80926 2012-03-12 19:13:25 -0700 Fix some compiler warnings (miscellaneous) 2012-03-14 10:35:43 -0700 1 1 1 Unclassified WebKit JavaScriptCore 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 staikos staikos ap barraclough fpizlo ggaren japhet manyoso mhahnenberg webkit.review.bot zimmermann oldest_to_newest 576901 0 131491 staikos 2012-03-12 19:13:25 -0700 Created attachment 131491 Fix signed/unsigned mismatches +++ This bug was initially created as a clone of Bug #80790 +++ Patches to come - mostly signed/unsigned changes. 577556 1 ap 2012-03-13 11:04:58 -0700 CC'ing some JSC folks in case these warnings point to something more interesting than just a need to cast. 577557 2 131491 ap 2012-03-13 11:05:36 -0700 Comment on attachment 131491 Fix signed/unsigned mismatches As mentioned in original bug, please use "int" instead of "signed", and please use C++ casts. 577571 3 131674 staikos 2012-03-13 11:16:07 -0700 Created attachment 131674 Don't define the static if it isn't used (ie: SA_RESTART defined) 577586 4 staikos 2012-03-13 11:22:10 -0700 (In reply to comment #2) > (From update of attachment 131491 [details]) > As mentioned in original bug, please use "int" instead of "signed", and please use C++ casts. http://www.cplusplus.com/doc/tutorial/typecasting/ Those are C++ casts. I can change to int. 577590 5 131678 staikos 2012-03-13 11:23:51 -0700 Created attachment 131678 Fix signed/unsigned mismatches 577625 6 ap 2012-03-13 11:57:55 -0700 > Those are C++ casts. I can change to int. It's a long-standing and agreed upon rule in WebKit, although it doesn't seem to be mentioned in <http://www.webkit.org/coding/coding-style.html>. Perhaps because it's more of safe coding rule than a style one. There is no doubt that C-style casts are still supported in C++, but that doesn't make them welcome in WebKit. 577627 7 131678 ap 2012-03-13 11:58:17 -0700 Comment on attachment 131678 Fix signed/unsigned mismatches Still wrong casts. 577668 8 staikos 2012-03-13 12:31:13 -0700 (In reply to comment #7) > (From update of attachment 131678 [details]) > Still wrong casts. Please enlighten me. 577682 9 ap 2012-03-13 12:43:38 -0700 > Please enlighten me. Does comment 6 not help? 577712 10 staikos 2012-03-13 13:07:42 -0700 (In reply to comment #9) > > Please enlighten me. > > Does comment 6 not help? Those are not C-style casts. They're C++-style. If you want a different C++ cast, please select: 1) static_cast<> 2) reinterpret_cast<> 3) dynamic_cast<> (of course it will fail, as will const_cast<>) Really, they're c++ casts. They also pass the style checker and the style guideline referenced. http://www.cplusplus.com/doc/tutorial/typecasting/ explains. I also think it looks a LOT nicer than an unneeded big long string of text. Saying "use C++ cast" is meaningless since it is a C++ cast - and that's my point. 577736 11 ap 2012-03-13 13:23:43 -0700 If you want WebKit coding style to change, please e-mail webkit-dev. If you can suggest a more clear description for C++ style casts to use in future reviews, please do so. I think that you understood my comment from the first time, and are just trying to be difficult. 577751 12 staikos 2012-03-13 13:31:07 -0700 (In reply to comment #11) > If you want WebKit coding style to change, please e-mail webkit-dev. If you can suggest a more clear description for C++ style casts to use in future reviews, please do so. > > I think that you understood my comment from the first time, and are just trying to be difficult. No, I want you to explain it. 577826 13 131678 fpizlo 2012-03-13 14:41:17 -0700 Comment on attachment 131678 Fix signed/unsigned mismatches View in context: https://bugs.webkit.org/attachment.cgi?id=131678&action=review I agree with you that it is really enjoyable to have philosophical discussions about which of the various styles of casting to use. There are many of them and their differences are fabulous. However, what's important for us is that we stick to one style. We use the foo_cast<bar>(baz) style, and we generally stay away from bar(baz) or (bar)baz. That's just our convention. > Source/JavaScriptCore/assembler/AssemblerBufferWithConstantPool.h:218 > + if (isForced || 5 * unsigned(m_numConsts) > 3 * maxPoolSize / sizeof(uint32_t)) This should be static_cast<m_numConsts> > Source/JavaScriptCore/assembler/MacroAssemblerARM.h:463 > + ARMWord tmp = (right.m_value == int(0x80000000)) ? ARMAssembler::INVALID_IMM : m_assembler.getOp2(-right.m_value); This should be static_cast<int>(0x80000000). > Source/JavaScriptCore/parser/Lexer.cpp:632 > + ASSERT(unsigned(c) <= USHRT_MAX); This should be static_cast<unsigned>(c) 577828 14 131678 fpizlo 2012-03-13 14:42:11 -0700 Comment on attachment 131678 Fix signed/unsigned mismatches View in context: https://bugs.webkit.org/attachment.cgi?id=131678&action=review >> Source/JavaScriptCore/assembler/AssemblerBufferWithConstantPool.h:218 >> + if (isForced || 5 * unsigned(m_numConsts) > 3 * maxPoolSize / sizeof(uint32_t)) > > This should be static_cast<m_numConsts> Correction, shjould be static_cast<unsigned>(m_numConstants) 577834 15 staikos 2012-03-13 14:52:04 -0700 (In reply to comment #13) > (From update of attachment 131678 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=131678&action=review > > I agree with you that it is really enjoyable to have philosophical discussions about which of the various styles of casting to use. There are many of them and their differences are fabulous. > > However, what's important for us is that we stick to one style. We use the foo_cast<bar>(baz) style, and we generally stay away from bar(baz) or (bar)baz. That's just our convention. > > > Source/JavaScriptCore/assembler/AssemblerBufferWithConstantPool.h:218 > > + if (isForced || 5 * unsigned(m_numConsts) > 3 * maxPoolSize / sizeof(uint32_t)) > > This should be static_cast<m_numConsts> So in reality this has nothing to do with C++, but more "write it my way". That's fine. Let's just call it such. 577835 16 131674 webkit.review.bot 2012-03-13 14:52:49 -0700 Comment on attachment 131674 Don't define the static if it isn't used (ie: SA_RESTART defined) Clearing flags on attachment: 131674 Committed r110617: <http://trac.webkit.org/changeset/110617> 577836 17 webkit.review.bot 2012-03-13 14:52:55 -0700 All reviewed patches have been landed. Closing bug. 577840 18 fpizlo 2012-03-13 14:58:08 -0700 (In reply to comment #15) > (In reply to comment #13) > > (From update of attachment 131678 [details] [details]) > > View in context: https://bugs.webkit.org/attachment.cgi?id=131678&action=review > > > > I agree with you that it is really enjoyable to have philosophical discussions about which of the various styles of casting to use. There are many of them and their differences are fabulous. > > > > However, what's important for us is that we stick to one style. We use the foo_cast<bar>(baz) style, and we generally stay away from bar(baz) or (bar)baz. That's just our convention. > > > > > Source/JavaScriptCore/assembler/AssemblerBufferWithConstantPool.h:218 > > > + if (isForced || 5 * unsigned(m_numConsts) > 3 * maxPoolSize / sizeof(uint32_t)) > > > > This should be static_cast<m_numConsts> > > So in reality this has nothing to do with C++, but more "write it my way". That's fine. Let's just call it such. To be clear, it's about consistency. It's better to have all of the code do it one way. It may be the wrong way; who knows. Doesn't really matter, since the semantics differences are minor. I prefer (bar)baz, you prefer bar(baz), but the code already uses foo_cast<bar>(baz) everywhere so it's better if we stick to it. 578248 19 zimmermann 2012-03-14 01:51:43 -0700 (In reply to comment #15) > So in reality this has nothing to do with C++, but more "write it my way". That's fine. Let's just call it such. Indeed, it's really just our style. I've been yelled at the past in my patches, to stop using unsigned(foo) and use static_cast<unsigned>(foo) instead, always with the argument that C style casts should be avoided. I was under the impression that (unsigned) foo and unsigned(foo) are C++ style casts, and it almost loos like Alexey is/was under the same impression. We should nail this down and force static_cast<unsigned>() usage in the stye guide, to avoid confusion. 578401 20 manyoso 2012-03-14 08:13:33 -0700 (In reply to comment #19) > (In reply to comment #15) > > So in reality this has nothing to do with C++, but more "write it my way". That's fine. Let's just call it such. > > Indeed, it's really just our style. I've been yelled at the past in my patches, to stop using unsigned(foo) and use static_cast<unsigned>(foo) instead, always with the argument that C style casts should be avoided. > I was under the impression that (unsigned) foo and unsigned(foo) are C++ style casts, and it almost loos like Alexey is/was under the same impression. > > We should nail this down and force static_cast<unsigned>() usage in the stye guide, to avoid confusion. Or quit trying to enforce unwritten style guide rules. I see *no* reason that the style guide should be updated to include this unwritten and informal 'style guideline'. And arguing consistency... does this really make the code more readable/hackable? Not at all in my opinion. And that is what consistency is for -> to make the code more readable/hackable/followable. 578479 21 ap 2012-03-14 09:21:49 -0700 It's incorrect that this is just for consistency. C-style casts (whether you write them using old C syntax "(type)variable" or constructor-style syntax "type(variable)" that you like to call C++ despite it having the same semantics) are more powerful and thus error-prone. Again, if you want to change WebKit coding style, please go to webkit-dev. 578542 22 manyoso 2012-03-14 10:35:43 -0700 (In reply to comment #21) > It's incorrect that this is just for consistency. C-style casts (whether you write them using old C syntax "(type)variable" or constructor-style syntax "type(variable)" that you like to call C++ despite it having the same semantics) are more powerful and thus error-prone. > > Again, if you want to change WebKit coding style, please go to webkit-dev. But I'm not trying to change webkit coding style: http://www.webkit.org/coding/coding-style.html If this is not for consistency, then please tell me how the casts in this patch were error prone. What future change do you think was likely to have an error because of this usage. 131491 2012-03-12 19:13:25 -0700 2012-03-13 11:23:51 -0700 Fix signed/unsigned mismatches 80790_1.patch text/plain 2775 staikos SW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKPT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291 cmNlL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkocmV2aXNpb24gMTEwMzk1KQorKysgU291cmNl L0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE3IEBA CisyMDEyLTAzLTExICBHZW9yZ2UgU3RhaWtvcyAgPHN0YWlrb3NAd2Via2l0Lm9yZz4KKworICAg ICAgICBGaXggc2lnbmVkL3Vuc2lnbmVkIG1pc21hdGNoIGNvbXBpbGVyIHdhcm5pbmdzLgorICAg ICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9ODA3OTAKKworICAg ICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICAqIGFzc2VtYmxlci9B c3NlbWJsZXJCdWZmZXJXaXRoQ29uc3RhbnRQb29sLmg6CisgICAgICAgIChKU0M6OkFzc2VtYmxl ckJ1ZmZlcldpdGhDb25zdGFudFBvb2w6OmZsdXNoV2l0aG91dEJhcnJpZXIpOgorICAgICAgICAq IGFzc2VtYmxlci9NYWNyb0Fzc2VtYmxlckFSTS5oOgorICAgICAgICAoSlNDOjpNYWNyb0Fzc2Vt YmxlckFSTTo6YnJhbmNoMzIpOgorICAgICAgICAqIHBhcnNlci9MZXhlci5jcHA6CisgICAgICAg IChKU0M6Ojo6cmVjb3JkMTYpOgorCiAyMDEyLTAzLTEwICBGaWxpcCBQaXpsbyAgPGZwaXpsb0Bh cHBsZS5jb20+CiAKICAgICAgICAgTExJbnQgc2hvdWxkIHN1cHBvcnQgSlNWQUxVRTY0CkluZGV4 OiBTb3VyY2UvSmF2YVNjcmlwdENvcmUvYXNzZW1ibGVyL0Fzc2VtYmxlckJ1ZmZlcldpdGhDb25z dGFudFBvb2wuaAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvSmF2YVNjcmlwdENvcmUvYXNzZW1ibGVy L0Fzc2VtYmxlckJ1ZmZlcldpdGhDb25zdGFudFBvb2wuaAkocmV2aXNpb24gMTEwMzkzKQorKysg U291cmNlL0phdmFTY3JpcHRDb3JlL2Fzc2VtYmxlci9Bc3NlbWJsZXJCdWZmZXJXaXRoQ29uc3Rh bnRQb29sLmgJKHdvcmtpbmcgY29weSkKQEAgLTIxNSw3ICsyMTUsNyBAQCBwdWJsaWM6CiAgICAg dm9pZCBmbHVzaFdpdGhvdXRCYXJyaWVyKGJvb2wgaXNGb3JjZWQgPSBmYWxzZSkKICAgICB7CiAg ICAgICAgIC8vIEZsdXNoIGlmIGNvbnN0YW50IHBvb2wgaXMgbW9yZSB0aGFuIDYwJSBmdWxsIHRv IGF2b2lkIG92ZXJ1c2Ugb2YgdGhpcyBmdW5jdGlvbi4KLSAgICAgICAgaWYgKGlzRm9yY2VkIHx8 IDUgKiBtX251bUNvbnN0cyA+IDMgKiBtYXhQb29sU2l6ZSAvIHNpemVvZih1aW50MzJfdCkpCisg ICAgICAgIGlmIChpc0ZvcmNlZCB8fCA1ICogdW5zaWduZWQobV9udW1Db25zdHMpID4gMyAqIG1h eFBvb2xTaXplIC8gc2l6ZW9mKHVpbnQzMl90KSkKICAgICAgICAgICAgIGZsdXNoQ29uc3RhbnRQ b29sKGZhbHNlKTsKICAgICB9CiAKSW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9hc3NlbWJs ZXIvTWFjcm9Bc3NlbWJsZXJBUk0uaAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvSmF2YVNjcmlwdENv cmUvYXNzZW1ibGVyL01hY3JvQXNzZW1ibGVyQVJNLmgJKHJldmlzaW9uIDExMDM5MykKKysrIFNv dXJjZS9KYXZhU2NyaXB0Q29yZS9hc3NlbWJsZXIvTWFjcm9Bc3NlbWJsZXJBUk0uaAkod29ya2lu ZyBjb3B5KQpAQCAtNDYwLDcgKzQ2MCw3IEBAIHB1YmxpYzoKICAgICAgICAgICAgIG1fYXNzZW1i bGVyLmxkcl91bl9pbW0oQVJNUmVnaXN0ZXJzOjpTMCwgcmlnaHQubV92YWx1ZSk7CiAgICAgICAg ICAgICBtX2Fzc2VtYmxlci5jbXBfcihsZWZ0LCBBUk1SZWdpc3RlcnM6OlMwKTsKICAgICAgICAg fSBlbHNlIHsKLSAgICAgICAgICAgIEFSTVdvcmQgdG1wID0gKHJpZ2h0Lm1fdmFsdWUgPT0gMHg4 MDAwMDAwMCkgPyBBUk1Bc3NlbWJsZXI6OklOVkFMSURfSU1NIDogbV9hc3NlbWJsZXIuZ2V0T3Ay KC1yaWdodC5tX3ZhbHVlKTsKKyAgICAgICAgICAgIEFSTVdvcmQgdG1wID0gKHJpZ2h0Lm1fdmFs dWUgPT0gc2lnbmVkKDB4ODAwMDAwMDApKSA/IEFSTUFzc2VtYmxlcjo6SU5WQUxJRF9JTU0gOiBt X2Fzc2VtYmxlci5nZXRPcDIoLXJpZ2h0Lm1fdmFsdWUpOwogICAgICAgICAgICAgaWYgKHRtcCAh PSBBUk1Bc3NlbWJsZXI6OklOVkFMSURfSU1NKQogICAgICAgICAgICAgICAgIG1fYXNzZW1ibGVy LmNtbl9yKGxlZnQsIHRtcCk7CiAgICAgICAgICAgICBlbHNlCkluZGV4OiBTb3VyY2UvSmF2YVNj cmlwdENvcmUvcGFyc2VyL0xleGVyLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvSmF2YVNjcmlw dENvcmUvcGFyc2VyL0xleGVyLmNwcAkocmV2aXNpb24gMTEwMzkzKQorKysgU291cmNlL0phdmFT Y3JpcHRDb3JlL3BhcnNlci9MZXhlci5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTYyOSw3ICs2Mjks NyBAQCB0ZW1wbGF0ZSA8dHlwZW5hbWUgVD4KIGlubGluZSB2b2lkIExleGVyPFQ+OjpyZWNvcmQx NihpbnQgYykKIHsKICAgICBBU1NFUlQoYyA+PSAwKTsKLSAgICBBU1NFUlQoYyA8PSBVU0hSVF9N QVgpOworICAgIEFTU0VSVCh1bnNpZ25lZChjKSA8PSBVU0hSVF9NQVgpOwogICAgIG1fYnVmZmVy MTYuYXBwZW5kKHN0YXRpY19jYXN0PFVDaGFyPihjKSk7CiB9CiAK 131674 2012-03-13 11:16:07 -0700 2012-03-13 14:52:49 -0700 Don't define the static if it isn't used (ie: SA_RESTART defined) 80926_2.patch text/plain 1315 staikos SW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKPT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291 cmNlL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkocmV2aXNpb24gMTEwNTgxKQorKysgU291cmNl L0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE0IEBA CisyMDEyLTAzLTEzICBHZW9yZ2UgU3RhaWtvcyAgPHN0YWlrb3NAd2Via2l0Lm9yZz4KKworICAg ICAgICBUaGUgY2FsbGJhY2sgaXMgb25seSB1c2VkIGlmIFNBX1JFU1RBUlQgaXMgZGVmaW5lZC4g IENvbXBpbGUgaXQgb3V0CisgICAgICAgIG90aGVyd2lzZSB0byBhdm9pZCBhIHdhcm5pbmcuCisg ICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD04MDkyNgorCisg ICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgICogaGVhcC9NYWNo aW5lU3RhY2tNYXJrZXIuY3BwOgorICAgICAgICAoSlNDKToKKwogMjAxMi0wMy0xMiAgSG9qb25n IEhhbiAgPGhvam9uZy5oYW5Ac2Ftc3VuZy5jb20+CiAKICAgICAgICAgRml4IHRlc3QgY2FzZXMg Zm9yIFJlZ0V4cCBtdWx0aWxpbmUKSW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9oZWFwL01h Y2hpbmVTdGFja01hcmtlci5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291cmNlL0phdmFTY3JpcHRDb3Jl L2hlYXAvTWFjaGluZVN0YWNrTWFya2VyLmNwcAkocmV2aXNpb24gMTEwNTc5KQorKysgU291cmNl L0phdmFTY3JpcHRDb3JlL2hlYXAvTWFjaGluZVN0YWNrTWFya2VyLmNwcAkod29ya2luZyBjb3B5 KQpAQCAtOTYsNiArOTYsNyBAQCB0eXBlZGVmIEhBTkRMRSBQbGF0Zm9ybVRocmVhZDsKIHR5cGVk ZWYgcHRocmVhZF90IFBsYXRmb3JtVGhyZWFkOwogc3RhdGljIGNvbnN0IGludCBTaWdUaHJlYWRT dXNwZW5kUmVzdW1lID0gU0lHVVNSMjsKIAorI2lmIGRlZmluZWQoU0FfUkVTVEFSVCkKIHN0YXRp YyB2b2lkIHB0aHJlYWRTaWduYWxIYW5kbGVyU3VzcGVuZFJlc3VtZShpbnQgc2lnbm8pCiB7CiAg ICAgc2lnc2V0X3Qgc2lnbmFsU2V0OwpAQCAtMTA0LDYgKzEwNSw3IEBAIHN0YXRpYyB2b2lkIHB0 aHJlYWRTaWduYWxIYW5kbGVyU3VzcGVuZFIKICAgICBzaWdzdXNwZW5kKCZzaWduYWxTZXQpOwog fQogI2VuZGlmCisjZW5kaWYKIAogY2xhc3MgTWFjaGluZVRocmVhZHM6OlRocmVhZCB7CiBwdWJs aWM6Cg== 131678 2012-03-13 11:23:51 -0700 2012-03-13 14:42:11 -0700 Fix signed/unsigned mismatches 80790_1.patch text/plain 2772 staikos SW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKPT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291 cmNlL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkocmV2aXNpb24gMTEwMzk1KQorKysgU291cmNl L0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE3IEBA CisyMDEyLTAzLTExICBHZW9yZ2UgU3RhaWtvcyAgPHN0YWlrb3NAd2Via2l0Lm9yZz4KKworICAg ICAgICBGaXggc2lnbmVkL3Vuc2lnbmVkIG1pc21hdGNoIGNvbXBpbGVyIHdhcm5pbmdzLgorICAg ICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9ODA5MjYKKworICAg ICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICAqIGFzc2VtYmxlci9B c3NlbWJsZXJCdWZmZXJXaXRoQ29uc3RhbnRQb29sLmg6CisgICAgICAgIChKU0M6OkFzc2VtYmxl ckJ1ZmZlcldpdGhDb25zdGFudFBvb2w6OmZsdXNoV2l0aG91dEJhcnJpZXIpOgorICAgICAgICAq IGFzc2VtYmxlci9NYWNyb0Fzc2VtYmxlckFSTS5oOgorICAgICAgICAoSlNDOjpNYWNyb0Fzc2Vt YmxlckFSTTo6YnJhbmNoMzIpOgorICAgICAgICAqIHBhcnNlci9MZXhlci5jcHA6CisgICAgICAg IChKU0M6Ojo6cmVjb3JkMTYpOgorCiAyMDEyLTAzLTEwICBGaWxpcCBQaXpsbyAgPGZwaXpsb0Bh cHBsZS5jb20+CiAKICAgICAgICAgTExJbnQgc2hvdWxkIHN1cHBvcnQgSlNWQUxVRTY0CkluZGV4 OiBTb3VyY2UvSmF2YVNjcmlwdENvcmUvYXNzZW1ibGVyL0Fzc2VtYmxlckJ1ZmZlcldpdGhDb25z dGFudFBvb2wuaAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvSmF2YVNjcmlwdENvcmUvYXNzZW1ibGVy L0Fzc2VtYmxlckJ1ZmZlcldpdGhDb25zdGFudFBvb2wuaAkocmV2aXNpb24gMTEwMzkzKQorKysg U291cmNlL0phdmFTY3JpcHRDb3JlL2Fzc2VtYmxlci9Bc3NlbWJsZXJCdWZmZXJXaXRoQ29uc3Rh bnRQb29sLmgJKHdvcmtpbmcgY29weSkKQEAgLTIxNSw3ICsyMTUsNyBAQCBwdWJsaWM6CiAgICAg dm9pZCBmbHVzaFdpdGhvdXRCYXJyaWVyKGJvb2wgaXNGb3JjZWQgPSBmYWxzZSkKICAgICB7CiAg ICAgICAgIC8vIEZsdXNoIGlmIGNvbnN0YW50IHBvb2wgaXMgbW9yZSB0aGFuIDYwJSBmdWxsIHRv IGF2b2lkIG92ZXJ1c2Ugb2YgdGhpcyBmdW5jdGlvbi4KLSAgICAgICAgaWYgKGlzRm9yY2VkIHx8 IDUgKiBtX251bUNvbnN0cyA+IDMgKiBtYXhQb29sU2l6ZSAvIHNpemVvZih1aW50MzJfdCkpCisg ICAgICAgIGlmIChpc0ZvcmNlZCB8fCA1ICogdW5zaWduZWQobV9udW1Db25zdHMpID4gMyAqIG1h eFBvb2xTaXplIC8gc2l6ZW9mKHVpbnQzMl90KSkKICAgICAgICAgICAgIGZsdXNoQ29uc3RhbnRQ b29sKGZhbHNlKTsKICAgICB9CiAKSW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9hc3NlbWJs ZXIvTWFjcm9Bc3NlbWJsZXJBUk0uaAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvSmF2YVNjcmlwdENv cmUvYXNzZW1ibGVyL01hY3JvQXNzZW1ibGVyQVJNLmgJKHJldmlzaW9uIDExMDM5MykKKysrIFNv dXJjZS9KYXZhU2NyaXB0Q29yZS9hc3NlbWJsZXIvTWFjcm9Bc3NlbWJsZXJBUk0uaAkod29ya2lu ZyBjb3B5KQpAQCAtNDYwLDcgKzQ2MCw3IEBAIHB1YmxpYzoKICAgICAgICAgICAgIG1fYXNzZW1i bGVyLmxkcl91bl9pbW0oQVJNUmVnaXN0ZXJzOjpTMCwgcmlnaHQubV92YWx1ZSk7CiAgICAgICAg ICAgICBtX2Fzc2VtYmxlci5jbXBfcihsZWZ0LCBBUk1SZWdpc3RlcnM6OlMwKTsKICAgICAgICAg fSBlbHNlIHsKLSAgICAgICAgICAgIEFSTVdvcmQgdG1wID0gKHJpZ2h0Lm1fdmFsdWUgPT0gMHg4 MDAwMDAwMCkgPyBBUk1Bc3NlbWJsZXI6OklOVkFMSURfSU1NIDogbV9hc3NlbWJsZXIuZ2V0T3Ay KC1yaWdodC5tX3ZhbHVlKTsKKyAgICAgICAgICAgIEFSTVdvcmQgdG1wID0gKHJpZ2h0Lm1fdmFs dWUgPT0gaW50KDB4ODAwMDAwMDApKSA/IEFSTUFzc2VtYmxlcjo6SU5WQUxJRF9JTU0gOiBtX2Fz c2VtYmxlci5nZXRPcDIoLXJpZ2h0Lm1fdmFsdWUpOwogICAgICAgICAgICAgaWYgKHRtcCAhPSBB Uk1Bc3NlbWJsZXI6OklOVkFMSURfSU1NKQogICAgICAgICAgICAgICAgIG1fYXNzZW1ibGVyLmNt bl9yKGxlZnQsIHRtcCk7CiAgICAgICAgICAgICBlbHNlCkluZGV4OiBTb3VyY2UvSmF2YVNjcmlw dENvcmUvcGFyc2VyL0xleGVyLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvSmF2YVNjcmlwdENv cmUvcGFyc2VyL0xleGVyLmNwcAkocmV2aXNpb24gMTEwMzkzKQorKysgU291cmNlL0phdmFTY3Jp cHRDb3JlL3BhcnNlci9MZXhlci5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTYyOSw3ICs2MjksNyBA QCB0ZW1wbGF0ZSA8dHlwZW5hbWUgVD4KIGlubGluZSB2b2lkIExleGVyPFQ+OjpyZWNvcmQxNihp bnQgYykKIHsKICAgICBBU1NFUlQoYyA+PSAwKTsKLSAgICBBU1NFUlQoYyA8PSBVU0hSVF9NQVgp OworICAgIEFTU0VSVCh1bnNpZ25lZChjKSA8PSBVU0hSVF9NQVgpOwogICAgIG1fYnVmZmVyMTYu YXBwZW5kKHN0YXRpY19jYXN0PFVDaGFyPihjKSk7CiB9CiAK