From: Jeff King <peff@peff.net>
Date: Sun, 29 Sep 2013 19:32:45 GMT
Subject: Re: [PATCH 6/9] http: update base URLs when we see redirects
Message-ID: <20130929193245.GB17644@sigill.intra.peff.net>
In-Reply-To: <CAPig+cQVQ6iY8onvb--88W0XHofC815NZHTb=L744hANtT8mVQ@mail.gmail.com>

On Sun, Sep 29, 2013 at 03:26:45PM -0400, Eric Sunshine wrote:

> On Sat, Sep 28, 2013 at 4:34 AM, Jeff King <peff@peff.net> wrote:
> > diff --git a/http.c b/http.c
> > index 65a0048..8775b5c 100644
> > --- a/http.c
> > +++ b/http.c
> > @@ -921,11 +921,71 @@ static int http_request_reauth(const char *url,
> > +static int update_url_from_redirect(struct strbuf *base,
> > +                                   const char *asked,
> > +                                   const struct strbuf *got)
> > +{
> > +       const char *tail;
> > +       size_t tail_len;
> > +
> > +       if (!strcmp(asked, got->buf))
> > +               return 0;
> > +
> > +       if (strncmp(asked, base->buf, base->len))
> > +               die("BUG: update_url_from_redirect: %s is not a superset of %s",
> > +                   asked, base->buf);
> 
> Is there something non-obvious going on here? die(...,base->buf) takes
> advantage of the terminating NUL promised by strbuf, but then
> strncmp(...,base->buf,base->len) is used rather than the simpler
> strcmp(...,base->buf).

Yes, we are not checking for equality, but rather making sure that
"asked" begins with "base->buf". It might be more clearly written as:

  if (prefixcmp(asked, base->buf))

I was trying to take advantage of the fact that we know base->len
already, but this it not a particularly performance-critical code path.
We can afford the extra strlen that prefixcmp will do.

-Peff