| msg110234 - (view) |
Author: Alexander Belopolsky (belopolsky) *  |
Date: 2010-07-13 21:49 |
| On Tue, Jul 13, 2010 at 4:52 PM, Brett Cannon wrote [1]: .. > I say this every time I give an import talk and it has been brought up here > before but obviously not everyone catches it (which is understandable as I > think when it came up on python-dev it was at the tail end of a discussion), > so I am just going to repeat myself: > Do not put junk in fromlist if you call __import__ directly! Use > importlib.import_module! Or if you have a *really* good reason to not use > it, then use ``__import__(name); module = sys.modules[name]``. I think one of the reasons the message does not sink in (at least this is the reason in my case) is that a user who discovers that __import__('foo.bar') returns foo instead of bar, looks up help(__import__) to find that """ When importing a module from a package, note that __import__('A.B', ...) returns package A when fromlist is empty, but its submodule B when fromlist is not empty. """ Passing fromlist=["dummy"] seems like a natural solution after reading this. The ReST documentation [2] is slightly better as it contain a recommendation that says: "If you simply want to import a module (potentially within a package) by name, you can call __import__() and then look it up in sys.modules." However this still fails to mention the (better IMO) alternative of using importlib.import_module(). I believe the __import__ docstring should start with a recommendation not to use it directly and use importlib.import_module() instead, while the ReST documentation should grow a warning not to use dummy fromlist and the recommendation to use __import__() followed by sys.modules lookup should be changed to a recommendation to use mportlib.import_module(). [1] "Peculiar import code in pickle.py" <http://mail.python.org/pipermail/python-dev/2010-July/101906.html>. [2] http://docs.python.org/dev/library/functions.html?#__import__ |
|
|
| msg110241 - (view) |
Author: Éric Araujo (eric.araujo) * |
Date: 2010-07-13 23:24 |
| More rationale for not using a dummy fromlist, from Brett Cannon: “Pulling from sys.modules is the correct way to do this. There are subtle issues when using a bunk fromlist argument (empty modules, double initialization, etc.). If one does not use importlib.import_module -- written *specifically* to prevent people from doing the nasty hack with the fromlist -- then you should use the sys.modules approach, period” |
|
|
| msg140183 - (view) |
Author: Éric Araujo (eric.araujo) * |
Date: 2011-07-12 13:35 |
| The docstring of __import__ was updated to mention importlib in 3d490c3a019e, for #7397. Attached patch edits the docs. |
|
|
| msg140207 - (view) |
Author: Brett Cannon (brett.cannon) * |
Date: 2011-07-12 19:20 |
| Patch looks good to me. |
|
|
| msg141391 - (view) |
Author: Éric Araujo (eric.araujo) * |
Date: 2011-07-29 15:54 |
| Patch committed in 3.2. Attached patch ports the docstring change to 2.7 and edits the reST docs, please approve. |
|
|
| msg141392 - (view) |
Author: Roundup Robot (python-dev)  |
Date: 2011-07-29 16:11 |
| New changeset 7bfc0a45542c by Éric Araujo in branch '3.2': Let the doc of __import__ link to importlib (#9254). http://hg.python.org/cpython/rev/7bfc0a45542c New changeset 4a6cb2d9e906 by Éric Araujo in branch 'default': Merge from 3.2 (#9254, #8982, #9788) http://hg.python.org/cpython/rev/4a6cb2d9e906 |
|
|
| msg165820 - (view) |
Author: Roundup Robot (python-dev) |
Date: 2012-07-18 23:44 |
| New changeset 751f28564a45 by R David Murray in branch '2.7': Closes #9254: backport __import__ docstring/doc mentions of importlib. http://hg.python.org/cpython/rev/751f28564a45 |
|
|