illwieckz commented Nov 20, 2024

CMake provide downloadable cmake binaries that runs on the old distro used to build the releases, so I don't mind bumping the requirement.

illwieckz commented Nov 20, 2024

Also, we have M_SQRT2.

slipher commented Nov 20, 2024

A freetype submodule is only needed as an RmlUi dependency, so shouldn't that be in the Unvanquished repository rather than Daemon? To build Lua f…

sweet235 commented Nov 20, 2024

The only thing I am currently doing is #2780. Need to check if that PR is up to date with my server.

DolceTriade commented Nov 20, 2024

fair enough :(

illwieckz commented Nov 20, 2024

Nice!

VReaperV commented Nov 20, 2024

Nope, it's just vertexSimple.

VReaperV commented Nov 20, 2024

This reveals another bug: the main menu, connecting, downloading, and possibly loading screens are all rendered on top of each other when connectin…

illwieckz commented Nov 20, 2024

Lua is very fast to build, Freetype a bit less I guess, but likely not in a way it is that slow. The Saigo toolchain will skip the very slow pexe-t…

illwieckz commented Nov 20, 2024

Ah I see, for some reasons my mind misread glConfig2.specularMapping instead. 😅️

slipher commented Nov 20, 2024

You can use specular mapping while reflection mapping is disabled. That's the broken case.

illwieckz commented Nov 20, 2024

Maybe there is a way to test for glConfig2.reflectionMapping when setting material->enableSpecularMapping earlier? Anyway, this is not blocking.

slipher commented Nov 20, 2024

LGTM

slipher commented Nov 20, 2024

The name radius is a bit misleading since it suggests a half-extent, not the full extent. I'd call it width

slipher commented Nov 20, 2024

Could you comment the geometrical motivation behind square root of 2?

slipher commented Nov 20, 2024

Would be good to include CONTENTS_PLAYERCLIP

illwieckz commented Nov 20, 2024

Done.

slipher commented Nov 20, 2024

Technically action rush also targets enemy buildables. But I'm not really worried about the standard BTs. This could be annoying if you are trying …

illwieckz commented Nov 20, 2024

Done.

illwieckz commented Nov 20, 2024

Removed.

illwieckz commented Nov 20, 2024

Removed.

slipher commented Nov 20, 2024

Does this mean the bot will start flying almost every time there is an enemy near?

slipher commented Nov 20, 2024

I don't think we have any code for using jetpack in master? We have to start somewhere, don't we. The PR title/commit message only described bu…

illwieckz commented Nov 20, 2024

PNaCl still needs that file, but we are lucky that PNaCl doesn't require this change, so yes we can probably skip this commit, this whole folder wi…

illwieckz commented Nov 20, 2024

Something I committed by mistake when I tested: #1312 Fixed.

illwieckz commented Nov 20, 2024

Done.

slipher commented Nov 20, 2024

Maybe for bot management strategies more complex than bot fill we should just use a Lua script...

slipher commented Nov 20, 2024

The variable vm defined a bit above is PNaCl-specific (and redundant with the newly added NACL_VMS_PROJECT).

slipher commented Nov 20, 2024

Let's just turn it off uniformly? I can't imagine they actually implemented multiple addressing modes in the special compiler backend.

slipher commented Nov 20, 2024

You said we can do without zlib so this is not needed, right?

slipher commented Nov 20, 2024

What's this change?

slipher commented Nov 20, 2024

saigo_arch is unused

slipher commented Nov 20, 2024

How about adding the project name in the comment so the message isn't identical every time?

illwieckz commented Nov 19, 2024

sgame status The sgame builds for all architectures. I have tested the amd64 sgame and it runs. cgame status The cgame builds for all architectures. …

illwieckz commented Nov 19, 2024

So, this looks ready to me, it is known to be able to build a full game (cgame+sgame) that runs once other PRs taking care of other topic are merge…

illwieckz commented Nov 19, 2024

I think we should be able to avoid zlib by adjusting the Freetype configuration. zlib is only depended on from functions concerning gzip support. …

illwieckz commented Nov 19, 2024

The bug affects both the engine and the cgame, but it looks different if both engine and games are compiled with affected compilers than only one o…

illwieckz commented Nov 19, 2024

With current Saigo (LLVM 19): So that looks to be a game code bug.

illwieckz commented Nov 19, 2024

Does 2D transformation require some vertex complex code?

slipher commented Nov 19, 2024

This shader would have less than 10 lines in its main functions so no need to share anything.

illwieckz commented Nov 19, 2024

The new insert mechanism may be usable to share possible common code if needed, the macro creep was basically a workaround for no insert.

slipher commented Nov 19, 2024

Since it turned out that 2D drawing doesn't actually need any special GLSL, I might make a very basic shader to be shared by 2D and debug drawing. …

illwieckz commented Nov 19, 2024

So with the commit from: #3196 It works!

illwieckz commented Nov 19, 2024

It works! ❤️

VReaperV commented Nov 19, 2024

I've been thinking that by quantising the stage data and offloading textures to a buffer with a fixed layout might improve this further. For contex…

Viech commented Nov 19, 2024

This would make sense lore-wise. But from a gameplay perspective it's humans with their lower mobility who are in greater need of this information …

illwieckz commented Nov 19, 2024

If I comment out the <if> tag in ui/options_ui.rml, the execution continue, but then I get stuck in ui/callvote_fillbots.rml and there is now <if> …

VReaperV commented Nov 19, 2024

I believe #3061 is a fix for this issue.

VReaperV commented Nov 19, 2024

Could this just be an asymmetrical aspect between the 2 teams? It makes sense that OM does this because it's some magic telepathic bullshit, howeve…

VReaperV commented Nov 19, 2024

I know that some form of this exists on @sweet235 Map&Bot Testing server. I don't know if it's in master though.

VReaperV commented Nov 19, 2024

model is not properly textured by itself Is this still the case? The other 2 issues have been fixed by converting the model to iqm.

VReaperV commented Nov 19, 2024

To me, making them blue and red sounds like they would blend in too much with the background?

illwieckz commented Nov 19, 2024

If I comment out the <if> tag in ui/options_ui.rml, the execution continue, but then I get stuck in ui/callvote_fillbots.rml and there is now <if> …

illwieckz commented Nov 19, 2024

It is stuck with an <if> tag.

illwieckz commented Nov 19, 2024

So in libs/RmlUi/Source/Core/XMLParser.cpp in XMLParser::HandleElementStart() it is stuck in node_handler->ElementStart()…

illwieckz commented Nov 19, 2024

Ah no, BaseXMLParser::HandleElementStart() is overriden by XMLParser::HandleElementStart()… 🤦‍♀️️

illwieckz commented Nov 19, 2024

Well, because both parameters are RMLUI_UNUSED_PARAMETER, I guess the compiler can just skip the function.

illwieckz commented Nov 19, 2024

So I did this: void BaseXMLParser::HandleElementStartInternal(const String& name, const XMLAttributes& attributes) { line_number_open_tag = line_nu…

illwieckz commented Nov 19, 2024

In BaseXMLParser::HandleElementStartInternal() it is stuck in HandleElementStart().

illwieckz commented Nov 19, 2024

It is stuck in some HandleElementStartInternal() in ReadOpenTag(): bool BaseXMLParser::ReadOpenTag() { // Increase the open depth open_tag_depth++; //

illwieckz commented Nov 19, 2024

I now did that in ReadBody(): else { Log::Message(Log::LT_INFO, "BaseXMLParser::ReadBody 1.5.1"); if (!ReadOpenTag()) { Log::Message(Log::LT_INFO, "…

illwieckz commented Nov 19, 2024

So I did that: void BaseXMLParser::ReadBody() { RMLUI_ZoneScoped; open_tag_depth = 0; line_number_open_tag = 0; Log::Message(Log::LT_INFO, "BaseXML…

illwieckz commented Nov 19, 2024

I just noticed that the counter restart when doing ReadBody(), so maybe the function is recursive.