[packages/spdlog] add note about implications of cxx20 bcond

[Jakub Bogusz]

On Tue, Nov 03, 2026 at 01:34:42AM +0100, Jan Rękorajski wrote:
> On Tue, 03 Nov 2026, atler wrote:
> 
> > On 02.11.2026 23:54, atler wrote:
> > > commit 0bee8bcc2cf48d27effeb15df5d845995f511524
> > > Author: Jan Palus <atler at pld-linux.org>
> > > Date:   Mon Nov 2 23:53:51 2026 +0100
> > > 
> > >     add note about implications of cxx20 bcond
> > > 
> > >  spdlog.spec | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > ---
> > > diff --git a/spdlog.spec b/spdlog.spec
> > > index 53f8787..3750e02 100644
> > > --- a/spdlog.spec
> > > +++ b/spdlog.spec
> > > @@ -1,6 +1,6 @@
> > >  #
> > >  # Conditional build:
> > > -%bcond_without	cxx20	# C++ 20 std::format instead of libfmt
> > > +%bcond_without	cxx20	# C++ 20 std::format instead of libfmt (users of library must be built with C++20 too)
> > 
> > Does anyone have strong preference for keeping this bcond enabled? C++
> > standard matching dance between library X which enforces -std=c++17 in its
> > CXXFLAGS, library Y which requires C++20 but does not enforce it in any
> > way, all when compiling application Z unaware of any of this is becoming
> > tiring. spdlog built with libfmt is more lenient which makes me want to
> > do the switch.
> 
> I'd say keep it simple and turn it off. C++ standards are already a
> pain, let's not add more.

OK, let's stick to libfmt for now. I changed the default in spec.


-- 
Jakub Bogusz    http://qboosh.pl/