Arthur de Jong

Open Source / Free Software developer

summaryrefslogtreecommitdiffstats
path: root/tests/test_util.doctest
Commit message (Collapse)AuthorAgeFilesLines
* Convert various reasonable unicode digitsArthur de Jong2019-04-291-1/+1
| | | | | This converts many of the "reasonable" unicode digits that are just variations on ASCII 0-9 to their ASCII counterparts.
* Use an internal isdigits() function instead of str.isdigit()Arthur de Jong2019-04-291-2/+27
| | | | | | | | | | | The problem with the latter is that it will also accept all kinds of unicode digits that are not the ASCII 0-9 digits causing all kinds of problems in check digit calculations. Some of these unicode characters are also considered digits by int() but some are not (such as the SUPERSCRIPT TWO unicode character). Closes https://github.com/arthurdejong/python-stdnum/issues/96
* Make unicode conversion standardArthur de Jong2018-05-011-1/+15
| | | | | | A few modules use non-ASCII characters in numbers. This introduces a to_unicode() function in util so that it can be used by multiple modules.
* Handle unicode arguments in get_cc_module()David Arnold2017-10-141-0/+19
| | | | Closes https://github.com/arthurdejong/python-stdnum/issues/54
* Ensure 100% branch coverageArthur de Jong2017-08-281-0/+50
This ensures that the tests fail if 100% branch coverage is not achieved. It also adds some pragma statements for code that cannot be covered or is Python version dependent. Furthermore, the get_module_list() function was removed from stdnum.util and more tests were made from stdnum.util and stdnum.numdb. The functionality to call format() in a country-specific IBAN implementation was also dropped because it was not used.