Add documentation for `register_strwidth_error` and `string` functions

This commit is contained in:
ZyX 2014-12-06 15:02:07 +03:00
parent f697d9ef67
commit b80da891a1
1 changed files with 35 additions and 0 deletions

View File

@ -58,6 +58,27 @@ last_swe_idx = 0
def register_strwidth_error(strwidth):
'''Create new encode errors handling method similar to ``replace``
Like ``replace`` this method uses question marks in place of the characters
that cannot be represented in the requested encoding. Unlike ``replace`` the
amount of question marks is identical to the amount of display cells
offending character occupies. Thus encoding ```` (U+2026, HORIZONTAL
ELLIPSIS) to ``latin1`` will emit one question mark, but encoding ````
(U+FF21, FULLWIDTH LATIN CAPITAL LETTER A) will emit two question marks.
Since width of some characters depends on the terminal settings and
powerline knows how to respect them a single error handling method cannot be
used. Instead of it the generator function is used which takes ``strwidth``
function (function that knows how to compute string width respecting all
needed settings) and emits new error handling method name.
:param function strwidth:
Function that computs string width measured in display cells the string
occupies when displayed.
:return: New error handling method name.
'''
global last_swe_idx
last_swe_idx += 1
@ -141,6 +162,20 @@ else:
return s
string.__doc__ = (
'''Transform ``unicode`` or ``bytes`` object into ``str`` object
On Python-2 this encodes ``unicode`` to ``bytes`` (which is ``str``) using
UTF-8 encoding; on Python-3 this decodes ``bytes`` to ``unicode`` (which is
``str``) using UTF-8 encoding.
Useful for functions that expect an ``str`` object in both unicode versions,
not caring about the semantic differences between them in Python-2 and
Python-3.
'''
)
def surrogate_pair_to_character(high, low):
'''Transform a pair of surrogate codepoints to one codepoint
'''