Writing own memmem for Windows - c

I notcied that memmem is not available in MSVC for Windows, so I tried to write something for it. I have the following code:
void *memmem(const void *haystack_start, size_t haystack_len, const void *needle_start, size_t needle_len)
{
const unsigned char *haystack = (const unsigned char *)haystack_start;
const unsigned char *needle = (const unsigned char *)needle_start;
const unsigned char *h = NULL;
const unsigned char *n = NULL;
size_t x = needle_len;
/* The first occurrence of the empty string is deemed to occur at
the beginning of the string. */
if (needle_len == 0) {
return (void *)haystack_start;
}
/* Sanity check, otherwise the loop might search through the whole
memory. */
if (haystack_len < needle_len) {
return NULL;
}
for (; *haystack && haystack_len--; haystack++) {
x = needle_len;
n = needle;
h = haystack;
if (haystack_len < needle_len)
break;
if ((*haystack != *needle) || (*haystack + needle_len != *needle + needle_len))
continue;
for (; x; h++, n++) {
x--;
if (*h != *n)
break;
if (x == 0)
return (void *)haystack;
}
}
return NULL;
}
But, I do not think it works correctly. If I try something like this:
static const char haystack[24] = {
0x4e, 0x65, 0x76, 0x65, 0x72, 0x20, 0x67, 0x6f,
0x6e, 0x6e, 0x61, 0x20, 0x67, 0x69, 0x76, 0x65,
0x20, 0x79, 0x6f, 0x75, 0x20, 0x75, 0x70, 0x2c,
};
static const char needle[8] = {
0x20, 0x79, 0x6f, 0x75, 0x20, 0x75, 0x70, 0x2c
};
char *res = memmem(haystack, sizeof(haystack), needle, sizeof(needle));
printf("%s", res);
The result is null. Any ideas where the problem might be?

I think you are overcomplicating this.
void *memmem(const void *haystack, size_t haystack_len,
const void * const needle, const size_t needle_len)
{
if (haystack == NULL) return NULL; // or assert(haystack != NULL);
if (haystack_len == 0) return NULL;
if (needle == NULL) return NULL; // or assert(needle != NULL);
if (needle_len == 0) return NULL;
for (const char *h = haystack;
haystack_len >= needle_len;
++h, --haystack_len) {
if (!memcmp(h, needle, needle_len)) {
return h;
}
}
return NULL;
}
Until haystack_len is greater or equal to needle_len, you should memory compare needle with current position in haystack. If it's true, return haystack.
There is no need to explicitly cast a const void * pointer const unsigned char *haystack = (const unsigned char *)haystack_start; is just const unsigned char *haystack = haystack_start;
As said in comments by #molbdnilo (*haystack != *needle) || (*haystack + needle_len != *needle + needle_len)) is just the same think. It becomes obvious, once you use [] operator rather then *: haystack[0] != needle[0] || haystack[0] + needle_len != needle[0] + needle_len. Even if you meant ... != needle[needle_len] this is out-of-bound access of needle.
The for is just strange:
for (; *haystack && haystack_len--; haystack++) {
if (haystack_len < needle_len)
break;
Why not:
for (; *haystack && haystack_len < needle_len; haystack_len--, haystack++)
?
And the expression *haystack is just invalid, you are not checking null-terminated string like in case of strstr. haystack points to any bytes in memory and may have zero as values. The haystack_len keeps the length of haystack.
You can use memcmp to compare memory, no need to write that part yourself.

The accepted answer is pretty good, but there are easy optimizations that can be made. For example, windows does provide a memchr function, and it is probably written in assembly language and is much more efficient than manually looping one character at a time from C and calling memcmp at every character. I didn't perform any benchmarks, but almost certainly this is not ideal.
Here's a solution the makes use of this optimization. We do the primary search with memchr to find the first byte of the needle, then do a memcmp to check the rest. Since there are 256 possible values for a byte, on average this will reduce the number of calls to memcmp by a factor of 255.
void *os_memmem(const void *src,int srclen,const void *trg,int trglen)
{
unsigned char *csrc = (unsigned char *)src;
unsigned char *ctrg = (unsigned char *)trg;
unsigned char *tptr,*cptr;
int searchlen,ndx=0;
/* add some initial error checking if you want */
while (ndx<=srclen) {
cptr = &csrc[ndx];
if ((searchlen = srclen-ndx-trglen+1) <= 0) {
return NULL;
} /* if */
if ((tptr = memchr(cptr,*ctrg,searchlen)) == NULL) {
return NULL;
} /* if */
if (memcmp(tptr,ctrg,trglen) == 0) {
return tptr;
} /* if */
ndx += tptr-cptr+1;
} /* while */
return NULL;
}

The following statements are the same
char needle[] = "a test";
char needle[7] = { 'a', ' ', 't', 'e', 's', 't', '\0' };
Your haystack does not have a '\0' at the same place as needle.
Try with
// needle is NOT a string
char needle[6] = "a test"; // needle is NOT a string

for (; *haystack && haystack_len--; haystack++) {
Here you're decrementing haystack_len at the start of the loop. This means that when you reach your match haystack_len = 7 but needle_len = 8. Therefore you fail the f (haystack_len < needle_len) check and continue. As Kamil notes in the comments you probably don't want a nul check on haystack either. So I'd suggest
for (; haystack_len > 0; ++haystack, --haystack_len) {

Related

Search for byte pattern in a buffered data stream

I want to search for a pattern of bytes that I receive in chunks (serially) as and when such data is available. For example, a pattern of bytes 0xbbffbbffbb. There is no guarantee that this pattern will be received in full, so doing a simple strnstrn may not be the solution. What algorithm can I use to look for this pattern? My approach is to look for the first byte (in this case 0xbb), then ensure I have 4 more bytes and then compare it with the string. Though it fails if there is some garbage data after two bytes, say 0xbbff01[bbffbbffbb].
My code(sorry if shabby) looks like this:
char* pattern_search(char* buff, size_t *bytes_read)
{
char* ptr = buff;
uint16_t remaining_length = *bytes_read;
while(1) {
// look for one byte in the stream
char* pattern_start = memmem((void*)ptr, remaining_length, 0xbb, 1);
if (pattern_start == NULL) {
// printf("nothing found\n");
return NULL;
}
int pos = pattern_start - ptr;
remaining_length = remaining_length - pos;
ptr = pattern_start;
// see if you have 5 bytes to compare, if not get more
remaining_length += get_additional_bytes();
// compare 5 bytes for pattern
pattern_start = memmem((void*)ptr, remaining_length, start_flag, PATTERN_LEN);
if (pattern_start == NULL) {
// move one step and continue search
ptr++;
remaining_length--;
// move these bytes back to beginning of the buffer
memcpy(buff, ptr, remaining_length);
ptr = buff;
*bytes_read = remaining_length;
if (remaining_length > 0) {
continue;
} else {
return NULL;
}
} else {
// found!
printf("pattern found!\n");
ptr = pattern_start;
break;
}
}
return ptr;
}
One can certainly find many different solutions here. One could be:
specify the pattern as an unsigned char array
calling of an 'input_received' function with received data blocks and a pointer to a callback function, which is called whenever the pattern is found
It could look like this:
#include <stdio.h>
static unsigned const char PATTERN[] = {0xbb, 0xff, 0xbb, 0xff, 0xbb};
static void found(size_t pos) {
printf("pattern found at index %zu\n", pos);
}
static void input_received(const unsigned char *const data,
int n,
void (*callback)(size_t)) {
static int match_count;
static size_t position;
for (int i = 0; i < n; i++, position++) {
if (data[i] == PATTERN[match_count]) {
match_count++;
} else {
match_count = data[i] == PATTERN[0] ? 1 : 0;
}
if (match_count == sizeof PATTERN) {
(*callback)(position - sizeof PATTERN + 1);
match_count = 0;
}
}
}
int main(void) {
unsigned char input[] = {0xff, 0x01, 0x02, 0xff, 0x00,
0xbb, 0xff, 0xbb, 0xff, 0xbb,
0xbb, 0xff, 0xbb, 0xff, 0xbb};
input_received(input, 2, found);
input_received(&input[2], 3, found);
input_received(&input[5], 2, found);
input_received(&input[7], 2, found);
input_received(&input[9], 5, found);
input_received(&input[14], 1, found);
return 0;
}
Test
This would then output the following in the debug console:
pattern found at index 5
pattern found at index 10

How to generate a file from C-hexvalues?

I have a C-array that looks like this:
char hexc[] = {
0x41, 0x80, 0x7a, 0x39, 0xea, 0x7e, 0x27, 0xfc,
0xe6, 0x45, 0x9c, 0x8b, 0xb5, 0xce, 0xa7, 0x35,
0x5f, 0xf2, 0x43, 0xcf, 0x89, 0xd8, 0x61, 0xec,
0xe7, 0xed, 0x2e, 0x34, 0x45, 0x0c, 0x32, 0xae,
0x71, 0x4f, 0x1c, 0xd8, 0xb5, 0x8c, 0x1e, 0xdd,
0x5d, 0x90, 0xf3, 0xf2, 0xe7, 0xa6, 0x4f, 0xef,
0xec, 0x96, 0xe3, 0xca, 0x8e, 0xeb, 0x64, 0x1d,
0x18, 0xa9, 0x95, 0xec, 0x64, 0x02, 0xf8, 0x26,
};
I knew that behind this hex-representations is a .GIF-file, what is the best way to generate from this hex-values a viewable file again? And how to add the missing GIF-header?
You would just open the file and write to it:
FILE *f = fopen("filename.gif", "wb");
if (!f) return; // or do something else
fwrite(hexc, 1, sizeof(hexc), f);
fclose(f);
Make sure to #include <stdio.h>.
open the file as binary
ofstream outfile ("new.gif",ofstream::binary);
and then write your buffer
outfile.write (hexc, sizeof hexc);
Turbo C DOS code
/****************************************************************************
** This support Compuserve 256 colour GIF87a and GIF89a image up to **
** 320x200 in size. **
****************************************************************************/
//This program requires a stack of at least 19.5K!!
#include "stdio.h"
typedef
struct GIFHeader {
char Signature [7];
unsigned int ScreenWidth, ScreenHeight;
unsigned char Depth, Background, Zero;
};
struct GIFDescriptor {
char Separator;
unsigned int ImageLeft, ImageTop, ImageWidth, ImageHeight;
unsigned char Depth;
};
char far *Screen = (char far *)0xA0000000L;
//For loading from the file
FILE *GIFFile;
unsigned int BPointer;
unsigned char Buffer [257];
//GIF data is stored in blocks of a certain size
unsigned char BlockSize;
//For loading the code
unsigned char CodeSize;
char BitsIn;
unsigned char Temp;
//Coordinates
unsigned int X, Y, tlX, tlY, brX, brY;
//The string table
unsigned int Prefix [4096];
unsigned char Suffix [4096];
//This sets the display to VGA 320x200 in 256 colours
void VGAScreen ()
{
asm {
mov ax, 0x13
int 0x10
}
}
//This resets the display to text mode
void TextScreen ()
{
asm {
mov ax, 0x3
int 0x10
}
}
//This sets a DAC register to a specific Red Green Blue-value
void SetDAC(unsigned char DAC, unsigned char R, unsigned char G, unsigned char B)
{
outportb (0x3C8, DAC);
outportb (0x3C9, R);
outportb (0x3C9, G);
outportb (0x3C9, B);
}
//This sets one pixel on the screen
void PutPixel (unsigned int x, unsigned int y, unsigned char c)
{
Screen [(y << 8) + (y << 6) + x] = c;
}
//Function to read from the buffer
unsigned char LoadByte ()
{
//Read next block}
if (BPointer == BlockSize) {
fread (Buffer, BlockSize + 1, 1, GIFFile);
BPointer = 0;
}
//Return byte
return Buffer [BPointer++];
}
//Procedure to read the next code from the file
unsigned int ReadCode ()
{
int Counter;
unsigned int Code;
Code = 0;
//Read the code, bit by bit
for (Counter = 0; Counter < CodeSize; Counter++) {
//Maybe, a new byte needs to be loaded with a further 8 bits
if (++BitsIn == 9) {
Temp = LoadByte ();
BitsIn = 1;
}
//Add the current bit to the code
if (Temp & 1) Code += 1 << Counter;
Temp >>= 1;
}
return Code;
}
//Procedure to draw a pixel
void NextPixel (unsigned int c)
{
//Actually draw the pixel on screen
PutPixel (X, Y, c & 255);
//Move to next row, if necessary
if (++X == brX) {
X = tlX;
Y++;
}
}
//Local function to output a string. Returns the first character.
unsigned char OutString (unsigned int CurCode)
{
unsigned int OutCount;
unsigned char OutCode [1024];
//If it's a single character, output that
if (CurCode < 256) {
NextPixel (CurCode);
} else {
OutCount = 0;
//Store the string, which ends up in reverse order
do {
OutCode [OutCount++] = Suffix [CurCode];
CurCode = Prefix [CurCode];
} while (CurCode > 255);
//Add the last character
OutCode [OutCount++] = CurCode;
//Output all the string, in the correct order
do {
NextPixel (OutCode [--OutCount]);
} while (OutCount);
}
//Return 1st character
return CurCode;
}
//This actually loads the GIF
void LoadGIF (char *Filename)
{
//For loading from the GIF file
struct GIFHeader Header;
struct GIFDescriptor Descriptor;
//Colour information
unsigned char BitsPerPixel,
NumOfColours;
unsigned int DAC;
unsigned char Palette [256][3];
//For indexing the string table
unsigned int FirstFree, FreeCode;
//All the code information
unsigned char InitCodeSize;
unsigned int Code, OldCode, MaxCode;
//Special codes
unsigned int ClearCode, EOICode;
//Check whether the GIF file exists, and open it
GIFFile = fopen (Filename, "rb");
if (GIFFile == 0) {
TextScreen ();
printf ("Could not open file %s", Filename);
return;
}
//Read header
fread (&Header, 6, 1, GIFFile);
Header.Signature [6] = 0;
fread (&Header.ScreenWidth, sizeof (Header) - 7, 1, GIFFile);
//Check signature and terminator
if ((strcmp (Header.Signature, "GIF87a")
&& strcmp (Header.Signature, "GIF89a"))
|| Header.Zero) {
TextScreen ();
printf ("Not a valid GIF file\n");
return;
}
//Get amount of colours in image
BitsPerPixel = 1 + (Header.Depth & 7);
NumOfColours = (1 << BitsPerPixel) - 1;
//Load global colour map
fread (Palette, 3, (NumOfColours + 1), GIFFile);
for (DAC = 0; DAC <= NumOfColours; DAC++)
SetDAC (DAC, Palette [DAC][0] >> 2,
Palette [DAC][1] >> 2,
Palette [DAC][2] >> 2);
//Load the image descriptor
fread (&Descriptor, sizeof (Descriptor), 1, GIFFile);
if (Descriptor.Separator != ',') {
TextScreen ();
printf ("Incorrect image descriptor.\n");
return;
}
//Get image corner coordinates
tlX = Descriptor.ImageLeft;
tlY = Descriptor.ImageTop;
brX = tlX + Descriptor.ImageWidth;
brY = tlY + Descriptor.ImageHeight;
//Some restrictions apply
if (Descriptor.Depth & 128) {
TextScreen ();
printf ("Local colour maps not supported\n");
return;
}
if (Descriptor.Depth & 64) {
TextScreen ();
printf ("Interlaced images not supported\n");
return;
}
//Get initial code size
fread (&CodeSize, 1, 1, GIFFile);
//GIF data is stored in blocks, so it's necessary to know the size
fread (&BlockSize, 1, 1, GIFFile);
//Start loader
BPointer = BlockSize;
//Special codes used in the GIF spec
ClearCode = 1 << CodeSize; //Code to reset
EOICode = ClearCode + 1; //End of file
//Initialize the string table
FirstFree = ClearCode + 2; //Strings start here
FreeCode = FirstFree; //Strings can be added here
//Initial size of the code and its maximum value
InitCodeSize = ++CodeSize;
MaxCode = 1 << CodeSize;
BitsIn = 8;
//Start at top left of image
X = Descriptor.ImageLeft;
Y = Descriptor.ImageTop;
do {
//Read next code
Code = ReadCode ();
//If it's an End-Of-Information code, stop processing
if (Code == EOICode) break;
//If it's a clear code...
else if (Code == ClearCode) {
//Clear the string table
FreeCode = FirstFree;
//Set the code size to initial values
CodeSize = InitCodeSize;
MaxCode = 1 << CodeSize;
//The next code may be read
Code = ReadCode ();
OldCode = Code;
//Set pixel
NextPixel (Code);
//Other codes
} else {
/*If the code is already in the string table, it's string is displayed,
and the old string followed by the new string's first character is
added to the string table.*/
if (Code < FreeCode)
Suffix [FreeCode] = OutString (Code);
else {
/*If it is not already in the string table, the old string followed by
the old string's first character is added to the string table and
displayed.*/
Suffix [FreeCode] = OutString (OldCode);
NextPixel (Suffix [FreeCode]);
}
//Finish adding to string table
Prefix [FreeCode++] = OldCode;
//If the code size needs to be adjusted, do so
if (FreeCode >= MaxCode && CodeSize < 12) {
CodeSize++;
MaxCode <<= 1;
}
//The current code is now old
OldCode = Code;
}
} while (Code != EOICode);
//Close the GIF file
fclose (GIFFile);
}
void main (int argcount, char *argvalue[])
{
char FileName [80];
//Check if a filename was passed as a parameter, otherwise ask for one
if (argcount > 1) {
strcpy (FileName, argvalue [1]);
} else {
printf ("Enter filename:");
gets (FileName);
}
//Switch to graphics screen
VGAScreen ();
//Load GIF file
LoadGIF (FileName);
//Wait for keypress
getch ();
//Switch back to text mode
TextScreen ();
}

algorithm for bejeweled (3-in-a-row all the way until 5-in-a-row) [duplicate]

Given a (char *) string, I want to find all occurrences of a substring and replace them with an alternate string. I do not see any simple function that achieves this in <string.h>.
The optimizer should eliminate most of the local variables. The tmp pointer is there to make sure strcpy doesn't have to walk the string to find the null. tmp points to the end of result after each call. (See Shlemiel the painter's algorithm for why strcpy can be annoying.)
// You must free the result if result is non-NULL.
char *str_replace(char *orig, char *rep, char *with) {
char *result; // the return string
char *ins; // the next insert point
char *tmp; // varies
int len_rep; // length of rep (the string to remove)
int len_with; // length of with (the string to replace rep with)
int len_front; // distance between rep and end of last rep
int count; // number of replacements
// sanity checks and initialization
if (!orig || !rep)
return NULL;
len_rep = strlen(rep);
if (len_rep == 0)
return NULL; // empty rep causes infinite loop during count
if (!with)
with = "";
len_with = strlen(with);
// count the number of replacements needed
ins = orig;
for (count = 0; tmp = strstr(ins, rep); ++count) {
ins = tmp + len_rep;
}
tmp = result = malloc(strlen(orig) + (len_with - len_rep) * count + 1);
if (!result)
return NULL;
// first time through the loop, all the variable are set correctly
// from here on,
// tmp points to the end of the result string
// ins points to the next occurrence of rep in orig
// orig points to the remainder of orig after "end of rep"
while (count--) {
ins = strstr(orig, rep);
len_front = ins - orig;
tmp = strncpy(tmp, orig, len_front) + len_front;
tmp = strcpy(tmp, with) + len_with;
orig += len_front + len_rep; // move to next "end of rep"
}
strcpy(tmp, orig);
return result;
}
This is not provided in the standard C library because, given only a char* you can't increase the memory allocated to the string if the replacement string is longer than the string being replaced.
You can do this using std::string more easily, but even there, no single function will do it for you.
There isn't one.
You'd need to roll your own using something like strstr and strcat or strcpy.
You could build your own replace function using strstr to find the substrings and strncpy to copy in parts to a new buffer.
Unless what you want to replace_with is the same length as what you you want to replace, then it's probably best to use a new buffer to copy the new string to.
Here's some sample code that does it.
#include <string.h>
#include <stdlib.h>
char * replace(
char const * const original,
char const * const pattern,
char const * const replacement
) {
size_t const replen = strlen(replacement);
size_t const patlen = strlen(pattern);
size_t const orilen = strlen(original);
size_t patcnt = 0;
const char * oriptr;
const char * patloc;
// find how many times the pattern occurs in the original string
for (oriptr = original; patloc = strstr(oriptr, pattern); oriptr = patloc + patlen)
{
patcnt++;
}
{
// allocate memory for the new string
size_t const retlen = orilen + patcnt * (replen - patlen);
char * const returned = (char *) malloc( sizeof(char) * (retlen + 1) );
if (returned != NULL)
{
// copy the original string,
// replacing all the instances of the pattern
char * retptr = returned;
for (oriptr = original; patloc = strstr(oriptr, pattern); oriptr = patloc + patlen)
{
size_t const skplen = patloc - oriptr;
// copy the section until the occurence of the pattern
strncpy(retptr, oriptr, skplen);
retptr += skplen;
// copy the replacement
strncpy(retptr, replacement, replen);
retptr += replen;
}
// copy the rest of the string.
strcpy(retptr, oriptr);
}
return returned;
}
}
#include <stdio.h>
int main(int argc, char * argv[])
{
if (argc != 4)
{
fprintf(stderr,"usage: %s <original text> <pattern> <replacement>\n", argv[0]);
exit(-1);
}
else
{
char * const newstr = replace(argv[1], argv[2], argv[3]);
if (newstr)
{
printf("%s\n", newstr);
free(newstr);
}
else
{
fprintf(stderr,"allocation error\n");
exit(-2);
}
}
return 0;
}
As strings in C can not dynamically grow inplace substitution will generally not work. Therefore you need to allocate space for a new string that has enough room for your substitution and then copy the parts from the original plus the substitution into the new string. To copy the parts you would use strncpy.
// Here is the code for unicode strings!
int mystrstr(wchar_t *txt1,wchar_t *txt2)
{
wchar_t *posstr=wcsstr(txt1,txt2);
if(posstr!=NULL)
{
return (posstr-txt1);
}else
{
return -1;
}
}
// assume: supplied buff is enough to hold generated text
void StringReplace(wchar_t *buff,wchar_t *txt1,wchar_t *txt2)
{
wchar_t *tmp;
wchar_t *nextStr;
int pos;
tmp=wcsdup(buff);
pos=mystrstr(tmp,txt1);
if(pos!=-1)
{
buff[0]=0;
wcsncpy(buff,tmp,pos);
buff[pos]=0;
wcscat(buff,txt2);
nextStr=tmp+pos+wcslen(txt1);
while(wcslen(nextStr)!=0)
{
pos=mystrstr(nextStr,txt1);
if(pos==-1)
{
wcscat(buff,nextStr);
break;
}
wcsncat(buff,nextStr,pos);
wcscat(buff,txt2);
nextStr=nextStr+pos+wcslen(txt1);
}
}
free(tmp);
}
The repl_str() function on creativeandcritical.net is fast and reliable. Also included on that page is a wide string variant, repl_wcs(), which can be used with Unicode strings including those encoded in UTF-8, through helper functions - demo code is linked from the page. Belated full disclosure: I am the author of that page and the functions on it.
Here is the one that I created based on these requirements:
Replace the pattern regardless of whether is was long or shorter.
Not use any malloc (explicit or implicit) to intrinsically avoid memory leaks.
Replace any number of occurrences of pattern.
Tolerate the replace string having a substring equal to the search string.
Does not have to check that the Line array is sufficient in size to hold the replacement. e.g. This does not work unless the caller knows that line is of sufficient size to hold the new string.
avoid use of strcat() to avoid overhead of scanning the entire string to append another string.
/* returns number of strings replaced.
*/
int replacestr(char *line, const char *search, const char *replace)
{
int count;
char *sp; // start of pattern
//printf("replacestr(%s, %s, %s)\n", line, search, replace);
if ((sp = strstr(line, search)) == NULL) {
return(0);
}
count = 1;
int sLen = strlen(search);
int rLen = strlen(replace);
if (sLen > rLen) {
// move from right to left
char *src = sp + sLen;
char *dst = sp + rLen;
while((*dst = *src) != '\0') { dst++; src++; }
} else if (sLen < rLen) {
// move from left to right
int tLen = strlen(sp) - sLen;
char *stop = sp + rLen;
char *src = sp + sLen + tLen;
char *dst = sp + rLen + tLen;
while(dst >= stop) { *dst = *src; dst--; src--; }
}
memcpy(sp, replace, rLen);
count += replacestr(sp + rLen, search, replace);
return(count);
}
Any suggestions for improving this code are cheerfully accepted. Just post the comment and I will test it.
i find most of the proposed functions hard to understand - so i came up with this:
static char *dull_replace(const char *in, const char *pattern, const char *by)
{
size_t outsize = strlen(in) + 1;
// TODO maybe avoid reallocing by counting the non-overlapping occurences of pattern
char *res = malloc(outsize);
// use this to iterate over the output
size_t resoffset = 0;
char *needle;
while (needle = strstr(in, pattern)) {
// copy everything up to the pattern
memcpy(res + resoffset, in, needle - in);
resoffset += needle - in;
// skip the pattern in the input-string
in = needle + strlen(pattern);
// adjust space for replacement
outsize = outsize - strlen(pattern) + strlen(by);
res = realloc(res, outsize);
// copy the pattern
memcpy(res + resoffset, by, strlen(by));
resoffset += strlen(by);
}
// copy the remaining input
strcpy(res + resoffset, in);
return res;
}
output must be free'd
a fix to fann95's response, using in-place modification of the string, and assuming the buffer pointed to by line is large enough to hold the resulting string.
static void replacestr(char *line, const char *search, const char *replace)
{
char *sp;
if ((sp = strstr(line, search)) == NULL) {
return;
}
int search_len = strlen(search);
int replace_len = strlen(replace);
int tail_len = strlen(sp+search_len);
memmove(sp+replace_len,sp+search_len,tail_len+1);
memcpy(sp, replace, replace_len);
}
/*замена символа в строке*/
char* replace_char(char* str, char in, char out) {
char * p = str;
while(p != '\0') {
if(*p == in)
*p == out;
++p;
}
return str;
}
This function only works if ur string has extra space for new length
void replace_str(char *str,char *org,char *rep)
{
char *ToRep = strstr(str,org);
char *Rest = (char*)malloc(strlen(ToRep));
strcpy(Rest,((ToRep)+strlen(org)));
strcpy(ToRep,rep);
strcat(ToRep,Rest);
free(Rest);
}
This only replaces First occurrence
Here goes mine, make them all char*, which makes calling easier...
char *strrpc(char *str,char *oldstr,char *newstr){
char bstr[strlen(str)];
memset(bstr,0,sizeof(bstr));
int i;
for(i = 0;i < strlen(str);i++){
if(!strncmp(str+i,oldstr,strlen(oldstr))){
strcat(bstr,newstr);
i += strlen(oldstr) - 1;
}else{
strncat(bstr,str + i,1);
}
}
strcpy(str,bstr);
return str;
}
There is a function in string.h but it works with char [] not char* but again it outputs a char* and not a char []
It is simple and beautiful
Supposing we want to replace 'and' in 'TheandQuickandBrownandFox'.
We first split with strtok and then join with snprintf defined in the stdio.h
char sometext[] = "TheandQuickandBrownandFox";
char* replaced = malloc(1024);
// split on the substring, here I am using (and)
char* token = strtok(sometext, "and");
snprintf(replaced, 1, "%s", ""); // initialise so we can compare
while(token) {
if (strcmp(replaced, "") < 1) {
// if it is the first one
snprintf(replaced, 1024, "%s", token);
token = NULL;
} else {
// put the space between the existing and new
snprintf(replaced, 1024, "%s %s", replaced, token);
token = NULL;
}
}
free(replaced);
This should give us:
The Quick Brown Fox
You can use this function (the comments explain how it works):
void strreplace(char *string, const char *find, const char *replaceWith){
if(strstr(string, find) != NULL){
char *temporaryString = malloc(strlen(strstr(string, find) + strlen(find)) + 1);
strcpy(temporaryString, strstr(string, find) + strlen(find)); //Create a string with what's after the replaced part
*strstr(string, find) = '\0'; //Take away the part to replace and the part after it in the initial string
strcat(string, replaceWith); //Concat the first part of the string with the part to replace with
strcat(string, temporaryString); //Concat the first part of the string with the part after the replaced part
free(temporaryString); //Free the memory to avoid memory leaks
}
}
DWORD ReplaceString(__inout PCHAR source, __in DWORD dwSourceLen, __in const char* pszTextToReplace, __in const char* pszReplaceWith)
{
DWORD dwRC = NO_ERROR;
PCHAR foundSeq = NULL;
PCHAR restOfString = NULL;
PCHAR searchStart = source;
size_t szReplStrcLen = strlen(pszReplaceWith), szRestOfStringLen = 0, sztextToReplaceLen = strlen(pszTextToReplace), remainingSpace = 0, dwSpaceRequired = 0;
if (strcmp(pszTextToReplace, "") == 0)
dwRC = ERROR_INVALID_PARAMETER;
else if (strcmp(pszTextToReplace, pszReplaceWith) != 0)
{
do
{
foundSeq = strstr(searchStart, pszTextToReplace);
if (foundSeq)
{
szRestOfStringLen = (strlen(foundSeq) - sztextToReplaceLen) + 1;
remainingSpace = dwSourceLen - (foundSeq - source);
dwSpaceRequired = szReplStrcLen + (szRestOfStringLen);
if (dwSpaceRequired > remainingSpace)
{
dwRC = ERROR_MORE_DATA;
}
else
{
restOfString = CMNUTIL_calloc(szRestOfStringLen, sizeof(CHAR));
strcpy_s(restOfString, szRestOfStringLen, foundSeq + sztextToReplaceLen);
strcpy_s(foundSeq, remainingSpace, pszReplaceWith);
strcat_s(foundSeq, remainingSpace, restOfString);
}
CMNUTIL_free(restOfString);
searchStart = foundSeq + szReplStrcLen; //search in the remaining str. (avoid loops when replWith contains textToRepl
}
} while (foundSeq && dwRC == NO_ERROR);
}
return dwRC;
}
char *replace(const char*instring, const char *old_part, const char *new_part)
{
#ifndef EXPECTED_REPLACEMENTS
#define EXPECTED_REPLACEMENTS 100
#endif
if(!instring || !old_part || !new_part)
{
return (char*)NULL;
}
size_t instring_len=strlen(instring);
size_t new_len=strlen(new_part);
size_t old_len=strlen(old_part);
if(instring_len<old_len || old_len==0)
{
return (char*)NULL;
}
const char *in=instring;
const char *found=NULL;
size_t count=0;
size_t out=0;
size_t ax=0;
char *outstring=NULL;
if(new_len> old_len )
{
size_t Diff=EXPECTED_REPLACEMENTS*(new_len-old_len);
size_t outstring_len=instring_len + Diff;
outstring =(char*) malloc(outstring_len);
if(!outstring){
return (char*)NULL;
}
while((found = strstr(in, old_part))!=NULL)
{
if(count==EXPECTED_REPLACEMENTS)
{
outstring_len+=Diff;
if((outstring=realloc(outstring,outstring_len))==NULL)
{
return (char*)NULL;
}
count=0;
}
ax=found-in;
strncpy(outstring+out,in,ax);
out+=ax;
strncpy(outstring+out,new_part,new_len);
out+=new_len;
in=found+old_len;
count++;
}
}
else
{
outstring =(char*) malloc(instring_len);
if(!outstring){
return (char*)NULL;
}
while((found = strstr(in, old_part))!=NULL)
{
ax=found-in;
strncpy(outstring+out,in,ax);
out+=ax;
strncpy(outstring+out,new_part,new_len);
out+=new_len;
in=found+old_len;
}
}
ax=(instring+instring_len)-in;
strncpy(outstring+out,in,ax);
out+=ax;
outstring[out]='\0';
return outstring;
}
Using only strlen from string.h
sorry for my English
char * str_replace(char * text,char * rep, char * repw){//text -> to replace in it | rep -> replace | repw -> replace with
int replen = strlen(rep),repwlen = strlen(repw),count;//some constant variables
for(int i=0;i<strlen(text);i++){//search for the first character from rep in text
if(text[i] == rep[0]){//if it found it
count = 1;//start searching from the next character to avoid repetition
for(int j=1;j<replen;j++){
if(text[i+j] == rep[j]){//see if the next character in text is the same as the next in the rep if not break
count++;
}else{
break;
}
}
if(count == replen){//if count equals to the lenght of the rep then we found the word that we want to replace in the text
if(replen < repwlen){
for(int l = strlen(text);l>i;l--){//cuz repwlen greater than replen we need to shift characters to the right to make space for the replacement to fit
text[l+repwlen-replen] = text[l];//shift by repwlen-replen
}
}
if(replen > repwlen){
for(int l=i+replen-repwlen;l<strlen(text);l++){//cuz replen greater than repwlen we need to shift the characters to the left
text[l-(replen-repwlen)] = text[l];//shift by replen-repwlen
}
text[strlen(text)-(replen-repwlen)] = '\0';//get rid of the last unwanted characters
}
for(int l=0;l<repwlen;l++){//replace rep with repwlen
text[i+l] = repw[l];
}
if(replen != repwlen){
i+=repwlen-1;//pass to the next character | try text "y" ,rep "y",repw "yy" without this line to understand
}
}
}
}
return text;
}
if you want strlen code to avoid calling string.h
int strlen(char * string){//use this code to avoid calling string.h
int lenght = 0;
while(string[lenght] != '\0'){
lenght++;
}
return lenght;
}
There you go....this is the function to replace every occurance of char x with char y within character string str
char *zStrrep(char *str, char x, char y){
char *tmp=str;
while(*tmp)
if(*tmp == x)
*tmp++ = y; /* assign first, then incement */
else
*tmp++;
// *tmp='\0'; -> we do not need this
return str;
}
An example usage could be
Exmaple Usage
char s[]="this is a trial string to test the function.";
char x=' ', y='_';
printf("%s\n",zStrrep(s,x,y));
Example Output
this_is_a_trial_string_to_test_the_function.
The function is from a string library I maintain on Github, you are more than welcome to have a look at other available functions or even contribute to the code :)
https://github.com/fnoyanisi/zString
EDIT:
#siride is right, the function above replaces chars only. Just wrote this one, which replaces character strings.
#include <stdio.h>
#include <stdlib.h>
/* replace every occurance of string x with string y */
char *zstring_replace_str(char *str, const char *x, const char *y){
char *tmp_str = str, *tmp_x = x, *dummy_ptr = tmp_x, *tmp_y = y;
int len_str=0, len_y=0, len_x=0;
/* string length */
for(; *tmp_y; ++len_y, ++tmp_y)
;
for(; *tmp_str; ++len_str, ++tmp_str)
;
for(; *tmp_x; ++len_x, ++tmp_x)
;
/* Bounds check */
if (len_y >= len_str)
return str;
/* reset tmp pointers */
tmp_y = y;
tmp_x = x;
for (tmp_str = str ; *tmp_str; ++tmp_str)
if(*tmp_str == *tmp_x) {
/* save tmp_str */
for (dummy_ptr=tmp_str; *dummy_ptr == *tmp_x; ++tmp_x, ++dummy_ptr)
if (*(tmp_x+1) == '\0' && ((dummy_ptr-str+len_y) < len_str)){
/* Reached end of x, we got something to replace then!
* Copy y only if there is enough room for it
*/
for(tmp_y=y; *tmp_y; ++tmp_y, ++tmp_str)
*tmp_str = *tmp_y;
}
/* reset tmp_x */
tmp_x = x;
}
return str;
}
int main()
{
char s[]="Free software is a matter of liberty, not price.\n"
"To understand the concept, you should think of 'free' \n"
"as in 'free speech', not as in 'free beer'";
printf("%s\n\n",s);
printf("%s\n",zstring_replace_str(s,"ree","XYZ"));
return 0;
}
And below is the output
Free software is a matter of liberty, not price.
To understand the concept, you should think of 'free'
as in 'free speech', not as in 'free beer'
FXYZ software is a matter of liberty, not price.
To understand the concept, you should think of 'fXYZ'
as in 'fXYZ speech', not as in 'fXYZ beer'
You can use strrep()
char* strrep ( const char * cadena,
const char * strf,
const char * strr
)
strrep (String Replace). Replaces strf with strr in cadena and returns the new string. You need to free the returned string in your code after using strrep.
Parameters:
cadena: The string with the text.
strf: The text to find.
strr: The replacement text.
Returns
The text updated wit the replacement.
Project can be found at https://github.com/ipserc/strrep

How can I define an arbitrarily-sized 2D array and then determine its dimensions at compile-time?

Executive summary:
How can I define an arbitrarily-sized 2D array in C?
How can I determine the dimensions of that array at compile-time?
Full disclosure:
I'm writing code for an embedded controller. My application requires several lookup tables with different sizes which will all be used by one lookup function (a binary search). Here is what I have so far:
typedef struct
{
unsigned char count; // number of rows in the table
unsigned char width; // number of bytes in each row
const unsigned char * data; // pointer to table data[count][width]
}
LookupTable;
// returns the index of a value from within a table
unsigned char Lookup(unsigned long value, const LookupTable * table);
This part is working. What I would like to do now is define these tables in my source without having to manually enter the count and width constants. Here is what I am doing now:
#define T1_count 100
#define T1_width 3
const unsigned char table1_data[T1_count][T1_width] =
{
{ 0x12, 0x34, 0x56 },
{ 0x12, 0x38, 0x12 },
...
};
const LookupTable table1 = { T1_count, T1_width, table1_data };
Here is what I would like to be able to do (pseudo-code, since this array definition will not actually compile):
const unsigned char table1_data[] =
{
{ 0x12, 0x34, 0x56 },
{ 0x12, 0x38, 0x12 },
...
};
const LookupTable table1 =
{
get_count_expr(table1_data),
get_width_expr(table1_data),
table1_data
};
Obviously, get_count_expr and get_width_expr would have to be constant expressions of some sort, based on the size of the table, and not actual function calls.
To be clear, no part of this design is cast in stone. I'm just posting what I have so far, in the hopes that my intent is clear. Any ideas for improvement would be appreciated.
The "why":
These tables will be changed often, and it would make maintenance easier if entries could be added and removed, or the width of a table changed without having to manually adjust the constants each time. Having to keep track of the sizes manually can be error-prone and violates DRY. I'm looking for a better way.
Hmmm ... you can leave the leftmost size to the compiler:
#define T1_WIDTH 3
const unsigned char table1_data[][T1_WIDTH] =
{
{ 0x12, 0x34, 0x56 },
{ 0x12, 0x38, 0x12 },
/* ... */
};
T1_count = sizeof table1_data / sizeof *table1_data;
T1_width = sizeof *table1_data;
Well, it's ugly as hell, but I think the only way to do it within the constraints you've listed is to include the data in a string, and than have initialization code parse the string and generate the table. Ideally you'd do that in a script rather than use C to do it, but if it has to be in C, it has to be in C..
Note that in no way do I claim the following to be production code, but it's just a proof of concept...
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define datatable "\
{ 0x12, 0x34, 0x56 },\
{ 0x14, 0x36, 0x10 },\
{ 0x13, 0x37, 0x11 },\
{ 0x12, 0x38, 0x12 }"
typedef struct
{
unsigned char count; // number of rows in the table
unsigned char width; // number of bytes in each row
unsigned char **data; // pointer to table data[count][width]
}
LookupTable;
int parsedatatable(char *data, LookupTable **table) {
char *p, *sp, save;
unsigned char *tabledata;
int count = 0, width = 0;
unsigned int tmp;
int i,j;
/* find count */
p = strstr(data,"{");
while (p) {
p++;
p = strstr(p, "{");
count++;
}
/* find width */
p = strstr(data, "{");
p++;
sp = strstr(p, "}");
if (sp != NULL) {
save = *sp;
*sp = '\0';
}
while (p) {
p = strstr(p, ",");
width++;
if (p != NULL) p++;
}
if (sp != NULL) {
*sp = save;
}
printf("Count = %d, width = %d\n",count, width);
tabledata = (unsigned char *)malloc(width*count*sizeof(unsigned char));
*table = (LookupTable *)malloc(sizeof(LookupTable));
(*table)->data = (unsigned char **)malloc(count*sizeof(unsigned char*));
for (i=0; i<count; i++) {
(*table)->data[i] = &(tabledata[i*width]);
}
(*table)->count = count;
(*table)->width = width;
p = data;
for (i=0; i<count; i++) {
p = strstr(p,"{");
if (!p) {
fprintf(stderr,"Fail (a) reading in data!: %s\n",data);
free((*table)->data);
free(tabledata);
free(*table);
return -1;
}
p++;
for (j=0; j<width; j++) {
printf("Scanning <%s>, ",p);
sscanf(p,"%x",&tmp);
printf("got %d\n",tmp);
(*table)->data[i][j] = tmp;
p = strstr(p,",");
if (!p && j<width-1) {
fprintf(stderr,"Fail (b) reading in data!: %d, %d, %s\n",i,j,data);
free((*table)->data);
free(tabledata);
free(*table);
return -1;
}
p++;
}
}
return 0;
}
void printtable(LookupTable *table) {
unsigned char i,j;
for (i=0; i<table->count; i++) {
printf("{");
for (j=0; j<table->width; j++) {
printf("%x ",table->data[i][j]);
}
printf("}\n");
}
return;
}
int main(int argc, char **argv) {
char *data;
LookupTable *table;
data = (char *)malloc(strlen(datatable)+1);
strcpy(data,datatable);
parsedatatable(data,&table);
printtable(table);
return 0;
}
Well, but who fills these tables with data? I think that generated sources are better solution.
Define table1_data inside a header. You can auto-generate that header using a script. I do something similar to that for some of my projects. I have a CSV file with data and a Ruby or Python script that generates a header from it.

What function is to replace a substring from a string in C?

Given a (char *) string, I want to find all occurrences of a substring and replace them with an alternate string. I do not see any simple function that achieves this in <string.h>.
The optimizer should eliminate most of the local variables. The tmp pointer is there to make sure strcpy doesn't have to walk the string to find the null. tmp points to the end of result after each call. (See Shlemiel the painter's algorithm for why strcpy can be annoying.)
// You must free the result if result is non-NULL.
char *str_replace(char *orig, char *rep, char *with) {
char *result; // the return string
char *ins; // the next insert point
char *tmp; // varies
int len_rep; // length of rep (the string to remove)
int len_with; // length of with (the string to replace rep with)
int len_front; // distance between rep and end of last rep
int count; // number of replacements
// sanity checks and initialization
if (!orig || !rep)
return NULL;
len_rep = strlen(rep);
if (len_rep == 0)
return NULL; // empty rep causes infinite loop during count
if (!with)
with = "";
len_with = strlen(with);
// count the number of replacements needed
ins = orig;
for (count = 0; tmp = strstr(ins, rep); ++count) {
ins = tmp + len_rep;
}
tmp = result = malloc(strlen(orig) + (len_with - len_rep) * count + 1);
if (!result)
return NULL;
// first time through the loop, all the variable are set correctly
// from here on,
// tmp points to the end of the result string
// ins points to the next occurrence of rep in orig
// orig points to the remainder of orig after "end of rep"
while (count--) {
ins = strstr(orig, rep);
len_front = ins - orig;
tmp = strncpy(tmp, orig, len_front) + len_front;
tmp = strcpy(tmp, with) + len_with;
orig += len_front + len_rep; // move to next "end of rep"
}
strcpy(tmp, orig);
return result;
}
This is not provided in the standard C library because, given only a char* you can't increase the memory allocated to the string if the replacement string is longer than the string being replaced.
You can do this using std::string more easily, but even there, no single function will do it for you.
There isn't one.
You'd need to roll your own using something like strstr and strcat or strcpy.
You could build your own replace function using strstr to find the substrings and strncpy to copy in parts to a new buffer.
Unless what you want to replace_with is the same length as what you you want to replace, then it's probably best to use a new buffer to copy the new string to.
Here's some sample code that does it.
#include <string.h>
#include <stdlib.h>
char * replace(
char const * const original,
char const * const pattern,
char const * const replacement
) {
size_t const replen = strlen(replacement);
size_t const patlen = strlen(pattern);
size_t const orilen = strlen(original);
size_t patcnt = 0;
const char * oriptr;
const char * patloc;
// find how many times the pattern occurs in the original string
for (oriptr = original; patloc = strstr(oriptr, pattern); oriptr = patloc + patlen)
{
patcnt++;
}
{
// allocate memory for the new string
size_t const retlen = orilen + patcnt * (replen - patlen);
char * const returned = (char *) malloc( sizeof(char) * (retlen + 1) );
if (returned != NULL)
{
// copy the original string,
// replacing all the instances of the pattern
char * retptr = returned;
for (oriptr = original; patloc = strstr(oriptr, pattern); oriptr = patloc + patlen)
{
size_t const skplen = patloc - oriptr;
// copy the section until the occurence of the pattern
strncpy(retptr, oriptr, skplen);
retptr += skplen;
// copy the replacement
strncpy(retptr, replacement, replen);
retptr += replen;
}
// copy the rest of the string.
strcpy(retptr, oriptr);
}
return returned;
}
}
#include <stdio.h>
int main(int argc, char * argv[])
{
if (argc != 4)
{
fprintf(stderr,"usage: %s <original text> <pattern> <replacement>\n", argv[0]);
exit(-1);
}
else
{
char * const newstr = replace(argv[1], argv[2], argv[3]);
if (newstr)
{
printf("%s\n", newstr);
free(newstr);
}
else
{
fprintf(stderr,"allocation error\n");
exit(-2);
}
}
return 0;
}
As strings in C can not dynamically grow inplace substitution will generally not work. Therefore you need to allocate space for a new string that has enough room for your substitution and then copy the parts from the original plus the substitution into the new string. To copy the parts you would use strncpy.
// Here is the code for unicode strings!
int mystrstr(wchar_t *txt1,wchar_t *txt2)
{
wchar_t *posstr=wcsstr(txt1,txt2);
if(posstr!=NULL)
{
return (posstr-txt1);
}else
{
return -1;
}
}
// assume: supplied buff is enough to hold generated text
void StringReplace(wchar_t *buff,wchar_t *txt1,wchar_t *txt2)
{
wchar_t *tmp;
wchar_t *nextStr;
int pos;
tmp=wcsdup(buff);
pos=mystrstr(tmp,txt1);
if(pos!=-1)
{
buff[0]=0;
wcsncpy(buff,tmp,pos);
buff[pos]=0;
wcscat(buff,txt2);
nextStr=tmp+pos+wcslen(txt1);
while(wcslen(nextStr)!=0)
{
pos=mystrstr(nextStr,txt1);
if(pos==-1)
{
wcscat(buff,nextStr);
break;
}
wcsncat(buff,nextStr,pos);
wcscat(buff,txt2);
nextStr=nextStr+pos+wcslen(txt1);
}
}
free(tmp);
}
The repl_str() function on creativeandcritical.net is fast and reliable. Also included on that page is a wide string variant, repl_wcs(), which can be used with Unicode strings including those encoded in UTF-8, through helper functions - demo code is linked from the page. Belated full disclosure: I am the author of that page and the functions on it.
Here is the one that I created based on these requirements:
Replace the pattern regardless of whether is was long or shorter.
Not use any malloc (explicit or implicit) to intrinsically avoid memory leaks.
Replace any number of occurrences of pattern.
Tolerate the replace string having a substring equal to the search string.
Does not have to check that the Line array is sufficient in size to hold the replacement. e.g. This does not work unless the caller knows that line is of sufficient size to hold the new string.
avoid use of strcat() to avoid overhead of scanning the entire string to append another string.
/* returns number of strings replaced.
*/
int replacestr(char *line, const char *search, const char *replace)
{
int count;
char *sp; // start of pattern
//printf("replacestr(%s, %s, %s)\n", line, search, replace);
if ((sp = strstr(line, search)) == NULL) {
return(0);
}
count = 1;
int sLen = strlen(search);
int rLen = strlen(replace);
if (sLen > rLen) {
// move from right to left
char *src = sp + sLen;
char *dst = sp + rLen;
while((*dst = *src) != '\0') { dst++; src++; }
} else if (sLen < rLen) {
// move from left to right
int tLen = strlen(sp) - sLen;
char *stop = sp + rLen;
char *src = sp + sLen + tLen;
char *dst = sp + rLen + tLen;
while(dst >= stop) { *dst = *src; dst--; src--; }
}
memcpy(sp, replace, rLen);
count += replacestr(sp + rLen, search, replace);
return(count);
}
Any suggestions for improving this code are cheerfully accepted. Just post the comment and I will test it.
i find most of the proposed functions hard to understand - so i came up with this:
static char *dull_replace(const char *in, const char *pattern, const char *by)
{
size_t outsize = strlen(in) + 1;
// TODO maybe avoid reallocing by counting the non-overlapping occurences of pattern
char *res = malloc(outsize);
// use this to iterate over the output
size_t resoffset = 0;
char *needle;
while (needle = strstr(in, pattern)) {
// copy everything up to the pattern
memcpy(res + resoffset, in, needle - in);
resoffset += needle - in;
// skip the pattern in the input-string
in = needle + strlen(pattern);
// adjust space for replacement
outsize = outsize - strlen(pattern) + strlen(by);
res = realloc(res, outsize);
// copy the pattern
memcpy(res + resoffset, by, strlen(by));
resoffset += strlen(by);
}
// copy the remaining input
strcpy(res + resoffset, in);
return res;
}
output must be free'd
a fix to fann95's response, using in-place modification of the string, and assuming the buffer pointed to by line is large enough to hold the resulting string.
static void replacestr(char *line, const char *search, const char *replace)
{
char *sp;
if ((sp = strstr(line, search)) == NULL) {
return;
}
int search_len = strlen(search);
int replace_len = strlen(replace);
int tail_len = strlen(sp+search_len);
memmove(sp+replace_len,sp+search_len,tail_len+1);
memcpy(sp, replace, replace_len);
}
/*замена символа в строке*/
char* replace_char(char* str, char in, char out) {
char * p = str;
while(p != '\0') {
if(*p == in)
*p == out;
++p;
}
return str;
}
This function only works if ur string has extra space for new length
void replace_str(char *str,char *org,char *rep)
{
char *ToRep = strstr(str,org);
char *Rest = (char*)malloc(strlen(ToRep));
strcpy(Rest,((ToRep)+strlen(org)));
strcpy(ToRep,rep);
strcat(ToRep,Rest);
free(Rest);
}
This only replaces First occurrence
Here goes mine, make them all char*, which makes calling easier...
char *strrpc(char *str,char *oldstr,char *newstr){
char bstr[strlen(str)];
memset(bstr,0,sizeof(bstr));
int i;
for(i = 0;i < strlen(str);i++){
if(!strncmp(str+i,oldstr,strlen(oldstr))){
strcat(bstr,newstr);
i += strlen(oldstr) - 1;
}else{
strncat(bstr,str + i,1);
}
}
strcpy(str,bstr);
return str;
}
There is a function in string.h but it works with char [] not char* but again it outputs a char* and not a char []
It is simple and beautiful
Supposing we want to replace 'and' in 'TheandQuickandBrownandFox'.
We first split with strtok and then join with snprintf defined in the stdio.h
char sometext[] = "TheandQuickandBrownandFox";
char* replaced = malloc(1024);
// split on the substring, here I am using (and)
char* token = strtok(sometext, "and");
snprintf(replaced, 1, "%s", ""); // initialise so we can compare
while(token) {
if (strcmp(replaced, "") < 1) {
// if it is the first one
snprintf(replaced, 1024, "%s", token);
token = NULL;
} else {
// put the space between the existing and new
snprintf(replaced, 1024, "%s %s", replaced, token);
token = NULL;
}
}
free(replaced);
This should give us:
The Quick Brown Fox
You can use this function (the comments explain how it works):
void strreplace(char *string, const char *find, const char *replaceWith){
if(strstr(string, find) != NULL){
char *temporaryString = malloc(strlen(strstr(string, find) + strlen(find)) + 1);
strcpy(temporaryString, strstr(string, find) + strlen(find)); //Create a string with what's after the replaced part
*strstr(string, find) = '\0'; //Take away the part to replace and the part after it in the initial string
strcat(string, replaceWith); //Concat the first part of the string with the part to replace with
strcat(string, temporaryString); //Concat the first part of the string with the part after the replaced part
free(temporaryString); //Free the memory to avoid memory leaks
}
}
DWORD ReplaceString(__inout PCHAR source, __in DWORD dwSourceLen, __in const char* pszTextToReplace, __in const char* pszReplaceWith)
{
DWORD dwRC = NO_ERROR;
PCHAR foundSeq = NULL;
PCHAR restOfString = NULL;
PCHAR searchStart = source;
size_t szReplStrcLen = strlen(pszReplaceWith), szRestOfStringLen = 0, sztextToReplaceLen = strlen(pszTextToReplace), remainingSpace = 0, dwSpaceRequired = 0;
if (strcmp(pszTextToReplace, "") == 0)
dwRC = ERROR_INVALID_PARAMETER;
else if (strcmp(pszTextToReplace, pszReplaceWith) != 0)
{
do
{
foundSeq = strstr(searchStart, pszTextToReplace);
if (foundSeq)
{
szRestOfStringLen = (strlen(foundSeq) - sztextToReplaceLen) + 1;
remainingSpace = dwSourceLen - (foundSeq - source);
dwSpaceRequired = szReplStrcLen + (szRestOfStringLen);
if (dwSpaceRequired > remainingSpace)
{
dwRC = ERROR_MORE_DATA;
}
else
{
restOfString = CMNUTIL_calloc(szRestOfStringLen, sizeof(CHAR));
strcpy_s(restOfString, szRestOfStringLen, foundSeq + sztextToReplaceLen);
strcpy_s(foundSeq, remainingSpace, pszReplaceWith);
strcat_s(foundSeq, remainingSpace, restOfString);
}
CMNUTIL_free(restOfString);
searchStart = foundSeq + szReplStrcLen; //search in the remaining str. (avoid loops when replWith contains textToRepl
}
} while (foundSeq && dwRC == NO_ERROR);
}
return dwRC;
}
char *replace(const char*instring, const char *old_part, const char *new_part)
{
#ifndef EXPECTED_REPLACEMENTS
#define EXPECTED_REPLACEMENTS 100
#endif
if(!instring || !old_part || !new_part)
{
return (char*)NULL;
}
size_t instring_len=strlen(instring);
size_t new_len=strlen(new_part);
size_t old_len=strlen(old_part);
if(instring_len<old_len || old_len==0)
{
return (char*)NULL;
}
const char *in=instring;
const char *found=NULL;
size_t count=0;
size_t out=0;
size_t ax=0;
char *outstring=NULL;
if(new_len> old_len )
{
size_t Diff=EXPECTED_REPLACEMENTS*(new_len-old_len);
size_t outstring_len=instring_len + Diff;
outstring =(char*) malloc(outstring_len);
if(!outstring){
return (char*)NULL;
}
while((found = strstr(in, old_part))!=NULL)
{
if(count==EXPECTED_REPLACEMENTS)
{
outstring_len+=Diff;
if((outstring=realloc(outstring,outstring_len))==NULL)
{
return (char*)NULL;
}
count=0;
}
ax=found-in;
strncpy(outstring+out,in,ax);
out+=ax;
strncpy(outstring+out,new_part,new_len);
out+=new_len;
in=found+old_len;
count++;
}
}
else
{
outstring =(char*) malloc(instring_len);
if(!outstring){
return (char*)NULL;
}
while((found = strstr(in, old_part))!=NULL)
{
ax=found-in;
strncpy(outstring+out,in,ax);
out+=ax;
strncpy(outstring+out,new_part,new_len);
out+=new_len;
in=found+old_len;
}
}
ax=(instring+instring_len)-in;
strncpy(outstring+out,in,ax);
out+=ax;
outstring[out]='\0';
return outstring;
}
Using only strlen from string.h
sorry for my English
char * str_replace(char * text,char * rep, char * repw){//text -> to replace in it | rep -> replace | repw -> replace with
int replen = strlen(rep),repwlen = strlen(repw),count;//some constant variables
for(int i=0;i<strlen(text);i++){//search for the first character from rep in text
if(text[i] == rep[0]){//if it found it
count = 1;//start searching from the next character to avoid repetition
for(int j=1;j<replen;j++){
if(text[i+j] == rep[j]){//see if the next character in text is the same as the next in the rep if not break
count++;
}else{
break;
}
}
if(count == replen){//if count equals to the lenght of the rep then we found the word that we want to replace in the text
if(replen < repwlen){
for(int l = strlen(text);l>i;l--){//cuz repwlen greater than replen we need to shift characters to the right to make space for the replacement to fit
text[l+repwlen-replen] = text[l];//shift by repwlen-replen
}
}
if(replen > repwlen){
for(int l=i+replen-repwlen;l<strlen(text);l++){//cuz replen greater than repwlen we need to shift the characters to the left
text[l-(replen-repwlen)] = text[l];//shift by replen-repwlen
}
text[strlen(text)-(replen-repwlen)] = '\0';//get rid of the last unwanted characters
}
for(int l=0;l<repwlen;l++){//replace rep with repwlen
text[i+l] = repw[l];
}
if(replen != repwlen){
i+=repwlen-1;//pass to the next character | try text "y" ,rep "y",repw "yy" without this line to understand
}
}
}
}
return text;
}
if you want strlen code to avoid calling string.h
int strlen(char * string){//use this code to avoid calling string.h
int lenght = 0;
while(string[lenght] != '\0'){
lenght++;
}
return lenght;
}
There you go....this is the function to replace every occurance of char x with char y within character string str
char *zStrrep(char *str, char x, char y){
char *tmp=str;
while(*tmp)
if(*tmp == x)
*tmp++ = y; /* assign first, then incement */
else
*tmp++;
// *tmp='\0'; -> we do not need this
return str;
}
An example usage could be
Exmaple Usage
char s[]="this is a trial string to test the function.";
char x=' ', y='_';
printf("%s\n",zStrrep(s,x,y));
Example Output
this_is_a_trial_string_to_test_the_function.
The function is from a string library I maintain on Github, you are more than welcome to have a look at other available functions or even contribute to the code :)
https://github.com/fnoyanisi/zString
EDIT:
#siride is right, the function above replaces chars only. Just wrote this one, which replaces character strings.
#include <stdio.h>
#include <stdlib.h>
/* replace every occurance of string x with string y */
char *zstring_replace_str(char *str, const char *x, const char *y){
char *tmp_str = str, *tmp_x = x, *dummy_ptr = tmp_x, *tmp_y = y;
int len_str=0, len_y=0, len_x=0;
/* string length */
for(; *tmp_y; ++len_y, ++tmp_y)
;
for(; *tmp_str; ++len_str, ++tmp_str)
;
for(; *tmp_x; ++len_x, ++tmp_x)
;
/* Bounds check */
if (len_y >= len_str)
return str;
/* reset tmp pointers */
tmp_y = y;
tmp_x = x;
for (tmp_str = str ; *tmp_str; ++tmp_str)
if(*tmp_str == *tmp_x) {
/* save tmp_str */
for (dummy_ptr=tmp_str; *dummy_ptr == *tmp_x; ++tmp_x, ++dummy_ptr)
if (*(tmp_x+1) == '\0' && ((dummy_ptr-str+len_y) < len_str)){
/* Reached end of x, we got something to replace then!
* Copy y only if there is enough room for it
*/
for(tmp_y=y; *tmp_y; ++tmp_y, ++tmp_str)
*tmp_str = *tmp_y;
}
/* reset tmp_x */
tmp_x = x;
}
return str;
}
int main()
{
char s[]="Free software is a matter of liberty, not price.\n"
"To understand the concept, you should think of 'free' \n"
"as in 'free speech', not as in 'free beer'";
printf("%s\n\n",s);
printf("%s\n",zstring_replace_str(s,"ree","XYZ"));
return 0;
}
And below is the output
Free software is a matter of liberty, not price.
To understand the concept, you should think of 'free'
as in 'free speech', not as in 'free beer'
FXYZ software is a matter of liberty, not price.
To understand the concept, you should think of 'fXYZ'
as in 'fXYZ speech', not as in 'fXYZ beer'
You can use strrep()
char* strrep ( const char * cadena,
const char * strf,
const char * strr
)
strrep (String Replace). Replaces strf with strr in cadena and returns the new string. You need to free the returned string in your code after using strrep.
Parameters:
cadena: The string with the text.
strf: The text to find.
strr: The replacement text.
Returns
The text updated wit the replacement.
Project can be found at https://github.com/ipserc/strrep

Resources