Hacker News new | past | comments | ask | show | jobs | submit login
Mono 4.0.0 Release Notes (mono-project.com)
171 points by tizoc on April 4, 2015 | hide | past | favorite | 62 comments



And with this release we see the beginning of the convergence of the Mono and Microsoft codebases. I'm glad Mono will be more reliable as a result, there have always been persistent reliability and compatibility issues that the Mono team hasn't been able to really fix.

It's also a little sad, in a way. It relegates Xamarin to maintainers of a distribution for another operating system. At the end of the day, C# and .NET are now officially first class citizens on Linux. Speaking as someone that actually ran Mono in production for around a year before realizing it was a ridiculous waste of time and money and switching to Windows Server 2012 after finding and filling dozens of untrackable deployment and stability issues in ASP.NET under Mono, I'm definitely not ready to switch back anytime soon, but Mono is definitely foreseeably a real option.


The Mono team did an amazing job, and they actually had a lot of support deep within Microsoft all along. But yeah, I think it's okay to say "it never really worked."

(And there's nothing wrong with being a port factory that deals with obscure path issues all day. That's why they give you a paycheck...)

Still, Xamarin is doing a LOT more than just Mono, much of it extraordinary.

Also delivering something that actually works is an ego boost. Probably attracts the "right" kind of employee for the task, too.


> But yeah, I think it's okay to say "it never really worked."

I don't. I mean yes mono <4.0 has a lot of problems, but it works well enough for me to shell out money for applications running on mono and be happy about it.


I literally saved two businesses as favors to friends. One by replacing a Mono app on Mac with Windows running in a VM. The second by doing the two hours of config work they were afraid of to host on Windows Server instead of a Linux box. (At the same damn price!)

Xamarin did an immense amount of work pushing that boulder up a hill. Glad you got results, but frankly it did a lot of damage to a lot of people. I say this as a Microsoft fanboy and someone who deployed sites you've heard of on Mono + Postgres + CentOS.

In the last year or two many .NET vendors went from "yes, it works on Mono!" to "it should work but no support." This indicates the formal death of the ecosystem to anyone but apologist hobbyists and thankfully we've gotten to a better place now.


> This indicates the formal death of the ecosystem to anyone but apologist hobbyists and thankfully we've gotten to a better place now.

A couple of things. First: "This", so far, is anecdotal and doesn't indicate much of anything, really. For something concrete, look at Unity. Maybe you haven't used anything built with it, but everything I've seen come out of it works just fine.

Second: Apologist hobbyists? That's not true and needlessly insulting. If your aim is to just marginalize anyone with a differing opinion please don't expect me to take you seriously.


Um... Unity spent the last three years trying to find a way to pry themselves off Mono. Unity "works" because devs using Unity work their asses off tiptoeing around a really old version of Mono. You picked about the worst possible example to make your point.

http://forum.unity3d.com/threads/unity-5-and-mono-runtime.23...


Unity works though. Unity games have a much better batting average in my experience than C++ ports, and I have more problems with "cross-platform" games and other applications written in Java.

These aren't hobbyists and they're not definitely not apologists. They're making good money on it. If it was as bad as you say, why bother with the Linux/Mac support to begin with? That's just noise compared to the Windows market.

I also have to point out that an ancient version of mono being problematic doesn't mean mono failed or doesn't work. By my measure, even that old version of mono works.

I don't know what your metric is. People ship working code and make money in parts of the market they wouldn't otherwise have. But I guess that doesn't count in the face of what? your decision not to use it?


I'm not sure what your point is here. Talented programmers have been releasing stuff that works great on shitty platforms forever. If the money is there, you work around the nonsense, warts and all.

It doesn't mean DirectX 6.1 is a great 3D game platform, nor AmigaDOS an awesome place to make a video editor, nor Mono under Unity the bees knees for an iOS game.

It feels like we're debating whether the Mono bullshit was waist deep or merely knee deep. Which doesn't matter because Xamarin and Microsoft just pulled out a giant firehose and washed everything clean.


> I'm not sure what your point is here. Talented programmers have been releasing stuff that works great on shitty platforms forever. If the money is there, you work around the nonsense, warts and all.

That is precisely my point. We're still doing it and we will be doing it for the foreseeable.

For example:

> It doesn't mean DirectX 6.1 is a great 3D game platform [...]

If your opinion is DX6.1 didn't work, I think what you really mean is "isn't the best option under ideal circumstances today." I would agree with that assessment for mono, but "cross-platform" is not something anyone puts next to "ideal circumstances" when it comes to the work involved in making it happen.

> It feels like we're debating whether the Mono bullshit was waist deep or merely knee deep.

There is certainly bullshit to contend with in mono, but there's bullshit in Microsoft's too. And every other platform throughout the ages as you helpfully illustrated. A topical example is porting a game in C++ vs. writing working code for mono. C++ is uncontested as a working technology, but I would pick Unity on mono in a heartbeat for that task. So where is the line, really? In development all you can do is keep above it and all the better if you can keep your socks dry.

I get the impression you are very unimpressed with mono's failings and have written it off and want to complain about it. That's fine but it doesn't mean the hyperbole you've chosen to express yourself is accurate.


If the goal of the Mono project were to provide a cross-platform scripting language for games then, with obvious caveats about all the well-known problems, sure, "that worked."

But the goals and history of the project were clear. Meanwhile if you tried to run a desktop app on Mac or a server app on Linux, it was a total fucking debacle.

Xamarin punted desktop, then they punted server, and they landed on mobile. And even then they couldn't keep up or keep compatible.

"It worked" much the same way "the Space Shuttle worked." Awesome effort, cool project, didn't meet spec, didn't meet goals, and it had one nine of uptime before it repeatedly crashed.

I remain grateful for the work they did. But I'm glad things are moving forward sensibly now. Because unlike WINE, they really never got close.


> It's also a little sad, in a way. It relegates Xamarin to maintainers of a distribution for another operating system.

I'm not sure I understand what you mean here. As I understand things, Mono is still keeping its runtime - JIT, garbage collector (the release notes here mention optimizations to SGen), etc. That means Mono isn't nor is it planning to be "just" a distro for Microsoft's .NET. (But it is using more and more Microsoft library and compiler code.)


>I'm not sure I understand what you mean here. As I understand things, Mono is still keeping its runtime - JIT, garbage collector (the release notes here mention optimizations to SGen), etc.

So? Who will still use those, if there's a fully supported MS based (+ some open source effort for some edges) alternative?


It will be interesting to see if anyone releases an open-source port of the MS runtime to ARM... Xamarin could wind up left with only their proprietary cross-platform UI toolkit!


For Android this could be possible.

For iOS I think you would have to figure out how to hook up MS runtime to a LLVM AOT compiler and then do all the work Mono team did to make sure libraries compile/work without JIT.

I really hope Xamarin guys follow Microsoft lead and make entry level version of their product free. It's not just about the ~25$ per month, it's about community as well - say some guy X is maintaining .NET library Y. I write up a port/pull for Xamarin - X goes "sweet but I can't test that since I don't own Xamarin and I can't even include it in my test suite to not break it further". Now I have to maintain a branch of Y for Xamarin and most other people are likely not going to find it or willing to use it.

I don't know which % of their profit comes from the indie licensing but if it's not significant as I suspect I would hope that they see this - it will help both their adoption and the OSS community build tools for their product.


That depends on what Microsoft plans to do with the upcoming .NET Native compiler, but I guess that would be kept closed.


Runtime is already open and runs on Linux, one just has to cross compile it, optimize for mobile and do the bindings and such that Xamarin already does. That's still quite a bit of legwork since Xamarin has been doing it for some years now. Not saying someone couldn't do it, but it would take a good amount of time to be on the level Xamarin is in stability and performance.

I pay for Xamarin personally and think it's well worth it if doing cross-platform compared to the alternatives out there. I'd continue to do so until there's a very stable open sourced alternative. However, I think Xamarin knows such a possibility could happen and that's why they primarily target enterprise and not individual developers.

https://github.com/dotnet/coreclr


> At the end of the day, C# and .NET are now officially first class citizens on Linux.

C# and .NET will never be first class citizens on unix. This is by design.

http://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish


Maybe not first class, but I think Microsoft's eaten a bit of humble pie (ala IBM ~15-20 years ago when they started ramping up Linux efforts) and generally playing nice.

Who ten years ago would have imagined so much of .NET would wind up MIT licensed, not just some 'shared source' malarky, on Sourceforge's ~2015 successor?

This isn't the Microsoft of the Halloween memos anymore.


No, we are making them first class. When we have a bug that breaks one of our three build queues (Windows, Linux, Mac), it will not be acceptable to break any of the platforms.


Sweet :)


You're aware the past can only inform the future, not predict it?


Some people are never going to get over the past...guessing if the poster is American, they must not trust the British either for burning down the White House 200 years ago. I lived through the late 90s and 00s Microsoft era and know as well any developer the shenanigans that went on for better or worse. I always had a fondness for C# over Java for its willingness to innovate while Java stagnated, but never wanted to commit myself to being stuck only in the Windows ecosystem.

That said, I am excited to see the direction Microsoft is heading. It enthused me enough to start using Xamarin when I want to build cross platform for personal projects and to recently switch my day job to a company that uses .net for most of the their software. Not overly fond of cmd.exe and powershell on Windows still, but for most things, there's always Cygwin and it too has come a long way. Main issue is Powershell inherited all of cmd.exe's bad user interaction design (tab completion, select/copy/paste, buffer, history up/down scrolling) when compared to a typical *nix terminal.


> guessing if the poster is American, they must not trust the British either for burning down the White House 200 years ago.

Huh, what a fantastic point you make. Because I can trust the British in spite of something they did before any current living human being was alive, I should be willing to forget things that Microsoft did 10 years ago in an industry that hasn't fundamentally changed.

Are you being serious?


> Are you being serious?

Just countering your argument with an equally absurd one.

I probably wouldn't even bother if you hadn't dragged out the tired old "embrace, extend, extinguish" meme/mantra that was popular 10-15 years ago (and then wiki linking to it like no one is aware of it). I respect the differences in our opinions, but having to convey your point by adding that line in 2015 make one's argument a bit hallow.


Fair point about Wikipedia. I suppose that was unnecessary. I do think that's a bit of an ad hominem though.

> equally absurd one.

You must be trolling me. You cannot possibly think that the statement "Britain cannot be trusted, after what they did in the 18th century" and "Microsoft cannot be trusted after what they did in the early 2000's" are on equal footing.

But let's not resort to making this argument about analogies. How about you respond to my statement in the highest comment: What Microsoft can possibly stand to gain, fiscally speaking, from open sourcing .NET? From future profits to the current market.


> You must be trolling me.

Okay you're correct about the comparison, it was slightly more absurd to highlight the absurdity I saw in your original remark.

> What Microsoft can possibly stand to gain, fiscally speaking, from open sourcing .NET?

What does Google, Facebook or any other public, for-profit entity have to gain from open sourcing? I don't see Microsoft's motives as being any different. Companies aren't individuals, despite what US Corporate Law states. They're a system composed of the good and bad individuals currently working there and are subject to change. I would keep going, but I would basically end up saying everything coldtea already said to you a couple days ago.

If you you're looking for an official answer, you might want to ask the core .net developer in this thread as I don't work for or have any affiliation with Microsoft.


That's a 15+ years off slogan from the Desktop/Browser wars era...

Nowhere near as relevant or enlightening as you think it is in 2015 (wasn't much of an analysis in 1998 either, but still...)


Oh please.

Microsoft hasn't had some crazy awakening like, "Hey, have you guys heard of this open source thing? We should embrace open source and free software!"

They've realized that they're beginning to fall from #1 in the eyes of developers. So they're throwing some bones to try to dissuade people from leaving their platforms.

Microsoft stands to gain NOTHING from open sourcing their platforms besides developer marketshare and perhaps a bit of PR from those who are stupid enough to think that Microsoft suddenly turned altruistic.

They're scared shitless. They shipped a bunch of products that are not so popular with consumers (Windows Phone 7, WP8, Windows 8, Surface Pro, etc.) and developers are starting to care less about targeting people on Windows. This is just damage control.

I hope it isn't enough.


Response #1:

Yeah man, it's all capitalism, they don't do it out of the goodness of their hearts!

What a revelation!

Pass the bong.

Response #2:

They're is no "Microsoft" as a person. That's a company, an aggregate entity which doesn't have either brains or a strategy.

What DOES exist is people running the company. Those people can change (and have, Gates->Ballmer->this new guy), and these new people can have different ideas about how to go about things.

If there are any propritary companies you like, then imagine that MS could be turned to behave like them. There is no "evil brain" behind it all making sure MS will always be like in the 1998 carricature you have in mind.

IBM circa 1999-2004 was hailed by OSS advocates for supporting Linux and OS -- yet the same company was the fearful Microsoft equivalent of its era, a few decades back.

Oh, and this evil AT&T that everybody feared and they had to break up in an anti-trust case? This is the company that gave us UNIX in the first place. (Not to mention Xerox Parc, another company's proprietary research facility, without which we'd have shit today).

And vice versa: those "nice" companies could change course if the wrong people came on board, or their current people saw an opportunity and got greedy.

Response #3:

A lot of us don't give a flying fuck about the success of OSS as some ideological venture.

We like some OSS software, and we like some proprietary software. Heck, some of us make a living creating proprietary software for proprietary companies, from small indie firms to Adobe to Microsoft to what have you.

And even as users, a lot of us won't settle for an inferior OSS product is there's an (affordable) superior proprietary one (and vice versa).


This is a draft. Mono 4.0 hasn't been released yet. Who keeps publishing mono's release-notes drafts as HN news?


I added a note to the page to make this clear.

HN mods: can we get a "(draft)" note added to the submission title as well?


thanks Alex


People who want to learn some details about this upcoming release before it happens?


This is on HN homepage now, so the question you should ask is not who published it, but who upvoted it. And the answer is, at least 158 different people who found it useful.


> We no longer build or package reference assemblies for the .NET 2.0, .NET 3.5 or .NET 4.0 APIs. We only support now the .NET 4.5 assemblies as well as the mobile-based profiles.

Does that go for the runtime as well? Does this mean I can no longer run a (compiled) .NET 2.0 application on Linux?


No, it should run on the 4.5 profile just fine as backwards compatibility in the .NET APIs is very good.

Please note that this is still a draft and the .NET 2.0/3.5 assemblies will likely reappear as reference assemblies (i.e. containing only method signatures in IL, so you can build against those profiles but not run on them).


I look forward to being able to use this all in Unity3D in 7 years.


Unity is moving away from Mono anyhow,

http://blogs.unity3d.com/2014/05/20/the-future-of-scripting-...

so updates in Mono are irrelevant to Unity.


What? That's what this post says:

> we will also be upgrading to recent versions of the Mono C# compiler, base class libraries, and runtime for use in the editor

> We will continue to use the Mono C# compiler (and perhaps later, Roslyn). We will continue to use the Mono class libraries. All currently supported features and 3rd party libraries which work with Mono AOT should continue to work with IL2CPP. We are only seeking to provide a replacement for the Mono VM and AOT compiler, and will keep on leveraging the wonderful Mono Project.


Currently they support both - il2cpp isn't done yet. But as the blogpost calls it, it is the future of scripting in Unity.

Also, yes, this is a replacement for the VM and AOT, but the libraries in Mono are still very useful. Many of them are currently being replaced by Microsoft code as well, though, so it looks like .NET, Mono and il2cpp will share lots of code that way. Just not the VMs.


They are (apparently) using Mono (C# compiler + base class libraries + runtime) on the desktop for their IDE, and they are using IL2CPP and parts of Mono (C# compiler + base class libraries, but not the runtime) on their targets (iOS, Android, ...)


> Unity is moving away from Mono anyhow,

I don't know man. Mono moved away from Boehm GC to a real gc (SGen) because it never really worked for them bespite feeding it a lot of metadata. Others before them have found out that tanspiling to C/C++ isn't really that great (some of the more recent ones where HipHop PHP).


Just curious.. Why cant the mono team replaces almost whole of their codebase with the code Microsoft open sourced? Is there anything stopping them to do that? Maybe licensing issues?


First, the code Microsoft released does not build and run on all platforms. They promised Linux support, and are working on it now in the open, but it's not there yet. So Mono can't just adopt the Microsoft .NET VM without losing features.

Second, Microsoft is open-sourcing their whole VM - JIT, GC, as well as compiler and libraries. Compiler and library code is already being utilized by Mono, and more all the time. However, Mono doesn't even have plans to replace their JIT or GC with .NET's, as far as I know (and the Trello board linked to in the article seems consistent with that).

Could Mono do that (once Microsoft's builds on Linux and other platforms Mono needs)? Yes in theory, but it would mean abandoning all the massive amount of work they put into their VM over many years. I don't know whether it would be rational or not to do so.

Licensing isn't a problem for Mono to use Microsoft code. However, Mono's runtime is LGPL licensed, and this has been a source of revenue for them (people pay Xamarin for licenses for platforms like consoles where LGPL is a problem). So to replace the Mono VM with Microsoft's could be a step back there.


I can't speak for mono, but converging makes sense where it makes sense -- if mono has a solid implementation of something they consider core and the MS implementation isn't trivial to port, I don't see a good reason to port it.


Because it is a lot of work and takes time to integrate the pieces MS open sourced :)

There are also a few cases where the code needs to be tweaked so it works correctly on Mono and the platforms it runs on.


We are working on replacing those parts that we can.

But there are several issues:

* reference source code not being portable and not working on Unix. * complex dependencies between the portable and non portable pieces * some bits were only released recently, and without those it was not possible to port other chunks * dynamic happy code high prevents some code from being adopted as Mono can be used as a full static compiler on iOS, Xbox and Playstation.

So as time permits we will bring those over.

We also plan on supporting the "core" model where our contributions will flow back upstream.

But "core" is still a bit undefined


You know that Mono is cross-platform and Microsoft code only runs in Windows right?


Is this true?


Not anymore. The .NET Core has been ported to OS X and Linux already. It still has a long way to go in stability though.


The OP is talking about the reference sources, not .NET Core (which is an unfinished effort, btw).


NetFx core contains many of the same things that NetFx desktop does, and everything in Netfx Core is required to be cross-platform, so I would say that many, if not most, pieces of the framework should be portable directly to mono.


Yes, especially for the IO code, which is C# code p/invoking into Windows kernel APIs.

Mono project has been mainly pulling Microsoft's code that is managed, which is the one that should run in any OS unmodified.


Yes. Its literally the point of the Mono project.


It is a shame that this got submitted before we actually published the code. It is Easter and many of our engineers are taking these days off.

The release notes are also incomplete and not ready for publishing

Miguel


Sorry, saw it on reddit as "Mono 4 Released"[1] and reposted here, without noticing it was just a draft (the "THIS IS A DRAFT" comment wasn't there at the moment).

[1] http://www.reddit.com/r/programming/comments/31ezp0/mono_4_r...


Tail call optimisation in F# ?


Wait...mono doesn't do tail call optimization with f#???


It is (or was) complicated, you could stack overflow on two mutually recursive functions or tail calling inside continuations.


I'm getting constant memory on this (mono v3.12.1)

    let rec beep x =
        if x = 0 then x else boop (x - 1)
    and boop y =
        if y = 0 then y else beep (y - 1)
    
    let main argv =
        printfn "%A" (beep 999999999)
        0


Those are mutually recursive functions, but they're not using continuations.


This draft looks great. We've been running Mono since December 2013 for a fairly big data processing framework - 2014 saw us move it into production. We've had our fair share of really horrible performance issues, runtime issues and the rest. 4.0 has some stuff in that fixes many of those nasty issues!

Awesome work by team Xamarin + Contributors.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: