45679 2010-09-13 09:51:31 -0700 r66156 broke AtlasCT library, formerly affected http://map.d.co.il/ 2011-01-04 21:16:12 -0800 1 1 1 Unclassified WebKit Plug-ins 528+ (Nightly build) All All RESOLVED FIXED InRadar, Regression P1 Normal --- 44567 51042 1 antonm aestes abarth adamarticulate+wk aestes ap cpu darin ddkilzer dmozealous eric ian mjs nirz playmobil progame+wk webkit.review.bot oldest_to_newest 278260 0 antonm 2010-09-13 09:51:31 -0700 http://trac.webkit.org/changeset/66156/ broke http://map.d.co.il/ I don't know if HTML of the site is valid or not---I hope you're in a better position to judge. The repro: build at revision 66155 and open the site: a map should appear. Build the next revision, and map doesn't show up. That affects both Safari and Chromium. 278315 1 playmobil 2010-09-13 11:04:33 -0700 map.d.co.il is one of the most popular maps sites in Israel so this is a pretty visible regression. 280146 2 mjs 2010-09-16 09:37:31 -0700 This site has a dynamically generated nested object/embed which looks like this (copied from Web Inspector): <object classid="clsid:D27CDB6E-AE6D-11cf-96B8-444553540000" codebase="http://download.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=6,0,0,0" width="100%25" height="100%25" id="AtlasMap1" align=""> <param name="movie" value="http://maps.d.co.il/sdk_v3_2_1/Kernel/RemoteFlash8New.swf?3_2_0"> <param name="zmenu" value="false"> <param name="FlashVars" value="Server=&amp;Dll=&amp;customCtrl=&amp;MapWidth=100%25&amp;MapHeight=100%25&amp;FlashId=11&amp;UD=c11a6c8821cdb246&amp;SessionId=AN-ATLF76112002&amp;MapVersion=3_2_0&amp;GuiLang=Heb&amp;MapUnits=metric&amp;"> <param name="bgcolor" value="#EAEADB"> <param name="quality" value="high"> <param name="allowScriptAccess" value="always"> <param name="wmode" value="transparent"> <embed src="http://maps.d.co.il/sdk_v3_2_1/Kernel/RemoteFlash8New.swf?3_2_0" quality="high" wmode="transparent" allowscriptaccess="always" flashvars="Server=&amp;Dll=&amp;customCtrl=&amp;MapWidth=100%25&amp;MapHeight=100%25&amp;FlashId=11&amp;UD=c11a6c8821cdb246&amp;SessionId=AN-ATLF76112002&amp;MapVersion=3_2_0&amp;GuiLang=Heb&amp;MapUnits=metric&amp;" bgcolor="#EAEADB" menu="true" width="100%25" height="100%25" name="AtlasMap1" align="" type="application/x-shockwave-flash" pluginspage="http://www.macromedia.com/go/getflashplayer"></object> 285214 3 aestes 2010-09-26 17:44:47 -0700 Here is what is going on: - r66156 causes WebKit to load this site's plug-in via the object element rather than the nested embed element (see the snippet posted by Maciej above). - The plug-in loads successfully, but then tries to call out to JavaScript to do some extra initialization work. The function it calls is <name>_DoFSCommand(), where <name> is the name attribute of the element that loaded the plug-in. - The object element has no name attribute (name only exists on the embed element), so the plug-in tries to call a non-existant function. This leaves the plug-in in a partially initialized state. For instance, the map works (you can right click to zoom in, zoom out and re-center), but you can't drag the map or use the scroll wheel to zoom. - Adding a name attribute to the object or removing the object all together fixes this issue. - The site assumes the only browser that will use the object element is IE, and follows a special initialization path in that case. It uses <script event='...' for='...'> for initialization, which only works in IE. I think what we're doing is correct according to HTML5, but we should consider implementing a workaround given how popular this site is in Israel (Alexa ranks it #53). There are several things we could do to fix this: 1) Add a site-specific hack that forces fallback to the embed element and perhaps reach out to the website to notify them of this issue. 2) Remove the mapping from "clsid:D27CDB6E-AE6D-11cf-96B8-444553540000" to "application/x-shockwave-flash". I believe the reason this site works in Firefox is that it doesn't understand the classid attribute (either in general or just for flash in particular) on the object so it falls back to the embed (removing the classid attribute and adding a type attribute causes the same issue to reproduce in Firefox). Doing this would probably result in new compatibility regressions, so it's probably a bad idea. 285215 4 abarth 2010-09-26 17:55:51 -0700 I don't understand the motivation for http://trac.webkit.org/changeset/66156. It seems like it has caused a bunch of regressions. (I might be misunderstanding since I'm not following these issues that closely.) Maybe there's more of an explanation in rdar? 285218 5 aestes 2010-09-26 18:09:12 -0700 (In reply to comment #4) > I don't understand the motivation for http://trac.webkit.org/changeset/66156. It seems like it has caused a bunch of regressions. (I might be misunderstanding since I'm not following these issues that closely.) Maybe there's more of an explanation in rdar? It's true that this has caused regressions, although I don't think it has caused "a bunch". The only two issues I'm aware of right now is this one and a PLT regression (https://bugs.webkit.org/show_bug.cgi?id=45524). I'm aware of https://bugs.webkit.org/show_bug.cgi?id=44933, but this issue will be fixed when chromium merges r66992. Are there other regressions you're aware of? The intent of this change was to bring WebKit in line with how HTML5 specifies handling the object element. There really isn't any information in Radar that isn't also in the ChangeLog entries for r66156 and r66992, as well as in the bugzilla. 285221 6 abarth 2010-09-26 18:14:42 -0700 Oh, I don't mean to be negative. Just seeking to understand. When making changes to comply with HTML5, we should also keep in mind that another option is to change the spec. 285224 7 aestes 2010-09-26 18:17:52 -0700 (In reply to comment #6) > Oh, I don't mean to be negative. Just seeking to understand. When making changes to comply with HTML5, we should also keep in mind that another option is to change the spec. I didn't think you were being negative :) My quotations were not meant to be hostile. I agree that changing the spec is also an option, although the changes I've made so far seem to make sense and also make us more closely match the behavior of other browsers, at least by my testing. I've been meaning to post something to webkit-dev about this change since it is probably significant enough to warrant some discussion. This might also shake out any additional regressions. 285300 8 ian 2010-09-26 21:54:23 -0700 Please let me know how you fix this so I can be sure to update the spec accordingly. Do we have data on whether supporting classid="" is actually more helpful than harmful? Do any other non-IE browsers support it like Webkit? 285392 9 progame+wk 2010-09-27 03:58:53 -0700 i will gladly point d.co.il site admin here once it's decided as the best plan of action. 285393 10 playmobil 2010-09-27 04:00:40 -0700 Thanks Yair, FYI: I've already contacted the site admins, pointed them at this bug, and asked them to add a name attribute to the object tag. 285405 11 progame+wk 2010-09-27 05:02:18 -0700 I discovered this issue doesn't affect d.co.il only, it affects all of AtlasCT customers http://www.atlasct.co.il/Our_Customers.html which includes maps.walla.co.il (#6 in Israel according to Alexa), ynet (#5 in Israel) and more... 285616 12 aestes 2010-09-27 11:27:59 -0700 (In reply to comment #8) > Please let me know how you fix this so I can be sure to update the spec accordingly. Will do. > > Do we have data on whether supporting classid="" is actually more helpful than harmful? Do any other non-IE browsers support it like Webkit? It looks like Gecko handles the classid attribute only if there is an ActiveX plug-in installed. WebKit understands a small set of classids that it maps to MIME types (http://trac.webkit.org/browser/trunk/WebCore/html/HTMLObjectElement.cpp?rev=67179#L121). I don't have good data on whether this helps or hurts. Arguably it hurts in this case, but I presume these mappings were added to solve previous compatibility problems. 285631 13 darin 2010-09-27 11:37:43 -0700 (In reply to comment #12) > (In reply to comment #8) > > Please let me know how you fix this so I can be sure to update the spec accordingly. > > Will do. > > > Do we have data on whether supporting classid="" is actually more helpful than harmful? Do any other non-IE browsers support it like Webkit? > > It looks like Gecko handles the classid attribute only if there is an ActiveX plug-in installed. WebKit understands a small set of classids that it maps to MIME types (http://trac.webkit.org/browser/trunk/WebCore/html/HTMLObjectElement.cpp?rev=67179#L121). > > I don't have good data on whether this helps or hurts. Arguably it hurts in this case, but I presume these mappings were added to solve previous compatibility problems. At least one of these mappings was already in KHTML, not added in the WebKit era. There’s a good chance these mappings don’t help much any more. It’s not going to be easy to get data on how changing these will affect things, but perhaps we can just remove these entirely. 285842 14 aestes 2010-09-27 16:01:20 -0700 (In reply to comment #13) > (In reply to comment #12) > > (In reply to comment #8) > > > Please let me know how you fix this so I can be sure to update the spec accordingly. > > > > Will do. > > > > > Do we have data on whether supporting classid="" is actually more helpful than harmful? Do any other non-IE browsers support it like Webkit? > > > > It looks like Gecko handles the classid attribute only if there is an ActiveX plug-in installed. WebKit understands a small set of classids that it maps to MIME types (http://trac.webkit.org/browser/trunk/WebCore/html/HTMLObjectElement.cpp?rev=67179#L121). > > > > I don't have good data on whether this helps or hurts. Arguably it hurts in this case, but I presume these mappings were added to solve previous compatibility problems. > > At least one of these mappings was already in KHTML, not added in the WebKit era. There’s a good chance these mappings don’t help much any more. Yea, it looks like mappings for 'application/x-shockwave-flash' and 'audio/x-pn-realaudio-plugin' came from KHTML. We added mappings for 'video/quicktime' and 'application/x-director' in http://trac.webkit.org/changeset/3123 and a mapping for 'application/x-mplayer2' in http://trac.webkit.org/changeset/7616. The bugs for these changesets mention several sites that either no longer exist or would now work even without a classid->mime mapping. An argument for removing these mappings might be that an author who writes an object element with a classid is intending the browser to load an ActiveX component since classid is a COM/OLE concept, and so the expectation is that browsers that don't support ActiveX will render the object's fallback content. This seems to be the assumption made by AtlasCT. Also, the lack of these mappings in Gecko might mean that the only regressions would be in content that was specifically targeting WebKit in this manner. That's just a guess. 285862 15 aestes 2010-09-27 16:25:17 -0700 Opera changed their object handling to be in line with HTML5 in 10.50. Like Gecko, they also render fallback content when an object has a non-empty classid attribute. They blogged about this at http://my.opera.com/sitepatching/blog/2010/02/25/new-java-support-and-object-parsing. The site mentions that this change caused a regression on some Norwegian banking sites and they were working with the banks to get it resolved. 288842 16 progame+wk 2010-10-04 01:53:32 -0700 Looks like map.d.co.il updated their code and it is now working fine in Chrome. Things are still broken in other AtlasCT customer such as http://maps.walla.co.il/ http://maps.yad2.co.il/ http://www.atlasct.com/israel/citymouse/ http://ymap.winwin.co.il/Navigate.aspx http://www.atlasct.com/israel/calcalist/ http://www.homeless.co.il/map/ and probably more 290866 17 nirz 2010-10-07 01:40:32 -0700 Hi All, We've updated our code for all websites and it is now working fine in Chrome. BR, Nir, AtlasCT. 291371 18 darin 2010-10-07 18:14:58 -0700 So maybe we don’t have fix anything. 300475 19 playmobil 2010-10-27 12:49:27 -0700 So since the various sites are fixed it seems that the only remaining open question is whether we want to match other browsers more closely by doing away with/modifying classid mappings. 300509 20 cpu 2010-10-27 13:28:40 -0700 Actually it seems a lot of stuff still broken. See the matching bug in chrome It seems many sites use the flash fscommand technique to communicate with the page.Let me copy here one of the last comments in the chrome bug: "The reason is because content producers have simply done as they were told when writing their code. Adobe has documented that in order for fscommand to work properly, you need only to set the value of the NAME attribute of the EMBED tag or the ID property of the OBJECT tag. Setting the value of the name property has never been a requirement of fscommand. The following documentation dates back to 2000: http://www.adobe.com/support/flash/action_scripts/actionscript_dictionary/actionscript_dictionary372.html Even today, ten years later, Adobe ships their fscommand template in the latest version of Flash CS5. Attached is the output that Flash CS5 produces and the screenshot in the tool for how it’s selected. It conforms to their published doc. As expected, this template works in other browsers.." http://code.google.com/p/chromium/issues/detail?id=57986 I advice we remove the hardcoded mappings as Andy Estes proposes. 300563 21 adamarticulate+wk 2010-10-27 14:15:02 -0700 (In reply to comment #20) > I advice we remove the hardcoded mappings as Andy Estes proposes. Yes, I'm in agreement. Otherwise, I can personally attest to millions of sites (online courses and quizzes) that break. (Not to mention any other implementation of FSCommand per Adobe's spec.) For the full background on just the online training sites that break, read my full comment here: http://code.google.com/p/chromium/issues/detail?id=57986#c15 Pissed off students who don't get their scores recorded wouldn't bode so well. :) 300597 22 aestes 2010-10-27 14:44:44 -0700 (In reply to comment #20) > Actually it seems a lot of stuff still broken. See the matching bug in chrome > > It seems many sites use the flash fscommand technique to communicate with the page.Let me copy here one of the last comments in the chrome bug: > > "The reason is because content producers have simply done as they were told when writing their code. Adobe has documented that in order for fscommand to work properly, you need only to set the value of the NAME attribute of the EMBED tag or the ID property of the OBJECT tag. Setting the value of the name property has never been a requirement of fscommand. The following documentation dates back to 2000: > > http://www.adobe.com/support/flash/action_scripts/actionscript_dictionary/actionscript_dictionary372.html > > Even today, ten years later, Adobe ships their fscommand template in the latest version of Flash CS5. Attached is the output that Flash CS5 produces and the screenshot in the tool for how it’s selected. It conforms to their published doc. As expected, this template works in other browsers.." > > http://code.google.com/p/chromium/issues/detail?id=57986 > > I advice we remove the hardcoded mappings as Andy Estes proposes. Thanks for this feedback, Carlos. This provides good justification for changing our behavior. 300671 23 cpu 2010-10-27 16:09:16 -0700 Andy, Awesome. Could you fix it soon? :) I got a lot of flash devs breathing down my neck, it must be that their conference (MAX) is going on now... I can write the patch, but is it seems to me trivial and I am not a webkit contributor so it would take me a while to get up to speed. 300796 24 cpu 2010-10-27 18:57:49 -0700 actually, I take that back, just removing flash clsid from createClassIdToTypeMap() does not solve the problem as webkit is clever and has other means to retrieve the mime type. 300804 25 aestes 2010-10-27 19:12:17 -0700 (In reply to comment #24) > actually, I take that back, just removing flash clsid from createClassIdToTypeMap() does not solve the problem as webkit is clever and has other means to retrieve the mime type. Yes, even though you removed the classid->MIME mapping, WebKit asked its plug-in database for a plug-in that can handle files ending in '.swf' (taken from the movie param), and flash said that it could. The HTML5 spec says that if a non-empty classid attribute is specified that can't be handled by the UA, fallback content should be rendered (which is the embed element in these cases). So, the correct change is a little more complicated than just removing the classid mapping for flash. 300889 26 72147 aestes 2010-10-27 23:03:16 -0700 Created attachment 72147 Patch 300891 27 aestes 2010-10-27 23:04:15 -0700 <rdar://problem/8603844> 300892 28 72147 abarth 2010-10-27 23:05:20 -0700 Comment on attachment 72147 Patch Thanks Andy. 300895 29 aestes 2010-10-27 23:15:29 -0700 Committed r70748: <http://trac.webkit.org/changeset/70748> 300909 30 webkit.review.bot 2010-10-28 00:01:24 -0700 http://trac.webkit.org/changeset/70748 might have broken Qt Linux Release The following tests are not passing: platform/qt/plugins/qt-qwidget-plugin.html 300923 31 72152 aestes 2010-10-28 00:37:33 -0700 Created attachment 72152 Patch 300925 32 72152 abarth 2010-10-28 00:40:07 -0700 Comment on attachment 72152 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=72152&action=review > WebCore/html/HTMLObjectElement.cpp:223 > + ASSERT(object); ASSERT_UNUSED 300926 33 aestes 2010-10-28 00:48:18 -0700 Committed r70754: <http://trac.webkit.org/changeset/70754> 302610 34 progame+wk 2010-11-01 05:53:14 -0700 this change probably broke videos in nana10.co.il Steps to repro: 1. Visit http://10tv.nana10.co.il/ (for example). 2. If you see a Hebrew message over the black video box, answer it by clicking on בטל (this button http://f.nanafiles.co.il/partner48/Common/Images//VideoPlayer/UI/CancelSaveProfile.gif ) 3. wait for the video to load... it won't 302611 35 progame+wk 2010-11-01 05:55:08 -0700 inChromium bug related to my last reply: http://code.google.com/p/chromium/issues/detail?id=61418 302646 36 playmobil 2010-11-01 07:03:33 -0700 Yair: thanks! Could you please file a new bug. Information on whether this works in Firefox and a Safari nightly would also be really helpful. 302693 37 progame+wk 2010-11-01 08:09:42 -0700 Bug 48757 to figure out how come that change possibly broke (still just a guess) that site for Chrome only (works fine using Safari) 72147 2010-10-27 23:03:16 -0700 2010-10-27 23:05:19 -0700 Patch bug-45679-20101027230315.patch text/plain 8195 aestes SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA3MDc0NykKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMjMgQEAKKzIwMTAtMTAtMjcgIEFuZHkgRXN0ZXMgIDxhZXN0ZXNAYXBwbGUuY29t PgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIHI2NjE1 NiBicm9rZSBBdGxhc0NUIGxpYnJhcnksIGZvcm1lcmx5IGFmZmVjdGVkIGh0dHA6Ly9tYXAuZC5j by5pbC8KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTQ1 Njc5CisKKyAgICAgICAgVGhpcyBwYXRjaCByZW1vdmVzIFdlYktpdCdzIG1hcCBvZiBjbGFzc2lk IHZhbHVlcyB0byBNSU1FIHR5cGVzLiBJdCBhbHNvCisgICAgICAgIGNoYW5nZXMgdGhlIGJlaGF2 aW9yIG9mIG9iamVjdCBlbGVtZW50cyB0byByZW5kZXIgZmFsbGJhY2sgY29udGVudCB3aGVuCisg ICAgICAgIGEgbm9uLWVtcHR5IGNsYXNzaWQgYXR0cmlidXRlIGlzIHNwZWNpZmllZCwgd2hpY2gg aXMgdGhlIGJlaGF2aW9yIEhUTUw1CisgICAgICAgIHNwZWNpZmllcyB3aGVuIGEgVUEgZW5jb3Vu dGVycyBhIGNsYXNzaWQgaXQgZG9lc24ndCB1bmRlcnN0YW5kLgorCisgICAgICAgIFRlc3Q6IGZh c3QvcmVwbGFjZWQvb2JqZWN0LXdpdGgtbm9uLWVtcHR5LWNsYXNzaWQtdHJpZ2dlcnMtZmFsbGJh Y2suaHRtbAorCisgICAgICAgICogaHRtbC9IVE1MT2JqZWN0RWxlbWVudC5jcHA6IFJlbW92ZSBz ZXJ2aWNlVHlwZUZvckNsYXNzSWQoKSwKKyAgICAgICAgY3JlYXRlQ2xhc3NJZFRvVHlwZU1hcCgp LCBhbmQgdGhlIENsYXNzSWRUb1R5cGVNYXAgdHlwZWRlZi4KKyAgICAgICAgKFdlYkNvcmU6OkhU TUxPYmplY3RFbGVtZW50Ojp1cGRhdGVXaWRnZXQpOiBEbyBub3QgY2FsbAorICAgICAgICBzZXJ2 aWNlVHlwZUZvckNsYXNzSWQoKSB3aGVuIHRoZXJlIGlzIG5vIHR5cGUgYXR0cmlidXRlLCBhbmQg cmVuZGVyCisgICAgICAgIGZhbGxiYWNrIGNvbnRlbnQgaWYgdGhlIGNsYXNzaWQgYXR0cmlidXRl IGlzIG5vbi1lbXB0eS4KKwogMjAxMC0xMC0yNyAgRXJpYyBVaHJoYW5lICA8ZXJpY3VAY2hyb21p dW0ub3JnPgogCiAgICAgICAgIFJldmlld2VkIGJ5IERhdmlkIExldmluLgpJbmRleDogV2ViQ29y ZS9odG1sL0hUTUxPYmplY3RFbGVtZW50LmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBXZWJDb3JlL2h0bWwv SFRNTE9iamVjdEVsZW1lbnQuY3BwCShyZXZpc2lvbiA3MDYyMCkKKysrIFdlYkNvcmUvaHRtbC9I VE1MT2JqZWN0RWxlbWVudC5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTExNiwzMSArMTE2LDYgQEAg dm9pZCBIVE1MT2JqZWN0RWxlbWVudDo6cGFyc2VNYXBwZWRBdHRyaQogICAgICAgICBIVE1MUGx1 Z0luSW1hZ2VFbGVtZW50OjpwYXJzZU1hcHBlZEF0dHJpYnV0ZShhdHRyKTsKIH0KIAotdHlwZWRl ZiBIYXNoTWFwPFN0cmluZywgU3RyaW5nLCBDYXNlRm9sZGluZ0hhc2g+IENsYXNzSWRUb1R5cGVN YXA7Ci0KLXN0YXRpYyBDbGFzc0lkVG9UeXBlTWFwKiBjcmVhdGVDbGFzc0lkVG9UeXBlTWFwKCkK LXsKLSAgICBDbGFzc0lkVG9UeXBlTWFwKiBtYXAgPSBuZXcgQ2xhc3NJZFRvVHlwZU1hcDsKLSAg ICBtYXAtPmFkZCgiY2xzaWQ6RDI3Q0RCNkUtQUU2RC0xMUNGLTk2QjgtNDQ0NTUzNTQwMDAwIiwg ImFwcGxpY2F0aW9uL3gtc2hvY2t3YXZlLWZsYXNoIik7Ci0gICAgbWFwLT5hZGQoImNsc2lkOkNG Q0RBQTAzLThCRTQtMTFDRi1CODRCLTAwMjBBRkJCQ0NGQSIsICJhdWRpby94LXBuLXJlYWxhdWRp by1wbHVnaW4iKTsKLSAgICBtYXAtPmFkZCgiY2xzaWQ6MDJCRjI1RDUtOEMxNy00QjIzLUJDODAt RDM0ODhBQkREQzZCIiwgInZpZGVvL3F1aWNrdGltZSIpOwotICAgIG1hcC0+YWRkKCJjbHNpZDox NjZCMUJDQS0zRjlDLTExQ0YtODA3NS00NDQ1NTM1NDAwMDAiLCAiYXBwbGljYXRpb24veC1kaXJl Y3RvciIpOwotICAgIG1hcC0+YWRkKCJjbHNpZDo2QkY1MkE1Mi0zOTRBLTExRDMtQjE1My0wMEMw NEY3OUZBQTYiLCAiYXBwbGljYXRpb24veC1tcGxheWVyMiIpOwotICAgIG1hcC0+YWRkKCJjbHNp ZDoyMkQ2RjMxMi1CMEY2LTExRDAtOTRBQi0wMDgwQzc0QzdFOTUiLCAiYXBwbGljYXRpb24veC1t cGxheWVyMiIpOwotICAgIHJldHVybiBtYXA7Ci19Ci0KLXN0YXRpYyBTdHJpbmcgc2VydmljZVR5 cGVGb3JDbGFzc0lkKGNvbnN0IFN0cmluZyYgY2xhc3NJZCkKLXsKLSAgICAvLyBSZXR1cm4gZWFy bHkgaWYgY2xhc3NJZCBpcyBlbXB0eSAoc2luY2Ugd2Ugd29uJ3QgZG8gYW55dGhpbmcgYmVsb3cp LgotICAgIC8vIEZ1cnRoZXJtb3JlLCBpZiBjbGFzc0lkIGlzIG51bGwsIGNhbGxpbmcgZ2V0KCkg YmVsb3cgd2lsbCBjcmFzaC4KLSAgICBpZiAoY2xhc3NJZC5pc0VtcHR5KCkpCi0gICAgICAgIHJl dHVybiBTdHJpbmcoKTsKLSAgICAKLSAgICBzdGF0aWMgQ2xhc3NJZFRvVHlwZU1hcCogbWFwID0g Y3JlYXRlQ2xhc3NJZFRvVHlwZU1hcCgpOwotICAgIHJldHVybiBtYXAtPmdldChjbGFzc0lkKTsK LX0KLQogc3RhdGljIHZvaWQgbWFwRGF0YVBhcmFtVG9TcmMoVmVjdG9yPFN0cmluZz4qIHBhcmFt TmFtZXMsIFZlY3RvcjxTdHJpbmc+KiBwYXJhbVZhbHVlcykKIHsKICAgICAvLyBTb21lIHBsdWdp bnMgZG9uJ3QgdW5kZXJzdGFuZCB0aGUgImRhdGEiIGF0dHJpYnV0ZSBvZiB0aGUgT0JKRUNUIHRh ZyAoaS5lLiBSZWFsIGFuZCBXTVAKQEAgLTI1NSwxNCArMjMwLDkgQEAgdm9pZCBIVE1MT2JqZWN0 RWxlbWVudDo6dXBkYXRlV2lkZ2V0KGJvbwogICAgIC8vIEZJWE1FOiBUaGlzIHNob3VsZCBBU1NF UlQgaXNGaW5pc2hlZFBhcnNpbmdDaGlsZHJlbigpIGluc3RlYWQuCiAgICAgaWYgKCFpc0Zpbmlz aGVkUGFyc2luZ0NoaWxkcmVuKCkpCiAgICAgICAgIHJldHVybjsKLQotICAgIFN0cmluZyB1cmwg PSB0aGlzLT51cmwoKTsKICAgICAKLSAgICAvLyBJZiB0aGUgb2JqZWN0IGRvZXMgbm90IHNwZWNp ZnkgYSBNSU1FIHR5cGUgdmlhIGEgdHlwZSBhdHRyaWJ1dGUsIGJ1dCBkb2VzCi0gICAgLy8gY29u dGFpbiBhIGNsYXNzaWQgYXR0cmlidXRlLCB0cnkgdG8gbWFwIHRoZSBjbGFzc2lkIHRvIGEgTUlN RSB0eXBlLgorICAgIFN0cmluZyB1cmwgPSB0aGlzLT51cmwoKTsKICAgICBTdHJpbmcgc2Vydmlj ZVR5cGUgPSB0aGlzLT5zZXJ2aWNlVHlwZSgpOwotICAgIGlmIChzZXJ2aWNlVHlwZS5pc0VtcHR5 KCkpCi0gICAgICAgIHNlcnZpY2VUeXBlID0gc2VydmljZVR5cGVGb3JDbGFzc0lkKGNsYXNzSWQo KSk7CiAKICAgICAvLyBGSVhNRTogVGhlc2Ugc2hvdWxkIGJlIGpvaW5lZCBpbnRvIGEgUGx1Z2lu UGFyYW1ldGVycyBjbGFzcy4KICAgICBWZWN0b3I8U3RyaW5nPiBwYXJhbU5hbWVzOwpAQCAtMjkx LDggKzI2MSwxMiBAQCB2b2lkIEhUTUxPYmplY3RFbGVtZW50Ojp1cGRhdGVXaWRnZXQoYm9vCiAg ICAgaWYgKCFyZW5kZXJlcigpKQogICAgICAgICByZXR1cm47CiAKKyAgICAvLyBIVE1MNSBzYXlz IHRoYXQgZmFsbGJhY2sgY29udGVudCBzaG91bGQgYmUgcmVuZGVyZWQgaWYgYSBub24tZW1wdHkK KyAgICAvLyBjbGFzc2lkIGlzIHNwZWNpZmllZCBmb3Igd2hpY2ggdGhlIFVBIGNhbid0IGZpbmQg YSBzdWl0YWJsZSBwbHVnLWluLgorICAgIGJvb2wgaGFzRW1wdHlDbGFzc0lkID0gY2xhc3NJZCgp LmlzRW1wdHkoKTsKKwogICAgIFN1YmZyYW1lTG9hZGVyKiBsb2FkZXIgPSBkb2N1bWVudCgpLT5m cmFtZSgpLT5sb2FkZXIoKS0+c3ViZnJhbWVMb2FkZXIoKTsKLSAgICBib29sIHN1Y2Nlc3MgPSBi ZWZvcmVMb2FkQWxsb3dlZExvYWQgJiYgbG9hZGVyLT5yZXF1ZXN0T2JqZWN0KHRoaXMsIHVybCwg Z2V0QXR0cmlidXRlKG5hbWVBdHRyKSwgc2VydmljZVR5cGUsIHBhcmFtTmFtZXMsIHBhcmFtVmFs dWVzKTsKKyAgICBib29sIHN1Y2Nlc3MgPSBiZWZvcmVMb2FkQWxsb3dlZExvYWQgJiYgaGFzRW1w dHlDbGFzc0lkICYmIGxvYWRlci0+cmVxdWVzdE9iamVjdCh0aGlzLCB1cmwsIGdldEF0dHJpYnV0 ZShuYW1lQXR0ciksIHNlcnZpY2VUeXBlLCBwYXJhbU5hbWVzLCBwYXJhbVZhbHVlcyk7CiAKICAg ICBpZiAoIXN1Y2Nlc3MgJiYgZmFsbGJhY2tDb250ZW50KQogICAgICAgICByZW5kZXJGYWxsYmFj a0NvbnRlbnQoKTsKSW5kZXg6IExheW91dFRlc3RzL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBM YXlvdXRUZXN0cy9DaGFuZ2VMb2cJKHJldmlzaW9uIDcwNzQ3KQorKysgTGF5b3V0VGVzdHMvQ2hh bmdlTG9nCSh3b3JraW5nIGNvcHkpCkBAIC0xLDMgKzEsMTYgQEAKKzIwMTAtMTAtMjcgIEFuZHkg RXN0ZXMgIDxhZXN0ZXNAYXBwbGUuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAo T09QUyEpLgorCisgICAgICAgIHI2NjE1NiBicm9rZSBBdGxhc0NUIGxpYnJhcnksIGZvcm1lcmx5 IGFmZmVjdGVkIGh0dHA6Ly9tYXAuZC5jby5pbC8KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtp dC5vcmcvc2hvd19idWcuY2dpP2lkPTQ1Njc5CisKKyAgICAgICAgKiBmYXN0L2RvbS9vYmplY3Qt ZW1iZWQtcGx1Z2luLXNjcmlwdGluZy5odG1sOiBDaGFuZ2VkIGNsYXNzaWQKKyAgICAgICAgYXR0 cmlidXRlcyB0byB0eXBlIGF0dHJpYnV0ZXMuCisgICAgICAgICogZmFzdC9kb20vb2JqZWN0LXBs dWdpbi1oaWRlcy1wcm9wZXJ0aWVzLmh0bWw6IERpdHRvLgorICAgICAgICAqIGZhc3QvcmVwbGFj ZWQvb2JqZWN0LXdpdGgtbm9uLWVtcHR5LWNsYXNzaWQtdHJpZ2dlcnMtZmFsbGJhY2stZXhwZWN0 ZWQudHh0OiBBZGRlZC4KKyAgICAgICAgKiBmYXN0L3JlcGxhY2VkL29iamVjdC13aXRoLW5vbi1l bXB0eS1jbGFzc2lkLXRyaWdnZXJzLWZhbGxiYWNrLmh0bWw6IEFkZGVkLgorCiAyMDEwLTEwLTI3 ICBLaW51a28gWWFzdWRhICA8a2ludWtvQGNocm9taXVtLm9yZz4KIAogICAgICAgICBVbnJldmll d2VkLCB1cGRhdGluZyBkcnRfZXhwZWN0YXRpb25zLnR4dC4KSW5kZXg6IExheW91dFRlc3RzL2Zh c3QvZG9tL29iamVjdC1lbWJlZC1wbHVnaW4tc2NyaXB0aW5nLmh0bWwKPT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0g TGF5b3V0VGVzdHMvZmFzdC9kb20vb2JqZWN0LWVtYmVkLXBsdWdpbi1zY3JpcHRpbmcuaHRtbAko cmV2aXNpb24gNzA2MjApCisrKyBMYXlvdXRUZXN0cy9mYXN0L2RvbS9vYmplY3QtZW1iZWQtcGx1 Z2luLXNjcmlwdGluZy5odG1sCSh3b3JraW5nIGNvcHkpCkBAIC0zNCw3ICszNCw3IEBAIGZ1bmN0 aW9uIHRlc3QoKSAKIAogPE9CSkVDVCAKICAgICBpZD0ibXlPIgotICAgIGNsYXNzaWQ9ImNsc2lk OjAyQkYyNUQ1LThDMTctNEIyMy1CQzgwLUQzNDg4QUJEREM2QiIKKyAgICB0eXBlPSJhdWRpby9t cDQiCiAgICAgd2lkdGggPSAwIGhlaWdodCA9IDAKICAgICA+CiAgICAgPFBBUkFNIG5hbWU9InNy YyIgdmFsdWU9InJlc291cmNlcy9hcnRpY2xlcy5tNGEiPgpAQCAtMTAwLDcgKzEwMCw3IEBAIGZ1 bmN0aW9uIHRlc3QoKSAKIAogPG9iamVjdAogICAgIG5hbWU9IlBsdWdpbiIKLSAgICBjbGFzc2lk PSJjbHNpZDowMkJGMjVENS04QzE3LTRCMjMtQkM4MC1EMzQ4OEFCRERDNkIiCisgICAgdHlwZT0i YXVkaW8vbXA0IgogICAgIHdpZHRoID0gMCBoZWlnaHQgPSAwCiAgICAgPgogICAgIDxwYXJhbSBu YW1lPSJzcmMiIHZhbHVlPSJyZXNvdXJjZXMvYXJ0aWNsZXMubTRhIj4KSW5kZXg6IExheW91dFRl c3RzL2Zhc3QvZG9tL29iamVjdC1wbHVnaW4taGlkZXMtcHJvcGVydGllcy5odG1sCj09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT0KLS0tIExheW91dFRlc3RzL2Zhc3QvZG9tL29iamVjdC1wbHVnaW4taGlkZXMtcHJvcGVydGll cy5odG1sCShyZXZpc2lvbiA3MDYyMCkKKysrIExheW91dFRlc3RzL2Zhc3QvZG9tL29iamVjdC1w bHVnaW4taGlkZXMtcHJvcGVydGllcy5odG1sCSh3b3JraW5nIGNvcHkpCkBAIC0xOSw2ICsxOSw2 IEBAIHRoZSBPQkpFQ1QgZWxlbWVudC4KIElmIHRoZSB0ZXN0IHBhc3NlcywgeW91IHdpbGwgc2Vl IGEgIlBBU1NFRCIgbWVzc2FnZSBiZWxvdy4KIDwvcD4KIDxwIGlkPSJyZXN1bHQiPjwvcD4KLTxv YmplY3QgaWQ9Im9iaiIgY2xhc3NpZD0iY2xzaWQ6MDJCRjI1RDUtOEMxNy00QjIzLUJDODAtRDM0 ODhBQkREQzZCIj48IS0tIHF1aWNrdGltZSAtLT4KKzxvYmplY3QgaWQ9Im9iaiIgdHlwZT0idmlk ZW8vcXVpY2t0aW1lIj48IS0tIHF1aWNrdGltZSAtLT4KIDwvb2JqZWN0PgogPC9ib2R5PgpJbmRl eDogTGF5b3V0VGVzdHMvZmFzdC9yZXBsYWNlZC9vYmplY3Qtd2l0aC1ub24tZW1wdHktY2xhc3Np ZC10cmlnZ2Vycy1mYWxsYmFjay1leHBlY3RlZC50eHQKPT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVz dHMvZmFzdC9yZXBsYWNlZC9vYmplY3Qtd2l0aC1ub24tZW1wdHktY2xhc3NpZC10cmlnZ2Vycy1m YWxsYmFjay1leHBlY3RlZC50eHQJKHJldmlzaW9uIDApCisrKyBMYXlvdXRUZXN0cy9mYXN0L3Jl cGxhY2VkL29iamVjdC13aXRoLW5vbi1lbXB0eS1jbGFzc2lkLXRyaWdnZXJzLWZhbGxiYWNrLWV4 cGVjdGVkLnR4dAkocmV2aXNpb24gMCkKQEAgLTAsMCArMSw0IEBACitUaGlzIHRlc3RzIHRoYXQg ZmFsbGJhY2sgY29udGVudCBpcyByZW5kZXJlZCBmb3Igb2JqZWN0cyB3aXRoIG5vbi1lbXB0eSBj bGFzc2lkIGF0dHJpYnV0ZXMuIFRoZSB0ZXN0IHBhc3NlcyBpZiB0d28gbGluZXMgYXJlIHByaW50 ZWQgYmVsb3cgY29udGFpbmluZyB0aGUgd29yayAnUEFTUycuCisKK29iamVjdCB3aXRoIGNsYXNz aWQgYXR0cmlidXRlIGJ1dCBubyB0eXBlIGF0dHJpYnV0ZSByZW5kZXJzIGZhbGxiYWNrOiBQQVNT IAorb2JqZWN0IHdpdGggY2xhc3NpZCBhbmQgdHlwZSBhdHRyaWJ1dGVzIHJlbmRlcnMgZmFsbGJh Y2s6IFBBU1MKSW5kZXg6IExheW91dFRlc3RzL2Zhc3QvcmVwbGFjZWQvb2JqZWN0LXdpdGgtbm9u LWVtcHR5LWNsYXNzaWQtdHJpZ2dlcnMtZmFsbGJhY2suaHRtbAo9PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBMYXlv dXRUZXN0cy9mYXN0L3JlcGxhY2VkL29iamVjdC13aXRoLW5vbi1lbXB0eS1jbGFzc2lkLXRyaWdn ZXJzLWZhbGxiYWNrLmh0bWwJKHJldmlzaW9uIDApCisrKyBMYXlvdXRUZXN0cy9mYXN0L3JlcGxh Y2VkL29iamVjdC13aXRoLW5vbi1lbXB0eS1jbGFzc2lkLXRyaWdnZXJzLWZhbGxiYWNrLmh0bWwJ KHJldmlzaW9uIDApCkBAIC0wLDAgKzEsMTMgQEAKKzxzY3JpcHQ+CisgICAgaWYgKHdpbmRvdy5s YXlvdXRUZXN0Q29udHJvbGxlcikKKyAgICAgICAgbGF5b3V0VGVzdENvbnRyb2xsZXIuZHVtcEFz VGV4dCgpOworPC9zY3JpcHQ+Cis8cD5UaGlzIHRlc3RzIHRoYXQgZmFsbGJhY2sgY29udGVudCBp cyByZW5kZXJlZCBmb3Igb2JqZWN0cyB3aXRoIG5vbi1lbXB0eSBjbGFzc2lkIGF0dHJpYnV0ZXMu IFRoZSB0ZXN0IHBhc3NlcyBpZiB0d28gbGluZXMgYXJlIHByaW50ZWQgYmVsb3cgY29udGFpbmlu ZyB0aGUgd29yayAnUEFTUycuPC9wPgorPG9iamVjdCBpZD0ib2JqIiBjbGFzc2lkPSJjbHNpZDpE MjdDREI2RS1BRTZELTExY2YtOTZCOC00NDQ1NTM1NDAwMDAiPgorICAgIG9iamVjdCB3aXRoIGNs YXNzaWQgYXR0cmlidXRlIGJ1dCBubyB0eXBlIGF0dHJpYnV0ZSByZW5kZXJzIGZhbGxiYWNrOiBQ QVNTCis8L29iamVjdD4KKzxicj4KKzxvYmplY3QgaWQ9Im9iaiIgY2xhc3NpZD0iY2xzaWQ6RDI3 Q0RCNkUtQUU2RC0xMWNmLTk2QjgtNDQ0NTUzNTQwMDAwIiB0eXBlPWFwcGxpY2F0aW9uL3gtd2Vi a2l0LXRlc3QtbmV0c2NhcGUiPgorICAgIG9iamVjdCB3aXRoIGNsYXNzaWQgYW5kIHR5cGUgYXR0 cmlidXRlcyByZW5kZXJzIGZhbGxiYWNrOiBQQVNTCis8L29iamVjdD4KKwo= 72152 2010-10-28 00:37:33 -0700 2010-10-28 00:40:07 -0700 Patch bug-45679-20101028003732.patch text/plain 2389 aestes SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA3MDc1MikKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTcgQEAKKzIwMTAtMTAtMjggIEFuZHkgRXN0ZXMgIDxhZXN0ZXNAYXBwbGUuY29t PgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIEZpeCBh IHRlc3QgZmFpbHVyZSBpbnRyb2R1Y2VkIGluIHI3MDc0OCBieSBzdXBwb3J0aW5nIFF0J3Mgbm9u LXN0YW5kYXJkCisgICAgICAgIHVzZSBvZiBjbGFzc2lkLgorICAgICAgICBodHRwczovL2J1Z3Mu d2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9NDU2NzkKKworICAgICAgICAqIGh0bWwvSFRNTE9i amVjdEVsZW1lbnQuY3BwOgorICAgICAgICAoV2ViQ29yZTo6b2JqZWN0SGFzU3VwcG9ydGVkQ2xh c3NJZCk6IFJldHVybiB0cnVlIGlmIHRoZSBvYmplY3QncworICAgICAgICBzZXJ2aWNlVHlwZSBp cyAnYXBwbGljYXRpb24veC1xdC1wbHVnaW4nLgorICAgICAgICAoV2ViQ29yZTo6SFRNTE9iamVj dEVsZW1lbnQ6OnVwZGF0ZVdpZGdldCk6IERvIG5vdCByZW5kZXIgZmFsbGJhY2sKKyAgICAgICAg Y29udGVudCBpZiBhIG5vbi1lbXB0eSBjbGFzc2lkIGlzIHNwZWNpZmllZCBmb3IgYSBRdCBwbHVn aW4gb2JqZWN0LgorCiAyMDEwLTEwLTI4ICBJdmFuIEtyc3RpxIcgIDxpa2VAYXBwbGUuY29tPgog CiAgICAgICAgIFJldmlld2VkIGJ5IE1hcmsgUm93ZS4KSW5kZXg6IFdlYkNvcmUvaHRtbC9IVE1M T2JqZWN0RWxlbWVudC5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gV2ViQ29yZS9odG1sL0hUTUxPYmplY3RF bGVtZW50LmNwcAkocmV2aXNpb24gNzA3NDgpCisrKyBXZWJDb3JlL2h0bWwvSFRNTE9iamVjdEVs ZW1lbnQuY3BwCSh3b3JraW5nIGNvcHkpCkBAIC0yMTcsNiArMjE3LDE3IEBAIGJvb2wgSFRNTE9i amVjdEVsZW1lbnQ6Omhhc0ZhbGxiYWNrQ29udGUKICAgICB9CiAgICAgcmV0dXJuIGZhbHNlOwog fQorICAgIAorc3RhdGljIGJvb2wgb2JqZWN0SGFzU3VwcG9ydGVkQ2xhc3NJZChIVE1MT2JqZWN0 RWxlbWVudCogb2JqZWN0KQoreworICAgIEFTU0VSVChvYmplY3QpOworI2lmIFBMQVRGT1JNKFFU KQorICAgIC8vIFF0IHBsdWctaW5zIHVzZSBjbGFzc2lkIHRvIHNwZWNpZnkgd2hpY2ggUU9iamVj dCB0byBsb2FkLgorICAgIHJldHVybiBlcXVhbHNJZ25vcmluZ0Nhc2Uob2JqZWN0LT5zZXJ2aWNl VHlwZSgpLCAiYXBwbGljYXRpb24veC1xdC1wbHVnaW4iKTsKKyNlbHNlCisgICAgcmV0dXJuIGZh bHNlOworI2VuZGlmCit9CiAKIC8vIEZJWE1FOiBUaGlzIHNob3VsZCBiZSB1bmlmaWVkIHdpdGgg SFRNTEVtYmVkRWxlbWVudDo6dXBkYXRlV2lkZ2V0IGFuZAogLy8gbW92ZWQgZG93biBpbnRvIEhU TUxQbHVnaW5JbWFnZUVsZW1lbnQuY3BwCkBAIC0yNjMsMTAgKzI3NCwxMCBAQCB2b2lkIEhUTUxP YmplY3RFbGVtZW50Ojp1cGRhdGVXaWRnZXQoYm9vCiAKICAgICAvLyBIVE1MNSBzYXlzIHRoYXQg ZmFsbGJhY2sgY29udGVudCBzaG91bGQgYmUgcmVuZGVyZWQgaWYgYSBub24tZW1wdHkKICAgICAv LyBjbGFzc2lkIGlzIHNwZWNpZmllZCBmb3Igd2hpY2ggdGhlIFVBIGNhbid0IGZpbmQgYSBzdWl0 YWJsZSBwbHVnLWluLgotICAgIGJvb2wgaGFzRW1wdHlDbGFzc0lkID0gY2xhc3NJZCgpLmlzRW1w dHkoKTsKKyAgICBib29sIGhhc1ZhbGlkQ2xhc3NJZCA9IGNsYXNzSWQoKS5pc0VtcHR5KCkgfHwg b2JqZWN0SGFzU3VwcG9ydGVkQ2xhc3NJZCh0aGlzKTsKIAogICAgIFN1YmZyYW1lTG9hZGVyKiBs b2FkZXIgPSBkb2N1bWVudCgpLT5mcmFtZSgpLT5sb2FkZXIoKS0+c3ViZnJhbWVMb2FkZXIoKTsK LSAgICBib29sIHN1Y2Nlc3MgPSBiZWZvcmVMb2FkQWxsb3dlZExvYWQgJiYgaGFzRW1wdHlDbGFz c0lkICYmIGxvYWRlci0+cmVxdWVzdE9iamVjdCh0aGlzLCB1cmwsIGdldEF0dHJpYnV0ZShuYW1l QXR0ciksIHNlcnZpY2VUeXBlLCBwYXJhbU5hbWVzLCBwYXJhbVZhbHVlcyk7CisgICAgYm9vbCBz dWNjZXNzID0gYmVmb3JlTG9hZEFsbG93ZWRMb2FkICYmIGhhc1ZhbGlkQ2xhc3NJZCAmJiBsb2Fk ZXItPnJlcXVlc3RPYmplY3QodGhpcywgdXJsLCBnZXRBdHRyaWJ1dGUobmFtZUF0dHIpLCBzZXJ2 aWNlVHlwZSwgcGFyYW1OYW1lcywgcGFyYW1WYWx1ZXMpOwogCiAgICAgaWYgKCFzdWNjZXNzICYm IGZhbGxiYWNrQ29udGVudCkKICAgICAgICAgcmVuZGVyRmFsbGJhY2tDb250ZW50KCk7Cg==