在执行 curl 多请求时,我只能得到空响应。没有抛出异常,但响应值没有内容(在下面的代码段中注释)
这是我的代码的简化版本:
from StringIO import StringIO
import pycurl
class CurlStream(object):
curl_count = 0
curl_storage = []
def __init__(self):
self.curl_multi = pycurl.CurlMulti()
def add_request(self, request, post_fields=None):
self.curl_count += 1
curl = self._create_curl(request, post_fields)
self.curl_multi.add_handle(curl)
def perform(self):
while self.curl_count:
while True:
response, self.curl_count = self.curl_multi.perform()
if response != pycurl.E_CALL_MULTI_PERFORM:
break
self.curl_multi.select(1.0)
def read_all(self):
for response in self.curl_storage:
print response.getvalue() # this does nothing --prints blank lines
def close(self):
self.curl_multi.close()
def _create_curl(self, request, post_fields):
curl = pycurl.Curl()
output = StringIO()
self.curl_storage.append(output)
curl.setopt(curl.URL, request)
curl.setopt(curl.WRITEFUNCTION, output.write)
curl.setopt(curl.TIMEOUT, 20)
return curl
def main():
curl_stream = CurlStream()
curl_stream.add_request('http://www.google.com')
curl_stream.add_request('http://www.example.com')
curl_stream.perform()
curl_stream.read_all()
curl_stream.close()
if __name__ == '__main__':
main()
我在没有使用 curl multi 的情况下使用相同的选项发出了单个请求并且它有效。
请您参考如下方法:
好的,当我将 _create_curl 方法更改为此(添加 write_out 用于调试)时,我发现它有效:
def _create_curl(self, request, post_fields):
curl = pycurl.Curl()
curl.setopt(curl.URL, request)
curl.setopt(curl.WRITEFUNCTION, self.write_out)
curl.setopt(curl.TIMEOUT, 20)
# Below is the important bit, I am now adding each curl object to a list
self.curl_storage.append(curl)
return curl
def write_out(self, data):
print data
return len(data)
问题是在将 curl 对象添加到 multicurl 对象时,我没有保留对单个 curl 对象的任何引用,因此它自动关闭了。
根据pycurl docs对于 curl close() 方法:
Corresponds to curl_easy_cleanup in libcurl. This method is automatically called by pycurl when a Curl object no longer has any references to it, but can also be called explicitly.
