188729 2018-08-18 17:12:06 -0700 svg/custom/glyph-selection-bidi-mirror.svg and other SVG tests abandon their document because of references via SVGFontFaceElement 2020-06-29 16:42:46 -0700 1 1 1 Unclassified WebKit SVG WebKit Nightly Build Unspecified Unspecified REOPENED InRadar P2 Normal --- 1 simon.fraser webkit-unassigned ap Lawrence.j rackler simon.fraser webkit-bot-watchers-bugzilla webkit-bug-importer zimmermann oldest_to_newest 1451434 0 simon.fraser 2018-08-18 17:12:06 -0700 svg/custom/glyph-selection-bidi-mirror.svg and other SVG tests abandon their document (bug 186214). Ref stacks on the C++ side seem reasonable, so this looks like an issue on the GC side. 1451962 1 simon.fraser 2018-08-20 21:30:55 -0700 svg/wicd/test-rightsizing-b.xhtml abandons 4 instances of svg/wicd/resources/test-svg-child-object-rightsizing.svg. They are retained through a referencing element, the <font-face> element: <font-face font-family="Sera Sans" units-per-em="2048" panose-1="2 11 6 3 3 8 4 2 2 4" ascent="1901" descent="-483" alphabetic="0" /> 1451964 2 simon.fraser 2018-08-20 21:31:05 -0700 *** Bug 188776 has been marked as a duplicate of this bug. *** 1455150 3 simon.fraser 2018-08-29 21:15:41 -0700 I think these tests all suffer in the same way: svg/W3C-SVG-1.1-SE/color-prop-05-t.svg svg/W3C-SVG-1.1-SE/painting-control-04-f.svg svg/W3C-SVG-1.1/animate-elem-03-t.svg svg/W3C-SVG-1.1/animate-elem-24-t.svg svg/W3C-SVG-1.1/animate-elem-36-t.svg svg/W3C-SVG-1.1/animate-elem-40-t.svg svg/W3C-SVG-1.1/fonts-desc-02-t.svg svg/W3C-SVG-1.1/fonts-elem-01-t.svg svg/W3C-SVG-1.1/fonts-elem-02-t.svg svg/W3C-SVG-1.1/fonts-elem-04-b.svg svg/W3C-SVG-1.1/fonts-elem-05-t.svg svg/W3C-SVG-1.1/fonts-elem-06-t.svg svg/W3C-SVG-1.1/fonts-glyph-02-t.svg svg/W3C-SVG-1.1/fonts-glyph-03-t.svg svg/W3C-SVG-1.1/fonts-kern-01-t.svg svg/W3C-SVG-1.1/masking-mask-01-b.svg svg/W3C-SVG-1.1/pservers-grad-08-b.svg svg/W3C-SVG-1.1/render-elems-06-t.svg svg/W3C-SVG-1.1/render-elems-07-t.svg svg/W3C-SVG-1.1/render-elems-08-t.svg svg/W3C-SVG-1.1/render-groups-01-b.svg svg/W3C-SVG-1.1/render-groups-03-t.svg svg/W3C-SVG-1.1/text-align-08-b.svg svg/W3C-SVG-1.1/text-altglyph-01-b.svg svg/W3C-SVG-1.1/text-fonts-03-t.svg svg/W3C-SVG-1.1/text-intro-01-t.svg svg/W3C-SVG-1.1/text-intro-02-b.svg svg/W3C-SVG-1.1/text-intro-03-b.svg svg/W3C-SVG-1.1/text-intro-04-t.svg svg/W3C-SVG-1.1/text-text-04-t.svg svg/W3C-SVG-1.1/text-text-05-t.svg svg/W3C-SVG-1.1/text-text-06-t.svg 1455152 4 simon.fraser 2018-08-29 21:22:16 -0700 In MiniBrowser WK2, we correctly destroy the two SVGFontFaceElements and one CachedSVGFont that are created for svg/W3C-SVG-1.1-SE/color-prop-05-t.svg. However, in WTR, we only destroy one of those SVGFontFaceElements and never the CachedSVGFont, which keeps a Document alive. 1455157 5 simon.fraser 2018-08-29 22:16:12 -0700 To fix this we need to clear font caches, something like: diff --git a/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundle.cpp b/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundle.cpp index 1a32c605f5f2b713b6430309e1dee833318311ee..4d22a40e70a19ed493fb8c9d7a2760d33af3f21d 100644 --- a/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundle.cpp +++ b/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundle.cpp @@ -45,6 +45,7 @@ #include "WebPageGroupProxy.h" #include <WebCore/DatabaseTracker.h> #include <WebCore/MemoryCache.h> +#include <WebCore/MemoryRelease.h> #include <WebCore/PageCache.h> #include <WebCore/ResourceLoadObserver.h> #include <WebCore/ServiceWorkerThreadProxy.h> @@ -268,6 +269,7 @@ void WKBundleClearPageCache(WKBundleRef bundle) void WKBundleClearMemoryCache(WKBundleRef bundle) { MemoryCache::singleton().evictResources(); + WebCore::releaseMemory(WTF::Critical::Yes, WTF::Synchronous::Yes); // Some redundant work here. } WKDataRef WKBundleCreateWKDataFromUInt8Array(WKBundleRef bundle, JSContextRef context, JSValueRef data) 1667321 8 rackler 2020-06-29 15:42:03 -0700 svg/W3C-SVG-1.1/fonts-elem-01-t.svg and svg/W3C-SVG-1.1/fonts-elem-02-t.svg and svg/W3C-SVG-1.1/fonts-elem-03-b.svg and svg/W3C-SVG-1.1/fonts-elem-07-b.svg is no longer failing - remove expectations Current history is green: https://results.webkit.org/?suite=layout-tests&suite=layout-tests&suite=layout-tests&suite=layout-tests&test=svg%2FW3C-SVG-1.1%2Ffonts-elem-01-t.svg&test=svg%2FW3C-SVG-1.1%2Ffonts-elem-02-t.svg&test=svg%2FW3C-SVG-1.1%2Ffonts-elem-03-b.svg&test=svg%2FW3C-SVG-1.1%2Ffonts-elem-07-b.svg These tests were isolated from the others in the group as ones that are unexpectedly passing. 1667323 9 403126 rackler 2020-06-29 15:45:53 -0700 Created attachment 403126 Update TestExpectations 1667324 10 webkit-bug-importer 2020-06-29 15:47:14 -0700 <rdar://problem/64911473> 1667341 11 ews-feeder 2020-06-29 16:31:51 -0700 Committed r263706: <https://trac.webkit.org/changeset/263706> All reviewed patches have been landed. Closing bug and clearing flags on attachment 403126. 1667352 12 ryanhaddad 2020-06-29 16:42:07 -0700 Reopening since some of the tests are still failing. 403126 2020-06-29 15:45:53 -0700 2020-06-29 16:42:46 -0700 Update TestExpectations bug-188729-20200629174551.patch text/plain 1826 rackler U3VidmVyc2lvbiBSZXZpc2lvbjogMjYzNjk4CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggYmM1ZjU3ZTc4ODllYWJhNzhlMDNi YmFmYTgxMTFhMjhlNzZkZmIyNS4uNmQ1YjU3N2MwMjQzYTM1YzQwNzI2YTEzMDUzMDFhYmNmZDFl YmExNCAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDEyIEBACisyMDIwLTA2LTI5ICBLYXJsIFJhY2tsZXIgIDxy YWNrbGVyQGFwcGxlLmNvbT4KKworICAgICAgICBSZW1vdmUgZXhwZWN0YXRpb24gZm9yIHN2Zy9X M0MtU1ZHLTEuMS9mb250cy1lbGVtLTAxLXQuc3ZnIGFuZCBzdmcvVzNDLVNWRy0xLjEvZm9udHMt ZWxlbS0wMi10LnN2ZyBhbmQgc3ZnL1czQy1TVkctMS4xL2ZvbnRzLWVsZW0tMDMtYi5zdmcgYW5k IHN2Zy9XM0MtU1ZHLTEuMS9mb250cy1lbGVtLTA3LWIuc3ZnIGFzIHRoZXkgYXJlIHBhc3Npbmcu IAorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTg4NzI5 CisKKyAgICAgICAgVW5yZXZpZXdlZCB0ZXN0IGdhcmRlbmluZy4KKworICAgICAgICAqIHBsYXRm b3JtL21hYy9UZXN0RXhwZWN0YXRpb25zOgorCiAyMDIwLTA2LTI5ICBLYXJsIFJhY2tsZXIgIDxy YWNrbGVyQGFwcGxlLmNvbT4KIAogICAgICAgICBSZW1vdmUgZXhwZWN0YXRpb24gZm9yIG1lZGlh L2Z1bGxzY3JlZW4tYXBpLWVuYWJsZWQtbWVkaWEtd2l0aC1wcmVzZW50YXRpb24tbW9kZS5odG1s IGFzIHRoZXkgYXJlIHBhc3NpbmcuIApkaWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0v bWFjL1Rlc3RFeHBlY3RhdGlvbnMgYi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9tYWMvVGVzdEV4cGVj dGF0aW9ucwppbmRleCBhNTlhYTIyNjIwNThlZjZhYjVmNThmZDdlZGYzNjcyY2EzM2FlOTFiLi45 Mjk1ZjgyMTJhMjdkODhiMTJlN2NkYmI4OGRiYjA0NTU5ZmJkMWFlIDEwMDY0NAotLS0gYS9MYXlv dXRUZXN0cy9wbGF0Zm9ybS9tYWMvVGVzdEV4cGVjdGF0aW9ucworKysgYi9MYXlvdXRUZXN0cy9w bGF0Zm9ybS9tYWMvVGVzdEV4cGVjdGF0aW9ucwpAQCAtMTM1NywxMCArMTM1Nyw2IEBAIHdlYmtp dC5vcmcvYi8xNjE0OTEgbWVkaWEvdmlkZW8tbWFpbi1jb250ZW50LWFsbG93LXRoZW4tc2Nyb2xs Lmh0bWwgWyBQYXNzIEZhaWx1CiB3ZWJraXQub3JnL2IvMTcyMDUyIFsgRGVidWcgXSBpbXBvcnRl ZC93M2Mvd2ViLXBsYXRmb3JtLXRlc3RzL2h0bWwvd2ViYXBwYXBpcy90aW1lcnMvdHlwZS1sb25n LXNldGludGVydmFsLmh0bWwgWyBQYXNzIEZhaWx1cmUgXQogCiAjIDxyZGFyOi8vcHJvYmxlbS8y OTAzMTUwOT4gUkVHUkVTU0lPTj8gKEZvbnRQYXJzZXItMTk1KTogc3ZnL1czQy1TVkctMS4xL2Zv bnRzLWVsZW0tKiBhbmQgc3ZnL1czQy1TVkctMS4xL3RleHQtaW50cm8tKiB0ZXN0cyBmYWlsaW5n Ci1zdmcvVzNDLVNWRy0xLjEvZm9udHMtZWxlbS0wMS10LnN2ZyBbIEZhaWx1cmUgXQotc3ZnL1cz Qy1TVkctMS4xL2ZvbnRzLWVsZW0tMDItdC5zdmcgWyBGYWlsdXJlIF0KLXN2Zy9XM0MtU1ZHLTEu MS9mb250cy1lbGVtLTAzLWIuc3ZnIFsgRmFpbHVyZSBdCi1zdmcvVzNDLVNWRy0xLjEvZm9udHMt ZWxlbS0wNy1iLnN2ZyBbIEZhaWx1cmUgXQogc3ZnL1czQy1TVkctMS4xL3RleHQtaW50cm8tMDEt dC5zdmcgWyBGYWlsdXJlIF0KIHN2Zy9XM0MtU1ZHLTEuMS90ZXh0LWludHJvLTAyLWIuc3ZnIFsg RmFpbHVyZSBdCiBzdmcvVzNDLVNWRy0xLjEvdGV4dC1pbnRyby0wMy1iLnN2ZyBbIEZhaWx1cmUg XQo=