处理由其他库导入的库中的异常的最佳实践是什么？

Question

在Python中，如何处理由其他库导入的库所产生的异常，应该怎么做呢？

举个例子，我有一个叫“pycontrol”的库，我把它导入到我的主程序中。而“pycontrol”又导入了“suds”库。然后，“suds”库又导入了“urllib2”库。我发现，当“suds”库在通过“urllib2”连接远程资源时遇到问题，这些异常就会传递到我的主程序里。

我现在的想法是，把“urllib2”和“suds”都导入到我的全局命名空间中，然后捕获它们抛出的那些常见异常，特别是那些在“pycontrol”中没有处理的异常。

还有没有其他更好的做法呢？

下面是代码片段的基本样子（没有把suds或urllib2导入到全局命名空间中）：

    import pycontrol.pycontrol as pc

    print "Connecting to iControl API on LTM %s..." % ltm
    try:
        b = pc.BIGIP(hostname=ltm, username=user, password=pw, 
            wsdls=wsdl_list, fromurl=True,
            debug=soap_debug)
    except (<whattocatch>), detail:
        print "Error: could not connect to iControl API on LTM %s... aborting!" % ltm
        print "Details: %s" % detail
        exitcode = 1
    else:
        print "Connection successfully established."

这是一个示例的错误追踪信息：

Connecting to iControl API on LTM s0-bigip1-lb2.lab.zynga.com...
Traceback (most recent call last):
  File "./register.py", line 507, in <module>
    main()
  File "./register.py", line 415, in main
    b = build_bigip_object(ltm, user, pw, WSDLS, soap_debug = False)
  File "./register.py", line 85, in build_bigip_object
    debug=soap_debug)
  File "build/bdist.macosx-10.6-universal/egg/pycontrol/pycontrol.py", line 81, in __init__
  File "build/bdist.macosx-10.6-universal/egg/pycontrol/pycontrol.py", line 103, in _get_clients
  File "build/bdist.macosx-10.6-universal/egg/pycontrol/pycontrol.py", line 149, in _get_suds_client
  File "/Library/Python/2.6/site-packages/suds/client.py", line 111, in __init__
    self.wsdl = reader.open(url)
  File "/Library/Python/2.6/site-packages/suds/reader.py", line 136, in open
    d = self.fn(url, self.options)
  File "/Library/Python/2.6/site-packages/suds/wsdl.py", line 136, in __init__
    d = reader.open(url)
  File "/Library/Python/2.6/site-packages/suds/reader.py", line 73, in open
    d = self.download(url)
  File "/Library/Python/2.6/site-packages/suds/reader.py", line 88, in download
    fp = self.options.transport.open(Request(url))
  File "/Library/Python/2.6/site-packages/suds/transport/https.py", line 60, in open
    return  HttpTransport.open(self, request)
  File "/Library/Python/2.6/site-packages/suds/transport/http.py", line 62, in open
    return self.u2open(u2request)
  File "/Library/Python/2.6/site-packages/suds/transport/http.py", line 118, in u2open
    return url.open(u2request, timeout=tm)
  File "/System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/urllib2.py", line 383, in open
    response = self._open(req, data)
  File "/System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/urllib2.py", line 401, in _open
    '_open', req)
  File "/System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/urllib2.py", line 361, in _call_chain
    result = func(*args)
  File "/System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/urllib2.py", line 1138, in https_open
    return self.do_open(httplib.HTTPSConnection, req)
  File "/System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/urllib2.py", line 1105, in do_open
    raise URLError(err)
urllib2.URLError: <urlopen error [Errno 8] nodename nor servname provided, or not known>

Answer 1

为什么你需要捕捉特定的异常呢？毕竟，从 b = pc.BIGIP(...) 这行代码抛出的任何异常（不仅仅是 URLError）都意味着你无法继续执行下去了。

我建议：

import traceback

try:
    b = pc.BIGIP(...)
except:
    traceback.print_exc()
    exitcode = 1
else:
    do_something_with_connection(b)

还有一个想法：为什么要费心去捕捉异常呢？当Python解释器遇到没有处理的异常时，它会把错误信息打印到标准错误输出，并结束程序：

b = bc.BIGIP(...)
do_something_with_connection(b)

或者如果你需要写入错误日志的话：

import logging
import sys

def main():
   b = bc.BIGIP(...)
   do_something_with_connection(b)

if __name__ == "__main__":
    try:
        main()
    except:
        logging.exception("An unexpected error occured")
        sys.exit(1)

Answer 2

我觉得你自己已经回答了自己的问题。导入urllib2模块，然后在你的代码中捕获异常。

from urllib2 import URLError

try:
    # something
except URLError, e:
    # Do something in case of error.