timeit on quick snippets causes UnicodeEncodeError

[Stephan Peijnik]

This bug report came in via bugs.debian.org. It seems to be present in ipython 0.9.1.

Quoting the reporter's description of the bug:

Here is the snippet (although I guess it needs to be present on any quick but
not uber-quick function ;)) and output:

*In [2]:import scipy.stats
In [3]:import numpy as N
In [4]:cpaste
Pasting code; enter '--' alone on the line to stop.
:def cr():
: try:
: scipy.stats.rdist(1.32, 0, 1).cdf(-1.0+N.finfo(float).eps)
: except ZeroDivisionError:
: pass
:--

In [5]:timeit cr()
---------------------------------------------------------------------------
UnicodeEncodeError Traceback (most recent call last)

/home/yoh/proj/pymvpa/pymvpa/mvpa/tests/<ipython console> in <module>()

/var/lib/python-support/python2.5/IPython/iplib.py in ipmagic(self, arg_s)
    951 else:
    952 magic_args = self.var_expand(magic_args,1)
--> 953 return fn(magic_args)
    954
    955 def ipalias(self,arg_s):

/var/lib/python-support/python2.5/IPython/Magic.py in magic_timeit(self, parameter_s)
   1829 precision,
   1830 best * scaling[order],
-> 1831 units[order])
   1832 if tc > tc_min:
   1833 print "Compiler time: %.2f s" % tc

UnicodeEncodeError: 'ascii' codec can't encode character u'\xb5' in position 27: ordinal not in range(128)

[Henry Prêcheur (henry-precheur)]

This is because the terminal doesn't support UTF-8. I have the same problem on OpenBSD with the stock xterm:

In [1]: print u'\xb5'
---------------------------------------------------------------------------
UnicodeEncodeError Traceback (most recent call last)

/home/henry/<ipython console> in <module>()

UnicodeEncodeError: 'ascii' codec can't encode character u'\xb5' in position 0: ordinal not in range(128)

I would advice to replace the u'\xb5s' with u'micros' since it can be displayed on every terminals.

[Fernando Perez (fdo.perez)]

http://bazaar.launchpad.net/~fdo.perez/ipython/trunk-dev/revision/1186

Contains a fix, though it's not ideal. I'd love to know how to do it so we don't lose the 'micro' character where it works. Input from unicode experts welcome.

[Brian Granger (ellisonbg)]

Should we call this "Fix Committed" or do you want to pursue further ways of resolving this?

[Fernando Perez (fdo.perez)]

Marking "Fix Committed", but I'd like to leave it without "Fix Released" in case we find a better solution.

Basically 'almost closed', but I'd like to have a better fix. I suspect once we start doing 3.0 work we'll have to revisit this and similar issues more closely.