163439 2016-10-14 01:58:47 -0700 [Cairo] REGRESSION(r206481): GraphicsContext3D::ImageExtractor fails to extract images 2026-01-12 09:05:09 -0800 1 1 1 Unclassified WebKit New Bugs WebKit Nightly Build Unspecified Unspecified REOPENED Regression P2 Normal --- 1 zan zan bugs-noreply commit-queue dino graouts kondapallykalyan sabouhallawa oldest_to_newest 1240173 0 zan 2016-10-14 01:58:47 -0700 [Cairo] GraphicsContext3D::ImageExtractor fails to extract images 1240178 1 291597 zan 2016-10-14 02:04:31 -0700 Created attachment 291597 Patch 1240180 2 291597 cgarcia 2016-10-14 02:06:56 -0700 Comment on attachment 291597 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=291597&action=review > Source/WebCore/ChangeLog:14 > + In the Cairo implementation of GraphicsContext3D::ImageExtractor, > + don't check for frame completeness at index 0. This information > + is now cached only after the frame for that index is decoded and > + marked as completed, which is done after this check. > + > + Becuase of this the current check forces extractImage() to return > + early and abort WebGL texture uploads from image sources. If this a recent regression after the image decoder refactorings? it would be great to mention a revision in the commit message, so that I'll know in the future that I don't need this in any stable branch. 1240185 3 zan 2016-10-14 02:13:39 -0700 It was probably introduced in r206481. https://trac.webkit.org/changeset/206481 1240187 4 291600 zan 2016-10-14 02:16:55 -0700 Created attachment 291600 Patch for landing 1240189 5 291600 zan 2016-10-14 02:20:34 -0700 Comment on attachment 291600 Patch for landing Clearing flags on attachment: 291600 Committed r207332: <http://trac.webkit.org/changeset/207332> 1240190 6 zan 2016-10-14 02:20:42 -0700 All reviewed patches have been landed. Closing bug. 1240315 7 sabouhallawa 2016-10-14 09:58:02 -0700 (In reply to comment #3) > It was probably introduced in r206481. > https://trac.webkit.org/changeset/206481 Yes this is true. In this change I made frameIsCompleteAtIndex() reads whatever in the cache without having to ask the decoder about the completeness of the frame and without trying to cache a new frame. The reason for this change is this information is mainly useful to the ImageFrameCache. It can be sued to decide whether we can use the current frame or we need to try recaching a new one. For other places of the code like GraphicsContext3D::ImageExtractor, it should not matter whether the frame is complete or not. The completeness of a frame represents whether the frame image can fully or partially displayed. And we need to display it even if it is partially decoded. 1240318 8 291597 sabouhallawa 2016-10-14 10:00:11 -0700 Comment on attachment 291597 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=291597&action=review > Source/WebCore/ChangeLog:11 > + marked as completed, which is done after this check. What are the repro steps of this bug? Can't we include a test case in this patch? 1240319 9 291597 sabouhallawa 2016-10-14 10:11:32 -0700 Comment on attachment 291597 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=291597&action=review > Source/WebCore/platform/graphics/cairo/GraphicsContext3DCairo.cpp:226 > - if (!decoder.frameCount() || !decoder.frameIsCompleteAtIndex(0)) > + if (!decoder.frameCount()) The function GraphicsContext3D::ImageExtractor::extractImage() in WebCore/platform/graphics/cg/GraphicsContext3D.cpp does not have the check decoder.frameIsCompleteAtIndex(). But the function GraphicsContext3D::ImageExtractor::extractImage() in WebCore/platform/graphics/efl/GraphicsContext3D.cpp does have the check decoder.frameIsCompleteAtIndex(). if (!decoder.frameCount() || !decoder.frameIsCompleteAtIndex(0)) return false; I think we need to apply the same fix there as well. Maybe we need to move some of the code of this function to the cross platform file: WebCore/platform/graphics/GraphicsContext3D.cpp to avoid repeating the code and avoid fixing the same bug in different places. 1240321 10 sabouhallawa 2016-10-14 10:15:41 -0700 Reopening the bug to apply the same fix to the elf port and to add a test if possible. 1241358 11 zan 2016-10-18 00:05:07 -0700 (In reply to comment #8) > Comment on attachment 291597 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=291597&action=review > > > Source/WebCore/ChangeLog:11 > > + marked as completed, which is done after this check. > > What are the repro steps of this bug? Can't we include a test case in this > patch? This is reproducible by running any WebGL example that does texture uploads from an image source (specifically image, canvas or video elements). There's plenty of tests for that, but the EFL and GTK+ ports probably didn't catch them because they either don't run those tests, or don't compare pixel results from these tests. This problem should have stored a relevant error on the GraphicsContext3D object, but I suppose again the ports don't run a test which would propagate this problem into a test failure. 291597 2016-10-14 02:04:31 -0700 2016-10-14 02:16:49 -0700 Patch bug-163439-20161014020103.patch text/plain 2035 zan U3VidmVyc2lvbiBSZXZpc2lvbjogMjA3MzIxCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggMWZiNWNiMDJjMzQ4ODhl ZGNkMGE5ZGZjMTRjNzM5MWQ1MDI3NzY5OC4uOTM0ZGM1ZmYzOWYwZmI1ZTJlYWZkYzZkMWY5ODI4 M2E2YjQ2YWYzNiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDIxIEBACisyMDE2LTEwLTE0ICBaYW4g RG9iZXJzZWsgIDx6ZG9iZXJzZWtAaWdhbGlhLmNvbT4KKworICAgICAgICBbQ2Fpcm9dIEdyYXBo aWNzQ29udGV4dDNEOjpJbWFnZUV4dHJhY3RvciBmYWlscyB0byBleHRyYWN0IGltYWdlcworICAg ICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTYzNDM5CisKKyAg ICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgSW4gdGhlIENhaXJv IGltcGxlbWVudGF0aW9uIG9mIEdyYXBoaWNzQ29udGV4dDNEOjpJbWFnZUV4dHJhY3RvciwKKyAg ICAgICAgZG9uJ3QgY2hlY2sgZm9yIGZyYW1lIGNvbXBsZXRlbmVzcyBhdCBpbmRleCAwLiBUaGlz IGluZm9ybWF0aW9uCisgICAgICAgIGlzIG5vdyBjYWNoZWQgb25seSBhZnRlciB0aGUgZnJhbWUg Zm9yIHRoYXQgaW5kZXggaXMgZGVjb2RlZCBhbmQKKyAgICAgICAgbWFya2VkIGFzIGNvbXBsZXRl ZCwgd2hpY2ggaXMgZG9uZSBhZnRlciB0aGlzIGNoZWNrLgorCisgICAgICAgIEJlY3Vhc2Ugb2Yg dGhpcyB0aGUgY3VycmVudCBjaGVjayBmb3JjZXMgZXh0cmFjdEltYWdlKCkgdG8gcmV0dXJuCisg ICAgICAgIGVhcmx5IGFuZCBhYm9ydCBXZWJHTCB0ZXh0dXJlIHVwbG9hZHMgZnJvbSBpbWFnZSBz b3VyY2VzLgorCisgICAgICAgICogcGxhdGZvcm0vZ3JhcGhpY3MvY2Fpcm8vR3JhcGhpY3NDb250 ZXh0M0RDYWlyby5jcHA6CisgICAgICAgIChXZWJDb3JlOjpHcmFwaGljc0NvbnRleHQzRDo6SW1h Z2VFeHRyYWN0b3I6OmV4dHJhY3RJbWFnZSk6CisKIDIwMTYtMTAtMTMgIEFsZXggQ2hyaXN0ZW5z ZW4gIDxhY2hyaXN0ZW5zZW5Ad2Via2l0Lm9yZz4KIAogICAgICAgICBIb3N0cyBvZiBVUkxzIHdp dGggbm9uLXNwZWNpYWwgc2NoZW1lcyBzaG91bGQgYmUgY2FzZS1zZW5zaXRpdmUsIGFuZCBub24t QVNDSUkgY2hhcmFjdGVycyBpbiBzdWNoIGhvc3RzIHNob3VsZCBiZSBwdW55Y29kZS1lbmNvZGVk CmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9wbGF0Zm9ybS9ncmFwaGljcy9jYWlyby9HcmFw aGljc0NvbnRleHQzRENhaXJvLmNwcCBiL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNz L2NhaXJvL0dyYXBoaWNzQ29udGV4dDNEQ2Fpcm8uY3BwCmluZGV4IDYzNmZlMWY4YzA4YmQ3ZDNi NWFhZmQyYjFiODA5ZjM4ZGJiMzE0MjAuLmMyZTY3ZDAwNDMxNzY0ZWVmYjRkYmFjY2NhNzA1NDA0 NWEwNzk0NjEgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL2Nh aXJvL0dyYXBoaWNzQ29udGV4dDNEQ2Fpcm8uY3BwCisrKyBiL1NvdXJjZS9XZWJDb3JlL3BsYXRm b3JtL2dyYXBoaWNzL2NhaXJvL0dyYXBoaWNzQ29udGV4dDNEQ2Fpcm8uY3BwCkBAIC0yMjMsNyAr MjIzLDcgQEAgYm9vbCBHcmFwaGljc0NvbnRleHQzRDo6SW1hZ2VFeHRyYWN0b3I6OmV4dHJhY3RJ bWFnZShib29sIHByZW11bHRpcGx5QWxwaGEsIGJvb2wKIAogICAgIGlmIChtX2ltYWdlLT5kYXRh KCkpIHsKICAgICAgICAgZGVjb2Rlci5zZXREYXRhKG1faW1hZ2UtPmRhdGEoKSwgdHJ1ZSk7Ci0g ICAgICAgIGlmICghZGVjb2Rlci5mcmFtZUNvdW50KCkgfHwgIWRlY29kZXIuZnJhbWVJc0NvbXBs ZXRlQXRJbmRleCgwKSkKKyAgICAgICAgaWYgKCFkZWNvZGVyLmZyYW1lQ291bnQoKSkKICAgICAg ICAgICAgIHJldHVybiBmYWxzZTsKICAgICAgICAgbV9pbWFnZVN1cmZhY2UgPSBkZWNvZGVyLmNy ZWF0ZUZyYW1lSW1hZ2VBdEluZGV4KDApOwogICAgIH0gZWxzZSB7Cg== 291600 2016-10-14 02:16:55 -0700 2016-10-14 02:16:55 -0700 Patch for landing bug-163439-20161014021327.patch text/plain 2108 zan U3VidmVyc2lvbiBSZXZpc2lvbjogMjA3MzMxCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggOWJhN2Y2Y2YyYTczYWY2 NTlhNGNjOTRiMzFkNGYwYTk0OTRlMmUxMC4uNDhjYzgwM2NhNWY5MTczNGU1NzkzMDVlYjAyN2E5 OGY0ODk0YmM4ZSAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDI0IEBACisyMDE2LTEwLTE0ICBaYW4g RG9iZXJzZWsgIDx6ZG9iZXJzZWtAaWdhbGlhLmNvbT4KKworICAgICAgICBbQ2Fpcm9dIEdyYXBo aWNzQ29udGV4dDNEOjpJbWFnZUV4dHJhY3RvciBmYWlscyB0byBleHRyYWN0IGltYWdlcworICAg ICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTYzNDM5CisKKyAg ICAgICAgUmV2aWV3ZWQgYnkgQ2FybG9zIEdhcmNpYSBDYW1wb3MuCisKKyAgICAgICAgSW4gdGhl IENhaXJvIGltcGxlbWVudGF0aW9uIG9mIEdyYXBoaWNzQ29udGV4dDNEOjpJbWFnZUV4dHJhY3Rv ciwKKyAgICAgICAgZG9uJ3QgY2hlY2sgZm9yIGZyYW1lIGNvbXBsZXRlbmVzcyBhdCBpbmRleCAw LiBUaGlzIGluZm9ybWF0aW9uCisgICAgICAgIGlzIG5vdyBjYWNoZWQgb25seSBhZnRlciB0aGUg ZnJhbWUgZm9yIHRoYXQgaW5kZXggaXMgZGVjb2RlZCBhbmQKKyAgICAgICAgbWFya2VkIGFzIGNv bXBsZXRlZCwgd2hpY2ggaXMgZG9uZSBhZnRlciB0aGlzIGNoZWNrLgorCisgICAgICAgIEJlY3Vh c2Ugb2YgdGhpcyB0aGUgY3VycmVudCBjaGVjayBmb3JjZXMgZXh0cmFjdEltYWdlKCkgdG8gcmV0 dXJuCisgICAgICAgIGVhcmx5IGFuZCBhYm9ydCBXZWJHTCB0ZXh0dXJlIHVwbG9hZHMgZnJvbSBp bWFnZSBzb3VyY2VzLgorCisgICAgICAgIFRoaXMgYnVnIHdhcyBtb3N0IGxpa2VseSBleHBvc2Vk IGFmdGVyIHRoZSByZWZhY3RvcmluZyBpbiByMjA2NDgxLgorICAgICAgICBodHRwczovL3RyYWMu d2Via2l0Lm9yZy9jaGFuZ2VzZXQvMjA2NDgxCisKKyAgICAgICAgKiBwbGF0Zm9ybS9ncmFwaGlj cy9jYWlyby9HcmFwaGljc0NvbnRleHQzRENhaXJvLmNwcDoKKyAgICAgICAgKFdlYkNvcmU6Okdy YXBoaWNzQ29udGV4dDNEOjpJbWFnZUV4dHJhY3Rvcjo6ZXh0cmFjdEltYWdlKToKKwogMjAxNi0x MC0xNCAgWW91ZW5uIEZhYmxldCAgPHlvdWVubkBhcHBsZS5jb20+CiAKICAgICAgICAgTWVkaWFS ZXNvdXJjZUxvYWRlciBzaG91bGQgdXNlIENhY2hlZFJlc291cmNlOjpyZXNzb3VyY2VFcnJvciB0 byBkbyBDT1JTIGNoZWNrcwpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYkNvcmUvcGxhdGZvcm0vZ3Jh cGhpY3MvY2Fpcm8vR3JhcGhpY3NDb250ZXh0M0RDYWlyby5jcHAgYi9Tb3VyY2UvV2ViQ29yZS9w bGF0Zm9ybS9ncmFwaGljcy9jYWlyby9HcmFwaGljc0NvbnRleHQzRENhaXJvLmNwcAppbmRleCA2 MzZmZTFmOGMwOGJkN2QzYjVhYWZkMmIxYjgwOWYzOGRiYjMxNDIwLi5jMmU2N2QwMDQzMTc2NGVl ZmI0ZGJhY2NjYTcwNTQwNDVhMDc5NDYxIDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29yZS9wbGF0 Zm9ybS9ncmFwaGljcy9jYWlyby9HcmFwaGljc0NvbnRleHQzRENhaXJvLmNwcAorKysgYi9Tb3Vy Y2UvV2ViQ29yZS9wbGF0Zm9ybS9ncmFwaGljcy9jYWlyby9HcmFwaGljc0NvbnRleHQzRENhaXJv LmNwcApAQCAtMjIzLDcgKzIyMyw3IEBAIGJvb2wgR3JhcGhpY3NDb250ZXh0M0Q6OkltYWdlRXh0 cmFjdG9yOjpleHRyYWN0SW1hZ2UoYm9vbCBwcmVtdWx0aXBseUFscGhhLCBib29sCiAKICAgICBp ZiAobV9pbWFnZS0+ZGF0YSgpKSB7CiAgICAgICAgIGRlY29kZXIuc2V0RGF0YShtX2ltYWdlLT5k YXRhKCksIHRydWUpOwotICAgICAgICBpZiAoIWRlY29kZXIuZnJhbWVDb3VudCgpIHx8ICFkZWNv ZGVyLmZyYW1lSXNDb21wbGV0ZUF0SW5kZXgoMCkpCisgICAgICAgIGlmICghZGVjb2Rlci5mcmFt ZUNvdW50KCkpCiAgICAgICAgICAgICByZXR1cm4gZmFsc2U7CiAgICAgICAgIG1faW1hZ2VTdXJm YWNlID0gZGVjb2Rlci5jcmVhdGVGcmFtZUltYWdlQXRJbmRleCgwKTsKICAgICB9IGVsc2Ugewo=