1249935 - Ambiguous language in documentation for btoa()

Status: RESOLVED FIXED

(Developer Documentation Graveyard :: General, defect, P5)

Description

[chrisv]

:: Developer Documentation Request

      Request Type: Correction
     Gecko Version: unspecified
 Technical Contact: 

:: Details

The 'Unicode Strings' section of the btoa() documentation has a rather vague statement "In most browsers, calling window.btoa() on a Unicode string will cause a Character Out Of Range exception".

What is meant by a 'Unicode string' is not specified. In fact, ASCII text is Unicode, and indeed ASCII text is valid UTF-8.

Naturally, with a 7-bit ASCII/Unicode string such as 'to', btoa('to') works ('dG8=').

Also, with an 8-bit ASCII/Unicode string such as 'ça', btoa('ça') works ('52E=').

Perhaps the statement should be "a multibyte Unicode string" rather than simply "a Unicode string".

With a multi-byte Unicode string such as '†' (U+2020), btoa('†') gives an InvalidCharacterError.

Comment 1

[Eric Shepherd [:sheppy]]

I've largely rewritten the text here, based off a review of the specification. Hopefully this is less ambiguous now. :)