Strings in C

[chrisa128]

Sorry about this question being so simple, I am just too tired to think straight.

If I wanted to compare text entered by the user would I have to check it bit by bit like the following

unsigned strcmp(unsigned usr_str,unsigned cmp_str)
{
unsigned c, u;

c = strlen(usr_str);
while (u =< c)
{
If (usr_str != cmp_str) return 0;
u++;
}
return 1;
}

Please help me quick as I will fall asleep soon....

[drizzt]

Certainly this is not the best strcmp procedure i have seen... First of all both strings passed to the procedure should be pointers and not unsigned vars... and you can check the end of a string in a faster way, simply checking the character '\0'.

Here is my strcmp routine:

Code: Select all

char strcmp(const char *s1, const char *s2)
{
   while ((*s2 != '\0') && (*s1==*s2))
   {
      s1++;
      s2++;
   }
   return (*s1 - *s2);
}

[Slasher]

You have the right idea, but this is better

int strcmp(char *str1,char *str2)
{
/*loops till one of the chars do not match*/

while(*str1++ == *str2++ );

if(*str1==*str2)
return 0;
else
if(*str1 < *str2)
return -1;
else
return 1;
}

try this, should work!
(ps Drizzt replied before i finished )

[klalafuda klalafu]

--- cut ---
1. while(*str1++ == *str2++ );
2. if(*str1==*str2)
...
--- cut ---

now immagine, that str1 == str2 == { 'a', b', 'c' '\0'};

so, at the line 2 you will reference an unallocated memory ahead of EOS char ('\0'). likely, nothing's too serious should happen, but possibly it's SIGSEGV

// wbr

[Pype.Clicker]

the 'segfault will probably be unfrequent, but i agree it could occur.
Moreover, with slasher's code, if you have
str1 = "TST\0a"; str2 = TST\0b";

the two strings will appear distinct (the they'll end with 'a'<'b') while they are equal at 'strcmp' semantic (strings end at first 0 character)

[chrisa128]

Thanks all, now I got another problem because I am using a structure for allmy Virtual Consoles. When I try to pass the field in the structure like the following

strcmp(_vcon[_curr_vc].cmd_buf[], strcmd1);

I get the passing interger as pointer warning which messes it up completely. I am going to try and replace it with

strcmp(_vcon[_curr_vc].cmd_buf, strcmd1);

as soon as I get to the computer which I use for development but it would saveme time if you could advise, me on this.

Thanks once again for your help so far, esspecially about the segfault, I will have a look at that later.

Chris

[klalafuda klalafu]

well, if finally walking the way above, WATCOM C RTL gives the following:

--- cut ---
strcmp(const char *s, const char *t ) {
for( ; *s == *t; s++, t++ )
if( *s == '\0' )
return 0;
return (*s - *t);
}
--- cut ---



// wbr

[Whatever5k]

strcmp(const char *s, const char *t ) {
for( ; *s == *t; s++, t++ )
if( *s == '\0' )
return 0;
return (*s - *t);
}

I'm not really sure if this strcmp() function is really good...
It's the right idea, but I wouldn't increment the "original" strings since this may cause problems...
for example:

Code: Select all

int isEqual;
char bla[] = "Hello";
char bla2[] = "Hello2";

printf("*bla = %c\n", *bla);
isEqual = strcmp(bla, bla2);
printf("*bla = %c\n", *bla);

The first printf() call would print a 'H' as it should. The second call would print a 'o' since you modified the pointer by incrementing it...
better is:

Code: Select all

int strcmp(const char *s, const char *t)
{
char *sPointer = s, *tPointer = t;
for (...)
...
}

best regards,
A. Blessing

[klalafuda klalafu]

sorry, but you'r wrong the case above would take place if the function was defined like:

strcmp(const char **s, const char **t );

// wbr

[Whatever5k]

You're right, I messed up some things...thanks