Sujet : Re: The joy of FORTRAN
De : antispam (at) *nospam* fricas.org (Waldek Hebisch)
Groupes : alt.folklore.computers comp.os.linux.miscDate : 26. Feb 2025, 15:10:15
Autres entêtes
Organisation : To protect and to server
Message-ID : <vpn7c5$3osdj$1@paganini.bofh.team>
References : 1 2 3 4 5 6 7 8 9 10
User-Agent : tin/2.6.2-20221225 ("Pittyvaich") (Linux/6.1.0-9-amd64 (x86_64))
In alt.folklore.computers rbowman <
bowman@montana.com> wrote:
On Tue, 25 Feb 2025 23:02:32 -0000 (UTC), Lawrence D'Oliveiro wrote:
I did once toy with the idea of, not quite a cross-assembler, but the
use of Python constructs to generate code
<https://bitbucket.org/ldo17/crosscode8/src/master/>. This tried to do
automatic handling of the necessary adjustments of addresses across page
boundaries; I think the algorithm works, though I never tested it very
thoroughly.
This is pretty obscure:
https://docs.micropython.org/en/v1.11/reference/asm_thumb2_hints_tips.html
I am not sure what you find obscure? ARM assembly is a pretty
conventional assembly. Somewhat nasty is limited range of
constants and offsets. Thumb subset is reduced and depending
on exact dialect may lack some instructions which are present
on other machines. Also, Cortex M0 (quite popular target)
theoretically has 15 usable registers, but in practice it is
more like 8 general purpose registers + stack pointer + link
register.
Using assembly from Micro Python leads to somewhat different syntax,
it that bothers you use it normal assembly.
I would say that every IBM mainfaimer should feel at home
using ARM assembly: ARM lacks baroque things like SS instructions,
but core is similar.
-- Waldek Hebisch