摘要：下面我們來(lái)看一下效果下次需要捕獲一個(gè)異常然后再拋出另一個(gè)異常的時(shí)候大家可以試試本文的方法。

一般實(shí)現(xiàn)捕獲異常然后再拋出另一個(gè)異常的方法類(lèi)似下面這樣:

def div():
    2 / 0

try:
    div()
except ZeroDivisionError as e:
    raise ValueError(e)

不知道大家有沒(méi)有注意到這樣拋出異常的方式有一個(gè)很?chē)?yán)重的問(wèn)題，那就是 在重新拋出另一個(gè)異常的時(shí)候，捕獲的上一個(gè)異常的 traceback 信息丟失了(python2): :

$ cat a.py
def div():
    2 / 0
try:
    div()
except ZeroDivisionError as e:
    raise ValueError(e)

$ python2 a.py
Traceback (most recent call last):
  File "a.py", line 6, in 
    raise ValueError(e)
ValueError: integer division or modulo by zero

這樣的話非常不利于查找問(wèn)題: 比如上面的例子中實(shí)際出錯(cuò)的代碼是第二行，但是 當(dāng)我們捕獲了第一個(gè)異常然后再拋出一個(gè)自定義異常的時(shí)候， 實(shí)際出錯(cuò)位置的信息就丟失了。

那么在 Python 2 下如果我們不想丟失捕獲的異常的 traceback 信息的話，應(yīng)該 怎樣重新拋出異常呢？

有兩種辦法, 還是用上面的例子舉例:

一種辦法是直接 raise: :

$ cat a.py
def div():
    2 / 0
try:
    div()
except ZeroDivisionError as e:
    raise

$ python2 a.py
Traceback (most recent call last):
  File "a.py", line 4, in 
    div()
  File "a.py", line 2, in div
    2 / 0
ZeroDivisionError: integer division or modulo by zero

另一種辦法就是 raise 另一個(gè)異常時(shí)指定上一個(gè)異常的 traceback 信息 (通過(guò) sys.exc_info() 獲取當(dāng)前捕獲的異常信息): :

$ cat a.py
import sys

def div():
    2 / 0
try:
    div()
except ZeroDivisionError as e:
    raise ValueError(e), None, sys.exc_info()[2]

$ python2 a.py
Traceback (most recent call last):
  File "a.py", line 6, in 
    div()
  File "a.py", line 4, in div
    2 / 0
ValueError: integer division or modulo by zero

這個(gè)是 raise 的高級(jí)用法:

raise exception, value, traceback

exception: 異常類(lèi)實(shí)例/異常類(lèi)

value: 初始化異常類(lèi)的參數(shù)值/異常類(lèi)實(shí)例（使用這個(gè)實(shí)例作為 raise 的異常實(shí)例）/元組/None

traceback: traceback 對(duì)象/None

下面我們來(lái)看看上面的方法是否可以應(yīng)對(duì)多層異常捕獲然后再拋出的情況: :

$ cat a.py
import sys

def div():
    2 / 0

def foo():
    try:
        div()
    except ZeroDivisionError as e:
        raise ValueError(e), None, sys.exc_info()[2]

def bar():
    try:
        foo()
    except ValueError as e:
        raise TypeError(e), None, sys.exc_info()[2]

def foobar():
    try:
        bar()
    except TypeError as e:
        raise
foobar()

$ python2 a.py
Traceback (most recent call last):
  File "a.py", line 23, in 
    foobar()
  File "a.py", line 20, in foobar
    bar()
  File "a.py", line 14, in bar
    foo()
  File "a.py", line 8, in foo
    div()
  File "a.py", line 4, in div
    2 / 0
TypeError: integer division or modulo by zero

從上面的結(jié)果可以看到這兩種方法是支持多層異常 traceback 信息傳遞的。

那么在 Python 3 下又怎么解決這個(gè)問(wèn)題呢？

在 Python 3 下默認(rèn)會(huì)附加上捕獲的上個(gè)異常的 trackback 信息（保存在異常實(shí)例的 __traceback__ 屬性中）: :

$ cat a.py
def div():
    2 / 0
try:
    div()
except ZeroDivisionError as e:
    raise ValueError(e)

$ python3 a.py
Traceback (most recent call last):
  File "a.py", line 4, in 
    div()
  File "a.py", line 2, in div
    2 / 0
ZeroDivisionError: division by zero

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "a.py", line 6, in 
    raise ValueError(e)
ValueError: division by zero

也支持指定使用哪個(gè)異常實(shí)例的 traceback 信息: raise ... from ... :

$ cat a.py
def div():
    2 / 0

try:
    div()
except ZeroDivisionError as e:
    raise ValueError(e) from e

$ python a.py
Traceback (most recent call last):
  File "a.py", line 5, in 
    div()
  File "a.py", line 2, in div
    2 / 0
ZeroDivisionError: division by zero

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "a.py", line 7, in 
    raise ValueError(e) from e
ValueError: division by zero

也可以指定使用的 traceback 對(duì)象: raise exception.with_traceback(traceback) :

$ cat a.py
import sys

def div():
    2 / 0

try:
    div()
except ZeroDivisionError as e:
    raise ValueError(e).with_traceback(sys.exc_info()[2])

$ python a.py
Traceback (most recent call last):
  File "a.py", line 7, in 
    div()
  File "a.py", line 4, in div
    2 / 0
ZeroDivisionError: division by zero

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "a.py", line 9, in 
    raise ValueError(e).with_traceback(sys.exc_info()[2])
  File "a.py", line 7, in 
    div()
  File "a.py", line 4, in div
    2 / 0
ValueError: division by zero

上面介紹了在 Python 2 和 Python 3 下的不同解決辦法，那么如何寫(xiě)一個(gè)兼容 Python 2 和 Python 3 的 reraise 函數(shù)呢？

下面將介紹一種方法:

PY3 = sys.version_info[0] == 3
if PY3:
    def reraise(tp, value, tb=None):
        if value.__traceback__ is not tb:
            raise value.with_traceback(tb)
        else:
            raise value
else:
    exec("""def reraise(tp, value, tb=None):
           raise tp, value, tb
    """)

這里的 reraise 函數(shù)我們約定了 vlaue 參數(shù)的值是一個(gè)異常類(lèi)的實(shí)例。 上面 else 中之所以用 exec 去定義 reraise 函數(shù)是因?yàn)?raise tp, value, tb 在 Python 3 下會(huì)報(bào)語(yǔ)法錯(cuò)誤，所以用 exec 來(lái) 繞過(guò) Python 3 下的語(yǔ)法錯(cuò)誤檢查。

下面我們來(lái)看一下效果: :

$ cat a.py

ef div():
    2 / 0

def foo():
    try:
        div()
    except ZeroDivisionError as e:
        reraise(ValueError, ValueError(e), sys.exc_info()[2])

def bar():
    try:
        foo()
    except ValueError as e:
        reraise(TypeError, TypeError(e), sys.exc_info()[2])

def foobar():
    try:
        bar()
    except TypeError:
        raise
foobar()

Python 2: :

$ python2 a.py
Traceback (most recent call last):
  File "a.py", line 34, in 
    foobar()
  File "a.py", line 31, in foobar
    bar()
  File "a.py", line 27, in bar
    reraise(TypeError, TypeError(e), sys.exc_info()[2])
  File "a.py", line 25, in bar
    foo()
  File "a.py", line 21, in foo
    reraise(ValueError, ValueError(e), sys.exc_info()[2])
  File "a.py", line 19, in foo
    div()
  File "a.py", line 15, in div
    2 / 0
TypeError: integer division or modulo by zero

Python 3: :

$ python3 a.py
Traceback (most recent call last):
  File "a.py", line 19, in foo
    div()
  File "a.py", line 15, in div
    2 / 0
ZeroDivisionError: division by zero

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "a.py", line 25, in bar
    foo()
  File "a.py", line 21, in foo
    reraise(ValueError, ValueError(e), sys.exc_info()[2])
  File "a.py", line 6, in reraise
    raise value.with_traceback(tb)
  File "a.py", line 19, in foo
    div()
  File "a.py", line 15, in div
    2 / 0
ValueError: division by zero

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "a.py", line 34, in 
    foobar()
  File "a.py", line 31, in foobar
    bar()
  File "a.py", line 27, in bar
    reraise(TypeError, TypeError(e), sys.exc_info()[2])
  File "a.py", line 6, in reraise
    raise value.with_traceback(tb)
  File "a.py", line 25, in bar
    foo()
  File "a.py", line 21, in foo
    reraise(ValueError, ValueError(e), sys.exc_info()[2])
  File "a.py", line 6, in reraise
    raise value.with_traceback(tb)
  File "a.py", line 19, in foo
    div()
  File "a.py", line 15, in div
    2 / 0
TypeError: division by zero

下次需要捕獲一個(gè)異常然后再拋出另一個(gè)異常的時(shí)候大家可以試試本文的方法。

6. Simple statements — Python 2.7.12 documentation

6. Built-in Exceptions — Python 2.7.12 documentation

7. Simple statements — Python 3.5.2 documentation

5. Built-in Exceptions — Python 3.5.2 documentation

PEP 3109 -- Raising Exceptions in Python 3000 | Python.org

bottle/bottle.py at cafc15419cbb4a6cb748e6ecdccf92893bb25ce5 · bottlepy/bottle

flask/_compat.py at 6e46d0cd3969f6c13ff61c95c81a975192232fed · pallets/flask

原文地址: https://mozillazg.com/2016/08...