使用netcat作为简单的CGI服务器
我试图使用以下脚本创建一个简单的 CGI 服务器（来自 http ://www.jfranken.de/homepages/johannes/vortraege/netcat_inhalt.en.html）。这更多的是关于它不起作用的技术原因，而不是这是否是一个好主意：
#!/bin/bash
export port=${port:-$1} # inherit $1 as $port
/usr/local/bin/nc -l -p $port -e $0 & # await further connections on this port
echo "hi"
有人知道为什么这不能正常工作吗？第一次执行时，它会在终端上打印“hi”，这是我所期望的。然后，我希望 netcat 在指定的端口上等待，准备生成该脚本的新副本，其中包含与网络套接字关联的 STDOUT（由于 -e 标志）。因此，对指定端口的 HTTP GET 请求将收到文本“hi”。
实际发生的情况是，每个 GET 请求都会导致生成一个新的 netcat 进程，一次一个（这是我所期望的），但在我杀死最新的 netcat 进程之前，所有 GET 请求都不会完成。此时，我发出的第一个 GET 请求以“hi”响应，所有其他请求都以失败的连接尝试返回。我怀疑这与我对 STDOUT 重定向细节的困惑有关，但我的理解是，每次有人连接时，该脚本都会生成一个新版本，并且该脚本的 STDIN/OUT 应该是网络套接字因为 -e netcat 标志。
I was trying to use the following script to create a simple CGI server (from http://www.jfranken.de/homepages/johannes/vortraege/netcat_inhalt.en.html). This more about the technical reason it's not working than whether this is a good idea:
#!/bin/bash
export port=${port:-$1} # inherit $1 as $port
/usr/local/bin/nc -l -p $port -e $0 & # await further connections on this port
echo "hi"
Does anyone know why this doesn't work properly? Upon first execution, it prints "hi" to the terminal, which I expect. I would then expect netcat to be waiting on the specified port, ready to spawn a new copy of this script with the STDOUT associated with the network socket (because of the -e flag). So HTTP GET requests to the specified port would then receive the text "hi".
What actually happens is that each GET request causes a new netcat process to be spawned, one at a time (which I expect), but none of the GET requests complete until I kill the latest netcat process. At that point, the first GET request I issued responds with "hi", and all the others come back as failed connection attempts. I suspect this has something to do with my confusion about the details of STDOUT redirection, but my understanding is that this script would just spawn a new version of itself every time someone connected, and the STDIN/OUT for the script should be the network socket because of the -e netcat flag.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
除非使用 -DGAPING_SECURITY_HOLE 编译，否则 netcat 的 -e 选项将不起作用，这应该会让您暂停一下以这种方式做事是否明智。这也是一个仅在 Hobbit 的 netcat 和 bsd-ish 衍生品中可用的选项，据我所知，在 GNU netcat 中不可用。
通过将数据传输到 netcat 进程中，您能否获得所需的结果？例如，
似乎具有大致相同的效果，但是数据可能是在建立连接之前生成的，而不是像使用 -e 标志那样在连接之后生成。如果您的数据不需要基于时间或交互式，那么这将起作用。
The -e option to netcat will not work unless it is compiled with -DGAPING_SECURITY_HOLE which should give you some pause as to the wisdom of doing things this way. It's also an option that is only available in Hobbit's netcat and bsd-ish derivatives, not afaik in GNU netcat.
Could you achieve the desired result by piping your data into the netcat process? For example
Seems to have largely the same effect, however the data is likely generated before the connection is made, rather than after as in using the -e flag. If your data doesn't need to be time-based or interactive this would work.