Re: Severe performance problem and proposed solution

[Zack Weinberg]

On Mon, Feb 25, 2002 at 12:44:32PM +0100, Akim Demaille wrote:
> >>>>> "Zack" == Zack Weinberg <address@hidden> writes:
> 
> Zack> Richard Henderson and I experimented with using echo instead of
> Zack> here documents[1].  The trouble with this is of course that echo
> Zack> may interpret backslashes and leading -, so both have to be
> Zack> escaped.  We got a two-page gobbet of M4 that sort of worked,
> Zack> but couldn't help at all with AC_DEFINE_UNQUOTED and the like.
> Zack> I think this approach is a dead end.
> 
> Hi Zack,
> 
> We are aware of this issue, and this is one of the reason why we now
> have configure re-exec'ing itself with a better shell, with `better'
> being defined as `supports $LINENO'.
> 
> If this is enough, then we're good.  But we need someone to test it.

It may well be good enough.  I cannot test it because (a) I have no
access to any of the affected machines, and (b) for other reasons,
libstdc++ cannot currently migrate away from autoconf 2.13.  (See the
discussion at
<http://gcc.gnu.org/ml/gcc-patches/2001-05/msg01986.html>; suggestions
would be appreciated.)

If you can show me a simple test script which determines if $LINENO
works, I can ask the people experiencing the problem to tell me what
the problem shells make of it.

zw