RFR (S) JDK-8160174: java.net.NetworkInterface - fixes and improvements for network interface listing

[Langer, Christoph]

Chris,

thanks for your comments. I'm kind of thinking the same - I actually also don't really like my result of merging the strings of code so much.

I'll go on and prepare a smaller change set only addressing the real bugs.

I also agree to switching the Linux implementation to getifaddrs at the time the Java 10 branch opens to gain experience with that without shipment pressure.

Stay tuned :)
Christoph


> -----Original Message-----
> From: Chris Hegarty [mailto:chris.hegarty at oracle.com]
> Sent: Dienstag, 12. Juli 2016 16:10
> To: Langer, Christoph <christoph.langer at sap.com>
> Cc: net-dev at openjdk.java.net
> Subject: Re: RFR (S) JDK-8160174: java.net.NetworkInterface - fixes and
> improvements for network interface listing
> 
> Christoph,
> 
> > On 11 Jul 2016, at 14:36, Langer, Christoph <christoph.langer at sap.com>
> wrote:
> >
> > Hi Chris (or anyone who is holding a stake in the NetworkInterface native
> implementation),
> >
> > I’ve spent some time on cleaning up NetworkInterface.c and came up with a
> bigger change:http://cr.openjdk.java.net/~clanger/webrevs/8160174.2/
> >
> > With this I attempted to consolidate the interface listing functionality of the 2
> main categories – one is ioctl (used on Linux IPv4, Solaris and AIX) and the
> other is getifaddrs (used on MacOS). I introduced some defines to switch
> between the implementations. I also consolidated the functionality for the ioctl
> based network interface listings by using an #ifdef section to distinguish
> between the Linux/AIX versus Solaris field and constant names.
> >
> > I’ve spent some time testing on the platforms and in principal it works. But as
> it is a matter of taste, I’d like to ask you if you support this type of change or
> have any hints or recommendations before going forward to finalize this.
> 
> I’m personally don’t like this approach much. I think it adds further
> complexity and difficulty to this code, which already has its fair share
> of both.
> 
> > For Linux I’m also suggesting to use the getifaddrs approach – I tested and
> found it working everywhere and with this we could get rid of the
> implementation to read /proc/net for IPv6.
> 
> You should probably break this type of change out, so that it can be
> evaluated independently, on its own merit. If it add nothing more than
> clean up, may it makes sense for this to happen early in 10, where it
> can have a longer time to bake.
> 
> > Furthermore I’m generally setting Null for the IPv6 broadcast address – which
> I think is common sense for IPv6.
> 
> Same as above.  This “smaller” changes can sometimes get lost in
> the noise of larger refactoring.
> 
> -Chris.
> 
> 
> > Thanks in advance and best regards
> > Christoph
> >
> >
> > From: Langer, Christoph
> > Sent: Donnerstag, 23. Juni 2016 16:37
> > To: 'net-dev at openjdk.java.net' <net-dev at openjdk.java.net>
> > Subject: RFR (S) JDK-8160174: java.net.NetworkInterface - fixes and
> improvements for network interface listing
> >
> > Hi,
> >
> > can I please get a review the following change:
> > Webrev: http://cr.openjdk.java.net/~clanger/webrevs/8160174.1/
> > Bug: https://bugs.openjdk.java.net/browse/JDK-8160174
> >
> > I made further fixes and cleanups for listing Unix type network interfaces,
> especially on AIX and Linux. I ran builds and verified also on Solaris and Mac.
> >
> > Thanks
> > Christoph