On Sat, Jun 26, 2021 at 08:06:58PM +0200, Theo Buehler wrote:
> On Sat, Jun 26, 2021 at 07:20:52PM +0200, Ingo Schwarze wrote:
> > Hi Jason and Theo,
> > 
> > Jason McIntyre wrote on Tue, Jun 22, 2021 at 06:37:27AM +0100:
> > > On Tue, Jun 22, 2021 at 04:48:39AM +0200, Theo Buehler wrote:
> > 
> > >> You have two overlong lines as indicated below. I would have thought
> > >> that mandoc -Tlint complains about that, but apparently it doesn't have
> > >> such a warning... With those wrapped,
> >  
> > > yes, there is no feedback on long lines. although we try to keep the
> > > source less than 80 width, there are some places where it is not
> > > possible.
> > > 
> > > i'm not sure whether adding a warning would be helpful or disruptive.
> > 
> > Here is a patch implementing such a style warning, leaning very
> > heavily into the direction of never producing false positives, that
> > is, not warning about long lines
> > 
> >  - in no-fill mode (e.g., .Bd -literal, .EX, .nf and the like)
> >  - that start with a dot (normal macro and request lines)
> >    or with a non-standard control character
> >  - that start with a space character or with an escape sequence
> >  - in tbl(7) context
> >  - in eqn(7) context
> >  - that do not contain a blank character before column 80
> 
> This sounds perfectly reasonable. While I don't think I'm likely to
> introduce overlong lines myself, I believe it's a useful warning for
> development. Since you've already done the work, I am in favor of
> committing this. It's not expensive at all in terms of runtime and code
> complexity. I agree that it won't be disruptive.
> 
> The code reads fine, thanks!

Agreed.

OK kn

Reply via email to