How to do reverse memcmp? - c

How can I do reverse memory comparison? As in, I give the ends of two sequences and I want the pointer to be decremented towards the beginning, not incremented towards the end.

There's no built-in function in the C standard library to do it. Here's a simple way to roll your own:
int memrcmp(const void *s1, const void *s2, size_t n)
{
if(n == 0)
return 0;
// Grab pointers to the end and walk backwards
const unsigned char *p1 = (const unsigned char*)s1 + n - 1;
const unsigned char *p2 = (const unsigned char*)s2 + n - 1;
while(n > 0)
{
// If the current characters differ, return an appropriately signed
// value; otherwise, keep searching backwards
if(*p1 != *p2)
return *p1 - *p2;
p1--;
p2--;
n--;
}
return 0;
}
If you something high-performance, you should compare 4-byte words at a time instead of individual bytes, since memory latency will be the bottleneck; however, that solution is significantly more complex and not really worth it.

Much like in a post (C memcpy in reverse) originally linked by Vlad Lazarenko, here is a solution based on that, that I haven't yet tested but should get you started.
int reverse_memcmp(const void *s1, const void *s2, size_t n)
{
unsigned char *a, *b;
a = s1;
b = s2;
size_t i = 0;
// subtracting i from last position and comparing
for (i = 0; i < n; i++) {
if (a[n-1-i] != b[n-1-i]) {
// return differences between different byte, strcmp()-style
return (a[n-1-i] - b[n-1-i]);
}
}
return 0;
}

All you need to do is specify your two ends and the size that you'd like to compare, as well as a step size. Please note especially that the step size may be the most important part for getting the expected results. It will greatly ease the implementation if you restrict the sizes. For the size of a char you could do something like:
int compare (void *one, void *two, size_t size)
{
char *one_char = (char *)one;
char *two_char = (char *)two;
size_t i;
for (i = 0; i < size; i++)
{
if (*(one_char - i) != *(two_char - i))
return(NOT_EQUAL);
}
return(EQUAL);
}

shorter code (C code doesn't need force pointer type cast):
int reverse_memcmp(const void *end1, const void *end2, size_t n) {
const unsigned char *a = end1, *b = end2;
for (; n; --n)
if (*--a != *--b) return *a - *b;
return 0;
}

Related

exercise 9.7 Kochan. strange output

I am working to learn C using Kochan's Programming in C 4th edition. problem 9.7 the goal is to insert a string of characters into another array. I am supposed to write a function to accomplish this. I have two problems.
When I have the algorithm print the result as it goes through the if statements, it produces the desired output, however when I change it to an %s, I only get a partial output. My hunch is that a null character is being placed where i do not want it, but I simply cannot see it.
To see what was happening, I added a printf that would track the letter and the array space it was occupying. I was surprised to see that the first letter was not 0, but was blank, and the next letter was assigned the 0. Any insight into this would be appreciated.
The funtion of interest is "insertString".
#include <stdio.h>
#include <stdbool.h>
char x[] = {"the wrong son was shot that day"};
char text[] = {"per"};
int countString (char x[])
{
int counter, z;
for (counter = 0; x[counter] != '\0'; ++counter)
z = counter+1;
return z;
}
void insertString (char text[],char x[],int n) //source, text to input, where
{
int count, clock, i = countString(text), q = countString(x);
int counter = 0;
char y[i + q];
for(count = 0; x[count] != '\0'; ++count){
if (count < n){
y[count] = x[count];
printf("%c %i", y[count], count); //The integer call is just to put a number next to the
//letter. This is where my second issue is shown.
}
else if (counter <= i){
y[count] = text[counter];
++counter;
printf("%c", y[count]);
}
else{
y[count]= x[count - counter];
printf("%c", y[count]);
}
}
printf("\n\n");
y[count-counter] = '\0';
printf("%s", y);
}
int main (void)
{
void insertString(char text[], char x[], int i);
int countString(char x[]);
int i;
insertString(text, x, 10);
return 0;
}
10 out of 10 times I post here it is because im doing something dumb, so I use SO as an absolute last resort if i am getting into the territory of just randomly trying stuff with no methodology. Thanks for your patience in advance.
Your condition is wrong in the for. It should be x[count - counter] != '\0'
In the second condition use just < to avoid overindexing. (else if (counter < i))
You put the terminating NULL char at wrong place. You should do this: y[count] = '\0'
printf inside a string routine like this is fine for debugging, but it's a poor way to write a general-purpose function because it makes it impossible to use its output for further programmatic manipulation. It can also make it difficult to reason about how the state of the function interacts in unpredictable ways with the state of the printed data.
I assume you haven't learned about dynamic memory allocation which is a prerequisite to returning strings from functions. You can inline the function logic into main or printf only at the end of the function in the meantime.
Adding to this point, a void function would need to reallocate space in the string to insert into and would be in-place. This seems likely less generally useful than allocating a new string to hold the result.
Using global variables like char x[] when there's no need is poor practice. It's better to put those strings scoped to main. Since your function can access these variables in addition to its parameters, confusion can ensue when scope and encapsulation is breached.
Use consistent formatting and avoid variable names like q that mean virtually nothing. Instead of adding comments to explain poor var names:
void insertString (char text[],char x[],int n) //source, text to input, where
You can simply name the variables exactly what they represent:
void insertString(char *dest, char *source, int add_index)
Also, now that you've mastered countString, you can abstract this by calling the builtin strlen.
Be sure to allocate enough space in buffers: char y[i + q]; should be y[i+q+1] to allow room for the null terminator '\0'.
As for the logic, I think it's easier to break into three loops without conditions instead of one loop with conditions. This makes it easier to break the problem down into the three constituent steps:
Add everything up until add_index from the dest string to the result.
Add everything in the source string to the result.
Add everything after add_index from the dest string to the result.
Using this approach, all that's left is figuring out how to map the indexes appropriately. Here it is in code:
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
char *insert_string(char *dest, char *source, int add_index) {
int source_len = strlen(source);
int dest_len = strlen(dest);
int result_size = source_len + dest_len + 1;
char *result = malloc(result_size);
for (int i = 0; i < add_index; i++) {
result[i] = dest[i];
}
for (int i = 0; i < source_len; i++) {
result[i+add_index] = source[i];
}
for (int i = add_index; i < dest_len; i++) {
result[i+add_index] = dest[i];
}
result[result_size-1] = '\0';
return result;
}
int main(void) {
char *result = insert_string("hello world", "cruel ", 6);
printf("%s\n", result);
free(result);
return 0;
}
Although this is likely for instructional purposes, these operations can be abstracted further using builtin string functions like strncpy and sprintf.
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
char *insert_string(char *dest, char *source, int add_index) {
int result_size = strlen(dest) + strlen(source) + 1;
char *result = malloc(result_size);
char pre[add_index+1];
pre[add_index] = '\0';
strncpy(pre, dest, add_index);
sprintf(result, "%s%s%s", pre, source, dest + add_index);
return result;
}
int main(void) {
char *result = insert_string("hello world", "cruel ", 6);
printf("%s\n", result);
free(result);
return 0;
}
Doing this in-place is more straightforward. Since the result already has the prefix, you can copy the destination postfix to create a source-sized gap in the middle and then overwrite the gap using the source string. It's up to the caller to make sure that the destination buffer is large enough to hold the insertion.
#include <stdio.h>
#include <string.h>
void insert_string(char *dest, char *source, int add_index) {
int source_len = strlen(source);
int dest_len = strlen(dest);
for (int i = add_index; i < dest_len; i++) {
dest[i+add_index] = dest[i];
}
for (int i = 0; i < source_len; i++) {
dest[i+add_index] = source[i];
}
}
int main(void) {
// allocate extra space in the string to hold the insertion
char greeting[32] = "hello world";
insert_string(greeting, "cruel ", 6);
printf("%s\n", greeting);
return 0;
}
A note of caution: none of these functions handle errors at all, so they're unsafe. Correct functions should check that the add_index falls within the bounds of the dest string. This is an exercise for the reader.
The original exercise is here:
Your function is not doing it. You need to insert the string into another string not to create a new one with both mixed. You can do it this way of course and then copy it into the original one - but it is the most uneficient way to archive it (memory & timewise).
Use the correct types.
size_t mystrlen(const char *str)
{
const char *end = str;
while(*end++);
return end - str - 1;
}
char *strinsert(char *dest, size_t pos, const char *istr)
{
char *temp = dest, *work;
size_t ilen = mystrlen(istr);
size_t nmove;
while(*temp) temp++;
nmove = temp - dest - pos + 1;
work = temp;
temp += ilen;
while(nmove--) *temp-- = *work--;
work = dest + pos;
while(*istr) *work++ = *istr++;
return dest;
}
int main()
{
char dest[128] = "0123456789012345678901234567890123456789";
printf("%s", strinsert(dest, 7, "ABCD"));
}
https://godbolt.org/z/KMnLU2

comparing if void * contains 0 numbytes?

I need generic way to check if void * contains 0 till num_bytes.
I came up with following approach. *p does not contain same type of data everytime hence cant do *(type*)p
bool is_pointer_0(void *p, int num) {
void *cmp;
cmp = (void*)malloc(num);
memset(cmp, 0, num);
if (memcmp(p, cmp, num)) {
free(cmp);
return false;
} else {
free(cmp);
return true;
}
}
The function allocates & frees up num bytes on every call, not pretty I think.
Please suggest faster approaches. Appreciate the help.
Update :
How about this approach ?
bool is_pointer_0(void *p, int num) {
void *a = NULL;
memcpy(&a, p, num);
return a == NULL;
}
This code casts the void pointer to a char pointer. This allows the memory pointed at to be treated as a sequence of bytes. Then cycles through the specified length looking for non zero bytes. I do not know if the standards guarantee that this would work (ie casting a void* to a char* will provide a pointer to the raw bytes), but in real life it works
bool is_pointer_0(void *p, int num) {
char *c = (char *)p;
for(int i = 0; i < num; i++)
if(c[i]) return false;
return true;
}
You can cast the pointer to a char* or unsigned char* and check the values of the elements.
unsigned char* cp = reinterpret_cast<unsigned char*>(p);
for (int i = 0; i < num; ++i )
{
if ( cp[i] != 0 )
{
return false;
}
}
return true;
Note: this approach may be better for long buffers well aligned. Yet this answer is fast due to simplicity.
Since the memory, if all zeros, must compare to itself, use memcmp(): a platform specific optimized function.
int memcmp0(const void *buf, size_t n) {
#define TESTVALUE 0
const char *cbuf = (const char *) buf;
while (n > 0) {
// If odd size, last byte not 0?
if (n % 2 && (cbuf[n - 1] != TESTVALUE)) return 0;
// 1st half matches 2nd half?
size_t half = n / 2;
if(memcmp(cbuf, &cbuf[half], half) != 0) return 0;
n = half;
}
return 1;
}
This is easy to extend to other values other than 0 by changing TESTVALUE.
Note: at most log2(n) iterations.
Another after accept answer:
Since the memory, if all zeros, must compare to itself, use memcmp(): a platform specific optimized function.
Check the first value and then use memcmp() to compare ptr[0],ptr[1], then ptr[1],ptr[2], then ptr[2],ptr[3], etc.
int memcmpfast(const void *ptr, size_t n, char testvalue) {
const char *cptr = (const char *) ptr;
if (n == 0) return 1;
if (cptr[0] != testvalue) return 0;
return memcmp(cptr, cptr + 1, n - 1) == 0;
}
I'd probably go with something like this:
bool is_pointer_0(void* p, int num)
{
return std::search_n((char*)p, (char*)p + num, num, 0) == p;
}
Or this:
bool is_pointer_0(void* p, int num)
{
return std::all_of((char*)p, (char*)p + num, [](char c){return c == 0;});
}

Convert float 32 bit big endian to little endian

My data is a 2d 32 bit 2-D pointer image which is in big endian and I wish to convert in to little endian. Please check the following code:
float** ReverseFloat( const float **inFloat )
{
float **retVal;
char *floatToConvert = ( char* ) && inFloat;
char *returnFloat = ( char* ) && retVal;
// swap the bytes into a temporary buffer
returnFloat[0] = floatToConvert[3];
returnFloat[1] = floatToConvert[2];
returnFloat[2] = floatToConvert[1];
returnFloat[3] = floatToConvert[0];
return retVal;
}
Hi. As per all of your advice, I tried the following code but it gives all zero values:
float **swapend( float **in ,float **out , int h, int v)
{
char *floatToConvert = ( char* ) & in;
char *returnFloat = ( char *) & out;
for (int m=0; m<h; m++)
for (int k=0; k<v; k++)
{
for (int i=0; i<sizeof(float); i++)
returnFloat[sizeof(float)-1-i] = floatToConvert[i];
}
return out;
}
Thank you.
This won't work reliably. You create a local variable retVal and at the end, you return it. The value of this variable will eventually be random after the function ends.
I think this approach would be better since the compiler will do most of the necessary copying:
float be2le( char * floatToConvert ) {
float buffer;
char * returnFloat = (char *) &buffer;
...swap...
return buffer;
}
&& is the logical and operator (it may also carry a pointer like semantic as a gnu extension, but this is non portable), while & is the "address of” operator. You don't need to use pointers of pointers either:
float ReverseFloat(float inFloat ){
float retVal;
char *floatToConvert = ( char* ) &inFloat;
char *returnFloat = ( char* ) &retVal;
// swap the bytes into a temporary buffer
returnFloat[0] = floatToConvert[3];
returnFloat[1] = floatToConvert[2];
returnFloat[2] = floatToConvert[1];
returnFloat[3] = floatToConvert[0];
return retVal;
}
Note that this function works both ways: it let you convert from one endianness to the other.
Following your update: if you wish to convert several values contained in an array, just apply this function on each of them sequentially:
void ReverseFloatArray(unsigned int count, float *in, float *out){
unsigned int i;
for (i = 0; i < count; i++)
out[i] = ReverseFloat(in[i]);
}

string_comparator in C

Okay so I need to several quite long strings in C. So I say to myself "why, you'd better use that handy dandy qsort function! Better write yourself a string_comparator for it!"
So of course I do and here she is:
int string_comparator(const void* el1, const void* el2) {
char* x = (char*) el1;
char* y = (char*) el2;
int str_len = strlen(x);
int i = 0;
for (; i < str_len; i++) {
//when there are non-equal chars
if (x[i] != y[i]) {
break;
}
}
return x[i] - y[i];
}
So of course I pass my handy dandy string_comparator function to the C qsort function as such:
qsort(list.words, list.num_words, sizeof(char*), string_comparator);
list is a struct that holds a char** (words) and ints which refer to the number of words held by it (such as num_words)
Now I have the problem where my list is not getting sorted alphabetically like I had hoped! I put a bunch of printf statements in my comparator and it printed out garbage values for the strings every time so I'm fairly sure that is the problem. But why is that the problem?? I've used qsort before (never to sort words..just sorting characters) and from what I understand this should work...What's going wrong here?
I appreciate any suggestions!
This is a common mistake when using qsort(). Here are the corrections:
char *x = *(char **) el1;
char *y = *(char **) el2;
Because list.words has type char **, not type char *, right?
Another example of qsort()
Here's how you sort an array of int with qsort():
int int_comparator(const void *el1, const void *el2)
{
int x = *(int *) el1;
int y = *(int *) el2;
return x - y;
}
void sort_ints(int *a, size_t n)
{
// these two lines are both "correct"
// the second line is more "obviously correct"
// qsort(a, n, sizeof(int), int_comparator);
qsort(a, n, sizeof(*a), int_comparator);
}
Now, if you go through and replace int with char *, you have to replace int * with char **.

What is the proper way of implementing a good "itoa()" function?

I was wondering if my implementation of an "itoa" function is correct. Maybe you can help me getting it a bit more "correct", I'm pretty sure I'm missing something. (Maybe there is already a library doing the conversion the way I want it to do, but... couldn't find any)
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
char * itoa(int i) {
char * res = malloc(8*sizeof(int));
sprintf(res, "%d", i);
return res;
}
int main(int argc, char *argv[]) {
...
// Yet, another good itoa implementation
// returns: the length of the number string
int itoa(int value, char *sp, int radix)
{
char tmp[16];// be careful with the length of the buffer
char *tp = tmp;
int i;
unsigned v;
int sign = (radix == 10 && value < 0);
if (sign)
v = -value;
else
v = (unsigned)value;
while (v || tp == tmp)
{
i = v % radix;
v /= radix;
if (i < 10)
*tp++ = i+'0';
else
*tp++ = i + 'a' - 10;
}
int len = tp - tmp;
if (sign)
{
*sp++ = '-';
len++;
}
while (tp > tmp)
*sp++ = *--tp;
return len;
}
// Usage Example:
char int_str[15]; // be careful with the length of the buffer
int n = 56789;
int len = itoa(n,int_str,10);
The only actual error is that you don't check the return value of malloc for null.
The name itoa is kind of already taken for a function that's non-standard, but not that uncommon. It doesn't allocate memory, rather it writes to a buffer provided by the caller:
char *itoa(int value, char * str, int base);
If you don't want to rely on your platform having that, I would still advise following the pattern. String-handling functions which return newly allocated memory in C are generally more trouble than they're worth in the long run, because most of the time you end up doing further manipulation, and so you have to free lots of intermediate results. For example, compare:
void delete_temp_files() {
char filename[20];
strcpy(filename, "tmp_");
char *endptr = filename + strlen(filename);
for (int i = 0; i < 10; ++i) {
itoa(endptr, i, 10); // itoa doesn't allocate memory
unlink(filename);
}
}
vs.
void delete_temp_files() {
char filename[20];
strcpy(filename, "tmp_");
char *endptr = filename + strlen(filename);
for (int i = 0; i < 10; ++i) {
char *number = itoa(i, 10); // itoa allocates memory
strcpy(endptr, number);
free(number);
unlink(filename);
}
}
If you had reason to be especially concerned about performance (for instance if you're implementing a stdlib-style library including itoa), or if you were implementing bases that sprintf doesn't support, then you might consider not calling sprintf. But if you want a base 10 string, then your first instinct was right. There's absolutely nothing "incorrect" about the %d format specifier.
Here's a possible implementation of itoa, for base 10 only:
char *itobase10(char *buf, int value) {
sprintf(buf, "%d", value);
return buf;
}
Here's one which incorporates the snprintf-style approach to buffer lengths:
int itobase10n(char *buf, size_t sz, int value) {
return snprintf(buf, sz, "%d", value);
}
A good int to string or itoa() has these properties;
Works for all [INT_MIN...INT_MAX], base [2...36] without buffer overflow.
Does not assume int size.
Does not require 2's complement.
Does not require unsigned to have a greater positive range than int. In other words, does not use unsigned.
Allows use of '-' for negative numbers, even when base != 10.
Tailor the error handling as needed. (needs C99 or later):
char* itostr(char *dest, size_t size, int a, int base) {
// Max text needs occur with itostr(dest, size, INT_MIN, 2)
char buffer[sizeof a * CHAR_BIT + 1 + 1];
static const char digits[36] = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ";
if (base < 2 || base > 36) {
fprintf(stderr, "Invalid base");
return NULL;
}
// Start filling from the end
char* p = &buffer[sizeof buffer - 1];
*p = '\0';
// Work with negative `int`
int an = a < 0 ? a : -a;
do {
*(--p) = digits[-(an % base)];
an /= base;
} while (an);
if (a < 0) {
*(--p) = '-';
}
size_t size_used = &buffer[sizeof(buffer)] - p;
if (size_used > size) {
fprintf(stderr, "Scant buffer %zu > %zu", size_used , size);
return NULL;
}
return memcpy(dest, p, size_used);
}
I think you are allocating perhaps too much memory. malloc(8*sizeof(int)) will give you 32 bytes on most machines, which is probably excessive for a text representation of an int.
i found an interesting resource dealing with several different issues with the itoa implementation
you might wanna look it up too
itoa() implementations with performance tests
I'm not quite sure where you get 8*sizeof(int) as the maximum possible number of characters -- ceil(8 / (log(10) / log(2))) yields a multiplier of 3*. Additionally, under C99 and some older POSIX platforms you can create an accurately-allocating version with sprintf():
char *
itoa(int i)
{
int n = snprintf(NULL, 0, "%d", i) + 1;
char *s = malloc(n);
if (s != NULL)
snprintf(s, n, "%d", i);
return s;
}
HTH
You should use a function in the printf family for this purpose. If you'll be writing the result to stdout or a file, use printf/fprintf. Otherwise, use snprintf with a buffer big enough to hold 3*sizeof(type)+2 bytes or more.
sprintf is quite slow, if performance matters it is probably not the best solution.
if the base argument is a power of 2 the conversion can be done with a shift and masking, and one can avoid reversing the string by recording the digits from the highest positions. For instance, something like this for base=16
int num_iter = sizeof(int) / 4;
const char digits[] = {'0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f'};
/* skip zeros in the highest positions */
int i = num_iter;
for (; i >= 0; i--)
{
int digit = (value >> (bits_per_digit*i)) & 15;
if ( digit > 0 ) break;
}
for (; i >= 0; i--)
{
int digit = (value >> (bits_per_digit*i)) & 15;
result[len++] = digits[digit];
}
For decimals there is a nice idea to use a static array big enough to record the numbers in the reversed order, see here
Integer-to-ASCII needs to convert data from a standard integer type
into an ASCII string.
All operations need to be performed using pointer arithmetic, not array indexing.
The number you wish to convert is passed in as a signed 32-bit integer.
You should be able to support bases 2 to 16 by specifying the integer value of the base you wish to convert to (base).
Copy the converted character string to the uint8_t* pointer passed in as a parameter (ptr).
The signed 32-bit number will have a maximum string size (Hint: Think base 2).
You must place a null terminator at the end of the converted c-string Function should return the length of the converted data (including a negative sign).
Example my_itoa(ptr, 1234, 10) should return an ASCII string length of 5 (including the null terminator).
This function needs to handle signed data.
You may not use any string functions or libraries.
.
uint8_t my_itoa(int32_t data, uint8_t *ptr, uint32_t base){
uint8_t cnt=0,sgnd=0;
uint8_t *tmp=calloc(32,sizeof(*tmp));
if(!tmp){exit(1);}
else{
for(int i=0;i<32;i++){
if(data<0){data=-data;sgnd=1;}
if(data!=0){
if(data%base<10){
*(tmp+i)=(data%base)+48;
data/=base;
}
else{
*(tmp+i)=(data%base)+55;
data/=base;
}
cnt++;
}
}
if(sgnd){*(tmp+cnt)=45;++cnt;}
}
my_reverse(tmp, cnt);
my_memcopy(tmp,ptr,cnt);
return ++cnt;
}
ASCII-to-Integer needs to convert data back from an ASCII represented string into an integer type.
All operations need to be performed using pointer arithmetic, not array indexing
The character string to convert is passed in as a uint8_t * pointer (ptr).
The number of digits in your character set is passed in as a uint8_t integer (digits).
You should be able to support bases 2 to 16.
The converted 32-bit signed integer should be returned.
This function needs to handle signed data.
You may not use any string functions or libraries.
.
int32_t my_atoi(uint8_t *ptr, uint8_t digits, uint32_t base){
int32_t sgnd=0, rslt=0;
for(int i=0; i<digits; i++){
if(*(ptr)=='-'){*ptr='0';sgnd=1;}
else if(*(ptr+i)>'9'){rslt+=(*(ptr+i)-'7');}
else{rslt+=(*(ptr+i)-'0');}
if(!*(ptr+i+1)){break;}
rslt*=base;
}
if(sgnd){rslt=-rslt;}
return rslt;
}
I don't know about good, but this is my implementation that I did while learning C
static int ft_getintlen(int value)
{
int l;
int neg;
l = 1;
neg = 1;
if (value < 0)
{
value *= -1;
neg = -1;
}
while (value > 9)
{
l++;
value /= 10;
}
if (neg == -1)
{
return (l + 1);
}
return (l);
}
static int ft_isneg(int n)
{
if (n < 0)
return (-1);
return (1);
}
static char *ft_strcpy(char *dest, const char *src)
{
unsigned int i;
i = 0;
while (src[i] != '\0')
{
dest[i] = src[i];
i++;
}
dest[i] = src[i];
return (dest);
}
char *ft_itoa(int n)
{
size_t len;
char *instr;
int neg;
neg = ft_isneg(n);
len = ft_getintlen(n);
instr = (char *)malloc((sizeof(char) * len) + 1);
if (n == -2147483648)
return (ft_strcpy(instr, "-2147483648"));
if (!instr)
return (NULL);
if (neg == -1)
n *= -1;
instr[len--] = 0;
if (n == 0)
instr[len--] = 48;
while (n)
{
instr[len--] = ((n % 10) + 48);
n /= 10;
}
if (neg == -1)
instr[len] = '-';
return (instr);
}
This should work:
#include <string.h>
#include <stdlib.h>
#include <math.h>
char * itoa_alloc(int x) {
int s = x<=0 ? 1 ? 0; // either space for a - or for a 0
size_t len = (size_t) ceil( log10( abs(x) ) );
char * str = malloc(len+s + 1);
sprintf(str, "%i", x);
return str;
}
If you don't want to have to use the math/floating point functions (and have to link in the math libraries) you should be able to find non-floating point versions of log10 by searching the Web and do:
size_t len = my_log10( abs(x) ) + 1;
That might give you 1 more byte than you needed, but you'd have enough.
There a couple of suggestions I might make. You can use a static buffer and strdup to avoid repeatedly allocating too much memory on subsequent calls. I would also add some error checking.
char *itoa(int i)
{
static char buffer[12];
if (snprintf(buffer, sizeof(buffer), "%d", i) < 0)
return NULL;
return strdup(buffer);
}
If this will be called in a multithreaded environment, remove "static" from the buffer declaration.
This is chux's code without safety checks and the ifs. Try it online:
char* itostr(char * const dest, size_t const sz, int a, int const base) {
bool posa = a >= 0;
char buffer[sizeof a * CHAR_BIT + 1];
static const char digits[36] = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ";
char* p = &buffer[sizeof buffer - 1];
do {
*(p--) = digits[abs(a % base)];
a /= base;
} while (a);
*p = '-';
p += posa;
size_t s = &buffer[sizeof(buffer)] - p;
memcpy(dest, p, s);
dest[s] = '\0';
return dest;
}
main()
{
int i=1234;
char stmp[10];
#if _MSC_VER
puts(_itoa(i,stmp,10));
#else
puts((sprintf(stmp,"%d",i),stmp));
#endif
return 0;
}

Resources