Convert author/message encoding if not UTF-8
Test Plan: used the reparse.php script for reparsing commit messages and saw the correct author name (and mapped correctly as a phabricator user) in diffusion. Reviewers: epriestley Reviewed By: epriestley CC: aran, epriestley Differential Revision: 1059
This commit is contained in:
@@ -26,10 +26,15 @@ class PhabricatorRepositoryGitCommitMessageParserWorker
|
||||
// NOTE: %B was introduced somewhat recently in git's history, so pull
|
||||
// commit message information with %s and %b instead.
|
||||
list($info) = $repository->execxLocalCommand(
|
||||
'log -n 1 --pretty=format:%%an%%x00%%s%%n%%n%%b %s',
|
||||
'log -n 1 --pretty=format:%%e%%x00%%an%%x00%%s%%n%%n%%b %s',
|
||||
$commit->getCommitIdentifier());
|
||||
|
||||
list($author, $message) = explode("\0", $info);
|
||||
list($encoding, $author, $message) = explode("\0", $info);
|
||||
|
||||
if ($encoding != "UTF-8") {
|
||||
$author = mb_convert_encoding($author, 'UTF-8', $encoding);
|
||||
$message = mb_convert_encoding($message, 'UTF-8', $encoding);
|
||||
}
|
||||
|
||||
// Make sure these are valid UTF-8.
|
||||
$author = phutil_utf8ize($author);
|
||||
|
||||
Reference in New Issue
Block a user