70,026
社区成员
发帖
与我相关
我的任务
分享free函数如何正确使用，求大神解答。。
这是我做的ACM的一道题，有个地方实在想不明白错在哪。。求大神帮看看
先上题：
It was said that when testing the first computer designed by von Neumann, people gave the following problem to both the legendary professor and the new computer: If the 4th digit of 2^n is 7, what is the smallest n? The machine and von Neumann began computing at the same moment, and von Neumann gave the answer first.
Now you are challenged with a similar but more complicated problem: If the K-th digit of M^n is 7, what is the smallest n?
Input
Each case is given in a line with 2 numbers: K and M (< 1,000).
Output
For each test case, please output in a line the smallest n.
You can assume:
The answer always exist.
The answer is no more than 100.
Sample Input
3 2
4 2
4 3
Sample Output
15
21
11
题目很简单，但用到大数乘法，以下是我的代码
疑问就出在我注释的那行代码上，由于乘法中间用到了加法，因此每次都会产生加和的临时变量，我想把这些内存free掉，但是一加上这行代码提交就报Runtime Error, 在本地运行没有问题，去掉后提交可以过，但是内存占用很大。想问问这个地方free使用哪里出错了？我free的变量都是前一次的sum结果。。还有最后main函数里面每次做乘法的临时变量内存也可以free，但是加了类似代码也报错。。自己实在想不明白，求大神们帮看看啊，新手分不多见谅
先上题：
It was said that when testing the first computer designed by von Neumann, people gave the following problem to both the legendary professor and the new computer: If the 4th digit of 2^n is 7, what is the smallest n? The machine and von Neumann began computing at the same moment, and von Neumann gave the answer first.
Now you are challenged with a similar but more complicated problem: If the K-th digit of M^n is 7, what is the smallest n?
Input
Each case is given in a line with 2 numbers: K and M (< 1,000).
Output
For each test case, please output in a line the smallest n.
You can assume:
The answer always exist.
The answer is no more than 100.
Sample Input
3 2
4 2
4 3
Sample Output
15
21
11
题目很简单，但用到大数乘法，以下是我的代码
#include<stdio.h>
#include<string.h>
#include<stdlib.h>
void reverse(char *s)
{
int len = strlen(s);
int i, j;
char c;
for (i = 0, j = len - 1; i < j; i++, j--)
{
c = *(s + i);
*(s + i) = *(s + j);
*(s + j) = c;
}
}
char *appendTailZero(char *s, int zeros)
{
int i, len = strlen(s);
char *r = malloc(len + zeros + 1);
for (i = 0; i < len; i++)
*(r + i) = *(s + i);
for (i = len; i < len + zeros; i++)
*(r + i) = '0';
*(r + len + zeros) = '\0';
return r;
}
char *add(char *s1, char *s2)
{
int l1 = strlen(s1);
int l2 = strlen(s2);
int len = l1 > l2 ? l1 : l2;
char *r = malloc(len + 2);
int i, prev = 0, a, b, sum;
for (i = 0; i < len; i++)
{
a = l1 - 1 - i >= 0 ? *(s1 + l1 - 1 - i) - '0' : 0;
b = l2 - 1 - i >= 0 ? *(s2 + l2 - 1 - i) - '0' : 0;
sum = a + b + prev;
*(r + i) = sum > 9 ? sum - 10 + '0' : sum + '0';
prev = sum > 9 ? 1 : 0;
}
if (prev)
{
*(r + len) = '1';
*(r + len + 1) = '\0';
}
else
*(r + len) = '\0';
reverse(r);
return r;
}
char *multiplyHelper(char *s1, int digit)
{
int i, res, prev = 0, len = strlen(s1);
char *r = malloc(len + 2);
if (!digit)
{
*r = '0';
*(r + 1) = '\0';
return r;
}
for (i = 0; i < len; i++)
{
res = (*(s1 + len - 1 - i) - '0') * digit + prev;
*(r + i) = res % 10 + '0';
prev = res / 10;
}
if (prev)
{
*(r + len) = prev + '0';
*(r + len + 1) = '\0';
}
else
*(r + len) = '\0';
reverse(r);
return r;
}
char *multiply(char *s1, char *s2)
{
if (strlen(s1) < strlen(s2))
{
char *temp;
temp = s1;
s1 = s2;
s2 = temp;
}
int l2 = strlen(s2);
int i;
char *t, *zt;
char *sum = malloc(2);
char *tp;
sum = "0";
for (i = 0; i < l2; i++)
{
tp = sum;
t = multiplyHelper(s1, *(s2 + l2 - 1 - i) - '0');
zt = appendTailZero(t, i);
sum = add(tp, zt);
free(t);
free(zt);
free(tp);//加入这行代码提交就报Runtime Error, 不加就Accepted,但是内存占用很大
}
return sum;
}
int main()
{
int k;
char m[5];
while (scanf("%d %s", &k, m) != EOF)
{
char *r = m;
int count = 1;
while (strlen(r) < k)
{
r = multiply(r, m);
count++;
}
r = r + strlen(r) - k;
while (*r != '7')
{
r = multiply(r, m);
r = r + strlen(r) - k;
count++;
}
printf("%d\n", count);
}
return 0;
}
疑问就出在我注释的那行代码上，由于乘法中间用到了加法，因此每次都会产生加和的临时变量，我想把这些内存free掉，但是一加上这行代码提交就报Runtime Error, 在本地运行没有问题，去掉后提交可以过，但是内存占用很大。想问问这个地方free使用哪里出错了？我free的变量都是前一次的sum结果。。还有最后main函数里面每次做乘法的临时变量内存也可以free，但是加了类似代码也报错。。自己实在想不明白，求大神们帮看看啊，新手分不多见谅
...全文
请发表友善的回复…
发表回复
心之痕 2014-01-10
- 打赏
- 举报
终于知道第一个free那边为什么会错了。。原来是去free了一个常量字符串，我以为malloc之后就能调free,中间指向常量0后那个malloc（2）完全就没用到了
ALNG的思路真的不错，针对题目给出了特定的省时省内存的解法，收获不少，感谢
。。其实我原先写了个通用点的大数加乘法看这题有用到就拷过来用了，好了结贴了
。。其实我原先写了个通用点的大数加乘法看这题有用到就拷过来用了，好了结贴了
肖邦之离歌 2014-01-10
- 打赏
- 举报
free函数只能释放通过malloc、realloc、calloc函数开辟的内存，你的代码里sum开始是malloc的，但是你紧接着就把sum指向了“0”（这里实际上已经造成了内存泄漏），后面tp又指向了sum指向的地址即“0”，这样的地址不是通过上述三个函数分配的，所以会报错。
孩皮妞野 2014-01-10
- 打赏
- 举报
闲的无聊，我实现了一遍，你先自己做在用我的对照吧。
好久不写C了，尽量按照记忆中的C写的，难保没有露马脚的地方 :)
/*
It was said that when testing the first computer designed by von Neumann, people gave the
following problem to both the legendary professor and the new computer: If the 4th digit
of 2^n is 7, what is the smallest n? The machine and von Neumann began computing at the
same moment, and von Neumann gave the answer first.
Now you are challenged with a similar but more complicated problem: If the K-th digit of
M^n is 7, what is the smallest n?
Input
Each case is given in a line with 2 numbers: K and M (< 1,000).
Output
For each test case, please output in a line the smallest n.
You can assume:
The answer always exist.
The answer is no more than 100.
Sample Input
3 2
4 2
4 3
Sample Output
15
21
11
*/
#include<stdio.h>
#include<string.h>
#include<stdlib.h>
int find_n(int k, int m);
int main()
{
int k,m;
while (scanf("%d %d", &k, &m) == 2)
printf("%d\n", find_n(m,k));
return 0;
}
/* test if the digit'th(starting from 0) digit of num's decimal representation is 7
* return 1 when test true, false otherwise.
*
* eg. test7(157434,2) is true, while test7(157432,3) is false
*
* note: num should be between 0 and 10^9-1
* and digit should be 0 to 8
*/
int test7(long num, int digit)
{
char _[10];
sprintf(_,"%ld", num);
return _[digit]=='7';
}
/* for debug only.
* I leave it behind because it sheds some light on how the large number
* should be interpreted
*
*/
void dump(long * p, int cnt)
{
printf("%d", p[cnt-1]);
for(int i=cnt-2; i>=0; --i)
printf("%09d",p[i]);
printf("\n");
}
void multiply(long* p, int* cnt_element, int m);
/* given integer k and m, find the smallest integer n such that
* the k'th digit of m^n (decimal representation) is 7
*
*/
int find_n(int k, int m)
{
// sizeof(long) is guaranteed to be 4
long array[34];
int cnt=1, // number of elements currently filled
n=1; // the value that we are looking for
array[0]=m; // we will put least significant int(s) in element 0
--k; // so that first digit is the 0th digit
while( cnt< k/9 || !test7(array[k/9], k%9) )
{
++n;
multiply(array, &cnt, m);
}
return n;
}
/* our ad hoc large number multiplication function
* param: p: points to array of long's which represent a large number. note the least
* significant element is the 0th element and each element should be between
* 0 and 10^9-1;
* pcnt: points to number of elements actually used in the above array
* m: number to multiply into the large number(no more than 100 according to problem description)
*
*
*/
void multiply(long* p, int* pcnt, int m)
{
long long v=0;
long carry=0;
for(int i=0; i<*pcnt; ++i)
{
v=carry+(long long)p[i]*m;
p[i]=(long)(v%1000000000);
carry=(long)(v/1000000000);
}
if(carry)
p[(*pcnt)++]=carry;
}
孩皮妞野 2014-01-10
- 打赏
- 举报
讨厌的CSDN不许连续回复超过3次！
回到这题，如果没有聪明的方法(估计有), 就用你的暴力法，那你的乘法实现还是太低效率了，即使写对了也10有8、9会timeout. 其实根本不需要那么多次动态分配内存，
已知M<1000, n<100, 所以M^n不大于10^300， 你直接在堆上开一个300的数组，就对它操作，这样一次malloc和free都不用，会大大提高你的程序的效率。
还可以再提高，你用一个字符来表示一个十进制的数位，这当然可以得到正确的结果，但效率还可以提高。
可以相信目标机器不会低于32位处理器，更准确的说法应该是目标机器上的c编译器的sizeof(int)不会低于4. 你完全可以用size_t或者int数组(instead of char[]), 在int间用10^9进位，每个int降级到10^9进位，即如果一个整数存的是10^9-1， 对它increment后，它会归0，而高位整数进一......希望你听懂了我的意思。这样的好处是计算更紧凑而保留了快速检查第k位是否为7的能力......
如果你能看懂我上面的意思，建议你用那个思路再实现一遍，甚至对比实现一遍，比较两者的效率，假定可以通过ACM的机器检测。
哦，你没必要定义乘法(*)，定义自乘(*=)就好了。形如:
void multiply_by(int self[], int num);
其中的self是一个足够装下10^300次方，成员间10^9进位的整型数组（啰嗦一下，每个成员装10^9, 10^300需要34个这样的int, 可以想一下你原来用成员用char，成员间10进位的方案要用300个元素, 哪一个会更快？)
你先试试，搞不定再问吧。
大奶兔白糖 2014-01-10
- 打赏
- 举报
第一次进循环的时候，sum="0",tp=sum，这时候tp指向了常量字符串“0”，这种怎么能free呢？
心之痕 2014-01-10
- 打赏
- 举报
谢谢你的解答，下面两个回复我能理解，这里还不是很清楚，以我的理解，在第一次进入循环开始，tp和sum都指向0字符串，而经过加和以后，sum指向add函数返回的新的内存区域，而我没有再tp=sum,这时候tp应该还是指向0字符串啊？难道这时tp和sum都是指向add后的新的内存区域，我free(tp)把加和的结果free掉了？
孩皮妞野 2014-01-10
- 打赏
- 举报
说一说你的multiply的设计，每一个函数原型都是一份合同，你的mutiply函数这份合同是这样的:
s1指向乘数1字符数组的指针，s2指向乘数2字符数组的指针；
返回指向存放两者相乘的运算结果的字符数组的指针，这个字符数组是动态用malloc从堆中分配的，在使用完毕后free之是调用者的责任.
再看你的代码
while (strlen(r) < k)
{
r = multiply(r, m);
count++;
}
孩皮妞野 2014-01-10
- 打赏
- 举报
int main()
{
int k;
char m[5];
while (scanf("%d %s", &k, m) != EOF)
{
char *r = m;
int count = 1;
while (strlen(r) < k)
{
char * tmp=multiply(r, m);
if(r!=m)
free(r);
r = tmp;
count++;
}
r += strlen(r) - k;
while (*r != '7')
{
char *tmp=multiply(r, m);
if(r!=m) // you should be able to skip this line!!
free(r);
r=tmp;
r += strlen(r) - k;
count++;
}
printf("%d\n", count);
}
return 0;
}
孩皮妞野 2014-01-10
- 打赏
- 举报
char *sum = malloc(2);
char *tp;
sum = "0";
for (i = 0; i < l2; i++)
{
tp = sum;
t = multiplyHelper(s1, *(s2 + l2 - 1 - i) - '0');
zt = appendTailZero(t, i);
sum = add(tp, zt);
free(t);
free(zt);
}
return sum;
}
