120682 2013-09-04 08:04:51 -0700 New test inspector-protocol/page/archive.html added in r154828 fails on EFL, Qt, GTK 2019-11-11 17:04:43 -0800 1 1 1 Unclassified WebKit Web Inspector 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED InRadar P2 Normal --- 87008 119774 1 zarvai yurys allan.jensen commit-queue hi inspector-bugzilla-changes joepeck ossy webkit-bug-importer yurys oldest_to_newest 924428 0 zarvai 2013-09-04 08:04:51 -0700 Test fails with timeout or diff. --- /mnt/buildbot/efl-linux-slave-1/efl-linux-64-release-wk1/build/layout-test-results/inspector-protocol/page/archive-expected.txt +++ /mnt/buildbot/efl-linux-slave-1/efl-linux-64-release-wk1/build/layout-test-results/inspector-protocol/page/archive-actual.txt @@ -1,2 +1,3 @@ -PASS: Received archive data. +CONSOLE MESSAGE: line 4: TypeError: undefined is not an object (evaluating 'event.result.data') +FAIL: Timed out waiting for notifyDone to be called 924442 1 zarvai 2013-09-04 08:38:00 -0700 Skipped on Qt in http://trac.webkit.org/changeset/155042. 924577 2 joepeck 2013-09-04 13:38:46 -0700 My guess is this is because the port does not support ENABLE(WEB_ARCHIVE). If that is the case, a minor tweak to the test can bail with the Page.archive(…) result error, and port specific results can be checked in for ports not implementing ENABLE(WEB_ARCHIVE). Is there a page archive that these ports do support? 1587078 3 382756 yurys 2019-11-04 12:13:50 -0800 Created attachment 382756 Patch 1587088 4 382756 hi 2019-11-04 12:37:08 -0800 Comment on attachment 382756 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=382756&action=review > LayoutTests/platform/gtk/TestExpectations:-2331 > -webkit.org/b/120682 inspector/page/archive.html [ Timeout ] If `Page.archive` isn't implemented (other than sending back "Not supported" as an error), why aren't we just skipping the test on these platforms? 1587106 5 382756 yurys 2019-11-04 13:05:30 -0800 Comment on attachment 382756 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=382756&action=review >> LayoutTests/platform/gtk/TestExpectations:-2331 >> -webkit.org/b/120682 inspector/page/archive.html [ Timeout ] > > If `Page.archive` isn't implemented (other than sending back "Not supported" as an error), why aren't we just skipping the test on these platforms? It still may crash or timeout for various reasons and we don't want to have a protocol method that crashes even if the functionality is not implemented. 1588435 6 382756 hi 2019-11-07 15:15:05 -0800 Comment on attachment 382756 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=382756&action=review >>> LayoutTests/platform/gtk/TestExpectations:-2331 >>> -webkit.org/b/120682 inspector/page/archive.html [ Timeout ] >> >> If `Page.archive` isn't implemented (other than sending back "Not supported" as an error), why aren't we just skipping the test on these platforms? > > It still may crash or timeout for various reasons and we don't want to have a protocol method that crashes even if the functionality is not implemented. How exactly would this crash? There are already general tests for protocol behavior. If it's "Not supported", unless you yourself put something inside of it that causes a crash, it shouldn't ever cause a crash. My concern with this is that it now requires any other ports that may want to add Web Inspector support to have to send back `errorString = "Not supported"_s;` instead of something that may be more descriptive/useful for that port. If this truly isn't supported, then we should instead add support for `featureGuard` in the protocol JSON on a per-command/per-event basis (right now it can only be done per-domain), allowing you to describe what platforms/ports actually can do this. 1588450 7 382756 yurys 2019-11-07 15:31:47 -0800 Comment on attachment 382756 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=382756&action=review >>>> LayoutTests/platform/gtk/TestExpectations:-2331 >>>> -webkit.org/b/120682 inspector/page/archive.html [ Timeout ] >>> >>> If `Page.archive` isn't implemented (other than sending back "Not supported" as an error), why aren't we just skipping the test on these platforms? >> >> It still may crash or timeout for various reasons and we don't want to have a protocol method that crashes even if the functionality is not implemented. > > How exactly would this crash? There are already general tests for protocol behavior. If it's "Not supported", unless you yourself put something inside of it that causes a crash, it shouldn't ever cause a crash. > > My concern with this is that it now requires any other ports that may want to add Web Inspector support to have to send back `errorString = "Not supported"_s;` instead of something that may be more descriptive/useful for that port. > > If this truly isn't supported, then we should instead add support for `featureGuard` in the protocol JSON on a per-command/per-event basis (right now it can only be done per-domain), allowing you to describe what platforms/ports actually can do this. Similar to any other test there is a possibility it may crash. Given that the crashes are rarely intentional I'd rather not speculate about how it can happen and have the tests catch it. This is also why we keep many tests marked as Failure instead of just skipping it. If any port decides to change error message it can easily change the message either for other platforms too or just platform specific expectations. Using guards would add even more fragmentation to the platforms and be more hassle to support. Ideally all commands should be the same across platforms unless inspected feature is platforms specific. In any case this is not the approach which is currently taken and it deserves a bigger discussion while this patch is cleaning up custom expectations and reduces time required to run the tests (Timeout always takes longest possible time to finish). 1588495 8 382756 hi 2019-11-07 16:45:29 -0800 Comment on attachment 382756 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=382756&action=review >>>>> LayoutTests/platform/gtk/TestExpectations:-2331 >>>>> -webkit.org/b/120682 inspector/page/archive.html [ Timeout ] >>>> >>>> If `Page.archive` isn't implemented (other than sending back "Not supported" as an error), why aren't we just skipping the test on these platforms? >>> >>> It still may crash or timeout for various reasons and we don't want to have a protocol method that crashes even if the functionality is not implemented. >> >> How exactly would this crash? There are already general tests for protocol behavior. If it's "Not supported", unless you yourself put something inside of it that causes a crash, it shouldn't ever cause a crash. >> >> My concern with this is that it now requires any other ports that may want to add Web Inspector support to have to send back `errorString = "Not supported"_s;` instead of something that may be more descriptive/useful for that port. >> >> If this truly isn't supported, then we should instead add support for `featureGuard` in the protocol JSON on a per-command/per-event basis (right now it can only be done per-domain), allowing you to describe what platforms/ports actually can do this. > > Similar to any other test there is a possibility it may crash. Given that the crashes are rarely intentional I'd rather not speculate about how it can happen and have the tests catch it. This is also why we keep many tests marked as Failure instead of just skipping it. > > If any port decides to change error message it can easily change the message either for other platforms too or just platform specific expectations. > > Using guards would add even more fragmentation to the platforms and be more hassle to support. Ideally all commands should be the same across platforms unless inspected feature is platforms specific. In any case this is not the approach which is currently taken and it deserves a bigger discussion while this patch is cleaning up custom expectations and reduces time required to run the tests (Timeout always takes longest possible time to finish). If another port/platform comes along, ideally they shouldn't have to change any test file to match their code, only test expectations. As such, how about we change the test to just log the error message, and have GTKs expectations file contain that log so it passes with the expected error message? LayoutTests/inspector/page/archive.html: ``` InspectorProtocol.sendCommand("Page.archive", {}, function(event) { if (event.error) { ProtocolTest.fail(`Page.archive sent back error '${event.error}'.`); return; } ... }); ``` LayoutTests/platform/gtk/inspector/page/archive-expected.txt" ``` FAIL: Page.archive sent back error 'Not supported'. ``` As an aside, I think this is exactly the kind of situation where we don't want the protocol to be the same across platforms. Although `Page.archive` may not be platform specific in principle, in implementation it only actually works on platforms with `ENABLE_CF` and `ENABLE_WEB_ARCHIVE`. What this means right now is that the frontend will show the archive button even on platforms where `Page.archive` is not supported, meaning that it will do nothing if any user clicks on it. We shouldn't even be showing the button at all for platforms that don't implement its functionality. The way Web Inspector has done this is to check whether protocol has that particular command/event exists or not, which can be controlled by build flags (as well as other things, like debuggableType/targetType as of <https://webkit.org/b/200384>). 1588728 9 382756 yurys 2019-11-08 11:11:01 -0800 Comment on attachment 382756 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=382756&action=review >>>>>> LayoutTests/platform/gtk/TestExpectations:-2331 >>>>>> -webkit.org/b/120682 inspector/page/archive.html [ Timeout ] >>>>> >>>>> If `Page.archive` isn't implemented (other than sending back "Not supported" as an error), why aren't we just skipping the test on these platforms? >>>> >>>> It still may crash or timeout for various reasons and we don't want to have a protocol method that crashes even if the functionality is not implemented. >>> >>> How exactly would this crash? There are already general tests for protocol behavior. If it's "Not supported", unless you yourself put something inside of it that causes a crash, it shouldn't ever cause a crash. >>> >>> My concern with this is that it now requires any other ports that may want to add Web Inspector support to have to send back `errorString = "Not supported"_s;` instead of something that may be more descriptive/useful for that port. >>> >>> If this truly isn't supported, then we should instead add support for `featureGuard` in the protocol JSON on a per-command/per-event basis (right now it can only be done per-domain), allowing you to describe what platforms/ports actually can do this. >> >> Similar to any other test there is a possibility it may crash. Given that the crashes are rarely intentional I'd rather not speculate about how it can happen and have the tests catch it. This is also why we keep many tests marked as Failure instead of just skipping it. >> >> If any port decides to change error message it can easily change the message either for other platforms too or just platform specific expectations. >> >> Using guards would add even more fragmentation to the platforms and be more hassle to support. Ideally all commands should be the same across platforms unless inspected feature is platforms specific. In any case this is not the approach which is currently taken and it deserves a bigger discussion while this patch is cleaning up custom expectations and reduces time required to run the tests (Timeout always takes longest possible time to finish). > > If another port/platform comes along, ideally they shouldn't have to change any test file to match their code, only test expectations. As such, how about we change the test to just log the error message, and have GTKs expectations file contain that log so it passes with the expected error message? > > LayoutTests/inspector/page/archive.html: > ``` > InspectorProtocol.sendCommand("Page.archive", {}, function(event) { > if (event.error) { > ProtocolTest.fail(`Page.archive sent back error '${event.error}'.`); > return; > } > > ... > }); > ``` > > LayoutTests/platform/gtk/inspector/page/archive-expected.txt" > ``` > FAIL: Page.archive sent back error 'Not supported'. > ``` > > As an aside, I think this is exactly the kind of situation where we don't want the protocol to be the same across platforms. Although `Page.archive` may not be platform specific in principle, in implementation it only actually works on platforms with `ENABLE_CF` and `ENABLE_WEB_ARCHIVE`. What this means right now is that the frontend will show the archive button even on platforms where `Page.archive` is not supported, meaning that it will do nothing if any user clicks on it. We shouldn't even be showing the button at all for platforms that don't implement its functionality. The way Web Inspector has done this is to check whether protocol has that particular command/event exists or not, which can be controlled by build flags (as well as other things, like debuggableType/targetType as of <https://webkit.org/b/200384>). Just printing FAIL: followed by actual error sounds fine to me. The reason I used assert is just to make it clear for a reader unfamiliar with this code that the only expected error is 'Not implemented'. But it is only marginally better than just printing output, so I don't mind changing it that way. Updated the test. As for 'archive' in particular, it could perhaps be more generic 'archive' in a platform-specific format. In any case there are very few platform-specific commands/features. If you don't want to show corresponding UI components you could add information about supported platforms to the protocol definition to hide the components on start. 1588729 10 383145 yurys 2019-11-08 11:11:21 -0800 Created attachment 383145 Patch 1588967 11 383145 hi 2019-11-08 21:45:08 -0800 Comment on attachment 383145 Patch r=me 1588970 12 383145 commit-queue 2019-11-08 22:28:04 -0800 Comment on attachment 383145 Patch Clearing flags on attachment: 383145 Committed r252305: <https://trac.webkit.org/changeset/252305> 1588971 13 commit-queue 2019-11-08 22:28:06 -0800 All reviewed patches have been landed. Closing bug. 1589393 14 webkit-bug-importer 2019-11-11 17:04:43 -0800 <rdar://problem/57099549> 382756 2019-11-04 12:13:50 -0800 2019-11-08 11:11:20 -0800 Patch bug-120682-20191104121349.patch text/plain 3543 yurys U3VidmVyc2lvbiBSZXZpc2lvbjogMjUyMDA2CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggYzg4ODAyZTdkZGRjMDcxYmFkZTAw YzJjMzMwOGZjNjg0YzgyZjZhNC4uMmQwZmIzZjUyNmZmZTU1NzIwYTc4MmU1NzgzNjIxZGI0MDZi NDhkMCAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDE3IEBACisyMDE5LTExLTA0ICBZdXJ5IFNlbWlraGF0c2t5 ICA8eXVyeXNAY2hyb21pdW0ub3JnPgorCisgICAgICAgIE5ldyB0ZXN0IGluc3BlY3Rvci1wcm90 b2NvbC9wYWdlL2FyY2hpdmUuaHRtbCBhZGRlZCBpbiByMTU0ODI4IGZhaWxzIG9uIEVGTCwgUXQs IEdUSworICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTIw NjgyCisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgR2Vu ZXJhdGUgY3VzdG9tIGV4cGVjdGF0aW9ucyBmb3IgaW5zcGVjdG9yLXByb3RvY29sL3BhZ2UvYXJj aGl2ZS5odG1sIG9uIHRoZSBwbGF0Zm9ybXMKKyAgICAgICAgd2hlcmUgUGFnZS5hcmNoaXZlIGlz IG5vdCBzdXBwb3J0ZWQgeWV0LgorCisgICAgICAgICogaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS5o dG1sOiAnTm90IHN1cHBvcnRlZCcgaXMgYSB2YWxpZCByZXNwb25zZSBvbiBzb21lIHBsYXRmb3Jt cy4KKyAgICAgICAgKiBwbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9uczoKKyAgICAgICAgKiBw bGF0Zm9ybS9ndGsvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS1leHBlY3RlZC50eHQ6IEFkZGVkLgor CiAyMDE5LTExLTA0ICBDaHJpcyBEdW1leiAgPGNkdW1lekBhcHBsZS5jb20+CiAKICAgICAgICAg TWVkaWFSZWNvcmRlciBzaG91bGQgbm90IHByZXZlbnQgZW50ZXJpbmcgdGhlIGJhY2svZm9yd2Fy ZCBjYWNoZQpkaWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS5o dG1sIGIvTGF5b3V0VGVzdHMvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS5odG1sCmluZGV4IDgwNjkx ZDBhNDJhYWZlYTEyYTdjYzUxZWRjZjMyODc5ZmFhZDVjNzAuLjA4YTI5MWJhODc3MGUxMWU2YzNi NzRhZWFlMDEwZjhhYWY2MWYyM2YgMTAwNjQ0Ci0tLSBhL0xheW91dFRlc3RzL2luc3BlY3Rvci9w YWdlL2FyY2hpdmUuaHRtbAorKysgYi9MYXlvdXRUZXN0cy9pbnNwZWN0b3IvcGFnZS9hcmNoaXZl Lmh0bWwKQEAgLTYsMTQgKzYsMTcgQEAgZnVuY3Rpb24gdGVzdCgpCiB7CiAgICAgSW5zcGVjdG9y UHJvdG9jb2wuc2VuZENvbW1hbmQoIlBhZ2UuZW5hYmxlIiwge30pOwogICAgIEluc3BlY3RvclBy b3RvY29sLnNlbmRDb21tYW5kKCJQYWdlLmFyY2hpdmUiLCB7fSwgZnVuY3Rpb24oZXZlbnQpIHsK LSAgICAgICAgdmFyIGRhdGEgPSBldmVudC5yZXN1bHQuZGF0YTsKLSAgICAgICAgaWYgKCFkYXRh KQotICAgICAgICAgICAgUHJvdG9jb2xUZXN0LmxvZygiRkFJTDogbm8gZGF0YSIpOwotICAgICAg ICBlbHNlIGlmIChkYXRhLmxlbmd0aCA8IDEwMDApCi0gICAgICAgICAgICBQcm90b2NvbFRlc3Qu bG9nKCJGQUlMOiB1bmV4cGVjdGVkbHkgc2hvcnQgZGF0YS4gQSBzZXJpYWxpemVkIGFyY2hpdmUg c2hvdWxkIGJlIHByZXR0eSBsYXJnZS4iKTsKLSAgICAgICAgZWxzZQotICAgICAgICAgICAgUHJv dG9jb2xUZXN0LmxvZygiUEFTUzogUmVjZWl2ZWQgYXJjaGl2ZSBkYXRhLiIpOwotCisgICAgICAg IGlmIChldmVudC5lcnJvcikKKyAgICAgICAgICAgIFByb3RvY29sVGVzdC5leHBlY3RFcXVhbChl dmVudC5lcnJvci5tZXNzYWdlLCAiTm90IHN1cHBvcnRlZCIsICJQYWdlLmFyY2hpdmUgaXMgbm90 IHN1cHBvcnRlZCBvbiB0aGlzIHBsYXRmb3JtIHlldC4iKTsKKyAgICAgICAgZWxzZSB7CisgICAg ICAgICAgICB2YXIgZGF0YSA9IGV2ZW50LnJlc3VsdC5kYXRhOworICAgICAgICAgICAgaWYgKCFk YXRhKQorICAgICAgICAgICAgICAgIFByb3RvY29sVGVzdC5sb2coIkZBSUw6IG5vIGRhdGEiKTsK KyAgICAgICAgICAgIGVsc2UgaWYgKGRhdGEubGVuZ3RoIDwgMTAwMCkKKyAgICAgICAgICAgICAg ICBQcm90b2NvbFRlc3QubG9nKCJGQUlMOiB1bmV4cGVjdGVkbHkgc2hvcnQgZGF0YS4gQSBzZXJp YWxpemVkIGFyY2hpdmUgc2hvdWxkIGJlIHByZXR0eSBsYXJnZS4iKTsKKyAgICAgICAgICAgIGVs c2UKKyAgICAgICAgICAgICAgICBQcm90b2NvbFRlc3QubG9nKCJQQVNTOiBSZWNlaXZlZCBhcmNo aXZlIGRhdGEuIik7CisgICAgICAgIH0KICAgICAgICAgUHJvdG9jb2xUZXN0LmNvbXBsZXRlVGVz dCgpOwogICAgIH0pOwogfQpkaWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0vZ3RrL1Rl c3RFeHBlY3RhdGlvbnMgYi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9u cwppbmRleCBhZDNkZWNhOGVhMGQ4YTFhMmUxZGZhMGM3NGFiNzAwNmEwNzg4YjY0Li4wYWIwYjAx NGVhMDUxZWMyMDI3MjlhYWM5MjBiMDc0NWZiZThhMmNiIDEwMDY0NAotLS0gYS9MYXlvdXRUZXN0 cy9wbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9ucworKysgYi9MYXlvdXRUZXN0cy9wbGF0Zm9y bS9ndGsvVGVzdEV4cGVjdGF0aW9ucwpAQCAtMjMyOCw4ICsyMzI4LDYgQEAgd2Via2l0Lm9yZy9i LzEwODkyNSBodHRwL3Rlc3RzL21lZGlhL3ZpZGVvLXBsYXktc3RhbGwuaHRtbCBbIEZhaWx1cmUg VGltZW91dCBdCiAKIHdlYmtpdC5vcmcvYi8xMTY5NTcgbWVkaWEvdHJhY2svdHJhY2stYXV0b21h dGljLXN1YnRpdGxlcy5odG1sIFsgVGltZW91dCBdCiAKLXdlYmtpdC5vcmcvYi8xMjA2ODIgaW5z cGVjdG9yL3BhZ2UvYXJjaGl2ZS5odG1sIFsgVGltZW91dCBdCi0KIHdlYmtpdC5vcmcvYi8xNDc1 MTggaW5zcGVjdG9yL2RlYnVnZ2VyL25lc3RlZC1pbnNwZWN0b3JzLmh0bWwgWyBUaW1lb3V0IF0K IHdlYmtpdC5vcmcvYi8xMjI1NzEgaHR0cC90ZXN0cy9pbnNwZWN0b3IvZG9tL2Nyb3NzLWRvbWFp bi1pbnNwZWN0ZWQtbm9kZS1hY2Nlc3MuaHRtbCBbIFRpbWVvdXQgUGFzcyBdCiB3ZWJraXQub3Jn L2IvMTIyNTcxIGluc3BlY3Rvci9kb20vZm9jdXMuaHRtbCBbIEZhaWx1cmUgVGltZW91dCBQYXNz IF0KZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL3BsYXRmb3JtL2d0ay9pbnNwZWN0b3IvcGFnZS9h cmNoaXZlLWV4cGVjdGVkLnR4dCBiL0xheW91dFRlc3RzL3BsYXRmb3JtL2d0ay9pbnNwZWN0b3Iv cGFnZS9hcmNoaXZlLWV4cGVjdGVkLnR4dApuZXcgZmlsZSBtb2RlIDEwMDY0NAppbmRleCAwMDAw MDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwLi4xYTdjMTMzNTQ4YzU4NDM3NTM1 OWMwMjRkMmU4ZmZhYmYwMjg1OTY4Ci0tLSAvZGV2L251bGwKKysrIGIvTGF5b3V0VGVzdHMvcGxh dGZvcm0vZ3RrL2luc3BlY3Rvci9wYWdlL2FyY2hpdmUtZXhwZWN0ZWQudHh0CkBAIC0wLDAgKzEs MiBAQAorUEFTUzogUGFnZS5hcmNoaXZlIGlzIG5vdCBzdXBwb3J0ZWQgb24gdGhpcyBwbGF0Zm9y bSB5ZXQuCisK 383145 2019-11-08 11:11:21 -0800 2019-11-08 22:28:04 -0800 Patch bug-120682-20191108111120.patch text/plain 3352 yurys U3VidmVyc2lvbiBSZXZpc2lvbjogMjUyMjA2CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggNmU5ZjNkYjNmNWJmYzVhMjQ5YWZi ZDkyODQ1YWU0YjQxZDRkZmE1Zi4uOTcxMTFlMGRhZWYyNzM3NDBkOTc0MjljNDVmZTZhYzIxZmY1 NTFjMSAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDE3IEBACisyMDE5LTExLTA0ICBZdXJ5IFNlbWlraGF0c2t5 ICA8eXVyeXNAY2hyb21pdW0ub3JnPgorCisgICAgICAgIE5ldyB0ZXN0IGluc3BlY3Rvci1wcm90 b2NvbC9wYWdlL2FyY2hpdmUuaHRtbCBhZGRlZCBpbiByMTU0ODI4IGZhaWxzIG9uIEVGTCwgUXQs IEdUSworICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MTIw NjgyCisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgR2Vu ZXJhdGUgY3VzdG9tIGV4cGVjdGF0aW9ucyBmb3IgaW5zcGVjdG9yLXByb3RvY29sL3BhZ2UvYXJj aGl2ZS5odG1sIG9uIHRoZSBwbGF0Zm9ybXMKKyAgICAgICAgd2hlcmUgUGFnZS5hcmNoaXZlIGlz IG5vdCBzdXBwb3J0ZWQgeWV0LgorCisgICAgICAgICogaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS5o dG1sOiAnTm90IHN1cHBvcnRlZCcgaXMgYSB2YWxpZCByZXNwb25zZSBvbiBzb21lIHBsYXRmb3Jt cy4KKyAgICAgICAgKiBwbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9uczoKKyAgICAgICAgKiBw bGF0Zm9ybS9ndGsvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS1leHBlY3RlZC50eHQ6IEFkZGVkLgor CiAyMDE5LTExLTA2ICBSeW9zdWtlIE5pd2EgIDxybml3YUB3ZWJraXQub3JnPgogCiAgICAgICAg IEludGVncmF0ZSBzY3JvbGwgZXZlbnQgaW50byBIVE1MNSBldmVudCBsb29wCmRpZmYgLS1naXQg YS9MYXlvdXRUZXN0cy9pbnNwZWN0b3IvcGFnZS9hcmNoaXZlLmh0bWwgYi9MYXlvdXRUZXN0cy9p bnNwZWN0b3IvcGFnZS9hcmNoaXZlLmh0bWwKaW5kZXggODA2OTFkMGE0MmFhZmVhMTJhN2NjNTFl ZGNmMzI4NzlmYWFkNWM3MC4uNjhjMmU5N2FlYzliYzU1NmRlYmRkOTljOTNhOWZiNzcyZTlmNmU5 YSAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS5odG1sCisr KyBiL0xheW91dFRlc3RzL2luc3BlY3Rvci9wYWdlL2FyY2hpdmUuaHRtbApAQCAtNiwxNCArNiwx NyBAQCBmdW5jdGlvbiB0ZXN0KCkKIHsKICAgICBJbnNwZWN0b3JQcm90b2NvbC5zZW5kQ29tbWFu ZCgiUGFnZS5lbmFibGUiLCB7fSk7CiAgICAgSW5zcGVjdG9yUHJvdG9jb2wuc2VuZENvbW1hbmQo IlBhZ2UuYXJjaGl2ZSIsIHt9LCBmdW5jdGlvbihldmVudCkgewotICAgICAgICB2YXIgZGF0YSA9 IGV2ZW50LnJlc3VsdC5kYXRhOwotICAgICAgICBpZiAoIWRhdGEpCi0gICAgICAgICAgICBQcm90 b2NvbFRlc3QubG9nKCJGQUlMOiBubyBkYXRhIik7Ci0gICAgICAgIGVsc2UgaWYgKGRhdGEubGVu Z3RoIDwgMTAwMCkKLSAgICAgICAgICAgIFByb3RvY29sVGVzdC5sb2coIkZBSUw6IHVuZXhwZWN0 ZWRseSBzaG9ydCBkYXRhLiBBIHNlcmlhbGl6ZWQgYXJjaGl2ZSBzaG91bGQgYmUgcHJldHR5IGxh cmdlLiIpOwotICAgICAgICBlbHNlCi0gICAgICAgICAgICBQcm90b2NvbFRlc3QubG9nKCJQQVNT OiBSZWNlaXZlZCBhcmNoaXZlIGRhdGEuIik7Ci0KKyAgICAgICAgaWYgKGV2ZW50LmVycm9yKSB7 CisgICAgICAgICAgICBQcm90b2NvbFRlc3QuZmFpbChgUGFnZS5hcmNoaXZlIHJldHVybmVkIGVy cm9yOiAnJHtldmVudC5lcnJvci5tZXNzYWdlfScuYCk7CisgICAgICAgIH0gZWxzZSB7CisgICAg ICAgICAgICBsZXQgZGF0YSA9IGV2ZW50LnJlc3VsdC5kYXRhOworICAgICAgICAgICAgaWYgKCFk YXRhKQorICAgICAgICAgICAgICAgIFByb3RvY29sVGVzdC5mYWlsKCJubyBkYXRhIik7CisgICAg ICAgICAgICBlbHNlIGlmIChkYXRhLmxlbmd0aCA8IDEwMDApCisgICAgICAgICAgICAgICAgUHJv dG9jb2xUZXN0LmZhaWwoInVuZXhwZWN0ZWRseSBzaG9ydCBkYXRhLiBBIHNlcmlhbGl6ZWQgYXJj aGl2ZSBzaG91bGQgYmUgcHJldHR5IGxhcmdlLiIpOworICAgICAgICAgICAgZWxzZQorICAgICAg ICAgICAgICAgIFByb3RvY29sVGVzdC5wYXNzKCJSZWNlaXZlZCBhcmNoaXZlIGRhdGEuIik7Cisg ICAgICAgIH0KICAgICAgICAgUHJvdG9jb2xUZXN0LmNvbXBsZXRlVGVzdCgpOwogICAgIH0pOwog fQpkaWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0vZ3RrL1Rlc3RFeHBlY3RhdGlvbnMg Yi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvVGVzdEV4cGVjdGF0aW9ucwppbmRleCBkMjI4N2Ex Y2YxMmZiMDY4YzZkNWNiNmU5ZTVhODhjZjg2YWJmOWU0Li4wM2VhNzNlMDA2MzFjODQ4OWU5OTY1 MDAwNDVlMjRmMGE0NWFiMDFmIDEwMDY0NAotLS0gYS9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsv VGVzdEV4cGVjdGF0aW9ucworKysgYi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvVGVzdEV4cGVj dGF0aW9ucwpAQCAtMjMxOCw4ICsyMzE4LDYgQEAgd2Via2l0Lm9yZy9iLzEwODkyNSBodHRwL3Rl c3RzL21lZGlhL3ZpZGVvLXBsYXktc3RhbGwuaHRtbCBbIEZhaWx1cmUgVGltZW91dCBdCiAKIHdl YmtpdC5vcmcvYi8xMTY5NTcgbWVkaWEvdHJhY2svdHJhY2stYXV0b21hdGljLXN1YnRpdGxlcy5o dG1sIFsgVGltZW91dCBdCiAKLXdlYmtpdC5vcmcvYi8xMjA2ODIgaW5zcGVjdG9yL3BhZ2UvYXJj aGl2ZS5odG1sIFsgVGltZW91dCBdCi0KIHdlYmtpdC5vcmcvYi8xNDc1MTggaW5zcGVjdG9yL2Rl YnVnZ2VyL25lc3RlZC1pbnNwZWN0b3JzLmh0bWwgWyBUaW1lb3V0IF0KIAogQnVnKEdUSykgcGx1 Z2lucy9yZWxvYWRwbHVnaW5zLWFuZC1wYWdlcy5odG1sIFsgVGltZW91dCBdCmRpZmYgLS1naXQg YS9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS1leHBlY3Rl ZC50eHQgYi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvaW5zcGVjdG9yL3BhZ2UvYXJjaGl2ZS1l eHBlY3RlZC50eHQKbmV3IGZpbGUgbW9kZSAxMDA2NDQKaW5kZXggMDAwMDAwMDAwMDAwMDAwMDAw MDAwMDAwMDAwMDAwMDAwMDAwMDAwMC4uNWQ5MDFkYjBjNzk2NGU3MzIzYzg0OTI3NDI0NTE5ZDgw ZjRkOWU2MQotLS0gL2Rldi9udWxsCisrKyBiL0xheW91dFRlc3RzL3BsYXRmb3JtL2d0ay9pbnNw ZWN0b3IvcGFnZS9hcmNoaXZlLWV4cGVjdGVkLnR4dApAQCAtMCwwICsxLDIgQEAKK0ZBSUw6IFBh Z2UuYXJjaGl2ZSByZXR1cm5lZCBlcnJvcjogJ05vdCBzdXBwb3J0ZWQnLgorCg==