Does anyone have some documentation they could link me on how lightmaps are stored inside the .bsp?

I'm trying to create a Quake 1 map renderer in Unity and lightmaps are sort of my final frontier.

I've found some conflicting documentation, most of which is vague. From what I gather the lightmaps themselves are simple arrays of values 0-255. I assume it's just a big lump of bytes, and the faces point into it to the start of their lightmap, but I do not know how to determine the size of any given face's lightmap. bspfile.h defines it as [numstyles*surfsize], but I'm not sure where to find or how to calculate a faces surfsize for this purpose.

Can anyone toss me a pointer? 

Yeah, it's convoluted because they're not stored, you have to compute the lightmap sizes.

For each vertex on the face, you get the X and Y texture coordinates like this:

texcoord_x = (vert_x, very_y, vert_z, 1) dot (tex->vecs[0])
texcoord_y = (vert_x, very_y, vert_z, 1) dot (tex->vecs[1])

tex->vecs[0] and tex->vecs[1] are both vectors of 4 floats, tex is the texinfo_t* that has that's face texture alignment info.

Then, to get how many bytes wide/tall the lightmap is:
lightmap_width = ceil(max(texcoord_x) / 16.0) - floor(min(texcoord_x) / 16.0) + 1
lightmap_height = ceil(max(texcoord_y) / 16.0) - floor(min(texcoord_y) / 16.0) + 1

max(texcoord_x) is the maximum texcoord_x across all verts on the face. Hopefully I didn't mess anything up and the notation makes sense.

For reference this is the light tool code:
https://github.com/ericwa/tyrutils-ericw/blob/master/light/ltface.c#L355 

the engine source is probably the best documentation since it has code that reads, converts, and renders the lightmaps. 

Once you know, https://quakewiki.org/ would be a great host for that kind of information. 

I assume the notation above is dealing directly in the x/y coords of worldspace? I've already tamed the textures and mapping, if what I'm understanding is right, I should be able to use my existing code for decoding the texture UVs to do the lightmaps as well, with a little tweaking.

It was confusing because a lot of scraps I was finding were saying that lightmapping information/decoding was totally unrelated to texture UVs/composition, but it looks like that's not entirely true.

I'll give it a shot when I get off work.

Glad to see this place is still active. It was a huge help last time I worked on this project. It's the reason I've made it this far. 

I assume the notation above is dealing directly in the x/y coords of worldspace?
Yeah, the vert_x, very_y, vert_z are worldspace.

texcoord_x/y are in texture space, and dividing these by the texture width/height gives the texture U/V coordinates normalized to [0,1] (except they go outside that range for repeating.) So yeah you should be able to reuse some of the texturing code for this :-)

btw as a sanity check, the lightmap_width/height values you get should be between 1 and 18. 

Going further into properly integrating Scourge of Armagon's code into my mod, this time its sprites can't be properly compiled.
FTEqcc says "line9 - Unknown frame macro $spritename"

Here's the hipspr.qc file: http://pastebin.com/t0VypXAU

And yes, I made sure I also copy pasted the sprite files in the progs directory. 

that reminds me, i love that Jim Dose called it "QuickC" 

Daya, skip that file (delete it or something), its not qc code.
(note how its not included in hipnotic's progs.src either) 

Gotta thank CZG for that 

When the compilation went with no errors this time, bullet hole sprites still don't appear on brushes.
Even though I made sure to place the bullethole code in weapons.qc as well, so I really don't know what's wrong here.
In the end I can overlook that while developping the mod, but it's still annoying. 

Ok so i'm trying to compile a hexen 2 and csqc project with fteqcc.
So, i can compile a standard hexen2 code normally, but if i try to put some csqc in it and link it with

#pragma sourcefile csprogs.src

in 'global.hc' the compiler complains with the following error:
": werror Q208: progs CRC not recognised from quake nor clones"
The thing is that if i compile the CSQC separately its all good, but i really dont wanna to have to do two compiles each time i want to compile the whole project.
Btw i'm using the command line fteqcc (same happens with GUI)
my compilation command:

>.fteqcc -src hc -Thexen2

And also, i'm using Hammer of Thyrion's HCode since i couldn't find anything better or updated... 

put this in your hexen2 code (before any functions are processed):
#pragma target h2 //or fteh2, if you're going fteqw-only.
#pragma keyword enable thinktime
#pragma keyword enable loop
#pragma keyword enable until
and then remove -Thexen2 from your commandline.

alternatively, you can use
#pragma target id //or fte, because you might as well
in your csprogs.src, again before any statements are execed (keep the -Th2 bit if you're leaving the progs.dat code alone).

Alternatively, as you're not using fteqccgui anyway, you might as well just run fteqcc twice, with a batchfile or whatever.
fteqccgui is good for debugging though (which pragma sourcefile was implemented to simplify). 

I appended

#pragma target id

to the beginning of csdefs and it all went fine.
Thanks man! 

I can't for the life of me figure out how to enable the correct software quake style of rendering water with all its twisting and writhing in the Darkplaces engine. All the water does is scroll back and forth. I'm trying to dig through the engine code to find it. Its got to be somewhere...just disabled or commented out, right?

Anyone know where it is? 

gl_subdividesize to 16 maybe? 

...cause right now I'm working on that BFG, but:
I'm planning on doing a Berserk power-up where attack speed is increased by 3 (weapon animation would speed up to reflect that) and ammo consumption divided by 2.
The question is, would such a thing be possible with something similar to SuperDamage? Especially with guns who are tied with their model animations (nailgun, super nailgun, thunderbolt)? 

Was just reviewing the old Qc src here: http://www.gamers.org/dEngine/quake/Qc/combat.htm
the part where it mentions combining shotgun blasts (in T_damage ()), seems to be related to the dmg_take field I was working on earlier. I know there is a multidamage work around on another forum.....but what was the original intent of multidamage? Its spawning an ent that basicly collects the damage and distributes it? 

the whole code is designed to reduce the amount of data being sent from clients and servers. instead of sending 50~ damage messages for small amounts, it sends 1 damage message with a large amount. if a shotgun blast hits multiple entities, then multiple damage messages are needed, but they still wanted to collect everything up as much as possible.

as a side effect, it basically lets shotguns gib monsters. 

edit: sorry, should have actually looked at the code you were quoting.

this is to make sure the screen tint looks the right brightness.

imagine you took damage from 2 sources in the same frame, A was 50 damage and B was 5 damage.

if you just set .dmg_take to whatever the damage was, and B was processed after A, then you'd take 55 damage but have a very faint red screen effect. 

Going back to my last request, I know one way would be to have a (if BERSERK is active) check for each self.attack_finished, self.nextthink and ammo_counts for each weapons, but isn't there a more flexible option where potential added weapons could just benefit of just a function call, like what vanilla weapons do with the Quad Damage with SuperDamage(); ? 

Yes, I was wondering more about the multidamage function moreso, but you answered both questions, thanks.

I'm futzing a bit with some QC that checks each bullets impact point on the client, and determines a rough "hitlocation" guesstimate based on some code that Muavebib made and posted a long while back. Its counting up the correct number of impacts to each body part ok, however I am also wanting to separate out the possibility its contacting multiple targets.

Basicly Im floating 2 local ents - a,b and c and put this code in the loop if we hit a .FL_CLIENT ent:

if (a == world)
a = trace_ent;
else
{
if (trace_ent != a && b == world)
b = trace_ent;
else if (c != a && c != b && c == world)
c = trace_ent;
}

...it sorta crudely works and isolates the correct trace_ent near the end of the function where I count up location impacts, but I suspect this wont work well on more than 2 clients. Its hard to hit a group of more than 2 usually , but just in case Id like to cover that possibility somehow.

Is it easier to modify the multi_ent in multidamage to do this? 

So I'm in the final stretch on doing that BFG, and that's its coding.

The idea here for the attack is for the sphere itself to drill through enemies who gib through its direct impact, and to explode against a non-gibbed dying enemy on impact.
During its travel, the sphere would spawn lightning attacks to nearby enemies (let's say 200/250 map units max).
And the explosion impact, instead of generating an explosion radius, would spawn a different lightning attack, one that would last an instant and generate more damage and an impact sprite at the end of the traceline, and its radius would be as long as vanilla thunderbolt.

Thing is, even though I have all the assets in place (models, sounds, sprites for the bfg sphere, big impact, and blast impact) I don't know how to properly code all of that.
I'm following these: http://www.insideqc.com/qctut/lesson-22.shtml http://www.insideqc.com/qctut/lesson-51.shtml
But I don't think I see anything related to "while bfgshot is in world" as a check to launch the nearby-lightning code. 

The trick is similar to the problem of the looping animation you had previously. You need to create a think function like

void() bfg_ball_think =
{

//insert lightning bolt code here

self.nextthink = time + 0.3; //how often lightning strikes
self.think = bfg_ball_think; //keep repeating
}

Then when you create the ball you need to add

missile.think = bfg_ball_think;
missile.nextthink = time + 0.2;

This gets you lightning shooting out as it flies, which is a good start. 

Thanks again for Preach's help.

Now, for the lightning blast coming from the impact, the idea is to generate an impact sprite at the end of the lightning line generated. I know trace_endline has a role to this, and this is where the blast sprite function will be called, but I doubt "trace_endline = BFGBlast();" would work, as trace_endpos seems to be more attached to particle work.

Just for reference, here's what I wrote for the entire BFG (not including weapon animation, which aren't made yet): http://pastebin.com/E82MgGTy