Re: why is replace.mk complaining about gsed?

[John Marino <marino%netbsd.org@localhost>]

On 12/10/2011 12:46 AM, David Holland wrote:
> On Fri, Dec 09, 2011 at 10:50:01PM +0100, John Marino wrote:
>   >  >>That depends on what is considered the bug.  I consider comparing
>   >  >>string variables
>   >  >>without quotes a dodgy decision at best.  One could consider these
>   >  >>string comparisons
>   >  >>without double quotes to be a bug.
>   >  >You are wrong. If the variables are allowed to be undefined, the correct
>   >  >approach would be to use :U. Adding random "" is not what make fragments
>   >  >should use.
>   >
>   >  Fine.  Would some makefile expert please fix replace.mk in the
>   >  sanctioned way?
>
> I think the point is that it this combination of defined/undefined
> shouldn't be happening, so papering it over so it doesn't fail at this
> point is only asking for something more mysterious to go wrong.
I understand the point.
The fact is that it is happening, and it's breaking packages.
An obvious possibility is that the original author screwed up by 
thinking defined/undefined
couldn't happen when it legitimately can.
To iterate:
1) It's wrong to assume that replace.mk is correct and defined/undefined 
combo is wrong
2) Even if defined/undefined indicates a bug, it is a harmless one and 
preferable to a
broken package.

The code needs beefing up.  Something like "bmake -V '${DEPENDS}'" 
should never result in makefile
error messages.