On Wed, 2020-08-05 at 13:32 +0200, Pavel Hrdina wrote:
> On Wed, Aug 05, 2020 at 01:16:01PM +0200, Andrea Bolognani wrote:
> > On Wed, 2020-08-05 at 12:51 +0200, Pavel Hrdina wrote:
> > > + correct_rl = cc.has_header_symbol('readline/readline.h', 'rl_completion_quote_character', prefix: '#include <stdio.h>')
> > > + if not correct_rl
> > > + if get_option('readline').enabled()
> > > + error('readline is missing rl_completion_quote_character')
> >
> > This matches the original error message, but thinking about it again
> > I would suggest changing it to something like
> >
> > error('readline >= @0@ was not found'.format(readline_version))
> >
> > because the specifics of the symbol are not important here, we just
> > happen to be using it as a witness.
>
> Most of the error messages could use some rewording and I feel that it
> should be done separately.
Right. For some reason I mistakenly thought you were re-introducing
an error message that was missing from the Meson rewrite, but I see
now you're just moving it, so I completely agree with you.
--
Andrea Bolognani / Red Hat / Virtualization
