Reviewers: Jakob,
Message:
This just renames any old test data directory after a new archive is
downloaded.
There are several reasons why to do it this way:
* Preserve data directories the user might have around.
* On buildbots any untracked directory will be deleted before the next run
by
gclient.
* It is hard(TM) to recursively remove directories on Windows buildbots,
better
let gclient do it.
If you have a more elegant solution, let me know. But I want to keep the
"DownloadData" routine as simple as possible. So I am hesitant to adding a
stamp
file or something like that.
Description:
Fix test262 harness to purge old data directory.
[email protected]
Please review this at https://chromiumcodereview.appspot.com/10097006/
SVN Base: https://v8.googlecode.com/svn/branches/bleeding_edge
Affected files:
M test/test262/testcfg.py
Index: test/test262/testcfg.py
diff --git a/test/test262/testcfg.py b/test/test262/testcfg.py
index
fc03c4d02e56f833487506be9f3c8a18a719ce0a..2c9bf06f1428741a935406c769bf085f16f5a21c
100644
--- a/test/test262/testcfg.py
+++ b/test/test262/testcfg.py
@@ -104,9 +104,12 @@ class Test262TestConfiguration(test.TestConfiguration):
archive_url = TEST_262_URL % revision
archive_name = join(self.root, 'test262-%s.tar.bz2' % revision)
directory_name = join(self.root, 'data')
+ directory_old_name = join(self.root, 'data.old')
if not exists(archive_name):
print "Downloading test data from %s ..." % archive_url
urllib.urlretrieve(archive_url, archive_name)
+ if exists(directory_name):
+ os.rename(directory_name, directory_old_name)
if not exists(directory_name):
print "Extracting test262-%s.tar.bz2 ..." % revision
md5 = hashlib.md5()
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
