Re: why is re-defining register names specially treated in gas

[Nick Clifton <nickc at redhat dot com>]

Hi Jan,

(Sorry for the long delay in replying to this email)

> In assign_symbol(), the error for re-defining an already defined symbol
> is suppressed when the symbol is in reg_section. Can anyone explain why
> this has to be that way? It would seem to me that this is OK for 'real'
> register names (i.e. those produced by target specific code) 

I believe that this is the reason why it is permitted.

> but not for
> those resulting from equating custom symbols to register names (which
> currently appear to be indistinguishable).

Which I think must be considered to be a mis-feature.  Ideally we need 
some way to distinguish between these two types of register name symbols.

Cheers
  Nick