On Fri, 3 Jun 2005, Hajimu UMEMOTO wrote:
> Hi,
>
> >>> On Wed, 01 Jun 2005 21:44:03 +0900
> >>> Hajimu UMEMOTO <ume@mahoroba.org> said:
>
> ume> I've attached the patch to correct 1st argument of getnetbyaddr(3) in
> ume> this mail. It is subset of my previous patch. Since it breaks ABI
> ume> compatibility of getnetbyaddr(3), I think it is better to correct
> ume> n_net member of struct netent, too. Since there is objection, the
> ume> patch leaves struct addrinfo as is. So, it doesn't need to bump any
> ume> shlib major. Is it okay?
>
> Ultimately, I wish to correct struct addrinfo, too. Since correcting
> getnetbyaddr(3) breaks ABI compatibility after all, it seems storange
> to me to leave struct addrinfo alone as is. It is better to take this
> occasion to correct struct addrinfo as well.
> This breakage is only on 64 bit arch. The influence will grow as 64
> bit arch spreads. So, I believe it should be done as soon as
> possible.
Just leave it alone for now. When symbol versioning comes, you
should be able to remove the padding without bumping library versions
and producing imcompatibilities.
--
DE