AVR Freaks

I managed to solve my problem gracefuly!:_

const char PROGMEM ErrorArray1[] = "testa";
const char PROGMEM ErrorArray2[] = "testaaa";
const char PROGMEM ErrorArray3[] = "test";
const char PROGMEM ErrorArray4[] = "test";
const char PROGMEM ErrorArray5[] = "testdddddd";
const char PROGMEM ErrorArray6[] = "test";
const char PROGMEM ErrorArray7[] = "test";
PGM_P const ErrorArray8[][3] PROGMEM = {{ErrorArray1,ErrorArray2},{ErrorArray3,ErrorArray4},{ErrorArray5,ErrorArray6,ErrorArray7}};

I have no shame in posting on old threads when they're tutorials.

Thanks for this; cleared up a lot for me. Clear, concise and very helpful. 

I have an application I threw an ATMega2560 in because I needed high io count and a good chunk of memory for data acquisition of many sensors, and to write that data to an SDHC card. I didn't have a lot of time to spend researching MCU's but I'm hoping to go with a smaller chip once I'm done writing the program. This goes a long way in helping me reduce my memory footprint along the way. PTSD from university CS classes 10 years ago still has me often dragging my feet any time I need to do memory allocation (*cough* segfault *cough*) so all the little things I can do to save space first are very welcome ;)

Yes! I learned this in the few hours after submitting my comment. Took a while to figure out what I needed to do but I landed on still using _p functions for in-line string literals. Looking through the rest now to see what else could be set as __flash. I tend to be a #define guy, but I’m looking for places where I’m potentially using those defines repeatedly and moving them instead to constants in flash. My degree is CE and as such I have good CS knowledge and embedded really is my favorite but I’ve only done this work sporadically the last 8 years since college. All that to say I’m geeking out a bit on increasing my code efficiency with these methods. 
 

A couple questions I still have:

how would I know when I need to use __memx vs __flash? I understand that memx is for data that crosses the 64 Kb boundary, but am I going to know when that happens?

I was also running into trouble trying to debug a variable, I thought the issue was due to it being in flash, but ended up being because it only lived in a register. So I guess my question is, if it’s static and constant, shouldn’t the debugger still know it’s location in flash? Or is the answer “maybe yes maybe no, it depends”. Which I think is the case…when I scattered my variable around a bit more I was able to break and see it in my watch list. 

I made a flash card program from a long list of English two-three word expressions.  I used this format:
 

const char PROGMEM englishStr0000[] = "a babe in arms";     // extends up to 5850 expression strings

const char * const englishStrItems[ 1 ] PROGMEM = {englishStr0000};  // extends up to 5850 array elements                                                                                         

String    proverbString;      // construct an object of class String in order to word-wrap on OLED screen

uint16_t stringIndex = random(5850);                                  
proverbString = String(FPSTR( englishStrItems[stringIndex]));

The final code for 5850 expression strings was about 100 K bytes.  It loads and runs in a 32-bit CPU through Arduino. 

This is a good tutorial.

just for my curiosity......

if they do exactly the same and you still need the <avr/progmem.h> to get to the special _P functions why use __Flash instead of PROGMEM al together?

or is this a thing with portability of the code to other platforms that in the future you no longer need the visible <avr/progmem.h> but after defining 'variables' to be '__flash char' during compile time the right function is called

well that is then why I never understood the advantages of using __flash.......

Till now I always was under the impression that __flash only replaced the "progmem"  part but that in the end you still needed to use the special _P functions to have access....

That will make life a lot easier indeed.

so if I have a:

const __flash char txt1[] = "Hello";

and a

volatile char txt2[10] = "world/r/n/0";

can I then acces them both using the same function like

void sendit(* char tosend)

{

 UDR = tosend[1];

}

Or do you on that end still do need the separation between the two as the pointers are different?

note I always have been strugling with pointers and I have not been coding for a while now as my time has been consumed by other hobbies.