76699 2012-01-20 04:30:10 -0800 [WTR] Visited link tracking is not disabled properly 2012-05-16 02:58:16 -0700 1 1 1 Unclassified WebKit Tools / Tests 528+ (Nightly build) All All RESOLVED FIXED Qt, QtTriaged P2 Normal --- 76619 79666 1 ossy abecsi abecsi ap kbalazs ossy rafael.lobo zan oldest_to_newest 539417 0 ossy 2012-01-20 04:30:10 -0800 http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20WebKit2/r105504%20%2818894%29/results.html fast/block/margin-collapse/056.html: - RenderInline {A} at (0,0) size 56x21 [color=#0000EE] + RenderInline {A} at (0,0) size 56x21 [color=#551A8B] tables tests: - RenderInline {A} at (0,0) size 255x17 [color=#FFFF00] + RenderInline {A} at (0,0) size 255x17 [color=#999900] 539443 1 ossy 2012-01-20 05:50:20 -0800 And additionally it made 30 tests flakey ... :-/ 539449 2 ossy 2012-01-20 05:56:39 -0800 Shouldn't we roll out r105461? 539562 3 ap 2012-01-20 09:08:16 -0800 Changed results are quite likely to be improvements - did you find any actual regressions? Are you sure that this change made tests flaky? It seems quite unlikely. 539568 4 ossy 2012-01-20 09:14:14 -0800 (In reply to comment #3) > Changed results are quite likely to be improvements - did you find any actual regressions? Without this patch WK1 and WK2 results for these tests were same, but with this patch now the results are diverging. > Are you sure that this change made tests flaky? It seems quite unlikely. I'm absolutely sure. Check the following two test: r105460 - 1 failures 7 flakes 28 missing results - http://build.webkit.sed.hu/builders/x86-32%20Linux%20Qt%20Release%20WebKit2/builds/18859 r105461 - 9 failures 36 flakes 28 missing results - http://build.webkit.sed.hu/builders/x86-32%20Linux%20Qt%20Release%20WebKit2/builds/18860 539612 5 ap 2012-01-20 10:39:57 -0800 > fast/block/margin-collapse/056.html This test, and all tables/mozilla tests pass on Mac WebKit2. r105461 was a fix for an obvious bug. While it could have caused changes in test results, I don't see how it could introduce failures or flakiness. Perhaps there is some Qt specific code that used to neuter this bug, and now misbehaves? I'd like to help, but at this point, I think that some investigation from Qt side is needed. 541369 6 rafael.lobo 2012-01-24 09:00:05 -0800 (In reply to comment #5) > > fast/block/margin-collapse/056.html > > This test, and all tables/mozilla tests pass on Mac WebKit2. > > r105461 was a fix for an obvious bug. While it could have caused changes in test results, I don't see how it could introduce failures or flakiness. Perhaps there is some Qt specific code that used to neuter this bug, and now misbehaves? > > I'd like to help, but at this point, I think that some investigation from Qt side is needed. It seems very unlikely that those changes are behind this. I will take a look and try to figure out why this is happening, but when I ran just fast/block/margin-collapse/056.html I've got more diff than that pasted here (some widths were different). 541736 7 rafael.lobo 2012-01-24 14:56:02 -0800 It's a bit hard to realize what could be wrong here, I'm getting the same diff for head at r105461 and the previous one: https://gist.github.com/3d7570e546cb05e5be68 (the same results I get from current head). Ossy, do you have any idea how a test that only uses colors red, aqua and yellow could have produced the color #551A8B (purple)? Maybe it is blending colors? It would be nice if there could be a way for us to setup an environment that would produce the same results as our buildbot or if we could access it to run tests remotely. I'm using x86-64, maybe this can only be reproduced on 32... Any ideas? 544767 8 kbalazs 2012-01-29 14:58:03 -0800 (In reply to comment #7) > It's a bit hard to realize what could be wrong here, I'm getting the same diff for head at r105461 and the previous one: https://gist.github.com/3d7570e546cb05e5be68 (the same results I get from current head). Ossy, do you have any idea how a test that only uses colors red, aqua and yellow could have produced the color #551A8B (purple)? Maybe it is blending colors? It would be nice if there could be a way for us to setup an environment that would produce the same results as our buildbot or if we could access it to run tests remotely. I'm using x86-64, maybe this can only be reproduced on 32... Any ideas? You could use QBAT (http://webkit.sed.hu/blog/20101028/qtwebkit-builder-and-tester-virtual-machine) or copy the amazon ec2 vm (azbest_hu can help you with that). If you use a distro other than Debian or Ubuntu that can also be a reason for mismatches, so you could set up an Ubuntu chroot that should produce the same results. 588686 9 ossy 2012-03-27 02:35:19 -0700 I skipped tables/mozilla/marvin/backgr_simple-table-cell.html because it fails always on Qt-WK2 because of this bug - http://trac.webkit.org/changeset/112232. 612248 10 zan 2012-04-28 04:37:54 -0700 These failures occur simply because of visited links being colored as visited - #551A8B in fast/block/margin-collapse/056.html and #999900 in tables tests. Some tests then turn out as flaky as the links haven't been visited in the rerun due to a more narrow test list while others fail again because of specific tests being tested before again. To stop this problem from occurring, the back-forward list should be cleared after each test. A similar fix was required for the Gtk WebKit1 port at bug #71052. 622798 11 ossy 2012-05-14 03:54:33 -0700 Szilárd started working on fix suggested in Comment #10. 624059 12 abecsi 2012-05-15 07:23:06 -0700 Taking over as discussed with Ossy. The issue seems a bit more complex than described in Comment #10. Looks like the m_visitedLinkTable which is shared between the WebContext and WebProcess does not get cleaned properly / in time before style recalc, so links to previous tests appear in visited color. I'll try to find a solution and upload a patch. 624314 13 142012 abecsi 2012-05-15 11:33:01 -0700 Created attachment 142012 proposed patch 624334 14 abecsi 2012-05-15 11:46:42 -0700 (In reply to comment #13) > Created an attachment (id=142012) [details] > proposed patch Looks like before r105461 local files were not tracked in visited links properly because of the missing file:// prefix. This concealed the issue that WebKitTestRunner always tracked visited links, thus tests which have local links to previous tests began to fail, producing links with visited color. Not sure why this does not show up on Mac, though. WTR on Mac might not use PLATFORM_STRATEGIES. 624383 15 142012 jturcotte 2012-05-15 12:20:37 -0700 Comment on attachment 142012 proposed patch Looks good to me from what I've checked. It would be nice to know how this end up being disabled on the Mac port, assuming that it doesn't use PageGroup::is/addVisitedLink either. 625205 16 142012 abecsi 2012-05-16 02:58:07 -0700 Comment on attachment 142012 proposed patch Clearing flags on attachment: 142012 Committed r117248: <http://trac.webkit.org/changeset/117248> 625206 17 abecsi 2012-05-16 02:58:16 -0700 All reviewed patches have been landed. Closing bug. 142012 2012-05-15 11:33:01 -0700 2012-05-16 02:58:07 -0700 proposed patch 0001-Bug-76699-WTR-Visited-link-tracking-is-not-disabled-.patch text/plain 7369 abecsi RnJvbSAzNzc0MWNhYWViMzlkZDQ0Y2ZmYjNjZGQ1NTVlYTU3MWQ2YjNiNWU4IE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBBbmRyYXMgQmVjc2kgPGFuZHJhcy5iZWNzaUBub2tpYS5jb20+ CkRhdGU6IFR1ZSwgMTUgTWF5IDIwMTIgMjA6MzE6MzIgKzAyMDAKU3ViamVjdDogW1BBVENIXSBb QnVnIDc2Njk5XSBbV1RSXSBWaXNpdGVkIGxpbmsgdHJhY2tpbmcgaXMgbm90IGRpc2FibGVkCiBw cm9wZXJseQoKLS0tCiBMYXlvdXRUZXN0cy9DaGFuZ2VMb2cgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICB8ICAgIDkgKysrKysrKwogTGF5b3V0VGVzdHMvcGxhdGZvcm0vcXQtNS4wLXdrMi9T a2lwcGVkICAgICAgICAgICAgfCAgIDE1IC0tLS0tLS0tLS0tCiBTb3VyY2UvV2ViS2l0Mi9DaGFu Z2VMb2cgICAgICAgICAgICAgICAgICAgICAgICAgICB8ICAgMjggKysrKysrKysrKysrKysrKysr KysKIC4uLi9XZWJQcm9jZXNzL0luamVjdGVkQnVuZGxlL0luamVjdGVkQnVuZGxlLmNwcCAgIHwg ICAgMiArLQogU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9XZWJQcm9jZXNzLmNwcCAgICAgICAg ICAgfCAgICA0ICsrLQogU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9XZWJQcm9jZXNzLmggICAg ICAgICAgICAgfCAgICAzICsrLQogNiBmaWxlcyBjaGFuZ2VkLCA0MyBpbnNlcnRpb25zKCspLCAx OCBkZWxldGlvbnMoLSkKCmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cgYi9MYXlv dXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggNDczMDgwOS4uZTJjNTAzNCAxMDA2NDQKLS0tIGEvTGF5 b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3RzL0NoYW5nZUxvZwpAQCAtMSwzICsx LDEyIEBACisyMDEyLTA1LTE1ICBBbmRyYXMgQmVjc2kgIDxhbmRyYXMuYmVjc2lAbm9raWEuY29t PgorCisgICAgICAgIFtXVFJdIFZpc2l0ZWQgbGluayB0cmFja2luZyBpcyBub3QgZGlzYWJsZWQg cHJvcGVybHkKKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lk PTc2Njk5CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAg KiBwbGF0Zm9ybS9xdC01LjAtd2syL1NraXBwZWQ6IFVuc2tpcCBwYXNzaW5nIHRlc3RzLgorCiAy MDEyLTA1LTE1ICBDaHJpc3RvcGhlIER1bWV6ICA8Y2hyaXN0b3BoZS5kdW1lekBpbnRlbC5jb20+ CiAKICAgICAgICAgVW5yZXZpZXdlZCBFRkwgZ2FyZGVuaW5nLgpkaWZmIC0tZ2l0IGEvTGF5b3V0 VGVzdHMvcGxhdGZvcm0vcXQtNS4wLXdrMi9Ta2lwcGVkIGIvTGF5b3V0VGVzdHMvcGxhdGZvcm0v cXQtNS4wLXdrMi9Ta2lwcGVkCmluZGV4IDRjMzExMWIuLjUyMTI5ZjggMTAwNjQ0Ci0tLSBhL0xh eW91dFRlc3RzL3BsYXRmb3JtL3F0LTUuMC13azIvU2tpcHBlZAorKysgYi9MYXlvdXRUZXN0cy9w bGF0Zm9ybS9xdC01LjAtd2syL1NraXBwZWQKQEAgLTM5NiwxNyArMzk2LDYgQEAgZmFzdC9ibG9j ay9mbG9hdC9mbG9hdC1pbi1mbG9hdC1oaXQtdGVzdGluZy5odG1sCiBmYXN0L2R5bmFtaWMvMDEy Lmh0bWwKIHBsYXRmb3JtL3F0L3BsdWdpbnMvcXQtcXdpZGdldC1wbHVnaW4uaHRtbAogCi0jIFtR dF1bV0syXVJFR1JFU1NJT04ocjEwNTQ2MSk6IEl0IG1hZGUgOCB0ZXN0cyBmYWlsCi0jIGh0dHBz Oi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD03NjY5OQotZmFzdC9ibG9jay9tYXJn aW4tY29sbGFwc2UvMDU2Lmh0bWwKLXRhYmxlcy9tb3ppbGxhL21hcnZpbi9iYWNrZ3Jfc2ltcGxl LXRhYmxlLWNvbHVtbi5odG1sCi10YWJsZXMvbW96aWxsYS9tYXJ2aW4vYmFja2dyX3NpbXBsZS10 YWJsZS1yb3ctZ3JvdXAuaHRtbAotdGFibGVzL21vemlsbGEvbWFydmluL2JhY2tncl9zaW1wbGUt dGFibGUtcm93Lmh0bWwKLXRhYmxlcy9tb3ppbGxhX2V4cGVjdGVkX2ZhaWx1cmVzL21hcnZpbi9i YWNrZ3JfYm9yZGVyLXRhYmxlLWNvbHVtbi5odG1sCi10YWJsZXMvbW96aWxsYV9leHBlY3RlZF9m YWlsdXJlcy9tYXJ2aW4vYmFja2dyX2JvcmRlci10YWJsZS1yb3ctZ3JvdXAuaHRtbAotdGFibGVz L21vemlsbGFfZXhwZWN0ZWRfZmFpbHVyZXMvbWFydmluL2JhY2tncl9sYXllcnMtc2hvdy5odG1s Ci10YWJsZXMvbW96aWxsYV9leHBlY3RlZF9mYWlsdXJlcy9tYXJ2aW4vYmFja2dyX3Bvc2l0aW9u LXRhYmxlLWNvbHVtbi5odG1sCi0KICMgW1F0XVtXSzJdIHNjcm9sbGJhcnMvc2Nyb2xsLXJ0bC1v ci1idC1sYXllci5odG1sIGZhaWxzCiAjIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVn LmNnaT9pZD03Njk4Ngogc2Nyb2xsYmFycy9zY3JvbGwtcnRsLW9yLWJ0LWxheWVyLmh0bWwKQEAg LTU0MSwxMCArNTMwLDYgQEAgZnVsbHNjcmVlbi9mdWxsLXNjcmVlbi1pZnJhbWUtbGVnYWN5Lmh0 bWwKICMgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTgwMjA5CiBodHRw L3Rlc3RzL3NlY3VyaXR5L3hzcy1ERU5JRUQteHNsLWRvY3VtZW50LXNlY3VyaXR5T3JpZ2luLnht bAogCi0jIFtRdF1bV0syXVJFR1JFU1NJT04ocjEwNTQ2MSk6IEl0IG1hZGUgOCB0ZXN0cyBmYWls IGFuZCAzMCB0ZXN0cyBmbGFrZXkKLSMgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcu Y2dpP2lkPTc2Njk5Ci10YWJsZXMvbW96aWxsYS9tYXJ2aW4vYmFja2dyX3NpbXBsZS10YWJsZS1j ZWxsLmh0bWwKLQogIyBbUXRdW1dLMl0gRmFpbGluZyBodHRwL3Rlc3RzL3BsdWdpbnMgdGVzdHMg YWZ0ZXIgcmVlbmFibGVkIHBsdWdpbnMKICMgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19i dWcuY2dpP2lkPTgyODk1CiBodHRwL3Rlc3RzL3BsdWdpbnMvY3JlYXRlLXY4LXNjcmlwdC1vYmpl Y3RzLmh0bWwKZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJLaXQyL0NoYW5nZUxvZyBiL1NvdXJjZS9X ZWJLaXQyL0NoYW5nZUxvZwppbmRleCBiYzY3YjNiLi5mYTA0MWZmIDEwMDY0NAotLS0gYS9Tb3Vy Y2UvV2ViS2l0Mi9DaGFuZ2VMb2cKKysrIGIvU291cmNlL1dlYktpdDIvQ2hhbmdlTG9nCkBAIC0x LDMgKzEsMzEgQEAKKzIwMTItMDUtMTUgIEFuZHJhcyBCZWNzaSAgPGFuZHJhcy5iZWNzaUBub2tp YS5jb20+CisKKyAgICAgICAgW1dUUl0gVmlzaXRlZCBsaW5rIHRyYWNraW5nIGlzIG5vdCBkaXNh YmxlZCBwcm9wZXJseQorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5j Z2k/aWQ9NzY2OTkKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAg ICAgICBJZiBQTEFURk9STV9TVFJBVEVHSUVTIGlzIGVuYWJsZWQgdmlzaXRlZCBsaW5rIHRyYWNr aW5nIGlzIGRvbmUgdGhyb3VnaAorICAgICAgICBWaXNpdGVkTGlua1Byb3ZpZGVyIGFuZCBub3Qg dGhyb3VnaCBQYWdlR3JvdXAuCisgICAgICAgIEluamVjdGVkQnVuZGxlIHNldHMgdGhlIHNob3Vs ZFRyYWNrVmlzaXRlZExpbmtzIGZsYWcgb24gdGhlIFBhZ2VHcm91cCwKKyAgICAgICAgdGh1cyB0 aGUgc2hhcmVkIFZpc2l0ZWRMaW5rVGFibGUgZG9lcyBnZXQgcG9wdWxhdGVkIHdpdGggdmlzaXRl ZCBsaW5rcworICAgICAgICByZWdhcmRsZXNzIG9mIHRoZSBmbGFnLgorCisgICAgICAgIFRoZSBX ZWJQcm9jZXNzIHNob3VsZCBvbmx5IHRyYWNrIHZpc2l0ZWQgbGlua3MgaWYgdHJhY2tpbmcgaXMg c3BlY2lmaWNhbGx5CisgICAgICAgIGVuYWJsZWQgZm9yIGEgdGVzdCB0aHJvdWdoIFdUUidzIExh eW91dFRlc3RDb250cm9sbGVyLgorCisgICAgICAgIFRoaXMgcGF0Y2ggZml4ZXMgc2V2ZXJhbCBm bGFja3kgYW5kIGZhaWxpbmcgbGF5b3V0IHRlc3RzIG9uIFF0LVdLMi4KKworICAgICAgICAqIFdl YlByb2Nlc3MvSW5qZWN0ZWRCdW5kbGUvSW5qZWN0ZWRCdW5kbGUuY3BwOgorICAgICAgICAoV2Vi S2l0OjpJbmplY3RlZEJ1bmRsZTo6c2V0U2hvdWxkVHJhY2tWaXNpdGVkTGlua3MpOiBMZXQgdGhl IFdlYlByb2Nlc3MKKyAgICAgICAga25vdyBhYm91dCB2aXNpdGVkIGxpbmsgdHJhY2tpbmcuCisg ICAgICAgICogV2ViUHJvY2Vzcy9XZWJQcm9jZXNzLmNwcDoKKyAgICAgICAgKFdlYktpdDo6V2Vi UHJvY2Vzczo6V2ViUHJvY2Vzcyk6CisgICAgICAgIChXZWJLaXQ6OldlYlByb2Nlc3M6OnNldFNo b3VsZFRyYWNrVmlzaXRlZExpbmtzKToKKyAgICAgICAgKFdlYktpdDo6V2ViUHJvY2Vzczo6YWRk VmlzaXRlZExpbmspOgorICAgICAgICAqIFdlYlByb2Nlc3MvV2ViUHJvY2Vzcy5oOgorICAgICAg ICAoV2ViUHJvY2Vzcyk6CisKIDIwMTItMDUtMTUgIERpbnUgSmFjb2IgIDxkaW51LmphY29iQG5v a2lhLmNvbT4KIAogICAgICAgICBbUXRdW1dLMl0gZmFzdC9sb2FkZXIvY3JlYXRlLWZyYW1lLWlu LURPTUNvbnRlbnRMb2FkZWQuaHRtbCBjcmFzaGVzCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViS2l0 Mi9XZWJQcm9jZXNzL0luamVjdGVkQnVuZGxlL0luamVjdGVkQnVuZGxlLmNwcCBiL1NvdXJjZS9X ZWJLaXQyL1dlYlByb2Nlc3MvSW5qZWN0ZWRCdW5kbGUvSW5qZWN0ZWRCdW5kbGUuY3BwCmluZGV4 IDhlNGYyYTEuLmVmYzhiYTkgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJLaXQyL1dlYlByb2Nlc3Mv SW5qZWN0ZWRCdW5kbGUvSW5qZWN0ZWRCdW5kbGUuY3BwCisrKyBiL1NvdXJjZS9XZWJLaXQyL1dl YlByb2Nlc3MvSW5qZWN0ZWRCdW5kbGUvSW5qZWN0ZWRCdW5kbGUuY3BwCkBAIC0xMTEsNyArMTEx LDcgQEAgV2ViQ29ubmVjdGlvbiogSW5qZWN0ZWRCdW5kbGU6OndlYkNvbm5lY3Rpb25Ub1VJUHJv Y2VzcygpIGNvbnN0CiAKIHZvaWQgSW5qZWN0ZWRCdW5kbGU6OnNldFNob3VsZFRyYWNrVmlzaXRl ZExpbmtzKGJvb2wgc2hvdWxkVHJhY2tWaXNpdGVkTGlua3MpCiB7Ci0gICAgUGFnZUdyb3VwOjpz ZXRTaG91bGRUcmFja1Zpc2l0ZWRMaW5rcyhzaG91bGRUcmFja1Zpc2l0ZWRMaW5rcyk7CisgICAg V2ViUHJvY2Vzczo6c2hhcmVkKCkuc2V0U2hvdWxkVHJhY2tWaXNpdGVkTGlua3Moc2hvdWxkVHJh Y2tWaXNpdGVkTGlua3MpOwogfQogCiB2b2lkIEluamVjdGVkQnVuZGxlOjpyZW1vdmVBbGxWaXNp dGVkTGlua3MoKQpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9XZWJQcm9j ZXNzLmNwcCBiL1NvdXJjZS9XZWJLaXQyL1dlYlByb2Nlc3MvV2ViUHJvY2Vzcy5jcHAKaW5kZXgg ZjI5MjMzNC4uODFmYmJkZiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9X ZWJQcm9jZXNzLmNwcAorKysgYi9Tb3VyY2UvV2ViS2l0Mi9XZWJQcm9jZXNzL1dlYlByb2Nlc3Mu Y3BwCkBAIC0xMzAsNiArMTMwLDcgQEAgc3RhdGljIGNvbnN0IGRvdWJsZSBzaHV0ZG93blRpbWVv dXQgPSA2MDsKIFdlYlByb2Nlc3M6OldlYlByb2Nlc3MoKQogICAgIDogQ2hpbGRQcm9jZXNzKHNo dXRkb3duVGltZW91dCkKICAgICAsIG1faW5EaWRDbG9zZShmYWxzZSkKKyAgICAsIG1fc2hvdWxk VHJhY2tWaXNpdGVkTGlua3ModHJ1ZSkKICAgICAsIG1faGFzU2V0Q2FjaGVNb2RlbChmYWxzZSkK ICAgICAsIG1fY2FjaGVNb2RlbChDYWNoZU1vZGVsRG9jdW1lbnRWaWV3ZXIpCiAjaWYgVVNFKEFD Q0VMRVJBVEVEX0NPTVBPU0lUSU5HKSAmJiBQTEFURk9STShNQUMpCkBAIC0yNjEsNiArMjYyLDcg QEAgdm9pZCBXZWJQcm9jZXNzOjppbml0aWFsaXplV2ViUHJvY2Vzcyhjb25zdCBXZWJQcm9jZXNz Q3JlYXRpb25QYXJhbWV0ZXJzJiBwYXJhbWUKIAogdm9pZCBXZWJQcm9jZXNzOjpzZXRTaG91bGRU cmFja1Zpc2l0ZWRMaW5rcyhib29sIHNob3VsZFRyYWNrVmlzaXRlZExpbmtzKQogeworICAgIG1f c2hvdWxkVHJhY2tWaXNpdGVkTGlua3MgPSBzaG91bGRUcmFja1Zpc2l0ZWRMaW5rczsKICAgICBQ YWdlR3JvdXA6OnNldFNob3VsZFRyYWNrVmlzaXRlZExpbmtzKHNob3VsZFRyYWNrVmlzaXRlZExp bmtzKTsKIH0KIApAQCAtMzQ0LDcgKzM0Niw3IEBAIGJvb2wgV2ViUHJvY2Vzczo6aXNMaW5rVmlz aXRlZChMaW5rSGFzaCBsaW5rSGFzaCkgY29uc3QKIAogdm9pZCBXZWJQcm9jZXNzOjphZGRWaXNp dGVkTGluayhXZWJDb3JlOjpMaW5rSGFzaCBsaW5rSGFzaCkKIHsKLSAgICBpZiAoaXNMaW5rVmlz aXRlZChsaW5rSGFzaCkpCisgICAgaWYgKGlzTGlua1Zpc2l0ZWQobGlua0hhc2gpIHx8ICFtX3No b3VsZFRyYWNrVmlzaXRlZExpbmtzKQogICAgICAgICByZXR1cm47CiAgICAgY29ubmVjdGlvbigp LT5zZW5kKE1lc3NhZ2VzOjpXZWJDb250ZXh0OjpBZGRWaXNpdGVkTGlua0hhc2gobGlua0hhc2gp LCAwKTsKIH0KZGlmZiAtLWdpdCBhL1NvdXJjZS9XZWJLaXQyL1dlYlByb2Nlc3MvV2ViUHJvY2Vz cy5oIGIvU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9XZWJQcm9jZXNzLmgKaW5kZXggZTZmYWMy Yy4uYzdhMzdlNSAxMDA2NDQKLS0tIGEvU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9XZWJQcm9j ZXNzLmgKKysrIGIvU291cmNlL1dlYktpdDIvV2ViUHJvY2Vzcy9XZWJQcm9jZXNzLmgKQEAgLTEx NSw2ICsxMTUsNyBAQCBwdWJsaWM6CiAjZW5kaWYKICNlbmRpZgogICAgIAorICAgIHZvaWQgc2V0 U2hvdWxkVHJhY2tWaXNpdGVkTGlua3MoYm9vbCk7CiAgICAgdm9pZCBhZGRWaXNpdGVkTGluayhX ZWJDb3JlOjpMaW5rSGFzaCk7CiAgICAgYm9vbCBpc0xpbmtWaXNpdGVkKFdlYkNvcmU6OkxpbmtI YXNoKSBjb25zdDsKIApAQCAtMTY3LDcgKzE2OCw2IEBAIHByaXZhdGU6CiAgICAgdm9pZCBpbml0 aWFsaXplV2ViUHJvY2Vzcyhjb25zdCBXZWJQcm9jZXNzQ3JlYXRpb25QYXJhbWV0ZXJzJiwgQ29y ZUlQQzo6QXJndW1lbnREZWNvZGVyKik7CiAgICAgdm9pZCBwbGF0Zm9ybUluaXRpYWxpemVXZWJQ cm9jZXNzKGNvbnN0IFdlYlByb2Nlc3NDcmVhdGlvblBhcmFtZXRlcnMmLCBDb3JlSVBDOjpBcmd1 bWVudERlY29kZXIqKTsKICAgICB2b2lkIHBsYXRmb3JtVGVybWluYXRlKCk7Ci0gICAgdm9pZCBz ZXRTaG91bGRUcmFja1Zpc2l0ZWRMaW5rcyhib29sKTsKICAgICB2b2lkIHJlZ2lzdGVyVVJMU2No ZW1lQXNFbXB0eURvY3VtZW50KGNvbnN0IFN0cmluZyYpOwogICAgIHZvaWQgcmVnaXN0ZXJVUkxT Y2hlbWVBc1NlY3VyZShjb25zdCBTdHJpbmcmKSBjb25zdDsKICAgICB2b2lkIHNldERvbWFpblJl bGF4YXRpb25Gb3JiaWRkZW5Gb3JVUkxTY2hlbWUoY29uc3QgU3RyaW5nJikgY29uc3Q7CkBAIC0y NjEsNiArMjYxLDcgQEAgcHJpdmF0ZToKIAogICAgIC8vIEZJWE1FOiBUaGUgdmlzaXRlZCBsaW5r IHRhYmxlIHNob3VsZCBub3QgYmUgcGVyIHByb2Nlc3MuCiAgICAgVmlzaXRlZExpbmtUYWJsZSBt X3Zpc2l0ZWRMaW5rVGFibGU7CisgICAgYm9vbCBtX3Nob3VsZFRyYWNrVmlzaXRlZExpbmtzOwog CiAgICAgYm9vbCBtX2hhc1NldENhY2hlTW9kZWw7CiAgICAgQ2FjaGVNb2RlbCBtX2NhY2hlTW9k ZWw7Ci0tIAoxLjcuMTAKCg==