Release proposal: v2.2.1

[rvagg]

• [c5a1009903] - build: avoid passing empty strings to build flags (Johan Bergström) #1789
• [5d83401086] - doc: put SEMVER-MINOR on pre-load module fix 2.2.0 (Rod Vagg)
• [4d6b768e5d] - http: revert deprecation of client property (Michaël Zasso) #1852

/R= @targos@mscdex@Fishrock123@thefourtheye@silverwind@ChALkeR (anyone else want to review?)

I can push this out within an hour or two given some eyes and perhaps even some actual testing!

[chrisdickinson]

This LGTM, I'll follow the thread and will be available to test on OSX / CentOS.

[rvagg]

tests @ https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/739/

nightly @ https://jenkins-iojs.nodesource.com/job/iojs+release+nightly/190/

[Fishrock123]

deps/npm/node_modules/request/node_modules/qs/Readme.md

uh oh, not this again??

LGTM though!

[rvagg]

OSX:

not ok 157 - test-dgram-send-callback-recursive.js #/Users/iojs/build/workspace/iojs+any-pr+multi/nodes/osx1010/test/parallel/test-dgram-send-callback-recursive.js:15 # throw new Error('infinite loop detected'); # ^ #Error: infinite loop detected # at SendWrap.onsend [as callback] (/Users/iojs/build/workspace/iojs+any-pr+multi/nodes/osx1010/test/parallel/test-dgram-send-callback-recursive.js:15:11) # at Immediate._onImmediate (dgram.js:337:10) # at processImmediate [as _immediateCallback] (timers.js:371:17) 

Anyone got any clues on this? I don't think I've seen this failure before.

Nightlies are here for testing if you have time: https://iojs.org/download/nightly/v2.2.1-nightly20150601c5a1009903/

[silverwind]

Anyone got any clues on this

Thought that was fixed in 4e2f999, cc: @yosuke-furukawa

If someone could confirm that test-npm works without ssl errors, that'd be great.

[yosuke-furukawa]

It's my test. I will investigate it .

[domenic]

Why was the qs readme deleted?

[Fishrock123]

Why was the qs readme deleted?

Weird git-ness originating several npm releases ago relating to case insensitive filesystems (maybe-ish).

[ChALkeR]

See 49bb7de

[rvagg]

updated to remove the readme change, so frustrating

[yosuke-furukawa]

hm... I can reproduce the error. sorry.
I should loosen the condition. I will pull request it soon.

[rvagg]

running again @ https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/740/

[rvagg]

building @ https://jenkins-iojs.nodesource.com/job/iojs+release/69/

[rvagg]

released @ https://iojs.org/dist/latest/

[thefourtheye]

reverts the removal of an undocumented client property on client connection

We didn't remove that, right?

[rvagg]

yeah, probably not the best description, feel free to PR to fix it to something more accurate

[silverwind]

'revert the improper deprecation' is probably best, like I've done in the commit log.

[thefourtheye]

Ya, that sounds better @silverwind 👍

[rvagg]

it wasn't deprecation that caused the problem though, it was changing the nature of it

[thefourtheye]

Then lets explain the actual problem itself. "Moving client to the IncomingMessage's prototype broke request, as it is expecting client to be the object's own property." Can we point the actual line which does this?

[silverwind]

changing the nature of it

That's pretty improper in my book 😛

[rvagg]

please look at #1859 and give a tick if you think that's accurate, everyone who's commenting here