Preparing MATLAB code for code generation to C code using MATLAB Coder. Error messages.

I have been writing a program which reads in all the characters from a text file. It then searches for a specific string within this character array, this is a register address. Then I want to take the following 8 characters, this is the register value, and use this in future.

The register value is a 8 character representation of a hexadecimal number. I need to convert this to a 32-bit binary number.

I understand that the read function doesn't read a specific number of characters, just keeps reading until the end of the file, hence the size of REGISTERS is [Inf:Inf]. Then the strfind function could in theory find the sub-string (address in this case) more than once. In my particular case the sub-string will only appear once, but it still has size [1x:?]. The problem with that is that I then try to reference to part of REGISTERS but it doesn't like it because the position offset is not a scalar.

How can I force these variables to behave correctly? I have tried defining them before using them, but I've had no success yet.

Thanks and here is my code:

% open file custom.txt and read all data into REGISTERS
filename = 'custom.txt';
fileID = fopen(filename,'r');
frewind(fileID);
REGISTERS = fread(fileID,[1,Inf],'char=>char');
fclose(fileID);
% Search for the register value at register address 'fa80005c'
DPASS_CONF = Get_Reg_Value_ng('fa80005c',REGISTERS);

The code for Get_Reg_Value() is as follows:

function [ REG_VAL ] = Get_Reg_Value_ng( REG_ADDR, REGISTERS )%#codegen 
%%Initial Values
k = 0;
REG_VALh = char(zeros(1,8));
REG_VALd = char(zeros(1,9));
REG_VAL = char(zeros(1,32));
k = strfind(REGISTERS,REG_ADDR);
k = k_(1);
REG_VALh = REGISTERS(k+9:k+17);
REG_VALd = char(hex2dec(REG_VALh));
REG_VAL = dec2bin(REG_VALd);

Ensure that the register value is passed as 32bits:

If REG_VAL is not 32 bits long...

if length(REG_VAL) < 32

Then right justify the value in a 32 bit variable.

  REG_VAL = (cat(2,zeros(1,32-length(REG_VAL)),REG_VAL));
elseif length(REG_VAL) > 32
  REG_VAL = REG_VAL(1:32);
end
end

The error I get is 'Expected a scalar value. This expression has size [1x:?].' This is an error in the line 'REG_VALh = REGISTERS(k+9:k+17);' and it seems to think the 'k+9' bit is the problem. Then I get a load of errors due to REG_VALh not being defined.

Answer 1

[Chinmayi Lanka]

On attempting to generate code for your function, I did not receive the error you mentioned.

I believe that it is a typo but ensure that the indexing is correct on line 9:

>> k = k(1);

Further, on line 11, there is no need to convert the result of 'hex2dec' to char. It should be:

>> REG_VALd = hex2dec(REG_VALh);