k8vavoom: no good thing ever dies!

ah. and of course, i found a weird bug (several) right after pushing the build out. how did they slipped... nevermind. expect a quickfix build soon (just give me a day or two for better testing this time... i hope).

On 7/2/2020 at 1:13 PM, Cacodemon345 said:

Ketmar uses a Nvidia card. Maybe an issue with your drivers? Remember that he uses Linux too.

I should be on the latest drivers. Might be because they're the "Game Ready" drivers provided using GeForce Experience rather than reference drivers?

On 7/2/2020 at 5:36 PM, Gunrock said:

I had that problem too. Strange however, it was fixed when i updated my gpu drivers and downloaded the latest K8Vavoom build.

Not certain on whether this is the latest build however. Is "k8vavoom_550627.7z" the correct download?

3 hours ago, Cambertian said:

I should be on the latest drivers. Might be because they're the "Game Ready" drivers provided using GeForce Experience rather than reference drivers?

sorry, i don't know. any version should work, but... i simply don't know.

3 hours ago, Cambertian said:

Not certain on whether this is the latest build however. Is "k8vavoom_550627.7z" the correct download?

yeah, the first post always contains the latest one.

still, i found and fixed some bugs in the code (and one bug in psprite rendering; probably unrelated, but who knows...), and i'm planning to publish that "quickfix" build very soon. maybe it will help.

Hi,

I recently updated my system, new CPU, motherboard, mem, video card and HDD, well actually it's an SSD.. Never the less, since then when I run k8vavoom, the player HUD graphic animates erratically on player movement: 

35 second video ~ 123mb

Strange huh?

Even though it looks to be some sort of joke wad/mod where Doomguy has had waay to much coffee, I assure you it's not.

It's the first time I've seen this, all other games and ports don't have this issue.

I've tried older versions with the same result. In the video I'm running the latest version out of the box with Doom2.wad.

Note* I too am running an Nvidia video card: 1660 GTX Super. ~ latest drivers.

Also, I'm running everything stock, no overclocking or anything like that going on.

O/S = Windows 10 Pro 64bit.

Let me know what you think.

@Mr.Rocket yeah, that was already reported above. thank you, now we know that there are at least two people with this bug. sadly, this is the only info we have. ;-) this is absolutely impossibiru (and i cannot reproduce it). as i said earlier, there should be either no dancing sprites, or all sprites should dance, because monsters and psprites are rendered with the same code. i haven't the slightest idea. that's one of the reasons i'm delaying the promised "quickfix build" -- i'm still trying to understand what could cause that caffeine overdose. ;-)

still not dead. but not much happens in k8vavoom now, sorry. that "quickfix" turned out to be very quick, amirite?

anyway, don't worry, nothing wrong happened with me or the project. i just sidetracked by some nostalgic 8-bit coding (a game in 9KB, with Z80 asm? yeah, i can do it the whole day! ;-).

consider this as a small "vacation". i am as full of ideas as ever. but there are alot of interesting projects around (including mines), and only 24 hours per a day. ;-) k8vavoom is still the main focus, though.

tl;dr: polyamorous vacation with another programming projects. will return soon.

No worries. ;)

I need to get back to working on that launcher too!

Just wanted to say great work on K8Vavoom!!! Quick question, voxels models work with K8Vavoom right?

On 7/8/2020 at 6:10 AM, ketmar said:

@Mr.Rocket yeah, that was already reported above. thank you, now we know that there are at least two people with this bug. sadly, this is the only info we have. ;-) this is absolutely impossibiru (and i cannot reproduce it). as i said earlier, there should be either no dancing sprites, or all sprites should dance, because monsters and psprites are rendered with the same code. i haven't the slightest idea. that's one of the reasons i'm delaying the promised "quickfix build" -- i'm still trying to understand what could cause that caffeine overdose. ;-)

Yeah, so far it seems to only affect the HUD player sprite. I'll let you know if enabling or disabling different things will change anything.

eg: lowering the res, testing with different offsets, changing the FOV, etc..

Thanks

@Gunrock, Not that I've noticed. I know normal models work, most of the time.

On 8/12/2020 at 4:37 AM, Gunrock said:

Quick question, voxels models work with K8Vavoom right?

sorry, but no, they don't. they're real PITA to render, because they generate ALOT of triangles. unless i switch to OpenGL 4+ with tesselation shaders, i don't think that i will implement them. this is one of those cases where something that is good for a pure software renderer is really-really bad for a hardware one.

just to give you some insight: each voxel generates two triangles for each visible side of the cube. so if you'll create a small box, let's say 3x3 voxels, you'll get 2*(3*3)*6 triangles, i.e. 108 of them. while the same box as a 3d model will take only 2*6 triangles. we didn'd created anything fancy yet, and we already wasted 96 triangles for nothing!

of course, it is possible to heavily preprocess voxel objects to merge some of the triangles, but it still won't be near as good as a 3d model.

don't get me wrong, i like "pixelated" look of voxel objects. but i still don't think that it worth all the time i may waste on making it work acceptably. sorry.

p.s.: i'm planning to upload updated build soon. nothing really interesting, but the current build was done with the wrong compiler flags. it mostly works, but there are some bugs introduced by SSE idiocity (like coercing NaNs to zeroes).

by the way, if you're interested what distracted me from k8vavoom... it is mostly Forth and assembler. ;-) one, two, ZX Spectrum emulator, and more Forth and assembler, this time for x86. i am so busy that i didn't even played Doom for all this time. ok, maybe once or twice. still have to finish several things before i could relax.

For some reason, monsters do not "flinch" or show pain in the latest build. This makes a lot of maps much harder because you cannot stun anything. I can't even stun a cyberdemon with an SSG reliably in this port. :( 

I'm experiencing the same issue that @Mr.Rocket and @Cambertian are having with the weapon sprites, and my computer has an RTX 2060. Whatever the problem, it's definitely exclusive to 20-series graphics cards. I have another computer with a GTX 1050, so I'll make an edit once I test it

@Csonicgo dunno if it is connected, but the latest official build is officially fubared, sorry. i built it with wrong compiler settings, and it broke alot of things in unexpected ways, which i didn't noticed on quick pre-release testing. i know that i promised an update several times already, and sorry for that too. next week, i guess. ;-) sorry again, dear marines, your equipment should be updated soon. shit happens. alas.

So, I had really brutal year and a half. A new job, buried a parent, broke my leg, not much time for doom and quake stuffs. Slowly getting back to dooming and quaking. Is this thing still going on? Seems that bit of momentum got lost.

BTW I pulled-in recent tree, built it and:

• paging in multi-page (too many lines to fit to single screen) menus seem fubared
  • once you get past the last visible line, redraw is broken, you can pick lines but don't see it, weird drawing issues.
• I have trouble binding MOUSE4 and MOUSE5 from menus
  • maybe off by one error or mouse buttons iteration ? this is pretty crucial for me as I started using "back" and "forward" buttons on mice to switch weapons (much better than wheel)
  • you can bind through config file, but every config save from menu it will break it
  • MOUSE4/5 are "invisible" in menu list (hard do describe - empty space but not ---) and are not detected when pressed on "press a key" prompt

@Gunrock I tried "wispers remastered" and it's really beautiful visually, but rather brutal on UV. Besides I hate archviles and these are inviso even. Spaces seem too cramped for so many monsters to me (at least), and movement is quite hard in the narrow dungeons with so many details.

@ketmar I am surprised that as a Linux person you are not using something reliable like ZFS (yaeh, I am that kind of heretic) or at least "fedora fanboy's btrfs", do me a favor and get it working ASAP. What rig do you have? A notebook, a desktop?

5 hours ago, 3t0 said:

So, I had really brutal year and a half. A new job, buried a parent, broke my leg, not much time for doom and quake stuffs. Slowly getting back to dooming and quaking. Is this thing still going on? Seems that bit of momentum got lost.

 

BTW I pulled-in recent tree, built it and:

• paging in multi-page (too many lines to fit to single screen) menus seem fubared
  • once you get past the last visible line, redraw is broken, you can pick lines but don't see it, weird drawing issues.
• I have trouble binding MOUSE4 and MOUSE5 from menus
  • maybe off by one error or mouse buttons iteration ? this is pretty crucial for me as I started using "back" and "forward" buttons on mice to switch weapons (much better than wheel)
  • you can bind through config file, but every config save from menu it will break it
  • MOUSE4/5 are "invisible" in menu list (hard do describe - empty space but not ---) and are not detected when pressed on "press a key" prompt

@Gunrock I tried "wispers remastered" and it's really beautiful visually, but rather brutal on UV. Besides I hate archviles and these are inviso even. Spaces seem too cramped for so many monsters to me (at least), and movement is quite hard in the narrow dungeons with so many details.

 

@ketmar I am surprised that as a Linux person you are not using something reliable like ZFS (yaeh, I am that kind of heretic) or at least "fedora fanboy's btrfs", do me a favor and get it working ASAP. What rig do you have? A notebook, a desktop?

Thanks for the feedback on "Wispers Remastered" I'm currently remastering "Project Slipgate" although mapping has slowed down due to my excitement of the new RTX 3000 series GPUs and I have a shit ton of games that need to be finished....;)

2 hours ago, 3t0 said:

Is this thing still going on?

absolutely. i distracted a little, but the project is not dead by any means. just some "vacation", because there are alot of things i want to do, and only 24 hours in a day. sighs

2 hours ago, 3t0 said:

once you get past the last visible line, redraw is broken, you can pick lines but don't see it, weird drawing issues.

strange. it works for me. still, it may be because the current public build is broken in several exciting ways due to me being an idiot experimenting with compiler settings.

2 hours ago, 3t0 said:

I have trouble binding MOUSE4 and MOUSE5 from menus

basically, only 3 buttons are supported. other buttons may work, but expect all kinds of glithces. i don't have the hardware to test that, so never really checked if the code is able to cope with more buttons. sorry.

2 hours ago, 3t0 said:

something reliable like ZFS

speed and memory hog, little value for personal use (highly subjective opinion, i know ;-). and let's not even start talking about btrfs: people that were using CRC32 as a hash function for hashtables cannot be trusted at all.

2 hours ago, 3t0 said:

What rig do you have? A notebook, a desktop?

~8 y.o. desktop.

still, it seems that the problem was with ext4 journals: the sectors with journals just "weared out" (sure, they were constantly overwritten, it's not something HDDs were made for, i guess; who in their sane mind would think that people will use HDD to constantly write new data?! ;-). i turned off journaling, and hope for the best now.

p.s.: new build is on its way. its not that exciting as some other builds, with very short changelog this time due to my Forth adventures. but don't worry, some of that Forth code will eventually find its way into k8vavoom as a new VM for VavoomC. it seems that everything i do sooner or later benefits k8vavoom in some way. ;-)

10 hours ago, Gunrock said:

I'm currently remastering "Project Slipgate"

I already played the demo and I dare to say it was much better than new whispers, getting stuck wise (no offense: whispers are really densely populated and tad claustrophobic). Still those archviles on skill 1 are killing me ;).

10 hours ago, Gunrock said:

mapping has slowed down due to my excitement of the new RTX 3000 series GPUs and I have a shit ton of games that need to be finished....;)

Understandable, (un)fortunately I cannot stomach modern games anymore, so should I ever buy desktop that won't bother me.

10 hours ago, ketmar said:

let's not even start talking about btrfs: people that were using CRC32 as a hash function for hashtables cannot be trusted at all.

Oh my god, I swear this is first time in my life, that somebody else who is linux user, and they aren't behaving like stupid starry eyed anime fan girl, hyping this Ora..., I mean, Cyberdemon made FS to the moon, and is looking at it with well deserved suspicion. And don't get me started with their insanely weird RAID1 scheme, which somehow is not RAID1. For a moment I almost spilled my coffee.

10 hours ago, ketmar said:

speed and memory hog, little value for personal use (highly subjective opinion, i know ;-).

Trust me it's not that bad, you "only" need to dedicate 1-2GB per TB to it, on usual workstation, and that's probably exaggerated. If you have 2GB+ mem like 4GB or 8GB it's good ratio of value and  convenience vs performance. I even think this rule is suggested only when you have dedup, so you could go lower, but by capping it reasonably, not that much memory is lost to zfs. It will return memory to OS if it needs it anyway, albeit it will cling to it as long as it can. I am running zfs for almost a decade now on everyhting where I can and it was always worth it, never let me down. How much mem do you have?

10 hours ago, ketmar said:

basically, only 3 buttons are supported. other buttons may work, but expect all kinds of glithces. i don't have the hardware to test that, so never really checked if the code is able to cope with more buttons. sorry.

From my playthroughs everything works besides menu display and config generation, is there any specific reason k8-vavoom autosaves config each time it exits? I had to make it immutable for my changes to stick. Any way I can help this getting fixed?
 

12 hours ago, ketmar said:

~8 y.o. desktop.

That means you should have SATA, I think. How big dataset do you have? Within 100 GB ? 1 TB? If you are in your starving artist period (I've been there), low on hardware (and I've been there) maybe I could donate pair of used 3,5" drives to you. Of course only if you can handle the paranoia :). I have a hunch you are not far away for them to be shipped by post.

17 hours ago, 3t0 said:

Oh my god, I swear this is first time in my life, that somebody else who is linux user, and they aren't behaving like stupid starry eyed anime fan girl, hyping this Ora...,

;-) i am using GNU/Linux not because it is best, but because other systems are worser. so i have a huge list of "this sux" points, and absolutely no "this is for linux, so it is great!" attitude. ;-)

17 hours ago, 3t0 said:

Trust me it's not that bad

oh, i didn't meant that it is bad. i just looked at its features and requirements, and it looks really excessive for personal use. any FS will go mad if its master structures are damaged. otherwise, there is a little need to have anything more advanced then ext4 for me. snapshots? meh, i'm using git for sources, and other things don't need that. compression? almost everything is already compressed. crc checks for data? hdd hardware already does it, there is little reason to "help" it that way. and so on. so, in the end of the day, i have very advanced FS with features i don't use, and much more resource-demanding than ext4.

17 hours ago, 3t0 said:

How much mem do you have?

8GB, but it is somewhat unstable. and 32-bit system, of course. ;-)

17 hours ago, 3t0 said:

From my playthroughs everything works besides menu display and config generation

there are alot of subtle bugs you may not notice first (that's how it slipped into "official build", despite my playtesting). most of them are because i switched to "fastmath", and it completely fubared NaNs. and NaNs are used quite extensively as "this is special case" flags. what a great invention that Intel SIMD was! let's silently coerce nan to zero, what could possibly go wrong?! and i bet it is stealthly broken in other places too.

17 hours ago, 3t0 said:

is there any specific reason k8-vavoom autosaves config each time it exits?

it should only do that if something was changed. the engine generates new config (sorting everything), and then compares it with the old one. it only writes new one if it differs (made it like that so it will be somewhat easier on SSDs and such ;-).

17 hours ago, 3t0 said:

Any way I can help this getting fixed?

if you want to play with menus and don't have it saved, there is already a way! "dev thought of everything". ;-) use "+cfg_saving_allowed 0" in command line, and k8vavoom won't even try to save it. of course, you can set that cvar from the console too, any time you want. it is not persisten, tho (it cannot be, because you disabled config updates ;-).

of course, you can change the default directory for configs and saves too, with "-configdir" cli arg. by default, saves will be put in configdir/saves, but you can change that with "-savedir", if you only want to move saves, but not config.

we desperately need all those things documented. sighs.

17 hours ago, 3t0 said:

How big dataset do you have? Within 100 GB ? 1 TB?

1+TB.

17 hours ago, 3t0 said:

maybe I could donate pair of used 3,5" drives to you. Of course only if you can handle the paranoia :).

ah, i'm not infected with that modern paranoidal trend. i don't believe in clearnet anonymity anyway, and i don't even have my intimate photos to steal. ;-) also, any even slightly dedicated person can find alot about me (including my so-called "real name" and address) in ten minutes or so, i believe. especially considering that i am not really using nicks (Ketmar is what i see as my name). but back to the topic. ;-)

17 hours ago, 3t0 said:

have a hunch you are not far away for them to be shipped by post.

Ukraine, EU. i really appreciate your... dammit, is "offering" the right word? sometimes i'm completely forgetting the right word to use. and i absolutely believe that our postal service will do it's best to deliver that hardware shattered to finest small pieces. ;-) you may as well simply throw it out of the window. of course, i'll be glad to accept it if i'll be *really* desperate, but i don't want to see it simply wasted for nothing by the postal service. but thank you alot, i really appreciate it.

p.s.: oh, i somehow missed the part where you said that you pulled the source from the repo, sorry. in repo, compiler options should be right, of course. my bad, you're using GNU/Linux, and i was still thinking about windows build for some strange reason, lol.

small update (eeh, that "quick" fix was really quick! ;-). nothing fancy, but you'd better update, because previous build was fubared by wrong compiler options. sorry.

* oops; -ffast-math broke NaN comparisons (and damage factors)
* fixed UB and segfault in `DrawPlayerSprites()` (thanks, id0)
* it was impossible to use things (inited UseThing twice instead of UseThingRange) (thanks, id0)
* DECORATE `THINGSPEC_Switch` had invalid value (thanks, id0)
* texture: added workaround for non-standard Heretic floor lumps (4160 bytes) (thanks, id0)
* fixed off-by-one error in special flat textures checker (thanks, id0)
* added some sanity checks to `A_LeafCheck()` (bad, bad leaves!) (thanks, id0)

tagging @-TDRR-, as usual. ;-)

Ohhhh good time to be alive!!!! Ketmar's back:)

3 hours ago, Gunrock said:

Ohhhh good time to be alive!!!! Ketmar's back:)

thank you! not at my usuall full speed yet, tho. i think i'll need month or two more, but we're definitely getting closer. ;-)

On 9/9/2020 at 12:07 PM, ketmar said:

... any FS will go mad if its master structures are damaged. otherwise, there is a little need to have anything more advanced then ext4 for me. snapshots? meh, i'm using git for sources, and other things don't need that. compression? almost everything is already compressed. crc checks for data? hdd hardware already does it, there is little reason to "help" it that way. and so on. so, in the end of the day, i have very advanced FS with features i don't use, and much more resource-demanding than ext4.

That's why you build at least zmirror zpools (thus you have at least 2 copies even of master structures). Those zmirrors are rather fast and not so computationally demanding, when compared to higher order duplication configs like zraid1 (raid5), zraid2(raid6) and zraid3 ("raid7"). Even those are not such hogs as popular opinion of gossip spewing swarm of million faceless parrot-drones on the internet want you to believe, and can run even on modest rigs.

For example I am almost exclusive zmirror user. Of course we are talking true mirrors here, like linux dm mirrors, freebsd's geom mirrors or openbsd's softraid mirrors. That means you can do 3-way or four-way mirrors, if you are paranoid, eg. setups where each element device is fully functional 1:1 copy of each other. Even if you loose 2 drives in 3-way mirror you can still boot.

Why I got sidetracked into this? Because for some reason btrfs has some weird shit raid1c{3,4} mirrors, that I don't understand, and I have hunch that people working on btrfs either :). Are devices in these special mirrors 1:1 full copies? Are they not? Who knows? Will it boot? It seems to me, btrfs is just bag of features thrown-in randomly and nothing makes sense. Yet the current generation of young linux users is telling me how awesome btrfs is. Then when you ask questions like this, they look at you, with a blank stares of deers in headlights. They seem to really be just kids playing with their osx lindows clones, arguing about unsubstantial details like graphical interfaces, while UAC corporation clones are running the whole show from shadows. It's really sad what we have come to :).

Regarding the hardware crc32 checks, it's even sadder. I know it's hard to believe, but those don't really work. It seems unbelievable, until you actually saw real disk "controller" mangle the data and zfs fix it. See it a couple of times, and you realize, that belief in checks in your drive is completely not grounded in reality. You wonder how thing even works. The builtin error correction is not that good (most likely not good at all), especially in consumer products.  The thing is, you are not "helping" the hard drive, you are actually taking defensive and adversarial stance to it - you don't believe that hard drive - hard drives lie. All of them and all the time.

One could say "I couldn't care less". But I am a data hoarder, so I visit some datasets only after very long pauses, sometimes even after years, and I want to be able to get that data out, perfectly, even if it's something completely banal, like catgirl pictures.  That's why I am genuinely surprised when somebody brings up valid points, not often talked about, like you did. btrfs doesn't really inspire strong belief in it's capability of getting data out after years of no use.

But I also used to run ext4 with data=journal on top of dm raid1 for longest time, so I actually understand where you are coming from. I just wanted to tell you, that if you are prone to data loss like I am (and it seems that you are ... because some of us are just "lucky") there are ways nowadays to make that an non issue, and they are not even that "expensive", so maybe you could give it a test ride.

Consider it like friendly advice "been there, done that, this was my way out". With this said, I think we don't have to revisit this theme ever again, ever.

On 9/9/2020 at 12:07 PM, ketmar said:

1+TB. 

That much is in my eyes some serious dataset to not be at least mirrored :). But that is just me, whatever floats your boat.

I asked because I was decommissioning one of my ancient machines and I have one pair of 1TBs 3.5" from it. You would have to trim your set or rather add it in as an secondary mirror.

On 9/9/2020 at 12:07 PM, ketmar said:

Ukraine, EU... i absolutely believe that our postal service will do it's best to deliver that hardware shattered to finest small pieces...

Slovakia, EU. So we share a border. Welcome to our new and amazing western world, shall you build great and real capitalism, just as we did. Just remember to not buy western food, for us western citizens of lower status, it's of lower quality only, not what true westernes are allowed to consume :). Regarding the packaging I believe that ton of bubble wrap could fix it, but yes I know very well these state operated services might be a tiny bit ... unreliable. Especially at complicated times, like we are living now.

On 9/9/2020 at 12:07 PM, ketmar said:

8GB, but it is somewhat unstable. and 32-bit system, of course. ;-)

So, you are like what, on 64-bit cpu in 32-bit OS in 32-bit mode? Capped at 4 gigs? Or what? 32-bit CPU using PAE (because last time I ran something like that, it was nice nice little bag full of cute surprises)?

On 9/9/2020 at 12:07 PM, ketmar said:

we desperately need all those things documented. sighs. 

Definitely. I tried some fiddling around codebase, but for some reason my QtCreator gets really confused by this repo (common occurrence with non QT-based, non-uber-c++-nerd projects). Just curious what code editing environment you are using (maybe I asked already but forgot, so please just remind me)?

23 hours ago, 3t0 said:

It seems to me, btrfs is just bag of features thrown-in randomly and nothing makes sense.

that's what i thought about it too last time i did some research. tbh, it looked like that from the start, and it didn't changed much since. ;-)

23 hours ago, 3t0 said:

Regarding the hardware crc32 checks, it's even sadder. I know it's hard to believe, but those don't really work. It seems unbelievable, until you actually saw real disk "controller" mangle the data and zfs fix it.

yeah, i know that "consumer devices" sux, and they barely do checking at all. but in my case -- without mirrors -- it doesn't really matter. if the data is broken, there is no copy to restore it from. so there's no serious reason to pay this price. for some "very sensitive" data i'm ok with external checksums, and everything other... i'm just hoping for the best. ;-) that's what i meant -- it doesn't help in my case, not that it's useless at all.

sure, i know that good mirroring is not that slow as some people tend to believe. and there is no need to mirror the things you're actively working on (because they are changing really fast; it is way better to setup backup cron task to sync everything once in a day/week/smth), so mirroring overhead can be reduced to almost zero. i just cannot stand the fact that i have a FULL HDD FILLED WITH DUPLICATE DATA!!1111! just think about all those wonderful new bytes you can store on it instead of duplicating your old ones! ;-)

23 hours ago, 3t0 said:

Consider it like friendly advice "been there, done that, this was my way out". With this said, I think we don't have to revisit this theme ever again, ever.

yeah, thank you. but it is sometimes very hard to do the things right, even when you know what it right, and know for sure that you're doing it wrong. at least for me. ;-) your words are wise, but it's hard to follow The Way Of The Wise. ;-)

23 hours ago, 3t0 said:

So, you are like what, on 64-bit cpu in 32-bit OS in 32-bit mode?

yep. my OS was installed more than ten years ago, and moved with me to new motherboards and hdds. it is very customised 32-bit Slackware, and i am used to it.

23 hours ago, 3t0 said:

Capped at 4 gigs? Or what? 32-bit CPU using PAE

yeah, it is in PAE mode. i don't really need more than 3GB per one process anyway. i strongly believe that if software needs more than 2GB to run, and it is not some very specialised software like DB server, or professional image/audio processing, than such software is just crap, and it should be kept off of my system. so far i haven't any problems with PAE.

right now, for example, system RAM is 1.92G/7.82G (used/total). and it runs browser with alot of tabs opened, mail server, http server, mysql server, mail client, IRC client, xmpp and tox clients, more than two dozens terminals (yeah, i am that crazy), mpv, and some other things i prolly forgot about. ;-)

23 hours ago, 3t0 said:

Just curious what code editing environment you are using

very simple "programmers editor" i wrote myself, and sometimes mcedit from Midnight Commander. my editor can create "projects", but those are just lists of files i can quickly switch between without restarting the editor itself. nothing really fancy. it has regexp searching over all "project" files, but this is prolly the only feature that knows about "projects" at all. basically, it is simple tabbed terminal notepad with some extra features, only tabs are vertical. not that i need anything better anyway (i'm not using debuggers, for example).

9 hours ago, ketmar said:

very simple "programmers editor" i wrote myself, and sometimes mcedit from Midnight Commander. my editor can create "projects", but those are just lists of files i can quickly switch between without restarting the editor itself. nothing really fancy. it has regexp searching over all "project" files, but this is prolly the only feature that knows about "projects" at all.

Explains why you don't have a problem with context sensitive smart editor syntactically live parsing stuff all the time, like the QtCreator does. I started using it because it has pretty decent C99 support (not C++, I don't know C++) and at the time I was hacking on php, and it has ability to navigate through C symbols very quickly. It was great when getting to know php codebase, awesome timesaver, mainly to look under hood how stuff works. Unfortunately most stuff in k8vavoom is messing it on my setup :(. So you are not using even cscope (not sure if it knows c++) or something like that? Guess I have to find something that can still intelligently jump around but does not caca itself on k8vavoom src.

9 hours ago, ketmar said:

dozens terminals (yeah, i am that crazy)

Who doesn't have them? Let me guess urxvt?

9 hours ago, ketmar said:

very customised 32-bit Slackware

I swear Slackware users are the strangest of the bunch, in a good sense.

5 hours ago, 3t0 said:

So you are not using even cscope

nope. i was thinking about it, but i am too lazy. it may help a little with alien codebase, of course, but not much too -- you still have to look what each function/method does, and how it interacts with other parts. so for me it rarely worth it.

5 hours ago, 3t0 said:

Let me guess urxvt?

not even "u", it was rxvt long time ago. until i wrote my own terminal emulator. i think you can see the trend here. ;-)

5 hours ago, 3t0 said:

Guess I have to find something that can still intelligently jump around but does not caca itself on k8vavoom src.

it's strange that QtCreator cannot cope with k8vavoom code. there is nothing fancy there, not even alot of templates -- it is mostly "C with classes". there are macros for VavoomC intergration, but if QtCreator becomes confused by such simple things... eh...

anyway, i tried to split the code to modules (more or less), and put those into separate subdirs. it's not done in the best way yet, but much better than flat "src/", like it was when i forked it.

p.s.: just thought that QtCreator may be confused by #includes. in the process of shuffling code around some #include pathes becomes slightly wrong. gcc is still able to find everything due to "-I" mess in cmake, but QtCreator may not be happy. also, alot of code is between #ifdef CLIENT or #ifdef SERVER (yeah, it seems that Janis planned to create "dumb client" too, and the remnants are still there).

Hey @ketmar, do you mind if I try porting the k8Vavoom source port to FreeBSD?

8 hours ago, Cacodemon345 said:

Hey @ketmar, do you mind if I try porting the k8Vavoom source port to FreeBSD?

yeah, be my guest! the only thing i should warn you about is that no clang patches will be accepted into the main repo, so you'd better go with gcc port. but other FreeBSD fixes are welcome.

p.s.: please, don't start "clang flames" here too. there is no way i will change my mind about it. sorry.

5 hours ago, ketmar said:

no clang patches will be accepted into the main repo

Given that all BSDs are trying to move away from "tainted" evil of GPL compilers as soon as possible, and clang becoming de-facto standard there, this is rather weird limit, but I digress. Any simple explanation? Even "religious reasons" is acceptable, to keep it simple and not start flames. Or it is "too complicated"?

Not that you can completely avoid gcc with ports currently, either.

Hey, sorry I don't want to derail or anything but is there a secret console command of some sort in which I can turn the HUD off?

~ the player weapon HUD.