2015-01-11 57 views
0

我有以下问题,在我的代码中,我试图在HTTP协议中构建响应头。我不明白为什么我的代码不起作用。在缓冲功能构建响应内容,并将其写入套接字文件描述符:C字符串构建

void write_fd(struct http_response* response, int client_socket) 
{ 
    int length = strlen(response->file_content + MAX_HEADER_LENGTH); 
    char response_content[length]; 
    response_content[0] = '\0'; 

    printf("-- Descriptor %i, start responding\n", client_socket); 

    write_fd_resp_line(response, response_content); 
    //printf("%s\n", response_content); - "HTTP/1.1 GET OK\n" 
    write_fd_date(response_content); 
    //printf("%s\n", response_content); - Segmentation Fault 
    write_fd_server_name(response_content); 
    write_fd_con_type(response, response_content); 
    write_fd_doc_content(response, response_content); 

    int sended = 0; 
    int content_length = strlen(response_content) + 1; 
    int n; 
    while(sended != content_length) { 
     n = write(client_socket, response_content + sended, content_length - sended); 
     if(n <= 0) break; 
     sended += n; 
     printf("-- Descriptor - %i, sended %i/%i\n", client_socket, sended, content_length); 
    } 

}

,但是当我变了:

char response_content[length]; 

char* response_content = malloc(length); 

功能工作,服务器写入响应内容到套接字,但之后,我得到分段错误。我不明白为什么。

功能的模式write_fd_*类似于:

void write_fd_resp_line(http_response* response, char* response_content) 
{ 
    char *tmp; 
    char code_str[4]; 
    tmp = (char*) get_status_code_name(response->code); 
    snprintf(code_str, 4, "%d", response->code); 
    strcat(response_content, HTTP_VERSION); 
    strcat(response_content, " "); 
    strcat(response_content, code_str); 
    strcat(response_content, " "); 
    strcat(response_content, tmp); 
    strcat(response_content, "\n"); 
} 
+5

is this typo'strlen(response-> file_content + MAX_HEADER_LENGTH);' - >'strlen(response-> file_content)+ MAX_HEADER_LENGTH;'? –

+0

@iharob我有些怀疑这是OP难题的重要来源。 – WhozCraig

+0

你是对的!我没有看到! – kris14an

回答

4

你有一个错字,但有趣的代码编译为确保

int length = strlen(response->file_content + MAX_HEADER_LENGTH); 

应该

int length = strlen(response->file_content) + MAX_HEADER_LENGTH; 

的代码编译的原因是因为这意味着

response->file_content + MAX_HEADER_LENGTH 

是通过response->file_content指针由MAX_HEADER_LENGTH递增,这是有效的,但很可能不正确,并且分割falult非常可能的原因。