Re: _AS_DETECT_BETTER_SHELL speedup

[Stepan Kasal]

Hello Paul,

OK, I back off my one line patch.  But:

On Thu, Jan 27, 2005 at 09:51:57PM -0800, Paul Eggert wrote:
> Ralf Wildenhues <address@hidden> writes:
> > And on windows, for example, every fork not done is good!
> 
> Minor efficiencies like this are not sufficient justification,

indeed, the speedup is not minor for Windows ports.
For example, on Cygwin, each fork can take ages.

> if there is a real risk of breaking things elsewhere.

I don't think there is a risk of actually _breaking_ the configure
script.

Nothing worse can happen than a few spurious 'not found' messages.

> > I do not understand this reasoning.  AFAIK Bourne shells fork with
> > redirected { }, newer shells don't.

So I propose the following:
Ralf volunteered to produce a complete patch; let him do it.
Let's apply the patch; if a problem with unexpected 'not found' messages
appears, let's revert that patch.

I think the risk of cosmetic problems is worth the speedup on platforms
like Cygwin.

Paul, please give your nod, and Ralf will produce a patch.

Have a nice day,
        Stepan