Re: [AD] configure script: feature or bug ?

[Eric Botcazou <ebotcazou@xxxxxxxxxx>]

> Silencing requires an "if" statement around the check anyway, doesn't
> it?

Yes it does, but the "if" statement is already there for every non-silenced 
check, followed by an "else value=no". Currently the value is displayed 
unconditionally, I propose that we display it conditionally (i.e move the 
AC_MSG_CHECKING inside the "if" statement).

> In that case I'd prefer to see an explicit message saying that the
> feature was disabled.  It doesn't matter to me if the message doesn't
> look like a "genuine" autoconf message (although it probably matters to
> you :-)

IMHO if the messages were not (strictly) identical, this feature would have 
little value, since you couldn't compare two runs of ./configure; for example

   checking for linux/fb.h...yes

and

   checking for linux fbcon support....disabled

I can do that, but I don't like it very much. Moreover, in that case, should 
the configure script display a message for features that can be 
enabled/disabled but don't have any associated check (dbglib for example) ?

-- 
Eric Botcazou