Regression in ash (crashing)

[Rich Felker]

On Mon, Dec 17, 2012 at 11:29:25AM +0400, Leonid Lisovskiy wrote:
> On Sat, Dec 15, 2012 at 9:23 PM, Rich Felker <dalias at aerifal.cx> wrote:
> > On Sat, Dec 15, 2012 at 12:45:13PM +0300, Leonid Lisovskiy wrote:
> >> Probably we should create bug report with reproducible minimal test-cases?
> >
> > Yes and no. It would be nice to have an example case that shows why
> > the stunalloc is not valid, but I'm not convinced that a test-case is
> > relevant to fixing the bug. The fix for the present bug is jsut
> > revering the commit that added it. That brings back the memory-leak
> > bug, and fixing the memory leak is a more complicated issue, but I
> > don't think the test-case for the present bug is helpful in fixing the
> > memory leak..
> >
> > If you think I'm wrong on this, please explain why.
> 
> The single reason to create bug report with test case is to convince
> busybox maintainers to revert commit 4ba6c5d3ba.

OK. Do you think you can come up with one? I tried writing a naive one
and couldn't get it to crash. It may be better to start with the
crashing configure script and simplify until it no longer crashes...

Rich