Re: [bug-gawk] Unicode hyphens in man pages

[Neil R. Ormos]

address@hidden wrote:
> Daniel Reimer <address@hidden> wrote:

>> I'm not familiar with formatting standards for
>> man page sources, but when I use 'man awk' (awk
>> is linked to gawk on my up-to-date arch linux,
>> I presume that the awk man page comes from
>> gawk), I get unicode hyphens or minus or
>> whatever it is called ('???', code 8722) before
>> some options like '-F'.

>> This is very annoying since it makes it
>> impossible to search for options reliably (or
>> at all, since I don't have unicode characters
>> on my keyboard). Is there any particular reason
>> for this? A quick search didn't find unicode
>> hyphens in any other man page (there are none
>> in the gawk man page).

> The gawk man page follows standard troff
> conventions of using \- as the dash for options
> like -F, -v and so on.  It is the GNU troff
> suite that turns that character into a Unicode
> hypen.

> I suspect that if you do

>       LC_ALL=C man gawk

> the Unicode characters will go away and come out
> as plain old ASCII minus signs.

It seems odd that the original poster reports
seeing Unicode characters on the awk man page but
not on the gawk man page.

On my Debian systems, awk.1.gz is symlinked via
the alternatives mechanism to gawk.1.gz, so I
would expect identical treatment by the formatting
chain invoked by main.  (I don't know how this is
handled in Arch.)

Does the differential display of hyphens suggest
that the awk man page the OP is viewing is really
not the gawk man page?