655 # from MAL:
 656 #   should BuiltinImporter exist in sys.path or hard-wired in ImportManager?
 657 #   need __path__ processing
 658 #   performance
 659 #   move chaining to a subclass [gjs: it's been nuked]
 660 #   deinstall should be possible
 661 #   query mechanism needed: is a specific Importer installed?
 662 #   py/pyc/pyo piping hooks to filter/process these files
 663 #   wish list:
 664 #     distutils importer hooked to list of standard Internet repositories
 665 #     module->file location mapper to speed FS-based imports
 666 #     relative imports
 667 #     keep chaining so that it can play nice with other import hooks
 668 #
 669 # from Gordon:
 670 #   push MAL's mapper into sys.path[0] as a cache (hard-coded for apps)
 671 #
 672 # from Guido:
 673 #   need to change sys.* references for rexec environs
 674 #   need hook for MAL's walk-me-up import strategy, or Tim's absolute strategy
 675 #   watch out for sys.modules[...] is None
 676 #   flag to force absolute imports? (speeds _determine_import_context and
 677 #       checking for a relative module)
 678 #   insert names of archives into sys.path  (see quote below)
 679 #   note: reload does NOT blast module dict
 680 #   shift import mechanisms and policies around; provide for hooks, overrides
 681 #       (see quote below)
 682 #   add get_source stuff
 683 #   get_topcode and get_subcode
 684 #   CRLF handling in _compile
 685 #   race condition in _compile
 686 #   refactoring of os.py to deal with _os_bootstrap problem
 687 #   any special handling to do for importing a module with a SyntaxError?
 688 #       (e.g. clean up the traceback)
 689 #   implement "domain" for path-type functionality using pkg namespace
 690 #       (rather than FS-names like __path__)
 691 #   don't use the word "private"... maybe "internal"
 692 #
 693 #
 694 # Guido's comments on sys.path caching:
 695 #
 696 # We could cache this in a dictionary: the ImportManager can have a
 697 # cache dict mapping pathnames to importer objects, and a separate
 698 # method for coming up with an importer given a pathname that's not yet
 699 # in the cache.  The method should do a stat and/or look at the
 700 # extension to decide which importer class to use; you can register new
 701 # importer classes by registering a suffix or a Boolean function, plus a
 702 # class.  If you register a new importer class, the cache is zapped.
 703 # The cache is independent from sys.path (but maintained per
 704 # ImportManager instance) so that rearrangements of sys.path do the
 705 # right thing.  If a path is dropped from sys.path the corresponding
 706 # cache entry is simply no longer used.
 707 #
 708 # My/Guido's comments on factoring ImportManager and Importer:
 709 #
 710 # > However, we still have a tension occurring here:
 711 # >
 712 # > 1) implementing policy in ImportManager assists in single-point policy
 713 # >    changes for app/rexec situations
 714 # > 2) implementing policy in Importer assists in package-private policy
 715 # >    changes for normal, operating conditions
 716 # >
 717 # > I'll see if I can sort out a way to do this. Maybe the Importer class will
 718 # > implement the methods (which can be overridden to change policy) by
 719 # > delegating to ImportManager.
 720 #
 721 # Maybe also think about what kind of policies an Importer would be
 722 # likely to want to change.  I have a feeling that a lot of the code
 723 # there is actually not so much policy but a *necessity* to get things
 724 # working given the calling conventions for the __import__ hook: whether
 725 # to return the head or tail of a dotted name, or when to do the "finish
 726 # fromlist" stuff.
 727 #