[mlir] Fix tests after Chris implemented string escaping in MLIR
We don't need to C-escape any more, so don't. Also, change the expected escaping syntax to be the slightly noisier version that LLVM emits.
PiperOrigin-RevId: 208989483
diff --git a/lib/Parser/Token.cpp b/lib/Parser/Token.cpp
index 1b14ae0..4c3f9e4 100644
--- a/lib/Parser/Token.cpp
+++ b/lib/Parser/Token.cpp
@@ -113,7 +113,7 @@
break;
}
- assert(i + 1 < e && "invalid string should be caught by lexer");
+ assert(i + 1 <= e && "invalid string should be caught by lexer");
auto c2 = bytes[i++];
assert(llvm::isHexDigit(c1) && llvm::isHexDigit(c2) && "invalid escape");