MichaelChirico
commented
4 years ago I agree that these two cases should be handled similarly. The reason for the difference is that they are currently handled by the string building code, and that's different for byte-sized chars versus wide chars, but the detection should probably happen earlier.
METADATA
- Comment author - Duncan Murdoch
- Timestamp - 2014-10-27 14:16:21 UTC
For a long time, embedding a nul character in a string using \0 has thrown an error.
"\0" ## Error: embedded nul in string: '\0'
However, it is still possible to enter a nul character using Unicode syntax.
"abc\u0000def" ## [1] "abc"
R's behaviour should be consistent between the two specifications of nul. That is, attempting to create strings containing "\u0000" should throw an error.
METADATA