punycode: encoder contains bugs

@mathiasbynens

Not all test cases from http://tools.ietf.org/html/rfc3492#section-7.1 pass. Tests added in bnoordhuis/node@1f217ee.

FAIL: expected "ihqwcrb4cv8a8dqg056pqjye", got "ihqwcrb2cv8a8dqg056pqjye"
FAIL: expected "他们为什么不说中文", got "诵为斈不他们中什么"
FAIL: expected "ihqwctvzc91f659drss3x8bo0yb", got "ihqwctvzcv8e659drss3x8bo0yb"
FAIL: expected "他們爲什麽不說中文", got "ꁈ他谵什朒不倠中玼"
FAIL: expected "b1abfaaepdrnnbgefbaDotcwatmq2g4l", got "b1abfaaepdrnnbgefbadotcwatmq2g4l"
FAIL: expected "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b9g", got "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b3g"
FAIL: expected "TạisaohọkhôngthểchỉnóitiếngViệt", got "TạisaohkhôngtọhểchỉnóitiếngViệt"
FAIL: expected "3B-ww4c5e180e575a65lsy2b", got "3B-ww4c5e708d575a65lsy2b"
FAIL: expected "3年B組金八先生", got "廿3総鉜B八疪先"

cc @mathiasbynens

I have a working implementation that passes all unit tests except the Russian (Cyrillic) example string. I suspect that’s a typo in the RFC.

https://github.com/bestiejs/punycode.js

To run the unit tests in Node.js, clone the repository, cd into it, and then run node tests/tests.js.

$ node tests/tests.js 
 PASS - Punycode.utf16.decode
 PASS - Punycode.utf16.encode
 PASS - Punycode.decode
 FAIL - Punycode.encode
    PASS | EQ | ASCII string that breaks the existing rules for host-name labels | 
    PASS | EQ | ok | 
    PASS | EQ | ok | 
    PASS | EQ | ok | 
    PASS | EQ | ok | 
    PASS | EQ | ok | 
    PASS | EQ | ok | 
    PASS | EQ | ok | 
    PASS | EQ | Vietnamese | 
    PASS | EQ | Spanish | 
    FAIL | EQ | Russian (Cyrillic) | Expected: b1abfaaepdrnnbgefbaDotcwatmq2g4l, Actual: b1abfaaepdrnnbgefbadotcwatmq2g4l
    PASS | EQ | Korean (Hangul syllables) | 
    PASS | EQ | Japanese (kanji and hiragana) | 
    PASS | EQ | Hindi (Devanagari) | 
    PASS | EQ | Hebrew | 
    PASS | EQ | Czech | 
    PASS | EQ | Chinese (traditional) | 
    PASS | EQ | Chinese (simplified) | 
    PASS | EQ | Arabic (Egyptian) | 
    PASS | EQ | long string with both ASCII and non-ASCII characters | 
    PASS | EQ | mix of ASCII and non-ASCII characters | 
    PASS | EQ | multiple non-ASCII characters | 
    PASS | EQ | a single non-ASCII character | 
    PASS | EQ | a single basic code point | 
 PASS - Punycode.toUnicode
 PASS - Punycode.toASCII
----------------------------------------
    PASS: 63  FAIL: 1  TOTAL: 64
    Finished in 20 milliseconds.
----------------------------------------

Fixed in 326b2cb. Thanks, Mathias.

Navigation Menu

Repository navigation

punycode: encoder contains bugs #2072

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

punycode: encoder contains bugs #2072

Description

Activity

mathiasbynens commented on Nov 11, 2011

bnoordhuis commented on Nov 11, 2011

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions