Re: [RFC] Docs: document silent make rules in a new chapter

[Stefano Lattarini]

On Saturday 20 November 2010, Ralf Wildenhues wrote:
> * Stefano Lattarini wrote on Thu, Nov 18, 2010 at 09:22:48PM CET:
> > On Thursday 18 November 2010, Nick Bowler wrote:
> > > On 2010-11-18 20:31 +0100, Stefano Lattarini wrote:
> > > > address@hidden @code{AM_V_GEN}
> > > > address@hidden FIXME: wouldn't $(AM_V_SILENT) be clearer?  Should we 
> > > > deprecate
> > > > address@hidden $(AM_V_at)?  It should be kept for 
> > > > backward-compatibility, of
> > > > address@hidden course.
> > > 
> > > AM_V_GEN is a long enough name as it is; AM_V_SILENT would be even worse
> > > in this regard.
> > > 
> > > AM_V_at is very useful for targets which have multiple commands.  It's
> > > not that interesting to see "GEN foo.bar" five times in a row.
> > > 
> > There's probably a misunderstanding here; I was suggesting to rename
> > `AM_V_at' to `AM_V_SILENT', for clarity; and deprecate *only* the old
> > name `AM_V_at'.  Does my proposal make sense now?
> 
> It makes sense, but it's a long name.  It's a close call I'd say but
> I wouldn't want to deprecate AM_V_at, simply because it is shorter.
> Other renaming suggestions have been made before, see e.g. this thread:
> <http://lists.gnu.org/archive/html/bug-automake/2010-04/msg00001.html>
> 
> But I'm quite hesitant to do any renames at all unless there is a clear
> advantage.  Automake has had a slightly bad reputation in the past for
> not being backward compatible, and I wouldn't want that to return.
>
In this case that shouldn't be a problem, since I'm not proposing to
remove AM_V_at, but only to deprecate it in favor of the new alternative.
Anyway ...

> (And I don't like overly verbose makefiles with lots of
> duplication either.)
> 
... I'm fine with this; I'll just rewrite the fixme comment to
reference the thread above and to be more "possibilist":

 @c FIXME: Could we find a better name than $(AM_V_at)?  $(AM_V_SILENT)
 @c is nice, but also a bit too long.
 @c If we find such a better name, we should then deprecate $(AM_V_at),
 @c but *not* remove it: it should be kept for backward-compatibility.

Thanks,
   Stefano