119633 2013-08-09 11:44:10 -0700 REGRESSION (r142755): window.open creates an invisible window when width and height are 0 2013-08-09 16:43:01 -0700 1 1 1 Unclassified WebKit DOM 528+ (Nightly build) Mac Unspecified RESOLVED FIXED InRadar, Regression P1 Normal --- 1 ap ap allan.jensen cdumez commit-queue gtk-ews gyuyoung.kim japhet rakuco xan.lopez oldest_to_newest 915891 0 ap 2013-08-09 11:44:10 -0700 javascript:window.open("http://www.apple.com", "", "height=0,width=0") creates an invisible window in Safari now. <rdar://problem/14693930> 915893 1 ap 2013-08-09 11:50:24 -0700 Treating (0, 0) as default size falls apart because windowRect height ends up being 24 here: if (features.heightSet) windowRect.setHeight(features.height + (windowRect.height() - viewportSize.height())); 915896 2 ap 2013-08-09 11:54:18 -0700 window.resizeTo is broken too. 915900 3 ap 2013-08-09 11:57:04 -0700 I'm thinking that r142755 should be rolled out entirely, and replaced with a fix in createWindow() function in FrameLoader.cpp: - if (features.widthSet) + if (features.widthSet && features.width) windowRect.setWidth(features.width + (windowRect.width() - viewportSize.width())); - if (features.heightSet) + if (features.heightSet && features.height) windowRect.setHeight(features.height + (windowRect.height() - viewportSize.height())); Allan, does this make sense? 915922 4 208451 ap 2013-08-09 13:42:21 -0700 Created attachment 208451 proposed fix 915926 5 208451 darin 2013-08-09 13:49:44 -0700 Comment on attachment 208451 proposed fix View in context: https://bugs.webkit.org/attachment.cgi?id=208451&action=review > Source/WebCore/ChangeLog:3 > + REGRESSION (r142755): window.open creates an invisible window when width and height are 0 Seems like the bug is when width *or* height is zero. What about small widths and heights that are not zero? 915950 6 208451 gtk-ews 2013-08-09 14:48:50 -0700 Comment on attachment 208451 proposed fix Attachment 208451 did not pass gtk-ews (gtk): Output: http://webkit-queues.appspot.com/results/1410299 915952 7 allan.jensen 2013-08-09 14:50:18 -0700 I don't understand how the browser is supposed to be able to guess if default size was requested when the get an openWindow call. The issue is this: In many browsers windows that request a specific size are opened in a separate window, but windows that request default size are opened in a tab. This is decided outside of WebKit, and thus WebKit needs to tell the browser that default was requested. If 0x0 is not used, what is? 915955 8 ap 2013-08-09 15:15:00 -0700 Looks like Gtk bot is broken now. > Seems like the bug is when width *or* height is zero. What about small widths and heights that are not zero? These are handled in DOMWindow::adjustWindowRect, which enforces a minimum size. Width and height are adjusted to be at least 100. > In many browsers windows that request a specific size are opened in a separate window, but windows that request default size are opened in a tab. When WebKit creates a window, it's before any of this code is run (oldPage->chrome().createWindow() in FrameLoader.cpp). The window is always created with a default size, by definition of default size, and is then resized by the code this patch touches. Afterwards, it's displayed with page->chrome().show(). This has to work in all WebKit based browsers as long as window.open(url, "", "") works with no explicit size. We just read the current size from client using page->chrome().windowRect() and page->chrome().pageRect(), adjust it to be on screen and to not be too small, and feed back to the client. I find it a bit strange how this logic is split between WebKit and client, but it works. Are you saying that there are WebKit based browsers that make the decision on whether to put the window in tab based on requested size? Presumably they make the decision in their show() implementation then. 915956 9 208451 ap 2013-08-09 15:17:15 -0700 Comment on attachment 208451 proposed fix View in context: https://bugs.webkit.org/attachment.cgi?id=208451&action=review > Source/WebKit/efl/WebCoreSupport/ChromeClientEfl.cpp:131 > - if (!ewk_view_setting_enable_auto_resize_window_get(m_view) || rect.isEmpty()) > + if (!ewk_view_setting_enable_auto_resize_window_get(m_view)) Actually, I think that this code used to break moving the window to screen if it had default size, but an offscreen position. But perhaps that can never happen in practice. 915964 10 ap 2013-08-09 15:35:59 -0700 > I don't understand how the browser is supposed to be able to guess if default size was requested when the get an openWindow call. To answer your question more directly, WebCore does pass all the requested features to the client, so it can do anything even upfront in createWindow(). It is kind of strange that WebCore enforces additional constraints after the fact, I don't know why we do this. 915965 11 allan.jensen 2013-08-09 15:39:02 -0700 (In reply to comment #10) > Are you saying that there are WebKit based browsers that make the decision on whether to put the window in tab based on requested size? Presumably they make the decision in their show() implementation then. Yes all of them except Safari apparently. 915979 12 allan.jensen 2013-08-09 16:02:36 -0700 (In reply to comment #11) > (In reply to comment #10) > > Are you saying that there are WebKit based browsers that make the decision on whether to put the window in tab based on requested size? Presumably they make the decision in their show() implementation then. > > Yes all of them except Safari apparently. Well, specifically the classic desktop browsers implementing tabs. Default size windows opening in tabs is expected behavior on a lot of sites that can control whether a new window should popup or not that way. WebKit for years requsted a window of size (0,0) if no size was set, so that became how the browsers implemented the standard behavior. The minimum size for unset window size was only enforced for a few months before I modified it to let 0 through again. Did something change in Safari to rely on the minimum in the timespan it was enforced? 915982 13 ap 2013-08-09 16:23:25 -0700 > WebKit for years requsted a window of size (0,0) if no size was set The windows is created by this method: Page* Chrome::createWindow(Frame* frame, const FrameLoadRequest& request, const WindowFeatures& features, const NavigationAction& action) const { Page* newPage = m_client->createWindow(frame, request, features, action); ... } How was the (0, 0) size passed? The size is in WindowFeatures, but neither your not my patch change anything about that structure content. We only change later calls that resize the already created window. > The minimum size for unset window size was only enforced for a few months before I modified it to let 0 through again. Do you know when this was added? As mentioned before, I'm surprised that any of this logic is in WebCore. 915983 14 208451 commit-queue 2013-08-09 16:24:03 -0700 Comment on attachment 208451 proposed fix Clearing flags on attachment: 208451 Committed r153913: <http://trac.webkit.org/changeset/153913> 915984 15 commit-queue 2013-08-09 16:24:06 -0700 All reviewed patches have been landed. Closing bug. 915987 16 allan.jensen 2013-08-09 16:36:42 -0700 (In reply to comment #13) > Do you know when this was added? As mentioned before, I'm surprised that any of this logic is in WebCore. The adjustment was there before but the part that overwrote default 0,0 with minimum came from http://trac.webkit.org/changeset/134586 915988 17 ap 2013-08-09 16:43:01 -0700 Thank you for the pointer. Based on what I learned looking at this code today, it's probably wrong to call adjustWindowRect when creating a new window, and in general, it seems that logic added in r134586 should be client's responsibility. 208451 2013-08-09 13:42:21 -0700 2013-08-09 16:24:03 -0700 proposed fix WindowZeroSize.txt text/plain 9055 ap SW5kZXg6IFNvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV2Vi Q29yZS9DaGFuZ2VMb2cJKHJldmlzaW9uIDE1MzkwNykKKysrIFNvdXJjZS9XZWJDb3JlL0NoYW5n ZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDIzIEBACisyMDEzLTA4LTA5ICBBbGV4ZXkg UHJvc2t1cnlha292ICA8YXBAYXBwbGUuY29tPgorCisgICAgICAgIFJFR1JFU1NJT04gKHIxNDI3 NTUpOiB3aW5kb3cub3BlbiBjcmVhdGVzIGFuIGludmlzaWJsZSB3aW5kb3cgd2hlbiB3aWR0aCBh bmQgaGVpZ2h0IGFyZSAwCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVn LmNnaT9pZD0xMTk2MzMKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKwor ICAgICAgICBUZXN0OiBmYXN0L2RvbS9XaW5kb3cvb3Blbi16ZXJvLXNpemUtYXMtZGVmYXVsdC5o dG1sCisKKyAgICAgICAgUmVseWluZyBvbiBlYWNoIFdlYktpdCB0byByZWZ1c2Ugc2V0dGluZyBz aXplIHRvIHplcm8gd2FzIGZyYWdpbGUgLSBiZWNhdXNlIHRoaXMKKyAgICAgICAgcmVxdWlyZWQg ZWFjaCBvbmUgdG8gaGF2ZSB0aGUgY2hlY2ssIGFuZCBiZWNhdXNlIGJ5IHRoZSB0aW1lIHRoZSBj bGllbnQgd2FzIGNhbGxlZCwKKyAgICAgICAgdGhlIGluaXRpYWxseSB6ZXJvIHNpemUgd2FzIG5v dCBuZWNlc3NhcmlseSB6ZXJvLgorCisgICAgICAgICogbG9hZGVyL0ZyYW1lTG9hZGVyLmNwcDog KFdlYkNvcmU6OmNyZWF0ZVdpbmRvdyk6IFdoZW4gc2l6ZXMgYXJlIHplcm8sIGtlZXAgdGhlCisg ICAgICAgIHNpemUgdGhlIHdpbmRvdyB3YXMgY3JlYXRlZCB3aXRoLCBiZWNhdXNlIHRoYXQncyB0 aGUgZGVmYXVsdCBvbmUgYnkgZGVmaW5pdGlvbi4KKworICAgICAgICAqIHBhZ2UvRE9NV2luZG93 LmNwcDogKFdlYkNvcmU6OkRPTVdpbmRvdzo6YWRqdXN0V2luZG93UmVjdCk6IEl0J3MgdG9vIGxh dGUgdG8KKyAgICAgICAgY2hlY2sgZm9yIHplcm8gc2l6ZSBub3csIGl0J3MgYmVlbiBtYW5nbGVk IHRvIGFkanVzdCBmb3IgdGhlIGRpZmZlcmVuY2UgYmV0d2VlbgorICAgICAgICB3aW5kb3cgYW5k IHZpZXdwb3J0IHNpemUuCisKIDIwMTMtMDgtMDkgIEJldGggRGFraW4gIDxiZGFraW5AYXBwbGUu Y29tPgogCiAgICAgICAgIEFYOiBOb3QgYWJsZSB0byB1c2UgYXJyb3cga2V5cyB0byByZWFkIHRl eHQgaW4gYSBXSzIgYXBwCkluZGV4OiBTb3VyY2UvV2ViQ29yZS9sb2FkZXIvRnJhbWVMb2FkZXIu Y3BwCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9XZWJDb3JlL2xvYWRlci9GcmFtZUxvYWRlci5jcHAJ KHJldmlzaW9uIDE1MzkwNSkKKysrIFNvdXJjZS9XZWJDb3JlL2xvYWRlci9GcmFtZUxvYWRlci5j cHAJKHdvcmtpbmcgY29weSkKQEAgLTM0NTIsOSArMzQ1MiwxMCBAQCBQYXNzUmVmUHRyPEZyYW1l PiBjcmVhdGVXaW5kb3coRnJhbWUqIG9wCiAgICAgICAgIHdpbmRvd1JlY3Quc2V0WChmZWF0dXJl cy54KTsKICAgICBpZiAoZmVhdHVyZXMueVNldCkKICAgICAgICAgd2luZG93UmVjdC5zZXRZKGZl YXR1cmVzLnkpOwotICAgIGlmIChmZWF0dXJlcy53aWR0aFNldCkKKyAgICAvLyBaZXJvIHdpZHRo IGFuZCBoZWlnaHQgbWVhbiB1c2luZyBkZWZhdWx0IHNpemUsIG5vdCBtaW51bXVtIG9uZS4KKyAg ICBpZiAoZmVhdHVyZXMud2lkdGhTZXQgJiYgZmVhdHVyZXMud2lkdGgpCiAgICAgICAgIHdpbmRv d1JlY3Quc2V0V2lkdGgoZmVhdHVyZXMud2lkdGggKyAod2luZG93UmVjdC53aWR0aCgpIC0gdmll d3BvcnRTaXplLndpZHRoKCkpKTsKLSAgICBpZiAoZmVhdHVyZXMuaGVpZ2h0U2V0KQorICAgIGlm IChmZWF0dXJlcy5oZWlnaHRTZXQgJiYgZmVhdHVyZXMuaGVpZ2h0KQogICAgICAgICB3aW5kb3dS ZWN0LnNldEhlaWdodChmZWF0dXJlcy5oZWlnaHQgKyAod2luZG93UmVjdC5oZWlnaHQoKSAtIHZp ZXdwb3J0U2l6ZS5oZWlnaHQoKSkpOwogCiAgICAgLy8gRW5zdXJlIG5vbi1OYU4gdmFsdWVzLCBt aW5pbXVtIHNpemUgYXMgd2VsbCBhcyBiZWluZyB3aXRoaW4gdmFsaWQgc2NyZWVuIGFyZWEuCklu ZGV4OiBTb3VyY2UvV2ViQ29yZS9wYWdlL0RPTVdpbmRvdy5jcHAKPT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291 cmNlL1dlYkNvcmUvcGFnZS9ET01XaW5kb3cuY3BwCShyZXZpc2lvbiAxNTM5MDUpCisrKyBTb3Vy Y2UvV2ViQ29yZS9wYWdlL0RPTVdpbmRvdy5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTM0MCwxMSAr MzQwLDggQEAgRmxvYXRSZWN0IERPTVdpbmRvdzo6YWRqdXN0V2luZG93UmVjdChQYQogICAgICAg ICB3aW5kb3cuc2V0SGVpZ2h0KHBlbmRpbmdDaGFuZ2VzLmhlaWdodCgpKTsKIAogICAgIEZsb2F0 U2l6ZSBtaW5pbXVtU2l6ZSA9IHBhZ2UtPmNocm9tZSgpLmNsaWVudCgpLT5taW5pbXVtV2luZG93 U2l6ZSgpOwotICAgIC8vIExldCBzaXplIDAgcGFzcyB0aHJvdWdoLCBzaW5jZSB0aGF0IGluZGlj YXRlcyBkZWZhdWx0IHNpemUsIG5vdCBtaW5pbXVtIHNpemUuCi0gICAgaWYgKHdpbmRvdy53aWR0 aCgpKQotICAgICAgICB3aW5kb3cuc2V0V2lkdGgobWluKG1heChtaW5pbXVtU2l6ZS53aWR0aCgp LCB3aW5kb3cud2lkdGgoKSksIHNjcmVlbi53aWR0aCgpKSk7Ci0gICAgaWYgKHdpbmRvdy5oZWln aHQoKSkKLSAgICAgICAgd2luZG93LnNldEhlaWdodChtaW4obWF4KG1pbmltdW1TaXplLmhlaWdo dCgpLCB3aW5kb3cuaGVpZ2h0KCkpLCBzY3JlZW4uaGVpZ2h0KCkpKTsKKyAgICB3aW5kb3cuc2V0 V2lkdGgobWluKG1heChtaW5pbXVtU2l6ZS53aWR0aCgpLCB3aW5kb3cud2lkdGgoKSksIHNjcmVl bi53aWR0aCgpKSk7CisgICAgd2luZG93LnNldEhlaWdodChtaW4obWF4KG1pbmltdW1TaXplLmhl aWdodCgpLCB3aW5kb3cuaGVpZ2h0KCkpLCBzY3JlZW4uaGVpZ2h0KCkpKTsKIAogICAgIC8vIENv bnN0cmFpbiB0aGUgd2luZG93IHBvc2l0aW9uIHdpdGhpbiB0aGUgdmFsaWQgc2NyZWVuIGFyZWEu CiAgICAgd2luZG93LnNldFgobWF4KHNjcmVlbi54KCksIG1pbih3aW5kb3cueCgpLCBzY3JlZW4u bWF4WCgpIC0gd2luZG93LndpZHRoKCkpKSk7CkluZGV4OiBTb3VyY2UvV2ViS2l0L2VmbC9DaGFu Z2VMb2cKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PQotLS0gU291cmNlL1dlYktpdC9lZmwvQ2hhbmdlTG9nCShyZXZpc2lv biAxNTM5MDcpCisrKyBTb3VyY2UvV2ViS2l0L2VmbC9DaGFuZ2VMb2cJKHdvcmtpbmcgY29weSkK QEAgLTEsMyArMSwxMyBAQAorMjAxMy0wOC0wOSAgQWxleGV5IFByb3NrdXJ5YWtvdiAgPGFwQGFw cGxlLmNvbT4KKworICAgICAgICBSRUdSRVNTSU9OIChyMTQyNzU1KTogd2luZG93Lm9wZW4gY3Jl YXRlcyBhbiBpbnZpc2libGUgd2luZG93IHdoZW4gd2lkdGggYW5kIGhlaWdodCBhcmUgMAorICAg ICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTE5NjMzCisKKyAg ICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgKiBXZWJDb3JlU3Vw cG9ydC9DaHJvbWVDbGllbnRFZmwuY3BwOiAoV2ViQ29yZTo6Q2hyb21lQ2xpZW50RWZsOjpzZXRX aW5kb3dSZWN0KToKKyAgICAgICAgT25jZSBhZ2FpbiwgdGhlIHBhc3NlZCByZWN0IGNhbm5vdCBi ZSBlbXB0eS4KKwogMjAxMy0wNy0yNyAgTWFyayBSb3dlICA8bXJvd2VAYXBwbGUuY29tPgogCiAg ICAgICAgIExvZ2dpbmcgc2hvdWxkIGJlIGNvbmZpZ3VyYWJsZSB1c2luZyBodW1hbi1yZWFkYWJs ZSBjaGFubmVsIG5hbWVzIHJhdGhlciB0aGFuIGNyYXp5IGJpdG1hc2tzCkluZGV4OiBTb3VyY2Uv V2ViS2l0L2VmbC9XZWJDb3JlU3VwcG9ydC9DaHJvbWVDbGllbnRFZmwuY3BwCj09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0K LS0tIFNvdXJjZS9XZWJLaXQvZWZsL1dlYkNvcmVTdXBwb3J0L0Nocm9tZUNsaWVudEVmbC5jcHAJ KHJldmlzaW9uIDE1MzkwNSkKKysrIFNvdXJjZS9XZWJLaXQvZWZsL1dlYkNvcmVTdXBwb3J0L0No cm9tZUNsaWVudEVmbC5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTEyOCw3ICsxMjgsNyBAQCBGbG9h dFJlY3QgQ2hyb21lQ2xpZW50RWZsOjp3aW5kb3dSZWN0KCkKIAogdm9pZCBDaHJvbWVDbGllbnRF Zmw6OnNldFdpbmRvd1JlY3QoY29uc3QgRmxvYXRSZWN0JiByZWN0KQogewotICAgIGlmICghZXdr X3ZpZXdfc2V0dGluZ19lbmFibGVfYXV0b19yZXNpemVfd2luZG93X2dldChtX3ZpZXcpIHx8IHJl Y3QuaXNFbXB0eSgpKQorICAgIGlmICghZXdrX3ZpZXdfc2V0dGluZ19lbmFibGVfYXV0b19yZXNp emVfd2luZG93X2dldChtX3ZpZXcpKQogICAgICAgICByZXR1cm47CiAKICAgICBFY29yZV9FdmFz KiBlZSA9IGVjb3JlX2V2YXNfZWNvcmVfZXZhc19nZXQoZXZhc19vYmplY3RfZXZhc19nZXQobV92 aWV3KSk7CkluZGV4OiBTb3VyY2UvV2ViS2l0L2d0ay9DaGFuZ2VMb2cKPT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0g U291cmNlL1dlYktpdC9ndGsvQ2hhbmdlTG9nCShyZXZpc2lvbiAxNTM5MDcpCisrKyBTb3VyY2Uv V2ViS2l0L2d0ay9DaGFuZ2VMb2cJKHdvcmtpbmcgY29weSkKQEAgLTEsMyArMSwxMyBAQAorMjAx My0wOC0wOSAgQWxleGV5IFByb3NrdXJ5YWtvdiAgPGFwQGFwcGxlLmNvbT4KKworICAgICAgICBS RUdSRVNTSU9OIChyMTQyNzU1KTogd2luZG93Lm9wZW4gY3JlYXRlcyBhbiBpbnZpc2libGUgd2lu ZG93IHdoZW4gd2lkdGggYW5kIGhlaWdodCBhcmUgMAorICAgICAgICBodHRwczovL2J1Z3Mud2Vi a2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTE5NjMzCisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9C T0RZIChPT1BTISkuCisKKyAgICAgICAgKiBXZWJDb3JlU3VwcG9ydC9DaHJvbWVDbGllbnRHdGsu Y3BwOiAoV2ViS2l0OjpDaHJvbWVDbGllbnQ6OnNldFdpbmRvd1JlY3QpOgorICAgICAgICBPbmNl IGFnYWluLCB0aGUgcGFzc2VkIHJlY3QgY2Fubm90IGJlIGVtcHR5LgorCiAyMDEzLTA3LTI3ICBN YXJrIFJvd2UgIDxtcm93ZUBhcHBsZS5jb20+CiAKICAgICAgICAgTG9nZ2luZyBzaG91bGQgYmUg Y29uZmlndXJhYmxlIHVzaW5nIGh1bWFuLXJlYWRhYmxlIGNoYW5uZWwgbmFtZXMgcmF0aGVyIHRo YW4gY3JhenkgYml0bWFza3MKSW5kZXg6IFNvdXJjZS9XZWJLaXQvZ3RrL1dlYkNvcmVTdXBwb3J0 L0Nocm9tZUNsaWVudEd0ay5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291cmNlL1dlYktpdC9ndGsvV2Vi Q29yZVN1cHBvcnQvQ2hyb21lQ2xpZW50R3RrLmNwcAkocmV2aXNpb24gMTUzOTA1KQorKysgU291 cmNlL1dlYktpdC9ndGsvV2ViQ29yZVN1cHBvcnQvQ2hyb21lQ2xpZW50R3RrLmNwcAkod29ya2lu ZyBjb3B5KQpAQCAtMTcxLDggKzE3MSw3IEBAIHZvaWQgQ2hyb21lQ2xpZW50OjpzZXRXaW5kb3dS ZWN0KGNvbnN0IEYKICAgICBHdGtXaWRnZXQqIHdpbmRvdyA9IGd0a193aWRnZXRfZ2V0X3RvcGxl dmVsKEdUS19XSURHRVQobV93ZWJWaWV3KSk7CiAgICAgaWYgKHdpZGdldElzT25zY3JlZW5Ub3Bs ZXZlbFdpbmRvdyh3aW5kb3cpKSB7CiAgICAgICAgIGd0a193aW5kb3dfbW92ZShHVEtfV0lORE9X KHdpbmRvdyksIGludHJlY3QueCgpLCBpbnRyZWN0LnkoKSk7Ci0gICAgICAgIGlmICghaW50cmVj dC5pc0VtcHR5KCkpCi0gICAgICAgICAgICBndGtfd2luZG93X3Jlc2l6ZShHVEtfV0lORE9XKHdp bmRvdyksIGludHJlY3Qud2lkdGgoKSwgaW50cmVjdC5oZWlnaHQoKSk7CisgICAgICAgIGd0a193 aW5kb3dfcmVzaXplKEdUS19XSU5ET1cod2luZG93KSwgaW50cmVjdC53aWR0aCgpLCBpbnRyZWN0 LmhlaWdodCgpKTsKICAgICB9CiB9CiAKSW5kZXg6IExheW91dFRlc3RzL0NoYW5nZUxvZwo9PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09Ci0tLSBMYXlvdXRUZXN0cy9DaGFuZ2VMb2cJKHJldmlzaW9uIDE1MzkwNykKKysrIExh eW91dFRlc3RzL0NoYW5nZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDEzIEBACisyMDEz LTA4LTA5ICBBbGV4ZXkgUHJvc2t1cnlha292ICA8YXBAYXBwbGUuY29tPgorCisgICAgICAgIFJF R1JFU1NJT04gKHIxNDI3NTUpOiB3aW5kb3cub3BlbiBjcmVhdGVzIGFuIGludmlzaWJsZSB3aW5k b3cgd2hlbiB3aWR0aCBhbmQgaGVpZ2h0IGFyZSAwCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJr aXQub3JnL3Nob3dfYnVnLmNnaT9pZD0xMTk2MzMKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JP RFkgKE9PUFMhKS4KKworICAgICAgICAqIGZhc3QvZG9tL1dpbmRvdy9vcGVuLXplcm8tc2l6ZS1h cy1kZWZhdWx0LWV4cGVjdGVkLnR4dDogQWRkZWQuCisgICAgICAgICogZmFzdC9kb20vV2luZG93 L29wZW4temVyby1zaXplLWFzLWRlZmF1bHQuaHRtbDogQWRkZWQuCisKIDIwMTMtMDgtMDkgIENo cmlzdG9waGUgRHVtZXogIDxjaC5kdW1lekBzaXNhLnNhbXN1bmcuY29tPgogCiAgICAgICAgIEFk ZCBzdXBwb3J0IGZvciBLZXlib2FyZEV2ZW50LmxvY2F0aW9uIGF0dHJpYnV0ZQpJbmRleDogTGF5 b3V0VGVzdHMvZmFzdC9kb20vV2luZG93L29wZW4temVyby1zaXplLWFzLWRlZmF1bHQtZXhwZWN0 ZWQudHh0Cj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT0KLS0tIExheW91dFRlc3RzL2Zhc3QvZG9tL1dpbmRvdy9vcGVuLXpl cm8tc2l6ZS1hcy1kZWZhdWx0LWV4cGVjdGVkLnR4dAkocmV2aXNpb24gMCkKKysrIExheW91dFRl c3RzL2Zhc3QvZG9tL1dpbmRvdy9vcGVuLXplcm8tc2l6ZS1hcy1kZWZhdWx0LWV4cGVjdGVkLnR4 dAkod29ya2luZyBjb3B5KQpAQCAtMCwwICsxLDMgQEAKK1Rlc3QgdGhhdCBwYXNzaW5nICJoZWln aHQ9MCx3aWR0aD0wIiB0byB3aW5kb3cub3BlbigpIHJlc3VsdHMgaW4gZGVmYXVsdCB3aW5kb3cg c2l6ZS4KKworUEFTUwoKUHJvcGVydHkgY2hhbmdlcyBvbjogTGF5b3V0VGVzdHMvZmFzdC9kb20v V2luZG93L29wZW4temVyby1zaXplLWFzLWRlZmF1bHQtZXhwZWN0ZWQudHh0Cl9fX19fX19fX19f X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18K QWRkZWQ6IHN2bjptaW1lLXR5cGUKIyMgLTAsMCArMSAjIwordGV4dC9wbGFpbgpcIE5vIG5ld2xp bmUgYXQgZW5kIG9mIHByb3BlcnR5CkFkZGVkOiBzdm46ZW9sLXN0eWxlCiMjIC0wLDAgKzEgIyMK K25hdGl2ZQpcIE5vIG5ld2xpbmUgYXQgZW5kIG9mIHByb3BlcnR5CkluZGV4OiBMYXlvdXRUZXN0 cy9mYXN0L2RvbS9XaW5kb3cvb3Blbi16ZXJvLXNpemUtYXMtZGVmYXVsdC5odG1sCj09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT0KLS0tIExheW91dFRlc3RzL2Zhc3QvZG9tL1dpbmRvdy9vcGVuLXplcm8tc2l6ZS1hcy1kZWZh dWx0Lmh0bWwJKHJldmlzaW9uIDApCisrKyBMYXlvdXRUZXN0cy9mYXN0L2RvbS9XaW5kb3cvb3Bl bi16ZXJvLXNpemUtYXMtZGVmYXVsdC5odG1sCSh3b3JraW5nIGNvcHkpCkBAIC0wLDAgKzEsMzIg QEAKKzxodG1sPgorPGJvZHk+Cis8cD5UZXN0IHRoYXQgcGFzc2luZyAiaGVpZ2h0PTAsd2lkdGg9 MCIgdG8gd2luZG93Lm9wZW4oKSByZXN1bHRzIGluIGRlZmF1bHQgd2luZG93IHNpemUuPC9wPgor PGRpdiBpZD1yZXN1bHQ+PC9kaXY+Cis8c2NyaXB0PgoraWYgKHdpbmRvdy50ZXN0UnVubmVyKSB7 CisgICAgdGVzdFJ1bm5lci53YWl0VW50aWxEb25lKCk7CisgICAgdGVzdFJ1bm5lci5kdW1wQXNU ZXh0KCk7CisgICAgdGVzdFJ1bm5lci5zZXRDYW5PcGVuV2luZG93cygpOworfQorCit2YXIgcmVz dWx0ID0gZG9jdW1lbnQuZ2V0RWxlbWVudEJ5SWQoInJlc3VsdCIpOworCit3aW5kb3cuYWRkRXZl bnRMaXN0ZW5lcigibWVzc2FnZSIsIGZ1bmN0aW9uIChlKSB7CisgICAgdHJ5IHsKKyAgICAgICAg dmFyIGRpbWVuc2lvbnMgPSBKU09OLnBhcnNlKGUuZGF0YSk7CisgICAgICAgIGlmIChkaW1lbnNp b25zWzBdID4gMTAwICYmIGRpbWVuc2lvbnNbMV0gPiAxMDApCisgICAgICAgICAgICByZXN1bHQu aW5uZXJIVE1MID0gIlBBU1MiOworICAgICAgICBlbHNlCisgICAgICAgICAgICByZXN1bHQuaW5u ZXJIVE1MID0gIkZBSUw6ICIgKyBlLmRhdGE7CisgICAgfSBjYXRjaCAoZXgpIHsKKyAgICAgICAg YWxlcnQoIkZBSUw6ICIgKyBleCArICIgKCIgKyBlLmRhdGEgKyAiKSIpOworICAgIH0KKworICAg IGlmICh3aW5kb3cudGVzdFJ1bm5lcikKKyAgICAgICAgdGVzdFJ1bm5lci5ub3RpZnlEb25lKCk7 Cit9LCBmYWxzZSk7CisKK3dpbmRvdy5vcGVuKCJkYXRhOnRleHQvaHRtbCw8c2NyaXB0Pm9wZW5l ci5wb3N0TWVzc2FnZSgnWycgKyB3aW5kb3cuaW5uZXJXaWR0aCArICcsJyArIHdpbmRvdy5pbm5l ckhlaWdodCArICddJywgJyonKTsgd2luZG93LmNsb3NlKCk7PCIgKyAiL3NjcmlwdD4iLCAiIiwg ImhlaWdodD0wLHdpZHRoPTAiKTsKKzwvc2NyaXB0PgorPC9ib2R5PgorPC9odG1sPgoKUHJvcGVy dHkgY2hhbmdlcyBvbjogTGF5b3V0VGVzdHMvZmFzdC9kb20vV2luZG93L29wZW4temVyby1zaXpl LWFzLWRlZmF1bHQuaHRtbApfX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19f X19fX19fX19fX19fX19fX19fX19fX19fX19fCkFkZGVkOiBzdm46bWltZS10eXBlCiMjIC0wLDAg KzEgIyMKK3RleHQvaHRtbApcIE5vIG5ld2xpbmUgYXQgZW5kIG9mIHByb3BlcnR5Cg==