unicode-org
NormalizationTest.txt additions
I stumbled over an email from 2017 that I sent to @macchiati asking how & where to track stuff like this. Well, now we have an issue tracker here...
Email body minus the where-to-track question:
These test cases exposed bugs in ICU normalization boundary detection that would have caused bad normalized strings in ICU 60:
http://bugs.icu-project.org/trac/changeset/40355#file10
Also, Jim found strings for which the old Google normalization code did not work right. They would probably be good for adding to the test data too. I can help you dig those up.
That SVN/Trac commit from 2017-aug-25 was migrated to GitHub as https://github.com/unicode-org/icu/commit/2f87cf4c46e23e2f61af2b9b64b75e16544e985b
ICU-10524 normalization one-way mapping with trailing ccc>1 has no compose-boundary-after
X-SVN-Rev: 40355
I assume that "file10" was either the C++ or Java test code file adding the TestComposeBoundaryAfter() function.
I am not sure that I can still figure out the set of “strings for which the old Google normalization code did not work right”.
