smink的主机名正则表达式没有遵守主机名中各个标签长度的限制。有效主机名中的每个标签长度不能超过63个字节。

ValidHostnameRegex="^([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])\
(\.([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}[a-zA-Z0-9]))*$"

请注意，第一行末尾的反斜杠(上面)是用于分隔长行的Unix shell语法。它不是正则表达式本身的一部分。

下面是一行中单独的正则表达式:

^([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])(\.([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}[a-zA-Z0-9]))*$

您还应该单独检查主机名的总长度不能超过255个字符。更多信息，请咨询RFC-952和RFC-1123。

Regarding IP addresses, it appears that there is some debate on whether to include leading zeros. It was once the common practice and is generally accepted, so I would argue that they should be flagged as valid regardless of the current preference. There is also some ambiguity over whether text before and after the string should be validated and, again, I think it should. 1.2.3.4 is a valid IP but 1.2.3.4.5 is not and neither the 1.2.3.4 portion nor the 2.3.4.5 portion should result in a match. Some of the concerns can be handled with this expression:

grep -E '(^|[^[:alnum:]+)(([0-1]?[0-9]{1,2}|2[0-4][0-9]|25[0-5])\.){3}([0-1]?[0-9]{1,2}|2[0-4][0-9]|25[0-5])([^[:alnum:]]|$)' 

The unfortunate part here is the fact that the regex portion that validates an octet is repeated as is true in many offered solutions. Although this is better than for instances of the pattern, the repetition can be eliminated entirely if subroutines are supported in the regex being used. The next example enables those functions with the -P switch of grep and also takes advantage of lookahead and lookbehind functionality. (The function name I selected is 'o' for octet. I could have used 'octet' as the name but wanted to be terse.)

grep -P '(?<![\d\w\.])(?<o>([0-1]?[0-9]{1,2}|2[0-4][0-9]|25[0-5]))(\.\g<o>){3}(?![\d\w\.])'

如果IP地址在一个包含句子形式文本的文件中，那么点号的处理实际上可能会产生错误的否定，因为句号可以跟在后面，而不是点号符号的一部分。上面的一个变体可以修复这个问题:

grep -P '(?<![\d\w\.])(?<x>([0-1]?[0-9]{1,2}|2[0-4][0-9]|25[0-5]))(\.\g<x>){3}(?!([\d\w]|\.\d))'

def isValidHostname(hostname):

    if len(hostname) > 255:
        return False
    if hostname[-1:] == ".":
        hostname = hostname[:-1]   # strip exactly one dot from the right,
                                   #  if present
    allowed = re.compile("(?!-)[A-Z\d-]{1,63}(?<!-)$", re.IGNORECASE)
    return all(allowed.match(x) for x in hostname.split("."))

值得注意的是，大多数语言都有这样的库，它们通常内置在标准库中。这些库的更新频率可能比你四年前从Stack Overflow的答案中复制而忘记的代码要高得多。当然，他们通常也会把地址解析成一些可用的形式，而不仅仅是给你一堆组的匹配。

例如，在(POSIX) C中检测和解析IPv4:

#include <arpa/inet.h>
#include <stdio.h>

int main(int argc, char *argv[]) {
  for (int i=1; i!=argc; ++i) {
    struct in_addr addr = {0};
    printf("%s: ", argv[i]);
    if (inet_pton(AF_INET, argv[i], &addr) != 1)
      printf("invalid\n");
    else
      printf("%u\n", addr.s_addr);
  }
  return 0;
}

显然，如果您试图查找聊天消息中的所有有效地址，这样的函数将不起作用——但即使这样，使用一个简单但过于热心的正则表达式来查找潜在的匹配，然后使用库来解析它们可能会更容易。

例如，在Python中:

>>> import ipaddress
>>> import re
>>> msg = "My address is 192.168.0.42; 192.168.0.420 is not an address"
>>> for maybeip in re.findall(r'\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}', msg):
...     try:
...         print(ipaddress.ip_address(maybeip))
...     except ValueError:
...         pass

>>> my_hostname = "testhostn.ame"
>>> print bool(re.match("^(([a-zA-Z]|[a-zA-Z][a-zA-Z0-9\-]*[a-zA-Z0-9])\.)*([A-Za-z]|[A-Za-z][A-Za-z0-9\-]*[A-Za-z0-9])$", my_hostname))
True
>>> my_hostname = "testhostn....ame"
>>> print bool(re.match("^(([a-zA-Z]|[a-zA-Z][a-zA-Z0-9\-]*[a-zA-Z0-9])\.)*([A-Za-z]|[A-Za-z][A-Za-z0-9\-]*[A-Za-z0-9])$", my_hostname))
False
>>> my_hostname = "testhostn.A.ame"
>>> print bool(re.match("^(([a-zA-Z]|[a-zA-Z][a-zA-Z0-9\-]*[a-zA-Z0-9])\.)*([A-Za-z]|[A-Za-z][A-Za-z0-9\-]*[A-Za-z0-9])$", my_hostname))
True

这个怎么样?

([0-9]{1,3}\.){3}[0-9]{1,3}