Add option append (-a) to %save

[doda-zz]

inspired by: http://www.reddit.com/r/Python/comments/wc0wp/ipython_how_to_save_code_entered_in_the_shell_to/

:)

I'm not sure if simply writing a \n if it isn't there is the best way to handle the case of multiple save -a after one another, but i'd think it's than to try and change find_user_code.

[bfroehle]

Is there a reason why we shouldn't end on a newline always anyway?

[doda-zz]

i don't think there's ever a reason not to end a file with a newline, doda/ipython@4a6d842 would implement that and simplify the code as a result

[takluyver]

I think that makes sense - let's bring that commit into this PR.

[takluyver]

This only writes the #coding: utf-8 header if -a isn't used. That makes sense, but is it worth also writing it if -a is specified and the file doesn't already exist?

[doda-zz]

i've added 2 commits, the magic comment now always gets written if the file didn't exist beforehand, no matter -a or not and files are always ended with newlines, which makes sense given it's in the C standard

[takluyver]

I think this description could be a bit clearer - how about "append to the file instead of overwriting it."

[takluyver]

Other than the docstring (and maybe a test), I think this looks fine.

[doda-zz]

grepping through the source didn't lead me to any tests for %save so i've written a test that covers both %save as well as %save -a. if there's anything else let me know! :)

[takluyver]

Great, thanks. I'm running the tests now.

[takluyver]

For some reason it's refusing to post the results automatically, but it's somehow causing a failure in the history tests:

======================================================================
FAIL: IPython.core.tests.test_history.test_history
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/nose/case.py", line 198, in runTest
    self.test(*self.arg)
  File "/home/thomas/.ipy_pr_tests/venv-python3.2/lib/python3.2/site-packages/ipython-0.14.dev-py3.2.egg/IPython/core/tests/test_history.py", line 98, in test_history
    "# coding: utf-8\n" + "\n".join(hist))
AssertionError: "# coding: utf-8\na=1\ndef f():\n    test = 1\n    return test\nb='\u20ac\xc6\xbe\xf7\xdf'\n" != "# coding: utf-8\na=1\ndef f():\n    test = 1\n    return test\nb='\u20ac\xc6\xbe\xf7\xdf'"
  # coding: utf-8
  a=1
  def f():
      test = 1
      return test
- b='\u20ac\xc6\xbe\xf7\xdf'
?          -
+ b='\u20ac\xc6\xbe\xf7\xdf'
    """Fail immediately, with the given message."""
>>  raise self.failureException('"# coding: utf-8\\na=1\\ndef f():\\n    test = 1\\n    return test\\nb=\'\u20ac\xc6\xbe\xf7\xdf\'\\n" != "# coding: utf-8\\na=1\\ndef f():\\n    test = 1\\n    return test\\nb=\'\u20ac\xc6\xbe\xf7\xdf\'"\n  # coding: utf-8\n  a=1\n  def f():\n      test = 1\n      return test\n- b=\'\u20ac\xc6\xbe\xf7\xdf\'\n?          -\n+ b=\'\u20ac\xc6\xbe\xf7\xdf\'')

-------------------- >> begin captured stdout << ---------------------
def f():
    test = 1
    return test
b='\u20ac\xc6\xbe\xf7\xdf'
The following commands were written to file `/tmp/tmpb3fnwk/test4.py`:
a=1
def f():
    test = 1
    return test
b='\u20ac\xc6\xbe\xf7\xdf'
The following commands were written to file `/tmp/tmpb3fnwk/test.py`:
a=1
def f():
    test = 1
    return test
b='\u20ac\xc6\xbe\xf7\xdf'

--------------------- >> end captured stdout << ----------------------

That's in Python 3.2, but it looks like it's failing in all versions of Python.

[takluyver]

Oh, I think it's just because %save is now adding a newline at the end of the file, which the test doesn't expect. Could you just change the test?

[doda-zz]

here you go!

[takluyver]

Test results for commit d5e7d79 merged into master
Platform: linux2

• python2.7: OK (libraries not available: oct2py rpy2)
• python3.1: OK (libraries not available: cython matplotlib numpy oct2py pymongo qt rpy2 wx wx.aui zmq)
• python3.2: Failed, log at https://gist.github.com/3113637 (libraries not available: cython oct2py pymongo rpy2 wx wx.aui)

Not available for testing: python2.6

[takluyver]

The timeout error is probably unrelated. This all looks OK now, but I'll give it a day or so in case anyone else wants to look over it.

[doda-zz]

sure thanks for the quick replies!

[takluyver]

Merged. Thanks for the contribution!