On the previous commit we changed text and subject from unicode
to str. After a better solution from @cassiobotaro, we're using
unicode again, selecting the encoding as the one passed via encoding
parameter.
On PR #964 @matclab forced the encoding of both subject and
text variables to unicode.
After merging it, matclab realized that when we send Unicode
text to the method it raises an exception and asked if he should
change the commit. Unfortunately this thing was kept untouched.
This problem exists because we previously encode the unicode variables
to utf-8 (for instance here https://github.com/web2py/web2py/blob/master/gluon/tools.py#L478-L481) and then force again to unicode. This piece of code shows what happens:
```
>>> a = u'áéí'
>>> a
u'\xe1\xe9\xed'
>>> b = a.encode('utf-8')
>>> b
'\xc3\xa1\xc3\xa9\xc3\xad'
>>> unicode(a)
u'\xe1\xe9\xed'
>>> unicode(b)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 0: ordinal not in range(128)
```
If we force to str, just like @matclab suggested, we solve this issue.
rafaelol