Added support for 'hg export' diff format

Summary: This diff format is used by de-facto mercurial GUI called "TortoiseHG".
It is available as the only way to copy diff into clipboard (right click commit,
select "export", select "copy patch". I added this format support into arcanist
so revisions in Differential can be created from TortoiseHG via simple copy-
paste. Unit test added, manually tested.

See: https://github.com/facebook/arcanist/pull/46

Reviewed by: epriestley

src/parser/ArcanistDiffParser.php

@@ -26,6 +26,7 @@ final class ArcanistDiffParser {
   protected $api;
   protected $text;
   protected $line;
+  protected $lineSaved;
   protected $isGit;
   protected $isMercurial;
   protected $detectBinaryFiles = false;
@@ -234,22 +235,28 @@ final class ArcanistDiffParser {
         '(?P<type>diff -r) (?P<hgrev>[a-f0-9]+) (?:-r [a-f0-9]+ )?(?P<cur>.+)',
       );
 
-      $ok = false;
       $line = $this->getLine();
       $match = null;
-      foreach ($patterns as $pattern) {
+      $ok = $this->tryMatchHeader($patterns, $line, $match);
-        $ok = preg_match('@^'.$pattern.'$@', $line, $match);
-        if ($ok) {
-          break;
-        }
-      }
 
-      if (!$ok) {
+      if (!$ok && $this->isFirstNonEmptyLine()) {
-        $this->didFailParse(
+        // 'hg export' command creates so called "extended diff" that
-          "Expected a hunk header, like 'Index: /path/to/file.ext' (svn), ".
+        // contains some meta information and comment at the beginning
-          "'Property changes on: /path/to/file.ext' (svn properties), ".
+        // (isFirstNonEmptyLine() to check for beginning). Actual mercurial
-          "'commit 59bcc3ad6775562f845953cf01624225' (git show), ".
+        // code detects where comment ends and unified diff starts by
-          "'diff --git' (git diff), or '--- filename' (unified diff).");
+        // searching "diff -r" in the text.
+        $this->saveLine();
+        $line = $this->nextLineThatLooksLikeDiffStart();
+        if (!$this->tryMatchHeader($patterns, $line, $match)) {
+          // Restore line before guessing to display correct error.
+          $this->restoreLine();
+          $this->didFailParse(
+            "Expected a hunk header, like 'Index: /path/to/file.ext' (svn), ".
+            "'Property changes on: /path/to/file.ext' (svn properties), ".
+            "'commit 59bcc3ad6775562f845953cf01624225' (git show), ".
+            "'diff --git' (git diff), '--- filename' (unified diff), or " .
+            "'diff -r' (hg diff or patch).");
+        }
       }
 
       if (isset($match['type'])) {
@@ -325,6 +332,15 @@ final class ArcanistDiffParser {
     return $this->changes;
   }
 
+  protected function tryMatchHeader($patterns, $line, &$match) {
+    foreach ($patterns as $pattern) {
+      if (preg_match('@^'.$pattern.'$@', $line, $match)) {
+        return true;
+      }
+    }
+    return false;
+  }
+
   protected function parseCommitMessage(ArcanistDiffChange $change) {
     $change->setType(ArcanistDiffChangeType::TYPE_MESSAGE);
 
@@ -1026,6 +1042,34 @@ final class ArcanistDiffParser {
     return $this->getLine();
   }
 
+  protected function nextLineThatLooksLikeDiffStart() {
+    while (($line = $this->nextLine()) !== null) {
+      if (preg_match('/^\s*diff\s+-r/', $line)) {
+        break;
+      }
+    }
+    return $this->getLine();
+  }
+
+  protected function saveLine() {
+    $this->lineSaved = $this->line;
+  }
+
+  protected function restoreLine() {
+    $this->line = $this->lineSaved;
+  }
+
+  protected function isFirstNonEmptyLine() {
+    $count = count($this->text);
+    for ($i = 0; $i < $count; $i++) {
+      if (strlen(trim($this->text[$i])) != 0) {
+        return ($i == $this->line);
+      }
+    }
+    // Entire file is empty.
+    return false;
+  }
+
   protected function didFinishParse() {
     $this->text = null;
   }

src/parser/__tests__/ArcanistDiffParserTestCase.php

@@ -537,6 +537,9 @@ EOTEXT
           'Test.java',
           $change->getCurrentPath());
         break;
+      case 'hg-patch.hgdiff':
+        $this->assertEqual(1, count($changes));
+        break;
       default:
         throw new Exception("No test block for diff file {$diff_file}.");
         break;

src/parser/__tests__/diff/hg-patch.hgdiff

@@ -0,0 +1,16 @@
+# Some meta information
+Some
+multiline comment
+
+with empty strings.
+
+diff -r b073ee763933 -r 36fbb0c6c48a vim_nix.rb
+--- a/vim_nix.rb	Wed Jul 04 11:28:53 2012 +0400
++++ b/vim_nix.rb	Wed Jul 25 23:14:44 2012 +0400
+@@ -11,6 +11,3 @@
+   `sudo apt-get install vim-gtk`
+ end
+ 
+-File.delete( File.expand_path( "~/.vimrc" ) ) rescue
+-File.link( "#{DirDropbox()}/cfg/vim/_vimrc", File.expand_path( "~/.vimrc" ) )
+-