执行curl多个请求时,我无法获得任何空响应。没有抛出异常,但响应值没有内容(在下面的代码段中注释)
以下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()
我使用相同的选项发出单个请求而不使用curl multi,它可以正常工作。
答案 0 :(得分:1)
好的,所以当我将_create_curl
方法更改为此(添加write_out
进行调试)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)
问题是当将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,因此它是自动关闭的。
根据curl close()方法的pycurl docs:
对应于libcurl中的curl_easy_cleanup。这个方法是 当Curl对象不再具有任何内容时,由pycurl自动调用 引用它,但也可以明确调用。