It should always produce well-formed strings
Instead, this is what happens on utf-16le and utf-16be:
✖ FAIL Common implementation mistakes > Unicode > Invalid input is replaced > utf-16le
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'ab\ud800wx'
- 'ab�wx'
^
✖ FAIL Common implementation mistakes > Unicode > Invalid input is replaced > utf-16be
AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:
+ actual - expected
+ 'ab\ud800wx'
- 'ab�wx'
^
Libraries/users might rightfully assume that TextDecoder always produces well-formed strings, and not doing so is problematic.
This might have security implications due to:
- how hashes and signatures behave on non-well-formed strings
- not following
fatal option
It should always produce well-formed strings
Instead, this is what happens on
utf-16leandutf-16be:Libraries/users might rightfully assume that TextDecoder always produces well-formed strings, and not doing so is problematic.
This might have security implications due to:
fataloption