python / python/cpython

ConfigParser read(write()) not idempotent

未关闭
#96,239 2 条评论 0 个 reaction 已指派 1 人 在 GitHub 查看

@ambv 已经在做这个了。

开始于 2022年8月25日。

stdlib type-bug
主要语言
Python
星标
77.2k
派生
36k
PR 合并指标
PR 指标待抓取

描述

Bug report

Due to the the way that multiline values are handled and the fact that initial whitespace is preserved in keys added with set(), it is possible for calling parser1.write(foo) followed by parser2.read_file(foo) to leave parser1 and parser2 in radically different configuration states. As illustrated below, if a key with leading spaces is added with set() after a key with fewer leading spaces, that whitespace will be understood as part of the key and written to the file in a way that will be read back as part of the value of the previous key.

I believe that this is a bug in set, which should strip whitespace from its key argument. No ConfigParser read method can produce a key with leading whitespace, so such keys should not be accepted programmatically either.

from io import StringIO
import configparser
p = configparser.ConfigParser()
p.add_section('foo')
p.set('foo', 'a', '5')
p.set('foo', '    b', '6')
w = StringIO()
p.write(w)
w.seek(0)
print(w.read())
p = configparser.ConfigParser()
w.seek(0)
p.read_string(w.read())
print(p.items('foo'))
print(p.get('foo', 'b'))

output:

[foo]
a = 5
    b = 6


[('a', '5\nb = 6')]
Traceback (most recent call last):
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/configparser.py", line 790, in get
    value = d[option]
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/collections/__init__.py", line 986, in __getitem__
    return self.__missing__(key)            # support subclasses that define __missing__
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/collections/__init__.py", line 978, in __missing__
    raise KeyError(key)
KeyError: 'b'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/jeakle/python/tipr/test.py", line 15, in <module>
    print(p.get('foo', 'b'))
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/configparser.py", line 793, in get
    raise NoOptionError(option, section)
configparser.NoOptionError: No option 'b' in section: 'foo'

For context, this bit me while trying to subclass ConfigParser to handle a slightly different format. Among other things, we don't allow multiline values and want leading whitespace to be irrelevant. While attempting to override read_string() with a method that does some of our own processing on each line and then calls set() on it, I initially just passed the part of the line to left of the separator to set() as the key, which caused both the problem described above, and the issue that calling set() again later on the "same" key can introduce duplicate keys if they have different amounts of leading whitespace:

In fact a similar example shows that the output of write() can crash when being read() back:

from io import StringIO
import configparser
p = configparser.ConfigParser()
p.add_section('foo')
p.set('foo', '    b', '6')
p.set('foo', 'b', '7')
w = StringIO()
p.write(w)
w.seek(0)
print(w.read())
p = configparser.ConfigParser()
w.seek(0)
p.read_string(w.read())
print(p.items('foo'))
print(p.get('foo', 'b'))

output:

[foo]
    b = 6
b = 7


Traceback (most recent call last):
  File "/Users/jeakle/python/tipr/test.py", line 13, in <module>
    p.read_string(w.read())
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/configparser.py", line 724, in read_string
    self.read_file(sfile, source)
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/configparser.py", line 719, in read_file
    self._read(f, source)
  File "/usr/local/Cellar/python@3.10/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/configparser.py", line 1097, in _read
    raise DuplicateOptionError(sectname, optname,
configparser.DuplicateOptionError: While reading from '<string>' [line  3]: option 'b' in section 'foo' already exists

Your environment

  • CPython versions tested on: 3.10.4, 3.9.5
  • Operating system and architecture: MacOS, but replicated in online REPLs as well
Linked PRs
  • gh-157729

贡献指南

打开贡献指南

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

评估

这个 Issue 还没有评估数据。

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。