 |
 |
| View previous topic :: View next topic |
| Author |
Message |
soundscu
Joined: 29 Jun 2007
Posts: 62
Location: Raleigh, NC
|
| large arrays of single bits - index limitation |
 Posted: Tue Jan 22, 2008 9:53 am |
 |
|
Hi.
I've created a large array of single bits (within a union):
| Code: | union changeFlags
{
int8 flagByte[66];
int1 flagBit[528];
} dmxChange; |
It seems that an int16 pointer won't count past 255. In this code, only the first 256 bits end up changed, regardless of the passed values for startChan and endChan:
| Code: | void force_change_flags(int16 startChan, int16 endChan, int1 value)
{
int16 i;
for (i=startChan;i<=endChan;i++)
{
dmxChange.flagBit[i] = value;
} |
I really need individual access to 528 elements in this array. Any suggestions?
Thanks! |
|
 |
treitmey
Joined: 23 Jan 2004
Posts: 1094
Location: Appleton,WI USA
|
|
| Posted: Tue Jan 22, 2008 2:09 pm |
|
|
Can you tell us why you need to access it in a union, and not just use a function to change a byte/bit ??
| Code: | #define FIRM_MINOR 8
#define FIRM_MAJOR 'L'
#define HARD_MAJOR 12
#define HARD_MINOR 07
#include <18F4620.h>
//#include <18F4525.h>
#device adc=8
#fuses hs,wdt32,noprotect,nolvp,put
#use delay(clock=18432000,RESTART_WDT)
#use rs232(baud=19200,xmit=PIN_B3,invert,stream=debug)
#use rs232(xmit=PIN_C6,rcv=PIN_C7,enable=PIN_C5,baud=1200,bits=8,parity=e,errors,stream=CIM)
#case
#zero_ram
//========================= Defines ================================//
//========================= Globals =================================//
int8 flagByte[66];
//========================= Prototypes ==============================//
void init(void);
void force_change_flags(int16 startChan, int16 endChan, int1 value);
//========================= Main ====================================//
void main(void)
{
int16 i,j;
init();
i=75;
j=500;
force_change_flags(i, j, TRUE);
force_change_flags(410, 420, 0);
for(i=0;i<sizeof(flagByte);i++){
fprintf(DEBUG,"0x%2X,",flagByte[i]);
restart_wdt();
}
while(1)
{
restart_wdt();
}
}
//========================= Functions ==============================//
//=== init ===// setup the initial settings
void init(void){
output_low(PIN_C5);//turn off the TX_enable
setup_timer_0(RTCC_INTERNAL);
setup_wdt(WDT_ON);
setup_adc_ports(NO_ANALOGS);
setup_adc(ADC_OFF);
set_tris_a(0xFF);set_tris_b(0xFF);set_tris_c(0xFF);set_tris_d(0xFF);set_tris_e(0xFF);
fprintf(DEBUG,"Firm=%C_%03u.c\n\r",FIRM_MAJOR,FIRM_MINOR);
fprintf(DEBUG,"Hard=%2u.%02u\n\r",HARD_MAJOR,HARD_MINOR);
enable_interrupts(GLOBAL);
restart_wdt();
}
void force_change_flags(int16 startChan, int16 endChan, int1 value)
{
int16 curChan,B_idx,b_idx;
for (curChan=startChan;curChan<=endChan;curChan++)
{
restart_wdt();
B_idx=curChan/8;//Byte index
b_idx=curChan%8;//bit index
if(value){
bit_set (flagByte[B_idx],b_idx);
fprintf(DEBUG,"S %lu 0x%2X 0x%2X\n\r",curChan,B_idx,b_idx);
}
else{
bit_clear(flagByte[B_idx],b_idx);
fprintf(DEBUG,"C %lu 0x%2X 0x%2X\n\r",curChan,B_idx,b_idx);
}
//dmxChange.flagBit[i] = value;
}
} |
Last edited by treitmey on Tue Jan 22, 2008 3:13 pm; edited 3 times in total |
|
|
soundscu
Joined: 29 Jun 2007
Posts: 62
Location: Raleigh, NC
|
|
| Posted: Tue Jan 22, 2008 3:01 pm |
|
|
Sure. The union is part of my current workaround -- I've written a function that goes into the same array in byte format and sets an individual bit within a byte, starting with the same 16-bit index value that doesn't work in the previously posted code. This function works fine, so I have a working application, albeit more verbose than I'd like.
Declaring the bit array by itself, not inside a union, does not make any difference -- I can only access the first 256 index positions if the index is a variable.
But if the index is a literal rather than a variable, it works with any value, right up to the required max of 528 elements.
So, my question is: have I made a coding error, or have I found a compiler bug, or is it a known spec that the indices of a bit array can only be 8 bits wide? |
|
|
soundscu
Joined: 29 Jun 2007
Posts: 62
Location: Raleigh, NC
|
|
| Posted: Tue Jan 22, 2008 3:49 pm |
|
|
Thanks, but I already have a workaround function, and it's pretty much what you've written. I can get the job done. :)
My question is about using a bit array in the form
Where index is an int16 variable with a value greater than 255. It doesn't work for me. But it does work if index is a literal. Have I found a compiler bug, or is this normal? |
|
|
Ttelmah
Guest
|
|
| Posted: Wed Jan 23, 2008 5:28 am |
|
|
Bit arrays in CCS, are relatively 'new' (only a few versions ago, we'd have been saying "can't do this"). Now, historically, CCS, for a long while had a limit on arrays, only allowing int (int8) indexes. This has been removed on latter versions, so the question is "has this limit 'come back', when dealing with int1 arrays"?.
Looking at the assembler generated for an int1 array access, shows that if you access a single bit in such an array, using a constant index, it does handle values beyond 255, so:
| Code: |
.................... bitarray[0x10]=0;
01CA: BCF 07.0
.................... bitarray[0x110]=0;
01CC: BCF 27.0
|
However if you do the same, with an int16 variable (in this case, counting through 512 entries), you get:
| Code: |
.................... for (count=0;count<512;count++)
01A8: CLRF 48
01AA: CLRF 47
01AC: MOVF 48,W
01AE: SUBLW 01
01B0: BNC 01CA
.................... bitarray[count]=1;
01B2: MOVFF 47,49
01B6: MOVLW 01
01B8: MOVWF 4A
01BA: CLRF 4C
01BC: MOVLW 05
01BE: MOVWF 4B
01C0: BRA 0004
01C2: INCF 47,F
01C4: BTFSC FD8.2
01C6: INCF 48,F
01C8: BRA 01AC
|
If you look carefully at the start of the section accessing the array, you will see that it takes the LSB of the 16bit value (stored in 47), and transfers it to a temporary variable, prior to calling the access routine (here at address 4), but it does not take the MSB at all for the access (only talks to this for the increment, and tests for the loop...).
So it appears as if the bit array code, at present is limited to 8bit, when variables are used.
I'd 'moan' slightly at CCS, since this should at the very least be documented, if not fixed....
Best Wishes |
|
|
soundscu
Joined: 29 Jun 2007
Posts: 62
Location: Raleigh, NC
|
|
| Posted: Wed Jan 23, 2008 7:04 am |
|
|
| Thanks! That's the confirmation I was hoping to find. I have contacted CCS. |
|
|
soundscu
Joined: 29 Jun 2007
Posts: 62
Location: Raleigh, NC
|
|
| Posted: Thu Jan 24, 2008 1:05 pm |
|
|
CCS reports that they have fixed my problem and it will be in the next compiler release. :)
Jim |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
