Evangelism: Weekly Update March 20th

[julianduque]

io.js 1.6 release

This week we had a two io.js releases v1.6.1 and v1.6.0, complete changelog can be found on GitHub.

Notable changes

1.6.1

• path: New type-checking on path.resolve() #1153 uncovered some edge-cases being relied upon in the wild, most notably path.dirname(undefined). Type-checking has been loosened for path.dirname(), path.basename(), and path.extname() (Colin Ihrig) #1216.
• querystring: Internal optimizations in querystring.parse() and querystring.stringify() #847 prevented Number literals from being properly converted via querystring.escape() #1208, exposing a blind-spot in the test suite. The bug and the tests have now been fixed (Jeremiah Senkpiel) #1213.

1.6.0

• node: a new -r or --require command-line option can be used to pre-load modules at start-up (Ali Ijaz Sheikh) #881.
• querystring: parse() and stringify() are now faster (Brian White) #847.
• http: the http.ClientRequest#flush() method has been deprecated and replaced with http.ClientRequest#flushHeaders() to match the same change now in Node.js v0.12 as per joyent/node#9048 (Yosuke Furukawa) #1156.
• net: allow server.listen() to accept a String option for port, e.g. { port: "1234" }, to match the same option being accepted in net.connect() as of joyent/node#9268 (Ben Noordhuis) #1116.
• tls: further work on the reported memory leak although there appears to be a minor leak remaining for the use-case in question, track progress at #1075.
• v8: backport a fix for an integer overflow when --max_old_space_size values above 4096 are used (Ben Noordhuis) #1166.
• platforms: the io.js CI system now reports passes on FreeBSD and SmartOS (Solaris).
• npm: upgrade npm to 2.7.1. See npm CHANGELOG.md for details. Summary:
  • 6823807 #7121 npm install --save for Git dependencies saves the URL passed in, instead of the temporary directory used to clone the remote repo. Fixes using Git dependencies when shrinkwwapping. In the process, rewrote the Git dependency caching code. Again. No more single-letter variable names, and a much clearer workflow. (@othiym23)
  • abdd040 [email protected]: Provide more helpful error messages when JSON parse errors are encountered by using a more forgiving JSON parser than JSON.parse. (@smikes)
  • c56cfcd #7525 npm dedupe handles scoped packages. (@KidkArolis)
  • 4ef1412 #7075 If you try to tag a release as a valid semver range, npm publish and npm tag will error early instead of proceeding. (@smikes)

Known Issues

• Possible remaining TLS-related memory leak(s), details at #1075.
• Surrogate pair in REPL can freeze terminal #690
• Not possible to build io.js as a static library #686
• process.send() is not synchronous as the docs suggest, a regression introduced in 1.0.2, see #760 and fix in #774
• Calling dns.setServers() while a DNS query is in progress can cause the process to crash on a failed assertion #894

Community Updates

• browserify supports io.js, you can check the announcement here
• express.js added support to io.js
• Over the last two weeks we got access to hardware from Joyent and upstreamed a patch to V8 so we got io.js building. After that we worked on passing tests for both SmartOS and FreeBSD which as of two days ago now pass, this was thanks to the amazing work of the build team and Johan Bergström
• Petka Antonov is proposing a workers implementation in io.js under an experimental flag, you can join the discussion here
• io.js upgraded openssl to 1.0.1m

Upcoming Events

• NodeConf tickets are on sale, June 8th and 9th at Oakland, CA and NodeConf Adventure for June 11th - 14th at Walker Creek Ranch, CA
• CascadiaJS tickets are on sale, July 8th - 10th at Washington State
• NodeConf EU tickets are on sale, September 6th - 9th at Waterford, Ireland

[yosuke-furukawa]

browserify supports io.js. https://twitter.com/yosuke_furukawa/status/577150547850969088

[yosuke-furukawa]

io.js supports SmartOS. 👏
nodejs/build#64

[yosuke-furukawa]

express supports io.js expressjs/express@1656608

[yosuke-furukawa]

@petkaantonov begins implementing light-weight process backed by OS-thread. https://github.com/iojs/io.js this feature is exciting 👯

[dotproto]

@yosuke-furukawa I looked through io.js' issues and couldn't find reference to @petkaantonov's work lightweight processes backed by OS threads. Do you have a link?

[rvagg]

nodejs/node#1159

You might want to be cautious using this a in a weekly update though, there's not guarantee it'll actually land or that it'll ever come out from behind a feature flag. In general, big changes like this have to go through a long process of bikeshedding before becoming landable, if they get there at all. Node+threads has a very rocky history and this is going to bump into all of those things.

IMO one of the most interesting parts of this is using a feature flag to expose the functionality for testing, something that we'll hopefully see more of. But again, no guarantee that we'll ever get to use this.

[rvagg]

Looks like we're nearing FreeBSD builds going from always red to always blue on Jenkins: https://jenkins-iojs.nodesource.com/job/iojs+any-pr+multi/327/ (note that's not on io.js/v1.x yet). @jbergstroem may have an update on that before the newsletter goes out depending on whether nodejs/node#1167 gets merged or not I think but this is great news.

[jbergstroem]

I just merged nodejs/node#1167. Lets give the bots a try or two before we assume I didn't mess things up.

[yosuke-furukawa]

@SVincent oops, I mis-pasted the link, thanks @rvagg.

Yes, this feature has lots of chages. I guess iojs team will discuss this on next meeting. we had better to wait this feature tagged like semver-minor or semver-major by @iojs/tc .

But I am so excited on this feature.

[yosuke-furukawa]

If evangelism WG will write this feature on next weekly news, we should note the feature is not guaranteed, but need more opinions from community.

I watched this pull request,
community has various opinions.

• workers: initial implementation node#1159 (comment)
• workers: initial implementation node#1159 (comment)
• workers: initial implementation node#1159 (comment)

@rvagg
IMO, iojs/tc members need to discuss about the feature. you might want to apply tc-meeting or tc-agenda label to the p-r.

[jbergstroem]

Just following up re above.

I'd go as far as stating that both the FreeBSD and SmartOS buildbots are passing.

[rvagg]

I don't believe this has made it to an a weekly roundup yet but the thing I mentioned in a previous thread about Joyent offering build machines has actually come to fruition and the SmartOS slaves that @jbergstroem mentioned and on which he has been working to bring them back up to full support are actually supplied by Joyent for io.js use.

• https://jenkins-iojs.nodesource.com/computer/iojs-joyent-smartos14-32-1/
• https://jenkins-iojs.nodesource.com/computer/iojs-joyent-smartos14-64-1/

If it hasn't been mentioned yet then it would be good to put this in there as a positive sign on Joyent's behalf.

No binaries for "sunos" for io.js yet but that may be coming if there's a call for it.

[yosuke-furukawa]

io.js upgrades their openssl. nodejs/node#1206
no need to worry about the vulnerability on latest io.js.

[julianduque]

@iojs/evangelism #46 (comment) looks good? any comment before pushing and publishing?

[jbergstroem]

@julianduque possibly nitpicking; but saying that build team added support to freebsd is not entirely true since we've had build-slaves running freebsd for a long time. The news is that they now pass all tests.

[julianduque]

@jbergstroem good to know, updated the weekly with that comment, anything else?

[jbergstroem]

@julianduque nah, that'd be it. To summarise: over the last two weeks we got access to hardware from Joyent and upstreamed a patch to v8 so we got io.js building. After that we worked on passing tests for both SmartOS and FreeBSD which as of two days ago now pass.

[julianduque]

@jbergstroem I like your copy, will use it as is ;)

[jbergstroem]

Just did some comments on the commit; two minor issues.

[julianduque]

@jbergstroem thanks, PR and comment fixed

[yosuke-furukawa]

@julianduque LGTM!!!
Great work.

[julianduque]

I forgot to publish it yesterday, sent the draft to Node.js and JavaScript publication, waiting for @mikeal approval

[mikeal]

didn't realize anyone was waiting for me on this, I've been out for a while. once you think it's ready publish :)

[julianduque]

@mikeal I published it under @iojs https://medium.com/@iojs/io-js-1-6-release-1df38cf64e6c since we don't have access to the Node.js and JavaScript publication, how can we get access to that? or should we be publishing under @iojs in future weeklys?

[mikeal]

Let's pull in @hij1nx, can you add @iojs as an editor of the Node & JavaScript publication?

[mikeal]

We should try to keep the titles consistent. Between the title difference and the lack of being in the publication we had way less traction with this weeks.

[heapwolf]

@mikeal 👍 done!

[piscisaureus]

I like the "weekly update" title better than the "release" story. Also, there's always plenty of stuff going on - a release doesn't have to be the top story.

I also think it might help to add some spoilers to the title. E.g. "io.js week of march 20th - workers support & 1.6 release"

[mikeal]

@piscisaureus we usually do that in the subtitle of each post.

[julianduque]

@piscisaureus @mikeal missed the title, I updated the last article to match the title.

I think we had less traction due to the late publication, this should went live on friday instead of monday, now with access to the publication it will be easier to publish without need an approval from @mikeal or @hij1nx - Thanks :)

[julianduque]

New weekly update issue created: #48

[diagramatics]

Wait, this is weird. If I may barge in, the post isn't formatted like the others do. Normally the title is in H2, this new post has a higher heading level overall. Is this intended?

[mikeal]

@diagramatics I just fixed the formatting differences you described. @julianduque I usually re-format a bit from the direct markdown because the headings on medium can be quite large.

[diagramatics]

Ah, okay. Seems like the Medium post still has a bigger heading, but a confirmation that this is not intended is enough. Thanks.