On Sun, Sep 05, 2004 at 11:30:55 +0200, Jörn Reder wrote:
Jan Hudec wrote:Ok. So I want one thing. If I encode a script in unicode, I want all string literals to behave the same and have the utf8 flag set. But some of them happen not to contain any non-ascii characters. Now, how do I do this?I fear that's not possible. 'use utf8' enables the utf8 flag only for literals which contain bytes with the 8th bit set. 7bit characaters are left without utf8 flag set. But why do you want to have all you variables utf8 tagged? In general you don't need to care about the utf8 flag, because Perl upgrades the variable when needed. (besides the fact, that there may still be bugs in Perl's internal utf8 handling... ;) The utf8 flag only becomes important when you're doing I/O, write to filehandles or pass arguments to external programs etc.. E.g. you need to apply the correct encoding layer to all your filehandles, otherwise Perl can't do the right thing for you.
Um, yes. It shouldn't matter whether the strings, that are pure ascii,
are marked utf8 or not. (If there were no bugs in the handling). Since
the conversion is nop on them.
What is a little problem is, that there is no standard module to do the
I/O stream setting in a reasonably automagic way, like eg. tcl does.
Basicaly you have to set all the streams manualy.
-------------------------------------------------------------------------------
Jan 'Bulb' Hudec <bulb ucw cz>
Attachment:
signature.asc
Description: Digital signature