On 26.02.2014 11:48, Mladen Turk wrote:
> On 02/26/2014 11:18 AM, Rainer Jung wrote:
>>
>> No other finding currently, but still checking some stuff. Just wanted
>> to let you know quickly.
>>
> 
> OK. Give it a try. I'll just retag 1.2.39 with all the "polishing" stuff
> found. It's been a while since last release, so I was prepared for few
> runs :)

NP.

A few polishing items and questions:

You might want to add something about your IPV6 improvements to the
change log. There's the prefer flag and also the possibility to use IPv6
addresses in the config, right? More?

configure.in now checks for netdb.h and netinet/in.h but I couldn't find
any use of those in the final code (except for old use in Netware case
which doesn't rely on configure detection). Should we remove the new
detection code for those two headers?

native/nt_service/jk_nt_service.c still uses jk_resolve() with
sockaddr_in instead of tne new jk struct. Should we remove
native/nt_service or apply the probably simple fix to make it compatible
with the new jk structs? I'd say we can drop it but should add it to
changes and news. I don't remember anyone having askes about it for many
years and I'm not even convinced it worked in 1.2.37.

JK_INET / JK_INET6: it looks a bit confusing to me.
- JK_INET is set to APR_INET in the APR case
- JK_INET6 is always set to AF_INET6 in the IPv6 case,
  even when APR is used
- jk_connect.c uses JK_INET / JK_INET6 in the non-APR case,
  but a mixture pf APR_INET / APR_INET6 and JK_INET / JK_INET6
  in the APR case. I didn't find a functional problem due to
  the JK and APR stuff at the end being the same,
  but it looks confusing. Shouldn't we replace all APR_INET and
  APR_INET6 by JK_INET resp. JK_INET6 in teh code and set JK_INET6
  similar to JK_INET depending on APR (=APR_INET6) or non-APR
  (=AF_INET6)?

There are two places in common/jk_connect.c which use APR_HAVE_IPV6 and
not JK_HAVE_IPV6. is this intentional?

Regards,

Rainer

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org

Reply via email to