The End Is Nigh

[BlueSkyBullet]

That has become so fucking aggrevating that I promised myself to throw in the towel every other day for the last fucking months... I even mucked around a few hours to try to fix that probem myself (and I have no idea about Scons at all).

 

This is ridiculous indeed. And alarming enough that I would be be willing to look into it, but I do not want to do so if it would be a waste of my time. If you have been struggling with this for months, it is a shame.

 

Please do create another thread and if someone versed in Scons doesn't come along soon enough I will become versed in Scons and get this working the way it should. No reason to have to compile in some sort of transfer here and there roundabout. No reason to have to recompile the whole project with every code change.

[BlueSkyBullet]

Well already with SCons I'm seeing evangelical hyperbole, some more level-headed praise, and some enthusiam dampening mixed messages (eg. the SCons wiki speaks of "Built-in support for fetching source files from revision control systems, such as SCCS, RCS, CVS, Subversion, BitKeeper and Perforce." ; while the SCons official site says "Built-in support for fetching source files from SCCS, RCS, CVS, BitKeeper and Perforce"... you'll notice something important missing from the latter?"

Edited April 11, 2012 by BlueSkyBullet

[Tels]

This is ridiculous indeed. And alarming enough that I would be be willing to look into it, but I do not want to do so if it would be a waste of my time. If you have been struggling with this for months, it is a shame.

 

Please do create another thread and if someone versed in Scons doesn't come along soon enough I will become versed in Scons and get this working the way it should. No reason to have to compile in some sort of transfer here and there roundabout. No reason to have to recompile the whole project with every code change.

 

Created thread:

 

http://forums.thedarkmod.com/topic/13702-fixing-the-tdm-build-system/page__view__findpost__p__283737

[DeTeEff]

Sorry to postulate more work for you Tels but I had a similar thought about map compiling.

 

It takes far more resources to compile a map than to play it. Melan is currently hampered from doing campaign work

because the maps won't compile on his system and he must pass his work to Bikerdude to dmap.

 

If there were a Dmap server this would solve this dilemma and would open the way for even more intensive compile

types and optimizations.

 

Since a "hosted dmap server" is something that would benefit all Id Tech 4 projects, I'll float this concept over

at the iodoom3 IRC room and see if there are any takers

 

Does this mean that you'll sit at home mapping and then ever once in a while upload the map to a server that dmaps and compiles an sends the proc etc back to you?

 

As a sidenote, I WILL release my project, dead mod or not!

[nbohr1more]

Yes, that is the concept.

 

If people ever get mod support for Id Tech 5 they would need a similar setup to compile the Mega-Textures.

 

Not that I'm a big fan of the paradigm but there is this creeping consensus that you can do lots of large scale

pre-compiling and have client systems run content that is way better than what you would expect to be possible.

 

Even without crazy stuff like Mega-Textures (etc) plain old Dmap compiling can be made more effective by allowing

more recursion and iteration. If there were some way to distribute Dmap compiling to a cluster then participating TDM

players could be like Folding@Home. They could grant spare cycles to help Dmap large projects (that's pretty pie-in-the-sky

though...)

[DeTeEff]

Would that concept be feasible when rendering stuff like movies/video also? Or in any process where this server renders the data faster than the upload time...

Edited April 11, 2012 by Fieldmedic

[nbohr1more]

I suppose so. The main reason for this line of investigation is that the dmap code can just crash on desktop PC's without enough RAM address space (or enough RAM). The 64bit patch can overcome this to some degree but as maps get larger new limits may be hit. It'd be cool if chunking and disk cache could be used for lower-end configurations (though that would be pretty slow...).

[MoroseTroll]

I'm not a mapper, so please let me ask you, guys, a silly question: how much time dmap takes on your CPUs when you're editing your maps, and how often do you execute this command?

[Sotha]

I think dmapping takes roughly about the same time it takes to load a mission. Few minutes.

Dmapping has to be done each time you want to see ingame what you've done, I do it very often. I change something like 10-20 things in the map, the go ingame to check it all works.

[Melan]

That depends on the size and complexity of the mission. For something small and relatively simple, maybe half a minute. For a medium-sized mission, a few minutes. For large and heavily detailed missions, it can take 15-20 minutes, and there are memory-based hard limits - my main campaign mission would not compile before I upgraded my RAM, even if I shot down all concurrent processes like my browser etc.

 

But that's size and complexity. There are missions which are very large, look excellent, but compile without a fuss. It is not terrain brushes which pose the major problem, but entities (interactive or sorta-interactive elements which includes lights, moveable objects, AI, sounds, but also certain kinds of detail architecture). A mission with 20,000+ brushes and/or patches will compile just fine, but a mission with 13,000 brushes and too many entities (like trying to make every single plate and spoon and fork a moveable) will just fail on a weaker PC. A mission requires fewer resources to run (MAP) than to compile (DMAP). But there are still limitations for weaker configs. These are not limits you run into very often - they crop up in missions which get really, really big.

 

As for the frequency of DMAPs, I tend to do it once after every editing session to check out my work, identify bugs as soon as possible, and come up with more ideas I can use.

Edited April 12, 2012 by Melan

[Tels]

Yes, that is the concept.

 

If people ever get mod support for Id Tech 5 they would need a similar setup to compile the Mega-Textures.

 

Not that I'm a big fan of the paradigm but there is this creeping consensus that you can do lots of large scale

pre-compiling and have client systems run content that is way better than what you would expect to be possible.

 

Even without crazy stuff like Mega-Textures (etc) plain old Dmap compiling can be made more effective by allowing

more recursion and iteration. If there were some way to distribute Dmap compiling to a cluster then participating TDM

players could be like Folding@Home. They could grant spare cycles to help Dmap large projects (that's pretty pie-in-the-sky

though...)

 

I'd say wo a crazy fast internet connection it will take longer to upload/download several megabytes than it will take to dmap the map. Somebody would have to run the numbers, but IIRC SL was already 10Mbytes compressed, and you'd to upload the 2..5 MB (whatever it is) compressed map file before you can download the finished .proc, .cm etc files.

[Xarg]

I'm not a mapper, so please let me ask you, guys, a silly question: how much time dmap takes on your CPUs when you're editing your maps, and how often do you execute this command?

 

Takes me 30-40 seconds with my map. Out of curiosity, I pulled the Heart of Lone Salvation .map out of the .pk4 (it's the most complex I can think of off the top of my head), and tried to dmap it, doom3 crashed. Applied the 4gb patch, and compiled it in 2m20s. I personally do my dmapping whenever I finish a bunch of stuff I want to look at, so that could be every 2-3 minutes (when trying to fix something) to every 2 hours (when I want a quick drink/smoke break and have nothing else to do with the pc while I'm afk), then a playtest usually follows to make sure what I did works, and back to mapping.

 

Correction, takes 17s, just timed it.

Edited April 14, 2012 by Xarg

[nbohr1more]

Yes, if the purpose of a dmap server is to speed-up the dmap then that won't work-out.

 

My concept was for doing things that "can't" be done on a home PC or older hardware. I'm pretty sure that dmap

could be made to be more forgiving and utilize disk cache on lower-end systems (etc) so the term "can't" is

relative to the skills of whoever looks into the problem.

[Xarg]

Well as my HoLS dmap experiment showed, and what I know of some of the hardware still floating around in the interwebnets, there *could* be a use for such a thing, if the mapper in question's machine just simply cannot dmap their work. However, that number of people may be quite small, and possibly not worth the effort of setting up such a thing for just that function. At worst, a list of people with solid machines that could do the dmapping could be created.

[Ken_Joshima]

Hello there!

I am new to this whole "stealth-gaming-thing". I didn't play the original Thief series, but after seeing those screenshots on your web page and the many available missions I want to give it a shot. I don't exactly know what to expect from this mod, but it looks very promising.

I hope your concerns about that it could be "dying" are non-valid and you continue building it up.

[SiyahParsomen]

welcome Ken. You are very late to party, but it's not over yet. Enjoy as much as you can.

[simplen00b]

Welcome Ken. From personal experience - if you've never played a stealth game before, you may have a bit of a learning curve ahead of you, but stick with it and you'll be rewarded with one of the most immersive game-playing experiences around. Start with the training mission (duh), then the recommended newcomer missions (The Crown of Penitence, The Outpost, The Parcel, and Too Late; I'd also add Special Delivery). If you haven't already checked it out, the wiki page on Dark Mod gameplay may be useful (I'd definitely recommend the sections on combat and AI - it'll give you an idea of what you're letting yourself in for!)

 

You are very late to party, but it's not over yet.

Seriously - I think TDM's glory days are still to come.

[Ken_Joshima]

Thank you for the welcome and the advice! I already played the tutorial mission. Once you get used to the controls and if you play carefully enough, it's possible to make it through. But it isn't as easy as I thought it would be. o__o

Now I need some excitement! I saw there are new maps for beginners in the Fan Mission forum. Gonna get my first real Dark Mod experience..

[demagogue]

Those FMs are actually 'made by' beginners. It's a contest for them. But the maps are so good you wouldn't know it. It's as good a place to start as any though. Then you can participate in the contest by playing them, voting, and joining the discussion. (My first FM experience at TTLG was actually Komag's Contest 3).

 

I wouldn't worry about TDM dying in the sense there's going to be FMs getting made for a long time to come. It may have ebbs and flows of getting updated in versions, but that's different than the mappers making FMs. The version it's at now is already a very advanced game, so we don't really need more for the FMs to keep coming.

[zergrush]

I believe TDM will spring back to an increased activity phase once the standalone version is complete. By the way, has the team ever considered applying for next year's Independent Games Festival awards? It might be a good way to get some well-deserved attention.

[grayman]

Welcome, Ken!!

[Mechs Delight]

Hello there!

I am new to this whole "stealth-gaming-thing". I didn't play the original Thief series, but after seeing those screenshots on your web page and the many available missions I want to give it a shot. I don't exactly know what to expect from this mod, but it looks very promising.

I hope your concerns about that it could be "dying" are non-valid and you continue building it up.

 

If you have the time and money, the first two Thief games are available at GOG.com. And there is a huge library of fan missions available at several websites; you'd be in for a real treat.

 

- David

[lost_soul]

While digging through other threads, I read that Greebo is leaving. So, I would like to take this opportunity to say thanks for your contributions to TDM over the years Greebo. While I do not have a list handy of all issues you've addressed, I know that you are a coder and have no doubt fixed lots of bugs which would otherwise still be annoying us players. IIRC he fixed the save game corruption/crash issue way back when.

Edited April 23, 2012 by lost_soul

[PPoe]

He is not leaving, he just stopped to be lead coder. Anyway, thanks for his great work are really earned. So Thank you (and the whole TDM team also=) ).

[nbohr1more]

Um...

 

Lost_Soul, you need to look at the bug tracker entries more closely. Greebo is practically TDM incarnate. He's also been responsible for the heavy infrastructure work like the release mechanics and about... oh.... 99.9999999999 percent of Dark Radiant's development.

 

If you love TDM, you love Greebo's work.