punycode codec raises IndexError in decode_generalized_number()

[VikramHegde]

BPO	30566
Nosy	@bitdancer, @berkerpeksag, @vikhegde, @miss-islington
PRs	bpo-30566: Fix IndexError in the punycode codec #1986 bpo-30566: Fix IndexError when using punycode codec #18632 [3.8] bpo-30566: Fix IndexError when using punycode codec (GH-18632) #18651 [3.7] bpo-30566: Fix IndexError when using punycode codec (GH-18632) #18652

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2020-02-25.04:10:48.128>
created_at = <Date 2017-06-04.15:49:54.222>
labels = ['3.7', '3.8', 'type-bug', 'library', '3.9']
title = 'punycode codec raises IndexError in decode_generalized_number()'
updated_at = <Date 2020-02-25.04:10:48.125>
user = 'https://bugs.python.org/VikramHegde'

bugs.python.org fields:

activity = <Date 2020-02-25.04:10:48.125>
actor = 'berker.peksag'
assignee = 'none'
closed = True
closed_date = <Date 2020-02-25.04:10:48.128>
closer = 'berker.peksag'
components = ['Library (Lib)']
creation = <Date 2017-06-04.15:49:54.222>
creator = 'Vikram Hegde'
dependencies = []
files = []
hgrepos = []
issue_num = 30566
keywords = ['patch']
message_count = 9.0
messages = ['295127', '295149', '295270', '295278', '302070', '362613', '362617', '362618', '362623']
nosy_count = 5.0
nosy_names = ['r.david.murray', 'berker.peksag', 'Vikram Hegde', 'vikhegde', 'miss-islington']
pr_nums = ['1986', '18632', '18651', '18652']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue30566'
versions = ['Python 3.7', 'Python 3.8', 'Python 3.9']

[VikramHegde]

Here is the relevant code snippet from decode_generalized_number() in punycode.py

try:
        char = ord(extended[extpos])
    except IndexError:
        if errors == "strict":
            raise UnicodeError("incomplete punicode string")
        return extpos + 1, None
    extpos += 1
    if 0x41 <= char <= 0x5A: # A-Z
        digit = char - 0x41
    elif 0x30 <= char <= 0x39:
        digit = char - 22 # 0x30-26
    elif errors == "strict":
        raise UnicodeError("Invalid extended code point '%s'"
                           % extended[extpos])

While raising the UnicodeError() in the last line above, it accesses extended[extpos]. However extpos was incremented by 1 a few lines above that. This causes two errors:

1. The UnicodeError() prints the wrong character (the one after the character we want)
2. If the previous extpos was the last character in the string, then attempting to print character at extpos+1 will raise an IndexError.

[bitdancer]

Can you provide a reproducer, please?

[VikramHegde]

I have a patch for this problem but my contributor agreement has not been accepted yet, so I can't do a pull request.

Use the python package tldextract to trigger the bug. Here is a sample

Python 3.6.0 |Anaconda 4.3.1 (64-bit)| (default, Dec 23 2016, 12:22:00) 
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tldextract
>>> tldextract.extract("xn--w&")
Traceback (most recent call last):
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/encodings/punycode.py", line 207, in decode
    res = punycode_decode(input, errors)
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/encodings/punycode.py", line 194, in punycode_decode
    return insertion_sort(base, extended, errors)
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/encodings/punycode.py", line 165, in insertion_sort
    bias, errors)
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/encodings/punycode.py", line 146, in decode_generalized_number
    % extended[extpos])
IndexError: string index out of range

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

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/site-packages/tldextract/tldextract.py", line 358, in extract
    return TLD_EXTRACTOR(url)
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/site-packages/tldextract/tldextract.py", line 237, in __call__
    translations = [decode_punycode(label).lower() for label in labels]
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/site-packages/tldextract/tldextract.py", line 237, in <listcomp>
    translations = [decode_punycode(label).lower() for label in labels]
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/site-packages/tldextract/tldextract.py", line 232, in decode_punycode
    return idna.decode(label.encode('ascii'))
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/site-packages/idna/core.py", line 384, in decode
    result.append(ulabel(label))
  File "/home/vikram-work/anaconda3/envs/pefeatextract-debug/lib/python3.6/site-packages/idna/core.py", line 302, in ulabel
    label = label.decode('punycode')
IndexError: decoding with 'punycode' codec failed (IndexError: string index out of range)
>>>

[bitdancer]

You don't need an eternal package, just decoding b'xn--w&' with punycode will produce the traceback.

[vikhegde]

Could someone please review my PR. It has been in the pending state for over three months.

[berkerpeksag]

New changeset ba22e8f by Berker Peksag in branch 'master':
bpo-30566: Fix IndexError when using punycode codec (GH-18632)
ba22e8f

[berkerpeksag]

New changeset daef21c by Miss Islington (bot) in branch '3.8':
bpo-30566: Fix IndexError when using punycode codec (GH-18632)
daef21c

[berkerpeksag]

New changeset 55be9a6 by Miss Islington (bot) in branch '3.7':
bpo-30566: Fix IndexError when using punycode codec (GH-18632)
55be9a6

[berkerpeksag]

Thanks for the report and for the initial patch!