Re: Declaring variables

[Kim F. Storm]

Richard Stallman <address@hidden> writes:

> Many byte compiler warnings are for reference to a global
> variable that has no visible definition.  It would be nice
> to get rid of them.  The way to do that with existing facilities
> is to write
> 
>   (defvar variable)
> 
> in the file.  But that has disadvantages, because some facilities will
> treat it as a real definition of the variable.  It would be good to
> have a clean solution that would have no disadvantages.  I recall
> there was a discussion of this about a year ago, but I don't recall
> what conclusion (if any) we reached.

I'm using the following to avoid warnings for dynamic variables:

        (eval-when-compile
          (defvar ...)
          (defvar ...)
        )

I don't know whether etags will ignore things in eval-when-compile,
but I guess it could do that.

++kfs