Hello mark,

On 07 Jun 15 19:46, mark lewis wrote to Nicholas Boel:

 NB>> News to me I suppose. I have MsgEd installed here and all it is
 NB>> provided with is translation tables that only support 128
 NB>> character translations, not the full 256. That's why I originally
 NB>> went with Golded in the first place.

 ml> if you look at the ID number in the chs files, it is the first
 ml> non-comment line, and it is greater than 65536, then the entire 256
 ml> set should be listed in the lines... if the ID is less than 65536,
 ml> then only the top 128 are listed... this because most sets have the
 ml> same lower 128 so there's no need to translate them...

Then I can only assume most of the translation tables provided with MsgEd are
only translating the upper 128 then? I may have to take another look at MsgEd
and see if I can accomplish the same thing with it then as I've done with
GoldEd.

 NB>> I didn't convert my golded config fully over to CP437 the last
 NB>> time, as I was in a hurry to leave town. Now I have it converting
 NB>> properly (note the CHRS kludge on this message) and all your
 NB>> stats posts are displaying correctly.

 ml> i haven't figured that part out yet... right now i'm using something
 ml> called cp437 which sets up whatever you have it execute so that all
 ml> the characters are run through iconv for conversion to utf8 for
 ml> display purposes... i don't think it works in reverse but i just
 ml> posted another character test in the FIDOTEST echo to see...

Okay, so what I noticed over in FIDOTEST of you posting your umlaut o was
indeed a conversion from CP437 to UTF-8? If so, you're flying the CP437 CHRS
kludge yet and probably shouldn't be (only if that is the case).

 NB>> Confirmed. They display fine here now.

 ml> with this cp437, i can now see them... all of the block graphics used
 ml> in the bar graphs, too :)

I haven't heard of that until someone mentioned it recently. However, I can
accomplish this without any external utilities. See below..

 ml>>> i might try that... it seems to be similar to the 437_437,
 ml>>> 850_850 and others of the same nature... except that those all
 ml>>> contain the upper 256 characters whereas yours doesn't have any
 ml>>> at all... hummm...

 NB>> That still most likely won't get your line and box drawing
 NB>> characters you're looking for. I have a config here that let's me
 NB>> stick with UTF-8 on my console, but displays CP437 just fine at
 NB>> the moment. I'll verify to make sure everything is in working
 NB>> order and then see what I can do to give you a hand.

 ml> that would be nice... thanks!

 NB>> You mentioned a couple different approaches already.. so I'm a
 NB>> little confused. Are you using KDE's "Konsole" or an actual
 NB>> gui-less TTY console?

 ml> on /this/ system, i'm in a GUI... on others, there is no GUI so
 ml> they're raw TTY...

 NB>> If the former, setfont may not work in a gui program as I think
 NB>> it's specifically for a gui-less console. You would have to use
 NB>> Konsole's settings for changing the font and/or xorg.conf
 NB>> settings as X handles fonts and character sets differently.

 ml> yeah, that's what i've found with setfont (again)... i had forgotten
 ml> that it was limited to raw TTYs...

 NB>> I don't use X at all here, so while my config seems to work for
 NB>> me it's only in a gui-less environment and utilizes the
 NB>> terminus-fonts package that were installed to
 NB>> /usr/share/kbd/consolefonts (via Archlinux's pacman package
 NB>> manager. More than likely a bit different in Kubuntu).

 ml> i think i lost them some time back... i remember hunting them down and
 ml> trying to work with them...

 NB>> If you're trying to do all of this in an X environment, what I've
 NB>> been saying may not help you at all. :(

 ml> we can only find out :)

So far this is what I've been doing for CP437 in a console TTY on a pure UTF-8
console.

echo -e "\033%@" && echo -e "\033(U"
setfont ter-i14b    # 256 glyph terminus font

In golded.cfg, I use xlatcharset to translate everything to and from CP437, ie:

xlatcharset    cp850    cp437    850_437.chs
xlatcharset    cp437    cp850    437_850.chs

And so on for every other possible character set available in the translation
tables, even including 437_437.chs at the end of the list. Then:

xlatlocalset cp437
xlatimport   cp437
xlatexport   cp437

..and everything displays as it should in regards to line and box drawing
characters. With this config, I don't use any UTF-8 conversions whatsoever.

Then when I do my UTF-8 configuration, I have a completely different set of
conversions, first:

reset
setfont ter-v14b   # 512 glyph terminus font

xlatcharset    cp1125    utf-8    1125_u8.chs
xlatcharset    ibmpc     utf-8    437_u8.chs
xlatcharset    cp437     utf-8    437_u8.chs
xlatcharset    cp850     utf-8    850_utf8.chs

* NOTE: If you don't have any of these files, I can easily send them over. I
don't remember if they were supplied with GoldEd or if I got them from
somewhere else.

And I also have tables for cp865, cp866, koi8-r, latin-1, and latin-9 (I'm
guessing you see my point so I don't have to type every one out. :). At the end
of that list I use my utf8_utf8 empty table to aquire the level 4 parameter in
the CHRS kludge.

I don't translate anything FROM UTF-8, as it's not necessary. Then:

xlatlocalset utf-8
xlatimport   utf-8
xlatexport   utf-8

And away I go. Obviously Russian or Ukranian characters don't display properly
in GoldEd itself, but when I shell out to nano they display perfectly, or at
least as far as the 512 glyph font will allow anyways. Otherwise umlaut,
cedilla, grave, etc. characters all display perfectly both inside and outside
of Golded for the most part. The only downfall is I now don't see any of the
CP437 line or box drawing characters. But since I rarely read the stats echos
unless I need to, it's not much of a burden, and if needbe I can always switch
my config back and forth on the fly without issue.

Also, when I switch to my CP437 config, I must comment out /all/ my UTF-8
settings above, otherwise it will somehow try to force UTF-8, while the other
way around (switching to my UTF-8 configuration) everything can be left
uncommented (including all the CP437 stuff, minus xlatlocalset, xlatimport, and
xlatexport) without issue. Kinda wierd but just something I've learned in the
process. I hadn't commented out my UTF-8 stuff the first time I looked at your
STATS postings, which is why it didn't show properly. Once I did that it was
fine.

Anyhow, hope that helps you get somewhere, although as I mentioned, this may
only work for me since I'm not using X.

Regards,
Nick

--- GoldED+/LNX 1.1.5-b20130910