Mojibake.Security (Mojibake v0.8.0)

Copy Markdown View Source

Confusable detection and mixed-script checks (UTS #39).

Strings can look alike while being different: "pаypal" with a Cyrillic "а" is not "paypal". confusable?/2 tells whether two strings look the same, and skeleton/1 returns the form they are compared in, which you can store to look up lookalikes later. resolved_scripts/1 and mixed_script?/1 detect strings that mix scripts.

These functions need the :security feature. Without it they return or raise Mojibake.Error with reason :feature_not_enabled.

Summary

Functions

Return true if first and second are visually confusable: their skeletons are equal.

Return true if string mixes scripts that no single writing system uses together, the UTS #39 definition of mixed-script. Raises Mojibake.Error if string is not well-formed UTF-8.

Return the resolved script set of string: the scripts that all its characters share.

Return the confusable skeleton of string (UTS #39 bidiSkeleton with left-to-right direction): its bidi visual order, decomposed, without default-ignorable characters, and with each character replaced by the prototype it is confusable with.

Functions

confusable?(first, second)

@spec confusable?(String.t(), String.t()) :: boolean()

Return true if first and second are visually confusable: their skeletons are equal.

Ill-formed UTF-8 compares as U+FFFD.

Examples

iex> Mojibake.Security.confusable?("paypal", "pаypal")
true

iex> Mojibake.Security.confusable?("rn", "m")
true

iex> Mojibake.Security.confusable?("paypal", "PayPal")
false

mixed_script?(string)

@spec mixed_script?(String.t()) :: boolean()

Return true if string mixes scripts that no single writing system uses together, the UTS #39 definition of mixed-script. Raises Mojibake.Error if string is not well-formed UTF-8.

Examples

iex> Mojibake.Security.mixed_script?("pаypal")
true

iex> Mojibake.Security.mixed_script?("paypal2")
false

resolved_scripts(string)

@spec resolved_scripts(String.t()) :: :all | [Mojibake.Script.t()]

Return the resolved script set of string: the scripts that all its characters share.

Characters of the Common and Inherited scripts, such as digits and punctuation, go with any script. Han, Hiragana, Katakana, Hangul, and Bopomofo also count as the writing systems that use them, :japanese, :korean, and :han_with_bopomofo, so Japanese text that mixes kanji and kana resolves to [:japanese], and Hangul text to [:hangul, :korean].

Returns :all for a string that has only Common and Inherited characters, or none, and [] for a mixed-script string. Raises Mojibake.Error if string is not well-formed UTF-8.

Examples

iex> Mojibake.Security.resolved_scripts("Ελλάδα")
[:greek]

iex> Mojibake.Security.resolved_scripts("日本語のテキスト")
[:japanese]

iex> Mojibake.Security.resolved_scripts("123!")
:all

iex> Mojibake.Security.resolved_scripts("pаypal")
[]

skeleton(string)

@spec skeleton(String.t()) :: {:ok, String.t()} | {:error, Mojibake.Error.t()}

Return the confusable skeleton of string (UTS #39 bidiSkeleton with left-to-right direction): its bidi visual order, decomposed, without default-ignorable characters, and with each character replaced by the prototype it is confusable with.

A skeleton is for comparison only; do not show it.

Examples

iex> Mojibake.Security.skeleton("hеllo")
{:ok, "hello"}

iex> Mojibake.Security.skeleton(<<0xFF>>)
{:error, %Mojibake.Error{reason: :malformed_input}}

skeleton!(string)

@spec skeleton!(String.t()) :: String.t()

Like skeleton/1, but raises Mojibake.Error.

Examples

iex> Mojibake.Security.skeleton!("ℌello")
"Hello"