Search the Community

Well, let's explore this a bit. How can this be solved? Currently, creeping overrides running (like you said). Here are a couple issues or considerations: What about players who want to keep the fine control of toggling each one independently? Perhaps some players want to go from creeping to running. At the moment, the code is written in such a way (due to its Doom 3 history) that toggling creep can set the toggled run state, but toggling run cannot set the toggled creep state. The toggle creep key can set the toggled run state to walk but only once. If the player presses the toggle run key again, it will toggle without regard to the toggled creep state. Fixing this would require a lot of code rewriting. Brainstorming: It almost sounds like increase and decrease speed keys are desired. Run key to go from creep to walk and from walk to run. Creep key to go from run to walk and from walk to creep. If you're curious to give this a try, here's a Linux test build that matches beta212-05 (rev 16950-10635) with the following change: The toggle creep key sets the toggled run state to walk but only once. If the player presses the toggle run key again, it will toggle without regard to the toggled creep state. https://drive.google.com/file/d/1osTCQRf7LQ5wPvhGl2uRU4NcFnJPEu9_/view?usp=sharing

Wasn't sure what blind and deaf actually mean to the code: I thought they only offset the alert level increase rate, not what AI does afterward. You're right that it's kind of silly I didn't think to check on higher difficulty settings as they didn't come to mind as a possible factor. I'll do that as well and see which of those points it improves: #1 might be offset by that... I think the others aren't difficulty related and just not implemented, I'll actually check before presuming though.

AI specifically have code to test the attack vector I think, so they should generally know where it came from. They're also supposed to prioritize possible hiding places and investigate them first. If you're playing with AI acuity turned down, it might cause them to lose you in darkness faster. I have them both on Challenging and they go to alert almost instantly even when I'm just a little lit up. It's very challenging on open maps when you are on a balcony.

I tried to debug this. It turns out that "global variables" includes all immediate values (including string literals which take 128 bytes each) and all global function declarations (4 byte per function, including even engine events). I see 119240 bytes before starting Iris, and 124924 bytes after starting it. On Bakery Job, I see 109680 before starting game, 109700 bytes after starting it. So It seems that core game takes about 110K, but Iris itself takes only about 10K, and the limit is at 196K. Does not sound like a big problem for missions. With addons, you have to compile all of them at once, and perhaps there is a lot of code, so you reach the limit. Ok, let's double MAX_GLOBALS for 2.12.

As someone who tends to alert guards often and occasionally stir trouble when going through a FM, I noticed some major issues when it comes to AI realism and awareness during combat or when guards face difficult situations. Everything's fine when AI go about their usual patrols... once trouble takes place however, the illusion falls apart as guards act like they're less self aware than a toddler. Indeed AI realism can't be improved past a certain point as there's a limit to the effort the team can put into something so complex... yet I do believe a few improvements can make AI behavior much more realistic and exciting. After analyzing this issue for a long time, I decided to put it all it into a few main points... I apologize for their length as I wanted to go in a bit of detail on each one, hope folks have the time and patience to read them. Biggest issue is AI are unaware of where attacks are coming from. I recently made another thread on how I climbed on a table and blackjacked two guards to death as they sat there doing nothing, something that also happens when shooting them with arrows as guards only explore the nearby area. The issue seems to be that AI don't account for the direction a hit comes from, they only know something hit them but act as if it must be some mystical force of nature: If you're sitting in a parking lot and an asshole neighbor throws a tomato at you from his balcony, you aren't going to cluelessly investigate the road in front of you when the projectile clearly came from behind and hit you in the back of the head, instead you'll storm into the building and start looking for which of the neighbors facing that side of the road may be the culprit. Despite voice barks existing for this exact scenario, we never see AI running to get help from other AI. NPC's will do one of two things: If armed and with enough health they will attack or search for you nearby, if hurt or unarmed flee to a random location. I've never seen an AI consciously run up to another AI asking for help and bringing them to where they spotted me, even when fleeing the AI seems to go to a random location. They don't share knowledge with each other generally speaking: The only awareness AI spread to other AI is alert level, meaning NPC A becomes alert if it sees or hears that NPC B is alert too... beyond that there's no coherence or actual cooperation, the voices may indicate some form of searching together but friendly NPC's are never seen actually engaging. Another big issue is voices being played (or not played) in disconnect with what's actually going on. There are AI voices for most important circumstances but they're very rarely activated: It's a miracle to hear a guard say "someone's been hurt" or "there's a body here" when noticing someone who's unconscious or dead. What seems to happen is if AI was already alerted by another peculiarity such as a noise, they're no longer surprised by anything else and won't play the voices designated for that scenario, so they'll only mention a body if that's the first thing to alert them in any way. Furthermore AI don't actively talk with each other while searching together, everyone acts as if they're on their own and not a team. What happens after a conflict is over. For this discussion I won't focus on better permanent alerts, that has greater difficulty implications and I think I made a separate thread on it a while back. The problem I noticed is once the immediate alert has gone down, AI return to full normality and act abnormally calm: The idle voices change from saying things like "it's a quiet night" to "we've got an intruder" but that's about it. In any realistic scenario even a trained guard would be shocked after being in a fight or finding a body. Below I'll list the immediate improvements I see to those problems, which without having an understanding the code myself am presuming can be changed without too much effort: When an enemy hits the AI with any weapon, the AI should be alert to the estimate location of the shooter. If you're standing atop a tower and fire an arrow at a guard, the guard shouldn't draw his sword and look around their nearby vicinity like a fool, but instead run up to the tower where you're standing granted they can pathfind their way to that location. If the player is far away the destination should be fuzzy and a random location nearby, thus the guard won't run to your exact location but will still climb the stairs and enter a room near it. AI need to learn how to ask for help instead of fleeing to random places when not attacking. If an ally who isn't already alert can be found nearby, the scared AI should explicitly run to their location tell them where you are then have the ally either run to your location (if armed) or go to another ally to get them to your location (if unarmed). Even if an AI is already alert, finding a body or dropped weapon or broken arrow should result in the AI speaking the voice line for that circumstance, only being engaged in combat should suppress it. I'd go further and support repeating those voice lines: A guard yelling "we have a dead body" several times during the first seconds of discovery would make them appear more shocked. Similarly talking to a nearby guard shouldn't be done just once when the two first meet: When multiple AI are searching for you, they should constantly alternate between single voice lines (eg: "I bet you're right over there in those shadows") and looking at another guard to talk to colleagues (eg: "I know I saw him here keep searching"), this would be a huge improvement since guards currently act like they're completely unaware of each other during a coordinated search. Making guards permanently affected after an incident is a trickier one but a few tweaks could improve it. The most immediate solution would be changing the idle animations: Instead of stretching or blowing their noses or eating candy, AI should be seen randomly cowering or face-palming or even playing the scout animation to look around carefully. One suggestion I'd absolutely throw here: If the AI found a dead body from an ally, have them cry occasionally... I think that would be an interesting and unexpected detail, that will also get players to think and feel more about the consequences of their actions and how they affect the world. There are other ones I could get into, but some would be more difficult and likely not worth trying to solve. Most notable and worthy of at least a mention is how AI walk over the bodies of fallen friends as if they're doormats: Obviously there's no way to have them drag bodies to the side, but maybe an avoidance mechanism so they don't look like jackasses trying to profane their dead friends by literally stepping all over them could be a fix for that as well! Let me know what you think of those points and if there are other AI issues you've noticed yourself or better solutions you can think of: I'm not sure if I got everything here but I definitely believe the problems exist and we could make the world more natural and immersive with some simple fixes.

Here's my first FM. A small and easy mission, inspired by Thief's Den and The Bakery Job, where you must find and steal a cook's recipe book in order to save a friend from going out of business. Download: Mediafire (sk_cooks.pk4) TDM Website's Mission Page The in-game mission downloader Thanks to: The people who helped me get this far, both in the forums and on Discord. The beta testers: MirceaKitsune, Mat99, Baal, wesp5, Cambridge Spy, jaxa, grodenglaive, Acolytesix ( Per the author in the beta testing thread. ) Skaruts has given permission to the TDM Team to add Subtitles or Localization Strings to this mission. (No EFX Reverb.) If anyone from the Community or TDM team wishes to create these we will gladly test them and update the mission database.

Thank you! Unfortunately it didn't fix the bug. Not worth it to keep tinkering since stgatilov fixed it on the code side. The next 2.12 beta build will be good to go. I committed this anyway since it has the new subtitles.

With TDM 2.12, after the credits finished, the "Mission Complete" screen did not display. I found that the screen was black and I could hear my footsteps when I tried to move around. I think the reason for the mission not completing successfully was that the "Do not kill or harm allies" objective was never marked as "1 = STATE_COMPLETE" instead it was left as "0 = STATE_INCOMPLETE". Note, I didn't use noclip throughout the mission. Same as: https://forums.thedarkmod.com/index.php?/topic/18054-fan-mission-the-accountant-2-new-in-town-by-goldwell-20160509/&do=findComment&comment=458491

How about using TDM automation framework (and maybe pcem/qemu)? More info see: https://forums.thedarkmod.com/index.php?/topic/19828-automation-features-and-discussion/

Since it's a one-off problem, only for the first ambient at game start, then the way you fixed it is really the right way, because the system-level alternative would be whole a little subsystem (a specialized fidelay and mandatory _z soundshader) just to handle the first 0.1 seconds of game start, and then the old system for all the maps already out there, which is a bit much. And I think it's not even that common because many ambients start quietly to begin with. That said, there might be an easy bit of code that could make sure the fade in always works for the first ambient that doesn't mess with anything else in the system, like a hardcoded initial 0.1 sec. delay only for it, and that may be worth doing. But it'd need experimenting and testing to make sure it works as intended and doesn't have unintended consequences. Anyway, it's good we have this documented for now as the way to fix the problem for other people in the future that run into it searching for a fix. In fact it'd be good to put in the wiki to make sure the fix doesn't get lost.

Thanks! I guess that was a red herring. The problem happens due to a change made in source rev 10498: May I kindly ask that you nudge the player start position? It seems the current start position relies on incorrect physics and stgatilov does not want to revert to the incorrect behavior. Edit: Looks there may be a "code side" fix after all.

What does the code fix actually fix? If fixing something breaks something else, it's i.m.o. not a (good) fix. Edit: I would agree that maybe it's better to fix the mission(s) instead.

Do you realize that the original code is wrong? In fact, it has been wrong since revision 2, i.e. since the very beginning. Absolutely no reason to do anything special for one platform.

Well I think I've gone as far as I can with this but libxml is simply not doing what it's supposed to. I can put code like this in the Game constructor: auto games = GlobalRegistry().findXPath("//game"); assert(!games.empty()); // SUCCESS assert(games[0].getAttributeValue("name") == _name); // SUCCESS assert(!GlobalRegistry().findXPath("//game[@name='" + _name + "']").empty()); // FAIL The "//game" node is there, it has an attribute "name" with value "Doom 3 Demo" (or whatever the first game to be loaded is), but passing that exact name string back to an XPath query using the [@name='blah'] syntax just doesn't work, even though it used to work fine, and should work according to the XPath specification. I even dug down into the Document::findXPath function to see if something was being set in one of the C structures to indicate what is going wrong, but even though there is a lastError struct, it is empty. If there is any indication of the problem, it is buried deep within impenetrable and poorly-documented C structures. So, I am done with libxml and will look at C++ alternatives. If we were going to integrate this, what is the best way of doing so? I seem to recall from previous discussions you would prefer not to have git submodules linked into the repo, and it looks like pugixml is just one .cpp and one header file, so I guess the simplest approach is just to download the files and add them directly to the build in a suitable directory (like we do with the libfmt library).

Yes, I would like to find out what the issue is even if the ultimate decision is to use a different library in future. So far I have established that this line in Game.cpp is apparently working (no exception is thrown): // Import the game file into the registry GlobalRegistry().import(fullPath, "", Registry::treeStandard); but this (and any subsequent attempt to getKeyValue) then fails: // Get the engine path _enginePath = getKeyValue(enginePath); I can only think of two hypotheses: It's a timing issue or race condition: the import call is returning but something in libxml2 has not "finalised" the tree so it isn't ready to do XPath queries yet. I don't think there is any asynchronous or threaded code in the libxml2 API but who knows. There's something different about how XPath queries need to be formatted after merging. E.g. perhaps "//game" no longer works, and it has to have some other prefix. Further debugging is therefore needed. Edit: OK, there's definitely something weird going on with encodings. Adding a Registry::dump call into the Game constructor gives me output like this: output error : string is not in UTF-8 output error : string is not in UTF-8 eCrosshairs" ="mU" ���mU="SHIFT"/>< ="ToggleGrid" ="(&#xB3;R"/>< ="ToggleView" ="U" ���mU="SHIFT+CONTROL"/>< ="NextView" ="" ���mU="CONTROL"/>< ="ZoomIn" ="Delete"/>< ="ZoomOut" ="Insert"/>< ="CenterXYViews" ="" ���mU="CONTROL+SHIFT"/>< ="CenterXYView" ="" ���mU="CONTROL+ALT"/>< ="ToggleCubicClip" ="" ���mU="SHIFT"/>< ="ToggleCamera" ="U" ���mU="SHIFT+CONTROL"/>< ="TogTexLock" ="" ���mU="SHIFT"/>< ="DragVertices" ="U"/>< ="DragEdges" =""/>< ="DragFaces" =""/>< ="ToggleSelectionFocus" ="" ���mU="CONTROL"/>< ="ThickenPatchDialog" ="" ���mU="CONTROL"/>< ="ToggleShowAllLightRadii" ="" ���mU="CONTROL+SHIFT+ALT"/>< ="ToggleClipper" ="mU"/>< ="MouseTranslate" =""/>< ="MouseRotate" =""/>< ="MouseDrag" =""/>< ="NewOrthoView" ="" ���mU="CONTROL+ALT"/>< ="SetGrid0.125" ="" ���mU=""/>< ="SetGrid0.25" ="" ���mU=""/>< ="SetGrid0.5" ="" ���mU=""/>< ="SetGrid1" ="" ���mU=""/>< ="SetGrid2" ="U" ���mU=""/>< ="SetGrid4" ="" ���mU=""/>< ="SetGrid8" ="" ���mU=""/>< ="SetGrid16" ="" ���mU=""/>< ="SetGrid32" ="" ���mU=""/>< ="SetGrid64" ="" ���mU=""/>< ="SetGrid128" ="" ���mU=""/>< ="SetGrid256" ="" ���mU=""/>< ="ToggleTextureTool" ="" ���mU="CONTROL+ALT"/>< ="ToggleMainControl_TextureTool" ="" ���mU="CONTROL+ALT+SHIFT"/>< ="NormaliseTexture" ="" ���mU=""/>< ="TexToolGridUp" ="plus" ���mU="SHIFT"/>< ="TexToolGridDown" ="minus" ���mU="SHIFT"/>< ="TexToolMergeItems" ="" ���mU="Y&#x11D;mU"/>< ="GroupCycleBackward" ="ISO_Left_Tab" ���mU="SHIFT"/>< ="GroupCycleForward" ="" ���mU=""/>< ="RevertToWorldspawn" ="" ���mU="SHIFT"/>< ="SavePosition1" ="" ���mU="CONTROL+ALT"/>< ="SavePosition10" ="(&#xB3;R" ���mU="CONTROL+ALT"/>< ="SavePosition2" ="U" ���mU="CONTROL+ALT"/>< ="SavePosition3" ="" ���mU="CONTROL+ALT"/>< ="SavePosition4" ="" ���mU="CONTROL+ALT"/>< ="SavePosition5" ="" ���mU="CONTROL+ALT"/>< ="SavePosition6" ="" ���mU="CONTROL+ALT"/>< ="SavePosition7" ="" ���mU="CONTROL+ALT"/>< ="SavePosition8" ="" ���mU="CONTROL+ALT"/>< ="SavePosition9" ="" ���mU="CONTROL+ALT"/>< ="LoadPosition1" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition10" ="(&#xB3;R" ���mU="Y&#x11D;mU"/>< ="LoadPosition2" ="U" ���mU="Y&#x11D;mU"/>< ="LoadPosition3" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition4" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition5" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition6" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition7" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition8" ="" ���mU="Y&#x11D;mU"/>< ="LoadPosition9" ="" ���mU="Y&#x11D;mU"/>< ="ToggleOrthoBackgroundPanel" ="U" ���mU="Y&#x11D;mU"/>< ="ToggleRotationPivot" ="" ���mU="CONTROL"/>< ="ToggleAasVisualisationPanel" ="" ���mU="CONTROL+SHIFT"/>< ="GroupSelected" ="" ���mU="CONTROL+SHIFT"/>< ="UngroupSelected" ="" ���mU="CONTROL+SHIFT"/></mU><> Whereas with xmlParseFile there are no weird characters and the output just looks like properly formatted XML. I wonder if we have some nonstandard characters in our XML files for some reason.

Oh boy. Thanks for investigating this, reading the issue description I suspected that the XPath queries might be failing, but I thought it was rather due to the libxml2 version used in newer Ubuntus. I recall now that the automated Ubuntu build on Github is not running the unit tests (couldn't get it to run and gave up at some point) - so I missed that this is breaking stuff. I searched around for existing code to find out what to pass for that encoding parameter in xmlReadFile, but I couldn't find anything useful, everybody seems to pass NULL as argument, which I ended up doing too. I'm all for that, I was that close of doing it when I upgraded libxml2 for Windows. As long as it's supporting XPath and XML Tree manipulations, I'd go for the most light-weight one, maybe even header-only. Our libxml2 usage is confined to xmlutil, so it should be possible to switch. edit: I'd vote for pugixml, it seems to be still alive and has XPath support. It's what I've been using in the TDM game code too. But still, if it's the tree manipulations that break the queries, we can either try to find out what we are doing wrong here, or move away from pushing the .game data into the registry trees - it might not be necessary after all, since most (if not all) code is relying on the GameManager interface to get do the queries. For the moment being, I can revert the change to xmlReadFile, so that my repo is functional again. Are you still investigating this? You seem to have gotten quite far already.

I compiled DR with no extra options on Ubuntu 23.10, and installed it, all without any error, but when I try to launch DR it comes out with "no game type selected" and after ok, the code Aborts due to a memory overflow, I opened a bug in: https://bugs.thedarkmod.com/view.php?id=6472 where I supplied the log, any idea of what might be going wrong?

Related to looking into the stray marks associated with Stone font characters... I've noticed that the nominal 12 pt DDS (which as discussed above only kicks in under 8 pt) has only ASCII glyphs... nothing with code points above 127. Maybe that's why the corresponding 12 pt DAT file was not distributed, as a workaround to force the engine to use 24pt instead, because tels never got around to finishing the low-priority aspects of the project. The 24 pt DDS does have the ANSI glyphs. At a glance, most are fine; a few are a bit ragged. I didn't check compliance and completeness with TDM's character map as given on the wiki. I don't feel ANSI improvement is a 2.12 item. I've got a test FM to show all the characters. At some point, after refinement, I'll post a link.

For an as-yet unknown reason, this commit seems to break XML parsing on Linux: #6439: Use xmlReadFile instead of xmlParseFile which has been deprecated and removed. Privatise Document() constructor accepting an xmlDocPtr. As far as I can see, the commit is entirely correct. xmlParseFile is indeed deprecated, and the new usage of xmlReadFile matches what the libxml2 examples are suggesting. But the result is that although the xmlDoc* returned from the function is not NULL, nothing XML-related works, the entire registry system returns only empty values, and almost all of the tests are broken (because the main radiant core cannot be initialised without any registry values available). Changing back to xmlParseFile makes the problem go away but is an unsatisfactory solution because it specifically reintroduces a deprecated function call. I am not sure whether this is a bug in the specific version of libxml2 on my Ubuntu system, or something incorrect about how we are calling xmlReadFile (i.e. perhaps it requires an encoding or a particular non-default option to correctly process our XML files). Unfortunately like many of the core GNOME C libraries, the documentation is bare-bones and explains almost nothing (like what any of the parsing options actually mean), and I cannot see an obvious way to ask libxml2 to return meaningful errors, or to query exactly what might be wrong with a constructed xmlDoc* object. It makes me wonder if it would be better in the long term to ditch the reliance on libxml2 and instead use one of the light-weight C++ XML parsing libraries like RapidXml or pugixml instead. Not exactly a trivial change but might not be too cumbersome since the existing XML code is wrapped in our own xmlutil classes and not generally used directly by the rest of the codebase.

Bumping this thread. I was trying to parse the code for LibreCoop recently, the multiplayer coop mod for Doom3, or Dhewm3 more exactly. The main alternative is OpenCoop, but I think LibreCoop is more developed. Anyway, it got me thinking how much work would have to go into a coop mod for TDM. It's still my biggest wish item. The idea I got was one has to basically walk system by system through the code and think about the client and server side of packet swapping. TDM has a lot more and more complicated systems than Doom3, but once you start getting a feel for it, I think the basic system doesn't change that much. In a way it reminded me a bit of a pared down save/load system, what you need to update a game state, except you're streaming it in in real game-time, and you using tricks to fill in gaps to ease the load. The other thing I noticed is that maps themselves need their scripts tweaked and anything else happening in the world. But I wonder if there's a way to procedurally do that when a map is loading, so one could just use the FM files as released. It looks like it'd take more than a year or two if one were working steadily through it, although I think one would get efficient at it over time. Like I was noticing, there's a consistent logic to it. But most of all I think it'd be worth it. I really like Thief coop, and I think it'd be great for TDM. I'm just FYI'ing about it now because I was browsing through the other coop mods. Not even soliciting opinions or anything. Just thinking aloud (avisible?) about it.

Yes, all images, models and sounds would be gone. Even if you made a barebones replacement that only provides a very limited selection of assets you would need to create thousands of files just to achieve basic game functionality (movement sounds, guard clothes and speech, menus, tools and weapons etc.). It's probably orders of magnitude more work than when TDM got rid of all Doom3 assets for going standalone in v2.0. Technically it's probably possible for an FM to contain a full game's worth of assets, except for the code itself. IIRC some Doom3 mods had custom .dll's to extend the base code, though.

Is "assets" synonymous to "media/gamedata"? And are you referring to the 2.3 GB media/gamedata included in TDM at install? If all 2.3 GB media/gamedata were removed from the "TDB-libre" version, then no license change would be needed. Say then we have a small fan mission that is entirely libre, built entirely from libre assets and created to intentionally avoid using any of the current 2.3 GB media/gamedata. If we wanted to play that mission using only the source code, what media/gamedata components would be missing to do that? * GUI graphics and music? * HUD elements? * Any in-game sounds? * Inventory objects? * ... anything else that can neither be included in the mission's own media/gamedata, nor avoided during mission design? I'm assuming here that a mission actually can include its own media/gamedata (textures, sound, models), but I may be wrong and I'm grateful for any explanation that helps me understand. If you ask me, the TDM-installer works perfectly already today, and the instructions are brief and easy to follow. Installation from the Debian repository would be somewhat easier, but I also see other (perhaps greater) benefits which I mentioned earlier.

Thank you. I have more questions but I would be wasting your time. I will have a look at the source code someday.

As I understand the TDM license there are roughly three types of contributions to the TMD project as a whole: 1) Contributions to the source code of TDM: These are licensed GPL or BSD and can therefore be used already today by commercial projects. 2) Contributions to the 2.3 GB media/gamedata included in TDM at install: These are licensed CC-BY-NC-SA and restricts commercial use. 3) Contributions of fan missions that can be played using TDM and are added by the end-user after the install (either by the ingame downloader, website, or other source): These are not part of the core product and the license says "Any missions [...] are the property of their respective authors, and different licensing may apply.". This means the FM creators can choose any license they want, anything between CC-0/PD and strict copyright. Possibly even put additional restrictions on its use (e.g. say "You may only download and play this on regular TDM"), right? It is up to the end-user to abide by the stipulated license. The included missions "Training Misson", "A New Job", and "Tears of St Lucia" appears to fall into category (2) according to "Unless explicitly stated otherwise, all [...] non-software components that are distributed with The Dark Mod are licensed under [CC-BY-NC-SA]". Does anyone know if their license says anything else?