Clarify space_between

compiler/rustc_ast_pretty/src/pprust/state.rs

@@ -146,37 +146,49 @@ pub fn print_crate<'a>(
     s.s.eof()
 }
 
-/// This makes printed token streams look slightly nicer,
-/// and also addresses some specific regressions described in #63896 and #73345.
-fn space_between(prev: &TokenTree, curr: &TokenTree) -> bool {
-    if let TokenTree::Token(token, _) = prev {
-        // No space after these tokens, e.g. `x.y`, `$e`
-        // (The carets point to `prev`.)       ^     ^
-        if matches!(token.kind, token::Dot | token::Dollar) {
-            return false;
-        }
-        if let token::DocComment(comment_kind, ..) = token.kind {
-            return comment_kind != CommentKind::Line;
-        }
-    }
-    match curr {
-        // No space before these tokens, e.g. `foo,`, `println!`, `x.y`
-        // (The carets point to `curr`.)          ^           ^     ^
+/// Should two consecutive tokens be printed with a space between them?
+///
+/// Note: some old proc macros parse pretty-printed output, so changes here can
+/// break old code. For example:
+/// - #63896: `#[allow(unused,` must be printed rather than `#[allow(unused ,`
+/// - #73345: `#[allow(unused)] must be printed rather than `# [allow(unused)]
+///
+fn space_between(tt1: &TokenTree, tt2: &TokenTree) -> bool {
+    use token::*;
+    use Delimiter::*;
+    use TokenTree::Delimited as Del;
+    use TokenTree::Token as Tok;
+
+    // Each match arm has one or more examples in comments. The default is to
+    // insert space between adjacent tokens, except for the cases listed in
+    // this match.
+    match (tt1, tt2) {
+        // No space after line doc comments.
+        (Tok(Token { kind: DocComment(CommentKind::Line, ..), .. }, _), _) => false,
+
+        // `.` + ANYTHING: `x.y`, `tup.0`
+        // `$` + ANYTHING: `$e`
+        (Tok(Token { kind: Dot | Dollar, .. }, _), _) => false,
+
+        // ANYTHING + `,`: `foo,`
+        // ANYTHING + `.`: `x.y`, `tup.0`
+        // ANYTHING + `!`: `foo! { ... }`
         //
-        // FIXME: having `Not` here works well for macro invocations like
-        // `println!()`, but is bad when `!` means "logical not" or "the never
-        // type", where the lack of space causes ugliness like this:
-        // `Fn() ->!`, `x =! y`, `if! x { f(); }`.
-        TokenTree::Token(token, _) => !matches!(token.kind, token::Comma | token::Not | token::Dot),
-        // No space before parentheses if preceded by these tokens, e.g. `foo(...)`
-        TokenTree::Delimited(_, Delimiter::Parenthesis, _) => {
-            !matches!(prev, TokenTree::Token(Token { kind: token::Ident(..), .. }, _))
-        }
-        // No space before brackets if preceded by these tokens, e.g. `#[...]`
-        TokenTree::Delimited(_, Delimiter::Bracket, _) => {
-            !matches!(prev, TokenTree::Token(Token { kind: token::Pound, .. }, _))
-        }
-        TokenTree::Delimited(..) => true,
+        // FIXME: Incorrect cases:
+        // - Logical not: `x =! y`, `if! x { f(); }`
+        // - Never type: `Fn() ->!`
+        (_, Tok(Token { kind: Comma | Dot | Not, .. }, _)) => false,
+
+        // IDENT + `(`: `f(3)`
+        //
+        // FIXME: Incorrect cases:
+        // - Let: `let(a, b) = (1, 2)`
+        (Tok(Token { kind: Ident(..), .. }, _), Del(_, Parenthesis, _)) => false,
+
+        // `#` + `[`: `#[attr]`
+        (Tok(Token { kind: Pound, .. }, _), Del(_, Bracket, _)) => false,
+
+        _ => true,
     }
 }
 

tests/ui/macros/stringify.rs

@@ -115,6 +115,7 @@ fn test_expr() {
         "a + b * c - d + -1 * -2 - -3",
         "a + b * c - d + - 1 * - 2 - - 3"
     );
+    c2!(expr, [ x = !y ], "x = !y", "x =! y"); // FIXME
 
     // ExprKind::Unary
     c2!(expr, [ *expr ], "*expr", "* expr");
@@ -137,6 +138,7 @@ fn test_expr() {
 
     // ExprKind::If
     c1!(expr, [ if true {} ], "if true {}");
+    c2!(expr, [ if !true {} ], "if !true {}", "if! true {}"); // FIXME
     c2!(expr,
         [ if ::std::blah() { } else { } ],
         "if ::std::blah() {} else {}",
@@ -799,3 +801,32 @@ fn test_vis() {
     assert_eq!(inherited_vis!(struct), "");
     assert_eq!(stringify!(), "");
 }
+
+macro_rules! p {
+    ([$($tt:tt)*], $s:literal) => {
+        assert_eq!(stringify!($($tt)*), $s);
+    };
+}
+
+#[test]
+fn test_punct() {
+    // For all these cases, we must preserve spaces between the tokens.
+    // Otherwise, any old proc macro that parses pretty-printed code might glue
+    // together tokens that shouldn't be glued.
+    p!([ = = < < <= <= == == != != >= >= > > ], "= = < < <= <= == == != != >= >= > >");
+    p!([ && && & & || || | | ! ! ], "&& && & & || || | |!!"); // FIXME
+    p!([ ~ ~ @ @ # # ], "~ ~ @ @ # #");
+    p!([ . . .. .. ... ... ..= ..=], ".... .. ... ... ..= ..="); // FIXME
+    p!([ , , ; ; : : :: :: ], ",, ; ; : : :: ::"); // FIXME
+    p!([ -> -> <- <- => =>], "-> -> <- <- => =>");
+    p!([ $ $ ? ? ' ' ], "$$? ? ' '"); // FIXME
+    p!([ + + += += - - -= -= * * *= *= / / /= /= ], "+ + += += - - -= -= * * *= *= / / /= /=");
+    p!([ % % %= %= ^ ^ ^= ^= << << <<= <<= >> >> >>= >>= ],
+        "% % %= %= ^ ^ ^= ^= << << <<= <<= >> >> >>= >>=");
+
+    // For these one we must insert spaces between adjacent tokens, again due
+    // to proc macros.
+    p!([ +! ?= |> >>@ --> <-- $$ =====> ], "+! ? = | > >> @ - -> <- - $$== == =>"); // FIXME
+    p!([ ,; ;, ** @@ $+$ >< <> ?? +== ], ", ; ;, * * @ @ $+ $> < < > ? ? += ="); // FIXME
+    p!([ :#!@|$=&*,+;*~? ], ": #! @ | $= & *, + ; * ~ ?"); // FIXME
+}