UTF8 coding in TCL ?

I believe that you are doing this the wrong way.
The syntax of the essage text includes the sequence [_ and that is an indication that the strings are included in the automatic language conversion system.

You should be editing (or creating) a new .po file, and those can definitely include non-latin characters, see for example:
github.com/LinuxCNC/linuxcnc/blob/master/src/po/zh_HK.po

Some instructions are here:
wiki.linuxcnc.org/cgi-bin/wiki.pl?Internationalization

It should be easier to create a translation this way, as you only need to edit one file, and all the messages are in the same place.
The guide says that you need to install git and compile LinuxCNC from source. I don't know for sure if that is the case. I would be tempted to try creating a .po file and translating a few messages to see what happens.

If you are willing to create a complete translation file then I am sure we can help you to include it in the LinuxCNC distribution.

If you open a terminal and type "locale" then you should get something like
LANG=tr_TR.UTF-8

So, you would need to re-name the linuxcnc.pot file to tr_TR.po and then translate all the messages in it.

Unfortunately it looks like the .pot files don't exist until after the compilation stage, and I am not at a Linux PC at the moment to create one.
You can work with one of the files that does exist here:
github.com/LinuxCNC/linuxcnc/tree/2957cc...14909c09a5fb7/src/po
But the problem with that is that any untranslated messages will be in neither Tukish nor English, but whatever source file you chose.