[issue37301] CGIHTTPServer doesn't handle long POST requests

[vsbogd]

vsbogd <vsbogd at gmail.com> added the comment:

Yeah, I have seen this comment as well. As I see this code for the first time I am not very familiar with it. 

It seems that this explanation is related to the case when non-Python CGI script is executed:

                os.dup2(self.rfile.fileno(), 0)
                os.dup2(self.wfile.fileno(), 1)
                os.execve(scriptfile, args, env)

https://github.com/python/cpython/blob/3a1d50e7e573efb577714146bed5c03b9c95f466/Lib/http/server.py#L1176-L1178

On one hand the comment seems reasonable: as fileno is duped one cannot just read from it ahead of time. On the other hand few lines earlier one can see:

                # throw away additional data [see bug #427345]
                while select.select([self.rfile], [], [], 0)[0]:
                    if not self.rfile.read(1):
                        break

https://github.com/python/cpython/blob/3a1d50e7e573efb577714146bed5c03b9c95f466/Lib/http/server.py#L1163-L1166

which also seems like reading ahead of time.

My opinion this case should be tested after fix and if it works then fix can be applied without hesitation.

----------

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue37301>
_______________________________________