<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[Unreal Engine Issues - Latest Fixes]]></title><description><![CDATA[Latest Bug Fixes For Unreal Engine.]]></description><link>https://issues.unrealengine.com</link><image><url>https://issues.unrealengine.com/favicon.png</url><title>Unreal Engine Issues - Latest Fixes</title><link>https://issues.unrealengine.com</link></image><generator>RSS for Node</generator><lastBuildDate>Sun, 22 Nov 2026 20:50:15 GMT</lastBuildDate><atom:link href="https://issues.unrealengine.com/fixes" rel="self" type="application/rss+xml"/><item><title><![CDATA[UIScene support to 4.27+]]></title><link>https://issues.unrealengine.com/issue/UE-405329</link><guid isPermaLink="false">UE-405329</guid><pubDate>Fri, 20 Nov 2026 15:02:06 GMT</pubDate></item><item><title><![CDATA[Export: TXT (just textures) export constructs the whole exportable mesh]]></title><description><![CDATA[<p>This has been there since the start. Because of the exporter design the textures are coupled to the mesh geometry so in order to export them the whole mesh is first prepared and then thrown away. </p>]]></description><link>https://issues.unrealengine.com/issue/RC-3744</link><guid isPermaLink="false">RC-3744</guid><pubDate>Thu, 19 Nov 2026 09:42:31 GMT</pubDate></item><item><title><![CDATA[Foot Placement: Division by 0]]></title><description><![CDATA[<p> </p>
<div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent">
<pre class="code-java">
<span class="code-keyword">static</span> FVector PointDirectionPlaneIntersection(<span class="code-keyword">const</span> FVector Point, <span class="code-keyword">const</span> FVector Direction, <span class="code-keyword">const</span> FPlane Plane)
{
   <span class="code-keyword">return</span> Point + Direction * ((Plane.W - (Point | Plane)) / (Direction | Plane)); 
}; </pre>
</div></div>
<p>Direction | Plane can result in a division by 0. Unsure in whch case.</p>

<p> </p>]]></description><link>https://issues.unrealengine.com/issue/UE-241483</link><guid isPermaLink="false">UE-241483</guid><pubDate>Tue, 17 Nov 2026 22:10:15 GMT</pubDate></item><item><title><![CDATA[Crash in FMovieSceneNumericVariant::Set when loading a Level Sequence with Custom Time Warp on Android]]></title><description><![CDATA[

<p>On Android, loading a Level Sequence fatals in the async loading thread:</p>

<p>Assertion failed: (NewValue &amp; ~PAYLOAD_Bits) == 0</p>

<p>MovieSceneNumericVariant.cpp:133</p>

<p>Unable to store a pointer outside of a 48 bit address space in this container</p>

<p>This fires while serializing a <tt>UMovieSceneSection</tt> whose Time Warp type is Custom (Play Rate Curve / Time Warp Curve). Windows does not hit this.</p>
]]></description><link>https://issues.unrealengine.com/issue/UE-392043</link><guid isPermaLink="false">UE-392043</guid><pubDate>Tue, 17 Nov 2026 02:56:15 GMT</pubDate></item><item><title><![CDATA[UCharacterMovementComponent::CapsuleTouched crashes on UGeometryCollectionComponent overlap (unchecked USkinnedMeshComponent cast)]]></title><description><![CDATA[<h2><a name="Summary%3A"></a>Summary:</h2>

<p>Confirmed that this exists in latest as well as 5.8 which is the origin or both tickets. This is a regression from 5.7 - although it looks like the guard was not aiming specifically to catch that.</p>

<ul>
	<li>5.6 / 5.7 — a plain two-line accessor. It returns the reinterpreted bytes, and GetBoneName null-checks the result (SkinnedMeshComponent.cpp:2241), so garbage-reading-null yields NAME_None. The bug is exercised on every overlap and stays latent.</li>
</ul>


<ul>
	<li>5.8 / FN Main — gained a KnownSkinnedAsset cache-validation block under #if WITH_EDITOR. When the two disagree it calls OutdatedKnownSkinnedAssetDetected(), whose ensureMsgf message arguments dereference the garbage via GetFullName(). That<br>
dereference has no null-check equivalent, so the latent bad cast becomes a hard crash.</li>
</ul>


<p><tt>UCharacterMovementComponent::CapsuleTouched</tt> casts the overlapped component to <tt>USkinnedMeshComponent*</tt> with a C-style cast, guarded only by <tt>OtherBodyIndex != INDEX_NONE</tt>.</p>

<p><tt>Engine/Source/Runtime/Engine/Private/Components/CharacterMovementComponent.cpp:11693</tt></p>

<div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent">
<pre class="code-cpp">if (OtherComp != <span class="code-keyword">NULL</span> &amp;&amp; OtherComp-&gt;IsAnySimulatingPhysics())
{
    ...
    FName BoneName = NAME_None;
    if (OtherBodyIndex != INDEX_NONE)
    {
        BoneName = ((USkinnedMeshComponent*)OtherComp)-&gt;GetBoneName(OtherBodyIndex);   <span class="code-comment">// line 11713
</span>    }</pre>
</div></div>

<p>A <tt>UGeometryCollectionComponent</tt> is not a <tt>USkinnedMeshComponent</tt>, but it satisfies both guards: it reports <tt>IsAnySimulatingPhysics() == true</tt> while simulating, and it delivers a per-particle <tt>OtherBodyIndex &gt;= 0</tt> rather than <tt>INDEX_NONE</tt>. The cast is therefore taken on an unrelated type and <tt>GetBoneName</tt> reads through a garbage <tt>SkinnedAsset</tt> pointer.</p>

<p><tt>OtherBodyIndex</tt> is only meaningful as a bone index for skinned meshes. For a geometry collection it is a particle/transform index and should not be passed to <tt>GetBoneName</tt> at all.</p>

<h2><a name="Callstack%3A"></a>Call stack:</h2>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>Unhandled Exception: EXCEPTION_ACCESS_VIOLATION reading address 0x0000000000000012

UCharacterMovementComponent::CapsuleTouched()               CharacterMovementComponent.cpp:11713
USkinnedMeshComponent::GetBoneName()                        SkinnedMeshComponent.cpp:2542
USkinnedMeshComponent::GetSkinnedAsset()                    SkinnedMeshComponent.cpp:2882
USkinnedMeshComponent::OutdatedKnownSkinnedAssetDetected()  SkinnedMeshComponent.cpp:6119
UObjectBaseUtility::GetFullName()                           UObjectBaseUtility.cpp:117</pre>
</div></div>

<p>Reached through the ordinary movement path:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>UCharacterMovementComponent::TickComponent()
UCharacterMovementComponent::ControlledCharacterMove()
UCharacterMovementComponent::PerformMovement()
FScopedMovementUpdate::~FScopedMovementUpdate()
USceneComponent::UpdateOverlaps()
UPrimitiveComponent::BeginComponentOverlap()
UCharacterMovementComponent::CapsuleTouched()</pre>
</div></div>]]></description><link>https://issues.unrealengine.com/issue/UE-392989</link><guid isPermaLink="false">UE-392989</guid><pubDate>Fri, 13 Nov 2026 16:44:40 GMT</pubDate></item><item><title><![CDATA[Crash when merging a file containing an Orbit Rig]]></title><description><![CDATA[<p>Reported in SF [Link Removed]</p>

<p>Scenes containing Orbit Rigs will cause crashes when attempting to Merge with other files.</p>]]></description><link>https://issues.unrealengine.com/issue/TM-23791</link><guid isPermaLink="false">TM-23791</guid><pubDate>Thu, 12 Nov 2026 19:07:16 GMT</pubDate></item><item><title><![CDATA[Meta “AllowClasses” doesn’t filter in TArray using class]]></title><description><![CDATA[<p>This occurred when we use a class and variable TSubclassOf&lt;UObject&gt;, TObjectPtr&lt;UClass&gt;, TSoftClassPtr&lt;UObject&gt;.</p>

<p><span class="image-wrap" style="">[Image Removed]</span></p>

<p> </p>]]></description><link>https://issues.unrealengine.com/issue/UE-186119</link><guid isPermaLink="false">UE-186119</guid><pubDate>Thu, 12 Nov 2026 00:34:15 GMT</pubDate></item><item><title><![CDATA[TitleProperty does not work with VisibleAnywhere properties]]></title><description><![CDATA[<p>In:</p>
<div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent">
<pre class="code-java">
[Editor\PropertyEditor\Private\PropertyEditorHelpers.cpp:118]</pre>
</div></div>
<p>function SPropertyValueWidget::ConstructPropertyEditorWidget checks what type of widget to create.</p>

<p> </p>

<p>In:</p>
<div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent">
<pre class="code-java">
[Editor\PropertyEditor\Private\UserInterface\PropertyEditor\SPropertyEditorArrayItem.cpp:125]</pre>
</div></div>
<p>function SPropertyEditorArrayItem::Supports() fails the test </p>
<div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent">
<pre class="code-java">
<span class="code-keyword">if</span> (!(Property-&gt;GetOwner&lt;FArrayProperty&gt;()-&gt;PropertyFlags &amp; CPF_EditConst)) </pre>
</div></div>
<p> </p>

<p>Interestingly, SpropertyEditorArrayItem seems to be prepared to handle non-editable properties, as seen in function SPropertyEditorArrayItem::CanEdit().</p>

<p>But, for properties marked as CPF_EditConst, what ends up being used is a basic SpropertyEditor</p>

<p>Unlike SpropertyEditorArrayItem, the basic SpropertyEditor does not handle the TitleProperty annotation.</p>]]></description><link>https://issues.unrealengine.com/issue/UE-193027</link><guid isPermaLink="false">UE-193027</guid><pubDate>Tue, 10 Nov 2026 22:19:57 GMT</pubDate></item><item><title><![CDATA[Widget animations queued before Slate creation are not carried out in UE5.8]]></title><description><![CDATA[<p>We are currently upgrading some of our systems from Unreal 5.6 to Unreal 5.8, and we've noticed an issue with Widget animations. In one of our commonly used actors, a Widget animation is queued in the actor's BeginPlay using QueuePlayAnimation (representing a monitor being switched on). In Unreal 5.6, once the Slate for this widget has been constructed, the animation plays out, and the actor behaves correctly. Unfortunately under Unreal 5.8 this does not happen - the queued animation is being dropped and never plays, meaning that the actor behaves incorrectly.</p>]]></description><link>https://issues.unrealengine.com/issue/UE-398754</link><guid isPermaLink="false">UE-398754</guid><pubDate>Tue, 10 Nov 2026 15:54:03 GMT</pubDate></item><item><title><![CDATA[Resolve Referencing Properties not working on sub-objects]]></title><description><![CDATA[

<p>Property resolution doesn't work in many many cases. Only works with dependencies with properties on the highest level of the asset (TObject/TSoftObject as a variable)</p>

<p>After some debugging, it appears that the property resolution is not recursive and cannot dig into instanced objects (may even not work with structures, unsure about that)</p>

<p>This is very annoying because the case where we'd need that feature the most is for nested properties hard to find</p>
]]></description><link>https://issues.unrealengine.com/issue/UE-400621</link><guid isPermaLink="false">UE-400621</guid><pubDate>Tue, 10 Nov 2026 12:08:51 GMT</pubDate></item><item><title><![CDATA[Shadows become disabled for Datasmith imported lights on resync]]></title><description><![CDATA[<p>Reported in SF [Link Removed]</p>

<p>For Datasmith models containing lights, resynchronizing the model after any changes will cause any lights that had shadows enabled prior to the sync to have that setting reverted to Disabled.</p>

<p>This is a regression and did not occur in Twinmotion 2025.2.</p>]]></description><link>https://issues.unrealengine.com/issue/TM-23756</link><guid isPermaLink="false">TM-23756</guid><pubDate>Mon, 09 Nov 2026 13:56:29 GMT</pubDate></item><item><title><![CDATA[Missing textures on FBX import]]></title><description><![CDATA[<p>Reported in SF [Link Removed]</p>

<p>When importing some FBX files, the materials may be missing textures when using the Interchange importer. The attached model was created in Bentley Microstation as a .dgn file and exported as an FBX file. When importing into TM 2026.1 or 2026.2 Preview (Interchange), the model will appear without the associated texture.</p>

<p>This issue does not occur using the legacy importer, and does not occur when using prior versions of Twinmotion, such as 2025.2.</p>

<p>This issue does not affect all FBX files.</p>]]></description><link>https://issues.unrealengine.com/issue/TM-23804</link><guid isPermaLink="false">TM-23804</guid><pubDate>Mon, 09 Nov 2026 13:49:15 GMT</pubDate></item><item><title><![CDATA[CA_VALID_POINTER and CA_CHECK_RETVAL static analysis annotations broken]]></title><description><![CDATA[<blockquote><p>We had some code using the CA_VALID_POINTER and CA_CHECK_RETVAL annotations which had been working fine up until the 5.3 engine release. In 5.3, we started getting the following compile errors in a /analyze (-staticanalyzer=VisualCpp) build:</p></blockquote>
<blockquote><p> </p>

<p>error C2337: 'Pre': attribute not found</p>

<p>error C2337: 'Post': attribute not found</p>

<p> </p>

<p>I'm assuming that CA_READ_ONLY and CA_WRITE_ONLY would have similar issues, but we aren't using those just yet. I've made our own defines to replace these for now:</p>

<p> </p>

<p>#define CA_SK_VALID_POINTER <em>In</em></p>

<p>#define CA_SK_CHECK_RETVAL <em>Must_inspect_result</em></p>

<p> </p>

<p>though I'm not 100% certain those are equivalent to the previous annotations. At any rate, from the documentation I was able to find, it looks like the <span class="error">&#91;Pre&#93;</span> and <span class="error">&#91;Post&#93;</span> stuff has been deprecated by Microsoft, so I'm assuming those existing SAL defines need to be changed.</p></blockquote>]]></description><link>https://issues.unrealengine.com/issue/UE-197962</link><guid isPermaLink="false">UE-197962</guid><pubDate>Mon, 09 Nov 2026 10:07:31 GMT</pubDate></item><item><title><![CDATA[delete_all_material_expressions_in_function Doesn't delete all expressions in one go.]]></title><description><![CDATA[<p>UDN: <a href="https://udn.unrealengine.com/s/case/500QP000009NxNBYA0/deleteallmaterialexpressionsinfunction-doesnt-delete-all-expressions-in-one-go?fromCase=1" class="external-link protected" rel="nofollow noreferrer">https://udn.unrealengine.com/s/case/500QP000009NxNBYA0/deleteallmaterialexpressionsinfunction-doesnt-delete-all-expressions-in-one-go?fromCase=1</a></p>

<p>A UDN user is reporting that the delete_all_material_expressions_in_function Python API does not delete all expressions in one go, but some expressions are left over. Calling the API repeatedly will eventually delete all nodes.</p>

<p>The user also found out that changing the delete_all_material_expressions_in_function  implementation to iterate through the array of function expressions backwards solves the issue.</p>

<p><b>Evidence</b></p>

<ul>
	<li>Flux [Link Removed] J-001: 2026-10-27 (engine <tt>6.0.0-59684868</tt>), <tt>M_FX_ExitHologram</tt> had 218 expressions, 108 after <tt>delete_all_material_expressions</tt>; deleting each from <tt>get_material_expressions</tt> left 0 (rebuild 145).</li>
</ul>
]]></description><link>https://issues.unrealengine.com/issue/UE-215725</link><guid isPermaLink="false">UE-215725</guid><pubDate>Tue, 03 Nov 2026 22:32:01 GMT</pubDate></item><item><title><![CDATA[No Interface to Assign Modular Control Rigs in Sequencer with Blueprint]]></title><description><![CDATA[<p>To procedurally spawn ControlRigs in the sequencer through custom commands (needed to populate cinematic scenes automatically) we are using the interface provided by Editor Function Libraries, specifically <b>Find or Create Control Rig Track</b> and <b>Expose Control Rig Components</b> nodes requires a ControlRigClass to instantiate with each ControlRig being its own class child of Control Rig. below:</p>




<p>But to instantiate Modular Control Rigs. Modular Control Rigs are all of type ControlRigRuntimeAsset, they are pure data payload, so it is not possible to cast them to a ControlRig class and pass them to this node.</p>




<p>There is no alternative interface provided to be able to do this with Modular Control Rig; not in Blueprint, not in Python.</p>]]></description><link>https://issues.unrealengine.com/issue/UE-400417</link><guid isPermaLink="false">UE-400417</guid><pubDate>Tue, 03 Nov 2026 16:24:14 GMT</pubDate></item><item><title><![CDATA[The speed of Animators is increased in exports]]></title><description><![CDATA[<p>Reported in SF [Link Removed]</p>

<p>When setting Animator tools to use Collision mode for video, the loop speed of the animator will be twice as fast as specified in the video export.</p>

<p>Attached a sample scene that demonstrates this behavior.</p>

<p>Unclear if this is a regression from 2026.1 to 2026.1 HF2. Last confirmed working correctly in 2025.2.</p>]]></description><link>https://issues.unrealengine.com/issue/TM-23748</link><guid isPermaLink="false">TM-23748</guid><pubDate>Tue, 03 Nov 2026 13:55:24 GMT</pubDate></item><item><title><![CDATA[Multiple setter functions in UMaterialEditingLibrary always return false even when the parameter is correctly set on the MaterialInstance.]]></title><description><![CDATA[<p>In Unreal Engine, materials can expose parameters that can then be modified via instances (MaterialInstanceConstant). The UMaterialEditingLibrary provides several functions to modify these parameters at runtime or through editor scripts.</p>

<p>However, these setter functions (e.g., SetMaterialInstanceTextureParameterValue) currently always return false, even when the parameter update is successful. This may mislead users into thinking the operation failed.</p>

<p>A user initially reported this behavior for one function, but upon further inspection, this issue affects multiple setter functions across UMaterialEditingLibrary.</p>

<p>It would be helpful to review and correct the return logic for these functions to reflect whether the parameter update was successfully applied, and optionally provide more diagnostic output in case of failure.</p>

<p>This also do happen for the following functions:</p>
<ul class="alternate" type="square">
	<li>SetMaterialInstanceScalarParameterValue</li>
	<li>SetMaterialInstanceTextureParameterValue</li>
	<li>SetMaterialInstanceRuntimeVirtualTextureParameterValue</li>
	<li>SetMaterialInstanceSparseVolumeTextureParameterValue</li>
	<li>SetMaterialInstanceVectorParameterValue</li>
	<li>SetMaterialInstanceStaticSwitchParameterValue</li>
	<li>SetMaterialInstanceEnumerationParameterValue</li>
</ul>
]]></description><link>https://issues.unrealengine.com/issue/UE-291403</link><guid isPermaLink="false">UE-291403</guid><pubDate>Fri, 30 Oct 2026 21:50:55 GMT</pubDate></item><item><title><![CDATA[MRQ render with a slow output format drives Niagara GPU tick flush into its dummy-view path, causing an access violation in SetShaderParameters]]></title><description><![CDATA[<p><b>Sorry for the AI ticket, but it did a better job than I could:</b><br>
Rendering a level sequence through Movie Render Queue crashes the editor with an <tt>EXCEPTION_ACCESS_VIOLATION</tt> when the scene contains a Niagara system using a <b>GPU simulation target</b> and the MRQ output format is <b>EXR or PNG</b>. The same render completes successfully with <b>JPG</b> output.</p>

<p>The crash is not in the image writer. It occurs in <tt>FNiagaraGpuComputeDispatch::ProcessPendingTicksFlush()</tt> — a safety-valve path that runs queued GPU simulation ticks against a <b>fabricated dummy view outside the scene renderer</b>. A Data Interface supplies a resource pointer that is not valid in that context; the D3D12 backend then makes a virtual call on it and jumps to address <tt>0x1</tt>.</p>

<p>The output format matters only because of <b>write latency</b>. EXR (float, ~16 MB/frame) and 16-bit PNG stall the pipeline long enough for world ticks to accumulate without a corresponding scene render, crossing the tick-flush threshold. JPG finalizes fast enough to stay under it.</p>

<hr>

<h2><a name="Environment"></a>Environment</h2>

<div class='table-wrap'>
<table class='confluenceTable'><tbody>
<tr>
<th class='confluenceTh'>&nbsp;</th>
<th class='confluenceTh'>&nbsp;</th>
</tr>
<tr>
<td class='confluenceTd'>Engine</td>
<td class='confluenceTd'>5.8.1-56057345 <tt>+++UE5+Release-5.8</tt>, Development Editor</td>
</tr>
<tr>
<td class='confluenceTd'>OS</td>
<td class='confluenceTd'>Windows 11 24H2 — 10.0.26100.9106</td>
</tr>
<tr>
<td class='confluenceTd'>GPU</td>
<td class='confluenceTd'>NVIDIA RTX A6000, driver 596.72 (2026-06-10), D3D12 / SM6</td>
</tr>
<tr>
<td class='confluenceTd'>CPU / RAM</td>
<td class='confluenceTd'>AMD Ryzen Threadripper PRO 3995WX (64-core) / 256 GB</td>
</tr>
<tr>
<td class='confluenceTd'>Project</td>
<td class='confluenceTd'><tt>Flossing</tt> — level <tt>/Game/Test</tt>, sequence <tt>/Game/NewLevelSequence</tt></td>
</tr>
<tr>
<td class='confluenceTd'>Niagara system</td>
<td class='confluenceTd'><tt>/Game/EasyRain/System/Niagara/N_EasyRain_Niagara</tt> (GPU sim target)</td>
</tr>
<tr>
<td class='confluenceTd'>MRQ</td>
<td class='confluenceTd'>MovieGraph pipeline, Deferred renderer, <tt>beauty</tt> layer, 1920×1080, 1 camera, TileCount <span class="error">&#91;1,1&#93;</span></td>
</tr>
</tbody></table>
</div>


<p>Note: MRQ logs <tt>Using the default MovieGraphPipeline; the pipeline specified in Project Settings (MoviePipeline) is not compatible with graphs</tt> — the project-configured pipeline class is being silently ignored. Probably unrelated, but worth a look.</p>

<hr>

<h2><a name="Exception"></a>Exception</h2>

<p>From the minidump exception stream:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>ExceptionCode    : 0xC0000005   EXCEPTION_ACCESS_VIOLATION
ExceptionAddress : 0x0000000000000001
Info[0]          : 8            -&gt; EXECUTE (DEP)
Info[1]          : 0x0000000000000001</pre>
</div></div>

<p><tt>AccessType 8</tt> is the important detail — this is an <b>executed jump to</b> <tt>0x1</tt>, not a bad read. The address lies in no loaded module. That is an indirect call through a corrupt vtable pointer, i.e. a virtual call on a freed or never-initialised object.</p>

<p>RHI breadcrumbs (both configurations):</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>Breadcrumbs 'RDGExecute_RenderThread'
 - NiagaraGpuComputeDispatch
 - RenderGraphExecute
 - WorldTick</pre>
</div></div>

<p>Nsight Aftermath was enabled (<tt>RHI.Aftermath true</tt>) and produced <b>no</b> GPU crash dump, confirming a CPU-side fault rather than a GPU fault or TDR. <tt>MemoryStats.bIsOOM = 0</tt>. Driver not denylisted.</p>

<hr>

<h2><a name="Callstack"></a>Callstack</h2>

<p>Default configuration crashes on <b>RHIThread</b>, which loses the recording context and makes the bug very hard to attribute — the only Niagara frame is a template instantiated in the Niagara module:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>D3D12RHI!SetShaderParametersOnContext()                        D3D12Commands.cpp:698
D3D12RHI!FD3D12CommandContext::RHISetShaderParameters()        D3D12Commands.cpp:759
RHI!FRHICommandSetShaderParameters&lt;FRHIComputeShader&gt;::Execute()
Niagara!FRHICommand&lt;...SetShaderParameters&lt;FRHIComputeShader&gt;...&gt;::ExecuteAndDestruct()
RHI!FRHICommandListBase::Execute()
RHI!FRHICommandListExecutor::FTranslateState::Translate_ExecuteCommandList()
RenderCore!FRHIThread::Run()</pre>
</div></div>

<p>Re-running with <tt>-norhithread -forcerhibypass</tt> moves the fault to <b>RenderThread</b> and exposes the recording site:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>D3D12RHI!SetShaderParametersOnContext()                             D3D12Commands.cpp:698
D3D12RHI!FD3D12CommandContext::RHISetShaderParameters()             D3D12Commands.cpp:759
Niagara!FRHIComputeCommandList::SetBatchedShaderParameters()        RHICommandList.h:2631
Niagara!FNiagaraGpuComputeDispatch::DispatchStage'::&lt;lambda_2&gt;()    NiagaraGpuComputeDispatch.cpp:1871
Niagara!TRDGLambdaPass&lt;...&gt;::Execute()                              RenderGraphPass.h:704
RenderCore!FRDGBuilder::ExecutePass()                               RenderGraphBuilder.cpp:3525
RenderCore!FRDGBuilder::ExecuteSerialPass()                         RenderGraphBuilder.cpp:3549
RenderCore!FRDGBuilder::Execute()                                   RenderGraphBuilder.cpp:2157
Niagara!FNiagaraGpuComputeDispatch::ProcessPendingTicksFlush()      NiagaraGpuComputeDispatch.cpp:521
Niagara!FNiagaraGpuComputeDispatch::Tick'::&lt;lambda_1&gt;
RenderCore!ExecuteCommand()                                         RenderingThread.cpp:1533
RenderCore!RenderingThreadMain()                                    RenderingThread.cpp:261</pre>
</div></div>

<p>Line attribution caveat: the lambda frame reports <tt>:1871</tt> (<tt>DispatchComputeShader</tt>), but under bypass the preceding <tt>RHICmdList.SetBatchedShaderParameters(...)</tt> at <tt>:1869</tt> is inlined into the same frame. The D3D12 frames confirm the fault is in the parameter set, not the dispatch.</p>

<hr>

<h2><a name="Rootcause"></a>Root cause</h2>

<p><tt>FNiagaraGpuComputeDispatch::ProcessPendingTicksFlush()</tt> — <tt>NiagaraGpuComputeDispatch.cpp:385</tt></p>

<p>When GPU ticks queue up without being consumed by a scene render, <tt>Tick</tt> fires a flush whose behaviour is selected by <tt>fx.Niagara.Batcher.TickFlush.Mode</tt> (default <b>1</b>). The CVar's own help text documents the hazard:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>fx.Niagara.Batcher.TickFlush.Mode   (default 1)
  0 = Keep ticks queued, can result in a long pause when gaining focus again.
  1 = (Default) Process all queued ticks with dummy view / buffer data,
      may result in incorrect simulation due to missing depth collisions, etc.
  2 = Kill all pending ticks, may result in incorrect simulation due to
      missing frames of data, i.e. a particle reset.</pre>
</div></div>

<p>Mode 1 fabricates a view with a <b>null render target</b> (<tt>:474</tt>):</p>

<p>cpp</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>FSceneViewFamilyContext ViewFamily(
    FSceneViewFamily::ConstructionValues(nullptr, GetSceneInterface(), FEngineShowFlags(ESFIM_Game))
    .SetTime(CachedViewInitOptions.GameTime));
...
GetRendererModule().CreateAndInitSingleView(RHICmdList, &amp;ViewFamily, &amp;ViewInitOptions);
TConstStridedView&lt;FSceneView&gt; DummyViews = MakeStridedView&lt;const FSceneView&gt;(0, ViewFamily.Views[0], 1);

bIsOutsideSceneRenderer = true;   // "Allow downstream logic to detect we are
                                  //  running pending ticks outside the scene renderer"</pre>
</div></div>

<p>It then drives the full <tt>PreInitViews</tt> / <tt>PostInitViews</tt> / <tt>PostRenderOpaque</tt> chain against that dummy view. <tt>DispatchStage</tt> populates the batched parameters via <tt>SetShaderParameters(...)</tt> (<tt>:1866</tt>), which walks the shader parameter metadata including every Data Interface binding, and submits them at <tt>:1869</tt>. One of those bindings is invalid in the dummy-view context, and <tt>SetShaderParametersOnContext</tt> dereferences it.</p>

<p><b>The core defect:</b> <tt>bIsOutsideSceneRenderer</tt> is only consulted in <b>two places in the entire file</b> (<tt>:1981</tt> and <tt>:2143</tt>). It is never plumbed through to Data Interfaces, so a DI that binds a scene texture, depth buffer, or distance field has no way to detect the dummy-view context and bail out. The comment on <tt>:496</tt> states the flag exists precisely so "downstream logic" can detect this — but downstream DI parameter population does not check it. The result is a corrupt pointer rather than the merely-incorrect simulation the CVar comment leads you to expect.</p>

<h3><a name="Triggerthresholds"></a>Trigger thresholds</h3>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>fx.Niagara.Batcher.TickFlush.MaxQueuedFrames   default 3
fx.Niagara.Batcher.TickFlush.MaxPendingTicks   default 10</pre>
</div></div>

<p>Either crossing trips the flush (<tt>:405</tt>, <tt>:418</tt>). The <tt>MaxQueuedFrames</tt> help text says it is "generally only a concern when the application does not have focus" — the path was designed for a minimised editor, not for MRQ, which legitimately ticks the world without rendering every tick (warm-up frames, temporal sub-samples, and finalization).</p>

<h3><a name="Whytheoutputformatcorrelates"></a>Why the output format correlates</h3>

<p>Not alpha, and not bit depth — <b>finalize/write duration</b>. EXR and 16-bit PNG hold the pipeline long enough for ≥3 world ticks to elapse with no scene render, crossing <tt>MaxQueuedFrames</tt>. JPG does not. This also explains why the first crash landed exactly at MRQ's <tt>Moving to Finalize to finish writing items to disk</tt> — MRQ stops rendering the scene there while the world keeps ticking, so the threshold is crossed within a few frames.</p>

<h3><a name="Frequency"></a>Frequency</h3>

<p>10 for 10 on 2026-08-18 between 10:59 and 11:40, all with an identical signature. Crash GUIDs in <tt>Flossing/Saved/Crashes</tt>:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>0BF6DEB74DCE9C7AB3F1528D7896E696    RHIThread
7D28BEFC49E34ABB06214E92DFDD13B8    RHIThread
CA32B18048CB92979C8949B4F1D8B80F    RHIThread
493E616E4BD781DFE0F03095951E1E4F    RHIThread
6623D4FA464F18F9CED47EA9575F85D5    RHIThread
12BDB7D44C1E68A13674C48FA4370000    RHIThread
4E2B020A448915A91F9EF28233C0A07B    RHIThread
7156B1F74D6994B31C5AA99A25EB1540    RHIThread
DAB4A6DF4DB07931347BBBB81FD5C428    RHIThread
B8737D1445050FC93078BBA71C33A599    RenderThread  (-norhithread -forcerhibypass)</pre>
</div></div>

<hr>

<h2><a name="Diagnosisaid"></a>Diagnosis aid</h2>

<p>The default RHIThread stack is effectively undebuggable for this class of bug — the resource is bound on the render thread and dereferenced later on the RHI thread, so the recording context is gone. Relaunching with the following moves the fault to the recording site and was what made this diagnosable:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>UnrealEditor.exe &lt;Project&gt;.uproject -norhithread -forcerhibypass</pre>
</div></div>

<p>(<tt>-forcerhibypass</tt> — <tt>RHICommandList.cpp:1886</tt>. <tt>r.RHICmdBypass</tt> alone is a silent no-op while the RHI thread runs, because <tt>LatchBypass()</tt> at <tt>:1897</tt> requires <tt>!IsRunningRHIInSeparateThread()</tt>.)</p>

<hr>

<h2><a name="Workarounds"></a>Workarounds</h2>

<p><b>Recommended</b> — prevents the flush from firing during a render, keeps GPU particles and correct simulation:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>fx.Niagara.Batcher.TickFlush.MaxQueuedFrames 100000
fx.Niagara.Batcher.TickFlush.MaxPendingTicks 100000</pre>
</div></div>

<p><b>Confirms the diagnosis</b> — avoids the dummy-view path entirely:</p>

<div class="preformatted panel" style="border-width: 1px;"><div class="preformattedContent panelContent">
<pre>fx.Niagara.Batcher.TickFlush.Mode 0     # keep ticks queued
fx.Niagara.Batcher.TickFlush.Mode 2     # discard pending ticks (visible particle reset)</pre>
</div></div>

<p><b>Blunt</b> — <tt>fx.NiagaraAllowGPUParticles 0</tt>. Avoids the crash but removes the GPU-simulated effect from the render entirely, so it is not viable for delivery.</p>

<hr>

<h2><a name="Suggestedfix"></a>Suggested fix</h2>

<ol>
	<li>Plumb <tt>IsOutsideSceneRenderer()</tt> into Data Interface shader-parameter population so DIs depending on scene textures, depth, or distance fields can substitute a safe fallback binding. The flag already exists and is documented as being for exactly this purpose, but is only read in two places.</li>
	<li>Validate resource pointers before binding in <tt>ProcessPendingTicksFlush</tt>'s dispatch path, or skip stages whose DI bindings cannot be satisfied outside the scene renderer. Degrading to an incorrect simulation — which the CVar comment already sanctions — is strictly better than an access violation.</li>
	<li>Consider having MRQ suppress or raise the Niagara tick-flush thresholds for the duration of a render job. MRQ intentionally ticks the world without rendering every tick, so it will always be liable to trip a heuristic tuned for an unfocused editor.</li>
	<li><tt>checkf(Handle.IsValid(), ...)</tt> at <tt>D3D12Commands.cpp:673</tt> catches this class of problem in the bindless path. The non-bindless path immediately below (<tt>:679</tt> onward) has no equivalent guard and dereferences straight away.</li>
</ol>




<p><span class="nobr">[Link Removed]</span><br>
<span class="nobr">[Link Removed]</span><br>
<span class="nobr">[Link Removed]</span><br>
<span class="nobr">[Link Removed]</span><br>
<span class="nobr">[Link Removed]</span></p>
]]></description><link>https://issues.unrealengine.com/issue/UE-392456</link><guid isPermaLink="false">UE-392456</guid><pubDate>Thu, 29 Oct 2026 16:39:19 GMT</pubDate></item><item><title><![CDATA[Iris - actors spawned/loaded before a connection exists will not use delta compression]]></title><description><![CDATA[<p>FDeltaCompressionBaselineManager::CreateBaseline will return early if the object has an invalid ObjectInfoIndex, which is allocated in FDeltaCompressionBaselineManager::UpdateScope (by calling FDeltaCompressionBaselineManager::AllocPerObjectInfoForObject). However, UpdateScope only allocates info for objects that have newly come into scope.</p>

<p>UReplicationSystem::NetUpdate will skip calling FDeltaCompressionBaselineManager::PreSendUpdate (which calls FDeltaCompressionBaselineManager::UpdateScope) until a valid connection is present; however, the actor's index will still be set in PrevFrameScopableInternalIndices. Because of this, when a valid connection is present, FDeltaCompressionBaselineManager::UpdateScope will not allocate object info for the actor due to it not being considered an object that was just added.</p>

<p>The issue does not occur with bAllowMinimalUpdateIfNoConnections set to false, as this causes FDeltaCompressionBaselineManager::PreSendUpdate to be called from UReplicationSystem::NetUpdate regardless of whether a connection is present.</p>]]></description><link>https://issues.unrealengine.com/issue/UE-333873</link><guid isPermaLink="false">UE-333873</guid><pubDate>Thu, 29 Oct 2026 11:48:33 GMT</pubDate></item><item><title><![CDATA[View Bindings Panel Undo Crash — Function Parameter Pin Not Properly Serialized]]></title><description><![CDATA[<p>Initial investigation: After looking at the code, it seems the issue is that when the Function is set, its parameter isn't serialized as a valid Pin into the transaction. As a result, on Undo the transaction system fails to properly record/remove this parameter Pin, so after Redo the UI state and the underlying data become inconsistent — leaving a stale variable entry that should have been cleaned up, which crashes when clicked.</p>]]></description><link>https://issues.unrealengine.com/issue/UE-391655</link><guid isPermaLink="false">UE-391655</guid><pubDate>Wed, 28 Oct 2026 15:15:19 GMT</pubDate></item></channel></rss>