Re: mysqlIndexerTest.test_unicode fails for me

In message <[email protected]>,
Christof Meerwald writes:
>On Wed, Apr 29, 2020 at 07:56:39PM -0400, John P. Rouillard wrote:
>> Then only special thing I see is:
>> 
>>        # HACK: workaround mysql bug: http://bugs.mysql.com/bug.php?id=74901
>>        #   needed for test_mysql.mysqlDBTest.testFilteringSpecialChars
>>        sed -i -e 's/CREATE DATABASE \%s/CREATE DATABASE \%s COLLATE utf8_general_ci/' roundup/backends/back_mysql.py
>> 
>> Could that have something to do with it?
>
>Yes, I copied that straight from the travis config
>
>> Also Christof, is the test coverage info pushed off to some app or is
>> it just visible as part of the circleci log?
>
>At the moment it's only available in circleci.
>
>BTW, I have now added a test run based on Ubuntu 20.04
>
>The reverse multilink tests seem to be failing on all platforms:
>
>https://circleci.com/bb/cmeerw/roundup-tracker/76
>https://circleci.com/bb/cmeerw/roundup-tracker/79
>https://circleci.com/bb/cmeerw/roundup-tracker/81

Also the mysql encoding change is failing in travis:

https://travis-ci.org/github/roundup-tracker/roundup/jobs/681522097#L6967


  =================================== FAILURES ===================================

  ________________________ mysqlDBTest.testStringUnicode _________________________

  self = <test.test_mysql.mysqlDBTest testMethod=testStringUnicode>

      def testStringUnicode(self):

	  # test set & retrieve

	  ustr = u2s(u'\xe4\xf6\xfc\u20ac')

	  nid = self.db.issue.create(title=ustr, status='1')

  >       self.assertEqual(self.db.issue.get(nid, 'title'), ustr)

  test/db_test_base.py:302: 

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
  ...
            if value is not None:

>               value = self.to_hyperdb_value(props[name].__class__)(value)

E               UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-3: ordinal not in range(128)

   roundup/backends/rdbms_common.py:1205: UnicodeEncodeError

which makes sense since it was working with deault (latin1??)
encoding, so making it utf8 changes that.

--
				-- rouilj
John Rouillard
===========================================================================
My employers don't acknowledge my existence much less my opinions.