Of course it would be nice, and one would think that IAR should be interested in solving problems with their software - but it might also be that their software works “in the usual case”.
You’re quite right, and from my explorations so far, some (not all) issues can be caused by the compiler optimization settings, and having the provided projects set to ‘maximum’ optimization out of the box. Code that behaves well ‘in the usual case’ as you put it, can go off the rails if the compiler is then set to maximum speed/size savings.
They provide samples, which is nice, and it seems they worked for you in debug mode (unless I misunderstood something in your post - your demand for 'Release" mode suggested that). To me, that means there is no fundamental problem with their hardware or the code they supply.
Well, I have to say you’re mostly right here. Yes, examples are nice, and debug mode does work very well 
My request for release mode however is related to useful code: IE the supplied projects help with the basics of a valid c program for the embedded target. What they don’t cover is how to burn code into the MCU in a permanent fashion, IE how to actually use the product, not just perpetually debug it. Since this product is billed as their ‘top of the line’ ARM platform, it’s suspicious that there is no attempt at actually utilizing it’s strength (specifically the 5003 is notable for a very generous amount of onboard flash).
Olimex is, in my opinion, mostly a PCB manufacturer. You could hardly build a prototype PCB yourself for the money Olimex charges you.
Never claimed I could, I simply claimed I could get more for my money. Heh.
You’re comparing apples and oranges. The GBA is a consumer product manufactured for a mass market (therefor cheap).
Oh very true, but, remember my comment was in reply to Tsvetan’s assertion that Olimex had the best value for money on the market, which is plainly wrong. Even way off the mark when you compare it to the GBA scene. Especially the DS model which comes with an ARM9 as well as an ARM7 and two screens for still less than the Olimex solution. 
So yes Olimex, does have a good value, but on careful review, there is much better available
es, there are a lot of developers, and yes, they’re sharing their experience, tools and software, but you don’t pay for that - if you had to pay for it, you’d probably have to pay as much as for one of the vendor-supplied development kits (OKI, Atmel, whatever)
Well, there’s a will and a wish about that that ties in with the above value assertions again. The value of a development platform is more than simply the sum of its components. It’s also the ease with which it can be utilized, and the degree to which assistance is available. True, I didn’t pay much for the olimex board compared to what the board costs direct from oki, but get this:
So far, of all of IAR, Olimex, and Oki Semiconductor, Oki and one of their engineers has been the most help in trying to get my hardware to work for me… It was an Oki engineer who confirmed my suspicions about the optimizations in the IAR compiler
I mean that is above and beyond, an engineer taking time to help a guy who owns one chip worth a few dollars, on someone elses development board, with someone elses software! Claaaaasy 
Contrast that with the IAR fumble (told me how to fix the wrong chip), and the way in which i’ve essentially been told to go choke on a “hotdog” by Tsvetan, and it really puts the relative reputability of the involved parties into perspective.
Those that want to program a GBA are lucky that a third party is offering support for free.
Far luckier than those of us who believe Olimex is capable of producing a single working IAR project. 
Dominic, please do write back, since I find it reassuring to have another perspective on this, even if you are mostly supporting Olimex with your statements Please consider my counterpoints and get back to me.
And as for Tsvetan, have you decided so fast that Olimex is incapable of accomplishing this, or is it just unwilling to support its product?[/b]
