209139 2020-03-16 09:10:38 -0700 [ iOS wk2 ] webgl/1.0.3/conformance/textures/copy-tex-image-2d-formats.html is failing. 2020-10-28 19:20:54 -0700 1 1 1 Unclassified WebKit WebGL WebKit Nightly Build iPhone / iPad iOS 13 RESOLVED FIXED https://bugs.webkit.org/show_bug.cgi?id=205736 https://bugs.webkit.org/show_bug.cgi?id=218318 InRadar P2 Normal --- 205618 212249 207858 215804 1 Lawrence.j kbr dino ews-watchlist graouts jdarpinian justin_fan kbr kkinnunen kondapallykalyan mmaxfield webkit-bot-watchers-bugzilla webkit-bug-importer oldest_to_newest 1630338 0 Lawrence.j 2020-03-16 09:10:38 -0700 webgl/1.0.3/conformance/textures/copy-tex-image-2d-formats.html Description: This test is failing on iOS wk2. This appears to be related to the solution provided here: https://trac.webkit.org/changeset/258462. History: https://results.webkit.org/?suite=layout-tests&test=webgl%2F1.0.3%2Fconformance%2Ftextures%2Fcopy-tex-image-2d-formats.html&platform=ios&limit=50000 Diff: --- /Volumes/Data/slave/ios-simulator-13-release-tests-wk2/build/layout-test-results/webgl/1.0.3/conformance/textures/copy-tex-image-2d-formats-expected.txt +++ /Volumes/Data/slave/ios-simulator-13-release-tests-wk2/build/layout-test-results/webgl/1.0.3/conformance/textures/copy-tex-image-2d-formats-actual.txt @@ -1,5 +1,56 @@ This test runs the WebGL Test listed below in an iframe and reports PASS or FAIL. Test: ../../resources/webgl_test_files/conformance/textures/copy-tex-image-2d-formats.html -[ PASS ] All tests passed +[ 1: PASS ] getError was expected value: NO_ERROR : During Initialization +[ 2: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D ALPHA from RGBA +[ 3: PASS ] should be 0,0,0,127 +[ 4: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 5: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE from RGBA +[ 6: PASS ] should be 64,64,64,255 +[ 7: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 8: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE_ALPHA from RGBA +[ 9: PASS ] should be 64,64,64,127 +[ 10: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 11: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D RGB from RGBA +[ 12: PASS ] should be 64,255,191,255 +[ 13: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 14: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D RGBA from RGBA +[ 15: PASS ] should be 64,255,191,127 +[ 16: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 17: PASS ] getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D ALPHA from RGB +[ 18: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE from RGB +[ 19: PASS ] should be 64,64,64,255 +[ 20: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 21: PASS ] getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D LUMINANCE_ALPHA from RGB +[ 22: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D RGB from RGB +[ 23: PASS ] should be 64,255,191,255 +[ 24: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 25: PASS ] getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D RGBA from RGB +[ 26: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D ALPHA from RGBA +[ 27: PASS ] should be 0,0,0,127 +[ 28: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 29: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE from RGBA +[ 30: PASS ] should be 64,64,64,255 +[ 31: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 32: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE_ALPHA from RGBA +[ 33: PASS ] should be 64,64,64,127 +[ 34: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 35: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D RGB from RGBA +[ 36: PASS ] should be 64,255,191,255 +[ 37: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 38: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D RGBA from RGBA +[ 39: PASS ] should be 64,255,191,127 +[ 40: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 41: PASS ] getError was expected value: NO_ERROR : During Initialization +[ 42: PASS ] getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D ALPHA from RGB +[ 43: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE from RGB +[ 44: FAIL ] at (0, 0) expected: 64,64,64,255 was 64,64,64,0 +[ 45: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 46: PASS ] getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D LUMINANCE_ALPHA from RGB +[ 47: PASS ] getError was expected value: NO_ERROR : should be able to copyTexImage2D RGB from RGB +[ 48: FAIL ] at (0, 0) expected: 64,255,191,255 was 64,255,191,0 +[ 49: PASS ] getError was expected value: NO_ERROR : should be no errors +[ 50: PASS ] getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D RGBA from RGB +[ 51: PASS ] successfullyParsed is true +[ FAIL ] 2 failures reported 1630339 1 webkit-bug-importer 2020-03-16 09:11:16 -0700 <rdar://problem/60496751> 1630340 2 Lawrence.j 2020-03-16 09:15:14 -0700 I have marked this test as failing while this issue is investigated. https://trac.webkit.org/changeset/258500/webkit 1655380 3 justin_fan 2020-05-22 11:45:48 -0700 On iOS, it looks like copyTexImage2D is altering the alpha channel for RGB/LUMINANCE copies when it should be leaving those bits untouched. 1669483 4 kbr 2020-07-07 12:11:32 -0700 jdarpinian@ pointed out that he just reimplemented CopyTexImage2D on iOS in Bug 212249 to work around other driver bugs; that might have already fixed this problem. 1669508 5 justin_fan 2020-07-07 13:28:19 -0700 unfortunately, this test is still failing on iOS 13 Simulators. I can try to reproduce again once I get another working simulator build. 1673627 6 justin_fan 2020-07-21 14:36:01 -0700 There are only two sub-failures occurring in this test; when copyTexImage2D is attempted from LUMINANCE and RGB formats to the RGB format with an RGB backbuffer, the alpha channel is written as 255 when it should be 0. 1673634 7 justin_fan 2020-07-21 14:54:47 -0700 Sorry, ignore the last. What I meant to say was "when copyTexImage2D is attempted from an RGB backbuffer to LUMINANCE or RGB, the alpha channel is written as 255 when it should be 0. 1673635 8 kbr 2020-07-21 14:58:04 -0700 Not 100% sure how the test reads back the texture's data, but it sounds like it's drawing the resulting texture back to an RGBA back buffer and then calling readPixels with RGBA format. Thanks Justin for the triage - we'll investigate. 1689599 9 kbr 2020-09-17 14:48:00 -0700 Investigating. 1689670 10 kbr 2020-09-17 17:52:58 -0700 The same failure as originally reported still happens in the iOS Simulator: PASS getError was expected value: NO_ERROR : should be able to copyTexImage2D LUMINANCE from RGB FAIL at (0, 0) expected: 64,64,64,255 was 64,64,64,0 PASS getError was expected value: NO_ERROR : should be no errors PASS getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D LUMINANCE_ALPHA from RGB PASS getError was expected value: NO_ERROR : should be able to copyTexImage2D RGB from RGB FAIL at (0, 0) expected: 64,255,191,255 was 64,255,191,0 PASS getError was expected value: NO_ERROR : should be no errors PASS getError was expected value: INVALID_OPERATION : should not be able to copyTexImage2D RGBA from RGB 1693927 11 kbr 2020-10-01 16:52:38 -0700 Still haven't had time to debug into ANGLE on the iOS Simulator yet. Potentially related bugs pointed out by jdarpinian@: copyTexImage2D to LUMINANCE/ALPHA fails with internal context error on mac http://anglebug.com/5051 copy-texture-image-same-texture.html fails on OpenGL backend http://anglebug.com/2994 1701187 12 kbr 2020-10-23 19:00:36 -0700 The bug is that readPixels readbacks from an {alpha:false} back buffer are coming back with the alpha channel zeroed out, rather than 0xFF. It has nothing to do with CopyTexImage2D from the RGB back buffer. 1701289 13 412260 kbr 2020-10-25 01:20:36 -0700 Created attachment 412260 Patch 1701359 14 412260 kkinnunen 2020-10-26 00:27:08 -0700 Comment on attachment 412260 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=412260&action=review looks good to me, not a reviewer. > Source/WebCore/platform/graphics/angle/ExtensionsGLANGLE.cpp:610 > + if (!readingToPixelBufferObject && !attrs.alpha && (format == GraphicsContextGL::RGBA || format == GraphicsContextGL::BGRA) && (type == GraphicsContextGL::UNSIGNED_BYTE) && (state.boundReadFBO == m_context->m_fbo || (attrs.antialias && state.boundReadFBO == m_context->m_multisampleFBO))) Out of curiosity, did we not want to extend ANGLE workarounds around this area instead of existing more unstructured workarounds? Or would this be more task for downstream core developers? It'd help maybe making this more perfect in cases like preserveDrawingBuffer, which probably shouldn't need the workaround on grounds of the backing being normal texture instead of an iosurface I guess part of the tedious work would be to understand what's the root cause of the problem. I did a bit of similar work in https://bugs.webkit.org/show_bug.cgi?id=215804 but it's not finished due to not being that important. It might or might not be same root cause but different trigger this time. E.g. in that bug, glclear is "buggy", but it might be that BlitFramebuffer is buggy. So in this case, it might be that read pixels work but then rendering from blitted-to texture would be black. Anyhow, this is an tangent to the actual question wrt do we want the fix eventually in ANGLE or not. 1701581 15 kbr 2020-10-26 13:10:33 -0700 Yes, ideally this fix would be inside ANGLE. There are a couple of weird things going on with this bug from my perspective: 1) On the iOS Simulator, the WebGL back buffer texture isn't actually bound to the IOSurface. The [EAGLContext texImageIOSurface:target:internalFormat:width:height:format:type:plane:] API doesn't exist in the iOS simulator. Earlier in the development cycle of the ANGLE backend for WebKit, we emulated this behavior using readbacks in Bug 205618. https://developer.apple.com/documentation/opengles/eaglcontext/2890259-teximageiosurface?language=objc This means that the backing texture of the {alpha:false} canvas is just a vanilla RGB texture. I'm not sure why readbacks are coming back with transparent alpha. Compositing of the canvas on the page works fine. 2) This workaround's been done at the WebKit level for some time on both iOS and macOS, just not on iOS in the ANGLE backend. It seems appropriate to do that as a short term fix. 3) It's possible ANGLE's Metal backend could address this. If you have a solid fix for Bug 215804 then let's certainly get it in to ANGLE - but I'm inclined to check as soon as possible whether the longstanding OpenGL texture/IOSurface interaction problems have been addressed in Metal's IOSurface integration APIs. 1701594 16 kkinnunen 2020-10-26 13:33:48 -0700 Thanks for the info and the observations! > This means that the backing texture of the {alpha:false} canvas is just a vanilla RGB texture. I'm not sure why readbacks are coming back with transparent alpha. The pbuffers with RGB internal format are mapped to native internal format RGBA textures on simulator. The initialise alpha should set the initial alpha ok but maybe the blit overrides this? 1701690 17 kbr 2020-10-26 16:57:38 -0700 (In reply to Kimmo Kinnunen from comment #16) > Thanks for the info and the observations! > > > This means that the backing texture of the {alpha:false} canvas is just a vanilla RGB texture. I'm not sure why readbacks are coming back with transparent alpha. > > The pbuffers with RGB internal format are mapped to native internal format > RGBA textures on simulator. The initialise alpha should set the initial > alpha ok but maybe the blit overrides this? I see - thanks for that reminder and clarification. Actually - the bug is in the alpha:false, antialias:false path. Setting antialias:true for the glNoAlpha context in https://trac.webkit.org/browser/webkit/trunk/LayoutTests/webgl/1.0.3/resources/webgl_test_files/conformance/textures/copy-tex-image-2d-formats.html makes the bug disappear. Would it be OK if I land this change and then we follow up to try to fix the underlying issue and remove this workaround for both macOS and iOS? 1701785 18 kkinnunen 2020-10-27 00:12:14 -0700 > Would it be OK if I land this change Not sure if you ask me or in general, but: certainly it's ok for my part. 1701786 19 412260 dino 2020-10-27 00:20:14 -0700 Comment on attachment 412260 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=412260&action=review >> Source/WebCore/platform/graphics/angle/ExtensionsGLANGLE.cpp:610 >> + if (!readingToPixelBufferObject && !attrs.alpha && (format == GraphicsContextGL::RGBA || format == GraphicsContextGL::BGRA) && (type == GraphicsContextGL::UNSIGNED_BYTE) && (state.boundReadFBO == m_context->m_fbo || (attrs.antialias && state.boundReadFBO == m_context->m_multisampleFBO))) > > Out of curiosity, did we not want to extend ANGLE workarounds around this area instead of existing more unstructured workarounds? Or would this be more task for downstream core developers? > > It'd help maybe making this more perfect in cases like preserveDrawingBuffer, which probably shouldn't need the workaround on grounds of the backing being normal texture instead of an iosurface > I guess part of the tedious work would be to understand what's the root cause of the problem. > > I did a bit of similar work in https://bugs.webkit.org/show_bug.cgi?id=215804 but it's not finished due to not being that important. It might or might not be same root cause but different trigger this time. E.g. in that bug, glclear is "buggy", but it might be that BlitFramebuffer is buggy. So in this case, it might be that read pixels work but then rendering from blitted-to texture would be black. Anyhow, this is an tangent to the actual question wrt do we want the fix eventually in ANGLE or not. I'm ok with this landing for now in WebKit, but I agree that it might want to live in ANGLE. I'll mark it as r+ and Ken can decide. 1701963 20 ews-feeder 2020-10-27 10:35:14 -0700 Committed r269051: <https://trac.webkit.org/changeset/269051> All reviewed patches have been landed. Closing bug and clearing flags on attachment 412260. 412260 2020-10-25 01:20:36 -0700 2020-10-27 10:35:15 -0700 Patch bug-209139-20201025012036.patch text/plain 5707 kbr U3VidmVyc2lvbiBSZXZpc2lvbjogMjY4OTUzCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggYWJlMGQwMWE1YjA5MjE1 MGRiZDVmM2ViYjljZTdlMzYzMTY3ODI1Yy4uZTM3NjNlNGNlYjc1OGRhNjExNGFjZDM5MWY5NWYz ZWM2ZThkYjk2NSAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDIyIEBACisyMDIwLTEwLTI1ICBLZW5u ZXRoIFJ1c3NlbGwgIDxrYnJAY2hyb21pdW0ub3JnPgorCisgICAgICAgIFsgaU9TIHdrMiBdIHdl YmdsLzEuMC4zL2NvbmZvcm1hbmNlL3RleHR1cmVzL2NvcHktdGV4LWltYWdlLTJkLWZvcm1hdHMu aHRtbCBpcyBmYWlsaW5nLgorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1 Zy5jZ2k/aWQ9MjA5MTM5CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisK KyAgICAgICAgT24gQU5HTEUgYmFja2VuZCwgdXNlIHdpcGVBbHBoYUNoYW5uZWxGcm9tUGl4ZWxz IG9uIGlPUyBmYW1pbHksCisgICAgICAgIHNpbWlsYXJseSB0byBtYWNPUywgd2hlbiByZWFkaW5n IGJhY2sgZnJvbSBhbHBoYTpmYWxzZSBXZWJHTAorICAgICAgICBjb250ZXh0cy4gT24gYm90aCBi YWNrZW5kcywgYXBwbHkgdGhpcyBvbmx5IGZvciBVTlNJR05FRF9CWVRFCisgICAgICAgIHJlYWRi YWNrcy4KKworICAgICAgICBDb3ZlcmVkIGJ5IGV4aXN0aW5nIFdlYkdMIGNvbmZvcm1hbmNlIHRl c3RzLgorCisgICAgICAgICogcGxhdGZvcm0vZ3JhcGhpY3MvYW5nbGUvRXh0ZW5zaW9uc0dMQU5H TEUuY3BwOgorICAgICAgICAoV2ViQ29yZTo6RXh0ZW5zaW9uc0dMQU5HTEU6OnJlYWRuUGl4ZWxz Um9idXN0QU5HTEUpOgorICAgICAgICAqIHBsYXRmb3JtL2dyYXBoaWNzL2FuZ2xlL0dyYXBoaWNz Q29udGV4dEdMQU5HTEUuY3BwOgorICAgICAgICAoV2ViQ29yZTo6R3JhcGhpY3NDb250ZXh0R0xP cGVuR0w6OnJlYWRQaXhlbHMpOgorCiAyMDIwLTEwLTI0ICBTaW1vbiBGcmFzZXIgIDxzaW1vbi5m cmFzZXJAYXBwbGUuY29tPgogCiAgICAgICAgIFtMRkMgRGlzcGxheV0gU2ltcGxpZnkgZGlzcGxh eSB0cmVlIGJ1aWxkaW5nCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9ncmFw aGljcy9hbmdsZS9FeHRlbnNpb25zR0xBTkdMRS5jcHAgYi9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9y bS9ncmFwaGljcy9hbmdsZS9FeHRlbnNpb25zR0xBTkdMRS5jcHAKaW5kZXggOTY5YmFmY2M4Nzlh MDkwYjAzY2JmNDVkMTg0Y2YxMjFmZjQxYWZiYS4uNGI0ZGZiYjNmNzNjNWYyZDI3YmYxN2Y5MjU1 OWE2Zjg3YWM1NWQyYiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vZ3JhcGhp Y3MvYW5nbGUvRXh0ZW5zaW9uc0dMQU5HTEUuY3BwCisrKyBiL1NvdXJjZS9XZWJDb3JlL3BsYXRm b3JtL2dyYXBoaWNzL2FuZ2xlL0V4dGVuc2lvbnNHTEFOR0xFLmNwcApAQCAtNTY4LDcgKzU2OCw3 IEBAIHZvaWQgRXh0ZW5zaW9uc0dMQU5HTEU6OmdldFBvaW50ZXJ2Um9idXN0QU5HTEVSb2J1c3RB TkdMRShHQ0dMZW51bSBwbmFtZSwgR0NHTHNpCiAgICAgZ2w6OkdldFBvaW50ZXJ2Um9idXN0QU5H TEVSb2J1c3RBTkdMRShwbmFtZSwgYnVmU2l6ZSwgbGVuZ3RoLCBwYXJhbXMpOwogfQogCi0jaWYg UExBVEZPUk0oTUFDKQorI2lmIFBMQVRGT1JNKE1BQykgfHwgUExBVEZPUk0oSU9TX0ZBTUlMWSkK IHN0YXRpYyB2b2lkIHdpcGVBbHBoYUNoYW5uZWxGcm9tUGl4ZWxzKGludCB3aWR0aCwgaW50IGhl aWdodCwgdW5zaWduZWQgY2hhciogcGl4ZWxzKQogewogICAgIC8vIFdlIGNhbiBhc3N1bWUgdGhp cyBkb2Vzbid0IG92ZXJmbG93IGJlY2F1c2UgdGhlIGNhbGxpbmcgZnVuY3Rpb25zCkBAIC02MDYs OCArNjA2LDggQEAgdm9pZCBFeHRlbnNpb25zR0xBTkdMRTo6cmVhZG5QaXhlbHNSb2J1c3RBTkdM RShpbnQgeCwgaW50IHksIEdDR0xzaXplaSB3aWR0aCwgR0MKICAgICAgICAgcmV0dXJuOwogICAg IH0KIAotI2lmIFBMQVRGT1JNKE1BQykKLSAgICBpZiAoIXJlYWRpbmdUb1BpeGVsQnVmZmVyT2Jq ZWN0ICYmICFhdHRycy5hbHBoYSAmJiAoZm9ybWF0ID09IEdyYXBoaWNzQ29udGV4dEdMOjpSR0JB IHx8IGZvcm1hdCA9PSBHcmFwaGljc0NvbnRleHRHTDo6QkdSQSkgJiYgKHN0YXRlLmJvdW5kUmVh ZEZCTyA9PSBtX2NvbnRleHQtPm1fZmJvIHx8IChhdHRycy5hbnRpYWxpYXMgJiYgc3RhdGUuYm91 bmRSZWFkRkJPID09IG1fY29udGV4dC0+bV9tdWx0aXNhbXBsZUZCTykpKQorI2lmIFBMQVRGT1JN KE1BQykgfHwgUExBVEZPUk0oSU9TX0ZBTUlMWSkKKyAgICBpZiAoIXJlYWRpbmdUb1BpeGVsQnVm ZmVyT2JqZWN0ICYmICFhdHRycy5hbHBoYSAmJiAoZm9ybWF0ID09IEdyYXBoaWNzQ29udGV4dEdM OjpSR0JBIHx8IGZvcm1hdCA9PSBHcmFwaGljc0NvbnRleHRHTDo6QkdSQSkgJiYgKHR5cGUgPT0g R3JhcGhpY3NDb250ZXh0R0w6OlVOU0lHTkVEX0JZVEUpICYmIChzdGF0ZS5ib3VuZFJlYWRGQk8g PT0gbV9jb250ZXh0LT5tX2ZibyB8fCAoYXR0cnMuYW50aWFsaWFzICYmIHN0YXRlLmJvdW5kUmVh ZEZCTyA9PSBtX2NvbnRleHQtPm1fbXVsdGlzYW1wbGVGQk8pKSkKICAgICAgICAgd2lwZUFscGhh Q2hhbm5lbEZyb21QaXhlbHMod2lkdGgsIGhlaWdodCwgc3RhdGljX2Nhc3Q8dW5zaWduZWQgY2hh cio+KGRhdGEpKTsKICNlbHNlCiAgICAgVU5VU0VEX1BBUkFNKHJlYWRpbmdUb1BpeGVsQnVmZmVy T2JqZWN0KTsKZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL2Fu Z2xlL0dyYXBoaWNzQ29udGV4dEdMQU5HTEUuY3BwIGIvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0v Z3JhcGhpY3MvYW5nbGUvR3JhcGhpY3NDb250ZXh0R0xBTkdMRS5jcHAKaW5kZXggNDY1ZWYyNDNm Mzg2ZWIyZjgyYzkyNjRlMTBiOGUzODA1ODFjOTFiMi4uZDI2ZDRlZjgxZTZiMzM0MzdjMTE1M2U4 OGViMzE5NjBmYjg3N2E3ZiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vZ3Jh cGhpY3MvYW5nbGUvR3JhcGhpY3NDb250ZXh0R0xBTkdMRS5jcHAKKysrIGIvU291cmNlL1dlYkNv cmUvcGxhdGZvcm0vZ3JhcGhpY3MvYW5nbGUvR3JhcGhpY3NDb250ZXh0R0xBTkdMRS5jcHAKQEAg LTQxNyw3ICs0MTcsNyBAQCB2b2lkIEdyYXBoaWNzQ29udGV4dEdMT3BlbkdMOjpyZWFkUGl4ZWxz KEdDR0xpbnQgeCwgR0NHTGludCB5LCBHQ0dMc2l6ZWkgd2lkdGgsCiAgICAgICAgIGdsOjpCaW5k RnJhbWVidWZmZXIoZnJhbWVidWZmZXJUYXJnZXQsIG1fbXVsdGlzYW1wbGVGQk8pOwogCiAjaWYg UExBVEZPUk0oTUFDKSB8fCBQTEFURk9STShJT1NfRkFNSUxZKQotICAgIGlmICghYXR0cnMuYWxw aGEgJiYgKGZvcm1hdCA9PSBHcmFwaGljc0NvbnRleHRHTDo6UkdCQSB8fCBmb3JtYXQgPT0gR3Jh cGhpY3NDb250ZXh0R0w6OkJHUkEpICYmIChtX3N0YXRlLmJvdW5kUmVhZEZCTyA9PSBtX2ZibyB8 fCAoYXR0cnMuYW50aWFsaWFzICYmIG1fc3RhdGUuYm91bmRSZWFkRkJPID09IG1fbXVsdGlzYW1w bGVGQk8pKSkKKyAgICBpZiAoIWF0dHJzLmFscGhhICYmIChmb3JtYXQgPT0gR3JhcGhpY3NDb250 ZXh0R0w6OlJHQkEgfHwgZm9ybWF0ID09IEdyYXBoaWNzQ29udGV4dEdMOjpCR1JBKSAmJiAodHlw ZSA9PSBHcmFwaGljc0NvbnRleHRHTDo6VU5TSUdORURfQllURSkgJiYgKG1fc3RhdGUuYm91bmRS ZWFkRkJPID09IG1fZmJvIHx8IChhdHRycy5hbnRpYWxpYXMgJiYgbV9zdGF0ZS5ib3VuZFJlYWRG Qk8gPT0gbV9tdWx0aXNhbXBsZUZCTykpKQogICAgICAgICB3aXBlQWxwaGFDaGFubmVsRnJvbVBp eGVscyh3aWR0aCwgaGVpZ2h0LCBzdGF0aWNfY2FzdDx1bnNpZ25lZCBjaGFyKj4oZGF0YSkpOwog I2VuZGlmCiB9CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cgYi9MYXlvdXRUZXN0 cy9DaGFuZ2VMb2cKaW5kZXggOWFhYmY3NjY3Zjk1NzBkMzBhNzkxY2NlNDY2NmQzZWFhNDEyMmY3 OC4uMTI3NDQyNGUwMjY3MGEyNDY0ZTA2ZjdlYzBkNDgwNzliY2RlYjJiOSAxMDA2NDQKLS0tIGEv TGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3RzL0NoYW5nZUxvZwpAQCAtMSwz ICsxLDE1IEBACisyMDIwLTEwLTI1ICBLZW5uZXRoIFJ1c3NlbGwgIDxrYnJAY2hyb21pdW0ub3Jn PgorCisgICAgICAgIFsgaU9TIHdrMiBdIHdlYmdsLzEuMC4zL2NvbmZvcm1hbmNlL3RleHR1cmVz L2NvcHktdGV4LWltYWdlLTJkLWZvcm1hdHMuaHRtbCBpcyBmYWlsaW5nLgorICAgICAgICBodHRw czovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MjA5MTM5CisKKyAgICAgICAgUmV2 aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgUmVtb3ZlIGlPUyBzdXBwcmVzc2lv biBmb3IKKyAgICAgICAgd2ViZ2wvMS4wLjMvY29uZm9ybWFuY2UvdGV4dHVyZXMvY29weS10ZXgt aW1hZ2UtMmQtZm9ybWF0cy5odG1sLgorCisgICAgICAgICogcGxhdGZvcm0vaW9zLXdrMi9UZXN0 RXhwZWN0YXRpb25zOgorCiAyMDIwLTEwLTI0ICBKZXIgTm9ibGUgIDxqZXIubm9ibGVAYXBwbGUu Y29tPgogCiAgICAgICAgIFtCaWdTdXJdIEFwcGVuZGluZyBhIG5ldyBXZWJNIGluaXQgc2VnbWVu dCBiZXR3ZWVuIENsdXN0ZXIgZWxlbWVudHMgdGhyb3dzIGFuIGVycm9yCmRpZmYgLS1naXQgYS9M YXlvdXRUZXN0cy9wbGF0Zm9ybS9pb3Mtd2syL1Rlc3RFeHBlY3RhdGlvbnMgYi9MYXlvdXRUZXN0 cy9wbGF0Zm9ybS9pb3Mtd2syL1Rlc3RFeHBlY3RhdGlvbnMKaW5kZXggMzhlMjUxOWIwZTkxNmU3 ZjdiY2NkMTVkNjBiMDJjZTEwMTY2NmQxOC4uZWE4NDhkODJlMzJkOGEzOWQ0ZTE4M2UxM2Q1ZmI3 NWUxYzMzYjc1NSAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0vaW9zLXdrMi9UZXN0 RXhwZWN0YXRpb25zCisrKyBiL0xheW91dFRlc3RzL3BsYXRmb3JtL2lvcy13azIvVGVzdEV4cGVj dGF0aW9ucwpAQCAtMTQwOSw4ICsxNDA5LDYgQEAgaW1wb3J0ZWQvdzNjL3dlYi1wbGF0Zm9ybS10 ZXN0cy9odG1sL3NlbWFudGljcy9lbWJlZGRlZC1jb250ZW50L3RoZS1pbWctZWxlbWVudC8KIAog d2Via2l0Lm9yZy9iLzIwOTIwNSBmYXN0L2V2ZW50cy9hdXRvc2Nyb2xsLWluLWlmcmFtZS5odG1s IFsgUGFzcyBGYWlsdXJlIF0KIAotd2Via2l0Lm9yZy9iLzIwOTEzOSB3ZWJnbC8xLjAuMy9jb25m b3JtYW5jZS90ZXh0dXJlcy9jb3B5LXRleC1pbWFnZS0yZC1mb3JtYXRzLmh0bWwgWyBQYXNzIEZh aWx1cmUgXQotCiB3ZWJraXQub3JnL2IvMjA5MDgwIGltcG9ydGVkL3czYy93ZWItcGxhdGZvcm0t dGVzdHMvY3NzL2Nzcy13cml0aW5nLW1vZGVzL3dtLXByb3BhZ2F0aW9uLWJvZHktc2Nyb2xsLW9m ZnNldC12ZXJ0aWNhbC1sci5odG1sIFsgRmFpbHVyZSBdCiB3ZWJraXQub3JnL2IvMjA5MDgwIGlt cG9ydGVkL3czYy93ZWItcGxhdGZvcm0tdGVzdHMvY3NzL2Nzcy13cml0aW5nLW1vZGVzL2Ficy1w b3Mtbm9uLXJlcGxhY2VkLXZsci0wMDMueGh0IFsgSW1hZ2VPbmx5RmFpbHVyZSBdCiB3ZWJraXQu b3JnL2IvMjA5MDgwIGltcG9ydGVkL3czYy93ZWItcGxhdGZvcm0tdGVzdHMvY3NzL2Nzcy13cml0 aW5nLW1vZGVzL2Ficy1wb3Mtbm9uLXJlcGxhY2VkLXZsci0wMDkueGh0IFsgSW1hZ2VPbmx5RmFp bHVyZSBdCg==