3579 2005-06-16 16:52:44 -0700 In-place style-switching is leaving junk behind, messing up styles 2005-12-29 22:45:43 -0800 1 1 1 Unclassified WebKit CSS 412 Mac OS X 10.4 RESOLVED FIXED http://wallace.karelia.com/ P2 Normal --- 1 dwood darin eric ttalbot oldest_to_newest 12303 0 dwood 2005-06-16 16:52:45 -0700 I am using a javascript style switcher, as described here <http://www.alistapart.com/articles/alternate/ >. I am seeing problems where switching to a particular stylesheet, and then back to a previously working stylesheet, is exhibiting strange display where a previously visible element is now moved up too far. Steps to reproduce. 1. Open index.html reduction test case (attached) in Safari. The initial stylesheet, "No Parking Anytime", is fine ... the title "Announcing Sandvox" looks fine. 2. Switch, using the popup menu, to "Gnarled." 3. Now switch back to "No Parking Anytime." EXPECTED: the page should look the same as it did before. ACTUAL: the h2 text and graphic have been moved up substantially, above the webview's bounds REGRESSION: * This works fine in FireFox * Don't need to view "No Parking Anytime" first; if you switch "stylesheet" and "alternate stylesheet" in the code to view Gnarled first, it still screws up No Parking Anytime when you switch to that. NOTES: * There is no common stylesheet that both designs are using; one should completely cancel out the other when you switch. * The entire page can be viewed in context at <http://wallace.karelia.com/>. In the full page, you see that the h2 element is moved up underneath the div to the "north" of it. * This problem was demonstrated to Maciej and SuperKevin at WWDC last week. 12306 1 dwood 2005-06-16 16:59:58 -0700 The reduced test case is available here for download: <http://forms.karelia.com/KareliaSiteReduced.zip> (It's big, but only becuase of the images with the style sheets; I didn't want to pick through them) 12319 2 c.petersen87 2005-06-16 21:11:40 -0700 Attached test case reproduces with Safari 2.0 (v412) and TOT webkit. 12471 3 dwood 2005-06-17 11:32:14 -0700 Here's an additional, non-reduced, test case, which may be related. Go to http://wallace.karelia.com/ and choose "Aqua" theme (if it's not already selected), then Iris Spring. The site menu (below the flowers) will be word-wrapped (slightly, at least -- it should be all on one line in this design.). Then choose "No Parking Anytime" and then back to Iris Spring, it should be OK again. This one doesn't reproduce as well as the main test case, unfortunately. But the Iris Spring design definitely gets messed with when switching to other designs. 24130 4 eric 2005-11-15 18:49:10 -0800 I've seen this bug in S5. Try going here: http://meyerweb.com/eric/tools/s5/s5-intro.html Once you get there and the slide show come up, hit "C" to reveal the controls (in the lower right-hand corner of the screen), then hit "T" to switch to outline view and then "T" again to go back to the slide show. The 'select' menu for navigating the slide show will be moved up from where it was, sitting on top of the symbol-links that used to be above it. Note: this is in 1.3.1 (v312.3.1), OS X 10.3.9. 25417 5 hyatt 2005-12-02 17:10:31 -0800 This is likely to be multiple bugs. Style switching without reloading the page when you do many changes is very difficult to get right. 25419 6 dwood 2005-12-02 17:21:49 -0800 Hmm, the site where I had the non-reduced test case is not there any more. Let me know if you want me to put it back up. Otherwise, the reduced test case is attached. 25423 7 Graham.Dennis 2005-12-02 21:48:36 -0800 (In reply to comment #6) Can you put the reduced test case back up? 'gnarled' is no longer an option at http://wallace.karelia.com 25562 8 Graham.Dennis 2005-12-05 04:42:53 -0800 (In reply to comment #3) The problem with the word-wrapped text in Iris Spring seems to be related to the order in which the style sheets are disabled and enabled. If the old style sheet is disabled before the new one is enabled, then you get the strange behaviour in 'Iris Spring', however if the new one is enabled before the old one is disabled, then it works. This is demonstrated by putting 'No Parking Anytime' above 'Iris Spring' in the list, and 'Aqua' below it. A workaround is to rewrite the setActiveStyleSheet function as: function setActiveStyleSheet(title) { var i, a, main; for(i=0; (a = document.getElementsByTagName("link")[i]); i++) { if(a.getAttribute("rel").indexOf("style") != -1 && a.getAttribute("title")) { if(a.getAttribute("title") == title) a.disabled = false; } } for(i=0; (a = document.getElementsByTagName("link")[i]); i++) { if(a.getAttribute("rel").indexOf("style") != -1 && a.getAttribute("title")) { if(a.getAttribute("title") != title) a.disabled = true; } } } I am doubtful as to whether this will fix the original bug with the h2 text and graphic. 25590 9 dwood 2005-12-05 12:58:53 -0800 (In reply to comment #7) I've put up "gnarled" back to wallace.karelia.com so you can now test out the original, non-reduced test- case. 25591 10 dwood 2005-12-05 13:04:58 -0800 (In reply to comment #8) I put an additional popup on wallace.karelia.com to test out Graham's suggested workaround, but it does not (as he suspects) fix the problem. 25598 11 4977 Graham.Dennis 2005-12-05 15:05:18 -0800 Created attachment 4977 patch The problem was that when the top or bottom margin was changed from positive to negative (Gnarled has a negative top margin for the h2 element, but Aqua has 0), only one of m_maxTopPosMargin and m_maxTopNegMargin were appropriately set. This patch makes sure both are set. 25599 12 4977 hyatt 2005-12-05 15:20:54 -0800 Comment on attachment 4977 patch Great find. That's the right fix indeed. Please write a little test case that dynamically changes the margin from positive to negative so we have a layout test that covers this. 25600 13 dwood 2005-12-05 16:13:25 -0800 (In reply to comment #11) Graham, does that also fix the test case in comment #3? 25601 14 4980 Graham.Dennis 2005-12-05 16:33:34 -0800 Created attachment 4980 testcase This page has two buttons, one to set negative margin-top and margin-bottom values for the second div, one to set positive values. To aid automation, these functions are called from onLoad. If they weren't called automatically, then the page originally displays correctly (with positive margins on the second div). Clicking 'Negative margin' will change margin-top and margin-bottom to negative values. Currently, this has the same effect as margin-top and margin-bottom = 0px. Clicking 'Positive margin' then does nothing. With the patch, 'Negative margin' then behaves correctly, and clicking 'Positive margin' then returns to the original layout. 25602 15 Graham.Dennis 2005-12-05 16:38:49 -0800 (In reply to comment #13) No, that appears to be a completely different bug caused by the ordering of enabling and disabling the style sheets. I have investigated it, but I'm stumped at present. It will need it's own bug report. 25603 16 Graham.Dennis 2005-12-05 16:42:14 -0800 (In reply to comment #14) A better check might be to comment out the setTimeout line that calls setPositiveMargin so that only changing positive -> negative margins is tested and not positive->negative->positive. 25637 17 4977 hyatt 2005-12-06 12:01:18 -0800 Comment on attachment 4977 patch r=me 25773 18 darin 2005-12-10 11:11:15 -0800 I'd like to land the fix, but I can't get the test case to do anything in Safari or even Firefox, so I'm not comfortable checking anything in. 25775 19 dwood 2005-12-10 11:32:39 -0800 Darin, I'm a bit confused ... were you unable to get the style switcher to work on <http:// wallace.karelia.com/> on either firefox or Safari? (Perhaps you have javascript disabled?) Or were you unable to *reproduce* the bug on Safari or firefox? I just ran the latest nightly build and viewed <http://wallace.karelia.com/> ... first with "Gnarled" then with "No Parking Anytime" and I confirmed that the problem is still there on TOT. Doing the same on Firefox 1.5, I did not have the problem. Let me know some more specifics so I can help you bring 'er in for a landing. :-) 25779 20 darin 2005-12-10 12:30:19 -0800 I was testing the attached test case. I need a test case to land with the bug as a layout test. 25830 21 Graham.Dennis 2005-12-11 14:03:38 -0800 (In reply to comment #20) The test case doesn't look like it does anything in current ToT because the bug is present. Also the rendering is incorrect. If patch is fixed, the rendering is different, and the buttons actually work. Because of the bug setting the margin to anything between +20px and -20px changes nothing. I'll attach a second test case (commenting out the line mentioned in comment #16) that tests what happens when the margin is changed from positive -> negative, instead of positive -> negative -> positive. In this case, the ''positive margin" button will do something the first time, but no more in current ToT. 25831 22 5044 Graham.Dennis 2005-12-11 14:05:09 -0800 Created attachment 5044 testcase 2 Second testcase with setTimeout line commented. In this patch only changing the margin from positive -> negative is tested automatically. 25833 23 Graham.Dennis 2005-12-11 14:12:09 -0800 (In reply to comment #14) I've just realised my explanation of the original testcase was slightly inadequate. My description of the testcase was describing what happens if the onLoad="test();" is removed. With it, what is described in comment #14 happens automatically, and the buttons appear to do nothing. To see what I described in comment #14, one would need to remove the onLoad. I added 'onLoad' to make automatic testing easier. 26075 24 darin 2005-12-17 08:58:56 -0800 Here are things that still confuse me: 1) When I open test case 2 <http://bugzilla.opendarwin.org/attachment.cgi?id=5044> in Safari, without applying the patch, I see what looks like a negative margin. That's good because the code sets a negative margin in its load handler. So I don't see how that illustrates the bug in a way suitable for use as an automated test. 2) When I open test case 1 <http://bugzilla.opendarwin.org/attachment.cgi?id=4980> in Safari, without applying the patch, I see what looks like a positive margin. That's good because the code sets a negative margin and then a positive margin in its load handler and the timer. So I don't see how that illustrates the bug in a way suitable for use as an automated test. 3) I assume that this works properly in Firefox, so I opened both test cases in Firefox 1.5 for comparison with Safari. In both cases, the blocks show up with what look like positive margins, the layout is different than in Safari (more space between elements) and the buttons don't do anything. Do the test cases take advantage of something that is not implemented in Firefox? Does it illustrate other bugs in Safari? 4) The test cases don't have any self-documentation; when I open them they don't say what they should look like or how they should behave. This is important for the understanding of people working on the project in the future. Good test cases should describe, within the test page, what to expect. I do see with both test cases that if I click the positive margin and then the negative margin button in Safari that it doesn't go back to negative margins. So that does look like a way to reproduce the bug. But that's a manual test. Can someone create an automated test for this bug? 26217 25 5159 Graham.Dennis 2005-12-19 16:04:02 -0800 Created attachment 5159 testcase 3 Darin, thanks for the feedback. This is my first testcase, so I'm still learning :-) 1) Point taken. It was actually rendering correctly and not displaying the bug. 2) It's not actually displaying with the 20px top and bottom margin that the code sets. Instead it displays as though the top and bottom margin are 0px. 3) The buttons didn't work in Firefox (I should have tested this) because I was accessing element.style["margin-top"] and element.style["margin-bottom"], while these do exist for Safari, only element.style["margin"] (or element.style.margin) exists in Firefox. So my testcase was taking advantage of something in Safari. 4) I've tried to rectify this with this testcase. I include the two examples of what the negative and positive margin cases should look like, as well as testing both changing the margin from positive -> negative (which works), and changing the margin from positive -> negative -> positive (which doesn't work). I have also implemented the testcase in such a way that the buttons work in Firefox. This bug is not present in Firefox, and so the behaviour after the patch is the same as Firefox's behaviour. 27188 26 5159 darin 2005-12-29 17:58:25 -0800 Comment on attachment 5159 testcase 3 OK, the dynamic cases look fine and seem to demonstrate the problem perfectly. The automatic testing does not seem to work. The automatic cases don't seem to fail or do anything wrong in TOT Safari or in the latest released Safari either. So at this point, this is fine for a manual test, but not for the layout tests directory. 27189 27 5159 darin 2005-12-29 18:02:50 -0800 Comment on attachment 5159 testcase 3 I take back what I said. Test case to be working fine. 27190 28 5159 darin 2005-12-29 18:14:09 -0800 Comment on attachment 5159 testcase 3 I'll land this now (a modified version to fit better in 800x600). 27191 29 darin 2005-12-29 18:14:31 -0800 I'll land this patch later tonight or tomorrow. 4977 2005-12-05 15:05:18 -0800 2005-12-06 12:01:18 -0800 patch patch.diff text/plain 1108 Graham.Dennis SW5kZXg6IGtodG1sL3JlbmRlcmluZy9yZW5kZXJfYmxvY2suaAo9PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09ClJDUyBmaWxl OiAvY3ZzL3Jvb3QvV2ViQ29yZS9raHRtbC9yZW5kZXJpbmcvcmVuZGVyX2Jsb2NrLmgsdgpyZXRy aWV2aW5nIHJldmlzaW9uIDEuNzMKZGlmZiAtdSAtcjEuNzMgcmVuZGVyX2Jsb2NrLmgKLS0tIGto dG1sL3JlbmRlcmluZy9yZW5kZXJfYmxvY2suaCAgICAgIDI3IE5vdiAyMDA1IDIyOjUyOjA4IC0w MDAwICAgICAgMS43MworKysga2h0bWwvcmVuZGVyaW5nL3JlbmRlcl9ibG9jay5oICAgICAgNSBE ZWMgMjAwNSAyMzowNzozNCAtMDAwMApAQCAtOTAsMTQgKzkwLDIyIEBACiAgICAgfQogCiAgICAg dm9pZCBpbml0TWF4TWFyZ2luVmFsdWVzKCkgewotICAgICAgICBpZiAobV9tYXJnaW5Ub3AgPj0g MCkKKyAgICAgICAgaWYgKG1fbWFyZ2luVG9wID49IDApIHsKICAgICAgICAgICAgIG1fbWF4VG9w UG9zTWFyZ2luID0gbV9tYXJnaW5Ub3A7Ci0gICAgICAgIGVsc2UKKyAgICAgICAgICAgIG1fbWF4 VG9wTmVnTWFyZ2luID0gMDsKKyAgICAgICAgfQorICAgICAgICBlbHNlIHsKICAgICAgICAgICAg IG1fbWF4VG9wTmVnTWFyZ2luID0gLW1fbWFyZ2luVG9wOwotICAgICAgICBpZiAobV9tYXJnaW5C b3R0b20gPj0gMCkKKyAgICAgICAgICAgIG1fbWF4VG9wUG9zTWFyZ2luID0gMDsKKyAgICAgICAg fQorICAgICAgICBpZiAobV9tYXJnaW5Cb3R0b20gPj0gMCkgewogICAgICAgICAgICAgbV9tYXhC b3R0b21Qb3NNYXJnaW4gPSBtX21hcmdpbkJvdHRvbTsKLSAgICAgICAgZWxzZQorICAgICAgICAg ICAgbV9tYXhCb3R0b21OZWdNYXJnaW4gPSAwOworICAgICAgICB9CisgICAgICAgIGVsc2Ugewog ICAgICAgICAgICAgbV9tYXhCb3R0b21OZWdNYXJnaW4gPSAtbV9tYXJnaW5Cb3R0b207CisgICAg ICAgICAgICBtX21heEJvdHRvbVBvc01hcmdpbiA9IDA7CisgICAgICAgIH0KICAgICB9CiAKICAg ICB2aXJ0dWFsIHZvaWQgYWRkQ2hpbGRUb0Zsb3coUmVuZGVyT2JqZWN0KiBuZXdDaGlsZCwgUmVu ZGVyT2JqZWN0KiBiZWZvcmVDaGlsZCk7Cg== 4980 2005-12-05 16:33:34 -0800 2005-12-19 16:04:02 -0800 testcase margin_top_bottom_dynamic.html text/html 889 Graham.Dennis PGh0bWw+PGhlYWQ+CjxzY3JpcHQgdHlwZT0idGV4dC9qYXZhc2NyaXB0Ij4KPCEtLQpmdW5jdGlv biBzZXROZWdhdGl2ZU1hcmdpbigpCnsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlkKCJmb28iKS5z dHlsZVsibWFyZ2luLXRvcCJdID0gIi0yMHB4IjsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlkKCJm b28iKS5zdHlsZVsibWFyZ2luLWJvdHRvbSJdID0gIi0yMHB4IjsKfQoKZnVuY3Rpb24gc2V0UG9z aXRpdmVNYXJnaW4oKQp7Cglkb2N1bWVudC5nZXRFbGVtZW50QnlJZCgiZm9vIikuc3R5bGVbIm1h cmdpbi10b3AiXSA9ICIyMHB4IjsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlkKCJmb28iKS5zdHls ZVsibWFyZ2luLWJvdHRvbSJdID0gIjIwcHgiOwp9CgpmdW5jdGlvbiB0ZXN0KCkKewoJc2V0TmVn YXRpdmVNYXJnaW4oKTsKCXNldFRpbWVvdXQoInNldFBvc2l0aXZlTWFyZ2luKCk7IiwgMSk7Cn0K Ly8tLT4KPC9zY3JpcHQ+PC9oZWFkPgoJPGJvZHkgb25Mb2FkPSJ0ZXN0KCk7Ij4KCQk8ZGl2IHN0 eWxlPSJtYXJnaW4tdG9wOiAyMHB4OyBib3JkZXI6MXB4IGRvdHRlZCBibHVlOyI+CgkJCTxwPkxv cmVtIGlwc3VtPC9wPgoJCTwvZGl2PgoJCTxkaXYgc3R5bGU9Im1hcmdpbi10b3A6IDIwcHg7IG1h cmdpbi1ib3R0b206IDIwcHg7IGJvcmRlcjoxcHggZG90dGVkIGJsdWU7IiBpZD0iZm9vIj4KCQkJ PHA+TG9yZW0gaXBzdW08L3A+CgkJPC9kaXY+CgkJPGlucHV0IHR5cGU9InN1Ym1pdCIgdmFsdWU9 Ik5lZ2F0aXZlIG1hcmdpbiIgb25DbGljaz0ic2V0TmVnYXRpdmVNYXJnaW4oKTsiLz4KCQk8aW5w dXQgdHlwZT0ic3VibWl0IiB2YWx1ZT0iUG9zaXRpdmUgbWFyZ2luIiBvbkNsaWNrPSJzZXRQb3Np dGl2ZU1hcmdpbigpOyIvPgoJPC9ib2R5Pgo8L2h0bWw+Cg== 5044 2005-12-11 14:05:09 -0800 2005-12-19 16:04:02 -0800 testcase 2 margin_top_bottom_dynamic.html text/html 891 Graham.Dennis PGh0bWw+PGhlYWQ+CjxzY3JpcHQgdHlwZT0idGV4dC9qYXZhc2NyaXB0Ij4KPCEtLQpmdW5jdGlv biBzZXROZWdhdGl2ZU1hcmdpbigpCnsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlkKCJmb28iKS5z dHlsZVsibWFyZ2luLXRvcCJdID0gIi0yMHB4IjsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlkKCJm b28iKS5zdHlsZVsibWFyZ2luLWJvdHRvbSJdID0gIi0yMHB4IjsKfQoKZnVuY3Rpb24gc2V0UG9z aXRpdmVNYXJnaW4oKQp7Cglkb2N1bWVudC5nZXRFbGVtZW50QnlJZCgiZm9vIikuc3R5bGVbIm1h cmdpbi10b3AiXSA9ICIyMHB4IjsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlkKCJmb28iKS5zdHls ZVsibWFyZ2luLWJvdHRvbSJdID0gIjIwcHgiOwp9CgpmdW5jdGlvbiB0ZXN0KCkKewoJc2V0TmVn YXRpdmVNYXJnaW4oKTsKLy8Jc2V0VGltZW91dCgic2V0UG9zaXRpdmVNYXJnaW4oKTsiLCAxKTsK fQovLy0tPgo8L3NjcmlwdD48L2hlYWQ+Cgk8Ym9keSBvbkxvYWQ9InRlc3QoKTsiPgoJCTxkaXYg c3R5bGU9Im1hcmdpbi10b3A6IDIwcHg7IGJvcmRlcjoxcHggZG90dGVkIGJsdWU7Ij4KCQkJPHA+ TG9yZW0gaXBzdW08L3A+CgkJPC9kaXY+CgkJPGRpdiBzdHlsZT0ibWFyZ2luLXRvcDogMjBweDsg bWFyZ2luLWJvdHRvbTogMjBweDsgYm9yZGVyOjFweCBkb3R0ZWQgYmx1ZTsiIGlkPSJmb28iPgoJ CQk8cD5Mb3JlbSBpcHN1bTwvcD4KCQk8L2Rpdj4KCQk8aW5wdXQgdHlwZT0ic3VibWl0IiB2YWx1 ZT0iTmVnYXRpdmUgbWFyZ2luIiBvbkNsaWNrPSJzZXROZWdhdGl2ZU1hcmdpbigpOyIvPgoJCTxp bnB1dCB0eXBlPSJzdWJtaXQiIHZhbHVlPSJQb3NpdGl2ZSBtYXJnaW4iIG9uQ2xpY2s9InNldFBv c2l0aXZlTWFyZ2luKCk7Ii8+Cgk8L2JvZHk+CjwvaHRtbD4K 5159 2005-12-19 16:04:02 -0800 2005-12-29 18:14:09 -0800 testcase 3 margin_top_bottom_dynamic.html text/html 2157 Graham.Dennis PGh0bWw+PGhlYWQ+CjxzY3JpcHQgdHlwZT0idGV4dC9qYXZhc2NyaXB0Ij4KPCEtLQpmdW5jdGlv biBzZXROZWdhdGl2ZU1hcmdpbihlbGVtZW50KQp7Cglkb2N1bWVudC5nZXRFbGVtZW50QnlJZChl bGVtZW50KS5zdHlsZVsibWFyZ2luIl0gPSAiLTIwcHggMHB4IC0yMHB4IDBweCI7Cn0KCmZ1bmN0 aW9uIHNldFBvc2l0aXZlTWFyZ2luKGVsZW1lbnQpCnsKCWRvY3VtZW50LmdldEVsZW1lbnRCeUlk KGVsZW1lbnQpLnN0eWxlWyJtYXJnaW4iXSA9ICIyMHB4IDBweCAyMHB4IDBweCI7Cn0KCmZ1bmN0 aW9uIHRlc3QoKQp7CglzZXROZWdhdGl2ZU1hcmdpbigiZm9vIik7CgoJc2V0TmVnYXRpdmVNYXJn aW4oImJhciIpOwoJc2V0VGltZW91dCgic2V0UG9zaXRpdmVNYXJnaW4oJ2JhcicpOyIsIDEpOwp9 Ci8vLS0+Cjwvc2NyaXB0PjwvaGVhZD4KCTxib2R5IG9uTG9hZD0idGVzdCgpOyI+CgkJPHA+CgkJ V2hhdCBpdCBzaG91bGQgbG9vayBsaWtlIChwb3NpdGl2ZSBjYXNlKToKCQk8L3A+CgkJPGRpdiBz dHlsZT0iYm9yZGVyOiAxMHB4IHNvbGlkIGdyZWVuOyAiPgoJCQk8ZGl2IHN0eWxlPSJtYXJnaW4t dG9wOiAyMHB4OyBib3JkZXI6MXB4IHNvbGlkIGJsdWU7Ij4KCQkJCTxwPkxvcmVtIGlwc3VtPC9w PgoJCQk8L2Rpdj4KCQkJPGRpdiBzdHlsZT0ibWFyZ2luLXRvcDogMjBweDsgbWFyZ2luLWJvdHRv bTogMjBweDsgYm9yZGVyOjFweCBkb3R0ZWQgYmx1ZTsiPgoJCQkJPHA+TG9yZW0gaXBzdW08L3A+ CgkJCTwvZGl2PgoJCTwvZGl2PgoKCQk8cD4KCQlXaGF0IGl0IHNob3VsZCBsb29rIGxpa2UgKG5l Z2F0aXZlIGNhc2UpOgoJCTwvcD4KCQk8ZGl2IHN0eWxlPSJib3JkZXI6IDEwcHggc29saWQgZ3Jl ZW47ICI+CgkJCTxkaXYgc3R5bGU9Im1hcmdpbi10b3A6IDIwcHg7IGJvcmRlcjoxcHggc29saWQg Ymx1ZTsiPgoJCQkJPHA+TG9yZW0gaXBzdW08L3A+CgkJCTwvZGl2PgoJCQk8ZGl2IHN0eWxlPSJt YXJnaW4tdG9wOiAtMjBweDsgbWFyZ2luLWJvdHRvbTogLTIwcHg7IGJvcmRlcjoxcHggZG90dGVk IGJsdWU7Ij4KCQkJCTxwPkxvcmVtIGlwc3VtPC9wPgoJCQk8L2Rpdj4KCQk8L2Rpdj4KCQk8cD4K CQlEeW5hbWljIGNhc2UgKGF1dG9tYXRpY2FsbHkgdGVzdGluZyBwb3NpdGl2ZSAtLSZndDsgbmVn YXRpdmUpOgoJCTwvcD4KCgkJPGRpdiBzdHlsZT0iYm9yZGVyOiAxMHB4IHNvbGlkIGdyZWVuOyAi PgoJCQk8ZGl2IHN0eWxlPSJtYXJnaW4tdG9wOiAyMHB4OyBib3JkZXI6MXB4IHNvbGlkIGJsdWU7 Ij4KCQkJCTxwPkxvcmVtIGlwc3VtPC9wPgoJCQk8L2Rpdj4KCQkJPGRpdiBzdHlsZT0ibWFyZ2lu LXRvcDogMjBweDsgbWFyZ2luLWJvdHRvbTogMjBweDsgYm9yZGVyOjFweCBkb3R0ZWQgYmx1ZTsi IGlkPSJmb28iPgoJCQkJPHA+TG9yZW0gaXBzdW08L3A+CgkJCTwvZGl2PgoJCTwvZGl2PgoJCTxi ciAvPgoJCTxpbnB1dCB0eXBlPSJzdWJtaXQiIHZhbHVlPSJOZWdhdGl2ZSBtYXJnaW4iIG9uQ2xp Y2s9InNldE5lZ2F0aXZlTWFyZ2luKCdmb28nKTsiLz4KCQk8aW5wdXQgdHlwZT0ic3VibWl0IiB2 YWx1ZT0iUG9zaXRpdmUgbWFyZ2luIiBvbkNsaWNrPSJzZXRQb3NpdGl2ZU1hcmdpbignZm9vJyk7 Ii8+CgkJPHA+CgkJRHluYW1pYyBjYXNlIChhdXRvbWF0aWNhbGx5IHRlc3RpbmcgcG9zaXRpdmUg LS0mZ3Q7IG5lZ2F0aXZlIC0tJmd0OyBwb3NpdGl2ZSk6CgkJPC9wPgoKCQk8ZGl2IHN0eWxlPSJi b3JkZXI6IDEwcHggc29saWQgZ3JlZW47ICI+CgkJCTxkaXYgc3R5bGU9Im1hcmdpbi10b3A6IDIw cHg7IGJvcmRlcjoxcHggc29saWQgYmx1ZTsiPgoJCQkJPHA+TG9yZW0gaXBzdW08L3A+CgkJCTwv ZGl2PgoJCQk8ZGl2IHN0eWxlPSJtYXJnaW4tdG9wOiAyMHB4OyBtYXJnaW4tYm90dG9tOiAyMHB4 OyBib3JkZXI6MXB4IGRvdHRlZCBibHVlOyIgaWQ9ImJhciI+CgkJCQk8cD5Mb3JlbSBpcHN1bTwv cD4KCQkJPC9kaXY+CgkJPC9kaXY+CgkJPGJyIC8+CgkJPGlucHV0IHR5cGU9InN1Ym1pdCIgdmFs dWU9Ik5lZ2F0aXZlIG1hcmdpbiIgb25DbGljaz0ic2V0TmVnYXRpdmVNYXJnaW4oJ2JhcicpOyIv PgoJCTxpbnB1dCB0eXBlPSJzdWJtaXQiIHZhbHVlPSJQb3NpdGl2ZSBtYXJnaW4iIG9uQ2xpY2s9 InNldFBvc2l0aXZlTWFyZ2luKCdiYXInKTsiLz4KCgk8L2JvZHk+CjwvaHRtbD4K