C getline() returns -1 在 EOF 之前

C getline() returns -1 before EOF

我在 c 程序中使用 getline() 函数从文件中读取行,将读取的信息放入结构中,创建一个以该结构作为参数的线程,然后重复。 但是,它 getline() returns -1 在 EOF 之前,在提供的文件中 "id: test2" 下的第二组参数之后。
由于 print errno 在 gdb returns 0 中,我假设没有错误并且 getline() 不知何故似乎相信它已经达到了 EOF。
创建线程部分已被注释掉,因为它与手头的问题无关。

代码:

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <signal.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <pthread.h>
#include <sys/time.h>

#define BUFFER_SIZE 256

time_t cur;
char* keypath = "home/chy/.ssh/id_rsa";
int logfreq = 1;
int hashfreq = 180;

struct getArgs
{
  const char* id;
  const char* hostaddr;
  uint16_t port;
  const char* uname;
  const char* path;
};

int main(int argc, char **argv)
{
  int th_max = 5;
  int th_count = 0;
  pthread_t* ths = (pthread_t*) malloc(th_max * sizeof(pthread_t));

  FILE* fp;
  fp = fopen("config.txt", "r");

  char* line = NULL;
  char buf[128];
  int r = 0;
  size_t len = 0;
  ssize_t read;

  system("mkdir logs && chmod -R a+r logs");

  while(1)
  {
    pthread_t th;
//Get ID
    while((read = getline(&line, &l, fp)) != -1)
    {
      if(!(read > 0))
        continue;
      if(sscanf(line, "id: %[^\n]\n", buf) == 1)
        break;
      if(sscanf(line, "keypath: %[^\n]\n", buf) == 1)
        keypath = strdup(buf);
      else if(sscanf(line, "logfreq: %d\n", &r) == 1)
        logfreq = r;
      else if(sscanf(line, "hashfreq: %d\n", &r) == 1)
        hashfreq = r;
    }
    if(read == -1)
      break;
    struct getArgs* args = (struct getArgs*)malloc(sizeof(struct getArgs));
    args->id = strdup(buf);
//Get Host Address
    if(getline(&line, &len, fp) == -1)
    {
      free_args(args);
      break;
    }
    if(sscanf(line, "hostaddr: %[^\n]\n", buf) != 1)
    {
      free_args(args);
      continue;
    }
    args->hostaddr = strdup(buf);
//Get Port Number
    if(getline(&line, &len, fp) == -1)
    {
      free_args(args);
      break;
    }
    if(sscanf(line, "port: %d\n", &r) != 1)
    {
      free_args(args);
      continue;
    }
    args->port = r;
//Get Username
    if(getline(&line, &len, fp) == -1)
    {
      free_args(args);
      break;
    }
    if(sscanf(line, "username: %[^\n]\n", buf) != 1)
    {
      free_args(args);
      continue;
    }
args->uname = strdup(buf);
//def Path
    if(getline(&line, &len, fp) == -1)
    {
      free_args(args);
      break;
    }
    if(sscanf(line, "path: %[^\n]\n", buf) != 1)
    {
      free_args(args);
      continue;
    }
    args->path = strdup(buf);

//    int err = pthread_create(&th, NULL, &getFiles, args);

    if(th_count > th_max)
    {
      th_max *= 2;
      ths = (pthread_t*)realloc(ths, th_max * sizeof(pthread_t));
    }
    *(ths+th_count * sizeof(pthread_t)) = th;
    th_count++;
  }

  fclose(fp);

  while(1);
}

文件:

keypath: /home/username/.ssh/id_rsaNOPASSWORD
id: test1
hostaddr: XXX.XXX.XXX.XXX
port: 22
username: hpc
path: /home/hpc/
id: test2
hostaddr: XXX.XXX.XXX.XXX
port: 22
username: hpc
path: /home/hpc/
id: test3
hostaddr: XXX.XXX.XXX.XXX
port: 22
username: hpc
path: /home/hpc/
id: test4
hostaddr: XXX.XXX.XXX.XXX
port: 3844
username: uname
path: /home/uname/hpc/

最后一次getline()之前的文件指针内容:

{_flags = -72539000,
  _IO_read_ptr = 0x7ffff7ff80d0 "    \nid: test3\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test4\nhostaddr: XXX.XXX.XXX.XXX\nport: 3844\nusername: uname\npath: /home/uname/hpc/\n\n",
  _IO_read_end = 0x401d10 "H1l$0L1d$0H5-7[=13=]6 ",
  _IO_read_base = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_write_base = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_write_ptr = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_write_end = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_buf_base = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"..., _IO_buf_end = 0x7ffff7ff9000 "P07777", _IO_save_base = 0x0, _IO_backup_base = 0x0, _IO_save_end = 0x0, _markers = 0x0, _chain = 0x7ffff7bbb880, _fileno = 7, _flags2 = 0, _old_offset = 0, _cur_column = 0,
  _vtable_offset = 0 '[=13=]0', _shortbuf = "", _lock = 0x603120, _offset = -1, __pad1 = 0x0, __pad2 = 0x603130, __pad3 = 0x0, __pad4 = 0x0, __pad5 = 0, _mode = -1, _unused2 = '[=13=]0' <repeats 19 times>}

之后:

{_flags = -72538984,
  _IO_read_ptr = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_read_end = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_read_base = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_write_base = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_write_ptr = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_write_end = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"...,
  _IO_buf_base = 0x7ffff7ff8000 "keypath: /home/username/.ssh/id_rsaNOPASSWORD\n    \nid: test1\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /home/hpc/\n    \nid: test2\nhostaddr: XXX.XXX.XXX.XXX\nport: 22\nusername: hpc\npath: /ho"..., _IO_buf_end = 0x7ffff7ff9000 "P07777", _IO_save_base = 0x0, _IO_backup_base = 0x0, _IO_save_end = 0x0, _markers = 0x0, _chain = 0x7ffff7bbb880, _fileno = 7, _flags2 = 0, _old_offset = 0, _cur_column = 0,
  _vtable_offset = 0 '[=14=]0', _shortbuf = "", _lock = 0x603120, _offset = -1, __pad1 = 0x0, __pad2 = 0x603130, __pad3 = 0x0, __pad4 = 0x0, __pad5 = 0, _mode = -1, _unused2 = '[=14=]0' <repeats 19 times>}

我也尝试用 fgets() 而不是 getline() 实现相同的结果,但结果相同。
可能导致此问题的原因是什么?

编辑:
我有 运行 同一个文件的代码,除了没有多余的 spaces/blank 行以及切换顺序以查看是否有任何影响,但结果是相同的。
我还可以确认 test1 和 test2 的结构已正确填充。

{id = 0x6032e0 "test1", hostaddr = 0x603300 "XXX.XXX.XXX.XXX", port = 22, uname = 0x603320 "hpc", path = 0x603340 "/home/hpc/"}
{id = 0x603390 "test2", hostaddr = 0x6033b0 "XXX.XXX.XXX.XXX", port = 22, uname = 0x6033d0 "hpc", path = 0x6033f0 "/home/hpc/"}

编辑: 代码和文件的版本更改

sscanf(line, "id: %[^\n]\n", buf) != 1

不会匹配空白行,导致您的读取循环在该点不同步。随着 id: test2 被吞噬,更多的问题从那里级联,最终你遇到了 EOF。

[      Matches a nonempty sequence of characters ...

除了评论中讨论的 unsigned/signed type 不匹配之外,您遇到的主要问题是测试中使用的逻辑冲突,以及对 getlinesscanf 的多个松散组织调用,您将在其中测试 if ((read = getline ... || sscanf ...),其中 getlinesscanf(在输入或匹配失败时)可以 return EOF (-1) 无法判断问题出在哪里。要清理逻辑,您需要读取配置文件,您 (1) 读取行 - 一次,(2) 解析标记和值 ,然后 (3) 测试 tagvalue 并采取所需的措施。

坚持这种方法,并忽略与您的读取问题无关的线程代码,我 debugged/re-wrote 您的输入例程为您提供了一个以理智的方式处理读取和分离的方法示例。这样做时,我将所有设置收集在指向结构的指针数组中(例如 struct getArgs **args;)。您可能不需要使用您的 pthread 方案将每个结构都收集到一个结构数组中,但出于示例目的,已进行了更改。

我还在下面提供了第二个示例,该示例显示了一种额外的方法,需要进行更多检查以确保在 id、hostaddr、端口、用户名和路径中读取您的值团体。第一个简单示例将处理跳过输入文件中的空行:

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdint.h>

#define NARGS  32
#define TSIZE  16
#define VSIZE 256

char *keypath = "home/chy/.ssh/id_rsa";
int logfreq = 1;
int hashfreq = 180;

struct getArgs {
    const char *id;
    const char *hostaddr;
    uint16_t port;
    const char *uname;
    const char *path;
};

void *xcalloc (size_t n, size_t s);
void *xrealloc_dp (void *ptr, size_t *n);

int main (int argc, char **argv)
{
    struct getArgs **args = NULL;
    char *line = NULL;
    size_t len = 0;
    size_t idx = 0;
    size_t nargs = NARGS;
    size_t i;
    ssize_t read;
    FILE *fp = argc > 1 ? fopen (argv[1], "r") : fopen ("config.txt", "r");

    if (!fp) { fprintf (stderr, "file open failed.\n"); return 1; }

    /* allocate NARGS pointer to struct getArgs */
    args = xcalloc (NARGS, sizeof *args);

    /* read each line in file */
    while ((read = getline (&line, &len, fp)) != -1) {

        if (read == 1) continue;    /* skip blank lines */

        char tag[TSIZE] = {0};
        char val[VSIZE] = {0};

        /* separate tag and value (val) */
        if (sscanf (line, "%s %[^\n]%*c", tag, val) != 2) {
            fprintf (stderr, "error: sscanf conversion failed.\n");
            break;
        }

        /* handle keypath, logfreq, hashfreq */
        if (strcmp (tag, "keypath:") == 0) {
            keypath = strdup (val);
            continue;
        }

        if (strcmp (tag, "logfreq:") == 0) {
            logfreq = (int)strtol (val, NULL, 10);
            continue;
        }

        if (strcmp (tag, "hashfreq:") == 0) {
            hashfreq = (int)strtol (val, NULL, 10);
            continue;
        }

        /* allocate space for args[idx] if "id:", then
        handle id, hostaddr, port, username, path */
        if (strcmp (tag, "id:") == 0) {
            args[idx] = calloc (1, sizeof **args);
            args[idx]->id = strdup (val);
            continue;
        }

        if (strcmp (tag, "hostaddr:") == 0) {
            args[idx]->hostaddr = strdup (val);
            continue;
        }

        if (strcmp (tag, "port:") == 0) {
            args[idx]->port = (uint16_t)strtoul (val, NULL, 10);
            continue;
        }

        if (strcmp (tag, "username:") == 0) {
            args[idx]->uname = strdup (val);
            continue;
        }

        /* increment idx on path */
        if (strcmp (tag, "path:") == 0)
            args[idx++]->path = strdup (val);

        if (idx == nargs) /* check idx, realloc */
            args = xrealloc_dp (args, &nargs);
    }
    fclose (fp);

    printf ("\n keypath  : %s\n logfreq  : %d\n hashfreq : %d\n",
            keypath, logfreq, hashfreq);

    for (i = 0; i < idx; i++)
        printf ("\n id       : %s\n hostaddr : %s\n port     : %hu\n"
                " username : %s\n path     : %s\n", args[i]->id, args[i]->hostaddr,
                args[i]->port, args[i]->uname, args[i]->path);

    for (i = 0; i < idx; i++)
        free (args[i]);
    free (args);

    return 0;
}

/* calloc with error check, exit on failure */
void *xcalloc (size_t n, size_t s)
{
    register void *memptr = calloc (n, s);
    if (memptr == 0) {
        fprintf (stderr, "xcalloc() error: virtual memory exhausted.\n");
        exit (EXIT_FAILURE);
    }

    return memptr;
}

/* reallocate memory for a double-pointer from 'n' to 2 * 'n'
 * returns pointer to reallocated block on success, exit on
 * failure
 */
void *xrealloc_dp (void *ptr, size_t *n)
{
    void **p = ptr;
    void *tmp = realloc (p, 2 * *n * sizeof tmp);
    if (!tmp) {
        fprintf (stderr, "xrealloc_dp() error: virtual memory exhausted.\n");
        exit (EXIT_FAILURE);
    }
    p = tmp;
    memset (p + *n, 0, *n * sizeof tmp); /* set new pointers NULL */
    *n *= 2;

    return p;
}

xcallocxrealloc_dp函数只是对callocrealloc(双指针)进行错误检查的函数,以保持主体和逻辑代码清晰。我 运行 以下测试输入的代码并收到以下输出:

输入测试文件(带空行)

$ cat ../dat/idhostaddrport.txt
keypath: /home/username/.ssh/id_rsaNOPASSWORD
logfreq: 2
hashfreq: 250

id: test1
hostaddr: XXX.XXX.XXX.XXX
port: 221
username: hpc1
path: /home/hpc1/

id: test2
hostaddr: XXX.XXX.XXX.XXX
port: 222
username: hpc2
path: /home/hpc2/

id: test3
hostaddr: XXX.XXX.XXX.XXX
port: 223
username: hpc3
path: /home/hpc3/

id: test4
hostaddr: XXX.XXX.XXX.XXX
port: 3844
username: uname
path: /home/uname/hpc/

输出

$ ./bin/getline_sscanf_dbg ../dat/idhostaddrport.txt

 keypath  : /home/username/.ssh/id_rsaNOPASSWORD
 logfreq  : 2
 hashfreq : 250

 id       : test1
 hostaddr : XXX.XXX.XXX.XXX
 port     : 221
 username : hpc1
 path     : /home/hpc1/

 id       : test2
 hostaddr : XXX.XXX.XXX.XXX
 port     : 222
 username : hpc2
 path     : /home/hpc2/

 id       : test3
 hostaddr : XXX.XXX.XXX.XXX
 port     : 223
 username : hpc3
 path     : /home/hpc3/

 id       : test4
 hostaddr : XXX.XXX.XXX.XXX
 port     : 3844
 username : uname

您应该努力的一件事是进一步验证您的输入的方法。确保如果您读取 id,您也读取了每个 id 的剩余值 hostaddrpath。添加最少检查的一种简单变体是对读取循环的以下更改:

    /* read each line in file */
    while ((read = getline (&line, &len, fp)) != -1) {

        if (read == 1) continue;    /* skip blank lines */

        char tag[TSIZE] = {0};
        char val[VSIZE] = {0};

        /* separate tag and value (val) */
        if (sscanf (line, "%s %[^\n]%*c", tag, val) != 2) {
            fprintf (stderr, "error: sscanf conversion failed.\n");
            break;
        }

        /* handle keypath, logfreq, hashfreq */
        if (strcmp (tag, "keypath:") == 0) {
            keypath = strdup (val);
            continue;
        }

        if (strcmp (tag, "logfreq:") == 0) {
            logfreq = (int)strtol (val, NULL, 10);
            continue;
        }

        if (strcmp (tag, "hashfreq:") == 0) {
            hashfreq = (int)strtol (val, NULL, 10);
            continue;
        }

        /* allocate space for args[idx] if "id:", then
        handle id, hostaddr, port, username, path */
        if (strcmp (tag, "id:") == 0) {

            args[idx] = calloc (1, sizeof **args);
            args[idx]->id = strdup (val);
            size_t tagseq = 0;

            while ((read = getline (&line, &len, fp)) != -1) {

                if (read == 1) continue;    /* skip blank lines */

                /* separate tag and value (val) */
                if (sscanf (line, "%s %[^\n]%*c", tag, val) != 2) {
                    fprintf (stderr, "error: sscanf conversion failed.\n");
                    break;
                }

                if (strcmp (tag, "hostaddr:") == 0) {
                    if (tagseq != 0) { 
                        fprintf (stderr, "error: tagseq failed for hostaddr.\n");
                        exit (EXIT_FAILURE);
                    }
                    args[idx]->hostaddr = strdup (val);
                    tagseq++;
                    continue;
                }

                if (strcmp (tag, "port:") == 0) {
                    if (tagseq != 1) { 
                        fprintf (stderr, "error: tagseq failed for port.\n");
                        exit (EXIT_FAILURE);
                    }
                    args[idx]->port = (uint16_t)strtoul (val, NULL, 10);
                    tagseq++;
                    continue;
                }

                if (strcmp (tag, "username:") == 0) {
                    if (tagseq != 2) { 
                        fprintf (stderr, "error: tagseq failed for username.\n");
                        exit (EXIT_FAILURE);
                    }
                    args[idx]->uname = strdup (val);
                    tagseq++;
                    continue;
                }

                /* increment idx on path */
                if (strcmp (tag, "path:") == 0) {
                    if (tagseq != 3) { 
                        fprintf (stderr, "error: tagseq failed for path.\n");
                        exit (EXIT_FAILURE);
                    }
                    args[idx++]->path = strdup (val);
                    break;
                }
            }
        }

        if (idx == nargs) /* check idx, realloc */
            args = xrealloc_dp (args, &nargs);
    }
    fclose (fp);

看看两者,如果您有任何其他问题,请告诉我。