267348 2024-01-10 09:16:39 -0800 AX: ASSERTION FAILED: m_table inside of AXIsolatedTree::create 2024-01-11 21:45:24 -0800 1 1 1 Unclassified WebKit Accessibility Safari Technology Preview All All RESOLVED FIXED InRadar P2 Normal --- 1 dm_mazzoni jhoffman23 aboxhall andresg_22 apinheiro cfleizach dmazzoni ews-watchlist jcraig jhoffman23 samuel_white webkit-bug-importer oldest_to_newest 2004123 0 dm_mazzoni 2024-01-10 09:16:39 -0800 When I build WebKit on tip-of-tree in debug mode, then visit any webpage (even a simple Hello, world html page) with VoiceOver running, I get an assertion failure: ASSERTION FAILED: m_table /Volumes/Code/safari/OpenSource/WebKitBuild/Debug/usr/local/include/wtf/HashTable.h(231) The crash is happening inside of AXIsolatedTree::create at the top of this for loop: for (auto& relatedObjectID : axObjectCache.relations().keys()) { That loop was added here, so I think it's likely a recent regression from that change: AX: Move objectsForIds off of the main thread https://bugs.webkit.org/show_bug.cgi?id=266608 rdar://116133586 2004124 1 webkit-bug-importer 2024-01-10 09:16:49 -0800 <rdar://problem/120788145> 2004334 2 469369 jhoffman23 2024-01-10 22:18:45 -0800 Created attachment 469369 Patch 2004335 3 469370 jhoffman23 2024-01-10 22:20:14 -0800 Created attachment 469370 Patch 2004342 4 469370 cfleizach 2024-01-10 22:28:31 -0800 Comment on attachment 469370 Patch How does this work? Won't getting a reference to the real thing still reference the underlying thing that can be modified? 2004343 5 jhoffman23 2024-01-10 22:32:31 -0800 (In reply to chris fleizach from comment #4) > Comment on attachment 469370 [details] > Patch > > How does this work? Won't getting a reference to the real thing still > reference the underlying thing that can be modified? When we would call axObjectCache.relations(), we always had the chance of modifying m_relations since updateRelationsIsNeeded is part of that method. Now we prevent that possibility. We can hold a value if we want to be extra safe, but there is some space/time complexity added there—do you have a preference? 2004350 6 cfleizach 2024-01-10 22:48:15 -0800 (In reply to Joshua Hoffman from comment #5) > (In reply to chris fleizach from comment #4) > > Comment on attachment 469370 [details] > > Patch > > > > How does this work? Won't getting a reference to the real thing still > > reference the underlying thing that can be modified? > > When we would call axObjectCache.relations(), we always had the chance of > modifying m_relations since updateRelationsIsNeeded is part of that method. > Now we prevent that possibility. We can hold a value if we want to be extra > safe, but there is some space/time complexity added there—do you have a > preference? I'm just saying isn't this a reference, so it can still be modified during iteration? 2004352 7 jhoffman23 2024-01-10 22:56:26 -0800 (In reply to chris fleizach from comment #6) > (In reply to Joshua Hoffman from comment #5) > > (In reply to chris fleizach from comment #4) > > > Comment on attachment 469370 [details] > > > Patch > > > > > > How does this work? Won't getting a reference to the real thing still > > > reference the underlying thing that can be modified? > > > > When we would call axObjectCache.relations(), we always had the chance of > > modifying m_relations since updateRelationsIsNeeded is part of that method. > > Now we prevent that possibility. We can hold a value if we want to be extra > > safe, but there is some space/time complexity added there—do you have a > > preference? > > I'm just saying isn't this a reference, so it can still be modified during > iteration? Yeah, that is true. I'll update it to `const auto relations =` so we use a copy instead. 2004353 8 469372 jhoffman23 2024-01-10 23:00:27 -0800 Created attachment 469372 Patch 2004420 9 andresg_22 2024-01-11 06:26:34 -0800 (In reply to Joshua Hoffman from comment #8) > Created attachment 469372 [details] > Patch diff --git a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp index f6f0622abe93..c003ae7462a9 100644 --- a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp +++ b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp @@ -160,9 +160,10 @@ Ref<AXIsolatedTree> AXIsolatedTree::create(AXObjectCache& axObjectCache) tree->updateLoadingProgress(axObjectCache.loadingProgress()); - tree->updateRelations(axObjectCache.relations()); + const auto relations = axObjectCache.relations(); + tree->updateRelations(relations); - for (auto& relatedObjectID : axObjectCache.relations().keys()) { + for (auto& relatedObjectID : relations.keys()) { RefPtr axObject = axObjectCache.objectForID(relatedObjectID); if (axObject && axObject->accessibilityIsIgnored()) tree->addUnconnectedNode(axObject.releaseNonNull()); AG: I don't think we need this for loop at all since the unconnected iso objects are added in AXObjectCache::addRelation when relations are updated. I think this is another left over from a previous change that wasn't necessary. But I may be missing something... 2004434 10 jhoffman23 2024-01-11 07:50:38 -0800 (In reply to Andres Gonzalez from comment #9) > (In reply to Joshua Hoffman from comment #8) > > Created attachment 469372 [details] > > Patch > > diff --git a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > index f6f0622abe93..c003ae7462a9 100644 > --- a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > +++ b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > @@ -160,9 +160,10 @@ Ref<AXIsolatedTree> > AXIsolatedTree::create(AXObjectCache& axObjectCache) > > tree->updateLoadingProgress(axObjectCache.loadingProgress()); > > - tree->updateRelations(axObjectCache.relations()); > + const auto relations = axObjectCache.relations(); > + tree->updateRelations(relations); > > - for (auto& relatedObjectID : axObjectCache.relations().keys()) { > + for (auto& relatedObjectID : relations.keys()) { > RefPtr axObject = axObjectCache.objectForID(relatedObjectID); > if (axObject && axObject->accessibilityIsIgnored()) > tree->addUnconnectedNode(axObject.releaseNonNull()); > > AG: I don't think we need this for loop at all since the unconnected iso > objects are added in AXObjectCache::addRelation when relations are updated. > I think this is another left over from a previous change that wasn't > necessary. But I may be missing something... We need this loop in addition to the code in addRelations so that if the isolated tree is enabled/initialized after relations have been added, we still create unconnected objects for the existing relations on the tree. 2004435 11 andresg_22 2024-01-11 07:58:40 -0800 (In reply to Joshua Hoffman from comment #10) > (In reply to Andres Gonzalez from comment #9) > > (In reply to Joshua Hoffman from comment #8) > > > Created attachment 469372 [details] > > > Patch > > > > diff --git a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > index f6f0622abe93..c003ae7462a9 100644 > > --- a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > +++ b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > @@ -160,9 +160,10 @@ Ref<AXIsolatedTree> > > AXIsolatedTree::create(AXObjectCache& axObjectCache) > > > > tree->updateLoadingProgress(axObjectCache.loadingProgress()); > > > > - tree->updateRelations(axObjectCache.relations()); > > + const auto relations = axObjectCache.relations(); > > + tree->updateRelations(relations); > > > > - for (auto& relatedObjectID : axObjectCache.relations().keys()) { > > + for (auto& relatedObjectID : relations.keys()) { > > RefPtr axObject = axObjectCache.objectForID(relatedObjectID); > > if (axObject && axObject->accessibilityIsIgnored()) > > tree->addUnconnectedNode(axObject.releaseNonNull()); > > > > AG: I don't think we need this for loop at all since the unconnected iso > > objects are added in AXObjectCache::addRelation when relations are updated. > > I think this is another left over from a previous change that wasn't > > necessary. But I may be missing something... > > We need this loop in addition to the code in addRelations so that if the > isolated tree is enabled/initialized after relations have been added, we > still create unconnected objects for the existing relations on the tree. Wouldn't it make more sense to dirty the relations in that scenario? 2004485 12 jhoffman23 2024-01-11 10:26:40 -0800 (In reply to Andres Gonzalez from comment #11) > (In reply to Joshua Hoffman from comment #10) > > (In reply to Andres Gonzalez from comment #9) > > > (In reply to Joshua Hoffman from comment #8) > > > > Created attachment 469372 [details] > > > > Patch > > > > > > diff --git a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > index f6f0622abe93..c003ae7462a9 100644 > > > --- a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > +++ b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > @@ -160,9 +160,10 @@ Ref<AXIsolatedTree> > > > AXIsolatedTree::create(AXObjectCache& axObjectCache) > > > > > > tree->updateLoadingProgress(axObjectCache.loadingProgress()); > > > > > > - tree->updateRelations(axObjectCache.relations()); > > > + const auto relations = axObjectCache.relations(); > > > + tree->updateRelations(relations); > > > > > > - for (auto& relatedObjectID : axObjectCache.relations().keys()) { > > > + for (auto& relatedObjectID : relations.keys()) { > > > RefPtr axObject = axObjectCache.objectForID(relatedObjectID); > > > if (axObject && axObject->accessibilityIsIgnored()) > > > tree->addUnconnectedNode(axObject.releaseNonNull()); > > > > > > AG: I don't think we need this for loop at all since the unconnected iso > > > objects are added in AXObjectCache::addRelation when relations are updated. > > > I think this is another left over from a previous change that wasn't > > > necessary. But I may be missing something... > > > > We need this loop in addition to the code in addRelations so that if the > > isolated tree is enabled/initialized after relations have been added, we > > still create unconnected objects for the existing relations on the tree. > > Wouldn't it make more sense to dirty the relations in that scenario? I worry that dirtying relations if they have already been created could add some time complexity issues, but definitely open to discussing that. In terms of why adding the unconnected nodes here would be needed if ITM is already on, I added logging in both AXObjectCache::addRelation and AXIsolatedTree::createEmpty, and because of timing, we begin to add relations while the isolated tree is also being created, meaning we would miss out on some relations: LOG: AXObjectCache::addRelation LOG: AXObjectCache::addRelation LOG: AXObjectCache::addRelation LOG: AXObjectCache::addRelation LOG: AXObjectCache::addRelation LOG: AXIsolatedTree::createEmpty LOG: AXIsolatedTree::create LOG: AXObjectCache::addRelation LOG: AXObjectCache::addRelation 2004534 13 andresg_22 2024-01-11 13:30:18 -0800 (In reply to Joshua Hoffman from comment #12) > (In reply to Andres Gonzalez from comment #11) > > (In reply to Joshua Hoffman from comment #10) > > > (In reply to Andres Gonzalez from comment #9) > > > > (In reply to Joshua Hoffman from comment #8) > > > > > Created attachment 469372 [details] > > > > > Patch > > > > > > > > diff --git a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > > b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > > index f6f0622abe93..c003ae7462a9 100644 > > > > --- a/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > > +++ b/Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp > > > > @@ -160,9 +160,10 @@ Ref<AXIsolatedTree> > > > > AXIsolatedTree::create(AXObjectCache& axObjectCache) > > > > > > > > tree->updateLoadingProgress(axObjectCache.loadingProgress()); > > > > > > > > - tree->updateRelations(axObjectCache.relations()); > > > > + const auto relations = axObjectCache.relations(); > > > > + tree->updateRelations(relations); > > > > > > > > - for (auto& relatedObjectID : axObjectCache.relations().keys()) { > > > > + for (auto& relatedObjectID : relations.keys()) { > > > > RefPtr axObject = axObjectCache.objectForID(relatedObjectID); > > > > if (axObject && axObject->accessibilityIsIgnored()) > > > > tree->addUnconnectedNode(axObject.releaseNonNull()); > > > > > > > > AG: I don't think we need this for loop at all since the unconnected iso > > > > objects are added in AXObjectCache::addRelation when relations are updated. > > > > I think this is another left over from a previous change that wasn't > > > > necessary. But I may be missing something... > > > > > > We need this loop in addition to the code in addRelations so that if the > > > isolated tree is enabled/initialized after relations have been added, we > > > still create unconnected objects for the existing relations on the tree. > > > > Wouldn't it make more sense to dirty the relations in that scenario? > > I worry that dirtying relations if they have already been created could add > some time complexity issues, but definitely open to discussing that. > > In terms of why adding the unconnected nodes here would be needed if ITM is > already on, I added logging in both AXObjectCache::addRelation and > AXIsolatedTree::createEmpty, and because of timing, we begin to add > relations while the isolated tree is also being created, meaning we would > miss out on some relations: > > LOG: AXObjectCache::addRelation > LOG: AXObjectCache::addRelation > LOG: AXObjectCache::addRelation > LOG: AXObjectCache::addRelation > LOG: AXObjectCache::addRelation > LOG: AXIsolatedTree::createEmpty > LOG: AXIsolatedTree::create > LOG: AXObjectCache::addRelation > LOG: AXObjectCache::addRelation Thanks for the additional confirmation. 2004616 14 ews-feeder 2024-01-11 21:45:21 -0800 Committed 272958@main (0c339289de37): <https://commits.webkit.org/272958@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 469372. 469369 2024-01-10 22:18:45 -0800 2024-01-10 22:20:12 -0800 Patch bug-267348-20240110221844.patch text/plain 2021 jhoffman23 RnJvbSA1MzA4N2VhM2FlNzA4OTkzZjAxM2EyNjExMGExODExMzBlNGMyZGRlIE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBKb3NodWEgSG9mZm1hbiA8amhvZmZtYW4yM0BhcHBsZS5jb20+ CkRhdGU6IFdlZCwgMTAgSmFuIDIwMjQgMjI6MDQ6MTEgLTA4MDAKU3ViamVjdDogW1BBVENIXSBB WDogQVNTRVJUSU9OIEZBSUxFRDogbV90YWJsZSBpbnNpZGUgb2YKIEFYSXNvbGF0ZWRUcmVlOjpj cmVhdGUgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTI2NzM0OAogcmRh cjovLzEyMDc4ODE0NQpNSU1FLVZlcnNpb246IDEuMApDb250ZW50LVR5cGU6IHRleHQvcGxhaW47 IGNoYXJzZXQ9VVRGLTgKQ29udGVudC1UcmFuc2Zlci1FbmNvZGluZzogOGJpdAoKUmV2aWV3ZWQg YnkgTk9CT0RZIChPT1BTISkuCgpXZSBjYW5ub3QgcmVseSBvbiB1c2luZyBheE9iamVjdENhY2hl LnJlbGF0aW9ucygpIHRvIGl0ZXJhdGUgdGhlIG5ldyByZWxhdGlvbnMgZHVyaW5nIHRoZSBjcmVh dGlvbiBvZiB0aGUKaXNvbGF0ZWQgdHJlZSwgYXMgdGhhdCBIYXNoTWFwIG1heSBiZSBtb2RpZmll ZCBvciBpbiB0aGUgcHJvY2VzcyBvZiBiZWluZyB1cGRhdGVkLiBUaGlzIHBhdGNoIHNhdmVzIHRo ZQpyZWxhdGlvbnMgZnJvbSB0aGUgY2FjaGUgdG8gYSBjb25zdCB2YXJpYWJsZSwgdG8gcHJldmVu dCB0aGlzIGlzc3VlLgoKVGhpcyBjaGFuZ2UgYWxzbyBwcmV2ZW50cyB0aGUgbmVlZCB0byBjYWxs IEFYT2JqZWN0Q2FjaGU6OnJlbGF0aW9ucygpIHR3aWNl4oCUd2hpY2ggY291bGQgc3Vic2VxdWVu dGx5IGNhbGwKdXBkYXRlUmVsYXRpb25zSWZOZWVkZWQoKSBkb3duc3RyZWFtLgoKKiBTb3VyY2Uv V2ViQ29yZS9hY2Nlc3NpYmlsaXR5L2lzb2xhdGVkdHJlZS9BWElzb2xhdGVkVHJlZS5jcHA6CihX ZWJDb3JlOjpBWElzb2xhdGVkVHJlZTo6Y3JlYXRlKToKLS0tCiBTb3VyY2UvV2ViQ29yZS9hY2Nl c3NpYmlsaXR5L2lzb2xhdGVkdHJlZS9BWElzb2xhdGVkVHJlZS5jcHAgfCA1ICsrKy0tCiAxIGZp bGUgY2hhbmdlZCwgMyBpbnNlcnRpb25zKCspLCAyIGRlbGV0aW9ucygtKQoKZGlmZiAtLWdpdCBh L1NvdXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNvbGF0ZWRUcmVl LmNwcCBiL1NvdXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNvbGF0 ZWRUcmVlLmNwcAppbmRleCBmNmYwNjIyYWJlOTMuLjE4YWM5NjNhYTRlMiAxMDA2NDQKLS0tIGEv U291cmNlL1dlYkNvcmUvYWNjZXNzaWJpbGl0eS9pc29sYXRlZHRyZWUvQVhJc29sYXRlZFRyZWUu Y3BwCisrKyBiL1NvdXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNv bGF0ZWRUcmVlLmNwcApAQCAtMTYwLDkgKzE2MCwxMCBAQCBSZWY8QVhJc29sYXRlZFRyZWU+IEFY SXNvbGF0ZWRUcmVlOjpjcmVhdGUoQVhPYmplY3RDYWNoZSYgYXhPYmplY3RDYWNoZSkKIAogICAg IHRyZWUtPnVwZGF0ZUxvYWRpbmdQcm9ncmVzcyhheE9iamVjdENhY2hlLmxvYWRpbmdQcm9ncmVz cygpKTsKIAotICAgIHRyZWUtPnVwZGF0ZVJlbGF0aW9ucyhheE9iamVjdENhY2hlLnJlbGF0aW9u cygpKTsKKyAgICBjb25zdCBhdXRvJiByZWxhdGlvbnMgPSBheE9iamVjdENhY2hlLnJlbGF0aW9u cygpOworICAgIHRyZWUtPnVwZGF0ZVJlbGF0aW9ucyhyZWxhdGlvbnMpOwogCi0gICAgZm9yIChh dXRvJiByZWxhdGVkT2JqZWN0SUQgOiBheE9iamVjdENhY2hlLnJlbGF0aW9ucygpLmtleXMoKSkg eworICAgIGZvciAoYXV0byYgcmVsYXRlZE9iamVjdElEIDogcmVsYXRpb25zLmtleXMoKSkgewog ICAgICAgICBSZWZQdHIgYXhPYmplY3QgPSBheE9iamVjdENhY2hlLm9iamVjdEZvcklEKHJlbGF0 ZWRPYmplY3RJRCk7CiAgICAgICAgIGlmIChheE9iamVjdCAmJiBheE9iamVjdC0+YWNjZXNzaWJp bGl0eUlzSWdub3JlZCgpKQogICAgICAgICAgICAgdHJlZS0+YWRkVW5jb25uZWN0ZWROb2RlKGF4 T2JqZWN0LnJlbGVhc2VOb25OdWxsKCkpOwo= 469370 2024-01-10 22:20:14 -0800 2024-01-10 23:00:25 -0800 Patch bug-267348-20240110222013.patch text/plain 2018 jhoffman23 RnJvbSAzYTA3MDE2Y2UzZTkyOTlkYmEwMGY1YmNhMGMzNWQzMGYwNWY3ODlkIE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBKb3NodWEgSG9mZm1hbiA8amhvZmZtYW4yM0BhcHBsZS5jb20+ CkRhdGU6IFdlZCwgMTAgSmFuIDIwMjQgMjI6MDQ6MTEgLTA4MDAKU3ViamVjdDogW1BBVENIXSBB WDogQVNTRVJUSU9OIEZBSUxFRDogbV90YWJsZSBpbnNpZGUgb2YKIEFYSXNvbGF0ZWRUcmVlOjpj cmVhdGUgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTI2NzM0OAogcmRh cjovLzEyMDc4ODE0NQpNSU1FLVZlcnNpb246IDEuMApDb250ZW50LVR5cGU6IHRleHQvcGxhaW47 IGNoYXJzZXQ9VVRGLTgKQ29udGVudC1UcmFuc2Zlci1FbmNvZGluZzogOGJpdAoKUmV2aWV3ZWQg YnkgTk9CT0RZIChPT1BTISkuCgpXZSBjYW5ub3QgcmVseSBvbiB1c2luZyBheE9iamVjdENhY2hl LnJlbGF0aW9ucygpIHRvIGl0ZXJhdGUgdGhlIG5ldyByZWxhdGlvbnMgZHVyaW5nIHRoZSBjcmVh dGlvbiBvZiB0aGUKaXNvbGF0ZWQgdHJlZSwgYXMgdGhhdCBIYXNoTWFwIG1heSBiZSBtb2RpZmll ZCBvciBpbiB0aGUgcHJvY2VzcyBvZiBiZWluZyB1cGRhdGVkLiBUaGlzIHBhdGNoIGNyZWF0ZXMg YQpjb25zdCByZWZlcmVuY2UgdmFyaWFibGUgdG8gdGhlIHJlbGF0aW9ucywgdG8gcHJldmVudCB0 aGlzIGlzc3VlLgoKVGhpcyBjaGFuZ2UgYWxzbyBwcmV2ZW50cyB0aGUgbmVlZCB0byBjYWxsIEFY T2JqZWN0Q2FjaGU6OnJlbGF0aW9ucygpIHR3aWNl4oCUd2hpY2ggY291bGQgc3Vic2VxdWVudGx5 IGNhbGwKdXBkYXRlUmVsYXRpb25zSWZOZWVkZWQoKSBkb3duc3RyZWFtLgoKKiBTb3VyY2UvV2Vi Q29yZS9hY2Nlc3NpYmlsaXR5L2lzb2xhdGVkdHJlZS9BWElzb2xhdGVkVHJlZS5jcHA6CihXZWJD b3JlOjpBWElzb2xhdGVkVHJlZTo6Y3JlYXRlKToKLS0tCiBTb3VyY2UvV2ViQ29yZS9hY2Nlc3Np YmlsaXR5L2lzb2xhdGVkdHJlZS9BWElzb2xhdGVkVHJlZS5jcHAgfCA1ICsrKy0tCiAxIGZpbGUg Y2hhbmdlZCwgMyBpbnNlcnRpb25zKCspLCAyIGRlbGV0aW9ucygtKQoKZGlmZiAtLWdpdCBhL1Nv dXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNvbGF0ZWRUcmVlLmNw cCBiL1NvdXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNvbGF0ZWRU cmVlLmNwcAppbmRleCBmNmYwNjIyYWJlOTMuLjE4YWM5NjNhYTRlMiAxMDA2NDQKLS0tIGEvU291 cmNlL1dlYkNvcmUvYWNjZXNzaWJpbGl0eS9pc29sYXRlZHRyZWUvQVhJc29sYXRlZFRyZWUuY3Bw CisrKyBiL1NvdXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNvbGF0 ZWRUcmVlLmNwcApAQCAtMTYwLDkgKzE2MCwxMCBAQCBSZWY8QVhJc29sYXRlZFRyZWU+IEFYSXNv bGF0ZWRUcmVlOjpjcmVhdGUoQVhPYmplY3RDYWNoZSYgYXhPYmplY3RDYWNoZSkKIAogICAgIHRy ZWUtPnVwZGF0ZUxvYWRpbmdQcm9ncmVzcyhheE9iamVjdENhY2hlLmxvYWRpbmdQcm9ncmVzcygp KTsKIAotICAgIHRyZWUtPnVwZGF0ZVJlbGF0aW9ucyhheE9iamVjdENhY2hlLnJlbGF0aW9ucygp KTsKKyAgICBjb25zdCBhdXRvJiByZWxhdGlvbnMgPSBheE9iamVjdENhY2hlLnJlbGF0aW9ucygp OworICAgIHRyZWUtPnVwZGF0ZVJlbGF0aW9ucyhyZWxhdGlvbnMpOwogCi0gICAgZm9yIChhdXRv JiByZWxhdGVkT2JqZWN0SUQgOiBheE9iamVjdENhY2hlLnJlbGF0aW9ucygpLmtleXMoKSkgewor ICAgIGZvciAoYXV0byYgcmVsYXRlZE9iamVjdElEIDogcmVsYXRpb25zLmtleXMoKSkgewogICAg ICAgICBSZWZQdHIgYXhPYmplY3QgPSBheE9iamVjdENhY2hlLm9iamVjdEZvcklEKHJlbGF0ZWRP YmplY3RJRCk7CiAgICAgICAgIGlmIChheE9iamVjdCAmJiBheE9iamVjdC0+YWNjZXNzaWJpbGl0 eUlzSWdub3JlZCgpKQogICAgICAgICAgICAgdHJlZS0+YWRkVW5jb25uZWN0ZWROb2RlKGF4T2Jq ZWN0LnJlbGVhc2VOb25OdWxsKCkpOwo= 469372 2024-01-10 23:00:27 -0800 2024-01-11 21:45:22 -0800 Patch bug-267348-20240110230027.patch text/plain 2007 jhoffman23 RnJvbSA0ZjY0YTc1MWY5ZWQ3YmNlYmRhMzhmYTU3YTZlMjJjYzk5NzIwODY3IE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBKb3NodWEgSG9mZm1hbiA8amhvZmZtYW4yM0BhcHBsZS5jb20+ CkRhdGU6IFdlZCwgMTAgSmFuIDIwMjQgMjI6MDQ6MTEgLTA4MDAKU3ViamVjdDogW1BBVENIXSBB WDogQVNTRVJUSU9OIEZBSUxFRDogbV90YWJsZSBpbnNpZGUgb2YKIEFYSXNvbGF0ZWRUcmVlOjpj cmVhdGUgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTI2NzM0OAogcmRh cjovLzEyMDc4ODE0NQpNSU1FLVZlcnNpb246IDEuMApDb250ZW50LVR5cGU6IHRleHQvcGxhaW47 IGNoYXJzZXQ9VVRGLTgKQ29udGVudC1UcmFuc2Zlci1FbmNvZGluZzogOGJpdAoKUmV2aWV3ZWQg YnkgTk9CT0RZIChPT1BTISkuCgpXZSBjYW5ub3QgcmVseSBvbiB1c2luZyBheE9iamVjdENhY2hl LnJlbGF0aW9ucygpIHRvIGl0ZXJhdGUgdGhlIG5ldyByZWxhdGlvbnMgZHVyaW5nIHRoZSBjcmVh dGlvbiBvZiB0aGUKaXNvbGF0ZWQgdHJlZSwgYXMgdGhhdCBIYXNoTWFwIG1heSBiZSBtb2RpZmll ZCBvciBpbiB0aGUgcHJvY2VzcyBvZiBiZWluZyB1cGRhdGVkLiBUaGlzIHBhdGNoIGNyZWF0ZXMg YQpjb3B5IG9mIHRoZXNlIHJlbGF0aW9ucyBpbnN0ZWFkLCB0byBwcmV2ZW50IHRoaXMgaXNzdWUu CgpUaGlzIGNoYW5nZSBhbHNvIHByZXZlbnRzIHRoZSBuZWVkIHRvIGNhbGwgQVhPYmplY3RDYWNo ZTo6cmVsYXRpb25zKCkgdHdpY2XigJR3aGljaCBjb3VsZCBzdWJzZXF1ZW50bHkgY2FsbAp1cGRh dGVSZWxhdGlvbnNJZk5lZWRlZCgpIGRvd25zdHJlYW0uCgoqIFNvdXJjZS9XZWJDb3JlL2FjY2Vz c2liaWxpdHkvaXNvbGF0ZWR0cmVlL0FYSXNvbGF0ZWRUcmVlLmNwcDoKKFdlYkNvcmU6OkFYSXNv bGF0ZWRUcmVlOjpjcmVhdGUpOgotLS0KIFNvdXJjZS9XZWJDb3JlL2FjY2Vzc2liaWxpdHkvaXNv bGF0ZWR0cmVlL0FYSXNvbGF0ZWRUcmVlLmNwcCB8IDUgKysrLS0KIDEgZmlsZSBjaGFuZ2VkLCAz IGluc2VydGlvbnMoKyksIDIgZGVsZXRpb25zKC0pCgpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYkNv cmUvYWNjZXNzaWJpbGl0eS9pc29sYXRlZHRyZWUvQVhJc29sYXRlZFRyZWUuY3BwIGIvU291cmNl L1dlYkNvcmUvYWNjZXNzaWJpbGl0eS9pc29sYXRlZHRyZWUvQVhJc29sYXRlZFRyZWUuY3BwCmlu ZGV4IGY2ZjA2MjJhYmU5My4uYzAwM2FlNzQ2MmE5IDEwMDY0NAotLS0gYS9Tb3VyY2UvV2ViQ29y ZS9hY2Nlc3NpYmlsaXR5L2lzb2xhdGVkdHJlZS9BWElzb2xhdGVkVHJlZS5jcHAKKysrIGIvU291 cmNlL1dlYkNvcmUvYWNjZXNzaWJpbGl0eS9pc29sYXRlZHRyZWUvQVhJc29sYXRlZFRyZWUuY3Bw CkBAIC0xNjAsOSArMTYwLDEwIEBAIFJlZjxBWElzb2xhdGVkVHJlZT4gQVhJc29sYXRlZFRyZWU6 OmNyZWF0ZShBWE9iamVjdENhY2hlJiBheE9iamVjdENhY2hlKQogCiAgICAgdHJlZS0+dXBkYXRl TG9hZGluZ1Byb2dyZXNzKGF4T2JqZWN0Q2FjaGUubG9hZGluZ1Byb2dyZXNzKCkpOwogCi0gICAg dHJlZS0+dXBkYXRlUmVsYXRpb25zKGF4T2JqZWN0Q2FjaGUucmVsYXRpb25zKCkpOworICAgIGNv bnN0IGF1dG8gcmVsYXRpb25zID0gYXhPYmplY3RDYWNoZS5yZWxhdGlvbnMoKTsKKyAgICB0cmVl LT51cGRhdGVSZWxhdGlvbnMocmVsYXRpb25zKTsKIAotICAgIGZvciAoYXV0byYgcmVsYXRlZE9i amVjdElEIDogYXhPYmplY3RDYWNoZS5yZWxhdGlvbnMoKS5rZXlzKCkpIHsKKyAgICBmb3IgKGF1 dG8mIHJlbGF0ZWRPYmplY3RJRCA6IHJlbGF0aW9ucy5rZXlzKCkpIHsKICAgICAgICAgUmVmUHRy IGF4T2JqZWN0ID0gYXhPYmplY3RDYWNoZS5vYmplY3RGb3JJRChyZWxhdGVkT2JqZWN0SUQpOwog ICAgICAgICBpZiAoYXhPYmplY3QgJiYgYXhPYmplY3QtPmFjY2Vzc2liaWxpdHlJc0lnbm9yZWQo KSkKICAgICAgICAgICAgIHRyZWUtPmFkZFVuY29ubmVjdGVkTm9kZShheE9iamVjdC5yZWxlYXNl Tm9uTnVsbCgpKTsK