Live data from Hacker News

CoreCLR ARM 32-bit progress

github.com

31–40 of 44 posts

Re: CoreCLR ARM 32-bit progress

#31
post #26

Earlier quoted context omitted.

Most devs I know not in .net are not gonna jump for this reason. It's so similar to what I can get from Python and Scala and Java... Why would I want to touch the Microsoft ecosystem? Seems like it will take 10 or 15 years for a new round of college students to switch over.. so it has to be a very long term play.

Many of the current generation don't have any issues with Microsoft. We do mostly Java and .NET projects, and enough juniors get on both stacks.

Last time I checked, the better colleges around taught some combo of C, Python, Java, and LISP.

Sometimes a few business classes with .net, but it seems very rare overall.

The professors have a lot of influence in the future..

Re: CoreCLR ARM 32-bit progress

#32
post #27
post #26

Earlier quoted context omitted.

Many of the current generation don't have any issues with Microsoft. We do mostly Java and .NET projects, and enough juniors get on both stacks.

Yea, inexperience with Microsoft would make them think it's okay to do that before these announcements. I agree with parent comment though, not enough to make me come back. source: C# and .NET were the first languages I taught myself after PHP in 2007-2008.

I went back in 2010. My first contact with it was when my employer, a MS partner got an early access to it, still on Alpha stage.

Nowadays I would rather do .NET projects than Java ones, as I already have the language and toolchain features today at my disposal that will only be available (with luck) on Java 10.

Also because neither Sun or Oracle really understand what native development should be like.

JavaFX has everything to compete with XAML, yet it really needs some love.

Re: CoreCLR ARM 32-bit progress

#33
Fantastic!

I was looking into this just yesterday to see how far away it was, so the timing is really fortunate for me.

Between this and the VS Code builds we should now have pretty solid support for developing .NET on Pi and/or Chromebook, so I'm really excited to get rolling with it, especially seeing as I have some free time this week. I'll try to write up my experiences too in case it helps anyone else.

Re: CoreCLR ARM 32-bit progress

#34
post #14

Earlier quoted context omitted.

And that invalidates my comment... how? Microsoft announcements appear here in a nicely-spaced way that seems to garner lots of interest and, yet, not induce fatigue. How and why this is happening is "an exercise left to the reader."

1995 - circa 2014 Action: Microsoft is hostile to Open Source movement for many years. Reaction: developers have a deep distrust of Microsoft. 2014 - current day Action: Microsoft open sources a series of core projects and starts contributing to major Open Source projects. Reaction: ???

> 1995 - circa 2014

Microsoft launched CodePlex for open source development in 2006, and it had released some bits of code before that.

Microsoft also started releasing .NET source code in January 2008 under its MS-RSL license [1], so there's been a long transition, not a sudden recent change.

[1] https://weblogs.asp.net/scottgu/releasing-the-source-code-fo...

Re: CoreCLR ARM 32-bit progress

#36

Nice work. It will be great to use ASP.NET Core on a Pi with official support. I tried .NET out on the Pi using Mono, years ago when it was relatively new but after the software floating point OS was deprecated. The weird hardware floating point CPU made it pretty much unusable. Even simple things like DateTime objects would fail to work correctly. Similar issues were in Java IIRC but they had a special build of that…

Interesting, imho Java makes no guarantees about FP portability if you don't enforce it with the strictfp keyword, you'd get 80bit on x87 and 64bit on other architectures. Why would this cause failing tests though? Does that mean they relied on extended precision?

Not sure of the specifics. Java SE is now included with the official OS (Raspbian/NOOBS) [0]. When the Pi first launched you had to use the soft FP version of the OS or Java wouldn't work but this had a significant performance penalty. They added official support for hard FP a few years ago [1]. I believe the .NET issue was similar in that the VM (CLR i.e. JVM) didn't support the hard FP version of the OS. FP on the Pi processor is a bit weird and a non-standard implementation from what I've read.

P.S. In researching this post I noticed that one of my projects (a globe wake lamp) has been included in the latest project book (p108) [2]. First I've heard of this! This is actually an old version and there is a newer write-up on my blog [3].

[0]: https://www.raspberrypi.org/downloads/noobs/

[1]: https://www.raspberrypi.org/blog/oracle-java-on-raspberry-pi...

[2]: https://www.raspberrypi.org/magpi-issues/Projects_Book_v2.pd... [pdf]

[3]: https://unop.uk/pi-glowbe-mark-2/

Re: CoreCLR ARM 32-bit progress

#37
post #5

This is a major step forwards. Having .NET on ARM is one of the most important milestones imho.

It was already there, but not the main version. .NET Compact and Micro had ARM versions. There is also an Arduino (AVR) that uses .NET Micro as their OS.

I just want to emphasize the fact that not only is the .NET Micro Framework something that exists, but that it has been available since 2009 and remains so, under the items of an arguably much more agreeable license: Apache 2.0

Re: CoreCLR ARM 32-bit progress

#38
post #5

This is a major step forwards. Having .NET on ARM is one of the most important milestones imho.

It was already there, but not the main version. .NET Compact and Micro had ARM versions. There is also an Arduino (AVR) that uses .NET Micro as their OS.

One thing: the Netduino devices are ARMv7-M (Cortex-M3 and Cortex-M4) devices, not AVR

Re: CoreCLR ARM 32-bit progress

#39
post #13

Earlier quoted context omitted.

This is not an official announcement, it is not even an announcement. What the fuck are you talking about?

I NEVER SAID IT WAS! Good grief, people. You'd think the strawman argument wouldn't get this kind of traction on HN. I had a thought that I've seen lots of announcements -- nicely timed -- along these lines over the past few months. That's it. MOST of it has come from official sources. I was just making a freaking observation, and making an inference about it. It wasn't particularly about this post at all. Sorry for…

You can't go off like this on Hacker News after you clearly alluded to organized PR (if you didn't, then you weren't saying anything at all). To then label the rest of the interested discussion a circle-jerk is even worse. Please comment civilly and substantively here or not at all.

https://news.ycombinator.com/newsguidelines.html

Re: CoreCLR ARM 32-bit progress

#40
post #13

Earlier quoted context omitted.

This is not an official announcement, it is not even an announcement. What the fuck are you talking about?

I NEVER SAID IT WAS! Good grief, people. You'd think the strawman argument wouldn't get this kind of traction on HN. I had a thought that I've seen lots of announcements -- nicely timed -- along these lines over the past few months. That's it. MOST of it has come from official sources. I was just making a freaking observation, and making an inference about it. It wasn't particularly about this post at all. Sorry for…

And now this has been flagged.

Again, for the guy who apparently mods this whole thing, and to whom I can't reply: it was a random thought, for which I've been thrashed. Super.

I absolutely alluded to organized PR; I never claimed this was part of it. The timing simply brought the idea to the foreground. I never claimed the author or the poster was a shill. It was just an unfortunate coincidence. I apologize to the internet gods for not doing a thorough background search on this particular post and vetting it so as not to imply that THIS post was part of a conspiracy, or clearly stating that this was NOT implied. The implication just didn't cross my stupid little peabrain mind. Perhaps understanding every nuance of how every single person could interpret a post is part of the terrible, implied responsibility of posting on a public forum, and I readily admit that I utter failed to do so.

By all means, sctb, if you have the power to delete my account, please do so. Because of previous examples of nonsense exactly like this, I've tried several times, but HN offers me no way to do it. So I wind up lurking, then slowly coming back, and then this sort of thing happens again. And I never seem to learn that someone, somewhere will find a reason in ANY post to be upset about it, even AFTER EXPLANATION, and either I find these people way more often than others, or I just have a harder time with the response.

Post reply on HN