The first artifact is in the build_stdcxx script.  It's the conditional 
that's being used to determine if we're building with cygwin or not.  I 
need to do some more work in this area anyways, so that we can get 
cygwin builds working.

--Andrew Black

Martin Sebor wrote:
> There are a couple of minor issues in our gcc builds (both Linux and
> Solaris). AFAICT, neither of them appears to be causing any build
> failures but they seem more than cosmetic.
> 
> First, we appear to be using the wrong option(s) in the build_stdcxx
> script:
> 
> -- Setting up Solaris GNU binutils 2.17 --uname: illegal option -- o
> usage:    uname [-snrvmapiX]
>     uname [-S system_name]
> ./build_stdcxx: line 22: ((: Cygwin == : syntax error: operand expected 
> (error token is " ")
> 
> Second, we get the error below from the first invocation of make:
> 
> ### gmake builddir BUILDMODE=pthreads,shared,optimized, CONFIG=gcc.config
> /bin/sh: cho: command not found
> rm: cannot remove `/tmp/longsize': Operation not permitted
> 
> I suspect the second problem is fallout from my change 523520 (see
> http://svn.apache.org/viewvc?view=rev&revision=523520) and I plan
> to work on fixing it.
> 
> I'm not sure what's behind the first one. Andrew, can you take a look
> at it when you have a chance?
> 
> Martin