IPV6_V6ONLY not recognized on Windows

[CetinSert]

http://msdn.microsoft.com/en-us/library/windows/desktop/ms738574(v=vs.85).aspx claims a same named option is actually exported on Windows but the configure script of the network package fails to recognize this and defines HAVE_DECL_IPV6_V6ONLY 0.

[hreinhardt]

This seems to be a bug in MingW, which doesn't define the IPV6_V6ONLY constant. Apparently there had been a patch submitted to fix this in 2008, but it was rejected due to licensing reasons: http://sourceforge.net/tracker/?func=detail&atid=302435&aid=2035902&group_id=2435

I guess we could try to build a workaround into the network library if this is an important issue to you,

[CetinSert]

(Very high quality answer. I was expecting I would first be asked to post all sorts of intricate details and excerpts before someone got to say anything on the matter.)

As I see no other way on Windows to have a single listening socket accept both IPv4 and IPv6 connections, I do believe it is a very important issue for me. If there is any other way to accept both IPv4 and IPv6 connections at the same port, I might use it as a temporary solution waiting for the workaround in the network library.

[hreinhardt]

I've created and pushed a patch to my repo: https://github.com/hreinhardt/network

To test it you have to:

• clone my repository
• switch to the 'stable' branch
• run 'autoreconf' (which is included with MSys/MingW)
• run 'cabal install'

[CetinSert]

I just had time to test this and can confirm that it is working and leads to the expected behavior of being able to handle both IPv4 and IPv6 using a single socket.

What is the chance this ends up in network-2.3.0.12 or above? Is there anything I can do on my end to make sure that it does? How did you know the socket option value of IPV6_V6ONLY to be 27?

[hreinhardt]

I've created a pull request so if Johan is happy with the patch, he will probably merge it in a few days.

(I got the socket option number 27 from the SourceForge page I linked to above)

[tibbe]

Fixed in d646278 and f68a808. Released as v.2.3.0.12

[CetinSert]

2.3.0.12 fails to build on Windows due to catchIOError not being in scope:

checking whether IPV6_V6ONLY is declared... yes
checking for sendfile in sys/sendfile.h... no
checking for sendfile in sys/socket.h... no
checking for gethostent... no
configure: creating ./config.status
config.status: creating network.buildinfo
config.status: creating include/HsNetworkConfig.h
Preprocessing library network-2.3.0.12...
Preprocessing test suites for network-2.3.0.12...
Building network-2.3.0.12...
[1 of 8] Compiling Network.URI      ( Network\URI.hs, dist\build\Network\URI.o )
[2 of 8] Compiling Network.Socket.ByteString.Internal ( Network\Socket\ByteString\Internal.hs, dist\build\Network\Socket\ByteString\Internal.o )
[3 of 8] Compiling Network.Socket.Internal ( dist\build\Network\Socket\Internal.hs, dist\build\Network\Socket\Internal.o )

Network\Socket\Internal.hsc:81:1:
    Warning: Module `GHC.IOBase' is deprecated: use GHC.IO instead
[4 of 8] Compiling Network.Socket   ( dist\build\Network\Socket.hs, dist\build\Network\Socket.o )

Network\Socket.hsc:431:13: Not in scope: `catchIOError'
cabal: Error: some packages failed to install:
network-2.3.0.12 failed during the building phase. The exception was:
ExitFailure 1

[tibbe]

Gah! We need a Windows build bot! Anyone feel like setting one up? In the mean time I'll fix this.

[tibbe]

@CetinSert Fixed on the stable branch. If you test it I'll make another release.

[CetinSert]

@tibbe I already have a Windows server I use for testing software. I am also meaning to set up build bots for Haskell projects targeting Windows, Linux and FreeBSD. Is there any build automation software already used by others on other operating systems?

[tibbe]

@CetinSert I build all my packages on the latest 3 versions of GHC on my Linux buildbot at http://ci.johantibell.com/ using Jenkins CI.

[CetinSert]

@tibbe Your fix 87c3baa came a bit too hasty, I was analyzing a possible fix here: #34.

What needs to be done now:

1. Revert 87c3baa: System.IO.Error was already imported wholly further up the same file.
2. Rename catchIOError to catch: catchIOError is only availabe in base >= 4.4.0; using catch satisfies the whole dependency range on base >= 3 && < 4.6 without any warnings.

Sorry for being a bit late with #34 causing 87c3baa and not knowing git enough to send pull requests ^_^". Thank you for your patience with this fix.

[CetinSert]

@tibbe Jenkins seems to support all systems I can support today and even Mac OS X. Thanks a lot for your tip!