What are my options on OS X?

Stefan Arentz

I would like to start playing with the Propeller but I am not on Windows. I'm mostly using OS X and Linux.

What are my options for those platforms?

Is there is an open source toolchain available?

I have no problems with compiling and uploading code from the command line.

S.

Ale

have a look at the thread on linux and the propeller from a couple days ago. I use normally vmware on linux and Mac OSX just for the propeller tool nothing else, well and the logic ports's front end.

Mike Green

There is a Python downloader. Look in the "sticky" thread at the top of the thread list with Graham Stabler's Good Thread Index.
Here: http://forums.parallax.com/showthread.php?p=622354

You can run the Propeller Compiler/Assembler under CodeWeavers' CrossOver Mac (their commercial Wine version). There are some bugs where you get an error dialog that you can just dismiss. This happens when you traverse the directory tree and doesn't seem to affect compilation. There's also a font problem in displaying the hex for the compiled program that also doesn't affect compilation. The Propellent Executable runs under the command line in Windows and should work with CrossOver Mac or possibly even Wine for compilation, but I haven't actually tried it.

Normally I run Windows XP under VMWare's Fusion to use the Propeller Tool. I've also used Parallel's Desktop for the same thing. Both work very well.

Stefan Arentz

Thanks for the replies. But seriously, it doesn't sound like a very nice solution.

Anyone interested in implementing an open source implementation of the compiler?

That would certainly make the Propeller more interesting for a lot of people I think.

S.

Spork Frog

There's no open source toolchain at the moment. However, I use a combination of Wine, the Python Downloader, and Propellent and have never had any issues. If you're interested, I've put together a guide for Linux here.

Making an open-source toolchain from scratch would not be a simple task by any stretch of imagination. However, we do have the source code of the Spin interpreter, as well as the complete spec for PASM, so it wouldn't be impossible.

Post Edited (Spork Frog) : 8/24/2008 7:27:54 PM GMT

Stefan Arentz

Hmmm. The hardware architecture of the Propeller looks interesting but the fact that the eco-system around it is completely proprietary and closed does not really motivate me to explore this platform. I just found out that the bytecode is not even documented. So there is really no way to work on your own tools or get things going on alternative platforms.

This is rather sad to see happening in 2008. I can understand it from a sales perspective of course. But I'm sure they would actually sell more Propeller chips if the tools and spec were simply free and open.

S.

Mike Green

Stefan,
The Spin bytecode is now documented (not well, but it is available). The Propeller toolchain is free, but not open. Mostly it's because it was not originally written to be open and it would take too much in the way of critical resources to document it sufficiently to be useful to anyone.

Paul Baker

Hi Stefan, I understand your concerns. Let me briefly explain why things are the way they are.

Parallax considers providing technical support as one of our most important duties as a company and we have found that it often means the difference between a "purchase once" customer and a lifelong customer. Almost without exception everything we release; be it hardware, software or information, we fully support and our customers have come to expect this. With respect to the bytecode and the source code of our IDEs, we simply do not have the resources to support questions from customers regarding these things.

Additionally, there is exactly one person who knows the bytecodes within the company, and the information isn't neatly recorded in some internal document but interspersed in various places like the comments of the compiler and source code for the spin interpreter. He is also the person who is designing the next Propeller and having him compile the information would cause a delay in the next chip's development, which many of our customers are eagerly awaiting. WRT the IDE we again have one person that writes it, and that is a fraction of his overall responsibilities. Writing, updating and maintaining multiple OS IDEs is something we cannot devote the necessary resources towards. Also we do not want to open source the IDE because it would lead to multiple slight variations, and our customers would to a degree expect us to support which again isn't possible. One final wrinkle to the equation is that the compiler is written in intel assembly which makes it very difficult to export to other OSs.

I know this isn't want you want to hear, and for that I appologize. You are far from the first person to ask for this and certainly won't be the last. I have come to believe that the type of person who is attracted to the novel architecture of the Propeller is also the type of person inclined to use OS X and linux because they are also different from the mainstream way of doing things. We fully encourage people to explore those options which we are unable to provide, and customers have done a good job at figuring out the bytecode and spin interpreter on thier own. The information is out there (plus using the Windows IDE to proof the generated bytecode) for a determined soul(s) to create toolchains for alternate OSs, but to date no-one has created a fully working IDE for OS X or linux. Many, if not most, of our customers find doing what was described by Mike and Spork Frog adequate (though certainly not ideal).

One final note, sales has no impact on these decisions, after all we don't charge anything for the IDE.

I hope you consider using one of the techniques provided. If·interrupt-less concurrent processing makes sense to you, it is a really fun and powerful microcontroller to use.

Best Regards,

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Post Edited (Paul Baker (Parallax)) : 8/25/2008 7:25:08 AM GMT

Beanie2k

@Paul;

There is something in your post that has me concerned. You say much of the bytecode documentation and other design aspects are stored in the head of one person. I don't mean to sound negative in any way but.... what if something happened to that person? Where would that leave the Prop? Parallax? Its customers?

Paul Baker

The information is stored, in backed up systems, and a person such as I could decipher the information. But we leave the sole control of that information to Chip Gracey the president of our company (and IT can access in some unforseen event).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Post Edited (Paul Baker (Parallax)) : 8/25/2008 7:30:47 AM GMT

Beanie2k

Thanks Paul.

Forrest

VirtualBox www.virtualbox.org/ is a free alternative to running x86 operating systems on a Mac, though you'll still need to purchase Windows XP.

Stefan Arentz

Paul, thanks for your long reply.

I think the situation that you describe is exactly what a much more open approach to Propeller development and architecture could change. If the documentation were online and if people could build an eco-system around it then you would not have to worry so much about having this info in just one person's head.

I fully understand that you don't have time to answer all kinds of questions that you are not waiting for. However, the Parallax/Propeller community is smart enough to deal with that. Others will have answers ready, just like what is happening now on this forum already.

If you guys ever decide to be more open then I'll definitely take a look at Propeller again. RIght now it will just not work for me.

S.

BradC

Personally short of the source code for the compiler, I'm not sure how much more open it could get.

We have the source code for the _interpreter_.. a careful read of that is enough to document _every_ possible behaviour if you look closely.

I have only 2 questions for Parallax.
1 ) Is the Prop-II going to use SPIN and PASM (fully acknowledging that the interpreter will need modification to run on the new chip) ...
2 ) Are we going to get access to the source for the interpreter for the new chip also. (Personally having the interpreter source has answered every question I might have had about SPIN)

I'd love to be able to use my existing SPIN code base with a simple re-compile for the new chip.
and number (2) is a clincher for me personally [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!

Paul Baker

Stefan, I'm sorry to hear that, but I respect your decision.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.