Re: [PATCH 108/143] meson: convert hw/intc

[Philippe Mathieu-Daudé]

On 8/6/20 11:20 PM, Paolo Bonzini wrote:
> 
> 
> Il gio 6 ago 2020, 23:02 Peter Maydell <peter.maydell@linaro.org
> <mailto:peter.maydell@linaro.org>> ha scritto:
> 
>     On Thu, 6 Aug 2020 at 21:10, Paolo Bonzini <pbonzini@redhat.com
>     <mailto:pbonzini@redhat.com>> wrote:
>     >
>     > From: Marc-André Lureau <marcandre.lureau@redhat.com
>     <mailto:marcandre.lureau@redhat.com>>
>     >
>     > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com
>     <mailto:marcandre.lureau@redhat.com>>
>     > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com
>     <mailto:pbonzini@redhat.com>>
> 
>     Compared to:
> 
>     > -common-obj-$(CONFIG_HEATHROW_PIC) += heathrow_pic.o
> 
>     > +softmmu_ss.add(when: 'CONFIG_HEATHROW_PIC', if_true:
>     files('heathrow_pic.c'))
> 
>     this feels pretty clunky. Doesn't Meson have a nicer
>     way of expressing this kind of thing? 

Don't forget we also use:

HELPERS-$(call land,$(CONFIG_SOFTMMU),$(CONFIG_LINUX)) +=
qemu-bridge-helper$(EXESUF)
ifeq ($(CONFIG_LINUX)$(CONFIG_VIRGL)$(CONFIG_GBM)$(CONFIG_TOOLS),yyyy)
HELPERS-y += vhost-user-gpu$(EXESUF)
vhost-user-json-y += contrib/vhost-user-gpu/50-qemu-gpu.json
endif

> 
> 
> Nope, in fact it turns out we had to implement it ourselves because
> there was no way to express it at all in the way that QEMU build
> multiple binaries. My initial version didn't have the when/if_true part
> but the maintainer was quite adamant on requiring them instead of just
> relying on positional arguments.
> 
> Unfortunately, while common in QEMU, and in Linux too, it doesn't seem
> to be a very common idiom in general; Makefiles do express it in a very
> readable and concise way.
> 
> I cannot deny it looks clunky. While Meson does have a huge advantage in
> terms of predictability and lack of surprises when doing harder things,
> the imperative syntax is heavier when dealing with the simplest. They do
> remain simple (it's still one line of coffee), but the quotes and commas
> seem to be enough to "switch" our brain to a different mode.
> 
> My opinion is that you don't really look at these things very much, so
> it doesn't really matter, but I can see people disagreeing.
> 
> (You had already pointed this out during the very first discussions, if
> I remember correctly. I can't add much except that with extra experience
> the balance does tilt in favor of Meson despite the heavier syntax for
> simple cases).
> 
> Paolo
> 
>     It's really
>     common and I'd hope that a build system language
>     would provide less long-winded ways to write common
>     operations...
> 
>     thanks
>     -- PMM
>