Search the Community

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/

Mandrasola is a small sized map in which aspiring thief Thomas Porter steals some herbal products from a smuggler. The mission was created by me, Sotha and I wish to thank Bikerdude, BrokenArts and Ocn for playtesting and voice acting. Thanks goes naturally to everyone contributing and making TDM possible. This mission occurs chronologically before the Knighton's Manor, making it the first mission in the Thomas Porter series. Events in chronological order are: Mandrasola, The Knighton's Manor, The Beleaguered Fence, The Glenham Tower and The Transaction. The winter came early and suddenly this year. Weeks of strong blizzards and extremely harsh cold weather hit Bridgeport hard. With the seas completely frozen, a rare occurence indeed, most of the City harbor commerce has stopped completely. Vessels are stuck in the ice and no ship can leave or enter the City, resulting in the availability imported goods declining and their prices skyrocketing. One of these imported items is Mandrasola, a rare herbal product, which is imported overseas from the far southern continents. Mandrasola has its uses in alchemical cures and poisons, but mostly this substance is used for its narcotic qualities by commoners and even the nobility. The problem with Mandrasola is that excessive use is extremely addicting and the withdrawal effects are most grievious. Many are utterly incapable of stopping using Mandrasola and are transformed into quivering human ruins if they do no get their daily dose. And now this expensive and rare substance is running out from the whole City. Me and my fence, Lark Butternose, would love to grab this monopoly to ourselves: selling the last few doses in the City would probably be worth a fortune. According to Lark's sources, there remains only one smuggling lord who still has Mandrasola in stock. The problem is that this individual maintains an exclusive clandestine operation and only supplies a few nobles. Despite our best information gathering efforts we couldn't learn who the smuggler is and where he or she operates. Luckily we have an alternate plan. While searching for Mandrasola related information, we learned that a noblewoman called Lady Ludmilla is addicted to the substance and has paid high prices for small amounts of it. We also know that she has visited frequently someone in the Tanner's Ward waterfront, and since she goes to the area personally we believe she is visiting the smuggler. The plan is simple: I must monitor Ludmilla's most likely entryway to the Waterfront and then follow her to the smugglers hideout. I'd better be very careful around Ludmilla. She must not realise I'm following her or she probably won't lead me to her dealer. Hurting her is also out of the question. After she leads me to the smuggler's hideout, I can take my time to break in carefully and steal all the Mandrasola I can find. While I'm there it wouldn't be a bad idea to grab some loose valuables as well. I've now waited in the blistering cold for a few hours already. Looks like there are a few city watch patrols in the area to complicate matters... I think I heard a womans voice beyond the north gate. That must be lady Ludmilla, I haven't seen many ladies in these parts. I'd better get ready.. Links: Use the ingame downloader to get it. WARNING! Someone always fails to use spoiler tags. I do not recommend reading any further until you've played the mission.

@chakkman Thanks for the detailed feedback, I appreciate it. Your points are well taken. Yeah, I do have a problem with rectilinear design - I just don't think in curves and angles! As for good deeds, my missions usually some form of them. Always optional - don't want to force you to be nice!

Sorry for digging this out, but, one thing I always wondered: Shouldn't a gamma value of 1 in Dark Mod give you the best experience when you're using a monitor with a gamma value close to 2.2 (which I think is the optimum in terms of realistic colour reproduction)? I have a monitor with very good out of the box colour reproduction, and almost ideal gamma. Yet I have to turn down the in-game gamma to about 0.9 to get a good result. I must say though that I prefer it darker than what the average player probably has, so, that might be the problem here. Still, I usually have very good result leaving the gamma on default setting in most games, which is probably calibrated towards monitors which display ideal gamma.

Yeah, I don't know what to say. Rev 10383 works fine and 10384 does not. It is always reproducible for me, meaning that every time I try to reproduce it, I can reproduce it. After some more testing, I found that r_multiSamples set at 2 or 4 has the glitch most often at that viewpos. When set to 8 or 16, it is still possible but less likely. Setting r_fboResolution 2 made no difference. If com_maxFPS is set to something lower, such as 30, the glitched frames display for a longer period of time. Yeah, I also wonder if it driver dependent or GPU dependent. I discovered the glitch while using graphics drivers Mesa 22.0.5, so I updated my OS and graphics drivers to Mesa 23.3.2 today as a test, and the problem is still there. It's strange that rev 10384 exposed this problem. My machine specs and settings: Linux, Ubuntu 22.04 AMD Ryzen 9 5900X AMD Radeon RX 6700 XT (Mesa 22.0.5 & Mesa 23.3.2 both exhibit the glitch) com_fixedTic 1 com_maxFPS 60 Anti-Aliasing 2x (r_multiSamples 2) The video I shared shows the viewpos. Even at the same viewpos, the glitched frame will pop in and out. I've only seen this glitch happen in this single room in Accountant 1. If I move around the room at that viewpos, I can find other view positions that also expose the glitch. Here's another viewpos for Anti-Aliasing 4x, 8x, and 16x. It requires moving the mouse slightly to get into the exact viewpos to expose the glitch. (1.7 -166.4 0.0) is not precise enough. -1460.29 1425.48 172.25 1.7 -166.4 0.0 Glitched frame (Shows the curtain texture, covering most of the frame. Also, a hint of the red carpet texture?): Good frame (Same viewpos and Anti-Aliasing settings, when the glitched frame disappeared for a moment.): (Before and after rev 10384) I wonder if this might also be related to: (Before and after rev 10384) Also, sometimes on the loading screen with Anti-Aliasing enabled, I get a glitched background, which does not show with Anti-Aliasing disabled. See left side of screenshot: I don't know if the last two examples (distorted first frame on launch & loading screen distorted background) are related, but I thought I'd mention them just in case it helps track down the issue.

A Problem Arises I've paused subtitling of the Lady02 vocal set, because of a problem with the voice clips described here: https://forums.thedarkmod.com/index.php?/topic/21741-subtitles-possibilities-beyond-211/&do=findComment&comment=490151 While a way forward is being determined, I'll work on a different vocal set. Maybe manbeast, for which Kingsal just provided me the voice script.

It's easy to imagine how #1 can be solved fairly: When the player or an enemy AI shoot an armed AI, the AI that was hit should be given a random destination near the attacker to run to. The further the player is from the target, the larger the radius around the player in which a random node may be picked; If you're far away the virtual sphere should be large, meaning AI may run to a room or passage near you but still in your general direction . This new behavior should be influenced by AI acuity and difficulty settings though I'm not sure which ones... also toggled by one, for those who prefer the old way and abusing clueless guards as they wonder what force of the universe must be hitting them. By the way, and I just realized this now: The problem doesn't only affect attacks, but also spotting the player before attacking. Think of the last time a guard standing outside spotted you on a balcony before you hid, enough to draw their swords and start searching but not attack you outright; The guard was alerted by something in a building, yet he draws his sword to search the area near him. If I'm standing outside and think I see a burglar in my house through the window, I'm going to go near my home if I plan to confront the possible intruder, not start looking around the road next to me and say "I know you're here somewhere" which would imply I have a serious problem Main reason I mention #2 is when I helped with transcribing the subtitles for the drunk voice, I noticed there are voice lines for AI returning with an ally. Two in fact: If the player is still there the AI has a voice line for telling the other person "that's the one let's get him", if the player hid there are lines such as "you were too slow now he got away". Those voices in particular are never played, or at least I don't remember ever hearing them in all those years: It occurred to me while I wrote this post that they're likely designed for a mechanic that was never implemented... or was but broke very long ago? Curious what its history is.

If the "mission fails as soon as stealth score turns non-zero," that would not be good for ghost players. They might need to find out "how" they failed and experiment to avoid alerting guards. They might need to take those score points as a "bust". They might need to take those score points to complete an objective. Then, mission authors would need to encode exceptions into their missions, which would be a lot of work (if they decide to do it at all). However, part of what makes ghosting challenging and fun is when mission authors do not create their missions with ghosting in mind. Please see: Official Ghosting Rules: https://www.ttlg.com/forums/showthread.php?t=148523 Writing code for these rules would be a huge undertaking. Ghost Rules Discussion: https://www.ttlg.com/forums/showthread.php?t=148487 Creating an official mode could alienate these dedicated ghost players, because it would clash with what is considered ghosting in the community. Including the Stealth Stat Tool mod in the official release would be more useful. Or, making the audible alert states of guards quick and easy to recognize could help as well. For these reasons, I don't agree with an official "Ghost" mode. If the dev team were to do it, we should consult with @Klatremus so we get it 100% correct or not pursue it at all. (This ghosting bit should probably be in its own thread.)

So id recently started dablling with linux derivatives as a main OS again after holding off for years due to some problems running my hardware when i stumbled upon Solus. Solus is a distro that was developed independently from mainstream linux (from the ground up) and uses its own desktop model called budgie (also comes with a plasma gnome and XFCE dektop). Solus uses its own package manager called eopkg and it handles flatpak and snap as well. I been very impressed with it so far as things just worked out of the box for most parts, only real problem i had was getting steam running which turned out to be due to my machine loosing internet connection when it was installing it which left some broken files and after an uninstall / reinstall it worked quite nicely. one warning though Solus comes with a linux steam integration tool which while cool has a rather annoying bug if you try to use the intercept library no games will launch (well alledgedly one will and it was made for that because it wont launch without it on groan...), so do not enable that setting. lutris also works pretty much from the get go though i had to do some reading to find out how to get it to synch up with steam (steam profile must be set to public in pretty much everything). So how does games work on it, well actually rather well proton has made strides in getting even the newest games to run on steam and the only games i so far have not been able to run is the division 1 and 2 but this a'int a problem with steam but with ubisoft connect which crashes and might get fixed in the future (damn launchers...). So far i been plaing baldurs gate 3, callisto protocol, crysis remastered trilogy, DOOM, starfield and a wealth of older games with it and no problems besides having to run baldurs gate in DX11 mode (not really sure why the vulkan driver is not working for this game on linux ???). I might actually step onto the linux platform for good when win10 goes out of support, and ill probably stay with solus since it is less cluttered than most other distros and main packages are built explicitly for it which again means stability as you dont get puckered into installing something which was newer meant for your OS (yeah i f.... up numerous linux distros over the years because you can install pretty much what you like even if it breaks everything, not so with Solus). This does not mean you cant have some nice app you need which was built against say gnome because you sit on a plasma desktop, you absolutely can but they are only avaliable as snap / flatpaks so each of them run in its own sandbox with all the dependencies only avaliable to said program and keept isolated from the system. try it out :) https://getsol.us/download/

You're right, libxml2 can't be fundamentally broken since it is such a core dependency on Linux, plus the exact same XPath queries work perfectly fine even within our own XmlTest, so it must be something specific to how XML is being used within the registry setup. I found a couple of online sources which suggested that XPath queries might fail if the DTD doesn't validate, but we don't have any DTDs in our XML files and the XML_PARSE_DTDVALID option is not set by default in any case. Encodings are another possible culprit (especially since the problem seems to be OS-specific), but as far as I know encodings wouldn't change the parsing of characters like "[" or "@", and if the encoding was so fundamentally wrong that even regular 7-bit ASCII failed to parse, then how would "//game" ever be found? It seems that whatever the cause of the problem is, it is very well hidden within the API and would probably require building libxml2 from source and diving into it with the debugger to find out what is going wrong. Right, that sounds like the best approach. I'll set the HEADER_ONLY flag on Linux too then for consistency, and integrate it as a header-only dependency like libfmt.

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.

To cater to both audiences. I mentioned LibreGameWiki as one example. nbohr1more mentioned other uses. Explicitly allowing reuse and spread will help TDM reach a wider audience and would hopefully attract more volunteers. More volunteers which can help improve both TDM versions. There are several benefits for a project of being in the Debian repo. One is that TDM Debian-users can report defects on any package directly to Debian (no need to register on separate forums). Debian may then fix the issue themselves (in their "TDM-libre" package) and will offer the patch upstream to TDM, who can then choose to accept or reject the patch. I envision "TDM-libre" to have the same capability of downloading any mission as regular TDM. The only difference is that "TDM-libre" would come packaged with the regular engine (which is GPL+BSD) and an included mission that has libre media/gamedata. When I play TDM by myself, I want the unlimited-play and can accept commercial restrictions. But if I were to promote it somewhere, or charge for a stream when playing online, or make a video, I would want a version without commercial restrictions (and can temporarily accept limited-play) to make sure I don't violate anyone's copyright. Perhaps. That's what I'm trying to find out.

Looks like we have one physics based regression? Linux users get stuck at the start of mission "One Step Too Far" : Bisecting... mission works on 2.12 dev binary 10370 works on 10455 broken on 10518 works on 10488 Something between 10488 and 10518 caused the problem Compiling binaries: Rev 10498 is also broken 10491 works 10494 works 10495 works 10496 works 10497 works 10498 still broken. WINNER If I revert this change, the problem is fixed: original : vel += (vel * gravityNormal) * gravityNormal; 10498 : vel -= (vel * gravityNormal) * gravityNormal; Maybe ifdef it for Linux @stgatilov ?

I suggest you use the term "I", to make clear that it is something YOU want, and that you speak for yourself. But, as wesp5 mentioned, I don't really know what this is about, at all. And, I'm also wondering about all the newly registered people lately, who just arrived at this forum, and already want to revolutionize this mod. This is a thing I noticed 2 or 3 years ago, and which hasn't been present in the 15 years I play this mod and frequent these forums now. Really seems like a common thing these days, to not knock on the door, but kick it in, and stomp right in.

This is a different issue than the click of sound at the start of every transition that I was talking about in earlier posts that you're referencing. This may be a problem that's always been there but nobody really noticed or commented on it, or maybe only for certain ambients? So the first problem is that the ambient starts abruptly at game start presumably because the sound system isn't online yet to process the fade in, and it just starts the ambient at full sound. So then you added a delay, but since you're not coming off of a previous fade out to sets the volume to zero, it plays the full volume ambient through that delay and then cuts to zero and fades in. My idea is that if you want a fade in from the start, don't start with that ambient as it is. Start with a .1 second of silence, and then fade in the ambient. There are two ways I can think to do that. (There might be other ways.) 1. Some ambient soundshaders start with a "leadin" spawnarg that starts it with a bit of silence, originally used to stop the click of sound (soundshaders ending in "_z", there were two versions, one with and one without the _z, unless that was later changed). You could confirm that the soundshader for that ambient uses a "leadin" arg, or if it does but it's not long enough, maybe use a custom soundshader to use a longer leadin blip of silence. (In that case you'd have to make the longer silence file yourself.) 2. Or another possibly easier way may be to start with an "override" sound, look at the Sound Override part of the tutorial with silence, and then after .1 second or whatever, transition that to the normal location-based ambient according to the instructions, and then the override sound will fade out and yours will fade in. I think you'd do that by setting the override arg to "1" from the start and then change it back to "0" after 0.1 seconds using Target_setKeyVal as the instructions say. Confirm that the "silence" sound is stereo. I noticed if you start it with a non-stereo sound, the system stays non-stereo. I think that was fixed, but good to double check. If it isn't, then you could do the same thing with a normal ambient that just starts off silent or very quiet for that first .1 second.

I am trying to install the latest version of TDM, I think the last version I played was 2.09a. I am installing to a new folder My user created the folder and has full write access to it I am running the extracted .exe file in this new folder Problem 1: Why is the installer using the .exe filename as the URL for updating? This is absolute nonsense. I name my installers by the date or version number of the installer. This wasn't a problem in the past for TDM, or for any other thousands of installer I've used over the past 35 years. Never once have I seen an app look for a local file or external URL based on the name of the installer executable! According to https://wiki.thedarkmod.com/index.php?title=Tdm_installer_and_zipsync#tdm_installer about the installer "The config file is automatically downloaded from hardcoded URL. It describes all official versions of TDM and mirrors used to download them." Clearly this is not the case, it is literally using a URL based on the executable name! I should not be forced to keep a stupid name like tdm_installer.exe that doesn't even specify the version number (in this case v1.11). I have the previous installers in my downloads folder named/dated appropriately. The installer won't even work if you change the case of the executable to title case "TDM_Installer.exe"! It's ridiculous. Problem 2: Why is the install directory listed in red? My user created the folder and has full write permissions to it. The installer has no difficulties creating log files in the folder, so it obviously has write permissions. Problem 3: Why does the installer not provide any way to specify a new folder to install to? There is no "New Folder" button anywhere in the GUI. I'm not sure if typing in a folder name works or not, because it remains red either way, and won't let me continue the install with permission problems. But regular users should not be required to know this hack. Also, when you select a folder, the installer does not add "The_Dark_Mod" or whatever as a default folder name like most apps do. It is left entirely up to the user to create an install folder beforehand. Kind of defeats the purpose of an installer being quick and easy. Problem 2 - Follow Up: I am still unable to install TDM after renaming the .exe to the original lowercase name. It still complains that I don't have permissions to the folder, but obviously I do! tdm_installer.exe is creating log files in the folder while telling me that it doesn't have permissions to write to that very same folder! It's beyond absurd. Why is the installer lying? I have solved the issues by changing Security settings so ALL LOGGED IN USERS have full write permissions to the folder, but that is not desired -- why is that necessary? My user already had full permissions to the folder when creating it, and the installer had write permissions to the folder. Possible Solution For Problem 1: The installer should never use the name of the executable for anything to do with local filenames or external URLs. Update the installer so it does what the Wiki says it does -- uses a fixed URL to get the list of TDM versions. Possible Solution For Problem 2: The installer should be updated so it automatically changes permissions to the install folder as-needed. Requesting UAC privilege escalation if needed to alter the folder permissions. At the very least it should tell the user what permissions it needs for the folder (it already had full write permissions!). Heck, just make the installer ATTEMPT to create a file to see if it has permissions. Clearly it is checking permissions incorrectly and misinforming the user about it, then refusing to continue. Possible Solution For Problem 3: Have a way to create a new folder from the installer GUI, such as an "Add Folder" button or a right-click menu with "New Folder".

Greets. I just setup VS 2022 Community/VC++/MFC on a fresh box, and did a Checkout with TortoiseSVC from the public repository of build 16481-9881 (matching a fresh 9881 dev install of TDM environment, as required). When I do the build, I get a link error: Can't find file libavcodec.lib I'm not finding that file in the working repository, although I do see the libavcodec folder and avcodec.lib If I select DarkModTools/Properties, I'm able to go to the Link/Input field and see where libavcodec.lib is listed, but I can't edit it to change it to avcodec.lib (says it's inherited) I'm kinda rusty, haven't used VS since 2015, so this may be a trivial thing. Help?

TDM has tons of textures from "free" texture resources that do not allow redistribution and cannot be incorporated into a commercial project. Someone would need to create a huge replacement pack of textures that do not break the look of existing missions and do not infringe on the copyrighted textures. Also, many artists who contributed to this project do not want 3rd party entities to use their work in commercial projects. They intended the models, textures, sounds, animations to be exclusively used for Darkmod content. You would either have to replace ALL assets or contact every contributor and ask them to re-license their assets. Many contributors are no longer active with the project and haven't visited the forums in years so it would be no easy feat. I cannot speak to Debian policy but I think that they treat installers that add non-free content the same as non-free content itself. One could argue that Steam is such an installer but I guess Debian would counter that there are a few fully Libre games on Steam. I think Debian, Ubuntu, or Linux Mint need to consider a repo that allows for games (etc) that include non-libre content but intentionally offer this content for free to the community with no stipulations other than "don't try to sell it as a product".

The gamepad implementation allows for a great degree of flexibility to personalize settings, aside from a few minor issues that I mentioned here: https://forums.thedarkmod.com/index.php?/topic/22337-gamepad-bindings/ I would say that playing TDM with a gamepad works very well, especially considering that it was implemented as experimental and hasn't been changed since then. If I could, I'd go back to 2021-you and congratulate you on buying that gamepad. I notice that your DarkmodPadbinds.cfg looks very different from mine...

It seems like more and more "thief" and "thief players" is becoming a short hand to dismiss community members earnest desire to improve the game - which happens to be a barely legally distinct "thief style" game which was made by thief fans for thief fans and is "designed to simulate the stealth gameplay of Thief". Who is the predominant player base of the game supposed to be beyond fans of the thief games? Is there some better avenue to find feedback for the game beyond this forum? FOSS and linux forums? I have seen maybe half a dozen posts from that segment. I am a thief fan, I play thief fms, my association with those games is what drives me to play and make things for this game. Are we supposed to pretend the original games are not a huge reason why most of us are here at all? TL;DR version:

I can confirm the problem in Ubuntu 23.10, and the solution proposed "GDK_BACKEND=x11" i confirm that to solves it partially, the clipper does bring back the problem.

Thanks! 1) Doing LONG_PRESS PAD_A (what I, for lack of knowledge, call "jump-mantle" or "_jumpmantle") differs from doing PRESS PAD_A ("_jump"). "_jumpmantle" differs from "_mantle", so they must be mapped to different button-calls. "_jumpmantle" differs from "_jump", so they must also be mapped to different button-calls. This appears to be the case, but it is not evident (or changeable) in DarkmodPadbinds.cfg. "_jumpmantle" seems to be hard coded to always connect to the same button as "_jump" but with a long press. It is as if bindPadButton PRESS PAD_A "_jump" is not actually just binding PRESS PAD_A to "_jump", but rather interpreted as "link PAD_A (regardless of button press time) to behave exactly like keyboard SPACE for short and long presses". I would have expected the default DarkmodPadbinds.cfg to explicitly read: bindPadButton PRESS PAD_A "_jump" bindPadButton LONG_PRESS PAD_A "_jumpmantle" bindPadButton PRESS PAD_B "_crouch" bindPadButton LONG_PRESS PAD_B "_mantle" ... but neither LONG_PRESS PAD_A or "_jumpmantle" is listed in the file. If there are actions "_jump" and "_mantle", I suppose there must also be an action "_jumpmantle" since it is possible for the player to do all those movements: * "_mantle" does the movements "crouch on the high surface, then stand up" * "_jumpmantle" idoes the movements "jump slightly forward, then land standing on the high surface" * "_jump" idoes the movements "jump up, then land exactly where you started" If the actions "_jump" and "_moveup" are not synonymous, then perhaps the action "_moveup" is what i call "_jumpmantle"? 2) Thanks for the link! It was useful in more than one way. I'll link to that page from https://wiki.thedarkmod.com/index.php?title=Bindings_and_User_Settings#Gamepad_Default_Bindings if I can get an account on the wiki, which proved more difficult than i thought (https://forums.thedarkmod.com/index.php?/topic/22327-how-can-i-create-an-account-on-the-tdm-wiki/). However, it does not answer my question how to find out the name ("<button>") used for a button on my gamepad. Basically, I would need to press the button on my gamepad and some program could tell me "That button is called 'PAD_A'". In my case, I have a gamepad "Logitech F310" (https://commons.wikimedia.org/wiki/File:Logitech_F310_Gamepad.jpg) which has a "Logitech button" (see image) that I want to use. I was hoping to find out the "button name" for that button and then edit DarkmodPadbinds.cfg to map it to a function. 3) ... but if that button has an "unusual name" that TDM does not recognize, then it may perhaps not work. E.g. if that button is called "PAD_LOGITECH" and TDM cannot recognize that name, then I cannot map anything to it via DarkmodPadbinds.cfg. Using QJoyPad I can map any keyboard key to it instead, as a workaround, but I cannot map MODIFIER to it (since MODIFIER cannot be set to a keyboard key). If current implementation is still called "experimental", then I must say it works very well; @cabalistic: kudos for that! I may not have continued playing TDM had it not worked with a gamepad.

I did some debugging in unit tests. The first problem is that although we have a basic XmlTest, it uses the full RadiantTest fixture which can only be constructed if the XML registry is working fine, so these basic XML tests are not runnable. I managed to fix that by changing the behaviour (on Linux only) to use TEST_BASE_PATH instead of _context.getTestResourcePath() to find the test resource files, so that RadiantTest is not required. This confirmed that the basic functionality of loading XML is working perfectly fine, even with the switch to xmlReadFile(). All of the XML tests pass, and I can load one of the game files in a unit test and examine its properties. So there is nothing fundamentally wrong with the XML structure being created by the new function call. The problem seems to be that within the Game class, any attempt to look up key values in the registry fail. Although each Game class is constructed successfully and imports its content, any searches for its own XPath root (e.g. "//game[@name='Doom 3 Demo']") return a list of 0 nodes, even though that exact XPath string can be used successfully within the basic XML test to find the <game> node which was loaded directly into an xml::Document. So there must be something going wrong with either when or how the .game file content is being merged into the global registry hierarchy.