关于 Pycurl 使用遇到的问题

import pycurl import uuid import hashlib import os def get_filename(url): if not url: return None return hashlib.md5(url.encode()).hexdigest() class Fetcher(object): def __init__(self, urls, path): self.urls = urls self.path = path self.m = pycurl.CurlMulti() def fetch(self): if not urls or len(urls) == 0: print('empty urls...') return for url in urls: fdir = './%s/%s' % (self.path, get_filename(url)) if os.path.exists(fdir): print('%s exits, skip it...' % url) continue f = open(fdir, 'wb') c = pycurl.Curl() c.setopt(pycurl.URL, url) c.setopt(pycurl.WRITEDATA, f) self.m.add_handle(c) while True: ret, num_handles = self.m.perform() if ret != pycurl.E_CALL_MULTI_PERFORM: break while num_handles: ret = self.m.select(3.0) if ret == -1: continue while 1: ret, num_handles = self.m.perform() if ret != pycurl.E_CALL_MULTI_PERFORM: break print('downloading complete...') urls = ['xa.nuomi.com/1000338', 'xa.nuomi.com/1000002', 'xa.nuomi.com/884'] fetcher = Fetcher(urls, 'download') fetcher.fetch()

PyCURL

urls

URL

ret

2 条回复 2017-05-11 21:50:47 +08:00

blackeeper

2017-05-11 18:21:26 +08:00

加 c.close() ?

pank

2017-05-11 21:50:47 +08:00

@blackeeper 谢谢回复，已经找到问题了: IMPORTANT NOTE: add_handle does not implicitly add a Python reference to the Curl object (and thus does not increase the reference count on the Curl object). 应该是引用被冲掉了。修改一下变量名不重复就好了：

```
for idx, url in enumerate(urls):
f = open('./%s/%s' % (self.path, hashlib.md5(url.encode()).hexdigest()), 'wb')
locals()['c'+str(idx)] = pycur.Curl()
locals()['c'+str(idx)].setopt(pycurl.URL, url)
locals()['c'+str(idx)].setopt(pycurl.WRITEDATA,f)
self.m.add_handle(locals()['c'+str(idx)])
```