[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Building Gforth for Apple's new Arm-based Macs
From: |
Craig Treleaven |
Subject: |
Re: Building Gforth for Apple's new Arm-based Macs |
Date: |
Wed, 20 Jan 2021 20:21:52 -0500 |
> On Jan 20, 2021, at 11:59 AM, Alexander Shpilkin <ashpilkin@gmail.com> wrote:
>
> it appears to me that you misunderstand what these settings are doing.
> (Sorry if I’m guessing wrong.) Both build and host are _triples_, that
> is, of the form CPU-VENDOR-OS (or sometimes CPU-VENDOR-KERNEL-LIBC),
> and anything else you can pass in the place of a triple is a convenient
> abbreviation (expanded by code in `config.sub`, which, unlike the
> Lovecraftian horror that is the generated `configure`, is completely
> readable). In particular, it is a convenient abbreviation that does
> not depend on the machine you’re using (i.e. `config.sub` does not
> invoke or otherwise touch `config.guess`).
>
Hi Alex:
I was aware that I was supposed to feed a triple to the -build flag.
Pragmatically, I noticed configure didn’t complain and that the Gforth built
and passed tests when supplied only with x86_64. Since what I’m doing is
essentially a hack, I didn’t feel too bad about it!
It looks like maybe it is time for a maintenance release of Gforth. Hopefully,
in that case, the need for such hacks will go away.
Thanks,
Craig