[quagga-dev 11496] Re: patch: require IPv6 API built-time support
paul at jakma.org
Fri Sep 12 13:47:35 BST 2014
David tells me he has other cleanups this clashes with. So those should go
in first of course.
Indeed, this kind of patch is generally going to cause conflicts for
people. It's an easy patch to rebase though, so it should go 'after'. Hard
to pick a good 'after' though! :)
On Fri, 12 Sep 2014, Vincent JARDIN wrote:
> Yes these ifdef have to be removed.
> Le 12 sept. 2014 13:07, "Paul Jakma" <paul at jakma.org> a écrit :
>> Yeah, the below is purely about built-time requirement for IPv6 APIs.
>> It doesn't mean the system must have working IPv6 networks - just that the
>> more modern APIs and code are present in the system.
>> On Fri, 12 Sep 2014, David Lamparter wrote:
>> This would be the time for anyone using --disable-ipv6 to speak up... if
>>> there are use cases / people need this, I guess we can keep it around,
>>> but that really needs to be based on good reasoning.
>>> In particular, "I don't use IPv6" isn't enough, since (a) you really
>>> should be using IPv6 these days, and (b) your OS probably supports it
>>> anyway, and the probability for bugs is higher with the switcheroo than
>>> it is with having IPv6 supported but dormant.
>>> P.S.: OS support refers to the C library, not to someone switching off
>>> knobs in their kernel for random reasons... the latter is no trouble at
>>> all, if you have no IPv6 in your kernel at runtime that's something we
>>> need to/should support. If *that* doesn't work, that would be a bug to
>>> be fixed.
>>> On Fri, Sep 12, 2014 at 11:44:15AM +0100, Paul Jakma wrote:
>>>> Sifting through old stuff. Below is something I had that I've rebased.
>>>> Only a semi-serious proposal to merge it. One of those "be nice to do,
>>>> is it worth the huge amount of churn?" things.
Paul Jakma paul at jakma.org @pjakma Key ID: 64A2FF6A
Moebius always does it on the same side.
More information about the Quagga-dev