Long time has passed since it was coded. So to refresh the context, whether you were following or not, the utf8-cmdline branch makes the command-line tools to first convert entire argv into UTF-8 and then assume normalised strings when handling parameters. A work that minimises the amount of potentially forgotten-to-convert strings and brings support for all Unicode symbols in the cmdline on Windows which is impossible right now.
There is a test called basic_tests.py:argv_with_best_fit_chars. It checks that svn rejects Unicode symbols. Functionality which was illegal before changes introduced in that branch. I believe the correct solution would be to adjust test cases to expect successful execution instead of an error. I've tried to implement that before by changing arguments in run_and_verify_svn2, but it made me stuck... I kept getting an error message coming from the logger (?) that it struggles to convert output to print it. I really lack expertise with Python and svn-test to figure out the correct way to do it. Could anyone who knows it better than I do step ahead and maybe give it a fresh look from another perspective? It's basically the last thing that keeps us from getting this branch merged so far. I think it's time we finalise it. Alternatively I think we might remove that test as "outdated" because it checks an edge case which is now completely normal after the utf8ization. I would personally prefer the other option, but it's worth mentioning this possibility as well. Thanks! -- Timofei Zhakov
