On Tue, Oct 28, 2008 at 02:32:48PM +0100, Justus Winter wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Justus Winter wrote:
Jorge Arellano Cid wrote:
Not yet committed, mainly because the feature freeze for release. (although the small fix went in).
The attached patch fixes this and also adds the function
void a_DilloHost_to_String(DilloHost *host, char *dst, size_t size) If you finally get to a complete fix, please name exported functions in dns.c with the "a_Dns_" prefix. e.g. a_Dns_host_to_string Sure thing. Updated patch is attached. *bump*
While tracking my ipv6 related problems through dillos source files I noticed *a lot* of ENABLE_IPV6 #ifdefs... Is this necessary for portability or do you want to be able to strip ipv6 support from certain builds?
If portability at compile time is not the issue I would vote for the removal of those #ifdefs, if there's worry about the executables size, let's see:
$ wc --bytes with{,out}/dillo2/src/dillo-fltk 3311349 with/dillo2/src/dillo-fltk 3309974 without/dillo2/src/dillo-fltk
Well, IPv6 costs 1375 bytes on my system, I don't see any problem here, but I think this could be reduced. Many of those ipv6 related #ifdefs duplicate quite a lot of code. One tiny little cleanup patch is attached (please note that it depends on the a_Dns_DilloHost_to_String function). Any thoughts on this one?
Good you reminded me... IIRC the ENABLE_IPV6 stuff was because resolving was slower for IPV4 with the IPV6 code. I don't know whether this is still the case. Maybe portability is also an issue. Please send me two patches. One with a_Dns_dillohost_to_string() and another removing the ifdefs. If we get no comments on this thread, a commit experiment may take place. -- Cheers Jorge.-