frontend test tweaks

[minrk]

add some encoding to readline history tests that would fail in casting locales (ascii), so that the expected output really matches what should happen.

also relaxes a few timeouts in another test, which were seen to fail while poking around, to avoid spurious failures.

closes #1589

[minrk]

Test results for commit 93f5f3d merged into master
Platform: darwin

• python2.6: OK (libraries not available: cython matplotlib oct2py pygments pymongo qt rpy2 tornado wx wx.aui)
• python2.7: OK (libraries not available: oct2py wx wx.aui)
• python3.2: OK (libraries not available: cython matplotlib oct2py pymongo qt rpy2 wx wx.aui)

Not available for testing: python3.1

[minrk]

The above test_pr was run in an ASCII locale on OS X, which revealed another py3-specific test failure due to ascii default encoding.

[cdeil]

Test results for commit 8c11e5c merged into master
Platform: darwin

• python2.7: Failed, log at https://gist.github.com/2965469 (libraries not available: oct2py pymongo wx wx.aui)
• python3.2: Failed, log at https://gist.github.com/2965471 (libraries not available: matplotlib numpy oct2py pygments pymongo qt rpy2 tornado wx wx.aui zmq)

Not available for testing: python2.6, python3.1

[bfroehle]

Speaking of frontend tests, I've been seeing this one recently. It's not causing anything to fail, however...

**********************************************************************
IPython test group: IPython.frontend
..............................
----------------------------------------------------------------------
Ran 30 tests in 2.755s

OK
Traceback (most recent call last):
  File "/home/bfroehle/src/ipython/IPython/frontend/qt/console/completion_html.py", line 152, in eventFilter
    if obj == self._text_edit:
AttributeError: 'CompletionHtml' object has no attribute '_text_edit'

[minrk]

interesting - I've never seen that error, and from the code, it makes no sense at all for it to happen - the event filter is registered after the attribute is assigned. It could be a garbage collection issue on exit, where the event filter is still registered, but the object has started to be deleted.

[bfroehle]

The garbage collection hypothesis seems pretty reasonable, given that the
error prints after the test group has finished.

[minrk]

I wrote, but apparently failed to post, a note about @cdeil's test_pr run: I believe all failures listed are due to libedit, and should not be considered real failures (libedit is fully expected to break all kinds of stuff). I can confirm that an ASCII locale no longer fails (I could reproduce exactly the failures reported in #1589, and they are fixed).

Getting virtualenv to install the right readline is a bit of a pain, and I don't quite remember how to do it.