238235 2022-03-22 16:10:47 -0700 Use ASCIILiteral in a few more places where it is useful 2022-04-01 19:56:49 -0700 1 1 1 Unclassified WebKit WebCore Misc. WebKit Nightly Build Unspecified Unspecified RESOLVED FIXED https://bugs.webkit.org/show_bug.cgi?id=238162 https://bugs.webkit.org/show_bug.cgi?id=238263 https://bugs.webkit.org/show_bug.cgi?id=238287 InRadar P2 Normal --- 238693 1 cdumez cdumez changseok darin esprehn+autocc ews-watchlist ggaren gyuyoung.kim keith_miller mark.lam mifenton Morningstar msaboff saam tzagallo webkit-bug-importer oldest_to_newest 1853971 0 cdumez 2022-03-22 16:10:47 -0700 Use ASCIILiteral in a few more places where it is useful. 1853978 1 455448 cdumez 2022-03-22 16:12:21 -0700 Created attachment 455448 Patch 1853992 2 455448 ggaren 2022-03-22 16:36:03 -0700 Comment on attachment 455448 Patch r=me 1854035 3 ews-feeder 2022-03-22 18:29:07 -0700 Committed r291731 (248763@main): <https://commits.webkit.org/248763@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 455448. 1854036 4 webkit-bug-importer 2022-03-22 18:30:18 -0700 <rdar://problem/90671659> 1854174 5 455448 darin 2022-03-23 08:11:29 -0700 Comment on attachment 455448 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=455448&action=review > Source/JavaScriptCore/runtime/FunctionExecutable.cpp:162 > + functionHeader = "function "_s; I don’t understand why we want a String for functionHeader instead of a const char*. Optimizing how efficiently we initialize a String that should not be created in the first place seems like a mistake. Not sure exactly how jsMakeNontrivialString works, but it seems a lot like makeString to me. > Source/WebCore/html/HTMLTextAreaElement.cpp:398 > + normalizedValue.replace("\r\n"_s, "\n"_s); Makes no logical sense to me that _s is useful here. Replacing one substring with another should be efficient when both are const char* since neither is being used as a String after the operation is over, and adding that overload would be better than adding _s. And using _s means that if we do overload for const char* that will not be sufficient. We’ll have to overload for ASCIILiteral as well. This points to the trouble we can run into with so many different ways to represent string literals. This code isn’t performance critical, but it really offends me that it creates and destroys two StringImpl, and optimizing how quickly it does that seems like we are sort of missing the plot. 1854178 6 cdumez 2022-03-23 08:32:44 -0700 (In reply to Darin Adler from comment #5) > Comment on attachment 455448 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=455448&action=review > > > Source/JavaScriptCore/runtime/FunctionExecutable.cpp:162 > > + functionHeader = "function "_s; > > I don’t understand why we want a String for functionHeader instead of a > const char*. Optimizing how efficiently we initialize a String that should > not be created in the first place seems like a mistake. Not sure exactly how > jsMakeNontrivialString works, but it seems a lot like makeString to me. > > > Source/WebCore/html/HTMLTextAreaElement.cpp:398 > > + normalizedValue.replace("\r\n"_s, "\n"_s); > > Makes no logical sense to me that _s is useful here. Replacing one substring > with another should be efficient when both are const char* since neither is > being used as a String after the operation is over, and adding that overload > would be better than adding _s. And using _s means that if we do overload > for const char* that will not be sufficient. We’ll have to overload for > ASCIILiteral as well. > > This points to the trouble we can run into with so many different ways to > represent string literals. This code isn’t performance critical, but it > really offends me that it creates and destroys two StringImpl, and > optimizing how quickly it does that seems like we are sort of missing the > plot. To be clear, _s is definitely useful for performance here, with the current code base. What you're pointing out is that we should probably refactor the code so that we don't need a String at all. That sounds fair and I can look into this in a follow-up. Also, if we added a String::replace(const char*, const char*) and passed in a _s, I suspect the compiler would complain about it being ambiguous since an ASCIILiteral can be implicitly converted to either a const char* or a String. 1854182 7 darin 2022-03-23 08:39:26 -0700 (In reply to Chris Dumez from comment #6) > To be clear, _s is definitely useful for performance here, with the current > code base. What you're pointing out is that we should probably refactor the > code so that we don't need a String at all. Yes, and in the first, non-replace, case, it’s a super-local, super-trivial refactor, almost as simple as adding the _s. > Also, if we added a String::replace(const char*, const char*) and passed in > a _s, I suspect the compiler would complain about it being ambiguous since > an ASCIILiteral can be implicitly converted to either a const char* or a > String. On reflection what we probably want is String::replace(StringView, StringView). That might have good enough performance that we could get rid of most or all of the other overloads. 1855808 8 Morningstar 2022-03-28 15:48:12 -0700 (In reply to EWS from comment #3) > Committed r291731 (248763@main): <https://commits.webkit.org/248763@main> > > All reviewed patches have been landed. Closing bug and clearing flags on > attachment 455448 [details]. Could this patch be causing a whole suite of constantly crashing tests? https://results.webkit.org/?platform=ios&suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&test=editing%2Fdeleting%2F5126166.html&test=editing%2Fdeleting%2F5206311-1.html&test=editing%2Fdeleting%2F5206311-2.html&test=editing%2Fdeleting%2F5433862-1.html&test=editing%2Fdeleting%2Fbackspace-at-table-cell-beginning.html&test=editing%2Fdeleting%2Fdelete-line-016.html&test=editing%2Fdeleting%2Fforward-delete-empty-table-cell.html 1855811 9 cdumez 2022-03-28 15:50:20 -0700 (In reply to Matteo Flores from comment #8) > (In reply to EWS from comment #3) > > Committed r291731 (248763@main): <https://commits.webkit.org/248763@main> > > > > All reviewed patches have been landed. Closing bug and clearing flags on > > attachment 455448 [details]. > > Could this patch be causing a whole suite of constantly crashing tests? > https://results.webkit.org/?platform=ios&suite=layout-tests&suite=layout- > tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout- > tests&suite=layout-tests&test=editing%2Fdeleting%2F5126166. > html&test=editing%2Fdeleting%2F5206311-1. > html&test=editing%2Fdeleting%2F5206311-2. > html&test=editing%2Fdeleting%2F5433862-1. > html&test=editing%2Fdeleting%2Fbackspace-at-table-cell-beginning. > html&test=editing%2Fdeleting%2Fdelete-line-016. > html&test=editing%2Fdeleting%2Fforward-delete-empty-table-cell.html Hmm, crashes seem to look like: ASSERTION FAILED: areVisiblePositionsInSameTreeScope(result, vp) => useDownstream ? (result > vp) : (result < vp) ./editing/VisibleUnits.cpp(1707) : WebCore::VisiblePosition WebCore::nextSentenceBoundaryInDirection(const WebCore::VisiblePosition &, WebCore::SelectionDirection) 1 0x1381f1d89 WTFCrash 2 0x163f1d5da WebCore::nextSentenceBoundaryInDirection(WebCore::VisiblePosition const&, WebCore::SelectionDirection) 3 0x163f1cb4a WebCore::positionOfNextBoundaryOfGranularity(WebCore::VisiblePosition const&, WebCore::TextGranularity, WebCore::SelectionDirection) 4 0x124cf43b6 WebKit::WebPage::autocorrectionContext() 5 0x124cf4c78 WebKit::WebPage::preemptivelySendAutocorrectionContext() 6 0x125a264eb WebKit::WebPage::didChangeSelection(WebCore::Frame&) 7 0x12571fb35 WebKit::WebEditorClient::respondToChangedSelection(WebCore::Frame*) 8 0x163e5d360 WebCore::Editor::respondToChangedSelection(WebCore::VisibleSelection const&, WTF::OptionSet<WebCore::FrameSelection::SetSelectionOption>) 9 0x163e68314 WebCore::FrameSelection::setSelectionWithoutUpdatingAppearance(WebCore::VisibleSelection const&, WTF::OptionSet<WebCore::FrameSelection::SetSelectionOption>, WebCore::FrameSelection::CursorAlignOnScroll, WebCore::TextGranularity) 10 0x163e4b190 WebCore::FrameSelection::setSelection(WebCore::VisibleSelection const&, WTF::OptionSet<WebCore::FrameSelection::SetSelectionOption>, WebCore::AXTextStateChangeIntent, WebCore::FrameSelection::CursorAlignOnScroll, WebCore::TextGranularity) 11 0x163e5bef6 WebCore::FrameSelection::moveTo(WebCore::Position const&, WebCore::Affinity, WebCore::EUserTriggered) 12 0x16498d294 WebCore::DOMSelection::collapse(WebCore::Node*, unsigned int) 13 0x160f9b8bd WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, JSC::CallFrame*, WebCore::JSDOMSelection*)::'lambda'()::operator()() const 14 0x160f9b5e1 JSC::JSValue WebCore::toJS<WebCore::IDLUndefined, WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, JSC::CallFrame*, WebCore::JSDOMSelection*)::'lambda'()>(JSC::JSGlobalObject&, JSC::ThrowScope&, WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, JSC::CallFrame*, WebCore::JSDOMSelection*)::'lambda'()&&) 15 0x160f9b51b WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, JSC::CallFrame*, WebCore::JSDOMSelection*) 16 0x160f9b0b5 long long WebCore::IDLOperation<WebCore::JSDOMSelection>::call<&(WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, JSC::CallFrame*, WebCore::JSDOMSelection*)), (WebCore::CastedThisErrorBehavior)0>(JSC::JSGlobalObject&, JSC::CallFrame&, char const*) 17 0x160f983a4 WebCore::jsDOMSelectionPrototypeFunction_collapse(JSC::JSGlobalObject*, JSC::CallFrame*) It does not look related to me. 1855817 10 cdumez 2022-03-28 15:57:32 -0700 (In reply to Chris Dumez from comment #9) > (In reply to Matteo Flores from comment #8) > > (In reply to EWS from comment #3) > > > Committed r291731 (248763@main): <https://commits.webkit.org/248763@main> > > > > > > All reviewed patches have been landed. Closing bug and clearing flags on > > > attachment 455448 [details]. > > > > Could this patch be causing a whole suite of constantly crashing tests? > > https://results.webkit.org/?platform=ios&suite=layout-tests&suite=layout- > > tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout- > > tests&suite=layout-tests&test=editing%2Fdeleting%2F5126166. > > html&test=editing%2Fdeleting%2F5206311-1. > > html&test=editing%2Fdeleting%2F5206311-2. > > html&test=editing%2Fdeleting%2F5433862-1. > > html&test=editing%2Fdeleting%2Fbackspace-at-table-cell-beginning. > > html&test=editing%2Fdeleting%2Fdelete-line-016. > > html&test=editing%2Fdeleting%2Fforward-delete-empty-table-cell.html > > Hmm, crashes seem to look like: > ASSERTION FAILED: areVisiblePositionsInSameTreeScope(result, vp) => > useDownstream ? (result > vp) : (result < vp) > ./editing/VisibleUnits.cpp(1707) : WebCore::VisiblePosition > WebCore::nextSentenceBoundaryInDirection(const WebCore::VisiblePosition &, > WebCore::SelectionDirection) > 1 0x1381f1d89 WTFCrash > 2 0x163f1d5da > WebCore::nextSentenceBoundaryInDirection(WebCore::VisiblePosition const&, > WebCore::SelectionDirection) > 3 0x163f1cb4a > WebCore::positionOfNextBoundaryOfGranularity(WebCore::VisiblePosition > const&, WebCore::TextGranularity, WebCore::SelectionDirection) > 4 0x124cf43b6 WebKit::WebPage::autocorrectionContext() > 5 0x124cf4c78 WebKit::WebPage::preemptivelySendAutocorrectionContext() > 6 0x125a264eb WebKit::WebPage::didChangeSelection(WebCore::Frame&) > 7 0x12571fb35 > WebKit::WebEditorClient::respondToChangedSelection(WebCore::Frame*) > 8 0x163e5d360 > WebCore::Editor::respondToChangedSelection(WebCore::VisibleSelection const&, > WTF::OptionSet<WebCore::FrameSelection::SetSelectionOption>) > 9 0x163e68314 > WebCore::FrameSelection::setSelectionWithoutUpdatingAppearance(WebCore:: > VisibleSelection const&, > WTF::OptionSet<WebCore::FrameSelection::SetSelectionOption>, > WebCore::FrameSelection::CursorAlignOnScroll, WebCore::TextGranularity) > 10 0x163e4b190 > WebCore::FrameSelection::setSelection(WebCore::VisibleSelection const&, > WTF::OptionSet<WebCore::FrameSelection::SetSelectionOption>, > WebCore::AXTextStateChangeIntent, > WebCore::FrameSelection::CursorAlignOnScroll, WebCore::TextGranularity) > 11 0x163e5bef6 WebCore::FrameSelection::moveTo(WebCore::Position const&, > WebCore::Affinity, WebCore::EUserTriggered) > 12 0x16498d294 WebCore::DOMSelection::collapse(WebCore::Node*, unsigned int) > 13 0x160f9b8bd > WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, > JSC::CallFrame*, WebCore::JSDOMSelection*)::'lambda'()::operator()() const > 14 0x160f9b5e1 JSC::JSValue WebCore::toJS<WebCore::IDLUndefined, > WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, > JSC::CallFrame*, > WebCore::JSDOMSelection*)::'lambda'()>(JSC::JSGlobalObject&, > JSC::ThrowScope&, > WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, > JSC::CallFrame*, WebCore::JSDOMSelection*)::'lambda'()&&) > 15 0x160f9b51b > WebCore::jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, > JSC::CallFrame*, WebCore::JSDOMSelection*) > 16 0x160f9b0b5 long long > WebCore::IDLOperation<WebCore::JSDOMSelection>::call<&(WebCore:: > jsDOMSelectionPrototypeFunction_collapseBody(JSC::JSGlobalObject*, > JSC::CallFrame*, WebCore::JSDOMSelection*)), > (WebCore::CastedThisErrorBehavior)0>(JSC::JSGlobalObject&, JSC::CallFrame&, > char const*) > 17 0x160f983a4 > WebCore::jsDOMSelectionPrototypeFunction_collapse(JSC::JSGlobalObject*, > JSC::CallFrame*) > > It does not look related to me. Wenson says this is a regression from https://github.com/WebKit/WebKit/commit/5bd94703d84ac79d2a2d3fefe454709c54145e84, not my change. 455448 2022-03-22 16:12:21 -0700 2022-03-22 18:29:10 -0700 Patch bug-238235-20220322161220.patch text/plain 5534 cdumez U3VidmVyc2lvbiBSZXZpc2lvbjogMjkxNzE1CmRpZmYgLS1naXQgYS9Tb3VyY2UvSmF2YVNjcmlw dENvcmUvQ2hhbmdlTG9nIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwppbmRleCBm MjhiZTQ2MDVlZjQ5MWFhYmRiNDU3Yzk0ZWQ3ZmYzNmMyMDU5MWZlLi5jN2E2YzNlYWJlOTE5ZDBl MmExMzI1NzFjN2NlOTNmYzlkZmIxNTZhIDEwMDY0NAotLS0gYS9Tb3VyY2UvSmF2YVNjcmlwdENv cmUvQ2hhbmdlTG9nCisrKyBiL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKQEAgLTEs MyArMSwxMyBAQAorMjAyMi0wMy0yMiAgQ2hyaXMgRHVtZXogIDxjZHVtZXpAYXBwbGUuY29tPgor CisgICAgICAgIFVzZSBBU0NJSUxpdGVyYWwgaW4gYSBmZXcgbW9yZSBwbGFjZXMgd2hlcmUgaXQg aXMgdXNlZnVsCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9p ZD0yMzgyMzUKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAg ICAqIHJ1bnRpbWUvRnVuY3Rpb25FeGVjdXRhYmxlLmNwcDoKKyAgICAgICAgKEpTQzo6RnVuY3Rp b25FeGVjdXRhYmxlOjp0b1N0cmluZ1Nsb3cpOgorCiAyMDIyLTAzLTIxICBZdXN1a2UgU3V6dWtp ICA8eXN1enVraUBhcHBsZS5jb20+CiAKICAgICAgICAgW0pTQ10gQ2hhbmdlIERhdGUucGFyc2Ug dG8gc3RvcCByZXR1cm5pbmcgbnVtYmVycyB3aXRoIGZyYWN0aW9uYWwgcGFydApkaWZmIC0tZ2l0 IGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nIGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCmlu ZGV4IDc3ZGNjYmY5NjBhZjBhNjBjNDQ5MTUxOGFmM2MyZGU4ODE0YzQwZDguLmFhYmMwMmU1YWQ4 MWFlMWVlYjRjNmY4OGM5YWMwMWQ2Y2JiZjI2MjQgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3Jl L0NoYW5nZUxvZworKysgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKQEAgLTEsMyArMSwxNSBA QAorMjAyMi0wMy0yMiAgQ2hyaXMgRHVtZXogIDxjZHVtZXpAYXBwbGUuY29tPgorCisgICAgICAg IFVzZSBBU0NJSUxpdGVyYWwgaW4gYSBmZXcgbW9yZSBwbGFjZXMgd2hlcmUgaXQgaXMgdXNlZnVs CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0yMzgyMzUK KworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICAqIGh0bWwv SFRNTFRleHRBcmVhRWxlbWVudC5jcHA6CisgICAgICAgIChXZWJDb3JlOjpIVE1MVGV4dEFyZWFF bGVtZW50OjpzZXRWYWx1ZUNvbW1vbik6CisgICAgICAgICogcGxhdGZvcm0vYW5pbWF0aW9uL1Rp bWluZ0Z1bmN0aW9uLmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OlRpbWluZ0Z1bmN0aW9uOjpjc3NU ZXh0IGNvbnN0KToKKwogMjAyMi0wMy0yMiAgRnVqaWkgSGlyb25vcmkgIDxIaXJvbm9yaS5GdWpp aUBzb255LmNvbT4KIAogICAgICAgICBbV2luQ2Fpcm9dIFJlbW90ZUdyYXBoaWNzQ29udGV4dEdM LmgoNzgpOiBlcnJvciBDMjAyNzogdXNlIG9mIHVuZGVmaW5lZCB0eXBlICdXZWJDb3JlOjpHcmFw aGljc0NvbnRleHRHTCcKZGlmZiAtLWdpdCBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9ydW50aW1l L0Z1bmN0aW9uRXhlY3V0YWJsZS5jcHAgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvcnVudGltZS9G dW5jdGlvbkV4ZWN1dGFibGUuY3BwCmluZGV4IGExNzY0N2FjOWFhMjNlNzU0NzU2NzJlNGM2ZWI1 MmZkMDU5ZTRiNmMuLmYxNDlkNGU1OWE3MjRhODQyNTdmYjY4Nzg0OTJjMmFhM2IxNjJlOGIgMTAw NjQ0Ci0tLSBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9ydW50aW1lL0Z1bmN0aW9uRXhlY3V0YWJs ZS5jcHAKKysrIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL3J1bnRpbWUvRnVuY3Rpb25FeGVjdXRh YmxlLmNwcApAQCAtMTQ1LDcgKzE0NSw3IEBAIEpTU3RyaW5nKiBGdW5jdGlvbkV4ZWN1dGFibGU6 OnRvU3RyaW5nU2xvdyhKU0dsb2JhbE9iamVjdCogZ2xvYmFsT2JqZWN0KQogICAgIHN3aXRjaCAo cGFyc2VNb2RlKCkpIHsKICAgICBjYXNlIFNvdXJjZVBhcnNlTW9kZTo6R2VuZXJhdG9yV3JhcHBl ckZ1bmN0aW9uTW9kZToKICAgICBjYXNlIFNvdXJjZVBhcnNlTW9kZTo6R2VuZXJhdG9yV3JhcHBl ck1ldGhvZE1vZGU6Ci0gICAgICAgIGZ1bmN0aW9uSGVhZGVyID0gImZ1bmN0aW9uKiAiOworICAg ICAgICBmdW5jdGlvbkhlYWRlciA9ICJmdW5jdGlvbiogIl9zOwogICAgICAgICBicmVhazsKIAog ICAgIGNhc2UgU291cmNlUGFyc2VNb2RlOjpOb3JtYWxGdW5jdGlvbk1vZGU6CkBAIC0xNTksNyAr MTU5LDcgQEAgSlNTdHJpbmcqIEZ1bmN0aW9uRXhlY3V0YWJsZTo6dG9TdHJpbmdTbG93KEpTR2xv YmFsT2JqZWN0KiBnbG9iYWxPYmplY3QpCiAgICAgY2FzZSBTb3VyY2VQYXJzZU1vZGU6OkFzeW5j R2VuZXJhdG9yQm9keU1vZGU6CiAgICAgY2FzZSBTb3VyY2VQYXJzZU1vZGU6OkFzeW5jRnVuY3Rp b25Cb2R5TW9kZToKICAgICBjYXNlIFNvdXJjZVBhcnNlTW9kZTo6QXN5bmNBcnJvd0Z1bmN0aW9u Qm9keU1vZGU6Ci0gICAgICAgIGZ1bmN0aW9uSGVhZGVyID0gImZ1bmN0aW9uICI7CisgICAgICAg IGZ1bmN0aW9uSGVhZGVyID0gImZ1bmN0aW9uICJfczsKICAgICAgICAgYnJlYWs7CiAKICAgICBj YXNlIFNvdXJjZVBhcnNlTW9kZTo6QXJyb3dGdW5jdGlvbk1vZGU6CkBAIC0xNjksMTYgKzE2OSwx NiBAQCBKU1N0cmluZyogRnVuY3Rpb25FeGVjdXRhYmxlOjp0b1N0cmluZ1Nsb3coSlNHbG9iYWxP YmplY3QqIGdsb2JhbE9iamVjdCkKIAogICAgIGNhc2UgU291cmNlUGFyc2VNb2RlOjpBc3luY0Z1 bmN0aW9uTW9kZToKICAgICBjYXNlIFNvdXJjZVBhcnNlTW9kZTo6QXN5bmNNZXRob2RNb2RlOgot ICAgICAgICBmdW5jdGlvbkhlYWRlciA9ICJhc3luYyBmdW5jdGlvbiAiOworICAgICAgICBmdW5j dGlvbkhlYWRlciA9ICJhc3luYyBmdW5jdGlvbiAiX3M7CiAgICAgICAgIGJyZWFrOwogCiAgICAg Y2FzZSBTb3VyY2VQYXJzZU1vZGU6OkFzeW5jQXJyb3dGdW5jdGlvbk1vZGU6Ci0gICAgICAgIGZ1 bmN0aW9uSGVhZGVyID0gImFzeW5jICI7CisgICAgICAgIGZ1bmN0aW9uSGVhZGVyID0gImFzeW5j ICJfczsKICAgICAgICAgYnJlYWs7CiAKICAgICBjYXNlIFNvdXJjZVBhcnNlTW9kZTo6QXN5bmNH ZW5lcmF0b3JXcmFwcGVyRnVuY3Rpb25Nb2RlOgogICAgIGNhc2UgU291cmNlUGFyc2VNb2RlOjpB c3luY0dlbmVyYXRvcldyYXBwZXJNZXRob2RNb2RlOgotICAgICAgICBmdW5jdGlvbkhlYWRlciA9 ICJhc3luYyBmdW5jdGlvbiogIjsKKyAgICAgICAgZnVuY3Rpb25IZWFkZXIgPSAiYXN5bmMgZnVu Y3Rpb24qICJfczsKICAgICAgICAgYnJlYWs7CiAgICAgfQogCmRpZmYgLS1naXQgYS9Tb3VyY2Uv V2ViQ29yZS9odG1sL0hUTUxUZXh0QXJlYUVsZW1lbnQuY3BwIGIvU291cmNlL1dlYkNvcmUvaHRt bC9IVE1MVGV4dEFyZWFFbGVtZW50LmNwcAppbmRleCAyZjEwY2VkODQxMzFlZDdlOWRkMTg3YTM0 MDFjZmRjZGVjNGJhNDNlLi5kNTc3ZWE2NmVhYWEwMzE3MjAyOTU1MzBiMWFjMGQ1OTM4MzZmMGZi IDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29yZS9odG1sL0hUTUxUZXh0QXJlYUVsZW1lbnQuY3Bw CisrKyBiL1NvdXJjZS9XZWJDb3JlL2h0bWwvSFRNTFRleHRBcmVhRWxlbWVudC5jcHAKQEAgLTM5 NSw3ICszOTUsNyBAQCB2b2lkIEhUTUxUZXh0QXJlYUVsZW1lbnQ6OnNldFZhbHVlQ29tbW9uKGNv bnN0IFN0cmluZyYgbmV3VmFsdWUsIFRleHRGaWVsZEV2ZW50QgogICAgIC8vIENvZGUgZWxzZXdo ZXJlIG5vcm1hbGl6ZXMgbGluZSBlbmRpbmdzIGFkZGVkIGJ5IHRoZSB1c2VyIHZpYSB0aGUga2V5 Ym9hcmQgb3IgcGFzdGluZy4KICAgICAvLyBXZSBub3JtYWxpemUgbGluZSBlbmRpbmdzIGNvbWlu ZyBmcm9tIEphdmFTY3JpcHQgaGVyZS4KICAgICBTdHJpbmcgbm9ybWFsaXplZFZhbHVlID0gbmV3 VmFsdWUuaXNOdWxsKCkgPyBlbXB0eVN0cmluZygpIDogbmV3VmFsdWU7Ci0gICAgbm9ybWFsaXpl ZFZhbHVlLnJlcGxhY2UoIlxyXG4iLCAiXG4iKTsKKyAgICBub3JtYWxpemVkVmFsdWUucmVwbGFj ZSgiXHJcbiJfcywgIlxuIl9zKTsKICAgICBub3JtYWxpemVkVmFsdWUucmVwbGFjZSgnXHInLCAn XG4nKTsKIAogICAgIC8vIFJldHVybiBlYXJseSBiZWNhdXNlIHdlIGRvbid0IHdhbnQgdG8gbW92 ZSB0aGUgY2FyZXQgb3IgdHJpZ2dlciBvdGhlciBzaWRlIGVmZmVjdHMKZGlmZiAtLWdpdCBhL1Nv dXJjZS9XZWJDb3JlL3BsYXRmb3JtL2FuaW1hdGlvbi9UaW1pbmdGdW5jdGlvbi5jcHAgYi9Tb3Vy Y2UvV2ViQ29yZS9wbGF0Zm9ybS9hbmltYXRpb24vVGltaW5nRnVuY3Rpb24uY3BwCmluZGV4IDk3 YjIwNTdmNmE0MGZkZWJlNGRkMzU2ZGYzYzM3NGIwODZlNWFmNzYuLjNkMjc1MzViYmMzYmQxZjk2 YjIzZDAxYzEzZWE1ZGE4OTRkODhjMDEgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3BsYXRm b3JtL2FuaW1hdGlvbi9UaW1pbmdGdW5jdGlvbi5jcHAKKysrIGIvU291cmNlL1dlYkNvcmUvcGxh dGZvcm0vYW5pbWF0aW9uL1RpbWluZ0Z1bmN0aW9uLmNwcApAQCAtMTk4LDEzICsxOTgsMTMgQEAg U3RyaW5nIFRpbWluZ0Z1bmN0aW9uOjpjc3NUZXh0KCkgY29uc3QKICAgICBpZiAobV90eXBlID09 IFRpbWluZ0Z1bmN0aW9uOjpDdWJpY0JlemllckZ1bmN0aW9uKSB7CiAgICAgICAgIGF1dG8mIGZ1 bmN0aW9uID0gZG93bmNhc3Q8Q3ViaWNCZXppZXJUaW1pbmdGdW5jdGlvbj4oKnRoaXMpOwogICAg ICAgICBpZiAoZnVuY3Rpb24ueDEoKSA9PSAwLjI1ICYmIGZ1bmN0aW9uLnkxKCkgPT0gMC4xICYm IGZ1bmN0aW9uLngyKCkgPT0gMC4yNSAmJiBmdW5jdGlvbi55MigpID09IDEuMCkKLSAgICAgICAg ICAgIHJldHVybiAiZWFzZSI7CisgICAgICAgICAgICByZXR1cm4gImVhc2UiX3M7CiAgICAgICAg IGlmIChmdW5jdGlvbi54MSgpID09IDAuNDIgJiYgIWZ1bmN0aW9uLnkxKCkgJiYgZnVuY3Rpb24u eDIoKSA9PSAxLjAgJiYgZnVuY3Rpb24ueTIoKSA9PSAxLjApCi0gICAgICAgICAgICByZXR1cm4g ImVhc2UtaW4iOworICAgICAgICAgICAgcmV0dXJuICJlYXNlLWluIl9zOwogICAgICAgICBpZiAo IWZ1bmN0aW9uLngxKCkgJiYgIWZ1bmN0aW9uLnkxKCkgJiYgZnVuY3Rpb24ueDIoKSA9PSAwLjU4 ICYmIGZ1bmN0aW9uLnkyKCkgPT0gMS4wKQotICAgICAgICAgICAgcmV0dXJuICJlYXNlLW91dCI7 CisgICAgICAgICAgICByZXR1cm4gImVhc2Utb3V0Il9zOwogICAgICAgICBpZiAoZnVuY3Rpb24u eDEoKSA9PSAwLjQyICYmICFmdW5jdGlvbi55MSgpICYmIGZ1bmN0aW9uLngyKCkgPT0gMC41OCAm JiBmdW5jdGlvbi55MigpID09IDEuMCkKLSAgICAgICAgICAgIHJldHVybiAiZWFzZS1pbi1vdXQi OworICAgICAgICAgICAgcmV0dXJuICJlYXNlLWluLW91dCJfczsKICAgICAgICAgcmV0dXJuIG1h a2VTdHJpbmcoImN1YmljLWJlemllcigiLCBmdW5jdGlvbi54MSgpLCAiLCAiLCBmdW5jdGlvbi55 MSgpLCAiLCAiLCBmdW5jdGlvbi54MigpLCAiLCAiLCBmdW5jdGlvbi55MigpLCAnKScpOwogICAg IH0KIAo=