C / C++ / MFC

Re: Timer interrupt handler on SAM3x8e – Arduino Due

4-Aug-15 4:25

I really do not know where to post this and I hope I can explain my question , so be gentle.

I am using SAM3x8e hardware timer (TC3) as an interrupt.
The “TC3_Handler” (callback?) global function name is “hardcoded” somewhere and fires as expected at desired intervals.
I do some verification of this by outputting simple message to LCD.
The LCD “class” is declared / instantiated as a global class to the entire application and works amicably.
Simple “lcd.print(“OK”)” is executed just fine in the TC3 handler.
I also do some hardware output – turn on / off LED – just for an additional indication the handler is being executed.

Now I want to execute a method / function of ANOTHER globally defined / instantiated class, again in same TC3 handler.

Adding cpal.Run(); did not compile – undefined “cpal”.

I had to ADD extern CPAL cpal; so it would compile and execute cpal.Run();

My question is
both LCD and CPAL classes are global, the LCD can be used OK, but the CPAL has to have the extern “reference” added into the TC3 handler. Why?

Thanks for reading, appreciate any help.
Cheers
Vaclav

CPallini4-Aug-15 6:29

SOLVED Re: Timer interrupt handler on SAM3x8e – Arduino Due

4-Aug-15 6:29

Probably because LCD global object is declared extern in one of the headers included by your source file containing the TC3 handler.

Vaclav_4-Aug-15 7:38

Re: SOLVED Re: Timer interrupt handler on SAM3x8e – Arduino Due

4-Aug-15 7:38

Thanks,
makes perfect sense.
I am not sure I'll spent time to actually look for such declaration, but I need to learn more about hardware interrupts so maybe I'll run into it sooner or later.
Thanks again.
Vaclav

CPallini4-Aug-15 9:50

Complex code in conditionals

4-Aug-15 9:50

You are welcome.

charlieg4-Aug-15 0:08

4-Aug-15 0:08

I need opinions on code like this:

"if ((++(m_pLF->m_iLogCnt)) >= MSG_QUEUE_FILE_SIZE)"

maybe it's just me and my past - I was taught never to depend on order of operations. This is handled with the parenthesis , but really, should it be that hard to analyze an IF statement?

Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...

"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759

CPallini4-Aug-15 0:20

4-Aug-15 0:20

That's not terribly complex, but did you really need it that way?

charlieg4-Aug-15 1:37

4-Aug-15 1:37

No, that one was not terribly complex, and I also selectively edited it. Smile | :)

Richard MacCutchan4-Aug-15 0:26

Richard MacCutchan

4-Aug-15 0:26

Obfuscation is the best way to annoy those who have to maintain your code.

charlieg4-Aug-15 1:41

4-Aug-15 1:41

You know, it's one thing to obfuscate, it's another to code in such a style as approaching enemy action. It's hard enough to write clear code. The style of code I posted just seems easy to break and difficult to debug. The debugger is going to treat that like one line of code.

Caveat: I freely admit that I choose to restrict how much code I put on a line. I like using ternary or conditional operator as it can make things much cleaner, as long as you don't get silly with it.

Richard MacCutchan4-Aug-15 2:44

Richard MacCutchan

4-Aug-15 2:44

Absolutely agree with you. Most of the time code like that is done by people who don't know any better (copied it from the internet), or who think they are so damn clever.

charlieg4-Aug-15 6:22

4-Aug-15 6:22

Well, I know who wrote the code - I'll go with clever. Bright individual, I just wanted to make sure it wasn't just me.

Smile | :)

strcpy dilemma - it works half way

Stefan_Lang11-Aug-15 2:34

Stefan_Lang

11-Aug-15 2:34

This is not too complex, but at the limit where I would consider breaking it up into multiple statements.

That said, if that were a while condition, I might be more inclined to leave it like that, because (a) the increment might be considered part of the loop iteration, and (b) moving part of the condition might require more than one additional line (e. g. once before the start of the loop and once inside).

Apart from that, I am more bothered with the naming Poke tongue | ;-P

GOTOs are a bit like wire coat hangers: they tend to breed in the darkness, such that where there once were few, eventually there are many, and the program's architecture collapses beneath them. (Fran Poretto)

Vaclav_3-Aug-15 15:19

3-Aug-15 15:19

I am going crazy over this simple test code.

C++

char *Result ={"Result init "};
char *Source = {"123"};
char *Destination  = {"456"};

I can verify that all strings are as initialized.

Now I run this strcpy

C++

Result = strcpy(Source,Destination);

Now I can verify, using same approach as before , that
Result = 123 - AKA orignal Source, it was Result init before strcpy was run.
and Source is STILL 123.

Result suppose to be Source ( source destination swap) , but I expect the Source to be COPY of the original Destination AKA 456

What am I doing wrong?
strcpy executes giving correct return, but unmodified Destination.
Am I doing the initialization correctly?

Am I nuts?

Standard stdio using GCC compiler /tools on Arduino IDE.
Thank for your help.
Vaclav

modified 3-Aug-15 22:30pm.

PIEBALDconsult3-Aug-15 16:12

PIEBALDconsult

3-Aug-15 16:12

I think you have the parameter names backward so it's confusing to me.
After the strcpy, Result and Source should contain the same address, the value there should be 465, and the value at Destination should also be 456.
Is that what you see? Is that not what you want?

Be aware that you are not implementing a "swap" at all.

Unless there are other concerns, if you want to "swap" then swap the pointers, not the values:

C++

Result = Destination;
Destination = Source;
Source = Result;

modified 3-Aug-15 23:29pm.

Richard MacCutchan3-Aug-15 20:48

Richard MacCutchan

3-Aug-15 20:48

It should not work at all. String constants declared as above are read-only, so you should not be able to use strcpy to overwrite one of them by another. When I try it I get an access violation on the write, as I would expect. I am not sure why your system does not give the fault, maybe a peculiarity of Arduino. But since strcpy is supposed to return the pointer to the destination field, Result will now point to the data that Source points to, i.e. "123", but the original data should be untouched.

CPallini3-Aug-15 21:08

3-Aug-15 21:08

You cannot do that, it is a mistake. Your variables point to statically allocated (i.e. constant) strings you must NOT try to change.
The following program

#include <stdio.h>
#include <string.h>


int main()
{
  char Result[] ="Result init ";
  char Source[] = "123";
  char Destination[]  = "456";

  strcpy(Destination, Source);

  printf("%s %s\n", Destination, Source);

  return 0;
}

on the other hand, outputs

123 123

Please note, parameter order in strcpy is important (like in any other C function, after all).

Vaclav_4-Aug-15 3:49

4-Aug-15 3:49

Thanks, one of these days I'll hope to learn.
I did change the initialization and it did the trick with the exception of Result.

Can you explain to me why Result needs to be declared this way?

I appreciate all you guys time and I am sorry about the poor choice of variable names I did.
Thanks.

I am enclosing my full test code, just FYI , so do not pay much attention to it.

char *Result = {"0007890"};
char Destination[] = {"456"};
char Temp[] = {" "};
char Source[] = {"123"};

#ifdef DEBUG
lcd_i2c.clear();
//lcd_i2c.print("BuildAttribute...");
lcd_i2c.setCursor(0, 0);
lcd_i2c.print("Result ");
lcd_i2c.print(Result );
//for(;;);
lcd_i2c.setCursor(0, 1);
lcd_i2c.print("Source " );
lcd_i2c.print(Source );
//strcpy(PassedAttribute, "attribute");
lcd_i2c.setCursor(0, 2);
lcd_i2c.print("Dest " );
lcd_i2c.print(Destination );
lcd_i2c.setCursor(0, 3);
//lcd_i2c.print(Attribute);
//delay(2000);
//for(;;);
//return PassedAttribute;
#endif

//Result = strncpy(Source, "TESTDestination",2 );
printf("%s %s\n", Destination, Source);
Result = strcpy(Destination, Source);
printf("%s %s\n", Destination, Source);

//for (;;);
#ifdef DEBUG
lcd_i2c.clear();
//lcd_i2c.print("BuildAttribute...");
lcd_i2c.setCursor(0, 0);
lcd_i2c.print("Result ");
lcd_i2c.print(Result );
//for(;;);
lcd_i2c.setCursor(0, 1);
lcd_i2c.print("Source " );
lcd_i2c.print(Source );
//strcpy(PassedAttribute, "attribute");
lcd_i2c.setCursor(0, 2);
lcd_i2c.print("Dest " );
lcd_i2c.print(Destination );
lcd_i2c.setCursor(0, 3);
lcd_i2c.print("Temp ");
lcd_i2c.print(Temp);
//delay(2000);
for (;;);
//return PassedAttribute;
#endif
for(;;);

CPallini4-Aug-15 6:54

SOLVED Re: strcpy dilemma - it works half way

4-Aug-15 6:54

Result must be declared as pointer to character (char *) because (as already explained by Richard) you are assigning it with strcpy return value. Please note such an assignment is useless since strcpy result is already in Destination array.

Vaclav_4-Aug-15 7:41

4-Aug-15 7:41

Thanks.
I know the results are already known, but I was just using it to learn about the strcpy.