BUG: revert 2/3 changes in vbsuite

[jtratner]

Fixes #4477, #4474

[ghost]

Can you take care of #4474 as well please?

[jtratner]

done. I just checked out vb_suite from 338f340 - I don't think there were any other changes aside from 2to3 related ones, so probably should be okay. You might want to confirm that (tried to look through the diff). I have to run but can look again tonight.

[jtratner]

@jreback this just popped up on Travis - any idea? Given that this only changes vb_suite, I'm pretty sure it's unrelated.

======================================================================
ERROR: test_string_select (pandas.io.tests.test_pytables.TestHDFStore)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/travis/virtualenv/python2.7_with_system_site_packages/local/lib/python2.7/site-packages/pandas-0.11.0_1052_g7ad2e48-py2.7-linux-x86_64.egg/pandas/io/tests/test_pytables.py", line 2424, in test_string_select
    result = store.select('df',Term('x!=none'))
  File "/home/travis/virtualenv/python2.7_with_system_site_packages/local/lib/python2.7/site-packages/pandas-0.11.0_1052_g7ad2e48-py2.7-linux-x86_64.egg/pandas/io/pytables.py", line 475, in select
    raise KeyError('No object named %s in the file' % key)
KeyError: 'No object named df in the file'
----------------------------------------------------------------------

[jreback]

sometimes I see a random failure on travis (but can't reproduce it outside)....maybe a writing to disk error of some kind...if it reproduces let me know

[jtratner]

@jreback just happened twice in a row.

[jtratner]

@jreback trying it again, maybe it'll work this time...

[jtratner]

Actually every commit I've pushed based off of master:

• https://travis-ci.org/jtratner/pandas/builds/9896736
• https://travis-ci.org/jtratner/pandas/builds/9896602
• https://travis-ci.org/jtratner/pandas/builds/9896231
• https://travis-ci.org/jtratner/pandas/builds/9896231

etc.

[jreback]

hmm...that is weird...let me pull down your branch

[jreback]

works fine for me (both 32/64 bit)

[jtratner]

This commit is just current master failing with the same issue - https://travis-ci.org/jtratner/pandas/builds/9897047

[jreback]

are you sure you are completely rebased, I did push a fix last night to master (and in between I don't know what it would do)

[jreback]

current master is fine

[jtratner]

@jreback can you checkout a branch off of master and force Travis to build? I keep getting issues with commits that are exactly the same as master. This time I deleted my master branch, checked out pydata/pandas@master and then push to my own branch - still failed.

[jreback]

@jtratner fyi...@cpcloud ran into the same problem....but noone can repro (even on a very travis like machine), then suddenly went away....weird

[jreback]

tell you what...why don't I just merge this branch....?

[jtratner]

@jreback haha, your call. But just be aware that this just failed about 2 minutes ago b/c of the same issue cropping up again :-/

[jreback]

let's see how master does....problem is I CAN't repoduce, nor could @cpcloud locally at all (even on a box exactl like travis), 32-bit, same exact versions of tables,numexpr,numpy,python....really odd

[jreback]

ok....failed (same test)...let me see

[jreback]

ok...its fixed, but I all I did was this (which I am now going to take out)....maybe its some weird whitespace issue????
f08fb89

[jreback]

@jtratner master looks good now....very weird error