19822 2008-06-29 23:39:34 -0700 REGRESSION (r30243): setting location.hash to "#" causes a reload 2008-07-18 22:48:25 -0700 1 1 1 Unclassified WebKit Page Loading 528+ (Nightly build) Mac OS X 10.5 RESOLVED FIXED http://today.ask.com/default HasReduction, InRadar, Regression P1 Normal --- 1 zwarich zwarich ap beidson bjorn.tipling darin ddkilzer jurkuipers khooyp oliver sam oldest_to_newest 84743 0 zwarich 2008-06-29 23:39:34 -0700 In Safari 3.1.1, http://today.ask.com/default loads fine, but in ToT (r34883) it keeps on reloading. This is because location.hash is set to "#" on each load. 84744 1 22004 zwarich 2008-06-29 23:40:12 -0700 Created attachment 22004 Reduction 84766 2 zwarich 2008-06-30 11:39:45 -0700 *** Bug 17627 has been marked as a duplicate of this bug. *** 84768 3 zwarich 2008-06-30 11:41:51 -0700 This seems related to bug 13067. 84777 4 ddkilzer 2008-06-30 12:46:47 -0700 <rdar://problem/6044293> 84778 5 ddkilzer 2008-06-30 12:47:39 -0700 It would be interesting to know when this regressed via WebKitTools/Scripts/bisect-builds. 84781 6 beidson 2008-06-30 13:18:30 -0700 I've narrowed this to between r30471 and r30584 84782 7 beidson 2008-06-30 13:23:26 -0700 nevermind, it broke earlier than that. sigh 84783 8 ddkilzer 2008-06-30 13:24:16 -0700 (In reply to comment #6) > I've narrowed this to between r30471 and r30584 Odd. The bisect-builds script with Safari DP 4 on Leopard 10.5.4 reports: Works: r30240 Fails: r30267 84784 9 zwarich 2008-06-30 13:33:50 -0700 (In reply to comment #8) > Odd. The bisect-builds script with Safari DP 4 on Leopard 10.5.4 reports: > > Works: r30240 Fails: r30267 The only suspicion revision is r30243: http://trac.webkit.org/changeset/30243 84785 10 zwarich 2008-06-30 13:34:29 -0700 (In reply to comment #9) > The only suspicion revision is r30243: Oops. That was a typo. I meant "suspicious" revision. 84786 11 beidson 2008-06-30 13:36:01 -0700 http://trac.webkit.org/changeset/30243 seems like a likely candidate 84787 12 beidson 2008-06-30 13:36:33 -0700 Dammit, people, if all 3 of us are working on this... I'm just gonna pull out and let you two hash it out :) 84790 13 beidson 2008-06-30 14:32:33 -0700 I confirmed, as suspected, that 30243 caused this regression. 84791 14 beidson 2008-06-30 14:33:47 -0700 Oh boy, it was a huge patch that fixed about a dozen different fixes. 84792 15 ddkilzer 2008-06-30 14:55:48 -0700 git-bisect confirms that the regression occurred in r30243. 84812 16 zwarich 2008-06-30 19:16:34 -0700 I'm assigning this to myself. I'll take a shot at this tonight, but I probably won't have time if it drags into tomorrow. I also apologize for the prior confusion about who was working on this. 84813 17 darin 2008-06-30 19:29:42 -0700 Sorry my patch broke this. Cameron, I'm glad you're going to fix this. Let me know if you need anything from me. 84819 18 zwarich 2008-06-30 21:28:10 -0700 Strangely enough, the problem is that "url.ref() == str" is false even though both strings are empty. The comparison is ultimately done in StringHash::equal() in platform/text/StringHash.h, so the bug is likely there. 84822 19 zwarich 2008-06-30 23:25:31 -0700 The problematic code in StringImpl.h:equal() is if (a == b) return true; if (!a || !b) return false; It is possible for a StringImpl* of an empty string to be non-null, so this code is incorrect. However, if I change it to take the length of a non-null StringImpl*, it passes the check in JSLocation::setHash(), but later on in loading it dies with this backtrace: #0 0x02b3dc3d in WebCore::StringImpl::length (this=0xffffffff) at text/StringImpl.h:86 #1 0x028dcd3b in WebCore::StringHash::equal (a=0xffffffff, b=0x0) at text/StringHash.h:42 #2 0x02e9f126 in WebCore::equal (a=0xffffffff, b=0x0) at /Users/Cameron/WebKit/WebCore/platform/text/StringImpl.cpp:879 #3 0x028dccdf in WebCore::operator== (a=@0x14d24968, b=@0xbfffe1cc) at text/PlatformString.h:224 #4 0x028f52b5 in WTF::HashTable<WebCore::String, WebCore::String, WTF::IdentityExtractor<WebCore::String>, WebCore::StringHash, WTF::HashTraits<WebCore::String>, WTF::HashTraits<WebCore::String> >::isEmptyBucket (value=@0x14d24968) at HashTable.h:327 I guess 0xffffffff is an empty or deleted value in the hash table. Should it be getting used here? What is the best way to fix this? 84823 20 22017 zwarich 2008-06-30 23:26:43 -0700 Created attachment 22017 Patch that should work Here is the patch that should work, but fails due to the problem mentioned in the previous post. 84826 21 ggaren 2008-06-30 23:58:32 -0700 It is a (correct) feature that the null string and the empty string do not compare equal. Either KURL::ref needs to change to return an empty string instead of a null string in the !hasRef() case, for the sake of compatibility with substring operations, which return the empty string when no such substring exists, or clients performing substring operations need to account for null vs empty. 84830 22 zwarich 2008-07-01 00:57:26 -0700 Geoff straightened me out on IRC. I guess I thought WebCore's Strings should work like UString. The problem is that KURL::ref() possibly returns a null string. The natural thing to do is to make it return the empty string, but we should check its callsites first. It has 7 callsites in WebCore: void JSLocation::setHash(ExecState* exec, JSValue* value) { ... if (url.ref() == str) return; ... } We want url.ref() here to be the empty string, not the null string, as that fixes the bug. String KURL::prettyURL() const { ... if (m_fragmentEnd != m_queryEnd) { result.append('#'); append(result, ref()); } ... } Clearly, null and empty are the same here. String HTMLAnchorElement::hash() const { return "#" + href().ref(); } It doesn't matter here, because the result is immediately being concatenated with another string. String Location::hash() const { if (!m_frame) return String(); const KURL& url = this->url(); return url.ref().isNull() ? "" : "#" + url.ref(); } We can simply change the isNull() here to an isEmpty(), and the result of the second call is immediately concatenated. Not changing it causes the following tests to fail, so there is already test coverage here: fast/dom/location-hash.html http/tests/security/local-user-CSS-from-remote.html http/tests/security/xss-DENIED-assign-location-hash.html AccessibilityObject* AccessibilityRenderObject::internalLinkElement() const { HTMLAnchorElement* anchor = anchorElement(); if (!anchor) return 0; KURL linkURL = anchor->href(); String ref = linkURL.ref(); if (ref.isEmpty()) return 0; // check if URL is the same as current URL linkURL.setRef(""); ... } It doesn't matter here, because String::isEmpty() is the same for null and empty strings. void FrameLoader::gotoAnchor() { // If our URL has no ref, then we have no place we need to jump to. // OTOH If CSS target was set previously, we want to set it to 0, recalc // and possibly repaint because :target pseudo class may have been // set (see bug 11321). if (!m_URL.hasRef() && !(m_frame->document() && m_frame->document()->getCSSTarget())) return; String ref = m_URL.ref(); if (gotoAnchor(ref)) return; // Try again after decoding the ref, based on the document's encoding. if (m_decoder) gotoAnchor(decodeURLEscapeSequences(ref, m_decoder->encoding())); } Calling gotoAnchor() with an empty ref (as opposed to a null ref) seems to be fine, as the test fast/css/target-fragment-match.html for bug 11321 still passes. 84833 23 22018 zwarich 2008-07-01 01:44:16 -0700 Created attachment 22018 Proposed patch Here is a patch that fixes the problem. My attempts to make a layout test have all failed, but I will keep on trying. If anyone has an idea about how to get DRT to keep on going with the scheduled location changes, please let me know. 84838 24 22020 zwarich 2008-07-01 03:00:26 -0700 Created attachment 22020 Proposed patch with test I found the http/tests/loading/ directory. 84840 25 22021 zwarich 2008-07-01 03:06:20 -0700 Created attachment 22021 Proposed patch with test I forgot to include a description of the test in the test output. 84867 26 ddkilzer 2008-07-01 07:15:42 -0700 Great analysis, Cameron! 84871 27 22021 beidson 2008-07-01 07:41:42 -0700 Comment on attachment 22021 Proposed patch with test r=me! 84877 28 22021 ap 2008-07-01 08:06:11 -0700 Comment on attachment 22021 Proposed patch with test I think that a better explanation of expected results in the test would be beneficial. It would be very hard to figure out which order of callbacks is right otherwise, and might even make someone land incorrect results on top of correct ones. 84878 29 22021 beidson 2008-07-01 08:11:53 -0700 Comment on attachment 22021 Proposed patch with test Changed my mind - Alexey raises a good point. Additionally, tests that rely on DRT's load delegate output would do well to point out that the test was designed for DRT and might not be effective in the browser. 84879 30 darin 2008-07-01 08:18:28 -0700 There's a logical that ref() returns null instead of empty when there is no ref. It allows the caller to distinguish the case of an empty ref "<url>#" from no ref at all "<url>". That seems like a good design to me. Does it need to change? Can we change the callers instead? 84893 31 zwarich 2008-07-01 11:15:45 -0700 (In reply to comment #30) > There's a logical that ref() returns null instead of empty when there is no > ref. It allows the caller to distinguish the case of an empty ref "<url>#" from > no ref at all "<url>". That seems like a good design to me. Does it need to > change? Can we change the callers instead? I agree with your point. It doesn't seem like it would ever a problem, because anyone that wants to know whether there is a fragment checks url.hasRef() instead of url.ref().isNull(), but Brady pointed out to me that conversions into platform-specific URL classes could mean that bugs get introduced elsewhere. I'll change the one caller to account for the case where they are both empty. I will also include a more detailed description of failure in the layout test, as well as pointing out that if it fails in the browser it will simply reload forever. 84898 32 22031 zwarich 2008-07-01 11:49:04 -0700 Created attachment 22031 Revised proposed patch Here is a new patch incorporating the comments posted in this bug. Here is the output when the test fails: main frame - didStartProvisionalLoadForFrame main frame - didCommitLoadForFrame main frame - willPerformClientRedirectToURL: http://127.0.0.1:8000/loading/location-hash-reload-cycle.html main frame - didCancelClientRedirectForFrame main frame - willPerformClientRedirectToURL: http://127.0.0.1:8000/loading/location-hash-reload-cycle.html main frame - didHandleOnloadEventsForFrame main frame - willPerformClientRedirectToURL: http://127.0.0.1:8000/loading/location-hash-reload-cycle.html main frame - didFinishDocumentLoadForFrame main frame - didFinishLoadForFrame This test checks that no loader actions occur when setting window.location.hash to the empty string or "#". If this test fails when run in a browser, it will reload continuously. If this test fails when run in DumpRenderTree, the FrameLoader callback output will contain willPerformClientRedirectToURL callbacks. Both of the window.location.hashes change the output independently of each other. When run in a browser, it is very noticeable that it is reloading forever. 84901 33 22031 beidson 2008-07-01 11:59:02 -0700 Comment on attachment 22031 Revised proposed patch r=me 84902 34 zwarich 2008-07-01 12:06:41 -0700 Landed in r34927. 84910 35 22031 darin 2008-07-01 15:06:51 -0700 Comment on attachment 22031 Revised proposed patch Cameron, I know you meant this: if (oldRef == str || (oldRef.isNull() && str.isEmpty())) You forgot to use oldRef in the old part of the expression. 84911 36 zwarich 2008-07-01 15:09:25 -0700 (In reply to comment #35) > (From update of attachment 22031 [edit]) > Cameron, I know you meant this: > > if (oldRef == str || (oldRef.isNull() && str.isEmpty())) > > You forgot to use oldRef in the old part of the expression. How did Brady and I miss that? r=you to fix it? 84919 37 darin 2008-07-01 16:05:08 -0700 (In reply to comment #36) > How did Brady and I miss that? r=you to fix it? As you probably already noticed, I fixed it. 86406 38 zwarich 2008-07-18 22:48:25 -0700 *** Bug 17832 has been marked as a duplicate of this bug. *** 22004 2008-06-29 23:40:12 -0700 2008-06-29 23:40:12 -0700 Reduction test.html text/html 391 zwarich PCFET0NUWVBFIEhUTUwgUFVCTElDICItLy9XM0MvL0RURCBIVE1MIDQuMDEvL0VOIgogICAgICAg ICAgICAiaHR0cDovL3d3dy53My5vcmcvVFIvaHRtbDQvc3RyaWN0LmR0ZCI+CjxodG1sPgo8aGVh ZD4KICAgICAgICA8dGl0bGU+VGVzdDwvdGl0bGU+CiAgICAgICAgPHNjcmlwdCB0eXBlPSJ0ZXh0 L2phdmFzY3JpcHQiPgogICAgICAgIHdpbmRvdy5tYWluID0gZnVuY3Rpb24oKXsKICAgICAgICAg ICAgICAgIGxvY2F0aW9uLmhhc2ggPSAiIyI7CiAgICAgICAgfQogICAgICAgIDwvc2NyaXB0Pgo8 L2hlYWQ+Cjxib2R5IG9ubG9hZD0ibWFpbigpOyI+CiAgICAgICAgPGgxPlRlc3Q8L2gxPgogICAg ICAgIDxwPlRlc3RpbmcgbG9jYXRpb24uaGFzaDwvcD4KPC9ib2R5Pgo8L2h0bWw+Cg== 22017 2008-06-30 23:26:43 -0700 2008-07-01 01:44:16 -0700 Patch that should work StringHash.diff text/plain 885 zwarich SW5kZXg6IHBsYXRmb3JtL3RleHQvU3RyaW5nSGFzaC5oCj09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIHBsYXRmb3Jt L3RleHQvU3RyaW5nSGFzaC5oCShyZXZpc2lvbiAzNDkwMikKKysrIHBsYXRmb3JtL3RleHQvU3Ry aW5nSGFzaC5oCSh3b3JraW5nIGNvcHkpCkBAIC0zOCwxMyArMzgsMTMgQEAgbmFtZXNwYWNlIFdl YkNvcmUgewogICAgICAgICB7CiAgICAgICAgICAgICBpZiAoYSA9PSBiKQogICAgICAgICAgICAg ICAgIHJldHVybiB0cnVlOwotICAgICAgICAgICAgaWYgKCFhIHx8ICFiKQotICAgICAgICAgICAg ICAgIHJldHVybiBmYWxzZTsKIAotICAgICAgICAgICAgdW5zaWduZWQgYUxlbmd0aCA9IGEtPmxl bmd0aCgpOwotICAgICAgICAgICAgdW5zaWduZWQgYkxlbmd0aCA9IGItPmxlbmd0aCgpOworICAg ICAgICAgICAgdW5zaWduZWQgYUxlbmd0aCA9IGEgPyBhLT5sZW5ndGgoKSA6IDA7CisgICAgICAg ICAgICB1bnNpZ25lZCBiTGVuZ3RoID0gYiA/IGItPmxlbmd0aCgpIDogMDsKICAgICAgICAgICAg IGlmIChhTGVuZ3RoICE9IGJMZW5ndGgpCiAgICAgICAgICAgICAgICAgcmV0dXJuIGZhbHNlOwor ICAgICAgICAgICAgaWYgKGFMZW5ndGggPT0gMCAmJiBiTGVuZ3RoID09IDApCisgICAgICAgICAg ICAgICAgcmV0dXJuIHRydWU7CiAKICAgICAgICAgICAgIGNvbnN0IHVpbnQzMl90KiBhQ2hhcnMg PSByZWludGVycHJldF9jYXN0PGNvbnN0IHVpbnQzMl90Kj4oYS0+Y2hhcmFjdGVycygpKTsKICAg ICAgICAgICAgIGNvbnN0IHVpbnQzMl90KiBiQ2hhcnMgPSByZWludGVycHJldF9jYXN0PGNvbnN0 IHVpbnQzMl90Kj4oYi0+Y2hhcmFjdGVycygpKTsK 22018 2008-07-01 01:44:16 -0700 2008-07-01 03:00:26 -0700 Proposed patch kurl.diff text/plain 1607 zwarich SW5kZXg6IENoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBDaGFuZ2VMb2cJKHJldmlzaW9uIDM0OTA5 KQorKysgQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBAIC0xLDMgKzEsMTggQEAKKzIwMDgtMDct MDEgIENhbWVyb24gWndhcmljaCAgPGN3endhcmljaEB1d2F0ZXJsb28uY2E+CisKKyAgICAgICAg UmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgQnVnIDE5ODIyOiBSRUdSRVNT SU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9uLmhhc2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9h ZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE5ODIy PgorCisgICAgICAgIENoYW5nZSBLVVJMOjpyZWYoKSBzbyBpdCByZXR1cm5zIGFuIGVtcHR5IHN0 cmluZyBpbnN0ZWFkIG9mIGEgbnVsbAorICAgICAgICBzdHJpbmcgd2hlbiB0aGVyZSBpcyBubyBm cmFnbWVudC4KKworICAgICAgICAqIHBhZ2UvTG9jYXRpb24uY3BwOgorICAgICAgICAoV2ViQ29y ZTo6TG9jYXRpb246Omhhc2gpOgorICAgICAgICAqIHBsYXRmb3JtL0tVUkwuY3BwOgorICAgICAg ICAoV2ViQ29yZTo6S1VSTDo6cmVmKToKKwogMjAwOC0wNy0wMSAgU2ltb24gSGF1c21hbm4gIDxo YXVzbWFubkB3ZWJraXQub3JnPgogCiAgICAgICAgIEJ1aWxkIGZpeCwgaW5jbHVkZSBEYXRlSW5z dGFuY2UuaC4KSW5kZXg6IHBhZ2UvTG9jYXRpb24uY3BwCj09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIHBhZ2UvTG9j YXRpb24uY3BwCShyZXZpc2lvbiAzNDkwMikKKysrIHBhZ2UvTG9jYXRpb24uY3BwCSh3b3JraW5n IGNvcHkpCkBAIC0xMjAsNyArMTIwLDcgQEAgU3RyaW5nIExvY2F0aW9uOjpoYXNoKCkgY29uc3QK ICAgICAgICAgcmV0dXJuIFN0cmluZygpOwogCiAgICAgY29uc3QgS1VSTCYgdXJsID0gdGhpcy0+ dXJsKCk7Ci0gICAgcmV0dXJuIHVybC5yZWYoKS5pc051bGwoKSA/ICIiIDogIiMiICsgdXJsLnJl ZigpOworICAgIHJldHVybiB1cmwucmVmKCkuaXNFbXB0eSgpID8gIiIgOiAiIyIgKyB1cmwucmVm KCk7CiB9CiAKIFN0cmluZyBMb2NhdGlvbjo6dG9TdHJpbmcoKSBjb25zdApJbmRleDogcGxhdGZv cm0vS1VSTC5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PQotLS0gcGxhdGZvcm0vS1VSTC5jcHAJKHJldmlzaW9uIDM0 OTAyKQorKysgcGxhdGZvcm0vS1VSTC5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTU3Miw3ICs1NzIs NyBAQCBTdHJpbmcgS1VSTDo6dXNlcigpIGNvbnN0CiBTdHJpbmcgS1VSTDo6cmVmKCkgY29uc3QK IHsKICAgICBpZiAobV9mcmFnbWVudEVuZCA9PSBtX3F1ZXJ5RW5kKQotICAgICAgICByZXR1cm4g U3RyaW5nKCk7CisgICAgICAgIHJldHVybiBTdHJpbmcoIiIpOwogCiAgICAgcmV0dXJuIG1fc3Ry aW5nLnN1YnN0cmluZyhtX3F1ZXJ5RW5kICsgMSwgbV9mcmFnbWVudEVuZCAtIChtX3F1ZXJ5RW5k ICsgMSkpOwogfQo= 22020 2008-07-01 03:00:26 -0700 2008-07-01 03:06:20 -0700 Proposed patch with test kurl.diff text/plain 3502 zwarich SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiAzNDkwOSkKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTggQEAKKzIwMDgtMDctMDEgIENhbWVyb24gWndhcmljaCAgPGN3endhcmljaEB1 d2F0ZXJsb28uY2E+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAg ICAgICAgQnVnIDE5ODIyOiBSRUdSRVNTSU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9uLmhh c2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9hZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtpdC5v cmcvc2hvd19idWcuY2dpP2lkPTE5ODIyPgorCisgICAgICAgIENoYW5nZSBLVVJMOjpyZWYoKSBz byBpdCByZXR1cm5zIGFuIGVtcHR5IHN0cmluZyBpbnN0ZWFkIG9mIGEgbnVsbAorICAgICAgICBz dHJpbmcgd2hlbiB0aGVyZSBpcyBubyBmcmFnbWVudC4KKworICAgICAgICAqIHBhZ2UvTG9jYXRp b24uY3BwOgorICAgICAgICAoV2ViQ29yZTo6TG9jYXRpb246Omhhc2gpOgorICAgICAgICAqIHBs YXRmb3JtL0tVUkwuY3BwOgorICAgICAgICAoV2ViQ29yZTo6S1VSTDo6cmVmKToKKwogMjAwOC0w Ny0wMSAgU2ltb24gSGF1c21hbm4gIDxoYXVzbWFubkB3ZWJraXQub3JnPgogCiAgICAgICAgIEJ1 aWxkIGZpeCwgaW5jbHVkZSBEYXRlSW5zdGFuY2UuaC4KSW5kZXg6IFdlYkNvcmUvcGFnZS9Mb2Nh dGlvbi5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PQotLS0gV2ViQ29yZS9wYWdlL0xvY2F0aW9uLmNwcAkocmV2aXNp b24gMzQ5MDIpCisrKyBXZWJDb3JlL3BhZ2UvTG9jYXRpb24uY3BwCSh3b3JraW5nIGNvcHkpCkBA IC0xMjAsNyArMTIwLDcgQEAgU3RyaW5nIExvY2F0aW9uOjpoYXNoKCkgY29uc3QKICAgICAgICAg cmV0dXJuIFN0cmluZygpOwogCiAgICAgY29uc3QgS1VSTCYgdXJsID0gdGhpcy0+dXJsKCk7Ci0g ICAgcmV0dXJuIHVybC5yZWYoKS5pc051bGwoKSA/ICIiIDogIiMiICsgdXJsLnJlZigpOworICAg IHJldHVybiB1cmwucmVmKCkuaXNFbXB0eSgpID8gIiIgOiAiIyIgKyB1cmwucmVmKCk7CiB9CiAK IFN0cmluZyBMb2NhdGlvbjo6dG9TdHJpbmcoKSBjb25zdApJbmRleDogV2ViQ29yZS9wbGF0Zm9y bS9LVVJMLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09Ci0tLSBXZWJDb3JlL3BsYXRmb3JtL0tVUkwuY3BwCShyZXZp c2lvbiAzNDkwMikKKysrIFdlYkNvcmUvcGxhdGZvcm0vS1VSTC5jcHAJKHdvcmtpbmcgY29weSkK QEAgLTU3Miw3ICs1NzIsNyBAQCBTdHJpbmcgS1VSTDo6dXNlcigpIGNvbnN0CiBTdHJpbmcgS1VS TDo6cmVmKCkgY29uc3QKIHsKICAgICBpZiAobV9mcmFnbWVudEVuZCA9PSBtX3F1ZXJ5RW5kKQot ICAgICAgICByZXR1cm4gU3RyaW5nKCk7CisgICAgICAgIHJldHVybiBTdHJpbmcoIiIpOwogCiAg ICAgcmV0dXJuIG1fc3RyaW5nLnN1YnN0cmluZyhtX3F1ZXJ5RW5kICsgMSwgbV9mcmFnbWVudEVu ZCAtIChtX3F1ZXJ5RW5kICsgMSkpOwogfQpJbmRleDogTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCj09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT0KLS0tIExheW91dFRlc3RzL0NoYW5nZUxvZwkocmV2aXNpb24gMzQ5MTEpCisrKyBM YXlvdXRUZXN0cy9DaGFuZ2VMb2cJKHdvcmtpbmcgY29weSkKQEAgLTEsMyArMSwxNSBAQAorMjAw OC0wNy0wMSAgQ2FtZXJvbiBad2FyaWNoICA8Y3d6d2FyaWNoQHV3YXRlcmxvby5jYT4KKworICAg ICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBUZXN0cyBmb3I6CisK KyAgICAgICAgQnVnIDE5ODIyOiBSRUdSRVNTSU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9u Lmhhc2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9hZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtp dC5vcmcvc2hvd19idWcuY2dpP2lkPTE5ODIyPgorCisgICAgICAgICogaHR0cC90ZXN0cy9sb2Fk aW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLWV4cGVjdGVkLnR4dDogQWRkZWQuCisgICAg ICAgICogaHR0cC90ZXN0cy9sb2FkaW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLmh0bWw6 IEFkZGVkLgorCiAyMDA4LTA2LTMwICBBZGVsZSBQZXRlcnNvbiAgPGFkZWxlQGFwcGxlLmNvbT4K IAogICAgICAgICBSZXZpZXdlZCBieSBPbGl2ZXIuCkluZGV4OiBMYXlvdXRUZXN0cy9odHRwL3Rl c3RzL2xvYWRpbmcvbG9jYXRpb24taGFzaC1yZWxvYWQtY3ljbGUtZXhwZWN0ZWQudHh0Cj09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT0KLS0tIExheW91dFRlc3RzL2h0dHAvdGVzdHMvbG9hZGluZy9sb2NhdGlvbi1oYXNoLXJl bG9hZC1jeWNsZS1leHBlY3RlZC50eHQJKHJldmlzaW9uIDApCisrKyBMYXlvdXRUZXN0cy9odHRw L3Rlc3RzL2xvYWRpbmcvbG9jYXRpb24taGFzaC1yZWxvYWQtY3ljbGUtZXhwZWN0ZWQudHh0CShy ZXZpc2lvbiAwKQpAQCAtMCwwICsxLDYgQEAKK21haW4gZnJhbWUgLSBkaWRTdGFydFByb3Zpc2lv bmFsTG9hZEZvckZyYW1lCittYWluIGZyYW1lIC0gZGlkQ29tbWl0TG9hZEZvckZyYW1lCittYWlu IGZyYW1lIC0gZGlkSGFuZGxlT25sb2FkRXZlbnRzRm9yRnJhbWUKK21haW4gZnJhbWUgLSBkaWRG aW5pc2hEb2N1bWVudExvYWRGb3JGcmFtZQorbWFpbiBmcmFtZSAtIGRpZEZpbmlzaExvYWRGb3JG cmFtZQorCkluZGV4OiBMYXlvdXRUZXN0cy9odHRwL3Rlc3RzL2xvYWRpbmcvbG9jYXRpb24taGFz aC1yZWxvYWQtY3ljbGUuaHRtbAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBMYXlvdXRUZXN0cy9odHRwL3Rlc3Rz L2xvYWRpbmcvbG9jYXRpb24taGFzaC1yZWxvYWQtY3ljbGUuaHRtbAkocmV2aXNpb24gMCkKKysr IExheW91dFRlc3RzL2h0dHAvdGVzdHMvbG9hZGluZy9sb2NhdGlvbi1oYXNoLXJlbG9hZC1jeWNs ZS5odG1sCShyZXZpc2lvbiAwKQpAQCAtMCwwICsxLDE2IEBACis8aHRtbD4KKzxoZWFkPgorPHNj cmlwdD4KK2Z1bmN0aW9uIHJ1blRlc3RzKCkKK3sKKyAgICBpZiAod2luZG93LmxheW91dFRlc3RD b250cm9sbGVyKQorICAgICAgICBsYXlvdXRUZXN0Q29udHJvbGxlci5kdW1wQXNUZXh0KCk7Cisg ICAgCisgICAgd2luZG93LmxvY2F0aW9uLmhhc2ggPSAiIjsKKyAgICB3aW5kb3cubG9jYXRpb24u aGFzaCA9ICIjIjsKK30KKzwvc2NyaXB0PgorPC9oZWFkPgorPGJvZHkgb25sb2FkPSJydW5UZXN0 cygpOyI+Cis8L2JvZHk+Cis8L2h0bWw+Cg== 22021 2008-07-01 03:06:20 -0700 2008-07-01 11:49:04 -0700 Proposed patch with test kurl.diff text/plain 3725 zwarich SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiAzNDkwOSkKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTggQEAKKzIwMDgtMDctMDEgIENhbWVyb24gWndhcmljaCAgPGN3endhcmljaEB1 d2F0ZXJsb28uY2E+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAg ICAgICAgQnVnIDE5ODIyOiBSRUdSRVNTSU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9uLmhh c2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9hZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtpdC5v cmcvc2hvd19idWcuY2dpP2lkPTE5ODIyPgorCisgICAgICAgIENoYW5nZSBLVVJMOjpyZWYoKSBz byBpdCByZXR1cm5zIGFuIGVtcHR5IHN0cmluZyBpbnN0ZWFkIG9mIGEgbnVsbAorICAgICAgICBz dHJpbmcgd2hlbiB0aGVyZSBpcyBubyBmcmFnbWVudC4KKworICAgICAgICAqIHBhZ2UvTG9jYXRp b24uY3BwOgorICAgICAgICAoV2ViQ29yZTo6TG9jYXRpb246Omhhc2gpOgorICAgICAgICAqIHBs YXRmb3JtL0tVUkwuY3BwOgorICAgICAgICAoV2ViQ29yZTo6S1VSTDo6cmVmKToKKwogMjAwOC0w Ny0wMSAgU2ltb24gSGF1c21hbm4gIDxoYXVzbWFubkB3ZWJraXQub3JnPgogCiAgICAgICAgIEJ1 aWxkIGZpeCwgaW5jbHVkZSBEYXRlSW5zdGFuY2UuaC4KSW5kZXg6IFdlYkNvcmUvcGFnZS9Mb2Nh dGlvbi5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PQotLS0gV2ViQ29yZS9wYWdlL0xvY2F0aW9uLmNwcAkocmV2aXNp b24gMzQ5MDIpCisrKyBXZWJDb3JlL3BhZ2UvTG9jYXRpb24uY3BwCSh3b3JraW5nIGNvcHkpCkBA IC0xMjAsNyArMTIwLDcgQEAgU3RyaW5nIExvY2F0aW9uOjpoYXNoKCkgY29uc3QKICAgICAgICAg cmV0dXJuIFN0cmluZygpOwogCiAgICAgY29uc3QgS1VSTCYgdXJsID0gdGhpcy0+dXJsKCk7Ci0g ICAgcmV0dXJuIHVybC5yZWYoKS5pc051bGwoKSA/ICIiIDogIiMiICsgdXJsLnJlZigpOworICAg IHJldHVybiB1cmwucmVmKCkuaXNFbXB0eSgpID8gIiIgOiAiIyIgKyB1cmwucmVmKCk7CiB9CiAK IFN0cmluZyBMb2NhdGlvbjo6dG9TdHJpbmcoKSBjb25zdApJbmRleDogV2ViQ29yZS9wbGF0Zm9y bS9LVVJMLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09Ci0tLSBXZWJDb3JlL3BsYXRmb3JtL0tVUkwuY3BwCShyZXZp c2lvbiAzNDkwMikKKysrIFdlYkNvcmUvcGxhdGZvcm0vS1VSTC5jcHAJKHdvcmtpbmcgY29weSkK QEAgLTU3Miw3ICs1NzIsNyBAQCBTdHJpbmcgS1VSTDo6dXNlcigpIGNvbnN0CiBTdHJpbmcgS1VS TDo6cmVmKCkgY29uc3QKIHsKICAgICBpZiAobV9mcmFnbWVudEVuZCA9PSBtX3F1ZXJ5RW5kKQot ICAgICAgICByZXR1cm4gU3RyaW5nKCk7CisgICAgICAgIHJldHVybiBTdHJpbmcoIiIpOwogCiAg ICAgcmV0dXJuIG1fc3RyaW5nLnN1YnN0cmluZyhtX3F1ZXJ5RW5kICsgMSwgbV9mcmFnbWVudEVu ZCAtIChtX3F1ZXJ5RW5kICsgMSkpOwogfQpJbmRleDogTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCj09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT0KLS0tIExheW91dFRlc3RzL0NoYW5nZUxvZwkocmV2aXNpb24gMzQ5MTEpCisrKyBM YXlvdXRUZXN0cy9DaGFuZ2VMb2cJKHdvcmtpbmcgY29weSkKQEAgLTEsMyArMSwxNSBAQAorMjAw OC0wNy0wMSAgQ2FtZXJvbiBad2FyaWNoICA8Y3d6d2FyaWNoQHV3YXRlcmxvby5jYT4KKworICAg ICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBUZXN0cyBmb3I6CisK KyAgICAgICAgQnVnIDE5ODIyOiBSRUdSRVNTSU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9u Lmhhc2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9hZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtp dC5vcmcvc2hvd19idWcuY2dpP2lkPTE5ODIyPgorCisgICAgICAgICogaHR0cC90ZXN0cy9sb2Fk aW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLWV4cGVjdGVkLnR4dDogQWRkZWQuCisgICAg ICAgICogaHR0cC90ZXN0cy9sb2FkaW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLmh0bWw6 IEFkZGVkLgorCiAyMDA4LTA2LTMwICBBZGVsZSBQZXRlcnNvbiAgPGFkZWxlQGFwcGxlLmNvbT4K IAogICAgICAgICBSZXZpZXdlZCBieSBPbGl2ZXIuCkluZGV4OiBMYXlvdXRUZXN0cy9odHRwL3Rl c3RzL2xvYWRpbmcvbG9jYXRpb24taGFzaC1yZWxvYWQtY3ljbGUtZXhwZWN0ZWQudHh0Cj09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT0KLS0tIExheW91dFRlc3RzL2h0dHAvdGVzdHMvbG9hZGluZy9sb2NhdGlvbi1oYXNoLXJl bG9hZC1jeWNsZS1leHBlY3RlZC50eHQJKHJldmlzaW9uIDApCisrKyBMYXlvdXRUZXN0cy9odHRw L3Rlc3RzL2xvYWRpbmcvbG9jYXRpb24taGFzaC1yZWxvYWQtY3ljbGUtZXhwZWN0ZWQudHh0CShy ZXZpc2lvbiAwKQpAQCAtMCwwICsxLDYgQEAKK21haW4gZnJhbWUgLSBkaWRTdGFydFByb3Zpc2lv bmFsTG9hZEZvckZyYW1lCittYWluIGZyYW1lIC0gZGlkQ29tbWl0TG9hZEZvckZyYW1lCittYWlu IGZyYW1lIC0gZGlkSGFuZGxlT25sb2FkRXZlbnRzRm9yRnJhbWUKK21haW4gZnJhbWUgLSBkaWRG aW5pc2hEb2N1bWVudExvYWRGb3JGcmFtZQorbWFpbiBmcmFtZSAtIGRpZEZpbmlzaExvYWRGb3JG cmFtZQorVGhpcyB0ZXN0IGNoZWNrcyB0aGF0IG5vIGxvYWRlciBhY3Rpb25zIG9jY3VyIHdoZW4g c2V0dGluZyB3aW5kb3cubG9jYXRpb24uaGFzaCB0byB0aGUgZW1wdHkgc3RyaW5nIG9yICIjIi4K SW5kZXg6IExheW91dFRlc3RzL2h0dHAvdGVzdHMvbG9hZGluZy9sb2NhdGlvbi1oYXNoLXJlbG9h ZC1jeWNsZS5odG1sCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIExheW91dFRlc3RzL2h0dHAvdGVzdHMvbG9hZGlu Zy9sb2NhdGlvbi1oYXNoLXJlbG9hZC1jeWNsZS5odG1sCShyZXZpc2lvbiAwKQorKysgTGF5b3V0 VGVzdHMvaHR0cC90ZXN0cy9sb2FkaW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLmh0bWwJ KHJldmlzaW9uIDApCkBAIC0wLDAgKzEsMTcgQEAKKzxodG1sPgorPGhlYWQ+Cis8c2NyaXB0Pgor ZnVuY3Rpb24gcnVuVGVzdHMoKQoreworICAgIGlmICh3aW5kb3cubGF5b3V0VGVzdENvbnRyb2xs ZXIpCisgICAgICAgIGxheW91dFRlc3RDb250cm9sbGVyLmR1bXBBc1RleHQoKTsKKyAgICAKKyAg ICB3aW5kb3cubG9jYXRpb24uaGFzaCA9ICIiOworICAgIHdpbmRvdy5sb2NhdGlvbi5oYXNoID0g IiMiOworfQorPC9zY3JpcHQ+Cis8L2hlYWQ+Cis8Ym9keSBvbmxvYWQ9InJ1blRlc3RzKCk7Ij4K KzxwPlRoaXMgdGVzdCBjaGVja3MgdGhhdCBubyBsb2FkZXIgYWN0aW9ucyBvY2N1ciB3aGVuIHNl dHRpbmcgd2luZG93LmxvY2F0aW9uLmhhc2ggdG8gdGhlIGVtcHR5IHN0cmluZyBvciAiIyIuPC9w PgorPC9ib2R5PgorPC9odG1sPgo= 22031 2008-07-01 11:49:04 -0700 2008-07-01 11:59:02 -0700 Revised proposed patch hash.diff text/plain 3947 zwarich SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiAzNDkyNSkKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTcgQEAKKzIwMDgtMDctMDEgIENhbWVyb24gWndhcmljaCAgPGN3endhcmljaEB1 d2F0ZXJsb28uY2E+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAg ICAgICAgQnVnIDE5ODIyOiBSRUdSRVNTSU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9uLmhh c2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9hZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtpdC5v cmcvc2hvd19idWcuY2dpP2lkPTE5ODIyPgorCisgICAgICAgIENoYW5nZSBKU0xvY2F0aW9uOjpz ZXRIYXNoKCkgdG8gbm90IHNjaGVkdWxlIG5hdmlnYXRpb24gaW4gdGhlIGNhc2UKKyAgICAgICAg d2hlcmUgdGhlIG9sZCBmcmFnbWVudCBpcyB0aGUgbnVsbCBTdHJpbmcgYW5kIHRoZSBuZXcgZnJh Z21lbnQgaXMgYW4KKyAgICAgICAgZW1wdHkgU3RyaW5nLgorCisgICAgICAgICogYmluZGluZ3Mv anMvSlNMb2NhdGlvbkN1c3RvbS5jcHA6CisgICAgICAgIChXZWJDb3JlOjpKU0xvY2F0aW9uOjpz ZXRIYXNoKToKKwogMjAwOC0wNy0wMSAgQWRhbSBSb2JlbiAgPGFyb2JlbkBhcHBsZS5jb20+CiAK ICAgICAgICAgUm9sbCBvdXQgcjM0OTEzLCBhcyBpdCBicm9rZSB0aGUgV2luZG93cyBidWlsZCBk dWUgdG8gYmFkIGNhc3RpbmcKSW5kZXg6IFdlYkNvcmUvYmluZGluZ3MvanMvSlNMb2NhdGlvbkN1 c3RvbS5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PQotLS0gV2ViQ29yZS9iaW5kaW5ncy9qcy9KU0xvY2F0aW9uQ3Vz dG9tLmNwcAkocmV2aXNpb24gMzQ5MjUpCisrKyBXZWJDb3JlL2JpbmRpbmdzL2pzL0pTTG9jYXRp b25DdXN0b20uY3BwCSh3b3JraW5nIGNvcHkpCkBAIC0yMTMsMTAgKzIxMywxMSBAQCB2b2lkIEpT TG9jYXRpb246OnNldEhhc2goRXhlY1N0YXRlKiBleGVjCiAgICAgQVNTRVJUKGZyYW1lKTsKIAog ICAgIEtVUkwgdXJsID0gZnJhbWUtPmxvYWRlcigpLT51cmwoKTsKKyAgICBTdHJpbmcgb2xkUmVm ID0gdXJsLnJlZigpOwogICAgIFN0cmluZyBzdHIgPSB2YWx1ZS0+dG9TdHJpbmcoZXhlYyk7CiAg ICAgaWYgKHN0ci5zdGFydHNXaXRoKCIjIikpCiAgICAgICAgIHN0ciA9IHN0ci5zdWJzdHJpbmco MSk7Ci0gICAgaWYgKHVybC5yZWYoKSA9PSBzdHIpCisgICAgaWYgKHVybC5yZWYoKSA9PSBzdHIg fHwgKG9sZFJlZi5pc051bGwoKSAmJiBzdHIuaXNFbXB0eSgpKSkKICAgICAgICAgcmV0dXJuOwog ICAgIHVybC5zZXRSZWYoc3RyKTsKIApJbmRleDogTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCj09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT0KLS0tIExheW91dFRlc3RzL0NoYW5nZUxvZwkocmV2aXNpb24gMzQ5MjYpCisrKyBMYXlv dXRUZXN0cy9DaGFuZ2VMb2cJKHdvcmtpbmcgY29weSkKQEAgLTEsMyArMSwxNSBAQAorMjAwOC0w Ny0wMSAgQ2FtZXJvbiBad2FyaWNoICA8Y3d6d2FyaWNoQHV3YXRlcmxvby5jYT4KKworICAgICAg ICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBUZXN0cyBmb3I6CisKKyAg ICAgICAgQnVnIDE5ODIyOiBSRUdSRVNTSU9OIChyMzAyNDMpOiBzZXR0aW5nIGxvY2F0aW9uLmhh c2ggdG8gIiMiIGNhdXNlcyBhIHJlbG9hZAorICAgICAgICA8aHR0cHM6Ly9idWdzLndlYmtpdC5v cmcvc2hvd19idWcuY2dpP2lkPTE5ODIyPgorCisgICAgICAgICogaHR0cC90ZXN0cy9sb2FkaW5n L2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLWV4cGVjdGVkLnR4dDogQWRkZWQuCisgICAgICAg ICogaHR0cC90ZXN0cy9sb2FkaW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLmh0bWw6IEFk ZGVkLgorCiAyMDA4LTA3LTAxICBBZGFtIFJvYmVuICA8YXJvYmVuQGFwcGxlLmNvbT4KIAogICAg ICAgICBSb2xsIG91dCByMzQ5MTMsIGFzIGl0IGJyb2tlIHRoZSBXaW5kb3dzIGJ1aWxkIGR1ZSB0 byBiYWQgY2FzdGluZwpJbmRleDogTGF5b3V0VGVzdHMvaHR0cC90ZXN0cy9sb2FkaW5nL2xvY2F0 aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLWV4cGVjdGVkLnR4dAo9PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBMYXlvdXRU ZXN0cy9odHRwL3Rlc3RzL2xvYWRpbmcvbG9jYXRpb24taGFzaC1yZWxvYWQtY3ljbGUtZXhwZWN0 ZWQudHh0CShyZXZpc2lvbiAwKQorKysgTGF5b3V0VGVzdHMvaHR0cC90ZXN0cy9sb2FkaW5nL2xv Y2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLWV4cGVjdGVkLnR4dAkocmV2aXNpb24gMCkKQEAgLTAs MCArMSw2IEBACittYWluIGZyYW1lIC0gZGlkU3RhcnRQcm92aXNpb25hbExvYWRGb3JGcmFtZQor bWFpbiBmcmFtZSAtIGRpZENvbW1pdExvYWRGb3JGcmFtZQorbWFpbiBmcmFtZSAtIGRpZEhhbmRs ZU9ubG9hZEV2ZW50c0ZvckZyYW1lCittYWluIGZyYW1lIC0gZGlkRmluaXNoRG9jdW1lbnRMb2Fk Rm9yRnJhbWUKK21haW4gZnJhbWUgLSBkaWRGaW5pc2hMb2FkRm9yRnJhbWUKK1RoaXMgdGVzdCBj aGVja3MgdGhhdCBubyBsb2FkZXIgYWN0aW9ucyBvY2N1ciB3aGVuIHNldHRpbmcgd2luZG93Lmxv Y2F0aW9uLmhhc2ggdG8gdGhlIGVtcHR5IHN0cmluZyBvciAiIyIuIElmIHRoaXMgdGVzdCBmYWls cyB3aGVuIHJ1biBpbiBhIGJyb3dzZXIsIGl0IHdpbGwgcmVsb2FkIGNvbnRpbnVvdXNseS4gSWYg dGhpcyB0ZXN0IGZhaWxzIHdoZW4gcnVuIGluIER1bXBSZW5kZXJUcmVlLCB0aGUgRnJhbWVMb2Fk ZXIgY2FsbGJhY2sgb3V0cHV0IHdpbGwgY29udGFpbiB3aWxsUGVyZm9ybUNsaWVudFJlZGlyZWN0 VG9VUkwgY2FsbGJhY2tzLgpJbmRleDogTGF5b3V0VGVzdHMvaHR0cC90ZXN0cy9sb2FkaW5nL2xv Y2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLmh0bWwKPT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMv aHR0cC90ZXN0cy9sb2FkaW5nL2xvY2F0aW9uLWhhc2gtcmVsb2FkLWN5Y2xlLmh0bWwJKHJldmlz aW9uIDApCisrKyBMYXlvdXRUZXN0cy9odHRwL3Rlc3RzL2xvYWRpbmcvbG9jYXRpb24taGFzaC1y ZWxvYWQtY3ljbGUuaHRtbAkocmV2aXNpb24gMCkKQEAgLTAsMCArMSwxNyBAQAorPGh0bWw+Cis8 aGVhZD4KKzxzY3JpcHQ+CitmdW5jdGlvbiBydW5UZXN0cygpCit7CisgICAgaWYgKHdpbmRvdy5s YXlvdXRUZXN0Q29udHJvbGxlcikKKyAgICAgICAgbGF5b3V0VGVzdENvbnRyb2xsZXIuZHVtcEFz VGV4dCgpOworICAgIAorICAgIHdpbmRvdy5sb2NhdGlvbi5oYXNoID0gIiI7CisgICAgd2luZG93 LmxvY2F0aW9uLmhhc2ggPSAiIyI7Cit9Cis8L3NjcmlwdD4KKzwvaGVhZD4KKzxib2R5IG9ubG9h ZD0icnVuVGVzdHMoKTsiPgorPHA+VGhpcyB0ZXN0IGNoZWNrcyB0aGF0IG5vIGxvYWRlciBhY3Rp b25zIG9jY3VyIHdoZW4gc2V0dGluZyB3aW5kb3cubG9jYXRpb24uaGFzaCB0byB0aGUgZW1wdHkg c3RyaW5nIG9yICIjIi4gSWYgdGhpcyB0ZXN0IGZhaWxzIHdoZW4gcnVuIGluIGEgYnJvd3Nlciwg aXQgd2lsbCByZWxvYWQgY29udGludW91c2x5LiBJZiB0aGlzIHRlc3QgZmFpbHMgd2hlbiBydW4g aW4gRHVtcFJlbmRlclRyZWUsIHRoZSBGcmFtZUxvYWRlciBjYWxsYmFjayBvdXRwdXQgd2lsbCBj b250YWluIHdpbGxQZXJmb3JtQ2xpZW50UmVkaXJlY3RUb1VSTCBjYWxsYmFja3MuPC9wPgorPC9i b2R5PgorPC9odG1sPgo=