Releasing 2.6.3.5

[kazu-yamamoto]

Cc @eborden @glguy

It appeared that fixing #271 is too radical for 2.6.3.4. MIN_VERSION_network cannot capture this change. It should be done in 2.7.0.0.

• Revert Do not closeFd within sendFd #271
• Release 2.6.3.5
• Disabling 2.6.3.4 on Hackage

[eborden]

This sounds like a reasonable path.

[kazu-yamamoto]

@eborden Reverting and bump-up is done. We can release 2.6.3.5 whenever you like.

[kazu-yamamoto]

We should resolve #313 before releasing 2.6.3.5.

[kazu-yamamoto]

@eborden May I ask you to release 2.6.3.5 now?
#313 is messy.
Now I believe that we should not wait for fixing it because we cannot reproduce it yet.
One case of AppVeyor did not pass.
But I do believe that is due to a bug of test framework.

[kazu-yamamoto]

Ping @eborden

[eborden]

I'll put up a candidate tonight.

[eborden]

Candidate uploaded here: https://hackage.haskell.org/package/network-2.6.3.5/candidate

[eborden]

Candidate released.

Cafe message sent:

Announcing network-2.6.3.5

• Reverting "Do not closeFd within sendFd"
  #271

Sorry for any disruption this versioning snafu may have caused.

https://hackage.haskell.org/package/network-2.6.3.5

[eborden]

2.6.3.4 has been deprecated

[kazu-yamamoto]

@eborden Thank you!