Recursively expand `TokenKind::Interpolated` (take 2)

Fixes #68430

This is a re-attempt of PR #72388, which was previously reverted due to
a large number of breakages. All of the known breakages should now be
patched upstream.

7 files changed, 152 insertions, 75 deletions

diff --git a/src/librustc_parse/lib.rs b/src/librustc_parse/lib.rs
index 7b7235eed36..bc857c97742 100644
--- a/src/librustc_parse/lib.rs
+++ b/src/librustc_parse/lib.rs
@@ -7,8 +7,8 @@
 #![feature(or_patterns)]
 
 use rustc_ast as ast;
-use rustc_ast::token::{self, DelimToken, Nonterminal, Token};
-use rustc_ast::tokenstream::{self, TokenStream, TokenTree};
+use rustc_ast::token::{self, DelimToken, Nonterminal, Token, TokenKind};
+use rustc_ast::tokenstream::{self, IsJoint, TokenStream, TokenTree};
 use rustc_ast_pretty::pprust;
 use rustc_data_structures::sync::Lrc;
 use rustc_errors::{Diagnostic, FatalError, Level, PResult};
@@ -309,7 +309,7 @@ pub fn nt_to_tokenstream(nt: &Nonterminal, sess: &ParseSess, span: Span) -> Toke
     // modifications, including adding/removing typically non-semantic
     // tokens such as extra braces and commas, don't happen.
     if let Some(tokens) = tokens {
-        if tokenstream_probably_equal_for_proc_macro(&tokens, &tokens_for_real) {
+        if tokenstream_probably_equal_for_proc_macro(&tokens, &tokens_for_real, sess) {
             return tokens;
         }
         info!(
@@ -327,7 +327,11 @@ pub fn nt_to_tokenstream(nt: &Nonterminal, sess: &ParseSess, span: Span) -> Toke
 //
 // This is otherwise the same as `eq_unspanned`, only recursing with a
 // different method.
-pub fn tokenstream_probably_equal_for_proc_macro(first: &TokenStream, other: &TokenStream) -> bool {
+pub fn tokenstream_probably_equal_for_proc_macro(
+    first: &TokenStream,
+    other: &TokenStream,
+    sess: &ParseSess,
+) -> bool {
     // When checking for `probably_eq`, we ignore certain tokens that aren't
     // preserved in the AST. Because they are not preserved, the pretty
     // printer arbitrarily adds or removes them when printing as token
@@ -408,9 +412,6 @@ pub fn tokenstream_probably_equal_for_proc_macro(first: &TokenStream, other: &To
                 }
             }
             token_trees = out.into_iter().map(TokenTree::Token).collect();
-            if token_trees.len() != 1 {
-                debug!("break_tokens: broke {:?} to {:?}", tree, token_trees);
-            }
         } else {
             token_trees = SmallVec::new();
             token_trees.push(tree);
@@ -418,10 +419,32 @@ pub fn tokenstream_probably_equal_for_proc_macro(first: &TokenStream, other: &To
         token_trees.into_iter()
     }
 
-    let mut t1 = first.trees().filter(semantic_tree).flat_map(break_tokens);
-    let mut t2 = other.trees().filter(semantic_tree).flat_map(break_tokens);
+    let expand_nt = |tree: TokenTree| {
+        if let TokenTree::Token(Token { kind: TokenKind::Interpolated(nt), span }) = &tree {
+            // When checking tokenstreams for 'probable equality', we are comparing
+            // a captured (from parsing) `TokenStream` to a reparsed tokenstream.
+            // The reparsed Tokenstream will never have `None`-delimited groups,
+            // since they are only ever inserted as a result of macro expansion.
+            // Therefore, inserting a `None`-delimtied group here (when we
+            // convert a nested `Nonterminal` to a tokenstream) would cause
+            // a mismatch with the reparsed tokenstream.
+            //
+            // Note that we currently do not handle the case where the
+            // reparsed stream has a `Parenthesis`-delimited group
+            // inserted. This will cause a spurious mismatch:
+            // issue #75734 tracks resolving this.
+            nt_to_tokenstream(nt, sess, *span).into_trees()
+        } else {
+            TokenStream::new(vec![(tree, IsJoint::NonJoint)]).into_trees()
+        }
+    };
+
+    // Break tokens after we expand any nonterminals, so that we break tokens
+    // that are produced as a result of nonterminal expansion.
+    let mut t1 = first.trees().filter(semantic_tree).flat_map(expand_nt).flat_map(break_tokens);
+    let mut t2 = other.trees().filter(semantic_tree).flat_map(expand_nt).flat_map(break_tokens);
     for (t1, t2) in t1.by_ref().zip(t2.by_ref()) {
-        if !tokentree_probably_equal_for_proc_macro(&t1, &t2) {
+        if !tokentree_probably_equal_for_proc_macro(&t1, &t2, sess) {
             return false;
         }
     }
@@ -433,13 +456,17 @@ pub fn tokenstream_probably_equal_for_proc_macro(first: &TokenStream, other: &To
 //
 // This is otherwise the same as `eq_unspanned`, only recursing with a
 // different method.
-fn tokentree_probably_equal_for_proc_macro(first: &TokenTree, other: &TokenTree) -> bool {
+pub fn tokentree_probably_equal_for_proc_macro(
+    first: &TokenTree,
+    other: &TokenTree,
+    sess: &ParseSess,
+) -> bool {
     match (first, other) {
         (TokenTree::Token(token), TokenTree::Token(token2)) => {
             token_probably_equal_for_proc_macro(token, token2)
         }
         (TokenTree::Delimited(_, delim, tts), TokenTree::Delimited(_, delim2, tts2)) => {
-            delim == delim2 && tokenstream_probably_equal_for_proc_macro(&tts, &tts2)
+            delim == delim2 && tokenstream_probably_equal_for_proc_macro(&tts, &tts2, sess)
         }
         _ => false,
     }
@@ -498,7 +525,7 @@ fn token_probably_equal_for_proc_macro(first: &Token, other: &Token) -> bool {
             b == d && (a == c || a == kw::DollarCrate || c == kw::DollarCrate)
         }
 
-        (&Interpolated(..), &Interpolated(..)) => false,
+        (&Interpolated(..), &Interpolated(..)) => panic!("Unexpanded Interpolated!"),
 
         _ => panic!("forgot to add a token?"),
     }
diff --git a/src/test/ui/proc-macro/input-interpolated.stdout b/src/test/ui/proc-macro/input-interpolated.stdout
index 9cf33ba4a9d..a9636cfef82 100644
--- a/src/test/ui/proc-macro/input-interpolated.stdout
+++ b/src/test/ui/proc-macro/input-interpolated.stdout
@@ -15,51 +15,63 @@ PRINT-ATTR INPUT (DISPLAY): const A : u8 = 0 ;
 PRINT-ATTR INPUT (DEBUG): TokenStream [
     Ident {
         ident: "const",
-        span: #0 bytes(0..0),
+        span: #3 bytes(416..421),
     },
-    Ident {
-        ident: "A",
-        span: #0 bytes(0..0),
+    Group {
+        delimiter: None,
+        stream: TokenStream [
+            Ident {
+                ident: "A",
+                span: #0 bytes(503..504),
+            },
+        ],
+        span: #3 bytes(422..424),
     },
     Punct {
         ch: ':',
         spacing: Alone,
-        span: #0 bytes(0..0),
+        span: #3 bytes(424..425),
     },
     Ident {
         ident: "u8",
-        span: #0 bytes(0..0),
+        span: #3 bytes(426..428),
     },
     Punct {
         ch: '=',
         spacing: Alone,
-        span: #0 bytes(0..0),
+        span: #3 bytes(429..430),
     },
     Literal {
         kind: Integer,
         symbol: "0",
         suffix: None,
-        span: #0 bytes(0..0),
+        span: #3 bytes(431..432),
     },
     Punct {
         ch: ';',
         spacing: Alone,
-        span: #0 bytes(0..0),
+        span: #3 bytes(432..433),
     },
 ]
 PRINT-DERIVE INPUT (DISPLAY): struct A { }
 PRINT-DERIVE INPUT (DEBUG): TokenStream [
     Ident {
         ident: "struct",
-        span: #0 bytes(0..0),
+        span: #3 bytes(468..474),
     },
-    Ident {
-        ident: "A",
-        span: #0 bytes(0..0),
+    Group {
+        delimiter: None,
+        stream: TokenStream [
+            Ident {
+                ident: "A",
+                span: #0 bytes(503..504),
+            },
+        ],
+        span: #3 bytes(475..477),
     },
     Group {
         delimiter: Brace,
         stream: TokenStream [],
-        span: #0 bytes(0..0),
+        span: #3 bytes(478..480),
     },
 ]
diff --git a/src/test/ui/proc-macro/macro-rules-derive.rs b/src/test/ui/proc-macro/macro-rules-derive.rs
index 5b4d577a1ac..e0c40bbc734 100644
--- a/src/test/ui/proc-macro/macro-rules-derive.rs
+++ b/src/test/ui/proc-macro/macro-rules-derive.rs
@@ -1,14 +1,13 @@
 // aux-build:first-second.rs
-// FIXME: The spans here are bad, see PR #73084
 
 extern crate first_second;
 use first_second::*;
 
 macro_rules! produce_it {
     ($name:ident) => {
-        #[first] //~ ERROR cannot find type
+        #[first]
         struct $name {
-            field: MissingType
+            field: MissingType //~ ERROR cannot find type
         }
     }
 }
diff --git a/src/test/ui/proc-macro/macro-rules-derive.stderr b/src/test/ui/proc-macro/macro-rules-derive.stderr
index 4b72d29fe8a..54a079e4e73 100644
--- a/src/test/ui/proc-macro/macro-rules-derive.stderr
+++ b/src/test/ui/proc-macro/macro-rules-derive.stderr
@@ -1,8 +1,13 @@
 error[E0412]: cannot find type `MissingType` in this scope
-  --> $DIR/macro-rules-derive.rs:9:9
+  --> $DIR/macro-rules-derive.rs:10:20
    |
-LL |         #[first]
-   |         ^^^^^^^^ not found in this scope
+LL |             field: MissingType
+   |                    ^^^^^^^^^^^ not found in this scope
+...
+LL | produce_it!(MyName);
+   | -------------------- in this macro invocation
+   |
+   = note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)
 
 error: aborting due to previous error
 
diff --git a/src/test/ui/proc-macro/nodelim-groups.stdout b/src/test/ui/proc-macro/nodelim-groups.stdout
index 2fcd41f6da0..cdf851b535a 100644
--- a/src/test/ui/proc-macro/nodelim-groups.stdout
+++ b/src/test/ui/proc-macro/nodelim-groups.stdout
@@ -71,7 +71,6 @@ PRINT-BANG INPUT (DEBUG): TokenStream [
     },
 ]
 PRINT-BANG INPUT (DISPLAY): "hi" "hello".len() + "world".len() (1 + 1)
-PRINT-BANG RE-COLLECTED (DISPLAY): "hi" "hello" . len() + "world" . len() (1 + 1)
 PRINT-BANG INPUT (DEBUG): TokenStream [
     Literal {
         kind: Str,
@@ -82,50 +81,62 @@ PRINT-BANG INPUT (DEBUG): TokenStream [
     Group {
         delimiter: None,
         stream: TokenStream [
-            Literal {
-                kind: Str,
-                symbol: "hello",
-                suffix: None,
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
-            },
-            Punct {
-                ch: '.',
-                spacing: Alone,
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
-            },
-            Ident {
-                ident: "len",
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
-            },
             Group {
-                delimiter: Parenthesis,
-                stream: TokenStream [],
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
+                delimiter: None,
+                stream: TokenStream [
+                    Literal {
+                        kind: Str,
+                        symbol: "hello",
+                        suffix: None,
+                        span: $DIR/nodelim-groups.rs:21:17: 21:24 (#0),
+                    },
+                    Punct {
+                        ch: '.',
+                        spacing: Alone,
+                        span: $DIR/nodelim-groups.rs:21:24: 21:25 (#0),
+                    },
+                    Ident {
+                        ident: "len",
+                        span: $DIR/nodelim-groups.rs:21:25: 21:28 (#0),
+                    },
+                    Group {
+                        delimiter: Parenthesis,
+                        stream: TokenStream [],
+                        span: $DIR/nodelim-groups.rs:21:28: 21:30 (#0),
+                    },
+                ],
+                span: $DIR/nodelim-groups.rs:15:49: 15:54 (#7),
             },
             Punct {
                 ch: '+',
                 spacing: Alone,
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
-            },
-            Literal {
-                kind: Str,
-                symbol: "world",
-                suffix: None,
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
-            },
-            Punct {
-                ch: '.',
-                spacing: Alone,
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
-            },
-            Ident {
-                ident: "len",
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
+                span: $DIR/nodelim-groups.rs:15:55: 15:56 (#7),
             },
             Group {
-                delimiter: Parenthesis,
-                stream: TokenStream [],
-                span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
+                delimiter: None,
+                stream: TokenStream [
+                    Literal {
+                        kind: Str,
+                        symbol: "world",
+                        suffix: None,
+                        span: $DIR/nodelim-groups.rs:21:33: 21:40 (#0),
+                    },
+                    Punct {
+                        ch: '.',
+                        spacing: Alone,
+                        span: $DIR/nodelim-groups.rs:21:40: 21:41 (#0),
+                    },
+                    Ident {
+                        ident: "len",
+                        span: $DIR/nodelim-groups.rs:21:41: 21:44 (#0),
+                    },
+                    Group {
+                        delimiter: Parenthesis,
+                        stream: TokenStream [],
+                        span: $DIR/nodelim-groups.rs:21:44: 21:46 (#0),
+                    },
+                ],
+                span: $DIR/nodelim-groups.rs:15:57: 15:62 (#7),
             },
         ],
         span: $DIR/nodelim-groups.rs:16:47: 16:51 (#8),
diff --git a/src/test/ui/proc-macro/weird-hygiene.rs b/src/test/ui/proc-macro/weird-hygiene.rs
index 3f48191b5b2..7ba3f98a7a9 100644
--- a/src/test/ui/proc-macro/weird-hygiene.rs
+++ b/src/test/ui/proc-macro/weird-hygiene.rs
@@ -1,6 +1,4 @@
 // aux-build:weird-hygiene.rs
-// check-pass
-// FIXME: This should actually error, see PR #73084
 
 #![feature(stmt_expr_attributes)]
 #![feature(proc_macro_hygiene)]
@@ -22,7 +20,7 @@ macro_rules! other {
 
         #[derive(WeirdDerive)]
         enum MyEnum {
-            Value = (stringify!($tokens + hidden_ident), 1).1
+            Value = (stringify!($tokens + hidden_ident), 1).1 //~ ERROR cannot find
         }
 
         inner!();
@@ -33,7 +31,7 @@ macro_rules! invoke_it {
     ($token:expr) => {
         #[recollect_attr] {
             $token;
-            hidden_ident
+            hidden_ident //~ ERROR cannot find
         }
     }
 }
diff --git a/src/test/ui/proc-macro/weird-hygiene.stderr b/src/test/ui/proc-macro/weird-hygiene.stderr
new file mode 100644
index 00000000000..b17dc28f840
--- /dev/null
+++ b/src/test/ui/proc-macro/weird-hygiene.stderr
@@ -0,0 +1,25 @@
+error[E0425]: cannot find value `hidden_ident` in this scope
+  --> $DIR/weird-hygiene.rs:23:43
+   |
+LL |             Value = (stringify!($tokens + hidden_ident), 1).1
+   |                                           ^^^^^^^^^^^^ not found in this scope
+...
+LL |     other!(50);
+   |     ----------- in this macro invocation
+   |
+   = note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)
+
+error[E0425]: cannot find value `hidden_ident` in this scope
+  --> $DIR/weird-hygiene.rs:34:13
+   |
+LL |             hidden_ident
+   |             ^^^^^^^^^^^^ not found in this scope
+...
+LL |     invoke_it!(25);
+   |     --------------- in this macro invocation
+   |
+   = note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)
+
+error: aborting due to 2 previous errors
+
+For more information about this error, try `rustc --explain E0425`.