On Wed, Jan 12, 2011 at 04:33:02AM +0100, Christof Mroz wrote: > On Tue, 11 Jan 2011 09:52:57 +0100, Stefan Götz > <stefan.goetz@xxxxxxxxxxxxxxxxx> wrote: > >> Hm, but doesn't that mean that this branch is not ready for a merge >> proposal, anyway? > > I'd rather mark it as "ready to test" and "I'm done looking for policy > violations" (believe it or not :), maybe a branch proposal was too heavy > an abuse. I'll just write a mail to the list next time, instead. Until > re-synchro with trunk, none of the above apply of course. Please delete your branch and use mine as the basis for further work, it incorporates a bunch of fixes and is synchronized to a more recent version of HIPL. >> Exactly. Precisely *because* the 'data' member is of type void*, you >> can assign its value to the variable 'conn' without a cast. > > Aight, didn't know that. (Fun fact: in C++ it's mandatory to cast from > void* to something, but not the other way around. Of course, new/delete > and templates are preferred in this case, so it's actually a useful > reminder that you're doing something wrong). Do not confuse C with C++ :) >>>>> === modified file 'hipd/hiprelay.c' >>>>> --- hipd/hiprelay.c 2010-11-30 14:50:30 +0000 >>>>> +++ hipd/hiprelay.c 2010-12-13 21:28:35 +0000 >>>>> @@ -1015,16 +1015,16 @@ >>>>> * >>>>> * @param r the HIP control message to be relayed >>>>> * @param type_hdr message type >>>>> - * @param r_saddr the original source address >>>>> - * @param r_daddr the original destination address >>>>> + * @param r_saddr (unused) the original source address >>>>> + * @param r_daddr (unused) the original destination address >>>> >>>> [M] what is the point of adding unused function arguments to an already >>>> crowded signature? >>> >>> For debug purposes. Touching the function body would be out of the >>> scope of this >>> branch (even this doc change was, sorry). >> >> I'm not quite following. So if they are necessary for debug purposes, >> why are they unused? Aren't they 'used for debugging'? > > Just noting that while these parameters seem important, they actually > have no impact on the function. Removing them sounds like a good idea then :) Diego