Yes, that's the problem. You need to maintain a close attention level to know which things are POSIX. And in the case of GNU grep, you actually need to set POSIXLY_CORRECT=1. Otherwise its behavior is not a subset.
POSIX also forbids greps from searching UTF-16 because it mandates that certain characters always use a single byte. ripgrep, for example, doesn't have this constraint and thus can transparently search UTF-16 correctly via BOM sniffing.
POSIX also forbids greps from searching UTF-16 because it mandates that certain characters always use a single byte. ripgrep, for example, doesn't have this constraint and thus can transparently search UTF-16 correctly via BOM sniffing.