Free C compiler with nice examples can be found here.
This compiler is not based on other compilers and I am writing it from scratch. So I can add any new commands.
Wow. A noble effort, give WinAVR/GCC is free and CodeVision is so low cost.
New “commands”, in C, can be merely library calls rather than an entirely new compiler. But, have at it! The more the merrier.
What I’d like to see is a compiler to native code of a modern structured Basic. We don’t have that with BASCOM and FASTAVR is kaput. There is one other high-end Basic for AVRs but it is cost prohibitive for those working avocations. ZBasic is superb but compiles to a virtual machine which is perfect for one-of projects, but not at scale .
Thanks stevech for your nice post. !!
Any plans on making this C compiler open source someday?
I also play around with IDEs and I’m working on a new design for a cross platform IDE. I haven’t made my own compiler yet.
Regarding BASIC, that’s part of my idea for an open VM multi-language design:
http://www.ericengler.com/downloads/EmbeddedVM.txt
I don’t have a desire to make a native code compiler because its much harder and can’t easily be ported to other target families. With a VM design I can also roll in an RTOS into the VM kernel. And all this can work with a small amount of memory.
My key goals are first-class in-circuit debugging support, and cross-platform development tools. I also value supporting more than 1 source language, and supporting many different targets. Highly-optimized target-specific machine code is just not compatible with my other goals given the limitations on my free time.
But group-oriented open source projects can open up other possibilities because more people can contribute. So I’m open to consider that…
Eric
Eric you have very nice goal to achieve. Good luck !!!
My goal is simple. To write simple c compiler useful for hobby in my free time.
And no plan for open source.
New example with AVR-MT-128and DS18S20 added.
englere:
I don’t have a desire to make a native code compiler because its much harder and can’t easily be ported to other target families. With a VM design I can also roll in an RTOS into the VM kernel. And all this can work with a small amount of memory.
A public domain VM would be interesting, as compared to the superb VM included free with purchase of one of the low cost AVR chips or modules from zbasic.com. Has a cooperative RTOS too.
zbasic is definitely cool but you have to use their hardware.
Yes, but you can buy the chip-only version rather than a module and it's very inexpensive, considering the sophistication of the language +VM and resulting rapid development capability.englere:
zbasic is definitely cool but you have to use their hardware.