140821 2015-01-23 04:46:11 -0800 [buildbot] Simplify jscore-test buildstep 2015-02-02 11:15:13 -0800 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 140884 141165 1 ossy ossy ap bfulgham commit-queue fpizlo ggaren lforschler ossy rniwa oldest_to_newest 1063654 0 ossy 2015-01-23 04:46:11 -0800 RunJavaScriptCoreTests can be TestWithFailureCount as other tester steps. We would need only one countFailures instead of evaluateCommand, getText and getText2. actual.html log can be removed too, it is used by only the old style Mozilla JSC tests (run-javascriptcore-tests --no-jsc-stress). Now only Windows bots run this test, but this log isn't useful for anything, and Windows bots should switch to the modern JSC stress test to get better test coverage. And the empty logs on other bots are confusing. 1063656 1 245222 ossy 2015-01-23 04:53:11 -0800 Created attachment 245222 Patch 1063657 2 commit-queue 2015-01-23 04:54:17 -0800 Attachment 245222 did not pass style-queue: ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125: [RunJavaScriptCoreTestsTest.test_mozilla_failure_old_output] Undefined variable 'FAILURE' [pylint/E0602] [5] ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:130: [RunJavaScriptCoreTestsTest.test_mozilla_failures_old_output] Undefined variable 'FAILURE' [pylint/E0602] [5] ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:135: [RunJavaScriptCoreTestsTest.test_jsc_stress_failure_new_output] Undefined variable 'FAILURE' [pylint/E0602] [5] ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:139: [RunJavaScriptCoreTestsTest.test_jsc_stress_failures_new_output] Undefined variable 'FAILURE' [pylint/E0602] [5] Total errors found: 4 in 3 files If any of these errors are false positives, please file a bug against check-webkit-style. 1063660 3 ossy 2015-01-23 05:23:23 -0800 Just a note: I proposed to enable JSC stress tests on Windows by default in bug128307. Once that patch landed actual.html won't exist on any bot. 1063707 4 245222 ap 2015-01-23 10:00:17 -0800 Comment on attachment 245222 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=245222&action=review > Tools/ChangeLog:3 > + [buildbot] Simplify jscore-test buildstep Longer term, we need to output a JSON file with results that build.webkit.org/dashboard could parse and display in a pop-up. Probably also split the failure output into separate files per test, as searching for the failures in a gigantic single log is a waste of time. This seems orthogonal to this patch though. Another orthogonal comment: we should probably consider not running JSC stress tests on regular bots all the time, they are super slow, and probably take away more value by delaying results than they add by running the tests. > Tools/ChangeLog:9 > + (RunJavaScriptCoreTests): Inherited from TestWithFailureCount and removed useless actual.html logfile. I tried to find any failing "Mozilla" tests in recent test runs, and couldn't. But isn't this file what is generated when Mozilla tests fail? > Tools/BuildSlaveSupport/build.webkit.org-config/master.cfg:303 > + if match: > + return int(match.group(1)) This looks surprising. We don't add the failure counts if both Mozilla and JSC tests failed? Or does this get invoked twice for each test type? > Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125 > + self.assertResults(FAILURE, ["1 JSC tests failed"], 1, """Results for Mozilla tests: "1 tests" is not good grammar. Given that one failing test is the most common failure case, I think that the error text should be nice. 1063725 5 ossy 2015-01-23 11:00:57 -0800 (In reply to comment #4) > Longer term, we need to output a JSON file with results that > build.webkit.org/dashboard could parse and display in a pop-up. Probably > also split the failure output into separate files per test, as searching for > the failures in a gigantic single log is a waste of time. JSC tests are almost always pass, regression happens very rarely and in this case the core JSC hackers fixes the issue immediately or roll out the patch and then fix offline. JSON results and exact output isn't really necessarry. > Another orthogonal comment: we should probably consider not running JSC > stress tests on regular bots all the time, they are super slow, and probably > take away more value by delaying results than they add by running the tests. On the release bots it takes only 3.5 minutes, but on the debug bots it takes near 30 minutes. > > Tools/ChangeLog:9 > > + (RunJavaScriptCoreTests): Inherited from TestWithFailureCount and removed useless actual.html logfile. > > I tried to find any failing "Mozilla" tests in recent test runs, and > couldn't. But isn't this file what is generated when Mozilla tests fail? It is, but only the Windows bot runs the Mozilla tests, other runs JSC stress test, which don't generate actual.html. > > Tools/BuildSlaveSupport/build.webkit.org-config/master.cfg:303 > > + if match: > > + return int(match.group(1)) > > This looks surprising. We don't add the failure counts if both Mozilla and > JSC tests failed? No, we can't run both of them in the same time, see http://trac.webkit.org/browser/trunk/Tools/Scripts/run-javascriptcore-tests?rev=175142#L204 > > Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125 > > + self.assertResults(FAILURE, ["1 JSC tests failed"], 1, """Results for Mozilla tests: > > "1 tests" is not good grammar. Given that one failing test is the most > common failure case, I think that the error text should be nice. I agree, it would be better to use the proper grammar here. The existing code do this, but the generic TestWithFailureCount can't handle it. To make it work, we should mark the position of the plural suffix in failedTestsFormatString somehow. It would fix the grammar of webkitpy, webkitperl, unit tests, ... too. 1063733 6 ap 2015-01-23 11:22:38 -0800 > JSC tests are almost always pass, regression happens very rarely They were a lot more flaky for several months this fall, and figuring out which ones failed most frequently was an important part of fixing the problem. Even now, we have random crashes on stress/poly-call-exit.js.ftl-eager. Looking at <https://build.webkit.org/builders/Apple%20Yosemite%20Release%20WK2%20%28Tests%29?numbuilds=200>, jsc failures are a very common occurrence, although this disproves my claim that single failure is the most common case. It used to be quite common recently though. > It is, but only the Windows bot runs the Mozilla tests, other runs JSC stress test I see! 1064136 7 245341 ossy 2015-01-26 04:36:40 -0800 Created attachment 245341 Patch It depends on bug140884, which fixes the grammar issue of TestWithFailureCount. 1064137 8 commit-queue 2015-01-26 04:38:00 -0800 Attachment 245341 did not pass style-queue: ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125: [RunJavaScriptCoreTestsTest.test_mozilla_failure_old_output] Undefined variable 'FAILURE' [pylint/E0602] [5] ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:130: [RunJavaScriptCoreTestsTest.test_mozilla_failures_old_output] Undefined variable 'FAILURE' [pylint/E0602] [5] ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:135: [RunJavaScriptCoreTestsTest.test_jsc_stress_failure_new_output] Undefined variable 'FAILURE' [pylint/E0602] [5] ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:139: [RunJavaScriptCoreTestsTest.test_jsc_stress_failures_new_output] Undefined variable 'FAILURE' [pylint/E0602] [5] Total errors found: 4 in 3 files If any of these errors are false positives, please file a bug against check-webkit-style. 1064479 9 ossy 2015-01-27 03:52:44 -0800 grammar fixed by http://trac.webkit.org/changeset/179116 and Windows bots already run stress tests since http://trac.webkit.org/changeset/179165 , so I think it's ready for landing now. 1064542 10 245341 ap 2015-01-27 09:49:03 -0800 Comment on attachment 245341 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=245341&action=review > Tools/ChangeLog:4 > + [buildbot] Simplify jscore-test buildstep > + https://bugs.webkit.org/show_bug.cgi?id=140821 How does this change Buildbot's JSON results, will that break the dashboard? I'm not sure if I can tell from the patch. 1064546 11 ossy 2015-01-27 10:07:13 -0800 (In reply to comment #10) > Comment on attachment 245341 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=245341&action=review > > > Tools/ChangeLog:4 > > + [buildbot] Simplify jscore-test buildstep > > + https://bugs.webkit.org/show_bug.cgi?id=140821 > > How does this change Buildbot's JSON results, will that break the dashboard? > I'm not sure if I can tell from the patch. The JSON result would be changed a little bit, but it shouldn't affect the dashboard. eg. missing links for the empty actual.html and actual.html (source) log files, a little bit different result summary. I checked the dashboard code, it finds steps by name, which isn't changed, it is still "jscore-test". And it counts failures with "fail" word, which is still present in the output. I'm 99% sure it won't break the dashboard. 1064951 12 ossy 2015-01-28 10:26:44 -0800 ping? 1064956 13 245341 ap 2015-01-28 11:01:10 -0800 Comment on attachment 245341 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=245341&action=review > Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125 > - self.assertResults(FAILURE, ["jscore-test", '1 failing Mozilla test '], 1, """Results for Mozilla tests: > + self.assertResults(FAILURE, ["1 JSC test failed"], 1, """Results for Mozilla tests: The dashboard looks at testStep.results[1], and this patch moves output to results[0] AFAICT. It seems to become different from all other steps. One complication here is that the dashboard needs to work with both internal and open source buildbots, so changes to format need to be made synchronously. 1064964 14 ossy 2015-01-28 11:20:27 -0800 (In reply to comment #13) > Comment on attachment 245341 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=245341&action=review > > > Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125 > > - self.assertResults(FAILURE, ["jscore-test", '1 failing Mozilla test '], 1, """Results for Mozilla tests: > > + self.assertResults(FAILURE, ["1 JSC test failed"], 1, """Results for Mozilla tests: > > The dashboard looks at testStep.results[1], and this patch moves output to > results[0] AFAICT. It seems to become different from all other steps. > > One complication here is that the dashboard needs to work with both internal > and open source buildbots, so changes to format need to be made > synchronously. Strange, only jsc tests was different, layout/perl/python tests dumps everything to the first line. Let me check it in the details. 1064965 15 ossy 2015-01-28 11:28:41 -0800 (In reply to comment #13) > Comment on attachment 245341 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=245341&action=review > > > Tools/BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:125 > > - self.assertResults(FAILURE, ["jscore-test", '1 failing Mozilla test '], 1, """Results for Mozilla tests: > > + self.assertResults(FAILURE, ["1 JSC test failed"], 1, """Results for Mozilla tests: > > The dashboard looks at testStep.results[1], and this patch moves output to > results[0] AFAICT. It seems to become different from all other steps. > One complication here is that the dashboard needs to work with both internal > and open source buildbots, so changes to format need to be made > synchronously. I checked it, results[0] is the status of the buildstep, SUCCESS, FAIL, ... and result[1] contains the text results line by line, so it won't break the dashbord anywhere. Here is a good example: https://build.webkit.org/json/builders/Apple%20Win%207%20Debug%20%28Tests%29/builds/64427?as_text=1 "name": "jscore-test", "results": [ 2, [ "jscore-test", "6 failing JSC stress tests " ] ], ... "name": "layout-test", "results": [ 2, [ "63 failures", "45 new passes", "3 flakes", "3 missing results" ] ], ... 1064967 16 245341 ap 2015-01-28 11:37:51 -0800 Comment on attachment 245341 Patch I see. 1065211 17 245341 commit-queue 2015-01-29 05:32:25 -0800 Comment on attachment 245341 Patch Clearing flags on attachment: 245341 Committed r179342: <http://trac.webkit.org/changeset/179342> 1065212 18 commit-queue 2015-01-29 05:32:30 -0800 All reviewed patches have been landed. Closing bug. 1065215 19 ossy 2015-01-29 05:58:45 -0800 Buildbot is died after is landed: https://build.webkit.org/waterfall Service Temporarily Unavailable The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later. Apache/2.2 Server at build.webkit.org Port 443 Lucas, could you check what happened with the master? I thought changes are pushed to it manually by you. Otherwise this change is good, I checked it locally, so I have no idea what happened with the master. 1065218 20 ossy 2015-01-29 06:32:41 -0800 It works again. 245222 2015-01-23 04:53:11 -0800 2015-01-26 04:36:32 -0800 Patch bug-140821-20150123135512.patch text/plain 6262 ossy U3VidmVyc2lvbiBSZXZpc2lvbjogMTc4OTg5CmRpZmYgLS1naXQgYS9Ub29scy9DaGFuZ2VMb2cg Yi9Ub29scy9DaGFuZ2VMb2cKaW5kZXggNWM5MjZiMzJhOTRlYTliYTRiNGU5NmIyNjFmZjNiMWVh MDYxZmQ4Zi4uNDVhMzYwMzUwMTEzNzAxYTIyZTVmYmUwOTU4N2RlZTJlMjc1NzY5MSAxMDA2NDQK LS0tIGEvVG9vbHMvQ2hhbmdlTG9nCisrKyBiL1Rvb2xzL0NoYW5nZUxvZwpAQCAtMSwzICsxLDIz IEBACisyMDE1LTAxLTIzICBDc2FiYSBPc3p0cm9nb27DoWMgIDxvc3N5QHdlYmtpdC5vcmc+CisK KyAgICAgICAgW2J1aWxkYm90XSBTaW1wbGlmeSBqc2NvcmUtdGVzdCBidWlsZHN0ZXAKKyAgICAg ICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE0MDgyMQorCisgICAg ICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgICogQnVpbGRTbGF2ZVN1 cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25maWcvbWFzdGVyLmNmZzoKKyAgICAgICAgKFJ1bkph dmFTY3JpcHRDb3JlVGVzdHMpOiBJbmhlcml0ZWQgZnJvbSBUZXN0V2l0aEZhaWx1cmVDb3VudCBh bmQgcmVtb3ZlZCB1c2VsZXNzIGFjdHVhbC5odG1sIGxvZ2ZpbGUuCisgICAgICAgIChSdW5KYXZh U2NyaXB0Q29yZVRlc3RzLmNvdW50RmFpbHVyZXMpOiBBZGRlZC4KKyAgICAgICAgKFJ1bkphdmFT Y3JpcHRDb3JlVGVzdHMuY29tbWFuZENvbXBsZXRlKTogRGVsZXRlZC4KKyAgICAgICAgKFJ1bkph dmFTY3JpcHRDb3JlVGVzdHMuZXZhbHVhdGVDb21tYW5kKTogRGVsZXRlZC4KKyAgICAgICAgKFJ1 bkphdmFTY3JpcHRDb3JlVGVzdHMuZ2V0VGV4dCk6IERlbGV0ZWQuCisgICAgICAgIChSdW5KYXZh U2NyaXB0Q29yZVRlc3RzLmdldFRleHQyKTogRGVsZXRlZC4KKyAgICAgICAgKiBCdWlsZFNsYXZl U3VwcG9ydC9idWlsZC53ZWJraXQub3JnLWNvbmZpZy9tYXN0ZXJjZmdfdW5pdHRlc3QucHk6IFVw ZGF0ZWQuCisgICAgICAgIChSdW5KYXZhU2NyaXB0Q29yZVRlc3RzVGVzdC50ZXN0X21vemlsbGFf ZmFpbHVyZV9vbGRfb3V0cHV0KToKKyAgICAgICAgKFJ1bkphdmFTY3JpcHRDb3JlVGVzdHNUZXN0 LnRlc3RfbW96aWxsYV9mYWlsdXJlc19vbGRfb3V0cHV0KToKKyAgICAgICAgKFJ1bkphdmFTY3Jp cHRDb3JlVGVzdHNUZXN0LnRlc3RfanNjX3N0cmVzc19mYWlsdXJlX25ld19vdXRwdXQpOgorICAg ICAgICAoUnVuSmF2YVNjcmlwdENvcmVUZXN0c1Rlc3QudGVzdF9qc2Nfc3RyZXNzX2ZhaWx1cmVz X25ld19vdXRwdXQpOgorCiAyMDE1LTAxLTIyICBEYXZpZCBLaWx6ZXIgIDxkZGtpbHplckBhcHBs ZS5jb20+CiAKICAgICAgICAgY29uZmlndXJlLXhjb2RlLWZvci1pb3MtZGV2ZWxvcG1lbnQgbmVl ZHMgdG8gaW1wb3J0IHdlYmtpdGRpcnMucG0KZGlmZiAtLWdpdCBhL1Rvb2xzL0J1aWxkU2xhdmVT dXBwb3J0L2J1aWxkLndlYmtpdC5vcmctY29uZmlnL21hc3Rlci5jZmcgYi9Ub29scy9CdWlsZFNs YXZlU3VwcG9ydC9idWlsZC53ZWJraXQub3JnLWNvbmZpZy9tYXN0ZXIuY2ZnCmluZGV4IDgyNmZi NDJiYjJjMjEwNTEwMjRhNzA5Zjc2ZDgyZjEyNzc4OTkwOTcuLjNjZmQ5NmEyMjQ5YTY2YjU0MTRk YTliYTQ2YWVlOWVjODliNTk0NDIgMTAwNjQ0Ci0tLSBhL1Rvb2xzL0J1aWxkU2xhdmVTdXBwb3J0 L2J1aWxkLndlYmtpdC5vcmctY29uZmlnL21hc3Rlci5jZmcKKysrIGIvVG9vbHMvQnVpbGRTbGF2 ZVN1cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25maWcvbWFzdGVyLmNmZwpAQCAtMjc3LDEyICsy NzcsMTIgQEAgY2xhc3MgRG93bmxvYWRCdWlsdFByb2R1Y3Qoc2hlbGwuU2hlbGxDb21tYW5kKToK ICAgICBmbHVua09uRmFpbHVyZSA9IFRydWUKIAogCi1jbGFzcyBSdW5KYXZhU2NyaXB0Q29yZVRl c3RzKHNoZWxsLlRlc3QpOgorY2xhc3MgUnVuSmF2YVNjcmlwdENvcmVUZXN0cyhUZXN0V2l0aEZh aWx1cmVDb3VudCk6CiAgICAgbmFtZSA9ICJqc2NvcmUtdGVzdCIKICAgICBkZXNjcmlwdGlvbiA9 IFsianNjb3JlLXRlc3RzIHJ1bm5pbmciXQogICAgIGRlc2NyaXB0aW9uRG9uZSA9IFsianNjb3Jl LXRlc3RzIl0KICAgICBjb21tYW5kID0gWyJwZXJsIiwgIi4vVG9vbHMvU2NyaXB0cy9ydW4tamF2 YXNjcmlwdGNvcmUtdGVzdHMiLCBXaXRoUHJvcGVydGllcygiLS0lKGNvbmZpZ3VyYXRpb24pcyIp XQotICAgIGxvZ2ZpbGVzID0geydhY3R1YWwuaHRtbCAoc291cmNlKSc6ICdTb3VyY2UvSmF2YVNj cmlwdENvcmUvdGVzdHMvbW96aWxsYS9hY3R1YWwuaHRtbCd9CisgICAgZmFpbGVkVGVzdHNGb3Jt YXRTdHJpbmcgPSAiJWQgSlNDIHRlc3RzIGZhaWxlZCIKIAogICAgIGRlZiBfX2luaXRfXyhzZWxm LCBidWlsZEpTQ1Rvb2w9VHJ1ZSwgKmFyZ3MsICoqa3dhcmdzKToKICAgICAgICAgc2VsZi5idWls ZEpTQ1Rvb2wgPSBidWlsZEpTQ1Rvb2wKQEAgLTI5NSw1MSArMjk1LDE4IEBAIGNsYXNzIFJ1bkph dmFTY3JpcHRDb3JlVGVzdHMoc2hlbGwuVGVzdCk6CiAgICAgICAgICAgICBzZWxmLnNldENvbW1h bmQoc2VsZi5jb21tYW5kICsgWyctLW5vLWJ1aWxkJ10pCiAgICAgICAgIHJldHVybiBzaGVsbC5U ZXN0LnN0YXJ0KHNlbGYpCiAKLSAgICBkZWYgY29tbWFuZENvbXBsZXRlKHNlbGYsIGNtZCk6Ci0g ICAgICAgIHNoZWxsLlRlc3QuY29tbWFuZENvbXBsZXRlKHNlbGYsIGNtZCkKKyAgICBkZWYgY291 bnRGYWlsdXJlcyhzZWxmLCBjbWQpOgogICAgICAgICBsb2dUZXh0ID0gY21kLmxvZ3NbJ3N0ZGlv J10uZ2V0VGV4dCgpCiAKLSAgICAgICAgZXhwcmVzc2lvbnMgPSBbCi0gICAgICAgICAgICAoJ2Zh aWxpbmcgTW96aWxsYSB0ZXN0JywgcmUuY29tcGlsZShyJ15SZXN1bHRzIGZvciBNb3ppbGxhIHRl c3RzOlxyP1xuXHMrKFxkKykgcmVncmVzc2lvbicsIHJlLk1VTFRJTElORSkpLAotICAgICAgICAg ICAgKCdmYWlsaW5nIEpTQyBzdHJlc3MgdGVzdCcsIHJlLmNvbXBpbGUocideUmVzdWx0cyBmb3Ig SlNDIHN0cmVzcyB0ZXN0czpccj9cblxzKyhcZCspIGZhaWx1cmUnLCByZS5NVUxUSUxJTkUpKSwK LSAgICAgICAgXQotICAgICAgICByZXN1bHRMaW5lcyA9IHt9Ci0KLSAgICAgICAgZm9yIG5hbWUs IGV4cHJlc3Npb24gaW4gZXhwcmVzc2lvbnM6Ci0gICAgICAgICAgICBtYXRjaCA9IGV4cHJlc3Np b24uc2VhcmNoKGxvZ1RleHQpCi0gICAgICAgICAgICByZXN1bHRMaW5lc1tuYW1lXSA9IGludCht YXRjaC5ncm91cCgxKSkgaWYgbWF0Y2ggZWxzZSAwCi0KLSAgICAgICAgc2VsZi5yZWdyZXNzaW9u TGluZSA9IFtdCi0gICAgICAgIGZvciBuYW1lIGluIHJlc3VsdExpbmVzOgotICAgICAgICAgICAg ZmFpbHVyZXMgPSByZXN1bHRMaW5lc1tuYW1lXQotICAgICAgICAgICAgaWYgbm90IGZhaWx1cmVz OgotICAgICAgICAgICAgICAgIGNvbnRpbnVlCi0KLSAgICAgICAgICAgIGlmIGZhaWx1cmVzID09 IDE6Ci0gICAgICAgICAgICAgICAgcGx1cmFsU3VmZml4ID0gJycKLSAgICAgICAgICAgIGVsc2U6 Ci0gICAgICAgICAgICAgICAgcGx1cmFsU3VmZml4ID0gJ3MnCi0gICAgICAgICAgICBzZWxmLnJl Z3Jlc3Npb25MaW5lLmFwcGVuZCgiJWQgJXMlcyAiICUgKGZhaWx1cmVzLCBuYW1lLCBwbHVyYWxT dWZmaXgpKQotCi0gICAgICAgIGlmICdhY3R1YWwuaHRtbCAoc291cmNlKScgaW4gY21kLmxvZ3M6 Ci0gICAgICAgICAgICBzZWxmLmFkZEhUTUxMb2coJ2FjdHVhbC5odG1sJywgY21kLmxvZ3NbJ2Fj dHVhbC5odG1sIChzb3VyY2UpJ10uZ2V0VGV4dCgpKQotCi0gICAgZGVmIGV2YWx1YXRlQ29tbWFu ZChzZWxmLCBjbWQpOgotICAgICAgICBpZiBzZWxmLnJlZ3Jlc3Npb25MaW5lOgotICAgICAgICAg ICAgcmV0dXJuIEZBSUxVUkUKLQotICAgICAgICBpZiBjbWQucmMgIT0gMDoKLSAgICAgICAgICAg IHJldHVybiBGQUlMVVJFCi0KLSAgICAgICAgcmV0dXJuIFNVQ0NFU1MKKyAgICAgICAgbWF0Y2gg PSByZS5zZWFyY2gocideUmVzdWx0cyBmb3IgSlNDIHN0cmVzcyB0ZXN0czpccj9cblxzKyhcZCsp IGZhaWx1cmUnLCBsb2dUZXh0LCByZS5NVUxUSUxJTkUpCisgICAgICAgIGlmIG1hdGNoOgorICAg ICAgICAgICAgcmV0dXJuIGludChtYXRjaC5ncm91cCgxKSkKIAotICAgIGRlZiBnZXRUZXh0KHNl bGYsIGNtZCwgcmVzdWx0cyk6Ci0gICAgICAgIHJldHVybiBzZWxmLmdldFRleHQyKGNtZCwgcmVz dWx0cykKKyAgICAgICAgbWF0Y2ggPSByZS5zZWFyY2gocideUmVzdWx0cyBmb3IgTW96aWxsYSB0 ZXN0czpccj9cblxzKyhcZCspIHJlZ3Jlc3Npb24nLCBsb2dUZXh0LCByZS5NVUxUSUxJTkUpCisg ICAgICAgIGlmIG1hdGNoOgorICAgICAgICAgICAgcmV0dXJuIGludChtYXRjaC5ncm91cCgxKSkK IAotICAgIGRlZiBnZXRUZXh0MihzZWxmLCBjbWQsIHJlc3VsdHMpOgotICAgICAgICByZXN1bHQg PSBbc2VsZi5uYW1lXQotICAgICAgICByZXN1bHQuZXh0ZW5kKHNlbGYucmVncmVzc2lvbkxpbmUp Ci0gICAgICAgIHJldHVybiByZXN1bHQKKyAgICAgICAgcmV0dXJuIDAKIAogY2xhc3MgUnVuV2Vi S2l0VGVzdHMoc2hlbGwuVGVzdCk6CiAgICAgbmFtZSA9ICJsYXlvdXQtdGVzdCIKZGlmZiAtLWdp dCBhL1Rvb2xzL0J1aWxkU2xhdmVTdXBwb3J0L2J1aWxkLndlYmtpdC5vcmctY29uZmlnL21hc3Rl cmNmZ191bml0dGVzdC5weSBiL1Rvb2xzL0J1aWxkU2xhdmVTdXBwb3J0L2J1aWxkLndlYmtpdC5v cmctY29uZmlnL21hc3RlcmNmZ191bml0dGVzdC5weQppbmRleCAwNDFhMDZhNjAzZjM2ZjE3MTI2 YjA2NjBkNjA3N2IxZWVjNmQyNmU0Li45YzYyNWJhOGE4MmEwNjg0YWEzMmRmZmE0ZTgzNTM1NGQ0 MTFhODE5IDEwMDc1NQotLS0gYS9Ub29scy9CdWlsZFNsYXZlU3VwcG9ydC9idWlsZC53ZWJraXQu b3JnLWNvbmZpZy9tYXN0ZXJjZmdfdW5pdHRlc3QucHkKKysrIGIvVG9vbHMvQnVpbGRTbGF2ZVN1 cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25maWcvbWFzdGVyY2ZnX3VuaXR0ZXN0LnB5CkBAIC0x MjIsMjEgKzEyMiwyMSBAQCBjbGFzcyBSdW5KYXZhU2NyaXB0Q29yZVRlc3RzVGVzdCh1bml0dGVz dC5UZXN0Q2FzZSk6CiAgICAgT0suIiIiKQogCiAgICAgZGVmIHRlc3RfbW96aWxsYV9mYWlsdXJl X29sZF9vdXRwdXQoc2VsZik6Ci0gICAgICAgIHNlbGYuYXNzZXJ0UmVzdWx0cyhGQUlMVVJFLCBb ImpzY29yZS10ZXN0IiwgJzEgZmFpbGluZyBNb3ppbGxhIHRlc3QgJ10sIDEsICIiIlJlc3VsdHMg Zm9yIE1vemlsbGEgdGVzdHM6CisgICAgICAgIHNlbGYuYXNzZXJ0UmVzdWx0cyhGQUlMVVJFLCBb IjEgSlNDIHRlc3RzIGZhaWxlZCJdLCAxLCAiIiJSZXN1bHRzIGZvciBNb3ppbGxhIHRlc3RzOgog ICAgIDEgcmVncmVzc2lvbiBmb3VuZC4KICAgICAwIHRlc3RzIGZpeGVkLiIiIikKIAogICAgIGRl ZiB0ZXN0X21vemlsbGFfZmFpbHVyZXNfb2xkX291dHB1dChzZWxmKToKLSAgICAgICAgc2VsZi5h c3NlcnRSZXN1bHRzKEZBSUxVUkUsIFsianNjb3JlLXRlc3QiLCAnMiBmYWlsaW5nIE1vemlsbGEg dGVzdHMgJ10sIDEsICIiIlJlc3VsdHMgZm9yIE1vemlsbGEgdGVzdHM6CisgICAgICAgIHNlbGYu YXNzZXJ0UmVzdWx0cyhGQUlMVVJFLCBbIjIgSlNDIHRlc3RzIGZhaWxlZCJdLCAxLCAiIiJSZXN1 bHRzIGZvciBNb3ppbGxhIHRlc3RzOgogICAgIDIgcmVncmVzc2lvbnMgZm91bmQuCiAgICAgMCB0 ZXN0cyBmaXhlZC4iIiIpCiAKICAgICBkZWYgdGVzdF9qc2Nfc3RyZXNzX2ZhaWx1cmVfbmV3X291 dHB1dChzZWxmKToKLSAgICAgICAgc2VsZi5hc3NlcnRSZXN1bHRzKEZBSUxVUkUsIFsianNjb3Jl LXRlc3QiLCAnMSBmYWlsaW5nIEpTQyBzdHJlc3MgdGVzdCAnXSwgMSwgICIiIlJlc3VsdHMgZm9y IEpTQyBzdHJlc3MgdGVzdHM6CisgICAgICAgIHNlbGYuYXNzZXJ0UmVzdWx0cyhGQUlMVVJFLCBb IjEgSlNDIHRlc3RzIGZhaWxlZCJdLCAxLCAgIiIiUmVzdWx0cyBmb3IgSlNDIHN0cmVzcyB0ZXN0 czoKICAgICAxIGZhaWx1cmUgZm91bmQuIiIiKQogCiAgICAgZGVmIHRlc3RfanNjX3N0cmVzc19m YWlsdXJlc19uZXdfb3V0cHV0KHNlbGYpOgotICAgICAgICBzZWxmLmFzc2VydFJlc3VsdHMoRkFJ TFVSRSwgWyJqc2NvcmUtdGVzdCIsICc1IGZhaWxpbmcgSlNDIHN0cmVzcyB0ZXN0cyAnXSwgMSwg ICIiIlJlc3VsdHMgZm9yIEpTQyBzdHJlc3MgdGVzdHM6CisgICAgICAgIHNlbGYuYXNzZXJ0UmVz dWx0cyhGQUlMVVJFLCBbIjUgSlNDIHRlc3RzIGZhaWxlZCJdLCAxLCAgIiIiUmVzdWx0cyBmb3Ig SlNDIHN0cmVzcyB0ZXN0czoKICAgICA1IGZhaWx1cmVzIGZvdW5kLiIiIikKIAogCg== 245341 2015-01-26 04:36:40 -0800 2015-01-29 05:32:25 -0800 Patch bug-140821-20150126133837.patch text/plain 6290 ossy U3VidmVyc2lvbiBSZXZpc2lvbjogMTc5MTExCmRpZmYgLS1naXQgYS9Ub29scy9DaGFuZ2VMb2cg Yi9Ub29scy9DaGFuZ2VMb2cKaW5kZXggYzlhYzY1N2JmMjY2YzkxMTNiYTkzOGU2Nzk0M2MwNDI0 Njg2NTVjYS4uNWM4ZThkY2VjOGMwZDliZjk5NzUyNGY4NjQyZmNkZTk0YWZhY2UzYyAxMDA2NDQK LS0tIGEvVG9vbHMvQ2hhbmdlTG9nCisrKyBiL1Rvb2xzL0NoYW5nZUxvZwpAQCAtMSwzICsxLDIz IEBACisyMDE1LTAxLTI2ICBDc2FiYSBPc3p0cm9nb27DoWMgIDxvc3N5QHdlYmtpdC5vcmc+CisK KyAgICAgICAgW2J1aWxkYm90XSBTaW1wbGlmeSBqc2NvcmUtdGVzdCBidWlsZHN0ZXAKKyAgICAg ICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE0MDgyMQorCisgICAg ICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgICogQnVpbGRTbGF2ZVN1 cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25maWcvbWFzdGVyLmNmZzoKKyAgICAgICAgKFJ1bkph dmFTY3JpcHRDb3JlVGVzdHMpOiBJbmhlcml0ZWQgZnJvbSBUZXN0V2l0aEZhaWx1cmVDb3VudCBh bmQgcmVtb3ZlZCB1c2VsZXNzIGFjdHVhbC5odG1sIGxvZ2ZpbGUuCisgICAgICAgIChSdW5KYXZh U2NyaXB0Q29yZVRlc3RzLmNvdW50RmFpbHVyZXMpOiBBZGRlZC4KKyAgICAgICAgKFJ1bkphdmFT Y3JpcHRDb3JlVGVzdHMuY29tbWFuZENvbXBsZXRlKTogRGVsZXRlZC4KKyAgICAgICAgKFJ1bkph dmFTY3JpcHRDb3JlVGVzdHMuZXZhbHVhdGVDb21tYW5kKTogRGVsZXRlZC4KKyAgICAgICAgKFJ1 bkphdmFTY3JpcHRDb3JlVGVzdHMuZ2V0VGV4dCk6IERlbGV0ZWQuCisgICAgICAgIChSdW5KYXZh U2NyaXB0Q29yZVRlc3RzLmdldFRleHQyKTogRGVsZXRlZC4KKyAgICAgICAgKiBCdWlsZFNsYXZl U3VwcG9ydC9idWlsZC53ZWJraXQub3JnLWNvbmZpZy9tYXN0ZXJjZmdfdW5pdHRlc3QucHk6IFVw ZGF0ZWQuCisgICAgICAgIChSdW5KYXZhU2NyaXB0Q29yZVRlc3RzVGVzdC50ZXN0X21vemlsbGFf ZmFpbHVyZV9vbGRfb3V0cHV0KToKKyAgICAgICAgKFJ1bkphdmFTY3JpcHRDb3JlVGVzdHNUZXN0 LnRlc3RfbW96aWxsYV9mYWlsdXJlc19vbGRfb3V0cHV0KToKKyAgICAgICAgKFJ1bkphdmFTY3Jp cHRDb3JlVGVzdHNUZXN0LnRlc3RfanNjX3N0cmVzc19mYWlsdXJlX25ld19vdXRwdXQpOgorICAg ICAgICAoUnVuSmF2YVNjcmlwdENvcmVUZXN0c1Rlc3QudGVzdF9qc2Nfc3RyZXNzX2ZhaWx1cmVz X25ld19vdXRwdXQpOgorCiAyMDE1LTAxLTI2ICBNaWNoYWVsIENhdGFuemFybyAgPG1jYXRhbnph cm9AaWdhbGlhLmNvbT4gYW5kIENhcmxvcyBHYXJjaWEgQ2FtcG9zICA8Y2dhcmNpYUBpZ2FsaWEu Y29tPgogCiAgICAgICAgIFtHVEtdIGd0a2RvYyBkb2VzIG5vdCBhcHBlYXIgaW4gRGV2SGVscApk aWZmIC0tZ2l0IGEvVG9vbHMvQnVpbGRTbGF2ZVN1cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25m aWcvbWFzdGVyLmNmZyBiL1Rvb2xzL0J1aWxkU2xhdmVTdXBwb3J0L2J1aWxkLndlYmtpdC5vcmct Y29uZmlnL21hc3Rlci5jZmcKaW5kZXggM2NmZWQ4NTM2YzI5NzQ3OTJjZmYzZmQyNjM5MDYzNzkz NmJiZDc1Ny4uNmJkNmM2NjZlYTk3ZWQ2MDVkMmVlN2Q1Mzc2YTYzNDkzMjQwZGNhZCAxMDA2NDQK LS0tIGEvVG9vbHMvQnVpbGRTbGF2ZVN1cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25maWcvbWFz dGVyLmNmZworKysgYi9Ub29scy9CdWlsZFNsYXZlU3VwcG9ydC9idWlsZC53ZWJraXQub3JnLWNv bmZpZy9tYXN0ZXIuY2ZnCkBAIC0yNzcsMTIgKzI3NywxMiBAQCBjbGFzcyBEb3dubG9hZEJ1aWx0 UHJvZHVjdChzaGVsbC5TaGVsbENvbW1hbmQpOgogICAgIGZsdW5rT25GYWlsdXJlID0gVHJ1ZQog CiAKLWNsYXNzIFJ1bkphdmFTY3JpcHRDb3JlVGVzdHMoc2hlbGwuVGVzdCk6CitjbGFzcyBSdW5K YXZhU2NyaXB0Q29yZVRlc3RzKFRlc3RXaXRoRmFpbHVyZUNvdW50KToKICAgICBuYW1lID0gImpz Y29yZS10ZXN0IgogICAgIGRlc2NyaXB0aW9uID0gWyJqc2NvcmUtdGVzdHMgcnVubmluZyJdCiAg ICAgZGVzY3JpcHRpb25Eb25lID0gWyJqc2NvcmUtdGVzdHMiXQogICAgIGNvbW1hbmQgPSBbInBl cmwiLCAiLi9Ub29scy9TY3JpcHRzL3J1bi1qYXZhc2NyaXB0Y29yZS10ZXN0cyIsIFdpdGhQcm9w ZXJ0aWVzKCItLSUoY29uZmlndXJhdGlvbilzIildCi0gICAgbG9nZmlsZXMgPSB7J2FjdHVhbC5o dG1sIChzb3VyY2UpJzogJ1NvdXJjZS9KYXZhU2NyaXB0Q29yZS90ZXN0cy9tb3ppbGxhL2FjdHVh bC5odG1sJ30KKyAgICBmYWlsZWRUZXN0c0Zvcm1hdFN0cmluZyA9ICIlZCBKU0MgdGVzdCVzIGZh aWxlZCIKIAogICAgIGRlZiBfX2luaXRfXyhzZWxmLCBidWlsZEpTQ1Rvb2w9VHJ1ZSwgKmFyZ3Ms ICoqa3dhcmdzKToKICAgICAgICAgc2VsZi5idWlsZEpTQ1Rvb2wgPSBidWlsZEpTQ1Rvb2wKQEAg LTI5NSw1MSArMjk1LDE4IEBAIGNsYXNzIFJ1bkphdmFTY3JpcHRDb3JlVGVzdHMoc2hlbGwuVGVz dCk6CiAgICAgICAgICAgICBzZWxmLnNldENvbW1hbmQoc2VsZi5jb21tYW5kICsgWyctLW5vLWJ1 aWxkJ10pCiAgICAgICAgIHJldHVybiBzaGVsbC5UZXN0LnN0YXJ0KHNlbGYpCiAKLSAgICBkZWYg Y29tbWFuZENvbXBsZXRlKHNlbGYsIGNtZCk6Ci0gICAgICAgIHNoZWxsLlRlc3QuY29tbWFuZENv bXBsZXRlKHNlbGYsIGNtZCkKKyAgICBkZWYgY291bnRGYWlsdXJlcyhzZWxmLCBjbWQpOgogICAg ICAgICBsb2dUZXh0ID0gY21kLmxvZ3NbJ3N0ZGlvJ10uZ2V0VGV4dCgpCiAKLSAgICAgICAgZXhw cmVzc2lvbnMgPSBbCi0gICAgICAgICAgICAoJ2ZhaWxpbmcgTW96aWxsYSB0ZXN0JywgcmUuY29t cGlsZShyJ15SZXN1bHRzIGZvciBNb3ppbGxhIHRlc3RzOlxyP1xuXHMrKFxkKykgcmVncmVzc2lv bicsIHJlLk1VTFRJTElORSkpLAotICAgICAgICAgICAgKCdmYWlsaW5nIEpTQyBzdHJlc3MgdGVz dCcsIHJlLmNvbXBpbGUocideUmVzdWx0cyBmb3IgSlNDIHN0cmVzcyB0ZXN0czpccj9cblxzKyhc ZCspIGZhaWx1cmUnLCByZS5NVUxUSUxJTkUpKSwKLSAgICAgICAgXQotICAgICAgICByZXN1bHRM aW5lcyA9IHt9Ci0KLSAgICAgICAgZm9yIG5hbWUsIGV4cHJlc3Npb24gaW4gZXhwcmVzc2lvbnM6 Ci0gICAgICAgICAgICBtYXRjaCA9IGV4cHJlc3Npb24uc2VhcmNoKGxvZ1RleHQpCi0gICAgICAg ICAgICByZXN1bHRMaW5lc1tuYW1lXSA9IGludChtYXRjaC5ncm91cCgxKSkgaWYgbWF0Y2ggZWxz ZSAwCi0KLSAgICAgICAgc2VsZi5yZWdyZXNzaW9uTGluZSA9IFtdCi0gICAgICAgIGZvciBuYW1l IGluIHJlc3VsdExpbmVzOgotICAgICAgICAgICAgZmFpbHVyZXMgPSByZXN1bHRMaW5lc1tuYW1l XQotICAgICAgICAgICAgaWYgbm90IGZhaWx1cmVzOgotICAgICAgICAgICAgICAgIGNvbnRpbnVl Ci0KLSAgICAgICAgICAgIGlmIGZhaWx1cmVzID09IDE6Ci0gICAgICAgICAgICAgICAgcGx1cmFs U3VmZml4ID0gJycKLSAgICAgICAgICAgIGVsc2U6Ci0gICAgICAgICAgICAgICAgcGx1cmFsU3Vm Zml4ID0gJ3MnCi0gICAgICAgICAgICBzZWxmLnJlZ3Jlc3Npb25MaW5lLmFwcGVuZCgiJWQgJXMl cyAiICUgKGZhaWx1cmVzLCBuYW1lLCBwbHVyYWxTdWZmaXgpKQotCi0gICAgICAgIGlmICdhY3R1 YWwuaHRtbCAoc291cmNlKScgaW4gY21kLmxvZ3M6Ci0gICAgICAgICAgICBzZWxmLmFkZEhUTUxM b2coJ2FjdHVhbC5odG1sJywgY21kLmxvZ3NbJ2FjdHVhbC5odG1sIChzb3VyY2UpJ10uZ2V0VGV4 dCgpKQotCi0gICAgZGVmIGV2YWx1YXRlQ29tbWFuZChzZWxmLCBjbWQpOgotICAgICAgICBpZiBz ZWxmLnJlZ3Jlc3Npb25MaW5lOgotICAgICAgICAgICAgcmV0dXJuIEZBSUxVUkUKLQotICAgICAg ICBpZiBjbWQucmMgIT0gMDoKLSAgICAgICAgICAgIHJldHVybiBGQUlMVVJFCi0KLSAgICAgICAg cmV0dXJuIFNVQ0NFU1MKKyAgICAgICAgbWF0Y2ggPSByZS5zZWFyY2gocideUmVzdWx0cyBmb3Ig SlNDIHN0cmVzcyB0ZXN0czpccj9cblxzKyhcZCspIGZhaWx1cmUnLCBsb2dUZXh0LCByZS5NVUxU SUxJTkUpCisgICAgICAgIGlmIG1hdGNoOgorICAgICAgICAgICAgcmV0dXJuIGludChtYXRjaC5n cm91cCgxKSkKIAotICAgIGRlZiBnZXRUZXh0KHNlbGYsIGNtZCwgcmVzdWx0cyk6Ci0gICAgICAg IHJldHVybiBzZWxmLmdldFRleHQyKGNtZCwgcmVzdWx0cykKKyAgICAgICAgbWF0Y2ggPSByZS5z ZWFyY2gocideUmVzdWx0cyBmb3IgTW96aWxsYSB0ZXN0czpccj9cblxzKyhcZCspIHJlZ3Jlc3Np b24nLCBsb2dUZXh0LCByZS5NVUxUSUxJTkUpCisgICAgICAgIGlmIG1hdGNoOgorICAgICAgICAg ICAgcmV0dXJuIGludChtYXRjaC5ncm91cCgxKSkKIAotICAgIGRlZiBnZXRUZXh0MihzZWxmLCBj bWQsIHJlc3VsdHMpOgotICAgICAgICByZXN1bHQgPSBbc2VsZi5uYW1lXQotICAgICAgICByZXN1 bHQuZXh0ZW5kKHNlbGYucmVncmVzc2lvbkxpbmUpCi0gICAgICAgIHJldHVybiByZXN1bHQKKyAg ICAgICAgcmV0dXJuIDAKIAogY2xhc3MgUnVuV2ViS2l0VGVzdHMoc2hlbGwuVGVzdCk6CiAgICAg bmFtZSA9ICJsYXlvdXQtdGVzdCIKZGlmZiAtLWdpdCBhL1Rvb2xzL0J1aWxkU2xhdmVTdXBwb3J0 L2J1aWxkLndlYmtpdC5vcmctY29uZmlnL21hc3RlcmNmZ191bml0dGVzdC5weSBiL1Rvb2xzL0J1 aWxkU2xhdmVTdXBwb3J0L2J1aWxkLndlYmtpdC5vcmctY29uZmlnL21hc3RlcmNmZ191bml0dGVz dC5weQppbmRleCAwNDFhMDZhNjAzZjM2ZjE3MTI2YjA2NjBkNjA3N2IxZWVjNmQyNmU0Li5lNTFk NTI4Yzk0ODVjNzEwZDM3YzU4M2Y4NTg3NzM5Y2E2N2I3ZTEwIDEwMDc1NQotLS0gYS9Ub29scy9C dWlsZFNsYXZlU3VwcG9ydC9idWlsZC53ZWJraXQub3JnLWNvbmZpZy9tYXN0ZXJjZmdfdW5pdHRl c3QucHkKKysrIGIvVG9vbHMvQnVpbGRTbGF2ZVN1cHBvcnQvYnVpbGQud2Via2l0Lm9yZy1jb25m aWcvbWFzdGVyY2ZnX3VuaXR0ZXN0LnB5CkBAIC0xMjIsMjEgKzEyMiwyMSBAQCBjbGFzcyBSdW5K YXZhU2NyaXB0Q29yZVRlc3RzVGVzdCh1bml0dGVzdC5UZXN0Q2FzZSk6CiAgICAgT0suIiIiKQog CiAgICAgZGVmIHRlc3RfbW96aWxsYV9mYWlsdXJlX29sZF9vdXRwdXQoc2VsZik6Ci0gICAgICAg IHNlbGYuYXNzZXJ0UmVzdWx0cyhGQUlMVVJFLCBbImpzY29yZS10ZXN0IiwgJzEgZmFpbGluZyBN b3ppbGxhIHRlc3QgJ10sIDEsICIiIlJlc3VsdHMgZm9yIE1vemlsbGEgdGVzdHM6CisgICAgICAg IHNlbGYuYXNzZXJ0UmVzdWx0cyhGQUlMVVJFLCBbIjEgSlNDIHRlc3QgZmFpbGVkIl0sIDEsICIi IlJlc3VsdHMgZm9yIE1vemlsbGEgdGVzdHM6CiAgICAgMSByZWdyZXNzaW9uIGZvdW5kLgogICAg IDAgdGVzdHMgZml4ZWQuIiIiKQogCiAgICAgZGVmIHRlc3RfbW96aWxsYV9mYWlsdXJlc19vbGRf b3V0cHV0KHNlbGYpOgotICAgICAgICBzZWxmLmFzc2VydFJlc3VsdHMoRkFJTFVSRSwgWyJqc2Nv cmUtdGVzdCIsICcyIGZhaWxpbmcgTW96aWxsYSB0ZXN0cyAnXSwgMSwgIiIiUmVzdWx0cyBmb3Ig TW96aWxsYSB0ZXN0czoKKyAgICAgICAgc2VsZi5hc3NlcnRSZXN1bHRzKEZBSUxVUkUsIFsiMiBK U0MgdGVzdHMgZmFpbGVkIl0sIDEsICIiIlJlc3VsdHMgZm9yIE1vemlsbGEgdGVzdHM6CiAgICAg MiByZWdyZXNzaW9ucyBmb3VuZC4KICAgICAwIHRlc3RzIGZpeGVkLiIiIikKIAogICAgIGRlZiB0 ZXN0X2pzY19zdHJlc3NfZmFpbHVyZV9uZXdfb3V0cHV0KHNlbGYpOgotICAgICAgICBzZWxmLmFz c2VydFJlc3VsdHMoRkFJTFVSRSwgWyJqc2NvcmUtdGVzdCIsICcxIGZhaWxpbmcgSlNDIHN0cmVz cyB0ZXN0ICddLCAxLCAgIiIiUmVzdWx0cyBmb3IgSlNDIHN0cmVzcyB0ZXN0czoKKyAgICAgICAg c2VsZi5hc3NlcnRSZXN1bHRzKEZBSUxVUkUsIFsiMSBKU0MgdGVzdCBmYWlsZWQiXSwgMSwgICIi IlJlc3VsdHMgZm9yIEpTQyBzdHJlc3MgdGVzdHM6CiAgICAgMSBmYWlsdXJlIGZvdW5kLiIiIikK IAogICAgIGRlZiB0ZXN0X2pzY19zdHJlc3NfZmFpbHVyZXNfbmV3X291dHB1dChzZWxmKToKLSAg ICAgICAgc2VsZi5hc3NlcnRSZXN1bHRzKEZBSUxVUkUsIFsianNjb3JlLXRlc3QiLCAnNSBmYWls aW5nIEpTQyBzdHJlc3MgdGVzdHMgJ10sIDEsICAiIiJSZXN1bHRzIGZvciBKU0Mgc3RyZXNzIHRl c3RzOgorICAgICAgICBzZWxmLmFzc2VydFJlc3VsdHMoRkFJTFVSRSwgWyI1IEpTQyB0ZXN0cyBm YWlsZWQiXSwgMSwgICIiIlJlc3VsdHMgZm9yIEpTQyBzdHJlc3MgdGVzdHM6CiAgICAgNSBmYWls dXJlcyBmb3VuZC4iIiIpCiAKIAo=