added: rule to objectpascal lexer (#888)

* added: another rule for objectpascal lexer * updated: objectpascal lexer testdata * fixed: processing of control (escape) characters * updated: test data for objectpascal lexer
alecthomas · Nov 23, 2023 · d9f6ed6 · d9f6ed6
1 parent c5948a6
commit d9f6ed6
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 6 deletions.
diff --git a/lexers/embedded/objectpascal.xml b/lexers/embedded/objectpascal.xml
@@ -28,7 +28,7 @@
         <token type="Text"/>
       </rule>
       <!-- Compiler Directive -->
-      <rule pattern="\{[$].*?\}|\{[-](NOD|EXT|OBJ).*?\}">
+      <rule pattern="\{[$].*?\}|\{[-](NOD|EXT|OBJ).*?\}|\([*][$].*?[*]\)">
         <token type="CommentPreproc" />
       </rule>
       <!-- Comment Single -->
@@ -50,8 +50,12 @@
       <rule pattern="(?i:(\.\.))">
         <token type="Operator" />
       </rule>
+      <!-- Control Character -->
+      <rule pattern="[\#][0-9a-fA-F]*|[0-9]+[xX][0-9a-fA-F]*">
+        <token type="LiteralStringEscape" />
+      </rule>
       <!-- Numbers -->
-      <rule pattern="[\$][0-9a-fA-F]*[xX][0-9a-fA-F]*|[\$][0-9a-fA-F]*|[\#][0-9a-fA-F]*|[0-9]+[xX][0-9a-fA-F]*|([0-9]+[0-9a-fA-F]+(?=[hH]))">
+      <rule pattern="[\$][0-9a-fA-F]*[xX][0-9a-fA-F]*|[\$][0-9a-fA-F]*|([0-9]+[0-9a-fA-F]+(?=[hH]))">
         <token type="LiteralNumberHex" />
       </rule>
       <rule pattern="[0-9]+(\&#39;[0-9]+)*\.[0-9]+(\&#39;[0-9]+)*[eE][+-]?[0-9]+(\&#39;[0-9]+)*|[0-9]+(\&#39;[0-9]+)*\.[0-9]+(\&#39;[0-9]+)*|\d+[eE][+-]?[0-9]+">

diff --git a/lexers/testdata/objectpascal.actual b/lexers/testdata/objectpascal.actual
@@ -83,6 +83,8 @@ var
 {-NODEFINE    string       'String'        } {$OBJTYPENAME string   'NUnicodeString'} { defined in ustring.h }
 {-EXTERNALSYM ShortInt     'signed char'   } {-OBJTYPENAME ShortInt 'Bzc'}
 
+(*$hints off*) // another valid compiler directive
+
 const
    CLineBreak = {$IFDEF POSIX} _AnsiStr(#10) {$ENDIF}
        {$IFDEF MSWINDOWS} _AnsiStr(#13#10) {$ENDIF};

diff --git a/lexers/testdata/objectpascal.expected b/lexers/testdata/objectpascal.expected
@@ -409,6 +409,11 @@
   {"type":"TextWhitespace","value":" "},
   {"type":"CommentPreproc","value":"{-OBJTYPENAME ShortInt 'Bzc'}"},
   {"type":"Text","value":"\n\n"},
+  {"type":"CommentPreproc","value":"(*$hints off*)"},
+  {"type":"TextWhitespace","value":" "},
+  {"type":"CommentSingle","value":"// another valid compiler directive"},
+  {"type":"TextWhitespace","value":"\n"},
+  {"type":"Text","value":"\n"},
   {"type":"KeywordReserved","value":"const"},
   {"type":"Text","value":"\n"},
   {"type":"TextWhitespace","value":"   "},
@@ -420,7 +425,7 @@
   {"type":"TextWhitespace","value":" "},
   {"type":"Text","value":"_AnsiStr"},
   {"type":"Operator","value":"("},
-  {"type":"LiteralNumberHex","value":"#10"},
+  {"type":"LiteralStringEscape","value":"#10"},
   {"type":"Operator","value":")"},
   {"type":"TextWhitespace","value":" "},
   {"type":"CommentPreproc","value":"{$ENDIF}"},
@@ -430,7 +435,7 @@
   {"type":"TextWhitespace","value":" "},
   {"type":"Text","value":"_AnsiStr"},
   {"type":"Operator","value":"("},
-  {"type":"LiteralNumberHex","value":"#13#10"},
+  {"type":"LiteralStringEscape","value":"#13#10"},
   {"type":"Operator","value":")"},
   {"type":"TextWhitespace","value":" "},
   {"type":"CommentPreproc","value":"{$ENDIF}"},
@@ -1725,7 +1730,7 @@
   {"type":"TextWhitespace","value":" "},
   {"type":"Text","value":"Ord"},
   {"type":"Operator","value":"("},
-  {"type":"LiteralNumberHex","value":"#10"},
+  {"type":"LiteralStringEscape","value":"#10"},
   {"type":"Operator","value":")"},
   {"type":"Punctuation","value":"."},
   {"type":"Text","value":"ToExtended"},
@@ -1947,7 +1952,7 @@
   {"type":"Operator","value":"%"},
   {"type":"Text","value":"fs"},
   {"type":"Operator","value":":("},
-  {"type":"LiteralNumberHex","value":"0x2c"},
+  {"type":"LiteralStringEscape","value":"0x2c"},
   {"type":"Operator","value":")"},
   {"type":"Punctuation","value":","},
   {"type":"Operator","value":"%"},