Move lexer unit tests to rustc_lexer

StringReader is an intornal abstraction which at the moment changes a
lot, so these unit tests cause quite a bit of friction.

Moving them to rustc_lexer and more ingerated-testing style should
make them much less annoying, hopefully without decreasing their
usefulness much.

Note that coloncolon tests are removed (it's unclear what those are
testing).

\r\n tests are removed as well, as we normalize line endings even
before lexing.

Cargo.lock

@@ -342,7 +342,7 @@ dependencies = [
 name = "cargo-miri"
 version = "0.1.0"
 dependencies = [
- "cargo_metadata 0.11.1",
+ "cargo_metadata 0.9.1",
  "directories",
  "rustc-workspace-hack",
  "rustc_version",
@@ -391,6 +391,18 @@ dependencies = [
  "serde_json",
 ]
 
+[[package]]
+name = "cargo_metadata"
+version = "0.9.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "46e3374c604fb39d1a2f35ed5e4a4e30e60d01fab49446e08f1b3e9a90aef202"
+dependencies = [
+ "semver 0.9.0",
+ "serde",
+ "serde_derive",
+ "serde_json",
+]
+
 [[package]]
 name = "cargo_metadata"
 version = "0.11.1"
@@ -1979,6 +1991,7 @@ dependencies = [
 name = "miri"
 version = "0.1.0"
 dependencies = [
+ "byteorder",
  "colored",
  "compiletest_rs",
  "env_logger 0.7.1",

compiler/rustc_expand/src/lib.rs

@@ -39,11 +39,6 @@ mod tests;
 mod parse {
     #[cfg(test)]
     mod tests;
-    #[cfg(test)]
-    mod lexer {
-        #[cfg(test)]
-        mod tests;
-    }
 }
 #[cfg(test)]
 mod tokenstream {