Richard Stallman is responsible for the shrinking economy

I bet that far more code is compiled by "gcc 4.3 from Debian Lenny for i386" than by "Keil C for ARM9".

Actually, you have to look at the whole system. For example, in theory, the compiler may behave different if the system's memory allocator hands out pages in a different order on your computer than on mine, so symbols end up in different hash buckets.

And whereas gcc requires just a small set of operating system primitives, commercial compilers tend to depend on dynamic loaders, USB drivers, network libraries, crypto services, etc., to manage their licenses and dongles.

This is not at all far-fetched. I've seen a linker that plainly crashes on Win2k but works in WinXP, and every few months, I'm seeing obscure problems that go away after reboot (such as compiler crashing when invoked from Emacs but not when invoked from cmd.exe, or vice versa). I've never seen that with un-dongled software.

Stefan

What is QT??? How can a tool license require that your source code be GPL? The fact that you run your code through a compiler can't put restrictions on the source. Why not just say that if you use these tools we own your first born male child? I thought GPL only required that if you distribute the compiled code you had to distribute the source. If you don't distribute the compiled code, then there are no restrictions. If you have not released any of the code previously, how would anyone know that the code had been compiled using a GPL tool? This not only sounds ridiculous, but it sounds unenforceable.

When you say, "their own libraries", what libraries would that be? Aren't libraries available for GCC?

Rick

QT is a library, not a tool. If you *link* with it, the result is a derived work, and that derived work must be licensed in a way that's compatible with all of its component parts, including QT. This is no different than any other third-party library - you must honor their terms, or not use them.

In message , DJ Delorie writes

That is for YOUR GCC not any one else's. This the problem GCC is a VERY wide group and somewhat fragmented.

I agree. However as you point out your GCC compiler is effectively commercial in the way you operate. But that is just one GCC out of many hundreds of different versions.

Negotiators? Who has a C compiler with a list price over 5K USD per seat?

OK.. I can probably guess who :-)

Email me off list.

OK.

In message , Stefan Reuther writes

AFAIK it is on free download still if not I have a full copy complete with all the manuals.

Don't know. Not seen it in years but these things seldom disappear completely

One of the people who wrote it is within walking distance from my office

Richard Stallman has leveled the playing field.

We use both Open Source (Linux, etc.) and Windows CE for embedded systems, depending on the application.

It is because of Open Source that we are now able to purchase a development kit for under USD 200:

Faizal Haji Founder and Principal Consultant Mirano Systems Inc, Canada Web:

Op Thu, 16 Apr 2009 19:02:30 +0200 schreef Stan Katz :

"You can drink this new beer we've made, but you can only tell us how it tastes."

Yeah right, I'd like to see things like that hold up in court. Even a U.S. court.

That thought had crossed my mind, as well.

Jon

Unless there is a clause that the customer of the product may not provide any information on the performance of their product to anyone other than the vendor.

In message , Walter Banks writes

SO you can get a copy of that... Not doing too well on this myth that commercial Sw disappears and you cant get it again

In article , Chris H writes: |> >

|> >One of the people who wrote it is within walking distance from my office |> |> SO you can get a copy of that... Not doing too well on this myth that |> commercial Sw disappears and you cant get it again

Oh please, stop spin-doctoring.

There's a huge difference between *anyone* being able to get a copy of deprecated, ancient, and no longer sold/maintained software (as it is easily the case with Open Source) -- and only those who know someone who knows someone who has certain connections.

And you very well know that.

Rainer

In that case, they'd have to prove the case well enough. And it is possible to keep traces of such communications quite private and inaccessible.

What is terrible about this discussion is that the situation may at times be such that it has to take place at all.

Jon

You are being very cautious in failing to identify the compiler. This seems, to me, to be a quite useful piece of information to release about them. I consider their behaviour atroctious.

... snip ...

Did you establish that it hadn't been transformed into:

i--; while (' ' == p[i]) i--;

and you may have missed the code executing the initial i--; ? Or the compiler may have done just that.

... snip ...

That is obviously a fault with the dongle handling code. It probably indicates the manufacturer is very lax in checking their library use. Thus pointing it out is worth while, and may induce them to drop the offensive dongle.

... snip ...

Your new myth is that W Banks is everybody. You could have identified at least one speciment of 'everybody' who doesn't meet the requirement of being W Banks if W Banks had not been careless in eliminating attributions. At any rate your logic is so flawed that I fear for your software.

FYI GCC generally separates the compiler and the code generator. The compiler is revised for compilation advances. The code generator need not be altered, barring insects. A defined gcc-specific language does the connection, and that has the minimum changes. No patches are necessary (but may occur).