Ruby 2.1.2 is released | Hacker News Reader

Ruby 2.1.2 is released

ruby-lang.org

73 pointsmerrua29 comments

While I imagine that most Windows Ruby users already know this, I found it surprising that there still has not been an installer for Ruby 2.1 released for Windows. It appears there is at least one test failing on Windows that no one has had the time to fix.

Considering that we are almost half way to the 2.2 release date (Dec 2014), it seems like Windows support may end up becoming a second-class citizen.

I'd say Windows has been a second class citizen for a long time for any of the unix-centric open source languages. Ruby, Python, even node - you're fighting an uphill battle trying to do anything on Windows.

I'm sort of amazed they've made it run at all, the systems are so different. I'd advise anyone trying to run any of these systems on Windows to use a linux VM.

>Ruby, Python, even node

Nope: Windows support in Python is good and well-maintained, afaik. It boils down to attention for neophytes, really, and applications in education... a field where Python has much more mindshare than Ruby or node (IMHO).

I concur. I've found Windows support in the Python community to be pretty good. The Ruby community seems to have more of a focus on posix platforms.

Point taken. However, like I said above, I know a few startups who use python, and the only Windows boxes there are for IE testing. Certainly no-one deploys on it. Academic computing seems to be a different world, though.

While I do lots of web work, the majority of my work with Python is as a scripting language embedded in a cross-platform app (Sublime Text). There are easily hundreds of thousands of developers who use that work on a daily basis.

The Windows support in Python is extensive enough I was able to write a web-client using ctypes and the native Windows network API to better support users with funky proxies and so forth.

Well, fair enough then, I take that back. Nonetheless, I know a few python web programmers, and zero of them use windows. I imagine that libraries would be a big issue too, not to mention the native lack of simple things like a proper command line, ssh, all the normal tools. Sure, you might be able to make it all kind-of work, but it's harder than it should be and decidedly outside of the mainstream.

Couldn't really comment on academic computing. I suppose, though, they are not really relying on the larger ecosystem as much as (say) web applications, so would not be affected so badly by the libraries issue.

Python is also used as a scripting and plugin language in several large commercial Windows applications.

While this is true, Microsoft is making a big push to get Ruby working more equitably on its platform (for example, they had a heavy presence at Chef Conf), so I wouldn't expect this to continue indefinitely.

Yeah, they say that every couple of years (eg IronRuby) - it never seems to amount to anything. As with anything MS says, I'll believe it when I see it.

> While this is true, Microsoft is making a big push to get Ruby working more equitably on its platform (for example, they had a heavy presence at Chef Conf)

Having heavy presence at a conference is clearly evidence of trying to get mindshare from users of the thing the conference focuses on, but its less clear that it is evidence of actual concrete work to get the related software working on any platform that you control. Is there really any concrete signs of Microsoft working to get Ruby working better on Windows?

That's why I'm switching to Owin + Nancy, It is really sad to see the lack of love Ruby on Windows gets.

The other side of the fence isn't much better.

Nancy's own wiki gives instructions for hosting Nancy on Ubuntu that start with downloading and compiling the latest version of Mono. They continue, "Open VS2012 and start a new console app NancyDemo.sln". Indeed, if anyone can successfully develop a Nancy application on Ubuntu, I'd love to know how. I spent two days yak-shaving NuGet (I even submitted a pull request to fix one bug) before concluding that it just doesn't work on Ubuntu.

It is good, but Im not interested in using Nancy on Linux, the reason I'm using Ruby/Sinatra on Windows is to not to depend of the IIS envyroment, now with Owin and Nancy I can acomplish the same.

> it seems like Windows support may end up becoming a second-class citizen.

It won't end up being a second-class citizen, it already is. I wrote about this 6 years ago(!) http://www.rubyinside.com/is-windows-a-first-class-platform-... .. shortly after that I ran a panel at Euruko where I asked both the audience and Matz their opinions for Windows support and no-one really cared. It's only through the efforts of a relatively small and dedicated group of people that we have Ruby running on Windows whatsoever.

> It appears there is at least one test failing on Windows that no one has had the time to fix.

The failing test for test_float wasn't Windows-specific (it looks from the tracker [1] that other packagers were hitting it even though it passing on the CI system used by core, its just that Windows users are more likely to be dependent on the packager rather than building from source without running the test suite, so RubyInstaller treating it as a blocker meant more on windows than other packagers treating it that way would mean on other platforms.)

Also, there've been a number of apparent fixes (the issue was closed 5 months ago, and then again 4 months ago), before it was (hopefully, finally) resolved four days ago. Its not -- well, for the last half or so of the bugs life -- an issue of no one taking time to fix it

> Considering that we are almost half way to the 2.2 release date (Dec 2014), it seems like Windows support may end up becoming a second-class citizen.

Irritating as it is for RubyInstaller users, this doesn't really seem to be a result of less attention being paid to Windows.

[1] https://bugs.ruby-lang.org/issues/8358

Frankly, it should be a second class citizen. Look, MS started it by using the its lack of standards compliance as a business weapon. So, they made their choices. I'd rather see the precious resources be spent on other things.

FYI this was released 10 days ago - https://news.ycombinator.com/item?id=7718581

Good point! I though I've checked ruby-lang last 10 days...

That's really good news. I hope the different GC issues that required tuning and patching of 2.1.1 have been fixed.

Readline definitely got fixed so you no longer have to apply patches to get it to compile.

Is there anything particularly notable about this release? The release notes make it sound quite minor, with a small handful of bug fixes.

These point releases are what the patch (-pNNN) versions used to be, so they are minor.

There are some GC bug fixes that address some edge cases.

It's more than just an edge case -- you're very likely to encounter the bugs that were fixed if you allocate large objects in a rails request handler. You should upgrade.

Reply on news.ycombinator.com