Re: Clang/LLVM migration roadmap

[Riccardo Mottola]

Hi,

On 2022-02-10 02:31:44 +0000 Gregory Casamento 
<greg.casamento@gmail.com> wrote:

> Riccardo,
>
> I don't believe that GNUstep should hold back features to remain 
> compatible
> with any given compiler.  Not implementing features that are widely 
> used
> (not even particularly "modern" ones) because the less capable 
> compiler (in
> this case the latest GCC) lacks support is not a healthy direction.

We did not discuss here specific features, we are discussing of 
switching over.
Which features are we holding back?
Not ARC - not for the end user, only for a contributor. So it is not 
specifically a "feature of gnustep" .

There is a lot - perhaps seeked - confusion about these "features" and 
where they need to be used and exposed.

> Like you, I believe in choice.  I think that GNUstep should remain
> compatible with GCC so long as GCC is able to keep up with the 
> features
> needed by the project.

I don't belive ARC or properties are a feature needed by the project. 
They are conveniences which we could use, we can not.
They might be features needed by users relying on our libraries and 
our APIs.
But are we really holding back those? I'm note aware of that from my 
understading, but if, we should discuss that ans even more important 
than the given long-term migration.

I do not reply to your other points in discussion because it was a 1:1 
opinion, no reason to continue, or you actually misinterpreted my 
point and an email discussion might only lead to flames.
But one thing for sure I do not share: coung architectures just by the 
number of users. If we were taling about market share, we should ditch 
everything and code for Windows and perhaps GTK on Linux. All the rest 
is minority. We are by definition a niche project and for that we need 
to differentiate us. We are not Cocotron.


Cheers,
Riccardo

--
Proudly sent with GNUMail on GNUstep on FreeBSD/x86-64