greebo

Ok. Testing my existing Arch VM with wxWidgets 3.2.0, mouse wheel is working. Maybe it's something related to XFCE, I'll try to install that version.

Which Linux distro are we talking about? I can try to repro it in a virtual machine.

No luck with this dump, but in this case I think the log file should show something that lead to the crash.

I've tried to look for similar reports like this, but there's nothing that sounds like the problem here. That might also be due to GLCanvas not being used that much (and wxWidgets 3.2.0 itself is not that widespread yet to begin with). At this point I can only speculate, since I haven't access to any AMD hardware at the moment. I can offer to look at a crashdump you produce with DR 3.0.0, and see if its trail is also leading to the AMD drivers.

@HMartThis is where it crashes: atio6axx.dll!00007ffe5c79cc74() Unknown atig6pxx.dll!00007ffe9665c18c() Unknown opengl32.dll!pgldrvLoadAndAllocDriverInfo() Unknown opengl32.dll!LoadAvailableDrivers() Unknown opengl32.dll!__DescribePrimaryPixelFormat() Unknown opengl32.dll!wglNumHardwareFormats() Unknown opengl32.dll!wglSetPixelFormat() Unknown gdi32full.dll!SetPixelFormat() Unknown wxmsw32u_gl_vc14x_x64.dll!wxGLCanvas::FindMatchingPixelFormat(const wxGLAttributes & dispAttrs, tagPIXELFORMATDESCRIPTOR * ppfd) Line 1068 C++ wxmsw32u_gl_vc14x_x64.dll!wxGLCanvas::Create(wxWindow * parent, ...) Line 769 C++ wxmsw32u_gl_vc14x_x64.dll!wxGLCanvas::Create(wxWindow * parent, int id, ...) Line 750 C++ wxmsw32u_gl_vc14x_x64.dll!wxGLCanvas::wxGLCanvas(wxWindow * ...) Line 690 C++ The code is trying to construct a render window, it's heading off somewhere into the ATI drivers where it dies. It appears that this is something that changed between DR 3.2.0 and DR 3.0.0, at least for you? DarkRadiant 3.2.0 uses the newly released wxWidgets 3.2.0, which seems to be the only relevant change here to me. The GLWidget code hasn't been changed at all for a couple of releases, so maybe it is a problem with wxWidgets 3.2.0 constructing the GLCanvas in a different way than before.

Yes, I can upload them to darkradiant.net, no problem. If you want to place them yourself, you can clone https://github.com/codereader/codereader.github.io and create a Pull Request with the screenshots you want to have added. There's already an images/screenshots/ directory in that repo.

It says I need to get permission to download that file, please.

Hm, can anybody with an AMD card confirm this? I'm afraid I don't have an AMD-based GPU around to try this.

Is DR 3.0.0 or DR 3.1.0 doing the same thing? If yes, it might be a renderer issue introduced during the rewrite.

Works for me. Is your colour scheme alright?

I do what I can, rest assured. And with all due respect, I don't have to give any reasons for not solving issues to anyone. As for the issue at hand, I've already started looking into it before you asked, so chances are high that I can fix it in time for the next release. It seems it's a problem with the entity colour wrongly affecting the editor preview pass. And one more thing, setting the priority on the bugtracker usually doesn't get anybody anywhere, this is not influencing my time nor my motivation. A higher priority might be justified in this case, but often people just want to get "their" reports prioritized and set "Major" or "Block" severity on almost everything. In the end I pick the problems I'm interested in and I can find enough energy for. edit: Don't get me wrong, I do like coding on DR because mission authors are using it, so for the most part I do it for them. But I do get a bit touchy when I feel that people are trying to "make use" of me and try to assign work to me.

Renderer support for materials with multiple interaction stages (in this example two diffusemaps combined with one bump and one specular):

Better support for "set X on Y" spawnargs to set values on attached entities:

Blend Light Rendering:

Model Selector now lists entityDefs using the selected model. You can double-click to create the entityDef instead of a regular func_static:

DarkRadiant 3.2.0 is ready for download. What's new: Feature: Show entityDefs related to selected models in Model Chooser Feature: Support for rendering blend lights Feature: Implement sorting of Interaction Stages Feature: Recognise type of "set x on y" spawnargs Feature: OBJ files: loader supports usemtl keywords directly referencing material names (without .mtl file) Fixed: Material editor: additional preview object in textures/glass/ materials Fixed: Every attempt to load a texture will trigger a SIGSEGV signal caught: 11 Improvement: Update to wxWidgets 3.2.0 Windows and Mac Downloads are available on Github: https://github.com/codereader/DarkRadiant/releases/tag/3.2.0 and of course linked from the website https://www.darkradiant.net Thanks to all the awesome people who keep using DarkRadiant to create Fan Missions - they are the main reason for me to keep going. Please report any bugs or feature requests here in these forums, following these guidelines: Bugs (including steps for reproduction) can go directly on the tracker. When unsure about a bug/issue, feel free to ask. If you run into a crash, please record a crashdump: Crashdump Instructions Feature requests should be suggested (and possibly discussed) here in these forums before they may be added to the tracker. The list of changes can be found on the our bugtracker changelog. Have fun mapping!

Coming back to this discussion from issue #5739: I'm not so sure about putting the entity classes in the tree as children to the object, but what about this mockup below? Is having a tab with the "Related Entity Classes" too well hidden to be useful or would this be ok?

I don't insist on doing that myself, so feel free I'm currently upgrading to wxWidgets 3.2 on Windows, and I had to fix a few things in terms of implicit wxIcon>wxBitmap>wxBitmapBundle conversions.

We got a bug report here which seems to happen using wxWidgets 3.2. @jonri Is this something that looks like problems you remember running into? https://bugs.thedarkmod.com/view.php?id=6080

Does this help? https://docs.microsoft.com/en-us/windows/win32/wer/collecting-user-mode-dumps It says if you create that registry key it will automatically save a dump file to somwhere below your AppData folder.

You need to sign up here: https://bugs.thedarkmod.com Once you have a login, you can create a bug using "Report Issue": https://bugs.thedarkmod.com/bug_report_page.php

Can you please record a crash dump from the moment it stops working? It's the only way I can find out what's going wrong. (Crashdump instructions)

My bad, I've been thinking about Selection Sets. Of course, I'm using groups all the time... Yes, I know about that feature request, to be able to select within groups without disbanding them - very useful, but more involved than it might appear at first glance.

This is exactly the use case I had in mind that got me to open this thread and ask. I recently moved back the undo queue to select an item I had deleted earlier, selected it, copied it to the clipboard, then redid all the steps and pasted the item into the map. A move like this is no longer possible when the redo stack is cleared by selecting the item. But then again, I'm used to selections being undoable from Blender, where this is quite a time saver. It's very easy to mis-select a vertex when moving fast, and being able to remove those wrongly selected items by hitting Ctrl-Z is really nice. For this one feature I'd be interested whether anyone is using it at all - are many people using selection groups in their maps? I found it less useful myself, and the workflow is clunky, having to think of a name for the group before saving. - edit: scratch that, I've been thinking about selection sets here.

I'd like to get some feedback about this old feature request, before I start working on it. What do mission authors think, would it be beneficial to have the selection status be included in the undo/redo operations? Like, you select two items, and by hitting Undo the most recently selected item is de-selected again (hitting undo again will proceed to de-select the first one). Thoughts?