[PATCH v2 2/2] Cygwin: uname: add host machine tag to sysname.

[Corinna Vinschen]

On Nov 26 14:27, Jeremy Drake via Cygwin-patches wrote:
> On Tue, 26 Nov 2024, Corinna Vinschen wrote:
> 
> > On Nov 25 11:24, Jeremy Drake via Cygwin-patches wrote:
> > > +      switch (wincap.host_machine ())
> > > +	{
> > > +	  case IMAGE_FILE_MACHINE_AMD64:
> > > +	    n = stpcpy (buf, "-x64") - buf;
> > > +	    break;
> > > +	  case IMAGE_FILE_MACHINE_ARM64:
> > > +	    n = stpcpy (buf, "-ARM64") - buf;
> > > +	    break;
> > > +	  default:
> > > +	    n = __small_sprintf (buf, "-%04y", (int) wincap.host_machine ());
> > > +	    break;
> > > +	}
> >
> 
> > You can greatly simplify this switch.  We don't support 32 bit systems
> > and we will never again support 32 bit systems.  Any combination
> > including a 32 bit system can just go away.  Theoretically, only
> > the IMAGE_FILE_MACHINE_ARM64 case should be left.
> 
> Is the above edit enough, or do I need to remove the x64 case entirely as
> well?  If you truly just want to handle ARM64, this whole function could
> be inlined into just an
> 
>   if (wincap.host_machine () != wincap.cygwin_machine ()
>       && wincap.host_machine () == IMAGE_FILE_MACHINE_ARM64)
>     strcat (name->sysname, "-ARM64");
> 
> For a little more future-proofing, nested if/else so non-ARM64 gets the
> hex fallback from the default case above.

I'm not opposed to a switch statement consisting of an
IMAGE_FILE_MACHINE_ARM64 case and a default case adding "-???" or
something.  Chances are so extremly slim that we'll ever see another
CPU emulated on x86_64, we can always add a case for that if it turns
out that I'm totally wrong, right?


Thanks,
Corinna