Re: affstools 0.1

[Roman Z. <zi...@fh...>]

Hi,

> Some of the definitions for AFFS_*MAY*() in amigaffs.h are not correct.
> Protection bits for other and group (which were added later in AmigaOS 3.0 for
> Envoy) use inverse logic w.r.t. user. I don't have the AmigaOS includes at
> hand, so I don't dare to supply a patch for this.

Me neither, can someone send me a copy of the relevant parts?

> The same bug is present in the kernel, even in 2.0.38 (the oldest version I
> have here). I'm wondering how it got there, since it once was OK (IIRC, I added
> support for muFS and other/group protection in 1994). Or am I missing
> something?

Was it before or after write support? According to the Changes, that was
somewhere in 1.3 and on the ftp sites I only find 2.0.x and up...

> I also compiled with -Wall and created the patch below. Personally I would use
> __u32 to store 32-bit values (i.e. `LONG' under AmigaOS), and __u16 for 16-bit
> values, to avoid portability problems.

I probably add some more configure tests for that. :)
Thanks for the patch. I'm using -Wall too, but I got no warnings for that 
(using gcc 2.95.2 here).

bye, Roman