177938 2017-10-05 05:57:04 -0700 [GTK] 2 text tests failing since r222838 2017-10-16 12:22:04 -0700 1 1 1 Unclassified WebKit WebKitGTK WebKit Nightly Build Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 magomez tpopela bugs-noreply cgarcia fred.wang mcatanzaro tpopela oldest_to_newest 1357002 0 magomez 2017-10-05 05:57:04 -0700 fast/text/woff2-totalsfntsize.html [ ImageOnlyFailure ] fast/text/woff2.html [ ImageOnlyFailure ] 1357542 1 tpopela 2017-10-06 03:01:26 -0700 OK, so the problem was that in OptionGTK.cmake we had: set(USE_WOFF2 ON) but in should use: SET_AND_EXPOSE_TO_BUILD(USE_WOFF2 ON) but the tests should be fixed now with bug 177804 (r222907): 01:36:41.601 16255 [22052/46035] fast/text/woff2-totalsfntsize.html passed unexpectedly 01:36:41.600 16368 worker/2 fast/text/woff2-totalsfntsize.html passed 01:36:41.949 16255 [22061/46035] fast/text/woff2.html passed unexpectedly 01:36:41.949 16368 worker/2 fast/text/woff2.html passed It will mark them as passing.. 1357543 2 323005 tpopela 2017-10-06 03:08:51 -0700 Created attachment 323005 Patch 1357544 3 mcatanzaro 2017-10-06 03:12:18 -0700 (In reply to Tomas Popela from comment #1) > OK, so the problem was that in OptionGTK.cmake we had: > > set(USE_WOFF2 ON) > > but in should use: > > SET_AND_EXPOSE_TO_BUILD(USE_WOFF2 ON) Is this fixed yet...? 1357546 4 tpopela 2017-10-06 03:15:30 -0700 (In reply to Michael Catanzaro from comment #3) > (In reply to Tomas Popela from comment #1) > > OK, so the problem was that in OptionGTK.cmake we had: > > > > set(USE_WOFF2 ON) > > > > but in should use: > > > > SET_AND_EXPOSE_TO_BUILD(USE_WOFF2 ON) > > Is this fixed yet...? It is, but differently.. now it is defined with WEBKIT_OPTION_DEFINE().. 1357547 5 323005 tpopela 2017-10-06 03:16:58 -0700 Comment on attachment 323005 Patch Clearing flags on attachment: 323005 Committed r222973: <http://trac.webkit.org/changeset/222973> 1357548 6 tpopela 2017-10-06 03:17:03 -0700 All reviewed patches have been landed. Closing bug. 1357558 7 mcatanzaro 2017-10-06 03:50:53 -0700 What about the 2.18 branch? And there's also USE_CAIRO and USE_XDGMIME that are set but not exposed to build. Hmmmmm. 1357560 8 tpopela 2017-10-06 03:58:34 -0700 (In reply to Michael Catanzaro from comment #7) > What about the 2.18 branch? That one needs to be fixed.. > And there's also USE_CAIRO and USE_XDGMIME that are set but not exposed to > build. Hmmmmm. Exactly.. Man you should really be on IRC (and especially on FreeNode) ;).. 1357561 9 mcatanzaro 2017-10-06 04:41:02 -0700 We need to figure this out. USE(WOFF2) needs fixed for 2.18. And it's hard to believe that USE(CAIRO) is not working, but I'm not sure how. 1357562 10 tpopela 2017-10-06 04:53:19 -0700 (In reply to Michael Catanzaro from comment #9) > We need to figure this out. > > USE(WOFF2) needs fixed for 2.18. > For 2.18 it will be fixed by backporting r222838 there together with: diff --git a/Source/cmake/OptionsGTK.cmake b/Source/cmake/OptionsGTK.cmake index 0caa7dce0b8..b84419fc581 100644 --- a/Source/cmake/OptionsGTK.cmake +++ b/Source/cmake/OptionsGTK.cmake @@ -56,7 +56,7 @@ WEBKIT_OPTION_BEGIN() include(GStreamerDefinitions) -set(USE_CAIRO ON) +SET_AND_EXPOSE_TO_BUILD(USE_CAIRO ON) set(USE_WOFF2 ON) set(USE_XDGMIME ON) SET_AND_EXPOSE_TO_BUILD(USE_GCRYPT TRUE) > And it's hard to believe that USE(CAIRO) is not working, but I'm not sure > how. it is not as it is the same situation as here.. It needs to use the SET_AND_EXPOSE_TO_BUILD() and we have to remove its override from Platform.h 1360891 11 mcatanzaro 2017-10-16 12:22:04 -0700 Let's leave USE_CAIRO alone for now... it's awkward that it's set separately in Platform.h and Options[GTK,WPE].cmake, but there is not currently any correctness issue and this will be a difficult cleanup, for another day. Not really sure how we want to handle it. It doesn't make much sense to me to move it from Platform.h to CMake while leaving all the other stuff defined next to it in Platform.h, but it's true that it's clearly needed in CMake as well. Maybe we should be moving a lot more into CMake. I don't think r222838 is needed for 2.18, since we don't support disabling WOFF2 there. In fact, backporting that would introduce the regression, so it must not be backported. So I think we're good here... sort of. 323005 2017-10-06 03:08:51 -0700 2017-10-06 03:16:58 -0700 Patch bug-177938-20171006120849.patch text/plain 1635 tpopela U3VidmVyc2lvbiBSZXZpc2lvbjogMjIyOTcyCmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggNTMwYWRlOTdhYzM2ZTlhMjdmMGI5 Mzg1MzJjNThlZGM2MDE3Njk2ZS4uZjZmYTgxMjAwMmIwYWNiYzE1OTIwNjM4YTM5OTFiODhlM2M2 MTA5NCAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDE0IEBACisyMDE3LTEwLTA2ICBUb21hcyBQb3BlbGEgIDx0 cG9wZWxhQHJlZGhhdC5jb20+CisKKyAgICAgICAgW0dUS10gMiB0ZXh0IHRlc3RzIGZhaWxpbmcg c2luY2UgcjIyMjgzOAorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5j Z2k/aWQ9MTc3OTM4CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAg ICAgICAgVGhlc2UgdGVzdHMgYXJlIHBhc3NpbmcgYWZ0ZXIgcjIyMjkwNworCisgICAgICAgICog cGxhdGZvcm0vZ3RrL1Rlc3RFeHBlY3RhdGlvbnM6CisKIDIwMTctMTAtMDUgIENhcmxvcyBHYXJj aWEgQ2FtcG9zICA8Y2dhcmNpYUBpZ2FsaWEuY29tPgogCiAgICAgICAgIFtHVEtdW1dQRV0gQWxs IHJlc291cmNlIGxvYWQgc3RhdGlzdGljcyB0ZXN0cyBhZGRlZCBpbiByMjEyMTgzIGNyYXNoIGlu IEdUSyBib3RzLCB0aW1lb3V0IGluIEdUSyBhbmQgV1BFIGJvdHMgc2luY2UgcjIxOTA0OQpkaWZm IC0tZ2l0IGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0vZ3RrL1Rlc3RFeHBlY3RhdGlvbnMgYi9MYXlv dXRUZXN0cy9wbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9ucwppbmRleCA2MDY5MTFhNjg2NjI4 ZjdjOGMxZjhhOTNmNzg5MTA5ZGQzOGZmZWU4Li4wOTkxYTVkZTJjOTgyZWQ1Y2M2Nzk0NTg3Njg2 NDI3MjRkMWI5ODgzIDEwMDY0NAotLS0gYS9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvVGVzdEV4 cGVjdGF0aW9ucworKysgYi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9u cwpAQCAtMzM2Miw5ICszMzYyLDYgQEAgd2Via2l0Lm9yZy9iLzE3NzkzNiBpbXBvcnRlZC9ibGlu ay9lZGl0aW5nL3NlbGVjdGlvbi9vZmZzZXQtZnJvbS1wb2ludC1jb21wbGV4LXMKIHdlYmtpdC5v cmcvYi8xNzc5MzcgZmFzdC90ZXh0L3NvZnQtaHlwaGVuLW1pbi1wcmVmZXJyZWQtd2lkdGguaHRt bCBbIEltYWdlT25seUZhaWx1cmUgXQogd2Via2l0Lm9yZy9iLzE3NzkzNyBmYXN0L3RleHQvdW5r bm93bi1jaGFyLW5vdGRlZi5odG1sIFsgSW1hZ2VPbmx5RmFpbHVyZSBdCiAKLXdlYmtpdC5vcmcv Yi8xNzc5MzggZmFzdC90ZXh0L3dvZmYyLXRvdGFsc2ZudHNpemUuaHRtbCBbIEltYWdlT25seUZh aWx1cmUgXQotd2Via2l0Lm9yZy9iLzE3NzkzOCBmYXN0L3RleHQvd29mZjIuaHRtbCBbIEltYWdl T25seUZhaWx1cmUgXQotCiB3ZWJraXQub3JnL2IvMTc3OTM5IGVkaXRpbmcvcGFzdGVib2FyZC9k YXRhLXRyYW5zZmVyLWdldC1kYXRhLW5vbi1ub3JtYWxpemVkLXR5cGVzLmh0bWwgWyBGYWlsdXJl IF0KIAogIy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8v Ly8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8vLy8K