为什么要这样枚举呢？

发布于 2024-12-07 08:40:01 字数 2044 浏览 1 评论 0原文

对于以下函数，

KEYS = {}    
def get(kind):
    "returns a new key of a particular kind"
    global KEYS

    try:
        return KEYS[kind].pop(0)
    except (KeyError, IndexError):
        handmade_key = Key.from_path(kind, 1)
        start, end = allocate_ids(handmade_key, 3)
        id_range = range(start, end+1)
        KEYS[kind] = [Key.from_path(kind, id) for id in id_range]
        for key in KEYS[kind]:
            print "within get() -> %s:%s"%(key, key.id())
        return get(kind)

我编写了以下单元测试

def testget2000(self):
    s = set()
    for i in range(0, 7):
        key = keyfactory.get("Model1")
        print "from get()   -> %s:%s"%(key, key.id())
        s.add(key)
    self.assertEqual(len(s), 7)
    self.assertEqual(len([k.id for k in s]), 2000)

并收到以下错误

失败：testget2000（keyfactory_test.ModelTest）

<小时>

回溯（最近一次调用最后一次）：
  文件“/home/vertegal/work/ei-sc/appengine/keyfactory_test.py”，

第 36 行，在 testget2000 中 self.assertEqual(len(s), 7) 断言错误：断言错误：5！= 7

-------------------->>开始捕获标准输出 <<

来自 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAgw:2
从 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAww:3
在 get() 内 -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAQw:1
在 get() 内 -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAgw:2
在 get() 内 -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAww:3
从 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAQw:1
从 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAgw:2
从 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAww:3
在 get() 内 -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBAw:4
在 get() 内 -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBQw:5
在 get() 内 -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBgw:6
从 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBAw:4
从 get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBQw:5

我真的不明白为什么第一次要在“within”之前打印“from”。另外，为什么它会两次分配相同的前几个 id 呢？我是否创建了一些奇怪的闭包？异常处理程序中的 KEYS 与外部的对象是否不同？我迷路了。

原文

For the following function

KEYS = {}    
def get(kind):
    "returns a new key of a particular kind"
    global KEYS

    try:
        return KEYS[kind].pop(0)
    except (KeyError, IndexError):
        handmade_key = Key.from_path(kind, 1)
        start, end = allocate_ids(handmade_key, 3)
        id_range = range(start, end+1)
        KEYS[kind] = [Key.from_path(kind, id) for id in id_range]
        for key in KEYS[kind]:
            print "within get() -> %s:%s"%(key, key.id())
        return get(kind)

I have written the following unit test

def testget2000(self):
    s = set()
    for i in range(0, 7):
        key = keyfactory.get("Model1")
        print "from get()   -> %s:%s"%(key, key.id())
        s.add(key)
    self.assertEqual(len(s), 7)
    self.assertEqual(len([k.id for k in s]), 2000)

And get the following error

FAIL: testget2000 (keyfactory_test.ModelTest)

Traceback (most recent call last):
  File "/home/vertegal/work/ei-sc/appengine/keyfactory_test.py",

line 36, in testget2000
self.assertEqual(len(s), 7)
AssertionError: AssertionError: 5 != 7

-------------------- >> begin captured stdout <<

from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAgw:2
from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAww:3
within get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAQw:1
within get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAgw:2
within get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAww:3
from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAQw:1
from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAgw:2
from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYAww:3
within get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBAw:4
within get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBQw:5
within get() -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBgw:6
from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBAw:4
from get()   -> agpkZXZ-cXVpenJ5cgwLEgZNb2RlbDEYBQw:5

I really don't understand why it is that "from" is being printed before "within" the first time. Also, why is it that it allocates the same first few ids twice? Am I creating some weird closure? Is KEYS a different object in the exception handler than outside it? I am lost.

分享到QQ

分享到微博

如果你对这篇内容有疑问，欢迎到本站社区发帖提问参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。

发布评论

需要登录才能够评论，你可以免费注册一个本站的账号。

风向决定发型 2024-12-14 08:40:01

因此，看起来当您捕获的输出开始时，数据存储区为空，但 KEYS[kind] 已填充两个值。数据存储不会两次分配相同的 ID，您只是拥有从未分配过的剩余 ID。要么追踪写入KEYS的其他内容，要么在测试开始时将其清除。

另外，您似乎正在传递实际的模型类。 Key.from_path 需要一个字符串，例如 'Model1' 而不是 Model1。

回复收藏 0 原文

无远思近则忧 2024-12-14 08:40:01

正如 Drew 所建议的，您可能会添加

global KEYS
KEYS = {}

到测试的顶部（或者更好的是添加到 setUp 方法），但问题的根源在于设计：使用延迟初始化、可变全局的函数状态很难测试。创建一个 KeyFactory 类，以 KEYS 作为实例变量，可能会更好。

您想通过这样做来实现什么目的？

As Drew suggests, you might add

global KEYS
KEYS = {}

to the top of your test (or better, to a setUp method), but the root of your problem is in the design: Functions that use use lazily initialized, mutable global state are difficult to test. Making a KeyFactory class, with KEYS as an instance variable, might serve you better.

What is it that you're trying to achieve by doing this?

回复收藏 0 原文