I rebased the patch on the current default branch and pushed it here:
https://hg.savannah.gnu.org/hgweb/octave/rev/08a01acc2fcf

Sorry it took more than five years.

Marking as ready for test

Attached is a patch that will use UTF-8 encoding of symbols in all gnuplot terminal cases except the PostScript class of terminals, 'postscript' and 'epscairo'.  This appears to work well on my system.  There may be some terminals like 'png' that don't function exactly right, and if that is the case I'll investigate the issue in the gnuplot code.

Here's a reference to the TeX utf-8 glyphs:

http://www.w3.org/2001/06/utf-8-test/TeX.html

I was tempted to write a little script that creates the sym structure automatically so that all of the entries on that page are accounted for, but I just added the utf-8 characters that were present along side the 'Symbol' definitions in a cell.

Here's a test sequence:

setenv('GNUTERM', 'qt')
graphics_toolkit gnuplot
h = text (0.5,0.6, 'H \0 \Gamma \aleph I')
set(h,'fontweight', 'bold')

(file #34373)