Re: [hatari-devel] natfeats installing __NF cookie

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 13.7.2022 0.50, Miro Kropáček wrote:
> I've noticed that while EmuTOS properly detects Natfeats and installs the
> cookie properly, on TOS, naturally, Natfeats is not detected and therefore
> cookie not installed.
>
> Shouldn't Hatari patch the TOS image and install the cookie, same as Aranym
> does?

Aranym may be patching 4.04 TOS version and supporting that in addition 
to EmuTOS, but Hatari supports all TOS versions.

It patches 1.00, 1.02, 1.04, 1.06, 1.62, 2.05, 2.06, 2.07, 2.08, 3.06, 
4.00, 4.01, 4.02 and 4.04 i.e. 14 different TOS versions.  I do not 
think *any* cookies to be added for any of them by Hatari.


> Otherwise this commit:
> https://github.com/freemint/freemint/commit/3661519b9cfbb0fc0498e75343ca734ac42a0986
> basically doesn't work properly with Hatari+TOS because it assumes EmuTOS
> and its cookie jar.

Hatari examples for NatFeats usage do not use "__NF" cookie:
https://git.tuxfamily.org/hatari/hatari.git/tree/tests/natfeats


I see "__NF" cookie mentioned in the original Aranym NF proposal from 2002:
https://github.com/aranym/aranym/wiki/natfeats-proposal

But even that has this exception:
"68k code that do not intend to (ever) run on unsupported processors is 
not obliged to check for this cookie ..."

And in 2009 there was e.g. this Aranym News item:
---------------------
2009/04/25 - version 0.9.8beta released
....
 o fixed several NF drivers to work under FreeMiNT (not rely on __NF 
cookie)
---------------------

Maybe MiNT itself should not rely on "__NF" cookie either?

Or maybe MiNT should be adding it for regular TOS, (when NF is available 
and support for that is enabled in MiNT)?


	- Eero

PS. I assume you're not volunteering to provide tested Hatari support 
for patching "__NF" cookie to 14 TOS versions. :-)