[Mimersbrunn-tcltk] msg files and encodings inequal iso8859-1

[Harald O. <har...@el...>]

Dear list,

I want to know your thoughts how to use msg files with other encodings
than iso8859-1.

I have done an application for greece with tcl using iso8859-7 as system
encoding, running on Windows (english and greek) Linux and sco-unix (4
codepages).
I typed anything in an iso-8859-7 editor which is not a problem.
Each time a string with greek characters apears I use:
[encoding convertfrom iso8859-7 "???"] or I use
"\u03a8\u03a9\u03aa"

In the message file I use two global functions, se for english entries (in
iso-8859-1) and sg for greec entries (in iso-8859-7). I have copied them
at the end of this message.

Anyway I am not satisfied with the solution.
I would prefer to:

- Generally type msg strings in UTF-8 and build a mechanism in msgcat that
it uses this encoding for the last argument of mcset (As tags I generally
use symbolic tags and no printable text which makes programming easier
(codepage horror)).

- specify the encoding of the mcset command somewhere, the best would be
with the language. When language is gr use iso8859-7. When language is en
or de or fr use iso8859-1. Or use the system encoding.

What do you think ?

Best regards,
	Harald

Appendix: Functions sg and se
 proc sg {args} {
	upvar l l
	if {1 == [llength $args]} {
		eval ::msgcat::mcset $l $args
	} else {
		eval ::msgcat::mcset $l [lindex $args 0]\
			[encoding convertfrom iso8859-7 [lindex $args 1]]
	}
}
proc se {args} {
	upvar l l
	eval ::msgcat::mcset $l $args
}