new-run-webkit-tests hung while acquiring http lock on snow leopard bots

https://bugs.webkit.org/show_bug.cgi?id=64886

Temporarily disable the http locking to work around the issue.
I'm not actually sure if this is going to work or improve things
much.

Reviewed by Ryosuke Niwa.

* Scripts/webkitpy/layout_tests/port/mac.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@93383 268f45cc-cd09-0410-ab3c-d52691b4dbfc

diff --git a/Tools/ChangeLog b/Tools/ChangeLog
index 61aa56b98ba12d4289eedf9370bb674478d2a6ac..67e37b749b73532c43c1de9c2bb3b1d339554788 100644 (file)
--- a/Tools/ChangeLog
+++ b/Tools/ChangeLog
@@ -1,3 +1,16 @@
+2011-08-18  Dirk Pranke  <dpranke@chromium.org>
+
+        new-run-webkit-tests hung while acquiring http lock on snow leopard bots
+        https://bugs.webkit.org/show_bug.cgi?id=64886
+
+        Temporarily disable the http locking to work around the issue.
+        I'm not actually sure if this is going to work or improve things
+        much.
+
+        Reviewed by Ryosuke Niwa.
+
+        * Scripts/webkitpy/layout_tests/port/mac.py:
+
 2011-08-18  Tony Chang  <tony@chromium.org>
 
         add embedded png checksums to WebKitTestRunner

diff --git a/Tools/Scripts/webkitpy/layout_tests/port/mac.py b/Tools/Scripts/webkitpy/layout_tests/port/mac.py
index 838ed2eb26a6b12a878f02e4d5f07c7aaf065849..d15353970ba03ed762595cc1796b6fa5d8bca658 100644 (file)
--- a/Tools/Scripts/webkitpy/layout_tests/port/mac.py
+++ b/Tools/Scripts/webkitpy/layout_tests/port/mac.py
@@ -155,3 +155,20 @@ class MacPort(ApplePort):
 
     def show_results_html_file(self, results_filename):
         self._run_script('run-safari', ['-NSOpen', results_filename])
+
+    # FIXME: The next two routines turn off the http locking in order
+    # to work around failures on the bots caused when the slave restarts.
+    # See https://bugs.webkit.org/show_bug.cgi?id=64886 for more info.
+    # The proper fix is to make sure the slave is actually stopping NRWT
+    # properly on restart. Note that by removing the lock file and not waiting,
+    # the result should be that if there is a web server already running,
+    # it'll be killed and this one will be started in its place; this
+    # may lead to weird things happening in the other run. However, I don't
+    # think we're (intentionally) actually running multiple runs concurrently
+    # on any Mac bots.
+
+    def acquire_http_lock(self):
+        pass
+
+    def release_http_lock(self):
+        pass