suds访问Webservice产生fileCache问题

    最近用Python中的suds访问webservice，由于之前发布的webservice有问题，最先采用soaplib来进行发布，后来发现其已经不更新不维护了，已经改为rpclib（官方地址：http://github.com/arskom/rpclib）进行后续开发，但在rpclib的readme中，介绍了rpclib已经更名为spyne，并将持续进行更新，所以最后就选用spyne进行开发了，基于web框架Flask。

     在不断变换发布方式的过程中，一直使用suds来进行访问，但是发布的URL不变。结果发现，在服务器上访问URL失败，而用其它URL则是成功。困惑中。。。。最后查之，是因为suds使用了filecache，所以访问的是第一次发布的URL，第一次发布的URL需要进行schema校验，而服务器不能访问外网导致的。

  主要代码如下：

def open(self, url):
    """
    Open an XML document at the specified I{url}.
    First, the document attempted to be retrieved from
    the I{object cache}.  If not found, it is downloaded and
    parsed using the SAX parser.  The result is added to the
    cache for the next open().
    @param url: A document url.
    @type url: str.
    @return: The specified XML document.
    @rtype: I{Document}
    """
    cache = self.cache()
    id = self.mangle(url, 'document')
    d = cache.get(id)
    if d is None:
        d = self.download(url)
        cache.put(id, d)
    self.plugins.document.parsed(url=url, document=d.root())
    return d

这是根据文件ID来获取文件内容的代码，首先从文件cache中获取，若获取不到则从网络获取。其中cache对象即为ObjectCache，继承自FileCache，而FileCache的初始化代码为：

def __init__(self, location=None, **duration):
    """
    @param location: The directory for the cached files.
    @type location: str
    @param duration: The cached file duration which defines how
        long the file will be cached.  A duration=0 means forever.
        The duration may be: (months|weeks|days|hours|minutes|seconds).
    @type duration: {unit:value}
    """
    if location is None:
        location = os.path.join(tmp(), 'suds')
    self.location = location
    self.duration = (None, 0)
    self.setduration(**duration)
    self.checkversion()

其中tmp()方法的定义为：

def gettempdir():
    """Accessor for tempfile.tempdir."""
    global tempdir
    if tempdir is None:
        _once_lock.acquire()
        try:
            if tempdir is None:
                tempdir = _get_default_tempdir()
        finally:
            _once_lock.release()
    return tempdir

def _get_default_tempdir():
    """Calculate the default directory to use for temporary files.
    This routine should be called exactly once.

    We determine whether or not a candidate temp dir is usable by
    trying to create and write to a file in that directory.  If this
    is successful, the test file is deleted.  To prevent denial of
    service, the name of the test file must be randomized."""

    namer = _RandomNameSequence()
    dirlist = _candidate_tempdir_list()
    flags = _text_openflags

    for dir in dirlist:
        if dir != _os.curdir:
            dir = _os.path.normcase(_os.path.abspath(dir))
        # Try only a few names per directory.
        for seq in xrange(100):
            name = namer.next()
            filename = _os.path.join(dir, name)
            try:
                fd = _os.open(filename, flags, 0o600)
                try:
                    try:
                        with _io.open(fd, 'wb', closefd=False) as fp:
                            fp.write(b'blat')
                    finally:
                        _os.close(fd)
                finally:
                    _os.unlink(filename)
                return dir
            except (OSError, IOError) as e:
                if e.args[0] != _errno.EEXIST:
                    break # no point trying more names in this directory
                pass
    raise IOError, (_errno.ENOENT,
                    ("No usable temporary directory found in %s" % dirlist))

至此，已经完全清楚，是找一个可写的目录，然后测试是否可读可写，并返回该目录。文件就写在响应的目录下。耗费了半个下午的时间，终于找出问题，Mark一下。