OrbWeaver

I tested it and it looks pretty good. Live camera synchronisation from DarkRadiant to the game works. The game position moves in realtime as I move the DR camera around. Synchronisation in the other direction works if I choose the Sync camera back now option, but not in realtime. I don't know if this is expected. Hot reload works. I can change the colour of a light and see the change immediately in game if I choose Enable map update mode. Something weird happened to the camera in game at one point — I could no longer rotate in place, because horizontal mouse motions s

I've only tested compilation so far; I'll try to do some actual testing of the functionality this week assuming I don't run into any issues compiling the latest game SVN.

I merged your PR into my own topic branch for testing on Linux, but I haven't merged it yet into my master branch because Greebo mentioned in his unit testing thread that he is still planning to merge my master branch into his own repo, and I was worried that integrating your PR first might make this more difficult. Although it probably doesn't affect the difficulty that much (in terms of conflicts etc) whether he merges my master followed by your changes (in my topic branch) as two separate steps, or if I merge it into my master and he then merges my master into his repo in a single step.

It's easy to forget that humanity isn't a monolithic civilisation. Have you ever thought about how bizarre it would be to take modern technology like airliners or even battery-powered flashlights back into the stone age? Then you realise that you don't even need time travel for that: there are literally societies living right now who are still in the stone age. Remote tribes in jungles or on islands who are still chucking flint-tipped spears at each other as if the past few millennia simply never happened.

Delighted to hear that a mapper is actually using my ambients. I guess this map will be next on my list to play.

Admittedly I haven't tried it yet for a second time, but I would tend to rate Prey quite highly for replay value because you can play the game in a different way (e.g. using more human skills vs more Typhon skills etc). There are also different objective choices you can make but if you didn't find the story that engaging maybe this aspect won't be of interest to you. I've heard that if you do another playthrough there is an area you can go to where you get back all of the neuromods you accumulated during the first run, so you can hit the ground running with upgrades.

This must be one of the most hilarious suggestions for a Thief-like game I've ever heard. Random "friendly thieves" appearing out of nowhere and telling the player to complete their objectives. Perhaps the player could be followed around through the entire mission by a squad of four friendly thieves, giving helpful hints at regular intervals. "Don't forget to use your water arrows, Dr Freeman!". That seems to be the standard mechanism and will be well understood by players. It's a very common mechanic from MMOs and other games — activate the three pillars to unlock the sealed door and

There is no connection whatsover between Prey (2006) and Prey (2017) other than the fact that they have the same name and are first-person shooters. I actually liked the 2006 game with its weird physics-defying portals and creative use of the Doom 3 engine, but I don't think it was anywhere near as well-known or popular as its 2017 namesake.

None of the official mission servers are backed up? Yikes. I would have thought you'd want one master server with full backups and maybe version control, and this is the only server you upload to, then a bunch of mirrors which just replicate whatever is on the master server without needing their own backups.

Psi powers come from neuromods. None of them are available at first so you might not have access to them yet. I can't remember exactly when it happens but there is some interaction with the aliens which gives you the power to scan Typhids and then unlock various neuromods which give you skills like mindjacking and various psi-based damaging abilities. The disadvantage of taking lots of psi neurmods is that the turrets identify you as alien and will shoot at you.

I more or less completed this one last weekend. I say more or less, because although I reached the Mission Complete screen, I didn't find the hall of the ancients or the map, which sounds like it might have been interesting. Based on what I did see: The initial sewer section was possibly my favourite area yet. I particularly like underground areas and this one felt very realistic (not that I have any idea what a "real" sewer would look like), with a noticeable Half-Life 2 feel. The heavy locked gates were impressive and gave the impression that you were trapped in part of a much l

Try using more Psi powers, if you have them. You can recharge psi quite cheaply (even for free if you have the suit mod that allows you to regenerate psi whenever standing in the yellow "coral"), and powers like kinetic blast and mindjack can really turn the tide of battle. For the Nightmare, I don't bother trying to kill it, but just hide from it. It is quite large and cannot get into small spaces or use lifts, so in many areas it is quite easy just to hunker down for 3 minutes until it goes away.

The thing I found about Prey is that the difficulty varies hugely based on your upgrades and equipment. For example, I found the military operators very challenging either to sneak past or defeat in combat (especially when they respawn), but if you get the Typhid upgrade which allows you to hijack them for 30 seconds, it becomes trivially easy. Hijack one, off it goes to fight amongst the others, wait a few seconds and you can hijack another one, repeat.

I've merged stgatilov's PR into a topic branch in my repo called topic/gameconnection with some initial Linux fixes, so as and when you do want to have a look at this, feel free to pull from this branch. Most of the fixes are trivial stuff like #include filename capitalisation and the odd missing standard library header, adding the new files to the Linux Makefile etc. More notable changes: The clsocket library files are now all in a subdirectory of radiant/gameconnection to eliminate the out-of-tree dependency on files inside lib. This mirrors the layout of the picomodel library w

There are three options available at the point of export: Full smooths everything, giving the same result as ASE models. If you want unsmoothed faces in this mode you would need to split them (although from what you say, this wouldn't actually help). None leaves the whole model flat shaded, irrespective of split or non-split faces. Use Autosmooth settings smooths the model based on the Autosmooth checkbox and angle field in the Blender mesh properties. Angles less than the threshold will be smoothed, greater angles will be flat-shaded. I assume this is the most useful setti

Blender 2.80 supports exporting LWO with full angle-based smoothing as of the latest version of my (building on the work of others) LWO export script for Blender. There is no support for explicit smoothgroups AFAIK, other than by splitting vertices. In terms of fully open formats, Alembic looks interesting. It is free, and supported by almost everything according to Wikipedia. I don't know how its features stack up against other formats though, or whether it would be suitable for both static and animated meshes.

You can download Blender 2.80 compatible versions of both LWO and ASE export scripts from the link in my signature. The older versions floating around in other forum threads will probably not work with Blender 2.80.

This seems like a very useful tool, but could a moderator move the thread to a more appropriate forum? It really isn't DarkRadiant-related.

That seems like a good idea, and not just for download size — compressing at load time is horribly slow on certain (mainly AMD?) GPUs, including mine, unless you manually disable image_useNormalCompression which currently defaults to 1. If we're going to default to using normal compression I think we should also supply pre-compressed images to hopefully avoid this delay.

In theory you could reduce the size of the mod to almost any arbitrary size if you were willing to sacrifice quality. For example you could re-encode all the sounds at Vorbis quality 0, reduce all textures to 32x32 pixels with maximum DXT compression, remove any embedded videos (if there are any). Probably not much you can do about models, but I don't suppose they take up all that much space particularly after Zip compression. Of course somebody would need to do all the work of producing this cut-down version, and the result would be like playing a game from 1996, but maybe some people wo

Ah right, that's definitely very non-standard and potentially confusing. If there are CPP files to be compiled inside a subdirectory of lib, this should be a separate sub-project which generates a dynamic library which is then linked into Radiant or whatever else uses it, rather than other projects directly compiling CPP files from inside lib. Alternatively, if you don't want the hassle of creating a library subproject, you could put the CPP files directly inside a subdirectory of radiant at which point they can be compiled directly into the radiant project. Indeed. I've updated the co

If it's only used by code in radiant, putting it in there is OK, but if you would rather put it in lib (or it might be needed elsewhere), you could create a new subdirectory in lib with both header and cpp files, much like wxutil. Of course that will need changes to the build scripts to build the new library and link it where appropriate. Typically I use 4 spaces with Allman style braces, although we have a lot of legacy code which uses different styles so you will see a fair amount of inconsistency. Note that I do strongly prefer function open braces in column 0 where possible,

Sounds pretty good. I look forward to merging all this when it's ready.

Before jumping into this, I would suggest running some CPU stress tests and thermal monitoring to determine if overheating actually is a problem. If your CPU temps are fine and you start removing coolers and re-applying thermal paste, you are at risk of making things worse (or even damaging the CPU) for no actual benefit, especially if you are not experienced at this kind of hardware maintenance.

Yes, that would be simplest I think. Although I am not on GitHub myself (my repo is on GitLab), this is no problem because I can do the merge using the command-line tools while pulling anonymously from any accessible repository. I think the fix for this would be for entities to properly update their observers when importing a Memento from the undo system, which I'm guessing they don't currently do. Yes, if you like the clsocket library I suggest switching to it now, before too much code is linked into ZeroMQ (which it sounds like we don't need and doesn't add any real value). I