From: Jeff King <peff@peff.net>
Date: Wed, 28 Mar 2012 19:40:45 GMT
Subject: Re: [PATCH 2/2] git: continue alias lookup on EACCES errors
Message-ID: <20120328194045.GC29019@sigill.intra.peff.net>
In-Reply-To: <7v62dowpdu.fsf@alter.siamese.dyndns.org>

On Wed, Mar 28, 2012 at 11:29:17AM -0700, Junio C Hamano wrote:

> > That sounds sensible to me. I think it involves writing our own
> > execvp, though, right? If we use stock execvp, we can't tell the
> > difference between the two cases.
> 
> The stock exec*p() will not hit "/bin/ls" in either case, so we will give
> "'ls' is not a git command", without having to differenciate it.  That is
> what I meant by "we follow the usual rule to ignore it".
> 
> We already have the code necessary to enumerate the possible commands from
> components of the PATH in order to give suggestion, so we can run it
> after seeing exec*p() failure to see if we did not see any "ls", or we saw
> "ls" but it was not executable.  No need to penalize the normal case, no?

Yes, we can differentiate after the fact. Though I think it ends up
being almost the same code as just implementing execvp in the first
place.

-Peff