[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Eric Snow
Changes by Eric Snow : -- status: pending -> closed ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mai

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Eric Snow
Eric Snow added the comment: Windows looks happy now. I'll look into the duplicate portions separately in issue19469. -- status: open -> pending ___ Python tracker ___

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Roundup Robot
Roundup Robot added the comment: New changeset 78d36d54391c by Eric Snow in branch 'default': Issue #19413: Disregard duplicate namespace portions during reload tests. http://hg.python.org/cpython/rev/78d36d54391c -- ___ Python tracker

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Eric Snow
Eric Snow added the comment: Looks like this broke on windows: == FAIL: test_reload_namespace_changed (test.test_importlib.test_api.Source_ReloadTests) -- Trac

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Eric Snow
Changes by Eric Snow : -- status: pending -> open ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Eric Snow
Eric Snow added the comment: As you can see, Nick, I came up with a test that did just about the same thing (which you had suggested earlier :-) ). For good measure I also added a test that replaces a namespace package with a normal one. -- assignee: -> eric.snow resolution: -> fixe

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Roundup Robot
Roundup Robot added the comment: New changeset 88c3a1a3c2ff by Eric Snow in branch 'default': Issue #19413: Restore pre-3.3 reload() semantics of re-finding modules. http://hg.python.org/cpython/rev/88c3a1a3c2ff -- nosy: +python-dev ___ Python tracker

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Brett Cannon
Brett Cannon added the comment: No, the fix can go into Python 3.4 right now. -- ___ Python tracker ___ ___ Python-bugs-list mailing l

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Nick Coghlan
Nick Coghlan added the comment: Failing test case showing that Python 3.3 can't reload a module that is converted to a package behind importlib's back (I like this better than the purely introspection based tests, since it shows a real, albeit obscure, regression due to this behavioural change

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-31 Thread Eric Snow
Eric Snow added the comment: I'm fine with not fixing this for 3.3. Does this need to wait on PEP 451 acceptance? -- versions: -Python 3.3 ___ Python tracker ___ _

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-30 Thread Brett Cannon
Brett Cannon added the comment: Fine with fixing it, but in context of PEP 451, not 3.3. -- ___ Python tracker ___ ___ Python-bugs-lis

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-29 Thread Nick Coghlan
Nick Coghlan added the comment: Just had a thought on a possible functional test case: - write a module file - load it - check for expected attributes - move it from name.py to name/__init__.py - reload it - check for new expected attributes -- ___ Py

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-29 Thread Eric Snow
Eric Snow added the comment: Brett: any opinions on fixing this? 3.3? Nick: I'll add another test when I get a chance. -- ___ Python tracker ___ ___

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-29 Thread Nick Coghlan
Nick Coghlan added the comment: Patch mostly looks good to me, but there should be a second test ensuring that the loader attribute gets *replaced*, even if it is already set to something else. A similar test structure to the existing one should work, but replacing the "del types.__loader__"

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-28 Thread Eric Snow
Changes by Eric Snow : Removed file: http://bugs.python.org/file32385/reload-semantics.diff ___ Python tracker ___ ___ Python-bugs-list mailin

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-28 Thread Eric Snow
Eric Snow added the comment: Here's an updated patch. -- Added file: http://bugs.python.org/file32406/reload-semantics.diff ___ Python tracker ___ ___

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-28 Thread Arfrever Frehtes Taifersar Arahesis
Changes by Arfrever Frehtes Taifersar Arahesis : -- nosy: +Arfrever ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscri

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-26 Thread Nick Coghlan
Nick Coghlan added the comment: It's actually even simpler than that - we can just go back to ignoring the __loader__ attribute entirely and always searching for a new one, since we want to pick up changes to the import hooks, even for modules with a __loader__ already set (which is almost all

[issue19413] Reload semantics changed unexpectedly in Python 3.3

2013-10-26 Thread Eric Snow
New submission from Eric Snow: PJE brought up concerns on python-dev regarding PEP 451 and module reloading. [1] However, the issue isn't with the PEP changing reload semantics (mostly). Those actually changed with the switch to importlib (and a pure Python reload function) in the 3.3 releas