RFR: 8170920 SO_RCVBUF and SO_SNDBUF options problem for network channels on MacOS
Langer, Christoph
christoph.langer at sap.com
Mon Dec 12 07:51:48 UTC 2016
Hi Michael,
yes, I understand and agree that it's probably not the right time to do larger refactoring at this time in JDK9. And, yes, the changes for 8167420 and 8167457 have evolved to turn the native Inet*AddressImpl files completely over, so we should therefore defer the full cleanup to JDK10.
However, I still like to get 2 things into the current repo - which were the initial triggers for my changes. One is to remove the duplicate coding for a rather obsolete BSD environment in Inet4AddresImpl and the other is to use getnameinfo/getaddrinfo in Windows Inet4AddressImpl. I'll open new bugs for that and post the reduced changesets for review.
Best regards
Christoph
> -----Original Message-----
> From: Michael McMahon [mailto:michael.x.mcmahon at oracle.com]
> Sent: Freitag, 9. Dezember 2016 17:19
> To: Langer, Christoph <christoph.langer at sap.com>
> Cc: Chris Hegarty <chris.hegarty at oracle.com>; OpenJDK Network Dev list
> <net-dev at openjdk.java.net>
> Subject: Re: RFR: 8170920 SO_RCVBUF and SO_SNDBUF options problem for
> network channels on MacOS
>
> Those suggestions are reasonable Christoph. I think given the subtle
> nature of the bug
> (removing the #include not causing a compile error) I'm a bit wary about
> pushing wide scale
> native code changes at this stage of JDK 9. We might therefore defer the
> other cleanups until 10
> opens up in the new year.
>
> Michael.
>
> On 09/12/2016, 11:16, Langer, Christoph wrote:
> > Hi Michael,
> >
> > the bug then obviously was a side effect of my change
> http://hg.openjdk.java.net/jdk9/dev/jdk/rev/9f0ab4b20ff7 for 8167481. Sorry
> for that.
> >
> > To follow the concept of my cleanups I'd prefer if you could use #if
> defined(MACOSX) rather than #ifdef MACOSX in net_util_md.c to be consistent
> within the file.
> >
> > I also followed the platform order 1. Linux, 2. AIX, 3. Solaris, 4. Mac in all
> parts of the native code, so your newly introduced section should go after the
> "#if defined(__solaris__)" system header include parts, e.g. line 47 of the
> original file.
> >
> > Thanks
> > Christoph
> >
> >> -----Original Message-----
> >> From: net-dev [mailto:net-dev-bounces at openjdk.java.net] On Behalf Of
> Chris
> >> Hegarty
> >> Sent: Freitag, 9. Dezember 2016 11:52
> >> To: Michael McMahon<michael.x.mcmahon at oracle.com>; OpenJDK
> Network
> >> Dev list<net-dev at openjdk.java.net>
> >> Subject: Re: RFR: 8170920 SO_RCVBUF and SO_SNDBUF options problem for
> >> network channels on MacOS
> >>
> >>
> >> On 09/12/16 10:33, Michael McMahon wrote:
> >>> Could I get the following change reviewed please?
> >>>
> >>> http://cr.openjdk.java.net/~michaelm/8170920/webrev.1/
> >> Looks good.
> >>
> >> You may just want to remove the "from" year from the copyright
> >> year range in the test, before pushing.
> >>
> >> -Chris.
More information about the net-dev
mailing list