ntdll/cdrom : implement CDROM_Verify to work on Mac Os

[Ken Thomases]

On Dec 30, 2010, at 9:36 AM, James McKenzie wrote:

> This has been critically examined by several people and if the solution were this simple it would have been put in place a long time ago.  Charles Davis has spend over a year looking at how to implement CDROM functionality for MacOSX. 

Rather than arguing in sweeping generalizations, can we confine ourselves to the technical specifics of this particular case, please?  When you say "this has been ... examined", are you referring specifically to CDROM_Verify()?

> So, here is my question:  How can the community be 100% certain that:
> 
> 1.  The device has EXACTLY the same name on every available MacOSX installation?

What does the device name have to do with anything?  Do you think that Loïc's patch somehow contains an implicit dependence on device names?  If so, why?

> 2.  How can you be 100% certain that if the 'device' exists that the code will end up in CDROM_Verify()?

Why does that matter?  The question is, under what circumstances does it end up in CDROM_Verify()?  What is the right behavior of CDROM_Verify() and how would that be implemented on Mac OS X?

Can you explain which aspect of CDROM_Verify()'s behavior is not properly implemented by Loïc's patch?

Regards,
Ken