Nekochan Net

So I've been programming C/C++/Visual C++ w/ MFC and ATL for about 10 years now. I'm finally just starting to learn .NET and some of its features seem very intriguing to me. Apparently .NET is a lot more like Java than it is like C++. It is actually compiled on the fly (with the "just in time" compiler). Since the .NET code and the JIT compiler are actually seperate elements, like Java, this seems to scream "port me to irix!"

Does anyone know if this is doable in irix? I realize IRIX is a dying OS and SGI is trying to shake it like a bad case of crabs, but does anyone know if it can be done, and if so, how much effort would it take? If it were possible to run .NET code on IRIX, this would mean that any new software written for Windows XP/Vista should run on IRIX. (Vista is shipping with WinFX which is basically a low-level .NET support layer) I also believe that Visual Studio was written in .NET, which means that if we could get .NET support in IRIX, you could even use Visual Studio to develop code(say what you will about microsoft, but I firmly believe VS 2005 is the BEST IDE out there for coding - once you use intellisense you'll never go back, and oh yea, VS 2005 Express is FREE)

Also, since all SGI MIPS Processors since the R4000 (correct me if im wrong) have been 64 bit, new applications should definetly run on the processor, but how fast they would run is anyones guess.

Does anyone have enough knowledge in this field to give a yay/nay to wheater its even feasible?

Links:
Wikipedia entry on MSIL
http://en.wikipedia.org/wiki/Microsoft_Intermediate_Language

MSDN Entry describing how .NET code can run on embedded devices with MIPS processors
http://msdn2.microsoft.com/en-us/library/ms836805.aspx

Anyone want to make comments directly to me, please email me at jeff-gilbert AT uiowa.edu

I once tried to port Mono to IRIX but failed. There are x86 specific assembly code in the source code, which makes it a hard job.

AFAIK VS is not written entirely in .NET. They might use it for UI but for performance the core functionalities are written in C++.

Hi,

I'm not to going to respond to the idea that VS is the best IDE ever. Lots of development tools out there. There is more than one way to do things.

However, as to porting to .NET to Irix, sure, it's possible. In fact, to a degree is can already be done courtesy of Mono. You are limited to .NET 1.1 and 2.0 likely will never materialize.

http://www.mono-project.com/Main_Page

The difficulty is in getting Mono compiled onto Irix properly. Easy or hard I do not know. Feel free to give it a crack and see what happens. If nothing else, you might be able to add MIPS/Irix to the list of supported platforms.

I think you are way oversimplifying the effort needed to make .NET apps run on Irix (or anything non-Windows for that matter). Embedded MIPS vs. a full-up workstation are two very different worlds. And trying to keep up with Microsoft's renaming of APIs and technologies is like hitting a fast-moving target. Good luck, you'll need it.

Since it is likely you are probably using 2.0 type .NET applications (knowingly or by default), running .NET on Irix will never happen. The reason is certain pieces of the .NET framework in terms of implementation and behavior are not published by Microsoft. Without those pieces, interoperability will be all but impossible.

You'd honestly be better off trying to port a piece of software from scratch that trying to get the .NET binaries to run. It could be done more quickly and with a lot less pain. Look into cross-platform toolkits. It will be a challenge but at least it is doable. Understand that Windows and Unix are very different worlds. Trying to use Microsoft tools and code on a Unix platform of any sort is like being a devout Satanic Cultist at a Evangelical Gathering. Don't expect it too go over to well. (No offense to any of the Satanic Cultists out there).

Heck, while you're at it, why not just write a cross-assembler for MSIL to MIPS? Still easier than porting .NET.

Matt

Yes, I am greatly oversimplifling the port from CE to .NET. Obviously this would not be easy at all. But MS is strongly encouraging developers to stay away from MFC and ATL calls in .NET programs (thus reducing the amount of native code that is present in the application). As time progresses, .NET applications will use less and less native code, making them more likely to run on another platform. Also, as long as people keep working on mono and DotGNU, the higher the likelihood that someone will figure out .NET 2.0 and 3.0. Once this is done and documented, the relevent features could then be added to the JIT compiler to support .NET 2.0. This will just take time.

This is obviously a big project, but I'm willing to try to tackle it as a learning experience. I know nothing about programming for IRIX, nor about writing compilers for MIPS processors. No guarentee that anything will ever come of this project, but if anyone else wants to try to tackle it with me they're welcome to help. now, i just need to start boning up on CIL and MIPS assembley...

As stated above, I will be beginning this project when I recieve my new Fuel workstation. If there are other interested parties that would want to contribute, I can enable remote login capability for my Fuel, Indigo, and Indy machines(might be gettin an o2 and octance/octane2 from someone as well). So even if you don't own an SGI system, you're welcome to participate.

I'd be willing to do some testing and some coding, and having access to your machine would be nice, although I wouldn't use it that often. Since my SGI is on a high latency connection, I only use it locally, which is why I'd personally like access. I can't help with C++ coding, just testing the .NET 1.1 language/vm implementation capability. I'm a noob obviously, but over time maybe I can help more.

I've been advising people similarly for years!

_________________
4xRS6K 2xHP9K 6xSUN 1xDEC 14xMAC 7xPC 2xPS2

This is both true and not true. You see, I have also started and abandoned a mono port to IRIX. I had it, err, somewhat working.

There's virtually no architecture-specific material in the source tree. There's very little OS-specific material. It built on IRIX with <100 lines of code changed, all but a few in IFDEFs and build scripts.

The only work that really needs to be done is in the code generator. The entire bytecode-to-native-code compiler is retargetable, using a scheme very similar to lcc. Basically, someone needs to write and test conversions between the compiler's basic operations and MIPS assembly. This is harder than it sounds. I had a minimal set mangled from looking at the SPARC code, and while it sort of worked, it was not at all stable.

Mono still has an interpreter library that is used to build the dissembler and a few tests, but the interpreter is no longer capable of executing code. Don't bother barking up that tree.

If someone were to port mono to bigendian Linux/MIPS, it would take about an hour to get it up and running on IRIX.

That said, a MIPSpro port is totally impossible. Mono relies on being able to use inline assembly. Even if it were possible to change that, no patches that change that would ever be accepted. One would be in the very unenviable position of maintaining a weird fork for a very, very unpopular OS.



VS2003 and later are based upon the old Visual Interdev codebase, from 1996. It predates .NET by quite a lot

Will the gods be angry at porting such a hardcore Micky$oft product ???

I'ld wear thick rubber shoes while working on the project just in case

I wonder if Microsoft will ever embrace Open Source IP like Apple (not that they're my idea of the ideal tech company. OS X was an easy way to get a canned mature OS off the shelf to start from. Although they've done plenty of work.). After the whole SCO -> IBM/Linux think (with Microsoft pulling SCO's strings) I thought OpenSource organizations should go on the offensive to product *their* IP and possibly even get compensation to support future/current OpenSource products. *Many* products have gotten caught (WRT54G, NSLU2) quitely ripping off Linux to save on the development costs and then keeping the source code protected/secret - many more have just escaped under the radar by disguises. It wouldn't surprise me to see large amounts of GPL'ed or otherwise OS'ed copyright code in allot of commercial products - espc the more complicated libraries that would be easy to rip off and hard to have a semi-clean room rewrite. Of course when companies like Microsoft would get caught ripping off code they just release an update and rewrite the small parts of the code they we're busted for

I think Office for Linux would secure Microsoft's future prospects (at the expense of current profits - slightly) because their customer based wouldn't be forced to use a different productivity suite just because they need a reliable extensible OS. If Micky where to release an Office suite for Linux they could leverage it against OpenOffice and greatly distract from the effort of improvement. Analyst's are have down rated MS for a long time now because of 'limited future prospects' which is almost entirely because they don't want to embarrass if they 'cross over' to the other side - even just running a closed app like Office under Linux would make then turn (Novell ?) red. But it would start to ensure at least a chance of staying in the future of OS. Vista's somewhat of a flop - and MS is getting very much scummier trying to make deals with OEM's to only sell hardware that runs on Vista - and Microsoft for their part requires band new hardware to run Vista for no good reason. For example the SBLive isn't Vista compatable and never will be - there's no reason for this other then Creative's using Vista tp drive the current line of sound cards into the ground hoping they get to replace them - their trashing hardware that's only months old and that's very capable of running Vista if the drivers we're released. This is a pattern that's being steadily repeated with all the Vendors & OEMS when it comes to Vista. Suprisingly the groups it pisses off the most are the 'evoirMENTAL groups' thats want us all to live in Teepee's and burn wood to cook instead of having CANDU reactors supplying power with clean non-enriched fuel (gee - like the good 'ole days. Too bad the good ole days used energy much less effectively then currently. The earth couldn't support the population density we have now if we went back to 'nature')

A good DX9/10 interface for X-Windows would rock. MS could the likely sell their 'OS for Linux' at the same price or even far more then normal Windows and have less work and they wouldn't be held hostage as much by their lack of design and engineering they put into products from a security/network/low level OS standpoint. If MS messed up the Office for Linux who cares it could/would easily be jailed so it wouldn't have the same effect as w'Doze

The earth could not support the current population at western standard of living. Though how we got there from a discussion on .NET, I'm not sure.

_________________
4xRS6K 2xHP9K 6xSUN 1xDEC 14xMAC 7xPC 2xPS2

Yeah I dunno how I babbled to there from .NET. Yeah but the earth wouldn't support the way of life (burning coal, wood etc ...) that many humans lived hundreds of years ago even. I mean can you image London if people we're burning wood and coal for heat with the population it has now. The only was to support the current population in a sustainable manner is buy prudent use of of are cool inventions and toys - but the consumer driven throw away attitude would have to go right away, as would driving 200 miles to work everyday and general disregard of the amount of energy we're using - I think energy use is one of the least sustainable parts of the 'Western' life. Although it would be a big change I think if there was the political will (and the population knew a little more) we could be on our way to being totally sustainable within a 5-10 years and still enjoy a good quality of life

The 'feel good' projects have to go too though. Reducing 'CO2' is a scam - we need to worry about everything *BUT* CO2. For the most part recycling is a scam - if you can't make money recycling that means overall your doing no benefit. With the exception of metals, glass and some paper (in the order that it makes sense to recycle ) not only does %40 of the stuff in the blue box just make in into the trash heap but when they do recycle stuff like plastic their using more energy then just breaking it down and replacing the product from scratch. We shouldn't be using plastic for packaging but if we do at least we shouldnt' try and kid our selves that it's okay because we're 'recycling it' - when industry is forced to recycle plastics we're doing *more* damage that good in energy usage and defective products that get trashed anyway (for the most part) because recycle plastics generally suck (and once again use more energy then just to reproduce the material from scratch)

I'm for the environment but I hate 'tree huggers' that are just looking for a reason to be radical and oppose the status quo (however the Western status quo does suck and should go) . I mean good god - even the first *FOUNDER* of Green Peace left because they we're becoming more fixated on being a radical social organization then protecting the environment. The current 'Al Gore' style of scare is typical of the radical ideals of the group - for the most part they haven't made a big dent in Western living. but 15 years ago they said in tens years if things didn't completely change we would all be dead and dying. Hasn't happened - to get the attention they seek sooooooo much they have to go more and more overboard with each 'study' and new paper interview or people just say 'yeah I already heard that' and dismiss it. So now allot of 'studies' say we're drowning tomorrow because of melting ice caps and the day after our corpse will catch on via because of global warming. Really though their timetable for destruction has gotten closer and closer and their proposed day of reckoning gets worse and worse to the point of grossly sad humor

The whole global warming scam has got to be one of the biggest distortions of science and it really pisses me off. I mean with all this good will for the environment, the money and energy that's being spent 'reducing CO2' could be doing things that would make a real difference like LOWERING SULFUR LEVELS IN FUELS, planting trees though out cities and shutting down all smoke stacks withour modern scrubbers. This would make far more of a difference then going after CO2 *which* after all is the product of PERFECTLY CLEAN combustion. CO2 accounts for %5 of all the 'greenhouse gases' man made CO2 accounts for another %5 of all CO2 released each year (yes nature is the biggest 'polluter' and must be stopped under CO2 rules that are being pushed - which is a real danger) is from man and activities of man. Oh yeah and when they freak out because the 'ice caps are melting' there's another problem the oceans are *so* large and are such a large moderator of heat/temp that for the ocean to register a 1 degree change would take a little under 1,000 years !!!!!!!!. I mean you can't easily/quickly warm the large amount of water on earth - there's just too much of it the oceans are huge and *deep* with incredible mass to sink heat

There's one big thing that can raise the temp of Earth - solar activity. period. It turns out that CO2 causes more green plant activity which then absorbs the energy from the sun and *lowers* the temp of the planet not raises it. Gore moved the 'temp and CO2' lines in his movie so that CO2 was in line with TEMP. The actual data puts the rise of CO2 roughly 700 years *AFTER* global temp rises which makes it clear it acts like a feedback cycle (planet warms because of sun, oceans release large amounts of CO2. Earth cools . and the cycle repeats). That little prick Gore knowingly falsified the data by moving the concentrations of CO2 measure via core samples back by 700 years - otherwise it would be clear CO2 had nothing todo with it and he wouldn't have a chance to stroke his ego as world saving avatar.

Shit I mean if he *really* cared about reducing CO2 HE WOULD TELECOMMUTE AND WOULDN'T FLY HIS ASS ALL AROUND THE WORLD. Screw telecommuting he could just use the post office. Al Gore (TM) on his 'world saving adventures' (TM) releases many hundreds of times more CO2 then the average human - so it's clear ether he doesn't care and doesn't buy the CO2 thing. And no we *don't* care he was too much of a whimp to challenge Bush when he stole the election in Florida that's mean that he's a stupid whimp - not the smart victim which he puts so much effort into trying to play the role of (he's a bad actor), What we really have to fear is idiots that are trying to absorb and lock up the CO2 to humans some dude think is right, humans start projects like that on a large scale we'll *really* see environment disaster very quickly. We don't have a freaking idea of how the earth's feedback cycles work and how complex they are - if we start messing with the CO2 feedback cycle on a large scale we're really doomed - removing the CO2 *would* without a doubt start warming up the earth - which would cause the lemmings to remove *even more* CO2 and I guess you can see where this is going. So it's easy to get infinite resources when your causing the problem you claim to be fixing (eh - war on terror anybody ?)

You forgot about the volcanoes friend!



Estimate CO2 emissions significantly less (depending on current eruption activity of course) See: http://volcano.und.edu/vwdocs/Gases/man.html

Don't know about causing the end of the world, but talking about .NET running on *NIX does seem to direct the conversation that way...

Cheers,

Shawn

_________________

If the current population can only be supported by drawdown of non renewable resources then by definition it is not sustainable.

_________________
4xRS6K 2xHP9K 6xSUN 1xDEC 14xMAC 7xPC 2xPS2

Figured this would make it a little more likely:

http://download.mono-project.com/third- ... ce/2.10.5/

It's been ported to SPARC!

_________________
33mhz R3k/48mb/XS24 150mhz R4400/256mb/XL24 600mhz R14kA/2gb/V10 8x1.4ghz Itanium 2/8GB 32x600mhz R14kA/24GB 4x700mhz R16k/8GB/V12/DCD/SAS/FC/DM5 2x700mhz R16k/4GB

It ran on SPARC five years ago, before your forums necromancy.

Solaris/SPARC was one of the first platforms for mono, because when they started working on it, in 2004, it was a pretty important platform.