Re: [tbl] [regression] Formatting glitch in table separators

[Alejandro Colomar]

Hi Branden,

On 4/30/23 17:00, G. Branden Robinson wrote:
> Hi Alex,
> 
> At 2023-04-30T16:40:11+0200, Alejandro Colomar wrote:
>> This one seems serious, and not just a minor inconvenience.  You can
>> reproduce it in the current ascii.7 page from the Linux Manual Pages.
>> The table separator is gone (except for some aberration at the end,
>> which even invades the following subsection).
> 
> I know about this one, and regard it as cosmetic.

Yep, it's cosmetic; bigger cosmetics than some innocuous extra vertical
space, but cosmetic.

> 
> Bjarni reported it back on 25 March.
> 
> https://savannah.gnu.org/bugs/?63960

Good.

> 
> It is assuredly ugly--and please have no doubt that I find it highly
> irritating--but it destroys no information.  On the bright side, there's
> no longer a spurious blank line in the midst of the table.

:)

> 
>> I don't know if the CW font selection right above the table is
>> somehow guilty).
> 
> Nope, nothing to do with it.  The ugliness arises from two factors:
> 
> 1.  The "continuous rendering" mode used by groff man(7) and mdoc(7)
>     uses a technique that is more complicated than simply setting the
>     page length to a huge value, but then shortening it when the page
>     text ends, leaving only enough room for format the page footer.
> 
>     (Long-term, I think making that change will resolve #63960, but not
>     all of the bad behavior.)
> 
> 2.  GNU tbl deliberately draws horizontal and vertical rules too long on
>     nroff devices.  This was a deliberate design choice, and it's there
>     for a good reason, but in my opinion it is too ugly to live and
>     should replaced with another method, for which I have ideas.
> 
>     https://savannah.gnu.org/bugs/?62471
> 
> If you can't stand for it to be in your man page, ditch the vertical
> rule from the table format.
> 
> -l l l l | l l l l.
> +l l l l l l l l.
> 
> People will still understand the table.  And you can put it back when I
> get the above 2 bugs fixed.

I prefer having it there.  It doesn't hurt more than not having the rule
at all, apart from being quite weird.  But it's clearly a bug, and I don't
think it will confuse anyone.

Cheers,
Alex

> 
> Regards,
> Branden

-- 
<http://www.alejandro-colomar.es/>
GPG key fingerprint: A9348594CE31283A826FBDD8D57633D441E25BB5

OpenPGP_signature
Description: OpenPGP digital signature