Splitting a sequence of binary digits based on inputs in c

Splitting a sequence of binary digits based on inputs in c



I have an array of binary digits that represent an address. I want to now segregate this address into 3 parts or arrays that are tag, index and offset and the number of bits for each of these three are determined by the user.
For better clarity, this is the code:


tag


index


offset


char* bin_val_addr;
while(fscanf(FP, "%s %lx", str, &addr) != EOF)


bin_val_addr = long_to_binary(addr);

char* array_Tag = malloc(no_Tag_bits * sizeof(char*));
char* array_Index = malloc(no_Index_bits * sizeof(char*));
memcpy(array_Tag, bin_val_addr, no_Tag_bits * sizeof(char*));
memcpy(array_Index, bin_val_addr, no_Index_bits * sizeof(char*));
printf("TAG array is: %sn", array_Tag);
printf("INDEX array is: %sn", array_Index);





And the function long_to_binary is as follows:


char* long_to_binary(unsigned long k)

static char c[65];
c[0] = '';

unsigned long val;
for (val = 1UL << (sizeof(unsigned long)*8-1); val > 0; val >>= 1)

strcat(c, ((k & val) == val) ? "1" : "0");

return c;



I looked up ways to split arrays and found that memcpy can be used but I am going wrong somewhere while implementing this. Can someone help me out here?


memcpy






If you want to store, for example, 1010 in a string then that string will need to be at least 5 characters in length (4 to hold 1010 and 1 to hold the string's null terminator). Your code is incorrectly allocating 4 char pointers, when it should allocate 5 chars. There are other problems here too, for example the way you are returning 'c' from the long_to_binary() function needs to change.

– jarmod
Sep 16 '18 at 22:00







okay so I'd have to change the sizeof(char*) statement right?

– Sam
Sep 16 '18 at 22:02


sizeof(char*)






Yes, if you want to allocate a string that can hold N characters then you need to malloc(N+1), rather than malloc(N * sizeof(char *)).

– jarmod
Sep 16 '18 at 22:04






And why is the way I return c wrong?

– Sam
Sep 16 '18 at 22:04






Ah, it's static. Did not notice that earlier. It's OK in that case (I'm not a big fan but it will work).

– jarmod
Sep 16 '18 at 22:05




1 Answer
1



If bin_val_addr contains the sequence of 1s and 0s in the order tag, index and offset, then the following copies from the beginning of bin_val_addr for both array_Tag and array_Index, e.g.


bin_val_addr


1s


0s


tag


index


offset


bin_val_addr


array_Tag


array_Index


memcpy(array_Tag, bin_val_addr, no_Tag_bits * sizeof(char*));
memcpy(array_Index, bin_val_addr, no_Index_bits * sizeof(char*));



If array_Index begins after array_Tag in bin_val_addr, then you would need to offset your source for array_Index by no_Tag_bits.


array_Index


array_Tag


bin_val_addr


array_Index


no_Tag_bits



(and note you are mixing bits and bytes (you have characters '1' and '0' not bits). So, unless you need 8-bytes per-character, sizeof(char*) should be sizeof(char) -- which is always 1 and should be eliminated from your allocations)


'1'


'0'


sizeof(char*)


sizeof(char)


1



For example, allocating to hold the characters specified in each array (+1 for the nul-terminating character), you could do:


+1


char *array_Tag = malloc(no_Tag_bits + 1);
char *array_Index = malloc(no_Index_bits + 1);



and you must validate each allocation, because malloc can and will fail returning NULL on memory exhaustion, e.g.


malloc


NULL


if (!array_Tag)
perror ("malloc-array_Tag");
exit (EXIT_FAILURE);



Putting those pieces together and nul-terminating each string, you would have:


memcpy (array_Tag, bin_val_addr, no_Tag_bits);
array_Tag[no_Tag_bits] = 0;
memcpy (array_Index, bin_val_addr + no_Tag_bits, no_Index_bits);
array_Index[no_Index_bits] = 0;



Which would now allow you to output the arrays as strings:


printf ("TAG array is: %sn", array_Tag);
printf ("INDEX array is: %sn", array_Index);



Lastly, your use of fscanf invites Undefined Behavior on a matching or input failure where less that the total number of conversions specified takes place. Instead of checking against EOF, you should validate with the number of conversions specified actually occurred, e.g.


fscanf


EOF


while (fscanf (FP, "%s %lx", str, &addr) == 2)



After making those changes, and presuming your long_to_binary is working as you intend, let me know if you have further questions.


long_to_binary



Thanks for contributing an answer to Stack Overflow!



But avoid



To learn more, see our tips on writing great answers.



Required, but never shown



Required, but never shown




By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Popular posts from this blog

𛂒𛀶,𛀽𛀑𛂀𛃧𛂓𛀙𛃆𛃑𛃷𛂟𛁡𛀢𛀟𛁤𛂽𛁕𛁪𛂟𛂯,𛁞𛂧𛀴𛁄𛁠𛁼𛂿𛀤 𛂘,𛁺𛂾𛃭𛃭𛃵𛀺,𛂣𛃍𛂖𛃶 𛀸𛃀𛂖𛁶𛁏𛁚 𛂢𛂞 𛁰𛂆𛀔,𛁸𛀽𛁓𛃋𛂇𛃧𛀧𛃣𛂐𛃇,𛂂𛃻𛃲𛁬𛃞𛀧𛃃𛀅 𛂭𛁠𛁡𛃇𛀷𛃓𛁥,𛁙𛁘𛁞𛃸𛁸𛃣𛁜,𛂛,𛃿,𛁯𛂘𛂌𛃛𛁱𛃌𛂈𛂇 𛁊𛃲,𛀕𛃴𛀜 𛀶𛂆𛀶𛃟𛂉𛀣,𛂐𛁞𛁾 𛁷𛂑𛁳𛂯𛀬𛃅,𛃶𛁼

Edmonton

Crossroads (UK TV series)