114807 2013-04-18 04:16:55 -0700 Division by zero in CSSGradientValue::addStops() 2013-04-21 23:53:24 -0700 1 1 1 Unclassified WebKit CSS 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 cgarcia webkit-unassigned anilsson benjamin commit-queue dino esprehn+autocc macpherson menard mlattanzio rwlbuis tabatkins oldest_to_newest 877811 0 cgarcia 2013-04-18 04:16:55 -0700 This causes an assertion when running fast/forms/type-after-focus-rule-shrink-width.html in BlackBerry port. The test contains the following: input { width: 200px; background: -webkit-linear-gradient(left, white 110px, green 100px); } input:focus { width: 100px; } when the input is focused, the first offset of the gradient is outside the input. These offsest are normalized as 1.078431 and 0.980392 in CSSGradientValue::addStops(). But then there's a loop that checks if there are color stop with a position that is less than others specified in the list and setting them to the largest one. In our case we end up with two color stops at the same position 1.078431. At the end of the method, positions are normalized to the 0..1 scale in case of being out of the scale. Since both color stops are at the same position, its current scale is 0, so every offset is computed as a value divided by zero. So platform gradients are filled with two color stops at 'nan' position. The BlackBerry port requires that the first color stop is at position 0 and the last one at position 1, there's code to ensure that, but it checks whether offset[0] > 0 which is the case, and in platform code there's an ASSERT that checks offset[0] == 0 and it crashes. 877812 1 198718 cgarcia 2013-04-18 04:21:56 -0700 Created attachment 198718 Patch 877976 2 tabatkins 2013-04-18 10:53:42 -0700 Looks good to me! 878075 3 198718 benjamin 2013-04-18 14:07:43 -0700 Comment on attachment 198718 Patch This is common code, not blackberry code. How can this be tested? A division by zero should be reproducible on any platform. 878225 4 198718 dino 2013-04-18 22:06:43 -0700 Comment on attachment 198718 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=198718&action=review > Source/WebCore/ChangeLog:13 > + Fixes a crash in BlackBerry port running > + fast/forms/type-after-focus-rule-shrink-width.html. As Ben said, this isn't Blackberry code even if that's where it was crashing. The test you mention does have a linear gradient, but you should make one that is specific to this change. 878264 5 cgarcia 2013-04-18 23:45:33 -0700 (In reply to comment #3) > (From update of attachment 198718 [details]) > This is common code, not blackberry code. > How can this be tested? A division by zero should be reproducible on any platform. The problem is that for some reason the division by zero doesn't fail, but the offset is set to a 'nan' value. I don't know why, maybe it depends on the compiler. The thing is that platform gradients receive color stops at invalid positions. The fact that this crashes in the BlackBerry port is just coincidence, because it requires that gradients have a color stop at positions 0 and 1 and the platform code has an assert to check it. The code that ensures gradients are correctly passed to the platform doesn't deal correctly with those 'nan' values. I've checked that the same happens in cairo backend, but those values are passed directly to cairo that internally handles them. In any case, I'll see how we can test it. 878324 6 cgarcia 2013-04-19 02:15:51 -0700 hmm, I've been checking what other browsers do, and maybe we should match what others do, see http://plexode.com/eval3/#s=aekVQXANJVQMbAwEBHXgfPU+alQEPQ1BZAVycnZV5GwFKT02rRg5DTVBETBylpoOpEhERUVm2T6aVgqkUvL7AwgGGur2/t52Eur4BhwGxQrTJwosOU0ZRRkJVqU9Q3uDiVduVXpq3D61P4lMSo9CV3UpOQkhGVIOLBGIZMJRVBzyegOJBTIRFJRGJ5VBJ/BoBLBJgxGAITXa+i0MI0RjjHLIKdidfT8f0AJxQLcIVxCKcLhsPiMTisXjJVjcdZ0gb8jCclk74AMqfr/AIOJxUmLzmc1h0QiUUi0YjUcjxZoMiYNEZ1GYKddybsiVeMyKYTe9olIOfdKgECgkGqEKhlUnFXndaClcr0jY1ilD5uMrpdNl94qV6m9WnVZnoBwFAAMhwdFw1IxFzplOxs0x9VnNYnkby0fzFCYOEX1jU1mUwdCCYaDQDpyAJ3Dw9AcJmb2UBZ34f2pyTNv3W833Atdt4nG5HKAYNAZQJJLSQDCIDLwAA In this example, both ff and ie10 render the second box with no background, so they ignore color stops outside the scale. The result in webkit depends on every port, since we are passing nan values to the platform it depends on how every platform deals with those values. With cairo backend I get a solid #80c080 background while with BlackBerry I get a solid green background. 878365 7 198841 cgarcia 2013-04-19 04:17:45 -0700 Created attachment 198841 Updated patch It includes a ref test now and the behaviour matches what other browsers do. 878622 8 tabatkins 2013-04-19 13:03:18 -0700 (In reply to comment #6) > hmm, I've been checking what other browsers do, and maybe we should match what others do, see > > http://plexode.com/eval3/#s=aekVQXANJVQMbAwEBHXgfPU+alQEPQ1BZAVycnZV5GwFKT02rRg5DTVBETBylpoOpEhERUVm2T6aVgqkUvL7AwgGGur2/t52Eur4BhwGxQrTJwosOU0ZRRkJVqU9Q3uDiVduVXpq3D61P4lMSo9CV3UpOQkhGVIOLBGIZMJRVBzyegOJBTIRFJRGJ5VBJ/BoBLBJgxGAITXa+i0MI0RjjHLIKdidfT8f0AJxQLcIVxCKcLhsPiMTisXjJVjcdZ0gb8jCclk74AMqfr/AIOJxUmLzmc1h0QiUUi0YjUcjxZoMiYNEZ1GYKddybsiVeMyKYTe9olIOfdKgECgkGqEKhlUnFXndaClcr0jY1ilD5uMrpdNl94qV6m9WnVZnoBwFAAMhwdFw1IxFzplOxs0x9VnNYnkby0fzFCYOEX1jU1mUwdCCYaDQDpyAJ3Dw9AcJmb2UBZ34f2pyTNv3W833Atdt4nG5HKAYNAZQJJLSQDCIDLwAA > > In this example, both ff and ie10 render the second box with no background, so they ignore color stops outside the scale. The result in webkit depends on every port, since we are passing nan values to the platform it depends on how every platform deals with those values. With cairo backend I get a solid #80c080 background while with BlackBerry I get a solid green background. FF and IE aren't rendering with no background, they're correctly handling the color-stops (correcting the second one's position to 310px), and then rendering the resulting gradient. It just so happens that, with the specific gradient you've constructed, the visible area is completely white. If you changed it to use a non-white color as the first color-stop, you'd see that they definitely do render things correctly. 878730 9 cgarcia 2013-04-20 00:32:36 -0700 (In reply to comment #8) > (In reply to comment #6) > > hmm, I've been checking what other browsers do, and maybe we should match what others do, see > > > > http://plexode.com/eval3/#s=aekVQXANJVQMbAwEBHXgfPU+alQEPQ1BZAVycnZV5GwFKT02rRg5DTVBETBylpoOpEhERUVm2T6aVgqkUvL7AwgGGur2/t52Eur4BhwGxQrTJwosOU0ZRRkJVqU9Q3uDiVduVXpq3D61P4lMSo9CV3UpOQkhGVIOLBGIZMJRVBzyegOJBTIRFJRGJ5VBJ/BoBLBJgxGAITXa+i0MI0RjjHLIKdidfT8f0AJxQLcIVxCKcLhsPiMTisXjJVjcdZ0gb8jCclk74AMqfr/AIOJxUmLzmc1h0QiUUi0YjUcjxZoMiYNEZ1GYKddybsiVeMyKYTe9olIOfdKgECgkGqEKhlUnFXndaClcr0jY1ilD5uMrpdNl94qV6m9WnVZnoBwFAAMhwdFw1IxFzplOxs0x9VnNYnkby0fzFCYOEX1jU1mUwdCCYaDQDpyAJ3Dw9AcJmb2UBZ34f2pyTNv3W833Atdt4nG5HKAYNAZQJJLSQDCIDLwAA > > > > In this example, both ff and ie10 render the second box with no background, so they ignore color stops outside the scale. The result in webkit depends on every port, since we are passing nan values to the platform it depends on how every platform deals with those values. With cairo backend I get a solid #80c080 background while with BlackBerry I get a solid green background. > > FF and IE aren't rendering with no background, they're correctly handling the color-stops (correcting the second one's position to 310px), and then rendering the resulting gradient. It just so happens that, with the specific gradient you've constructed, the visible area is completely white. > > If you changed it to use a non-white color as the first color-stop, you'd see that they definitely do render things correctly. Yes, this is exactly with my patch does, with the patch applied, changing the color to red instead of white makes the second gradient to be rendered solid red, at least with cairo backend. 878799 10 198841 dino 2013-04-20 14:07:31 -0700 Comment on attachment 198841 Updated patch View in context: https://bugs.webkit.org/attachment.cgi?id=198841&action=review > Source/WebCore/css/CSSGradientValue.cpp:347 > - for (size_t i = 0; i < numStops; ++i) > - stops[i].offset = (stops[i].offset - firstOffset) / scale; > + for (size_t i = 0; i < numStops; ++i) > + stops[i].offset = (stops[i].offset - firstOffset) / scale; I'm seeing darker green in the diff here. Are they tab characters? Maybe it's just highlighting the extra indentation. 878984 11 198841 cgarcia 2013-04-21 23:24:43 -0700 Comment on attachment 198841 Updated patch View in context: https://bugs.webkit.org/attachment.cgi?id=198841&action=review Thanks for the review. >> Source/WebCore/css/CSSGradientValue.cpp:347 >> + stops[i].offset = (stops[i].offset - firstOffset) / scale; > > I'm seeing darker green in the diff here. Are they tab characters? Maybe it's just highlighting the extra indentation. I don't know what it means, but there are not tabs there, I've just checked it. 878986 12 198841 commit-queue 2013-04-21 23:53:21 -0700 Comment on attachment 198841 Updated patch Clearing flags on attachment: 198841 Committed r148859: <http://trac.webkit.org/changeset/148859> 878987 13 commit-queue 2013-04-21 23:53:24 -0700 All reviewed patches have been landed. Closing bug. 198718 2013-04-18 04:21:56 -0700 2013-04-19 04:17:45 -0700 Patch css-gradient-normalization.diff text/plain 1728 cgarcia ZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZyBiL1NvdXJjZS9XZWJDb3JlL0No YW5nZUxvZwppbmRleCBjMzVjODhmLi45MmI2NmU2IDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29y ZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMjEg QEAKKzIwMTMtMDQtMTggIENhcmxvcyBHYXJjaWEgQ2FtcG9zICA8Y2dhcmNpYUBpZ2FsaWEuY29t PgorCisgICAgICAgIERpdmlzaW9uIGJ5IHplcm8gaW4gQ1NTR3JhZGllbnRWYWx1ZTo6YWRkU3Rv cHMoKQorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTE0 ODA3CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgVGhl IG5vcm1hbGl6YXRpb24gb2YgdGhlIGNvbG9yIHN0b3AgcG9zaXRpb25zIG9mIGEgbGluZWFyIGdy YWRpZW50CisgICAgICAgIGRvZXNuJ3QgdGFrZSBpbnRvIGFjY291bnQgdGhhdCB0aGUgZmlyc3Qg YW5kIGxhc3QgcG9zaXRpb24gY2FuIGJlCisgICAgICAgIHRoZSBzYW1lLiBJbiBzdWNoIGNhc2Ug dGhlIHBvc2l0aW9ucyBhcmUgY29tcHV0ZWQgZGl2aWRpbmcgYnkgMC4KKworICAgICAgICBGaXhl cyBhIGNyYXNoIGluIEJsYWNrQmVycnkgcG9ydCBydW5uaW5nCisgICAgICAgIGZhc3QvZm9ybXMv dHlwZS1hZnRlci1mb2N1cy1ydWxlLXNocmluay13aWR0aC5odG1sLgorCisgICAgICAgICogY3Nz L0NTU0dyYWRpZW50VmFsdWUuY3BwOgorICAgICAgICAoV2ViQ29yZTo6Q1NTR3JhZGllbnRWYWx1 ZTo6YWRkU3RvcHMpOiBDbGFtcCB0aGUgcG9zaXRpb24gdG8gMCBpZgorICAgICAgICB0aGUgc2Nh bGUgaXMgMCB3aGVuIG5vcm1hbGl6aW5nIHRoZSBjb2xvciBzdG9wIHBvc2l0aW9ucy4KKwogMjAx My0wNC0xNyAgSm9obiBHcmlnZ3MgIDxqZ3JpZ2dzQGJsYWNrYmVycnkuY29tPgogCiAgICAgICAg IFtCbGFja0JlcnJ5XSBBZGQgc3VwcG9ydCBmb3IgZmlsZXN5c3RlbTogVVJMcyB0byBCbGFja0Jl cnJ5IE1lZGlhIFBsYXllci4KZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL2Nzcy9DU1NHcmFk aWVudFZhbHVlLmNwcCBiL1NvdXJjZS9XZWJDb3JlL2Nzcy9DU1NHcmFkaWVudFZhbHVlLmNwcApp bmRleCA4ZjE2OGFiLi42YTM1ZWYzIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29yZS9jc3MvQ1NT R3JhZGllbnRWYWx1ZS5jcHAKKysrIGIvU291cmNlL1dlYkNvcmUvY3NzL0NTU0dyYWRpZW50VmFs dWUuY3BwCkBAIC0zNDMsNyArMzQzLDcgQEAgdm9pZCBDU1NHcmFkaWVudFZhbHVlOjphZGRTdG9w cyhHcmFkaWVudCogZ3JhZGllbnQsIFJlbmRlck9iamVjdCogcmVuZGVyZXIsIFJlbmQKICAgICAg ICAgICAgIGZsb2F0IHNjYWxlID0gbGFzdE9mZnNldCAtIGZpcnN0T2Zmc2V0OwogCiAgICAgICAg ICAgICBmb3IgKHNpemVfdCBpID0gMDsgaSA8IG51bVN0b3BzOyArK2kpCi0gICAgICAgICAgICAg ICAgc3RvcHNbaV0ub2Zmc2V0ID0gKHN0b3BzW2ldLm9mZnNldCAtIGZpcnN0T2Zmc2V0KSAvIHNj YWxlOworICAgICAgICAgICAgICAgIHN0b3BzW2ldLm9mZnNldCA9IHNjYWxlID4gMCA/IChzdG9w c1tpXS5vZmZzZXQgLSBmaXJzdE9mZnNldCkgLyBzY2FsZSA6IDA7CiAKICAgICAgICAgICAgIEZs b2F0UG9pbnQgcDAgPSBncmFkaWVudC0+cDAoKTsKICAgICAgICAgICAgIEZsb2F0UG9pbnQgcDEg PSBncmFkaWVudC0+cDEoKTsK 198841 2013-04-19 04:17:45 -0700 2013-04-21 23:53:21 -0700 Updated patch css-gradient-stop-invalid.diff text/plain 5175 cgarcia ZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL0NoYW5nZUxvZyBiL0xheW91dFRlc3RzL0NoYW5nZUxv ZwppbmRleCBmYjI2MzgxLi43OWZmODNjIDEwMDY0NAotLS0gYS9MYXlvdXRUZXN0cy9DaGFuZ2VM b2cKKysrIGIvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTMgQEAKKzIwMTMtMDQt MTkgIENhcmxvcyBHYXJjaWEgQ2FtcG9zICA8Y2dhcmNpYUBpZ2FsaWEuY29tPgorCisgICAgICAg IERpdmlzaW9uIGJ5IHplcm8gaW4gQ1NTR3JhZGllbnRWYWx1ZTo6YWRkU3RvcHMoKQorICAgICAg ICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTE0ODA3CisKKyAgICAg ICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgKiBmYXN0L2dyYWRpZW50 cy9jc3MzLWNvbG9yLXN0b3AtaW52YWxpZC1leHBlY3RlZC5odG1sOiBBZGRlZC4KKyAgICAgICAg KiBmYXN0L2dyYWRpZW50cy9jc3MzLWNvbG9yLXN0b3AtaW52YWxpZC5odG1sOiBBZGRlZC4KKwog MjAxMy0wNC0xOCAgTWFydGluIFJvYmluc29uICA8bXJvYmluc29uQGlnYWxpYS5jb20+CiAKICAg ICAgICAgW0dUS10gZmFzdC9jYW52YXMvRHJhd0ltYWdlU2luZ2xlUGl4ZWxTdHJldGNoLmh0bWwg ZmFpbHMKZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZyBiL1NvdXJjZS9XZWJD b3JlL0NoYW5nZUxvZwppbmRleCBiNmY2ZTJhLi5hMmUxMzhlIDEwMDY0NAotLS0gYS9Tb3VyY2Uv V2ViQ29yZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCkBAIC0xLDMg KzEsMjQgQEAKKzIwMTMtMDQtMTkgIENhcmxvcyBHYXJjaWEgQ2FtcG9zICA8Y2dhcmNpYUBpZ2Fs aWEuY29tPgorCisgICAgICAgIERpdmlzaW9uIGJ5IHplcm8gaW4gQ1NTR3JhZGllbnRWYWx1ZTo6 YWRkU3RvcHMoKQorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/ aWQ9MTE0ODA3CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAg ICAgVGhlIG5vcm1hbGl6YXRpb24gb2YgdGhlIGNvbG9yIHN0b3AgcG9zaXRpb25zIG9mIGEgbGlu ZWFyIGdyYWRpZW50CisgICAgICAgIGRvZXNuJ3QgdGFrZSBpbnRvIGFjY291bnQgdGhhdCB0aGUg Zmlyc3QgYW5kIGxhc3QgcG9zaXRpb24gY2FuIGJlCisgICAgICAgIHRoZSBzYW1lLiBJbiBzdWNo IGNhc2UgdGhlIHBvc2l0aW9ucyBhcmUgY29tcHV0ZWQgZGl2aWRpbmcgYnkgMC4KKyAgICAgICAg Q2xhbXAgdGhlIHBvc2l0aW9ucyB0byAxIHdpdGhvdXQgbW92aW5nIHRoZSBlbmQgcG9pbnRzIGlu IHN1Y2gKKyAgICAgICAgY2FzZSB0byBtYXRjaCB3aGF0IG90aGVyIGJyb3dzZXJzIGRvLiBUaGlz IGVuc3VyZXMgdGhhdCBwb3NpdGlvbnMKKyAgICAgICAgcGFzc2VkIHRvIHRoZSBwbGF0Zm9ybSBn cmFkaWVudCBhcmUgaW4gdGhlIDAtMSByYW5nZSB3aGljaCBmaXhlcyBhCisgICAgICAgIGNyYXNo IGR1ZSB0byBhbiBhc3NlcnQgaW4gQmxhY2tCZXJyeSBwb3J0IHJ1bm5pbmcgdGVzdAorICAgICAg ICBmYXN0L2Zvcm1zL3R5cGUtYWZ0ZXItZm9jdXMtcnVsZS1zaHJpbmstd2lkdGguaHRtbC4KKwor ICAgICAgICBUZXN0OiBmYXN0L2dyYWRpZW50cy9jc3MzLWNvbG9yLXN0b3AtaW52YWxpZC5odG1s CisKKyAgICAgICAgKiBjc3MvQ1NTR3JhZGllbnRWYWx1ZS5jcHA6CisgICAgICAgIChXZWJDb3Jl OjpDU1NHcmFkaWVudFZhbHVlOjphZGRTdG9wcyk6CisKIDIwMTMtMDQtMTcgIEFuZGVycyBDYXJs c3NvbiAgPGFuZGVyc2NhQGFwcGxlLmNvbT4KIAogICAgICAgICBDaGFuZ2Ugc3RvcmFnZSBmYWN0 b3J5IGZ1bmN0aW9ucyB0byB0YWtlIGEgUGFnZUdyb3VwIGFuZCBQYWdlIHJlc3BlY3RpdmVseQpk aWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvZmFzdC9ncmFkaWVudHMvY3NzMy1jb2xvci1zdG9wLWlu dmFsaWQtZXhwZWN0ZWQuaHRtbCBiL0xheW91dFRlc3RzL2Zhc3QvZ3JhZGllbnRzL2NzczMtY29s b3Itc3RvcC1pbnZhbGlkLWV4cGVjdGVkLmh0bWwKbmV3IGZpbGUgbW9kZSAxMDA2NDQKaW5kZXgg MDAwMDAwMC4uYjM0NmE0MQotLS0gL2Rldi9udWxsCisrKyBiL0xheW91dFRlc3RzL2Zhc3QvZ3Jh ZGllbnRzL2NzczMtY29sb3Itc3RvcC1pbnZhbGlkLWV4cGVjdGVkLmh0bWwKQEAgLTAsMCArMSwx NCBAQAorPCFET0NUWVBFIGh0bWw+Cis8c3R5bGU+CisuYm94IHsKKyAgICBkaXNwbGF5OiBpbmxp bmUtYmxvY2s7CisgICAgaGVpZ2h0OiAxMDBweDsKKyAgICB3aWR0aDogMzAwcHg7CisgICAgbWFy Z2luOiAxMHB4OworICAgIGJvcmRlcjogMXB4IHNvbGlkIGJsYWNrOworICAgIGJhY2tncm91bmQt cmVwZWF0OiBuby1yZXBlYXQ7Cit9Cis8L3N0eWxlPgorPGgyPlRoaXMgc2hvdWxkIGJlIGFuIGVt cHR5IGJveDwvaDI+Cis8ZGl2IGNsYXNzPSJib3giPjwvZGl2PgorCmRpZmYgLS1naXQgYS9MYXlv dXRUZXN0cy9mYXN0L2dyYWRpZW50cy9jc3MzLWNvbG9yLXN0b3AtaW52YWxpZC5odG1sIGIvTGF5 b3V0VGVzdHMvZmFzdC9ncmFkaWVudHMvY3NzMy1jb2xvci1zdG9wLWludmFsaWQuaHRtbApuZXcg ZmlsZSBtb2RlIDEwMDY0NAppbmRleCAwMDAwMDAwLi5lNjZkN2I4Ci0tLSAvZGV2L251bGwKKysr IGIvTGF5b3V0VGVzdHMvZmFzdC9ncmFkaWVudHMvY3NzMy1jb2xvci1zdG9wLWludmFsaWQuaHRt bApAQCAtMCwwICsxLDE5IEBACis8IURPQ1RZUEUgaHRtbD4KKzxzdHlsZT4KKy5ib3ggeworICAg IGRpc3BsYXk6IGlubGluZS1ibG9jazsKKyAgICBoZWlnaHQ6IDEwMHB4OworICAgIHdpZHRoOiAz MDBweDsKKyAgICBtYXJnaW46IDEwcHg7CisgICAgYm9yZGVyOiAxcHggc29saWQgYmxhY2s7Cisg ICAgYmFja2dyb3VuZC1yZXBlYXQ6IG5vLXJlcGVhdDsKK30KKworLmxpbmVhciB7CisgICAgYmFj a2dyb3VuZC1pbWFnZTogLXdlYmtpdC1saW5lYXItZ3JhZGllbnQobGVmdCwgd2hpdGUgMzEwcHgs IGdyZWVuIDMwMHB4KTsKKyAgICBiYWNrZ3JvdW5kLWltYWdlOiAtbW96LWxpbmVhci1ncmFkaWVu dChsZWZ0LCB3aGl0ZSAzMTBweCwgZ3JlZW4gMzAwcHgpOworfQorPC9zdHlsZT4KKzxoMj5UaGlz IHNob3VsZCBiZSBhbiBlbXB0eSBib3g8L2gyPgorPGRpdiBjbGFzcz0ibGluZWFyIGJveCI+PC9k aXY+CisKZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL2Nzcy9DU1NHcmFkaWVudFZhbHVlLmNw cCBiL1NvdXJjZS9XZWJDb3JlL2Nzcy9DU1NHcmFkaWVudFZhbHVlLmNwcAppbmRleCA4ZjE2OGFi Li4yNTdlYjMzIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29yZS9jc3MvQ1NTR3JhZGllbnRWYWx1 ZS5jcHAKKysrIGIvU291cmNlL1dlYkNvcmUvY3NzL0NTU0dyYWRpZW50VmFsdWUuY3BwCkBAIC0z NDAsMTUgKzM0MCwyMSBAQCB2b2lkIENTU0dyYWRpZW50VmFsdWU6OmFkZFN0b3BzKEdyYWRpZW50 KiBncmFkaWVudCwgUmVuZGVyT2JqZWN0KiByZW5kZXJlciwgUmVuZAogICAgICAgICBpZiAoaXNM aW5lYXJHcmFkaWVudCgpKSB7CiAgICAgICAgICAgICBmbG9hdCBmaXJzdE9mZnNldCA9IHN0b3Bz WzBdLm9mZnNldDsKICAgICAgICAgICAgIGZsb2F0IGxhc3RPZmZzZXQgPSBzdG9wc1tudW1TdG9w cyAtIDFdLm9mZnNldDsKLSAgICAgICAgICAgIGZsb2F0IHNjYWxlID0gbGFzdE9mZnNldCAtIGZp cnN0T2Zmc2V0OworICAgICAgICAgICAgaWYgKGZpcnN0T2Zmc2V0ICE9IGxhc3RPZmZzZXQpIHsK KyAgICAgICAgICAgICAgICBmbG9hdCBzY2FsZSA9IGxhc3RPZmZzZXQgLSBmaXJzdE9mZnNldDsK IAotICAgICAgICAgICAgZm9yIChzaXplX3QgaSA9IDA7IGkgPCBudW1TdG9wczsgKytpKQotICAg ICAgICAgICAgICAgIHN0b3BzW2ldLm9mZnNldCA9IChzdG9wc1tpXS5vZmZzZXQgLSBmaXJzdE9m ZnNldCkgLyBzY2FsZTsKKyAgICAgICAgICAgICAgICBmb3IgKHNpemVfdCBpID0gMDsgaSA8IG51 bVN0b3BzOyArK2kpCisgICAgICAgICAgICAgICAgICAgIHN0b3BzW2ldLm9mZnNldCA9IChzdG9w c1tpXS5vZmZzZXQgLSBmaXJzdE9mZnNldCkgLyBzY2FsZTsKIAotICAgICAgICAgICAgRmxvYXRQ b2ludCBwMCA9IGdyYWRpZW50LT5wMCgpOwotICAgICAgICAgICAgRmxvYXRQb2ludCBwMSA9IGdy YWRpZW50LT5wMSgpOwotICAgICAgICAgICAgZ3JhZGllbnQtPnNldFAwKEZsb2F0UG9pbnQocDAu eCgpICsgZmlyc3RPZmZzZXQgKiAocDEueCgpIC0gcDAueCgpKSwgcDAueSgpICsgZmlyc3RPZmZz ZXQgKiAocDEueSgpIC0gcDAueSgpKSkpOwotICAgICAgICAgICAgZ3JhZGllbnQtPnNldFAxKEZs b2F0UG9pbnQocDEueCgpICsgKGxhc3RPZmZzZXQgLSAxKSAqIChwMS54KCkgLSBwMC54KCkpLCBw MS55KCkgKyAobGFzdE9mZnNldCAtIDEpICogKHAxLnkoKSAtIHAwLnkoKSkpKTsKKyAgICAgICAg ICAgICAgICBGbG9hdFBvaW50IHAwID0gZ3JhZGllbnQtPnAwKCk7CisgICAgICAgICAgICAgICAg RmxvYXRQb2ludCBwMSA9IGdyYWRpZW50LT5wMSgpOworICAgICAgICAgICAgICAgIGdyYWRpZW50 LT5zZXRQMChGbG9hdFBvaW50KHAwLngoKSArIGZpcnN0T2Zmc2V0ICogKHAxLngoKSAtIHAwLngo KSksIHAwLnkoKSArIGZpcnN0T2Zmc2V0ICogKHAxLnkoKSAtIHAwLnkoKSkpKTsKKyAgICAgICAg ICAgICAgICBncmFkaWVudC0+c2V0UDEoRmxvYXRQb2ludChwMS54KCkgKyAobGFzdE9mZnNldCAt IDEpICogKHAxLngoKSAtIHAwLngoKSksIHAxLnkoKSArIChsYXN0T2Zmc2V0IC0gMSkgKiAocDEu eSgpIC0gcDAueSgpKSkpOworICAgICAgICAgICAgfSBlbHNlIHsKKyAgICAgICAgICAgICAgICAv LyBUaGVyZSdzIGEgc2luZ2xlIHBvc2l0aW9uIHRoYXQgaXMgb3V0c2lkZSB0aGUgc2NhbGUsIGNs YW1wIHRoZSBwb3NpdGlvbnMgdG8gMS4KKyAgICAgICAgICAgICAgICBmb3IgKHNpemVfdCBpID0g MDsgaSA8IG51bVN0b3BzOyArK2kpCisgICAgICAgICAgICAgICAgICAgIHN0b3BzW2ldLm9mZnNl dCA9IDE7CisgICAgICAgICAgICB9CiAgICAgICAgIH0gZWxzZSBpZiAoaXNSYWRpYWxHcmFkaWVu dCgpKSB7CiAgICAgICAgICAgICAvLyBSYXRoZXIgdGhhbiBzY2FsaW5nIHRoZSBwb2ludHMgPCAw LCB3ZSB0cnVuY2F0ZSB0aGVtLCBzbyBvbmx5IHNjYWxlIGFjY29yZGluZyB0byB0aGUgbGFyZ2Vz dCBwb2ludC4KICAgICAgICAgICAgIGZsb2F0IGZpcnN0T2Zmc2V0ID0gMDsK