gnupic: Re: [gnupic] Current state of Linux Dev Tools for Microchip
Subject:
Re: [gnupic] Current state of Linux Dev Tools for Microchip
From:
Robert Pearce ####@####.####
Date:
1 Aug 2008 22:25:13 -0000
Message-Id: <20080801232405.5e56c853.rob@bdt-home.demon.co.uk>
On Fri, 01 Aug 2008 12:52:47 -0700 John Battle wrote:
> A few years ago, I finally gave up trying to make Microchip tools work
> under Linux and just built a machine toi run (ugh) Windows XP just so I
> could run MPLAB.
<snip>
>
> But.....it started me to wondering if anyone has had any success in
> building a programming environment as a native Linux app for the
> Microchip parts.
Yes, absolutely, and for the very simple reason that the free GnuPIC tools for Linux work better than the Microchip ones.
Specifically, I started out with MPASM / MPSIM in the days before MPLAB. I discovered GPSim very shortly after I discovered Linux, and was amazed at how much more capable it was. Sure, it didn't support the particular chip I wanted to use, but being open source I was able to fix that quite quickly. What mattered was that GPSim allowed me to simulate my code in a simulated environment, not just in splendid isolation like MPSim. To my mind, that feature is essential to make a simulator worth having, and last I looked MPLAB still doesn't provide it.
> I see the mailing list is alive and well, so is there
> an App now or what? How do people develop microchip stuff (especially
> dsPIC) under Linux. I understand there is a port of the C compiler and
> tools but I still don't know how to program a part or debug.
Well, I don't use the dsPIC, in fact I've only recently moved up to an 18F series device. However, for what I do the combination of gpasm, gplink and gpsim does everything I need to write the code, and then I program the chip with a home-built device. There are plenty of designs for such devices on the web, and there's Linux support for some of the "official" programmers too. If you want to program in C you can use SDCC, though I don't know how good the code it generates is. Mind you, my experience with HiSoft PICC / MPLAB on a project at work wasn't 100% problem free.
Actually, while on the subject, I recently tried to run some of my code in MPLAB to confirm whether GPSim correctly implemented the CPU behaviour. I couldn't. First, MPLAB wouldn't allow me to set things up right. Then it wouldn't compile. And when I got build errors, MPLAB refused to show me any results windows. Was that a Wine problem? No, because I was trying to run it natively on a WinXP box. So it's an MPLAB bug, and a complete killer.
I shall be sticking with the free Linux tools (also work on BSD and Windows).
Rob