Eryk Sun <eryk...@gmail.com> added the comment:
> If it was the tests, they seem to have been fixed elsewhere. The cited tests haven't been changed to work with non-normalized sys.executable, sys.prefix, etc. But it's not common to run the test suite with a non-normalized path such as "Lib/../build/python". > you stated above that on Windows a \\?\ path is possible but > will not work for startup, and that you would open an issue Sorry, I don't remember the details of the bug that I mentioned. But it became a non-issue in 3.6+, which canonicalizes the executable path. For example: >>> exe = r'\\?\C:\Program Files\Python36\python.exe' >>> cmd = 'python -c "import sys; print(sys.executable)"' >>> subprocess.call(cmd, executable=exe) C:\Program Files\Python36\python.exe 0 Compare the latter to the output of GetModuleFileNameW(NULL, ...) in this case, which returns a \\?\ extended (verbatim) path: >>> exe = r'\\?\C:\Program Files\Python36\python.exe' >>> cmd = 'python -c "import _winapi; print(_winapi.GetModuleFileName(0))"' >>> subprocess.call(cmd, executable=exe) \\?\C:\Program Files\Python36\python.exe 0 > Do any of the non-normalized Linux (or *nix) executable paths fail for > startup? Not that I'm aware of. AFAIK, it's just about the particular tests failing in this case. ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue31030> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com