78038 2012-02-07 14:01:30 -0800 [Chromium] Assertion failure minX <= maxX in Region.cpp 2012-03-14 11:33:33 -0700 1 1 1 Unclassified WebKit Layout and Rendering 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 jchaffraix danakj andersca cc-bugs danakj enne jamesr webkit.review.bot oldest_to_newest 551353 0 jchaffraix 2012-02-07 14:01:30 -0800 The ASSERT is in Region.cpp - line 214: ASSERT(minX <= maxX); It seems to be hit frequently by our bots on test following test: compositing/iframes/invisible-nested-iframe-show.html compositing/iframes/layout-on-compositing-change.html I don't know if it's a Chromium only bug. 551365 1 jamesr 2012-02-07 14:13:02 -0800 Can we get a full stack or link to failing runs? 551393 2 andersca 2012-02-07 14:24:32 -0800 It would also be helpful if we could figure out which Region operations lead up to this invalid region being created. 551399 3 danakj 2012-02-07 14:34:42 -0800 I managed to get this running invisible-nested-iframe-show.html enough times (7 in a row seems to be a magic number): [1522:1522:530403859202:ERROR:process_util_posix.cc(142)] Received signal 11 base::debug::StackTrace::StackTrace() [0x7fe70a800c06] base::(anonymous namespace)::StackDumpSignalHandler() [0x7fe70a85e55d] 0x7fe6ffd07af0 WebCore::Region::Shape::bounds() [0x7fe70660409f] WebCore::Region::subtract() [0x7fe706604595] WebCore::TiledLayerChromium::updateBounds() [0x7fe70663b354] When I print out the two rects going into Regions in updateBounds() I see the following: old: 0 0 new: 800 600 old: 880 1879 new: 0 0 old: 0 0 new: 15 600 old: 0 0 new: 800 600 old: 0 0 new: 800 600 old: 0 0 new: 800 600 old: 0 0 new: 210 210 old: 0 0 new: 250 170 old: 0 0 new: 250 230 old: 0 0 new: 250 230 old: 0 0 new: 210 210 old: 0 0 new: 260 193 old: 0 0 new: -15 15 -15 isn't a good size width. contentBounds() is wrong here in this case. 551406 4 danakj 2012-02-07 14:41:53 -0800 Just in case it's of interest and others have difficulty reproing, here is the full backtrace. Summary: Region seems okay, though asserting sooner on negative widths might be useful. [3526:3526:531217440850:ERROR:process_util_posix.cc(142)] Received signal 11 base::debug::StackTrace::StackTrace() [0x7f6fd8a6bc06] base::(anonymous namespace)::StackDumpSignalHandler() [0x7f6fd8ac955d] 0x7f6fcdf71af0 WebCore::Region::Shape::bounds() [0x7f6fd486f30f] WebCore::Region::subtract() [0x7f6fd486f805] WebCore::TiledLayerChromium::updateBounds() [0x7f6fd48a662f] WebCore::TiledLayerChromium::invalidateRect() [0x7f6fd48a75c7] WebCore::TiledLayerChromium::setNeedsDisplayRect() [0x7f6fd48a753d] WebCore::LayerChromium::setNeedsDisplay() [0x7f6fd3dfb03c] WebCore::LayerChromium::setBounds() [0x7f6fd4887302] WebCore::GraphicsLayerChromium::updateLayerSize() [0x7f6fd4883d58] WebCore::GraphicsLayerChromium::setSize() [0x7f6fd4882e24] WebCore::positionScrollbarLayer() [0x7f6fd47df99a] WebCore::ScrollView::positionScrollbarLayers() [0x7f6fd47dfc00] WebCore::RenderLayerCompositor::updateOverflowControlsLayers() [0x7f6fd424d7ef] WebCore::RenderLayerCompositor::frameViewDidChangeSize() [0x7f6fd424ae0d] WebCore::RenderLayerCompositor::ensureRootLayer() [0x7f6fd424dd68] WebCore::RenderLayerCompositor::enableCompositingMode() [0x7f6fd4247bc5] WebCore::RenderLayerCompositor::updateBacking() [0x7f6fd42485e4] WebCore::RenderLayerCompositor::computeCompositingRequirements() [0x7f6fd424a416] WebCore::RenderLayerCompositor::computeCompositingRequirements() [0x7f6fd4249f63] WebCore::RenderLayerCompositor::computeCompositingRequirements() [0x7f6fd424a00a] WebCore::RenderLayerCompositor::updateCompositingLayers() [0x7f6fd42483a9] WebCore::FrameView::updateCompositingLayers() [0x7f6fd4e2753a] WebCore::FrameView::layout() [0x7f6fd4e28c5f] WebCore::FrameView::layoutTimerFired() [0x7f6fd4e2bb99] WebCore::Timer<>::fired() [0x7f6fd4e35728] WebCore::ThreadTimers::sharedTimerFiredInternal() [0x7f6fd47ede30] WebCore::ThreadTimers::sharedTimerFired() [0x7f6fd47edd67] webkit_glue::WebKitPlatformSupportImpl::DoTimeout() [0x7f6fd8eae30c] base::BaseTimer<>::TimerTask::Run() [0x7f6fd8eaeb7f] base::internal::RunnableAdapter<>::Run() [0x7f6fd8b07f61] base::internal::InvokeHelper<>::MakeItSo() [0x7f6fd8b07ee5] base::internal::Invoker<>::Run() [0x7f6fd8b07e91] base::Callback<>::Run() [0x7f6fd8a64267] MessageLoop::RunTask() [0x7f6fd8aa0710] MessageLoop::DeferOrRunPendingTask() [0x7f6fd8aa0827] MessageLoop::DoWork() [0x7f6fd8aa1049] base::MessagePumpGlib::RunWithDispatcher() [0x7f6fd8a47738] base::MessagePumpGlib::Run() [0x7f6fd8a47b18] MessageLoop::RunInternal() [0x7f6fd8aa03cb] MessageLoop::RunHandler() [0x7f6fd8aa027e] MessageLoop::Run() [0x7f6fd8a9fbb3] webkit_support::RunMessageLoop() [0x57299c] TestShell::waitTestFinished() [0x53826e] TestShell::runFileTest() [0x5308f9] runTest() [0x4fa214] main [0x4facd6] 0x7f6fcdf5cc4d 0x4e6569 551423 5 jchaffraix 2012-02-07 14:52:41 -0800 FYI, the traces on the bot are the same as in comment #4. 551477 6 danakj 2012-02-07 15:39:05 -0800 When I run ntyimes in a debugger I end up seeing this assertion as well. In this case the size in RenderLayerBacking::m_compositedBounds is bad: (gdb) print m_compositedBounds $7 = {m_location = {m_x = 0, m_y = 0}, m_size = {m_width = -180, m_height = 930}} And this comes all the way in through the GraphicsLayer to cause this ASSERT. 553488 7 jamesr 2012-02-09 14:07:27 -0800 We should clamp the rects somewhere or revert the change to pipe these rects into regions, this is causing an unacceptable number of ASSERT failures on the bots and with interactive debugging. 553490 8 danakj 2012-02-09 14:14:12 -0800 Would it be wrong to clamp them in GraphicsLayerChromium? What do you think about the fact this hides the problem in WebCore for others to discover somehow? 553493 9 jamesr 2012-02-09 14:17:24 -0800 (In reply to comment #8) > Would it be wrong to clamp them in GraphicsLayerChromium? I don't know. If in doubt, we should do something conservative (i.e. doesn't break rendering). > What do you think about the fact this hides the problem in WebCore for others to discover somehow? I don't see how making tests crash and debugging impossible helps anyone. 553509 10 126374 danakj 2012-02-09 14:30:33 -0800 Created attachment 126374 Patch 553513 11 126374 jamesr 2012-02-09 14:37:04 -0800 Comment on attachment 126374 Patch I think checking for size.isEmpty() and setting size to (0, 0) would be slightly preferable - I'm not sure I want to "grow" a -50,50 layer to 0,50, or if that'll have any effect further down the pipe. 553521 12 126377 danakj 2012-02-09 14:44:30 -0800 Created attachment 126377 Patch 553525 13 126377 danakj 2012-02-09 14:48:33 -0800 Comment on attachment 126377 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=126377&action=review > Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:179 > + // avoid assertions in the compositor. ill fix that redundant comment before landing 553527 14 126377 jamesr 2012-02-09 14:53:00 -0800 Comment on attachment 126377 Patch OK thanks, I think this will work for now. Can you file a bug to investigate if we need to do anything more on the RenderLayerBacking or earlier phases? I have some theories but haven't investigated them on these particular test cases. 553533 15 danakj 2012-02-09 14:59:32 -0800 http://code.google.com/p/chromium/issues/detail?id=113534 553570 16 126389 fsamuel 2012-02-09 15:32:47 -0800 Created attachment 126389 Patch for landing 553827 17 126389 webkit.review.bot 2012-02-09 22:08:54 -0800 Comment on attachment 126389 Patch for landing Clearing flags on attachment: 126389 Committed r107360: <http://trac.webkit.org/changeset/107360> 553828 18 webkit.review.bot 2012-02-09 22:08:59 -0800 All reviewed patches have been landed. Closing bug. 578596 19 enne 2012-03-14 11:33:33 -0700 Committed r110719: <http://trac.webkit.org/changeset/110719> 126374 2012-02-09 14:30:33 -0800 2012-02-09 14:44:26 -0800 Patch bug-78038-20120209173031.patch text/plain 1866 danakj U3VidmVyc2lvbiBSZXZpc2lvbjogMTA3MjYwCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggNjM5ZDcxNTliY2JhNTc1 MzA5Y2ZmZDBlZTVkMjNlMTViMGY1NDQwOC4uYjNhZTRjNzk3ZGU3NzIzMmUyMjgzYmVmZDYwM2Y3 YzFiNzRkYzliOCAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE3IEBACisyMDEyLTAyLTA5ICBEYW5h IEphbnNlbnMgIDxkYW5ha2pAY2hyb21pdW0ub3JnPgorCisgICAgICAgIFtDaHJvbWl1bV0gQXNz ZXJ0aW9uIGZhaWx1cmUgbWluWCA8PSBtYXhYIGluIFJlZ2lvbi5jcHAKKyAgICAgICAgaHR0cHM6 Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTc4MDM4CisKKyAgICAgICAgUmV2aWV3 ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgQ292ZXJlZCBieSBleGlzdGluZyB0ZXN0 cyAoc2hvdWxkIG1ha2UgdGhlbSBzdG9wIGFzc2VydGluZykuCisKKyAgICAgICAgQ2xhbXAgc2l6 ZXMgZm9yIGNvbXBvc2l0ZWQgbGF5ZXJzIGNvbWluZyBvdXQgb2YgV2ViQ29yZSB0byBtYWtlIHN1 cmUgdGhleSBhcmUgdmFsaWQgbm9uLW5lZ2F0aXZlIHZhbHVlcy4KKworICAgICAgICAqIHBsYXRm b3JtL2dyYXBoaWNzL2Nocm9taXVtL0dyYXBoaWNzTGF5ZXJDaHJvbWl1bS5jcHA6CisgICAgICAg IChXZWJDb3JlOjpHcmFwaGljc0xheWVyQ2hyb21pdW06OnNldFNpemUpOgorCiAyMDEyLTAyLTA5 ICBNaWtlIExhd3RoZXIgIDxtaWtlbGF3dGhlckBjaHJvbWl1bS5vcmc+CiAKICAgICAgICAgQ1NT MyBjYWxjKCkgLSByZW1vdmUgbW9kCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9y bS9ncmFwaGljcy9jaHJvbWl1bS9HcmFwaGljc0xheWVyQ2hyb21pdW0uY3BwIGIvU291cmNlL1dl YkNvcmUvcGxhdGZvcm0vZ3JhcGhpY3MvY2hyb21pdW0vR3JhcGhpY3NMYXllckNocm9taXVtLmNw cAppbmRleCAzMWE1MTE4ZmZkMDNkMTgyNDhkMDdlNTFlMjRhZWVkMzJhZmU5NzBlLi4xMTA3ZjRl YmU5YTg4OGVjM2Q4ZTU3NDg4YmY1M2VmZjVhMGFiNWNjIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2Vi Q29yZS9wbGF0Zm9ybS9ncmFwaGljcy9jaHJvbWl1bS9HcmFwaGljc0xheWVyQ2hyb21pdW0uY3Bw CisrKyBiL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL2Nocm9taXVtL0dyYXBoaWNz TGF5ZXJDaHJvbWl1bS5jcHAKQEAgLTE3NSwxMCArMTc1LDE0IEBAIHZvaWQgR3JhcGhpY3NMYXll ckNocm9taXVtOjpzZXRBbmNob3JQb2ludChjb25zdCBGbG9hdFBvaW50M0QmIHBvaW50KQogCiB2 b2lkIEdyYXBoaWNzTGF5ZXJDaHJvbWl1bTo6c2V0U2l6ZShjb25zdCBGbG9hdFNpemUmIHNpemUp CiB7Ci0gICAgaWYgKHNpemUgPT0gbV9zaXplKQorICAgIC8vIFdlIGFyZSBnZXR0aW5nIG5lZ2F0 aXZlIHNpemVzIGZyb20gV2ViQ29yZSB0aGF0IGNhdXNlIGFzc2VydGlvbnMgdG8gZmFpbCBpbiB0 aGUgY29tcG9zaXRvci4gTmVnYXRpdmUgc2l6ZXMgZG9uJ3QgbWFrZQorICAgIC8vIHNlbnNlLCBz byBjbGFtcCB0aGVtIHRvIDAgaGVyZS4KKyAgICBGbG9hdFNpemUgY2xhbXBlZFNpemUobWF4PGZs b2F0PihzaXplLndpZHRoKCksIDApLCBtYXg8ZmxvYXQ+KHNpemUuaGVpZ2h0KCksIDApKTsKKwor ICAgIGlmIChjbGFtcGVkU2l6ZSA9PSBtX3NpemUpCiAgICAgICAgIHJldHVybjsKIAotICAgIEdy YXBoaWNzTGF5ZXI6OnNldFNpemUoc2l6ZSk7CisgICAgR3JhcGhpY3NMYXllcjo6c2V0U2l6ZShj bGFtcGVkU2l6ZSk7CiAgICAgdXBkYXRlTGF5ZXJTaXplKCk7CiB9CiAK 126377 2012-02-09 14:44:30 -0800 2012-02-09 15:32:43 -0800 Patch bug-78038-20120209174429.patch text/plain 1873 danakj U3VidmVyc2lvbiBSZXZpc2lvbjogMTA3MjYwCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggNjM5ZDcxNTliY2JhNTc1 MzA5Y2ZmZDBlZTVkMjNlMTViMGY1NDQwOC4uYjNhZTRjNzk3ZGU3NzIzMmUyMjgzYmVmZDYwM2Y3 YzFiNzRkYzliOCAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE3IEBACisyMDEyLTAyLTA5ICBEYW5h IEphbnNlbnMgIDxkYW5ha2pAY2hyb21pdW0ub3JnPgorCisgICAgICAgIFtDaHJvbWl1bV0gQXNz ZXJ0aW9uIGZhaWx1cmUgbWluWCA8PSBtYXhYIGluIFJlZ2lvbi5jcHAKKyAgICAgICAgaHR0cHM6 Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTc4MDM4CisKKyAgICAgICAgUmV2aWV3 ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgQ292ZXJlZCBieSBleGlzdGluZyB0ZXN0 cyAoc2hvdWxkIG1ha2UgdGhlbSBzdG9wIGFzc2VydGluZykuCisKKyAgICAgICAgQ2xhbXAgc2l6 ZXMgZm9yIGNvbXBvc2l0ZWQgbGF5ZXJzIGNvbWluZyBvdXQgb2YgV2ViQ29yZSB0byBtYWtlIHN1 cmUgdGhleSBhcmUgdmFsaWQgbm9uLW5lZ2F0aXZlIHZhbHVlcy4KKworICAgICAgICAqIHBsYXRm b3JtL2dyYXBoaWNzL2Nocm9taXVtL0dyYXBoaWNzTGF5ZXJDaHJvbWl1bS5jcHA6CisgICAgICAg IChXZWJDb3JlOjpHcmFwaGljc0xheWVyQ2hyb21pdW06OnNldFNpemUpOgorCiAyMDEyLTAyLTA5 ICBNaWtlIExhd3RoZXIgIDxtaWtlbGF3dGhlckBjaHJvbWl1bS5vcmc+CiAKICAgICAgICAgQ1NT MyBjYWxjKCkgLSByZW1vdmUgbW9kCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9y bS9ncmFwaGljcy9jaHJvbWl1bS9HcmFwaGljc0xheWVyQ2hyb21pdW0uY3BwIGIvU291cmNlL1dl YkNvcmUvcGxhdGZvcm0vZ3JhcGhpY3MvY2hyb21pdW0vR3JhcGhpY3NMYXllckNocm9taXVtLmNw cAppbmRleCAzMWE1MTE4ZmZkMDNkMTgyNDhkMDdlNTFlMjRhZWVkMzJhZmU5NzBlLi4xMWE0MzU0 M2YyNGNiYTkxYmNlMmVmYWJjNDlhYTA1N2ExYTAwNzMxIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2Vi Q29yZS9wbGF0Zm9ybS9ncmFwaGljcy9jaHJvbWl1bS9HcmFwaGljc0xheWVyQ2hyb21pdW0uY3Bw CisrKyBiL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL2Nocm9taXVtL0dyYXBoaWNz TGF5ZXJDaHJvbWl1bS5jcHAKQEAgLTE3NSwxMCArMTc1LDE2IEBAIHZvaWQgR3JhcGhpY3NMYXll ckNocm9taXVtOjpzZXRBbmNob3JQb2ludChjb25zdCBGbG9hdFBvaW50M0QmIHBvaW50KQogCiB2 b2lkIEdyYXBoaWNzTGF5ZXJDaHJvbWl1bTo6c2V0U2l6ZShjb25zdCBGbG9hdFNpemUmIHNpemUp CiB7Ci0gICAgaWYgKHNpemUgPT0gbV9zaXplKQorICAgIC8vIFdlIGFyZSByZWNlaXZpbmcgbmVn YXRpdmUgc2l6ZXMgaGVyZSB0aGF0IGNhdXNlIGFzc2VydGlvbnMgdG8gZmFpbCBpbiB0aGUgY29t cG9zaXRvci4gQ2xhbXAgdGhlbSB0byAwIHRvCisgICAgLy8gYXZvaWQgYXNzZXJ0aW9ucyBpbiB0 aGUgY29tcG9zaXRvci4KKyAgICBGbG9hdFNpemUgY2xhbXBlZFNpemUgPSBzaXplOworICAgIGlm IChjbGFtcGVkU2l6ZS5pc0VtcHR5KCkpCisgICAgICAgIGNsYW1wZWRTaXplID0gRmxvYXRTaXpl KCk7CisKKyAgICBpZiAoY2xhbXBlZFNpemUgPT0gbV9zaXplKQogICAgICAgICByZXR1cm47CiAK LSAgICBHcmFwaGljc0xheWVyOjpzZXRTaXplKHNpemUpOworICAgIEdyYXBoaWNzTGF5ZXI6OnNl dFNpemUoY2xhbXBlZFNpemUpOwogICAgIHVwZGF0ZUxheWVyU2l6ZSgpOwogfQogCg== 126389 2012-02-09 15:32:47 -0800 2012-02-09 22:08:54 -0800 Patch for landing bug-78038-20120209183245.patch text/plain 1861 fsamuel U3VidmVyc2lvbiBSZXZpc2lvbjogMTA3MjYwCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggNjM5ZDcxNTliY2JhNTc1 MzA5Y2ZmZDBlZTVkMjNlMTViMGY1NDQwOC4uM2FhYjJlNWQ5NmRiN2ViNjkzNzViNDkxMDU4OGMw NjFiZjg1ZDc4MSAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE3IEBACisyMDEyLTAyLTA5ICBEYW5h IEphbnNlbnMgIDxkYW5ha2pAY2hyb21pdW0ub3JnPgorCisgICAgICAgIFtDaHJvbWl1bV0gQXNz ZXJ0aW9uIGZhaWx1cmUgbWluWCA8PSBtYXhYIGluIFJlZ2lvbi5jcHAKKyAgICAgICAgaHR0cHM6 Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTc4MDM4CisKKyAgICAgICAgUmV2aWV3 ZWQgYnkgSmFtZXMgUm9iaW5zb24uCisKKyAgICAgICAgQ292ZXJlZCBieSBleGlzdGluZyB0ZXN0 cyAoc2hvdWxkIG1ha2UgdGhlbSBzdG9wIGFzc2VydGluZykuCisKKyAgICAgICAgQ2xhbXAgc2l6 ZXMgZm9yIGNvbXBvc2l0ZWQgbGF5ZXJzIGNvbWluZyBvdXQgb2YgV2ViQ29yZSB0byBtYWtlIHN1 cmUgdGhleSBhcmUgdmFsaWQgbm9uLW5lZ2F0aXZlIHZhbHVlcy4KKworICAgICAgICAqIHBsYXRm b3JtL2dyYXBoaWNzL2Nocm9taXVtL0dyYXBoaWNzTGF5ZXJDaHJvbWl1bS5jcHA6CisgICAgICAg IChXZWJDb3JlOjpHcmFwaGljc0xheWVyQ2hyb21pdW06OnNldFNpemUpOgorCiAyMDEyLTAyLTA5 ICBNaWtlIExhd3RoZXIgIDxtaWtlbGF3dGhlckBjaHJvbWl1bS5vcmc+CiAKICAgICAgICAgQ1NT MyBjYWxjKCkgLSByZW1vdmUgbW9kCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9y bS9ncmFwaGljcy9jaHJvbWl1bS9HcmFwaGljc0xheWVyQ2hyb21pdW0uY3BwIGIvU291cmNlL1dl YkNvcmUvcGxhdGZvcm0vZ3JhcGhpY3MvY2hyb21pdW0vR3JhcGhpY3NMYXllckNocm9taXVtLmNw cAppbmRleCAzMWE1MTE4ZmZkMDNkMTgyNDhkMDdlNTFlMjRhZWVkMzJhZmU5NzBlLi5iYjg4NWVk OGVlZjlhNWM5MmIzZTFlNjBmNTczYWViZWIwMzM5ZWVkIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2Vi Q29yZS9wbGF0Zm9ybS9ncmFwaGljcy9jaHJvbWl1bS9HcmFwaGljc0xheWVyQ2hyb21pdW0uY3Bw CisrKyBiL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL2Nocm9taXVtL0dyYXBoaWNz TGF5ZXJDaHJvbWl1bS5jcHAKQEAgLTE3NSwxMCArMTc1LDE2IEBAIHZvaWQgR3JhcGhpY3NMYXll ckNocm9taXVtOjpzZXRBbmNob3JQb2ludChjb25zdCBGbG9hdFBvaW50M0QmIHBvaW50KQogCiB2 b2lkIEdyYXBoaWNzTGF5ZXJDaHJvbWl1bTo6c2V0U2l6ZShjb25zdCBGbG9hdFNpemUmIHNpemUp CiB7Ci0gICAgaWYgKHNpemUgPT0gbV9zaXplKQorICAgIC8vIFdlIGFyZSByZWNlaXZpbmcgbmVn YXRpdmUgc2l6ZXMgaGVyZSB0aGF0IGNhdXNlIGFzc2VydGlvbnMgdG8gZmFpbCBpbiB0aGUgY29t cG9zaXRvci4gQ2xhbXAgdGhlbSB0byAwIHRvCisgICAgLy8gYXZvaWQgdGhvc2UgYXNzZXJ0aW9u cy4KKyAgICBGbG9hdFNpemUgY2xhbXBlZFNpemUgPSBzaXplOworICAgIGlmIChjbGFtcGVkU2l6 ZS5pc0VtcHR5KCkpCisgICAgICAgIGNsYW1wZWRTaXplID0gRmxvYXRTaXplKCk7CisKKyAgICBp ZiAoY2xhbXBlZFNpemUgPT0gbV9zaXplKQogICAgICAgICByZXR1cm47CiAKLSAgICBHcmFwaGlj c0xheWVyOjpzZXRTaXplKHNpemUpOworICAgIEdyYXBoaWNzTGF5ZXI6OnNldFNpemUoY2xhbXBl ZFNpemUpOwogICAgIHVwZGF0ZUxheWVyU2l6ZSgpOwogfQogCg==