UTF-8 OK in Dev.Sytem but not in executable

[cbx]

I am using LabVIEW since version 5.0 (on Windows NT4) and have switched to Linux with version 7.0 and am bnow using 8.2.1. The system is very up to date (Kernel 2.6.20, glibc 2.5, Xorg 7.2). LabVIEW is really great, but one problem keeps annoying me ever since I've switched to Linux.

Our systems are running UTF-8 character encoding and the professional development system seems to support that well. But as soon as I build an executable, the texts somehow fall back to something like ISO-8859. I have attached a screenshot showing the same VI, once run in the dev system and once run from the project explorer. It can easily be seen that in the dev system, the string gets encoded correctly (escaped with $C3), whereas the compiled version reads input as ISO-8859 and misinterprets the UTF-8 encoded strings inside the indicators.

I know, Linux is tricky and our distribution (Gentoo) is not suppotred, but I'd really appreciate any hint pointing me in the right direction.

Thanks!
/cbx

[Arsala]

Hello cbx,
what a poehtic text.
Yes the LV professional development system support UTF-8. There is a  text to UTF-8.vi subVI in the SMTP Send Email Send Message.vi that you can search in LV.
That should be ok, I've attached it.
Could you post the whole project. I will see if the VI works fine on a Windows OS.

Regards
Arsala
National Instruments

[cbx]

Thanks Arsala,

in very rare and desperate cases I post to some forum as a last resort because in almost every case I get it fixed shortly afterwards. It also helped this time. Yesterday I installed Ubuntu inside a VM and checked. Both variants worked correctly! So the flaw had to be caused by my system setup.

What I found out so far: Some time ago, something must have damaged my "de_DE.utf8" GLIBC locale. Nobody cared, KDE was working properly as well as all my applications, including the LV development system. Only the runtime engine based programs fell back to some ISO-8859 sort of thing. After I rebuilt the GLIBC locale, the effect was gone and the executable was fully UTF-8 aware again.

OK, it was my fault in the first place, but one thing keeps nagging: Where does the dramatically different behaviour between development system and runtime engine come from? (OK, you can treat this as a rethorical question )

thx for your reply anyway
/cbx