tsoome_: LANG=C.UTF-8.. LC_CTYPE="C.UTF-8" etc

tsoome_: as in, like this: 0x0.st/XVcl.png

you can see the unicode blocks showing up as M instead

I asked, because LC_* varaibles will override LANG. However, what does stty -a report?

tsoome_: 0x0.st/XVmi.png

tsoome_: xterm on top, xfce4-terminal on bottom

Oh, interesting. It's not working in durdraw, but the characters work if I cat a unicode file

Hmm. Durdraw uses python curses

Yep. Must be something with how durdraw is using curses and/or python. 0x0.st/XVmX.png

Hm. Does that need ncursesw as opposed to regular ncurses?

Hmm. Maybe. :) It shouldn't really, though I'm not sure what happens if you try to use wide characters in regular curses. durdraw will draw wide charaters, but isn't aware of them

the calls should be just to regular curses, through the python stdlib

but it is using some panels features that aren't in older curses versions

that said, I probably developed it against ncursesw

IndyJoenz thats output from gnu stty, what does /bin/stty show?

tsoome_: 0x0.st/XVmq.png

ok, so terminal lines should pass data as expected...

we could also check that with dtrace but given stty output should be sufficient...

Yep, it needs ncursesw - just built a temporary copy for testing

ptribble: nice :)

Which means it would work in other distributions, which explicitly enable ncursesw, but Tribblix always starts with whatever upstream ships as defaults and only changes them when a problem shows up

Ahh. I was wondering why the system python build wasn't sufficient

I think python just looks for some curses library and isn't too bothered about the details

But anyway I'll switch ncurses to ncursesw and we'll see how much else I have to fix

there shouldn't be too much, I did the switch recently in pkgsrc, and most of that was mechanical to support NetBSD's native curses

What I meant was breakage without rebuilding, as the docs says that wide and non-wide aren't binary compatible

ah, that's a foreign concept to me ;)

[illumos-gate] 15279 lex: clang build error -- Toomas Soome <tsoome⊙mc>

[illumos-gate] 16687 iasl: storing the address of local variable -- Toomas Soome <tsoome⊙mc>

[illumos-gate] 16684 Solaris 10 zones and libproc broken by new parameter to fcntl -- Robert Mustacchi <rm⊙fo>

btw, the thread starting at lists.gnu.org/archive/html/autoconf/2024-08/msg00006.html is about illumos ksh93 not playing nicely with GNU autoconf

Thanks alanc. It's reported on a pre illumos 13405 system so I expect current systems are okay but I'll offer to help them test if the OP doesn't volunteer. (fenix)

FEATURE 13405: ksh93 update to 2012-08-01 (Closed)

↳ illumos.org/issues/13405 | code.illumos.org/c/illumos-gate/+/1176

Thanks for taking a look, Andy.

alanc: lol

I suppose that's how you know you've had a good idea, at least

if you combine both .Xr and .Lk in a SEE ALSO, it kind gets difficult to read. Do we have an idiomatic way to put a line break in to tidy things up?

I've used regular old `.sp`, but it feels weird in mdoc

.Pp ?

I think that's a full paragraph break, rather than just a newline?

mandoc.bsd.lv/mdoc/appendix/markup.html has some notes but it's hard to interpret. `.br` is mentioned

well, it works as well as `.sp`, and it was mentioned, so I'll do that. thanks!