Hi Grégoire,
> > troff takes ISO-8859-1 as its default input. If you want to feed it
> > UTF-8 then look at groff's -k option. Try
> >
> > printf 'testé\n' | groff -k -me -X
>
> thanks a lot for your answer. Your command works fine. I wondered
> where do you have this information from:
I've picked it up from subscribing to this list and seeing others'
answers over time. :-)
> I was searching in man groff about the -k option, than in man preconv.
> I understood that -k make use of preconv, which can convert utf-8 to
> Latin1.
Yes, they are the right places to read. GNU troff understands \[u1234]
for Unicode codepoint U+1234 and preconv produces these.
> I can find out by myself with which caracters I will have to use "\n".
> But try this:
>
> printf 'testé\n, ü\nber, ç\na' | groff -k -me -X
>
> There is a space too much between é and , because of \n
> Is there a solution for that?
That had me puzzled for a while but the penny has dropped. I wrote \n
not because it's needed for groff -k to handle the preceding character
but because printf(1), unlike echo(1), doesn't automatically add a
linefeed to its output and \n is the "escape" to do this. Try these to
see what I mean.
printf 'foo\nbar\n'
printf 'testé\n, ü\nber, ç\na'
printf foo
> I tried something myself: The following command is not the solution:
> printf 'testé\, ü\ber, ç\a' | groff -k -me -X
This should work just fine.
printf 'testé, über, ça\n' | groff -k -me -X
> ... further I was reading around in the manual about groff -k and
> after which characters you have to use \n. I didn't found anything.
\n does mean something to troff too, but the \n given to printf is
interpreted by printf and turned into a single byte, ASCII LF, value 10,
before troff gets to see it. \n in troff makes it interpolate the value
of the given numeric register.
$ printf '.nr x 42\nNumbers: \\nx \\n(fo \\n[bar]\n' |
> groff -rfo=314 -rbar=281 -kX
Again, do just the printf to see groff's input as if you had typed it in
a file. \\ is the escape to have printf produce a single \
Cheers, Ralph.
