in reply to
Re^3: Pragma to handle unicode characters
in thread Pragma to handle unicode characters
You misunderstood the situation. The victims of the problem never intended to use locale information in any way in their scripts, and the scripts were not written to use locale information.
Can you point me to this problem? Seems, i don't understand it throughly. My problem is that i can't rely on system locale even i use locale pragma. As far i understand, those victims were dependent of their system locale even without using locale pragma? Was it like this?
I agree, that such behaviour will break earlier code and such move needs long time to adapt before any deprecation (if needed). I need the middle way: locale does not spread without using pragma, but when using it, it should flawlessly spread.