Don't require PsiFile.getText() for detecting Python charset declaration

python/psi-api/src/com/jetbrains/python/PythonFileType.java

@@ -16,11 +16,15 @@
 package com.jetbrains.python;
 
 import com.intellij.lang.Language;
+import com.intellij.openapi.editor.Document;
 import com.intellij.openapi.fileTypes.LanguageFileType;
 import com.intellij.openapi.project.Project;
+import com.intellij.openapi.util.TextRange;
 import com.intellij.openapi.util.text.StringUtil;
 import com.intellij.openapi.vfs.CharsetToolkit;
 import com.intellij.openapi.vfs.VirtualFile;
+import com.intellij.psi.PsiDocumentManager;
+import com.intellij.psi.PsiFile;
 import icons.PythonPsiApiIcons;
 import org.jetbrains.annotations.NotNull;
 import org.jetbrains.annotations.Nullable;
@@ -41,6 +45,7 @@ import java.util.regex.Pattern;
  */
 public class PythonFileType extends LanguageFileType {
   private static final Pattern ENCODING_PATTERN = Pattern.compile("coding[:=]\\s*([-\\w.]+)");
+  public static final int MAX_CHARSET_ENCODING_LINE = 2;
 
   public static PythonFileType INSTANCE = new PythonFileType();
 
@@ -100,14 +105,28 @@ public class PythonFileType extends LanguageFileType {
   }
 
   @Nullable
-  public static String getCharsetFromEncodingDeclaration(String content) {
+  public static String getCharsetFromEncodingDeclaration(@NotNull PsiFile file) {
+    final Document document = PsiDocumentManager.getInstance(file.getProject()).getDocument(file);
+    final String content;
+    if (document != null && document.getLineCount() > MAX_CHARSET_ENCODING_LINE) {
+      final int offset = document.getLineEndOffset(MAX_CHARSET_ENCODING_LINE);
+      content = document.getText(TextRange.create(0, offset));
+    }
+    else {
+      content = file.getText();
+    }
+    return getCharsetFromEncodingDeclaration(content);
+  }
+
+  @Nullable
+  private static String getCharsetFromEncodingDeclaration(@Nullable String content) {
     if (content == null || content.isEmpty()) {
       return null;
     }
     try {
       final BufferedReader reader = new BufferedReader(new StringReader(content));
       try {
-        for (int i = 0; i < 2; i++) {
+        for (int i = 0; i < MAX_CHARSET_ENCODING_LINE; i++) {
           final String line = reader.readLine();
           if (line == null) {
             return null;

python/src/com/jetbrains/python/inspections/PyByteLiteralInspection.java

@@ -75,7 +75,7 @@ public class PyByteLiteralInspection extends PyInspection {
         );
       }
 
-      final String charsetString = PythonFileType.getCharsetFromEncodingDeclaration(file.getText());
+      final String charsetString = PythonFileType.getCharsetFromEncodingDeclaration(file);
       try {
         if (charsetString != null && !Charset.forName(charsetString).equals(Charset.forName("US-ASCII")))
           default_bytes = false;

python/src/com/jetbrains/python/inspections/PyMandatoryEncodingInspection.java

@@ -62,7 +62,7 @@ public class PyMandatoryEncodingInspection extends PyInspection {
 
     @Override
     public void visitPyFile(PyFile node) {
-      final String charsetString = PythonFileType.getCharsetFromEncodingDeclaration(node.getText());
+      final String charsetString = PythonFileType.getCharsetFromEncodingDeclaration(node);
       if (charsetString == null) {
         TextRange tr = new TextRange(0,0);
         ProblemsHolder holder = getHolder();

python/src/com/jetbrains/python/inspections/PyNonAsciiCharInspection.java

@@ -69,7 +69,7 @@ public class PyNonAsciiCharInspection extends PyInspection {
       if (LanguageLevel.forElement(node).isPy3K()) return;
       PsiFile file = node.getContainingFile(); // can't cache this in the instance, alas
       if (file == null) return;
-      final String charsetString = PythonFileType.getCharsetFromEncodingDeclaration(file.getText());
+      final String charsetString = PythonFileType.getCharsetFromEncodingDeclaration(file);
 
       boolean hasNonAscii = false;
 

python/testSrc/com/jetbrains/python/PyEncodingTest.java

@@ -15,12 +15,12 @@
  */
 package com.jetbrains.python;
 
-import junit.framework.TestCase;
+import com.jetbrains.python.fixtures.PyTestCase;
 
 /**
  * @author yole
  */
-public class PyEncodingTest extends TestCase {
+public class PyEncodingTest extends PyTestCase {
   public void testEncodingEmacs() {
     doTest("#!/usr/bin/python\n# -*- coding: iso-8859-15 -*-\nimport os, sys", "iso-8859-15");
   }
@@ -33,7 +33,8 @@ public class PyEncodingTest extends TestCase {
     doTest("#!/usr/local/bin/python\n# coding: latin-1\nimport os, sys", "iso-8859-1");
   }
 
-  private static void doTest(final String text, final String expected) {
-    assertEquals(expected, PythonFileType.getCharsetFromEncodingDeclaration(text));
+  private void doTest(final String text, final String expected) {
+    myFixture.configureByText(PythonFileType.INSTANCE, text);
+    assertEquals(expected, PythonFileType.getCharsetFromEncodingDeclaration(myFixture.getFile()));
   }
 }