Re: 1.23: UTF-8 device produces mysterious characters

[Dave Kemper]

On 9/13/22, Steffen Nurpmeso <steffen@sdaoden.eu> wrote:
> Dave Kemper wrote in 
> <CACRhBXOuDOhOaPzuz56-vY0AVH7Ybx5i9B97Cg-hk+Dw+NbDPA@mail.gmail.com>:
>  |Groff has converted an input U+002D to a proper hyphen in typeset
>  |output for decades.  It has done so in UTF-8 output since at least
>  |groff 1.19.2.
>
> This is not true.  I never have seen anything but hyphen-minus on
> my boxes, and i was running 1.22.3 until last week.

I will ask again that you please post specific example input, along
with what output you saw in groff 1.22.3 and what is different in
1.22.4.

I ran the test before making my statement, verifying that it's true.
It is, however, locale-dependent:

$ groff --version | head -1
GNU groff version 1.19.2
$ echo $LC_CTYPE
C
$ nroff <<< go-to | cat -s | hexdump -C
00000000  67 6f 2d 74 6f 0a 0a                              |go-to..|
00000007
$ LC_CTYPE=en_US.UTF-8
$ nroff <<< go-to | cat -s | hexdump -C
00000000  67 6f e2 80 90 74 6f 0a  0a                       |go...to..|
00000009

Did you change your system's locale settings around the same time you
upgraded your groff installation?  You might be seeing a change from
one and attributing it to the other.