locale information.
Use this collation backend if...
- - You are on a system without ICU or UCA datafiles for the locale,
- and DUCET results are not acceptable.
+ - You are on a system without ICU.
Avoid this backend if...
- - ICU or UCA support is available for the current locale.
+ - ICU is available for the current locale.
- You are sorting strings from alphabets outside the primary locale.
- You need to support collating multiple locales at once.
- You need the same results across multiple platforms.
"""
import locale
+import re
import collate.errors
import collate._abcollator
def __init__(self, locale_code, encoding=None):
locale_code, encoding = collate._locale.getpair(locale_code, encoding)
try:
- locale.setlocale(locale.LC_COLLATE, locale_code)
+ setlocale = locale_code + "." + encoding
+ locale.setlocale(locale.LC_COLLATE, setlocale)
except locale.Error:
- raise collate.errors.InvalidLocaleError(locale_code)
+ raise collate.errors.InvalidLocaleError(setlocale)
self.locale = locale.getlocale(locale.LC_COLLATE)[0]
self.encoding = collate._locale.encoding(encoding)
b = b.decode(self.encoding, "replace")
return locale.strcoll(a, b)
+ def words(self, string, sep=re.compile(r"\W+", re.UNICODE)):
+ """Split the string into separate words.
+
+ This split is done using the locale's notion of a word boundry.
+ """
+ return re.split(sep, string)
+