Re: the fate of Vala in OE

[Carlos Rafael Giani <dv@pseudoterminal.org>]

On 2016-06-09 12:28, Alexander Kanavin wrote:
> Hello,
>
> I'd like to propose the removal of Vala recipe and support for 
> generating Vala bindings from oe-core. The only software that is 
> written in Vala and makes sense in embedded environment is Rygel (an 
> implementation of various UPnP AV services), and even that can be 
> compiled from intermediate C sources without a need for Vala compiler 
> or bindings. The only benefit from compiling from Vala source would be 
> that Rygel could be more easily patched - and I think that doesn't 
> really justify maintaining the full Vala stack.
>
> Thoughts?
>
> Alex

The benefit of easy patching is considerable. If you only have the C 
sources, then the patches have to be done between versions of that 
generated C code. The resulting patches are often big and unreadable.

I was actually happy to see g-i added to OE-core so we can get rid of 
some custom C-source patches for Rygel and just patch the Vala code 
directly instead. Removing vala would nullify this benefit.