143760 2015-04-15 06:22:15 -0700 [W32] Compile-time assertion failure: RenderBlock_should_stay_small 2015-05-18 23:31:36 -0700 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 133028 1 lrn1986 webkit-unassigned cgarcia lrn1986 mcatanzaro mcrha tpopela oldest_to_newest 1085455 0 lrn1986 2015-04-15 06:22:15 -0700 Presumably as a result of the way bitfields are packed on Windows (-mms-bitfields is the default these days). 1085535 1 250805 lrn1986 2015-04-15 09:54:08 -0700 Created attachment 250805 Move enum LineLayoutPath in RenderBlock This removes the bitfield interruption, allowing more efficient packing. Otherwise compile-time assertion fails: In file included from ../webkitgtk-2.4.8/Source/WTF/wtf/PossiblyNull.h:29:0, from ../webkitgtk-2.4.8/Source/WTF/wtf/FastMalloc.h:27, from ../webkitgtk-2.4.8/Source/WebCore/config.h:74, from ../webkitgtk-2.4.8/Source/WebCore/rendering/RenderBlock.cpp:24: ../webkitgtk-2.4.8/Source/WTF/wtf/Assertions.h:326:35: error: static assertion failed: RenderBlock_should_stay_small #define COMPILE_ASSERT(exp, name) static_assert((exp), #name) ^ ../webkitgtk-2.4.8/Source/WebCore/rendering/RenderBlock.cpp:88:1: note: in expansion of macro 'COMPILE_ASSERT' COMPILE_ASSERT(sizeof(RenderBlock) == sizeof(SameSizeAsRenderBlock), RenderBlock_should_stay_small); ^ 1085574 2 250805 darin 2015-04-15 10:34:45 -0700 Comment on attachment 250805 Move enum LineLayoutPath in RenderBlock View in context: https://bugs.webkit.org/attachment.cgi?id=250805&action=review > Source/WebCore/rendering/RenderBlock.h:-614 > - enum LineLayoutPath { UndeterminedPath, SimpleLinesPath, LineBoxesPath, ForceLineBoxesPath }; It’s very strange the defining an enum type (not a data member, just a type) causes a “bitfield interruption”. Which compiler has this problem? 1085580 3 lrn1986 2015-04-15 10:41:01 -0700 (In reply to comment #2) > Comment on attachment 250805 [details] > Move enum LineLayoutPath in RenderBlock > > View in context: > https://bugs.webkit.org/attachment.cgi?id=250805&action=review > > > Source/WebCore/rendering/RenderBlock.h:-614 > > - enum LineLayoutPath { UndeterminedPath, SimpleLinesPath, LineBoxesPath, ForceLineBoxesPath }; > > It’s very strange the defining an enum type (not a data member, just a type) > causes a “bitfield interruption”. Which compiler has this problem? i686-w64-mingw32-gcc-4.9.2 is what i've used The patch originated from https://bugs.webkit.org/show_bug.cgi?id=133028 , where the original author used 4.8.something 1086471 4 mcrha 2015-04-17 16:20:39 -0700 (In reply to comment #3) > The patch originated from https://bugs.webkit.org/show_bug.cgi?id=133028 , > where the original author used 4.8.something I still use this change with the webkit-2.4 branch at revision 182543. @LRN: so you've split my change (and other changes), but still want to receive credits for it? I'd say it's not a good habit by any means... :-/ 1086498 5 lrn1986 2015-04-17 17:36:34 -0700 (In reply to comment #4) > (In reply to comment #3) > > The patch originated from https://bugs.webkit.org/show_bug.cgi?id=133028 , > > where the original author used 4.8.something > > I still use this change with the webkit-2.4 branch at revision 182543. > > @LRN: so you've split my change (and other changes), but still want to > receive credits for it? I'd say it's not a good habit by any means... :-/ The problem with the original bug was that it had multiple conflicting patches from 2 different authors, which made it difficult for upstream to sort through. So i've split the changes into separate, manageable chunks and filed a lot bugs. I do admit that i haven't tracked copyrights or credit though (partially because many patches had no authorship information in them, i.e. were plain diffs; partially because i was more interested in figuring out which changes to apply and which to drop), and it did bother me somewhat when i was filing the bugs (especially on patches with lots of non-trivial code that i'm incapable of writing myself). So if you feel slighted, do comment where appropriate, and hopefully webkit devs will do something (no idea what's the protocol for such things; i assume there is one, because webkit seems to have lots of unapparent rules about committing). Ironically, half of these changes probably won't make it past 2.4 anyway :( 1086514 6 mcatanzaro 2015-04-17 20:22:31 -0700 Just please credit the original author somewhere in the changelog entry. :) "Patch by Person" or "Based on work by Person," something like that. 1095847 7 250805 cgarcia 2015-05-18 23:31:01 -0700 Comment on attachment 250805 Move enum LineLayoutPath in RenderBlock I don't understand it either, but it's harmless in any case. 1095848 8 cgarcia 2015-05-18 23:31:36 -0700 Committed to 2.4 http://trac.webkit.org/changeset/184548 250805 2015-04-15 09:54:08 -0700 2015-05-18 23:31:01 -0700 Move enum LineLayoutPath in RenderBlock Move-enum-LineLayoutPath-in-RenderBlock.patch text/plain 3861 lrn1986 RnJvbSAxNzM1OWIxODVhZWZjNTYyMjI1NmM2NjdhY2UwYjI1ODA0NDZlZmNlIE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiA9P1VURi04P3E/PUQwPUEwPUQxPTgzPUQxPTgxPUQwPUJCPUQw PUIwPUQwPUJEPTIwPUQwPTk4PUQwPUI2PUQwPUIxPz0KID0/VVRGLTg/cT89RDE9ODM9RDA9QkI9 RDA9QjA9RDE9ODI9RDA9QkU9RDA9QjI/PSA8bHJuMTk4NkBnbWFpbC5jb20+CkRhdGU6IE1vbiwg MTMgQXByIDIwMTUgMTA6MjE6NDIgKzAwMDAKU3ViamVjdDogW1BBVENIXSBNb3ZlIGVudW0gTGlu ZUxheW91dFBhdGggaW4gUmVuZGVyQmxvY2sKClRoaXMgcmVtb3ZlcyB0aGUgYml0ZmllbGQgaW50 ZXJydXB0aW9uLCBhbGxvd2luZyBtb3JlIGVmZmljaWVudApwYWNraW5nLgpPdGhlcndpc2UgY29t cGlsZS10aW1lIGFzc2VydGlvbiBmYWlsczoKSW4gZmlsZSBpbmNsdWRlZCBmcm9tIC4uL3dlYmtp dGd0ay0yLjQuOC9Tb3VyY2UvV1RGL3d0Zi9Qb3NzaWJseU51bGwuaDoyOTowLAogICAgICAgICAg ICAgICAgIGZyb20gLi4vd2Via2l0Z3RrLTIuNC44L1NvdXJjZS9XVEYvd3RmL0Zhc3RNYWxsb2Mu aDoyNywKICAgICAgICAgICAgICAgICBmcm9tIC4uL3dlYmtpdGd0ay0yLjQuOC9Tb3VyY2UvV2Vi Q29yZS9jb25maWcuaDo3NCwKICAgICAgICAgICAgICAgICBmcm9tIC4uL3dlYmtpdGd0ay0yLjQu OC9Tb3VyY2UvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyQmxvY2suY3BwOjI0OgouLi93ZWJraXRn dGstMi40LjgvU291cmNlL1dURi93dGYvQXNzZXJ0aW9ucy5oOjMyNjozNTogZXJyb3I6IHN0YXRp YyBhc3NlcnRpb24gZmFpbGVkOiBSZW5kZXJCbG9ja19zaG91bGRfc3RheV9zbWFsbAogI2RlZmlu ZSBDT01QSUxFX0FTU0VSVChleHAsIG5hbWUpIHN0YXRpY19hc3NlcnQoKGV4cCksICNuYW1lKQog ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIF4KLi4vd2Via2l0Z3RrLTIuNC44L1Nv dXJjZS9XZWJDb3JlL3JlbmRlcmluZy9SZW5kZXJCbG9jay5jcHA6ODg6MTogbm90ZTogaW4gZXhw YW5zaW9uIG9mIG1hY3JvICdDT01QSUxFX0FTU0VSVCcKIENPTVBJTEVfQVNTRVJUKHNpemVvZihS ZW5kZXJCbG9jaykgPT0gc2l6ZW9mKFNhbWVTaXplQXNSZW5kZXJCbG9jayksIFJlbmRlckJsb2Nr X3Nob3VsZF9zdGF5X3NtYWxsKTsKIF4KCmh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVn LmNnaT9pZD0xNDM3NjAKLS0tCiBTb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cgICAgICAgICAgICAg ICB8IDI3ICsrKysrKysrKysrKysrKysrKysrKysrKysrKwogU291cmNlL1dlYkNvcmUvcmVuZGVy aW5nL1JlbmRlckJsb2NrLmggfCAgMiArLQogMiBmaWxlcyBjaGFuZ2VkLCAyOCBpbnNlcnRpb25z KCspLCAxIGRlbGV0aW9uKC0pCgpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9n IGIvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCmluZGV4IGUzMzViNjIuLjdhMWExNTQgMTAwNjQ0 Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZworKysgYi9Tb3VyY2UvV2ViQ29yZS9DaGFu Z2VMb2cKQEAgLTEsNSArMSwzMiBAQAogMjAxNS0wNC0xNSAg0KDRg9GB0LvQsNC9INCY0LbQsdGD 0LvQsNGC0L7QsiAgPGxybjE5ODZAZ21haWwuY29tPgogCisgICAgICAgIE1vdmUgZW51bSBMaW5l TGF5b3V0UGF0aCBpbiBSZW5kZXJCbG9jaworCisgICAgICAgIFRoaXMgcmVtb3ZlcyB0aGUgYml0 ZmllbGQgaW50ZXJydXB0aW9uLCBhbGxvd2luZyBtb3JlIGVmZmljaWVudAorICAgICAgICBwYWNr aW5nLgorICAgICAgICBPdGhlcndpc2UgY29tcGlsZS10aW1lIGFzc2VydGlvbiBmYWlsczoKKyAg ICAgICAgSW4gZmlsZSBpbmNsdWRlZCBmcm9tIC4uL3dlYmtpdGd0ay0yLjQuOC9Tb3VyY2UvV1RG L3d0Zi9Qb3NzaWJseU51bGwuaDoyOTowLAorICAgICAgICAgICAgICAgICAgICAgICAgIGZyb20g Li4vd2Via2l0Z3RrLTIuNC44L1NvdXJjZS9XVEYvd3RmL0Zhc3RNYWxsb2MuaDoyNywKKyAgICAg ICAgICAgICAgICAgICAgICAgICBmcm9tIC4uL3dlYmtpdGd0ay0yLjQuOC9Tb3VyY2UvV2ViQ29y ZS9jb25maWcuaDo3NCwKKyAgICAgICAgICAgICAgICAgICAgICAgICBmcm9tIC4uL3dlYmtpdGd0 ay0yLjQuOC9Tb3VyY2UvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyQmxvY2suY3BwOjI0OgorICAg ICAgICAuLi93ZWJraXRndGstMi40LjgvU291cmNlL1dURi93dGYvQXNzZXJ0aW9ucy5oOjMyNjoz NTogZXJyb3I6IHN0YXRpYyBhc3NlcnRpb24gZmFpbGVkOiBSZW5kZXJCbG9ja19zaG91bGRfc3Rh eV9zbWFsbAorICAgICAgICAgI2RlZmluZSBDT01QSUxFX0FTU0VSVChleHAsIG5hbWUpIHN0YXRp Y19hc3NlcnQoKGV4cCksICNuYW1lKQorICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgIF4KKyAgICAgICAgLi4vd2Via2l0Z3RrLTIuNC44L1NvdXJjZS9XZWJDb3JlL3Jl bmRlcmluZy9SZW5kZXJCbG9jay5jcHA6ODg6MTogbm90ZTogaW4gZXhwYW5zaW9uIG9mIG1hY3Jv ICdDT01QSUxFX0FTU0VSVCcKKyAgICAgICAgIENPTVBJTEVfQVNTRVJUKHNpemVvZihSZW5kZXJC bG9jaykgPT0gc2l6ZW9mKFNhbWVTaXplQXNSZW5kZXJCbG9jayksIFJlbmRlckJsb2NrX3Nob3Vs ZF9zdGF5X3NtYWxsKTsKKyAgICAgICAgIF4KKworICAgICAgICBbVzMyXSBDb21waWxlLXRpbWUg YXNzZXJ0aW9uIGZhaWx1cmU6IFJlbmRlckJsb2NrX3Nob3VsZF9zdGF5X3NtYWxsCisgICAgICAg IGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0xNDM3NjAKKworICAgICAg ICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBObyBuZXcgdGVzdHMgKE9P UFMhKS4KKworICAgICAgICAqIHJlbmRlcmluZy9SZW5kZXJCbG9jay5oOgorCisyMDE1LTA0LTE1 ICDQoNGD0YHQu9Cw0L0g0JjQttCx0YPQu9Cw0YLQvtCyICA8bHJuMTk4NkBnbWFpbC5jb20+CisK ICAgICAgICAgQWRkIG1pc3NpbmcgaW5jbHVkZXMgdG8gRmlsZVN5c3RlbUd0awogCiAgICAgICAg IE5lZWRlZCBmb3IgR2V0TW9kdWxlSGFuZGxlRXhBKCkgJiBDby4KZGlmZiAtLWdpdCBhL1NvdXJj ZS9XZWJDb3JlL3JlbmRlcmluZy9SZW5kZXJCbG9jay5oIGIvU291cmNlL1dlYkNvcmUvcmVuZGVy aW5nL1JlbmRlckJsb2NrLmgKaW5kZXggNWFhMjllMC4uYzI5YjA3OCAxMDA2NDQKLS0tIGEvU291 cmNlL1dlYkNvcmUvcmVuZGVyaW5nL1JlbmRlckJsb2NrLmgKKysrIGIvU291cmNlL1dlYkNvcmUv cmVuZGVyaW5nL1JlbmRlckJsb2NrLmgKQEAgLTYxMSw4ICs2MTEsOCBAQCBwdWJsaWM6CiAgICAg dW5zaWduZWQgbV9iZWluZ0Rlc3Ryb3llZCA6IDE7CiAgICAgdW5zaWduZWQgbV9oYXNNYXJrdXBU cnVuY2F0aW9uIDogMTsKICAgICB1bnNpZ25lZCBtX2hhc0JvcmRlck9yUGFkZGluZ0xvZ2ljYWxX aWR0aENoYW5nZWQgOiAxOwotICAgIGVudW0gTGluZUxheW91dFBhdGggeyBVbmRldGVybWluZWRQ YXRoLCBTaW1wbGVMaW5lc1BhdGgsIExpbmVCb3hlc1BhdGgsIEZvcmNlTGluZUJveGVzUGF0aCB9 OwogICAgIHVuc2lnbmVkIG1fbGluZUxheW91dFBhdGggOiAyOworICAgIGVudW0gTGluZUxheW91 dFBhdGggeyBVbmRldGVybWluZWRQYXRoLCBTaW1wbGVMaW5lc1BhdGgsIExpbmVCb3hlc1BhdGgs IEZvcmNlTGluZUJveGVzUGF0aCB9OwogICAgIAogICAgIC8vIFJlbmRlclJ1YnlCYXNlIG9iamVj dHMgbmVlZCB0byBiZSBhYmxlIHRvIHNwbGl0IGFuZCBtZXJnZSwgbW92aW5nIHRoZWlyIGNoaWxk cmVuIGFyb3VuZAogICAgIC8vIChjYWxsaW5nIG1vdmVDaGlsZFRvLCBtb3ZlQWxsQ2hpbGRyZW5U bywgYW5kIG1ha2VDaGlsZHJlbk5vbklubGluZSkuCi0tIAoxLjguNS4z