Add lexer option to keep shebangs

fmease · Aug 15, 2022 · 84611bb · 84611bb
1 parent 0a3c298
commit 84611bb
Show file tree

Hide file tree

Showing 4 changed files with 17 additions and 3 deletions.
diff --git a/compiler/driver/src/cli.rs b/compiler/driver/src/cli.rs
@@ -204,6 +204,7 @@ pub(crate) fn arguments() -> Result<(Command, GlobalOptions)> {
                 .arg(
                     Arg::new(argument::NAME)
                         .required(true)
+                        .value_parser(WordParser)
                         .help("The name of the package"),
                 )
                 .args(package_creation_arguments),

diff --git a/compiler/lexer/src/lib.rs b/compiler/lexer/src/lib.rs
@@ -25,6 +25,7 @@ pub struct Outcome {
 
 #[derive(Default)]
 pub struct Options {
+    pub keep_shebang: bool,
     pub keep_comments: bool,
 }
 
@@ -35,6 +36,7 @@ fn lex_string(source: String) -> Outcome {
         &map[file],
         &Options {
             keep_comments: true,
+            keep_shebang: true,
         },
     )
     .lex()
@@ -45,8 +47,9 @@ pub trait WordExt: Sized {
 }
 
 impl WordExt for Word {
-    // @Beacon @Bug this allows ids like `   foo-bar `
-    // (and `hey;;;wkwkwkwwkw`)! just write a custom lexer for this!
+    // @Bug this allows words like `   foo-bar ` (leading & trailing ws)
+    // @Note and not long before it used to allow `hey;;;wkwkwkwwkw`!
+    // @Task just write a custom lexer for this!
     fn parse(name: String) -> Result<Self, ()> {
         let Outcome { tokens, errors } = lex_string(name);
 
@@ -58,7 +61,7 @@ impl WordExt for Word {
 
         obtain!(
             (tokens.next().ok_or(())?, tokens.next().ok_or(())?),
-            (BareToken::Word(atom), BareToken::EndOfInput) => atom
+            (BareToken::Word(word), BareToken::EndOfInput) => word
         )
         .map(Self::new_unchecked)
         .ok_or(())
@@ -216,12 +219,19 @@ impl<'a> Lexer<'a> {
 
         if let Some('!') = self.peek() {
             while let Some(character) = self.peek() {
+                if self.options.keep_shebang {
+                    self.take();
+                }
                 self.advance();
 
                 if character == '\n' {
                     break;
                 }
             }
+
+            if self.options.keep_shebang {
+                self.add(Shebang);
+            }
         } else {
             self.lex_punctuation();
         }

diff --git a/compiler/lexer/src/test.rs b/compiler/lexer/src/test.rs
@@ -87,6 +87,7 @@ fn shebang() {
 #!/usr/bin/lushui run
 it",
         vec![
+            Token::new(span(1, 23), Shebang),
             Token::new(span(23, 25), Word("it".into())),
             Token::new(span(25, 25), EndOfInput),
         ],

diff --git a/compiler/token/src/lib.rs b/compiler/token/src/lib.rs
@@ -59,6 +59,7 @@ impl TokenExt for Token {
 #[derive(Clone, PartialEq, Eq, Discriminant, Debug)]
 #[discriminant(name: TokenName)]
 pub enum BareToken {
+    Shebang,
     Comment,
     DocumentationComment,
     Word(Atom),        // @Task use crate::Word
@@ -173,6 +174,7 @@ impl fmt::Display for TokenName {
         }
 
         f.write_str(match self {
+            Shebang => "shebang",
             Comment => "comment",
             DocumentationComment => "documentation comment",
             Word => "word",