snprintf not working correctly

This is likely the problem. Only problem now is with full support, it doesn't fit in the code space, yet I'm not even half way done implementing the rest of my program. I actually have almost nothing implemented at the moment other than a UART :(

Yeah, I know that. sprintf is no proper function for a micro. It creates much code size.

Any chance you know of something I can use as an alternative?

You can create your own sprintf like functions that only do exactly the things you expect.
Or you will need another processor.

Dennis

I guess I'll be writing my own. Thank you very much for the help!

Ethan, this was just a possible reason for your problem. I cannot guarantee for beeing the reason.

Dennis

It is, I just tested the function using only %d which is supported in minimal and the output works, just not with the precision I need.

OK. Bad thing. But I had the same problem some time ago. It came up early enough to switch over to a processor with larger memory.

Dennis

Yeah, I really thought 16KB would be more than enough. Might just switch to the 32KB version as that will likely have enough space. Any chance you know how to figure out how much code space I'm currently using?

Ha, I just switched to compiler 4.3.* and I can fit everything with full printf support now! I wonder how many lines of code short I am from full.

Derp, it tells you when you load. Sorry, it's late and I'm being blind. I'm using 15680 bytes, so I'm skirting pretty close to my 16k.

So it is a G2553 LaunchPad project? :)

Actually a F5308! Already on PCB too, but I can always get a F5310 as a drop in replacement which has 32KB ram. In hind sight, pretty dumb for me to not pay an extra few cents for the extra development space.

FLASH****

I need to stop working now :D

Yes, I prefer to assemble sligthly bigger processors than actual required for the project. Since we are not selling millions of units, this tiny amount of extra cost is OK. And there is always some functionality that has to be added later. Bad if your code size exceeds the processor's memory then.

Good night!

Thanks again for the help!

You're welcome!

" I prefer to assemble slightly bigger processors than actual required for the project."
Dennis, our energy meter has an F1232. Its 16k flash and 256 byte ram were plenty when this started. But then I had to add more and more functions, and now I'm hand-optimizing things just to get two or three more bytes. After already switching form a single-binary model to a specific binary for each variant. I tried to do a new F5438-based design in 2010 (which originally brought me here to E2E), but I then was redirected to implement even more features into the existing firmware (with a huge time overhead for doing this on the 1232) rather than developing a new hardware where all this would be easy to implement in a fraction of the time.
Just to save a few cents on the processor when the initial design was made (before my time).