Nikiee

Why aren't you blacklisting the objects if you're always derendering them? To re-render objects that have been temporarily derendered, simply leaving the sim and returning should recover them.

Hokay. After a lot of trial and error, and heavy use of decimal2binary(), I think I finally have an alright grasp of how the operators work! For anybody else who might stumble accross this thread, I'll leave the practice script I ended up with by this point. Personally I learn the fastest when I have a consistent base that I use as an example, so with slight modifications, this is how I ended up using the examples by all of you. Some extra formatting for readability, too. (But the comments don't look quite as good without color-coding.) // CREATE INITIAL STRUCTUREinteger db_record(integer class, integer faction, integer outfit, integer gender){ class = (class & 0x7) <<6; // 0x7 = binary 111xxxxxx (6 = width of GENDER and OUTFIT and FACTION) faction = (faction & 0x7) <<3; // 0x7 = binary 111xxx (3 = width of GENDER and OUTFIT) outfit = (outfit & 0x3) <<1; // 0x3 = binary 11x (1 = width of GENDER) gender = (gender & 0x1); // 0x1 = binary 1 (No shifting because it's the right-most) // // x's replaced by 0 when shifted, then replaced by next value return class | faction | outfit | gender;}// GET VALUEinteger get_flag(integer w, integer x){ // x = Amount of bits to shift in the x "axis." check db_record for x. return ( (db >>x) & ((1<<w)-1) );}// SET VALUEinteger set_flag(integer i, integer w, integer x){ // i = integer to set db = db & ~( (1<<w)-1 <<x); // Mask return db = db | (i & (1<<w)-1) <<x;}integer db;default{ state_entry() { //class 5, faction 7, fully clothed, male db = db_record(5, 7, 3, 1); set_flag(1, 3, 6); // 0x7 = 111 set_flag(1, 3, 3); // 0x7 = 111 set_flag(1, 2, 1); // 0x3 = 11 set_flag(0, 1, 0); // 0x1 = 1 llOwnerSay("current flags: " +(string)get_flag(3, 6)+" " +(string)get_flag(3, 3)+" " +(string)get_flag(2, 1)+" " +(string)get_flag(1, 0)); }} I also wrote this down because it was by far the most helpful set of sentences in the whole thread for me: //It would be handy to have a way to create that mask of 0x7.//If we shift a solitary one bit left by 3 places, we get 8.//The expression (1<<w)-1 produces a field of ones w bits wide.// 1 = 0001// 0001 <<3 = 1000 (8)// 8-1 = 7 = 0111 (0x7)

That is incredibly helpful and I'm so thankful that you took the time to go over all of that twice. <3 And I ended up surprising myself by having guessed w, p, and d right, and half-guessed the purpose of s. Once I get home, I'm going to play around with the function and apply it into some made-up code of my own to see what breaks.

Bah, the first time I delete a comment anywhere because I felt dumb and then this happens. :< Sorry! Anyway, if this was all done in a single small script, it wouldn't really be that much more optimal for just a couple integers, but the point of the exercise (for me) is to get an understanding of bitfields and learning to operate them. That said, having stared at the functions for a while, I'm having trouble making sense of them because I don't understand what the variables are for. (I get that it's on purpose, but!)

integer dbRecord(integer gender, integer class, integer outfit){ // longhand method so can see more easy what is happening // & mask to make sure data value is in range. then shift right into position gender = (gender & 0x1) << 5; // 0x1 = binary 1. 5 = bit width of class and outfit class = (class & 0x7) << 2; // 0x7 = binary 111. 2 = bitwidth of outfit outfit = (outfit & 0x3); // 0x3 = binary 11.} The way you explained how many values each bit can store, and the way you ordered dbRecord() had my head stuck for quite a while. But I think I understand what you're doing with the values now. Neat! Just as a note for myself (and clarification for others?), the function would be "better" (in this situation) formatted like this: integer dbRecord(integer class, integer outfit, integer gender){ class = (class & 0x7) << 3; // 0x7 = binary 111xxx (3 = width of GENDER and OUTFIT) outfit = (outfit & 0x3) << 1; // 0x3 = binary 11x (1 = width of GENDER) gender = (gender & 0x1); // 0x1 = binary 1 (No shifting because it's the first) return gender | class | outfit; // OR them together and return as integer dbRecord} There would never be a need for more than 2 values for gender, clothing could be 3 layers (as an example), and classes we could have 7 or more. That aside, this is a very nice way of dealing with any amount of properties, especially since you could store more than 8 values using this method.

Recently I've decided that I want to optimize my code further by collapsing several integers to a single bitfield (signed-32). I keep wanting to store properties or "states" as integers and check those when needed, for example: integer gender = 1; integer class = 5; integer outfit = 2; default { state_entry() { ... } }The globals and values are arbitrary, but this is what I do. This ends up using 48 bytes of memory for simple checks that are done sparingly. What I've tried doing to compress the memory usage is this: integer specs; default { state_entry() { integer gender = 1; integer class = 5 <<4; integer outfit = 2 <<8; specs = gender|class|outfit; // result: 519, binary: 0010 0000 0111 ... } }Since a 1-digit integer can be stored in 4 bits, I could compress up to 8 (128 bytes) separate integers into one (16 bytes). But now when I'm trying to figure out a way to check for those individual values, I feel like I'm doing an unnecessary amount of math and/or checks to get there. This is where I'm stumped, and hoping that someone more familiar with bitfields could explain a good way to do this.