Re: why is replace.mk complaining about gsed?

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

On 12/9/2011 11:58 AM, OBATA Akio wrote:
> It is bad idea to just hide the bug.
>
> Please let me know how to reproduce the problem.
> (I have no problem for editors/emacs20 on DragonFly-i386-2.10.1)

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.

Clearly the original author thinks that it's impossible for 
_USE_TOOLS.<class> to
be defined and _TOOLS_DEPMETHOD.<class> to be undefined is something 
that can't
happen, but what if it *can* happen?  We've now seen it on two unrelated 
packages.
One with flex that used flex and one with gsed that does not use gsed.

Back to the code in question: Adding quotes around the 
_TOOLS_DEPMETHOD.<class>
will not cause the wrong thing to happen (iow _TOOLS_DEPMETHOD.<class> 
will not
get redefined accidentally).

I would take a package that builds correctly and "hides a bug" all day 
long if
the alternative is a broken package.  That's the case here.  If there is 
no technical
reason why double quotes should be missing from these conditions of 
replace.mk,
then they should be added in.