This test-case tests inter alia whether the Ethiopic digits x1369 through x1371 match the regex \d which equates to the character category Nd. These digits were moved from category Nd to No in Unicode 4.1 (apparently on the grounds that the Ethiopic numbering system that uses these digits is not a standard decimal numbering). Since we do not mandate a specific version of Unicode, the results of the test are therefore implementation-defined.

I propose to fix this by removing the offending characters from the test case.