__future__ environments #1671
__future__ environments #1671
Conversation
|
I think this behavior does match my expectations, thanks! |
|
Can you rebase? One of the big merges touched the run_cell signature as well. |
|
Will do, although it ma |
Allow the %run magic with '-b' to specify a file.
… sharing __future__ imports.
|
I'm not sure why I closed this one - looks like I hit the wrong button somewhere. Reopened and rebased. |
|
The Travis failure appears to be a problem with Travis, not the code. |
|
Test results for commit 0de3a99 merged into master (cdc7e66)
Not available for testing: python2.6 |
|
That failure is some problem with PySide/PyQt, unrelated to this. |
|
This looks good to me. Want to add a test? |
nose.tools no longer has a fail() function
|
I added a test, and it turned up a minor issue, which required a couple more API changes. I guess it shows how important it is to test everything. I also noticed a coincidental failure (caused by a file in the working directory where I run the test) was trying to use a now-nonexistant |
| @@ -90,7 +90,7 @@ def __init__(self): | |||
| # Now, we must monkeypatch the linecache directly so that parts of the | |||
| # stdlib that call it outside our control go through our codepath | |||
| # (otherwise we'd lose our tracebacks). | |||
| linecache.checkcache = self.check_cache | |||
| linecache.checkcache = check_linecache_ipython | |||
There was a problem hiding this comment.
I've made this a standalone function, so that linecache isn't holding a reference to the last CachingCompiler() we created. It didn't need to be a method in the first place.
|
This looks good to me. Anything ore you want to add / confirm? |
|
Nope, that's great. Merging now. |
__future__ environments
So far, this just closes the .ipy hole, so .py and .ipy files have the same behaviour with respect to future, both for %run and as startup files.
There was some discussion on the mailing list about whether
%runshould share the shell's__future__, or only%run -i. Either way, this will need a bit more thought, because it interacts with the other options that can be passed to%run, especially for profiling and debugging.%rerunand macros will share the shell's futures - I think this is expected.