feat(package): hemar: antlr grammar, but still does not work

package/default.nix

@@ -142,7 +142,8 @@ in {
   onlinepubs2man               = pkgs.callPackage ./onlinepubs2man                    {};
   migrator                     = pkgs.callPackage ./migrator                          {};
   nbt2json                     = pkgs.callPackage ./nbt2json                          {};
-  hemar                        = pkgs.callPackage ./hemar                             {};
+  hemar-parser                 = pkgs.callPackage ./hemar/parser                      {};
+  hemar-grammar                = pkgs.callPackage ./hemar/grammar                     {};
   pg-17-ext-http               = buildHttpExt      pkgs "17";
   pg-17-ext-smtp-client        = buildSmtpExt      pkgs "17";
   pg-17-ext-plhaskell          = buildPlHaskellExt pkgs "17";

package/hemar/grammar/antlr/GoodTry.g4

@@ -0,0 +1,28 @@
+grammar Hemar;
+
+// ----------------- parser rules -----------------
+
+hemar: elements? EOF ;
+
+elements: element+ ;
+
+element
+  : segment
+  | interpoltion 
+  ;
+
+segment : for elements? end ;
+
+for : 'for' 'in' ;
+
+end: 'end' ;
+
+interpoltion : 'mcha' ;
+
+OPEN  : '{[' ;
+CLOSE : ']}' ;
+
+WS  :   [ \t\n\r]+ -> skip ;
+LEADING_TEXT : { getCharPositionInLine() == 0 }? (~'{'|'{'~'[')* OPEN -> skip;
+MIDLE_TEXT   : CLOSE (~'{'|'{'~'[')* OPEN -> skip;
+ENDING_TEXT : CLOSE (~'{'|'{'~'[')* EOF -> skip ;

package/hemar/grammar/antlr/HemarLexer.g4

@@ -0,0 +1,49 @@
+lexer grammar HemarLexer;
+
+// ---------- default mode: plain text ----------
+
+// Everything that is not the start of "{[" is TEXT
+TEXT
+  : ( ~'{' | '{' ~'[' )+
+  ;
+
+// When we see "{[", emit LeftBrace and enter TAG mode
+LeftBrace
+  : '{[' -> pushMode(TAG)
+  ;
+
+// skip whitespace in plain text if you want
+SKIP_WS
+  : [ \t\r\n]+ -> skip
+  ;
+
+// ---------- TAG mode: inside {[ ... ]} ----------
+
+mode TAG;
+
+fragment WS: [ \t\r\n] ;
+
+For : 'for';
+In  : 'in';
+End : 'end';
+
+// identifier inside tag
+Path
+  : String
+  | String '.' Path
+  ;
+
+String
+  : ( ~[.\] \t\r\n] | ']' ~[}. \t\r\n] )+
+  | '"' ( ~'"' | '\\' '"'  )+ '"'
+  ;
+
+// closing "]}": emit RightBrace and go back to default mode
+RightBrace
+  : ']}' -> popMode
+  ;
+
+// skip whitespace inside tag
+SKIP_TAG_WS
+  : WS+ -> skip
+  ;

package/hemar/grammar/antlr/HemarParser.g4

@@ -0,0 +1,18 @@
+parser grammar HemarParser;
+
+options { tokenVocab=HemarLexer; }
+
+hemar   : element*? EOF ;
+
+element
+  : TEXT
+  | segment
+  | interpoltion
+  ;
+
+segment      : for element*? end;
+
+for : LeftBrace For Path In Path RightBrace;
+end : LeftBrace End RightBrace;
+
+interpoltion : LeftBrace Path RightBrace;

package/hemar/grammar/antlr/_Hemar.g4

@@ -0,0 +1,149 @@
+grammar Hemar;
+
+// ----------------- parser rules -----------------
+
+hemar
+    : elements? EOF
+    ;
+
+elements
+    : element+
+    ;
+
+element
+    : tag
+    | TEXT
+    ;
+
+// tag
+tag
+    : OPEN path          CLOSE
+    | OPEN loopStatement CLOSE
+    | OPEN includeHeader CLOSE
+    | OPEN 'end'         CLOSE
+    | OPEN function      CLOSE
+    | OPEN OPEN          CLOSE      // literal "{[" output
+    ;
+
+// loop tag: "for" string "in" path
+loopStatement
+    : 'for' STRING 'in' path
+    ;
+
+// include tag: "include" path
+includeHeader
+    : 'include' path
+    ;
+
+// function tag
+function
+    : 'compute' language functionBody?   // "compute" language body
+    | 'compute' '-'      functionBody?   // "compute" - body
+    ;
+
+language
+    : 'dash'
+    | 'plpgsql'
+    ;
+
+// everything up to (but not including) "]}"
+// (raw body, including "{[" etc, at *token* level)
+functionBody
+    : ( ~CLOSE )*
+    ;
+
+// path
+path
+    : '.'
+    | segmentedPath
+    ;
+
+segmentedPath
+    : segment ('.' segment)*
+    ;
+
+segment
+    : STRING
+    | index
+    ;
+
+// index: \0 .. \9, \1.. \9\d*, and negative forms
+index
+    : '\\' DIGIT
+    | '\\' ONENINE DIGITS?
+    | '\\' '-' DIGIT
+    | '\\' '-' ONENINE DIGITS?
+    ;
+
+// ----------------- lexer rules -----------------
+
+OPEN  : '{[';
+CLOSE : ']}';
+
+// text outside tags: anything except the "{[" sequence
+TEXT
+    : TEXT_CHAR+
+    ;
+
+/*
+ * Strings used in paths/loop variables:
+ *   "..." with escapes similar to your spec.
+ */
+STRING
+    : '"' ( ESC | STRING_CHAR )* '"'
+    ;
+
+fragment STRING_CHAR
+    : ~["\\\r\n]
+    ;
+
+/*
+ * Escapes:
+ *   .   (literal dot)
+ *   ]}  (literal "]}")   -- note this is two chars after '\'
+ *   "   \" 
+ *   \   \\
+ *   /   \/
+ *   b f n r t
+ *   uXXXX (hex)
+ *   whitespace after backslash (your ws-in-escape)
+ */
+fragment ESC
+    : '\\'
+      (
+          '.'
+        | ']}'
+        | '"'
+        | '\\'
+        | '/'
+        | 'b'
+        | 'f'
+        | 'n'
+        | 'r'
+        | 't'
+        | 'u' HEX HEX HEX HEX
+        | WS_CHAR
+      )
+    ;
+
+// digits / hex
+DIGITS : DIGIT+ ;
+DIGIT  : [0-9] ;
+ONENINE: [1-9] ;
+HEX    : [0-9a-fA-F] ;
+
+// whitespace for normal lexing
+WS
+    : [ \t\r\n]+ -> skip
+    ;
+
+// whitespace used inside escapes
+fragment WS_CHAR
+    : [ \t\r\n]
+    ;
+
+
+fragment TEXT_CHAR
+    : ~'{'              // any except '{'
+    | '{' ~'['          // '{' only if not starting OPEN
+    ;

package/hemar/grammar/antlr/default.nix

@@ -0,0 +1,36 @@
+{ stdenv, symlinkJoin, jre, antlr4, runtimeShell, jdk }:
+
+let
+  hemar-grammar = stdenv.mkDerivation {
+    pname = "hemar-grammar";
+    version = "0.1.0";
+
+    src = ./.; # directory with Hemar.g4
+
+    nativeBuildInputs = [
+      antlr4
+      jdk
+    ];
+
+    buildPhase = ''
+      antlr4 HemarLexer.g4 HemarParser.g4
+      javac *.java
+    '';
+
+    installPhase = ''
+      mkdir -p "$out/lib" "$out/bin"
+      cp *.class *.tokens "$out/lib"
+
+      cat > "$out/bin/hemar-grammar" <<EOF
+#!${runtimeShell}
+CLASSPATH="$out/lib:${antlr4}/share/java/*"
+exec ${jre}/bin/java -cp "\$CLASSPATH" org.antlr.v4.gui.TestRig Hemar hemar "\$@"
+EOF
+      chmod +x "$out/bin/hemar-grammar"
+    '';
+  };
+in
+symlinkJoin {
+  name = "hemar-grammar";
+  paths = [ hemar-grammar ];
+}

package/hemar/parser/hemar.sh

@@ -310,12 +310,6 @@ find_close_pattern() {
           return 1
         fi
       ;;
-      '.'|'/'|b|f|n|r|t)
-      ;;
-      ']')
-      ;;
-      u)
-      ;;
       *)
         if [ "${TAG_escape_flag+x}" ]; then
           if is_ws "$char"; then