On Mon, Nov 06, 2006 at 02:18:06PM -0800, Krzysztof Kowalczyk wrote:
> Yes, but USE_LIBIDN is set if HAVE_LIBIDN is set and tld.h is present.
> See lib\setup.h:
> #if defined(HAVE_LIBIDN) && defined(HAVE_TLD_H)
> /* The lib was present and the tld.h header (which is missing in libidn
> 0.3.X
> but we only work with libidn 0.4.1 or later) */
> #define USE_LIBIDN
> #endif

Good point.

> Also, the question, in your case, is if idn support is really
> disabled. If --wihout-libidn doesn't disable idn support but libidn.so

It is, as curl --version attests. This is a better test than 'ldd' as
the latter won't show a statically-linked libidn.

> is present on your system, then you won't see the problem. It only
> manifests itself if libidn.so is not there. One way to verify whether
> libcurl.so links to libidn.so is to ldd libcurl.so. Or just check
> config.h and see if HAVE_LIBIDN is defined after configure step.
>
> It's also possible that on some systems tld.h is not present, which
> would also disable idn support. I should have mentioned that I'm using
> Ubuntu 6.10 that does have tld.h

I've tried it with and without libidn.so, and with and without
--without-libidn, and it still works fine. However, configure will still
pick up a static libidn.a if libidn.so is gone.

> I still think there's a bug in ./configure that #defines HAVE_LIBIDN
> even if --without-libidn is given, but I can't track it down or fix it
> (at least in the near future) because all I know about auto* tools is
> what they're supposed to do but not how they work.

The configure test is a bit confusing. For one thing, it outputs
"whether to build with libidn...yes" whether or not it actually found libidn
(the IDN support: line at the end of configure is correct). The IDN_ENABLED
macro doesn't seem to be used anywhere. But, I couldn't find any
true error in the test.

>>> Dan

-- 
http://www.MoveAnnouncer.com              The web change of address service
          Let webmasters know that your web site has moved