Smaller optimizations for custom parser since before incorrect force push

CHANGELOG.md

@@ -2,6 +2,9 @@
 
 ## [Unreleased]
 
+- more strict in a few cases where code that was incorrect code (according to the elm compiler) was successfully parsed: tuples with >3 parts, record access on most things
+- better performance, parsing should now be a lot faster
+
 ## [7.3.4] - 2024-07-26
 
 Parsing is faster by ~90%. A big thank you to [@lue-bird](https://github.com/lue-bird) for finding and introducing a huge amount of performance improvements.

benchmark/src/Main.elm

@@ -10905,4 +10905,4 @@ maybeToList maybe =
         Just content ->
             [ content ]
 
-"""
+"""

elm.json

@@ -34,11 +34,10 @@
         "elm/core": "1.0.0 <= v < 2.0.0",
         "elm/json": "1.0.0 <= v < 2.0.0",
         "elm/parser": "1.0.0 <= v < 2.0.0",
-        "miniBill/elm-unicode": "1.0.2 <= v < 2.0.0",
         "rtfeldman/elm-hex": "1.0.0 <= v < 2.0.0",
         "stil4m/structured-writer": "1.0.1 <= v < 2.0.0"
     },
     "test-dependencies": {
         "elm-explorations/test": "2.0.0 <= v < 3.0.0"
     }
-}
+}

src/Elm/Parser/Base.elm

@@ -1,25 +1,25 @@
 module Elm.Parser.Base exposing (moduleName)
 
-import Elm.Parser.Node as Node
 import Elm.Parser.Tokens as Tokens
 import Elm.Syntax.ModuleName exposing (ModuleName)
-import Elm.Syntax.Node exposing (Node)
+import Elm.Syntax.Node exposing (Node(..))
 import ParserFast
 
 
 moduleName : ParserFast.Parser (Node ModuleName)
 moduleName =
-    ParserFast.map2 (\head tail -> head :: tail)
+    ParserFast.map2WithRange
+        (\range head tail ->
+            Node range (head :: tail)
+        )
         Tokens.typeName
         moduleNameOrEmpty
-        |> Node.parserCore
 
 
 moduleNameOrEmpty : ParserFast.Parser ModuleName
 moduleNameOrEmpty =
-    ParserFast.orSucceed
-        (ParserFast.map2 (\head tail -> head :: tail)
-            (ParserFast.symbolFollowedBy "." Tokens.typeName)
-            (ParserFast.lazy (\() -> moduleNameOrEmpty))
-        )
+    ParserFast.map2OrSucceed
+        (\head tail -> head :: tail)
+        (ParserFast.symbolFollowedBy "." Tokens.typeName)
+        (ParserFast.lazy (\() -> moduleNameOrEmpty))
         []

src/Elm/Parser/Comments.elm

@@ -1,28 +1,31 @@
-module Elm.Parser.Comments exposing (declarationDocumentation, moduleDocumentation, multilineCommentString, singleLineCommentCore)
+module Elm.Parser.Comments exposing (declarationDocumentation, moduleDocumentation, multilineComment, singleLineComment)
 
-import Elm.Parser.Node as Node
+import Char.Extra
 import Elm.Syntax.Documentation exposing (Documentation)
-import Elm.Syntax.Node exposing (Node)
+import Elm.Syntax.Node exposing (Node(..))
 import ParserFast exposing (Parser)
 
 
-singleLineCommentCore : ParserFast.Parser String
-singleLineCommentCore =
+singleLineComment : ParserFast.Parser (Node String)
+singleLineComment =
     ParserFast.symbolFollowedBy "--"
-        (ParserFast.chompWhile (\c -> c /= '\u{000D}' && c /= '\n'))
-        |> ParserFast.getChompedString
+        (ParserFast.whileMap
+            (\c -> c /= '\u{000D}' && c /= '\n' && not (Char.Extra.isUtf16Surrogate c))
+            (\content -> "--" ++ content)
+        )
+        |> ParserFast.mapWithRange Node
 
 
-multilineCommentString : ParserFast.Parser String
-multilineCommentString =
+multilineComment : ParserFast.Parser (Node String)
+multilineComment =
     ParserFast.offsetSourceAndThen
         (\offset source ->
-            case source |> String.slice offset (offset + 3) of
-                "{-|" ->
+            case String.slice (offset + 2) (offset + 3) source of
+                "|" ->
                     problemUnexpectedDocumentation
 
                 _ ->
-                    multiLineCommentStringNoCheck
+                    multiLineCommentNoCheck
         )
 
 
@@ -31,10 +34,11 @@ problemUnexpectedDocumentation =
     ParserFast.problem "unexpected documentation comment"
 
 
-multiLineCommentStringNoCheck : Parser String
-multiLineCommentStringNoCheck =
-    ParserFast.nestableMultiComment "{-" "-}"
-        |> ParserFast.getChompedString
+multiLineCommentNoCheck : Parser (Node String)
+multiLineCommentNoCheck =
+    ParserFast.nestableMultiCommentMapWithRange Node
+        ( '{', "-" )
+        ( '-', "}" )
 
 
 moduleDocumentation : Parser (Node String)
@@ -47,17 +51,4 @@ declarationDocumentation =
     -- technically making the whole parser fail on multi-line comments would be "correct"
     -- but in practice, all declaration comments allow layout before which already handles
     -- these.
-    -- Here's how a "safe" version would look:
-    -- ParserFast.oneOf
-    --[ -- if the next symbol isn't "{-|", we commit to failure
-    --  (ParserFast.symbol "{-" |> ParserFast.ignore ParserFast.chompIf (\c -> c /= '|'))
-    --    |> ParserFast.backtrackable
-    --    |> ParserFast.Extra.continueWith (ParserFast.problem "multiline comment should be documentation comment")
-    --, ParserFast.multiComment "{-" "-}" Nestable
-    --    |> ParserFast.getChompedString
-    --    |> Node.parserCore
-    --]
-    --    |> ParserFast.backtrackable
-    ParserFast.nestableMultiComment "{-" "-}"
-        |> ParserFast.getChompedString
-        |> Node.parserCore
+    multiLineCommentNoCheck