Regression on 10/STABLE (Was: SOL_TCP def?)

[Dmitry Sivachenko]

> On 29 Dec 2015, at 23:48, Jonathan Chen <jonc at chen.org.nz> wrote:
> 
> On 30 December 2015 at 07:50, Dmitry Sivachenko <trtrmitya at gmail.com> wrote:
>>> Patch gsoap to use IPPROTO_IP instead of SOL_TCP.
>> I meant IPPROTO_TCP, sorry.
> 
> Thanks for the quick-fix. However, IMHO this should be classed as a
> regression on 10/STABLE.
> 

I think it is non-portable code in gsoap, which was revealed by introduction of TCP_FASTOPEN to 10/stable.