Release proposal v2.1.0 #1777
Release proposal v2.1.0 #1777
Conversation
mscdex
added
the
meta
|
👍 There's still problems with v8 so no 3.0 for now. |
|
Switched to running all of the windows build machines to manually from cmd since it's the running-as-a-service that's causing these failures. @wolfeidau is suggesting that it's likely a security-policy thing that's preventing some action from occurring and maybe the timeout is because there's a dialog box or a UAC-ish thing (although it's 2008 as well so probably not UAC exactly). Got this failure on run 701 on SmartOS, been seeing this occasionally on those machines: test-net-server-max-connections.js https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/701/nodes=smartos14-64, who do we have for SmartOS? @jbergstroem, @geek, @misterdjules? Anyone care to have a look and consider why this might be a recurring problem? New run that should have Windows greens, if not all-green: https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/702/ |
|
just remembered we have a @nodejs/platform-solaris; @misterdjules can we add you to that team as well? |
|
Opened #1779 as the last in the series of |
|
@rvagg I don't have the time right now to look at https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/701/nodes=smartos14-64 because I'm supposed to be on vacation and I still have a lot to do with node v0.12.x before I can actually get off my computer :) But definitely, please cc me on Solaris/SmartOS issues, hopefully I'll be in a better position to help next time. I would also like to be part of the @nodejs/platform-solaris team, thanks for bringing that up! |
|
@misterdjules I'm onto it! Just need to setup a new vm locally (and downloading with imgadm is currently painfully slow). |
|
@jbergstroem Thank you 👍 |
|
I unfortunately can't reliably reproduce it outside of Jenkins; just tried running it 100 times without a fail on our 32-bit builder (as well as a local one). I'll bring this up on our next build meeting. |
|
@rvagg: win32 + running as a service is pretty much what I was playing around with when tinkering around a few weeks ago – but we ended up with the same errors after a week passed. There's definitely something there though. Last time I stopped the windows service and ran it through |
|
All
|
|
this issue has been converted to a PR, see diff for the changelog entry for this release https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/705/ moving on to release within an hour or two, consider this a warning if you have any pending concerns |
|
|
||
| * **crypto**: Diffie-Hellman key exchange (DHE) parameters (`'dhparams'`) must now be 1024 bits or longer or an error will be thrown. A warning will also be printed to the console if you supply less than 2048 bits. See https://weakdh.org/ for further context on this security concern. (Shigeki Ohtsu) [#1739](https://github.com/nodejs/io.js/pull/1739). | ||
| * **node**: A new `--trace-sync-io` command line flag will print a warning and a stack trace whenever a synchronous API is used. This can be used to track down synchronous calls that may be slowing down an application. (Trevor Norris) [#1707](https://github.com/nodejs/io.js/pull/1707). | ||
| * **node**: To allow for chaining of methods, the `setTimeout()`, `setKeepAlive()`, `setNoDelay()`, `ref()` and `unref()` methods used in `'net'`, `'dgram'`, `'http'`, `'https'` and `'tls'` now return the current instance instead of `undefined` (Roman Reiss) [#1779](https://github.com/nodejs/io.js/pull/1779). |
|
Changelog LGTM besides those two points. |
|
@rvagg Do I have time to land [email protected] before we push this out? #1763 (Just need to add to the squashed commit's changelog, but will wait on the go ahead from you.) |
rvagg
force-pushed
the
v2.1.0-release
branch
2 times, most recently
from
7c92f3a
to
c71bad6
Compare
|
thanks for picking those up @silverwind, fixed now. @Fishrock123 I'll wait for the npm merge, no rush, it'd be good to get it in |
|
npm landed. 👍 |
PR-URL: nodejs#1532 Notable Changes: * crypto: Diffie-Hellman key exchange (DHE) parameters ('dhparams') must now be 1024 bits or longer or an error will be thrown. A warning will also be printed to the console if you supply less than 2048 bits. See https://weakdh.org/ for further context on this security concern. (Shigeki Ohtsu) nodejs#1739. * node: A new --trace-sync-io command line flag will print a warning and a stack trace whenever a synchronous API is used. This can be used to track down synchronous calls that may be slowing down an application. (Trevor Norris) nodejs#1707. * node: To allow for chaining of methods, the setTimeout(), setKeepAlive(), setNoDelay(), ref() and unref() methods used in 'net', 'dgram', 'http', 'https' and 'tls' now return the current instance instead of undefined (Roman Reiss & Evan Lucas) nodejs#1699 nodejs#1768 nodejs#1779. * npm: Upgraded to v2.10.1, release notes can be found in https://github.com/npm/npm/releases/tag/v2.10.1 and https://github.com/npm/npm/releases/tag/v2.10.0. * util: A significant speed-up (in the order of 35%) for the common-case of a single string argument to util.format(), used by console.log() (Сковорода Никита Андреевич) nodejs#1749.
|
https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/706/ Running another test run, there was an odd failure on PI1 which I suspect is just a time-related issue but I just want to be sure, plus the npm update was merged. Will push for a release after this. |
|
Windows 2008 failure - the particular machine rebooted with Windows Updates and went back in to Jenkins-as-a-service mode so the timeout failures are back! I've put it back in to cmd.exe mode. test-cluster-disconnect.js failed on ARMv7, I'm going to call this an anomaly but it probably needs investigation. Going ahead with release |
PR-URL: #1777 Notable Changes: * crypto: Diffie-Hellman key exchange (DHE) parameters ('dhparams') must now be 1024 bits or longer or an error will be thrown. A warning will also be printed to the console if you supply less than 2048 bits. See https://weakdh.org/ for further context on this security concern. (Shigeki Ohtsu) #1739. * node: A new --trace-sync-io command line flag will print a warning and a stack trace whenever a synchronous API is used. This can be used to track down synchronous calls that may be slowing down an application. (Trevor Norris) #1707. * node: To allow for chaining of methods, the setTimeout(), setKeepAlive(), setNoDelay(), ref() and unref() methods used in 'net', 'dgram', 'http', 'https' and 'tls' now return the current instance instead of undefined (Roman Reiss & Evan Lucas) #1699 #1768 #1779. * npm: Upgraded to v2.10.1, release notes can be found in https://github.com/npm/npm/releases/tag/v2.10.1 and https://github.com/npm/npm/releases/tag/v2.10.0. * util: A significant speed-up (in the order of 35%) for the common-case of a single string argument to util.format(), used by console.log() (Сковорода Никита Андреевич) #1749.
PR-URL: nodejs/node#1777 Notable Changes: * crypto: Diffie-Hellman key exchange (DHE) parameters ('dhparams') must now be 1024 bits or longer or an error will be thrown. A warning will also be printed to the console if you supply less than 2048 bits. See https://weakdh.org/ for further context on this security concern. (Shigeki Ohtsu) #1739. * node: A new --trace-sync-io command line flag will print a warning and a stack trace whenever a synchronous API is used. This can be used to track down synchronous calls that may be slowing down an application. (Trevor Norris) #1707. * node: To allow for chaining of methods, the setTimeout(), setKeepAlive(), setNoDelay(), ref() and unref() methods used in 'net', 'dgram', 'http', 'https' and 'tls' now return the current instance instead of undefined (Roman Reiss & Evan Lucas) #1699 #1768 #1779. * npm: Upgraded to v2.10.1, release notes can be found in https://github.com/npm/npm/releases/tag/v2.10.1 and https://github.com/npm/npm/releases/tag/v2.10.0. * util: A significant speed-up (in the order of 35%) for the common-case of a single string argument to util.format(), used by console.log() (Сковорода Никита Андреевич) #1749.
It's been more than a week, how could we be so slack?!
Only thing forcing semver-minor here is the
--trace-sync-ioflag. I have a vague memory of a discussion resulting in an decision that CLI changes shouldn't impact semver (I don't actually agree with this fwiw), however I see that--requirestill got in with semver-minor so it looks like we're saying that CLI changes do impact semver.a3ee6a5ad2] - (SEMVER-MAJOR) 2015-05-24 io.js v2.1.0 Release (Rod Vagg) iojs/io.js#15329da168b71f] - buffer: optimize Buffer.byteLength (Brendan Ashworth) #17132b1c01c2cc] - build: refactor pkg-config for shared libraries (Johan Bergström) #16033c44100558] - core: set PROVIDER type as Persistent class id (Trevor Norris) #1730c1de6d249e] - (SEMVER-MINOR) core: implement runtime flag to trace sync io (Trevor Norris) #17079e7099fa4e] - deps: make node-gyp work with io.js (cjihrig) iojs/io.js#990c54d057598] - deps: upgrade to npm 2.10.1 (Rebecca Turner) #1763367ffd167d] - doc: update AUTHORS list (Rod Vagg) #17762bb2f06b3e] - doc: fix typo in CONTRIBUTING.md (Rich Trott) #1755515afc6367] - doc: path is ignored in url.format (Maurice Butler) #1753f0a8bc3f84] - doc: fix spelling in CHANGELOG (Felipe Batista)86dd244d9b] - doc: add notes to child_process.fork() and .exec() (Rich Trott) #1718066274794c] - doc: update links from iojs/io.js to nodejs/io.js (Frederic Hemberger) #1715cb381fe3e0] - (SEMVER-MINOR) net: return this from setNoDelay and setKeepAlive (Roman Reiss) #177985d9983009] - net: persist net.Socket options before connect (Evan Lucas) #151839dde3222e] - (SEMVER-MINOR) net,dgram: return this from ref and unref methods (Roman Reiss) #17685773438913] - test: fix jslint error (Michaël Zasso) #1743867631986f] - test: fix test-sync-io-option (Santiago Gimeno) #1734f29762f4dd] - test: enable linting for tests (Roman Reiss) #17212a71f02988] - tls: emit errors happening before handshake finish (Malte-Thorben Bruns) #176980342f649d] - tls: use.destroy(err)instead of destroy+emit (Fedor Indutny) #17119b35be5810] - tls: make server not use DHE in less than 1024bits (Shigeki Ohtsu) #1739214d02040e] - util: speed up common case of formatting string (Сковорода Никита Андреевич) #1749d144e96fbf] - win,node-gyp: enable delay-load hook by default (Bert Belder) #17630d6d3dda95] - win,node-gyp: make delay-load hook C89 compliant (Sharat M R) TooTallNate/node-gyp#616I see no reason to delay this, I'm proposing that we push ahead with this within ~24h unless something major comes up.
Windows timeouts still happening but apart from that we're looking solid: https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/700/ (I'm going to have yet another go at understanding the Windows Jenkins problems, they are becoming a big problem for defining stability)