Re: CONFIG_KPROBES=y build requires gawk

[Rob Landley <rob@landley.net>]

On Wednesday 16 December 2009 20:43:01 H. Peter Anvin wrote:
> On 12/16/2009 05:39 PM, Roland Dreier wrote:
> > Is there any reason not to apply the patch below, to allow more awk
> > implementations to be used?  After all, it's not like we're going to put
> > non-ASCII characters into the map file...
>
> I guess the question is if it will break under any other circumstances,
> but I guess we can find those when we get to them.
>
> There was a long discussion about the use of awk on IRC today.
> Apparently mawk, in particular, is actively broken, because the
> maintainer believe that POSIX is crap.  There are quite a few issues
> with it, according to reports.

if the kernel specifies posix, and that implementation doesn't do posix, then 
that implementation doesn't build the kernel.  Blacklisting known broken 
implementations makes a certain amount of sense.

> We need a sane scripting language available to the kernel build, and
> given all the problems we have had with different versions or even just
> sometimes different builds of sh, awk, and even bc -- plus the fact that
> those utilities just don't necessarily do what we want makes it very
> frustrating.

1) Posix exists for a reason.

2) Busybox implements what the kernel has needed to build.  (I test this every 
release, and I fix it where necessary.)

> Personally I think a dependency on Perl is better than the
> mess we're in; I understand other people disagree.

Vehemently.

> What is definitely
> not acceptable, however, is the status quo.  The situation is, quite
> frankly, ridiculous enough that perhaps the right thing to do is to
> write a small scripting engine and bundle it with the kernel.  Something
> that does what we need it to do, but is only one implementation and
> something we can extend at will if need be.

*shrug*  That's one way to avoid environmental dependencies.

> 	-hpa

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds