| --- Log opened Sun Aug 18 00:00:50 2019 |
| 00:09 | | * Derakon stares at Unity, and this code: https://pastebin.com/7vVZx0fM |
| 00:09 | <&Derakon> | The function InstantiateWithCallback creates an object that contains a PlayerShipSelector instance. It sets the `callback` property of that PlayerShipSelector, and verifies that it is set. |
| 00:09 | <&Derakon> | However, in the Update function of that instance, `callback` is null. |
| 00:10 | <&McMartin> | Is it the same top-level object? |
| 00:11 | <&Derakon> | As best I can tell, yes. |
| 00:11 | <&Derakon> | Here's a screenshot of the console: Screenshot of my console: https://i.imgur.com/oodauac.png |
| 00:12 | <&Derakon> | Er, errant copy/paste there. |
| 00:12 | <&Derakon> | Or redundant, anyway. |
| 00:13 | <&Derakon> | Huh, but if I call GetInstanceID() from the static and non-static methods they give different values. |
| 00:14 | <&Derakon> | Aha! From #unity3d on freenode: " you're getting the PlayerShipSelector on the prefab, not the instantiated object" |
| 00:14 | <&McMartin> | static won't *have* an instance... |
| 00:14 | <&McMartin> | Oh, ick, Unity-specific stuff |
| 00:14 | <&Derakon> | The static method creates an instance. |
| 00:14 | <&McMartin> | OK, I'm going to be useless there~ |
| 00:14 | <&Derakon> | That's what I was referring to. |
| 00:14 | <&Derakon> | But yeah, the line "prefab.GetComponentInChildren" should have been "dialog.GetComponentInChildren". |
| 00:15 | <&Derakon> | Your instincts were right, anyway. I did have multiple objects in play. I'd forgotten the prefab could be interacted with that way. |
| 00:15 | <&Derakon> | And then got sloppy with variable names during a refactor. |
| 00:17 | <&Derakon> | Woo, you can now choose a ship to use when entering a mission. |
| 00:19 | <&Derakon> | I feel like I've been progressing really slowly lately, but I am making progress...next step: adding a complete-mission workflow. Then you'll be able to start the game at the main menu, go to the ship designer, create a new ship or load an existing one, edit it, save it, go back to the main menu, choose a mission, choose a ship to use, complete the mission, and return to the main menu. |
| 00:54 | <&[R]> | So... with LLVM and clang is there actually a way to compile to something you can run directly on LLVM, or is it purely meant as an intermediary? |
| 00:58 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 01:00 | <&[R]> | I was looking at Limbo+dis and was sort of wondering if LLVM could be used to do the same thing |
| 01:27 | <&McMartin> | [R]: I think you're looking for the output format they call "LLVM Bitcode", which can be directly executed with `lli`? |
| 01:29 | <&[R]> | Cool, thanks |
| 01:30 | <&McMartin> | That question mark isn't just for show; if you get some progress on this, do let me know |
| 01:30 | <&McMartin> | My confidence level here is p. low :) |
| 01:30 | <&[R]> | Heh |
| 01:30 | <&[R]> | FUUUUUUCK |
| 01:30 | <&[R]> | FUCKING PUREE CAT BUTTON |
| 01:30 | <&McMartin> | no |
| 01:30 | <&McMartin> | all kitties get 10s |
| 01:30 | <&[R]> | I was going to give you some links but I hit the puree cat button |
| 01:31 | <&McMartin> | :( poor kitty |
| 01:31 | <&McMartin> | Anyway, I mostly know of bitcode as the thing that would make my app bundles at work way smaller, but I'm not allowed to use it becuase we depend on libraries that want inline asm~ |
| 01:31 | <&[R]> | D: |
| 01:32 | <&McMartin> | Things may have improved since we made those decisions, though |
| 01:32 | <&[R]> | Anyways, I'll get you those links when FF starts back up in 5-10 minutes |
| 01:33 | | * McMartin notes that at this point it probably also matters way less because, like, every device he targets professionally these days is AArch64 exclusively |
| 01:34 | <&[R]> | https://stackoverflow.com/questions/9148890/how-to-make-clang-compile-to-llvm-ir |
| 01:34 | <&[R]> | http://llvm.org/docs/CommandGuide/lli.html |
| 01:35 | <&McMartin> | Hm. That makes me think bitcode may be a level one or two below where you're aiming. |
| 01:35 | <&[R]> | Potentially |
| 01:38 | <&McMartin> | I see two statements in my journeys that don't add up |
| 01:38 | <&McMartin> | From https://en.wikipedia.org/wiki/LLVM#Intermediate_representation |
| 01:38 | <&McMartin> | "LLVM supports three isomorphic (i.e., functionally equivalent) forms of IR: a human-readable assembly format, an in-memory format suitable for frontends, and a dense bitcode format for serializing." |
| 01:39 | <&McMartin> | From the lli docs: |
| 01:39 | <&McMartin> | "lli is not an emulator. It will not execute IR of different architectures and it can only interpret (or JIT-compile) for the host architecture." |
| 01:39 | <&[R]> | Ah |
| 01:39 | <&[R]> | Yeah, that's not useful |
| 01:39 | <&McMartin> | Maybe armv7-ios and arm64-ios are "the same host architecture" |
| 01:39 | <&McMartin> | But the whole point of shipping iOS apps as bitcode was "this way you do not need to send three copies" |
| 01:40 | <&McMartin> | ... um, yeah, from later in the lli docs, you can ovveride -march and -mcpu |
| 01:40 | <&McMartin> | So idkwtf |
| 01:41 | <&[R]> | But it references llc throughout all the command examples |
| 01:41 | <&McMartin> | This calls for adventures, imo |
| 01:44 | <&[R]> | .ll isn't even bytecode |
| 01:45 | <&[R]> | It's "human readable LLVM assembly language" |
| 01:46 | <&[R]> | Okay, so llvm-as produces bytecode |
| 01:47 | <&[R]> | clang -S -emit-llvm printf.c; llvm-as printf.ll |
| 01:48 | <&[R]> | Produces printf.bc |
| 01:48 | <&[R]> | Which lli can read and use |
| 01:48 | <&[R]> | 00000c00: 696e 7072 696e 7466 382e 302e 3178 3836 inprintf8.0.1x86 |
| 01:48 | <&[R]> | 00000c10: 5f36 342d 7063 2d6c 696e 7578 2d67 6e75 _64-pc-linux-gnu |
| 01:48 | <&[R]> | 00000c20: 7072 696e 7466 2e63 2e4c 2e73 7472 2e4c printf.c.L.str.L |
| 01:49 | <&[R]> | Hmm |
| 01:49 | <&[R]> | (select output of xxd printf.bc) |
| 01:52 | <&[R]> | Wish I had some ready to go alternative arch installs, since the want of this was to make building stuff for them easier |
| 01:52 | <&McMartin> | Sounds like a job for an RPi, imo |
| 01:52 | <&[R]> | I ahve one, just not immediately setup |
| 01:52 | <&[R]> | Hmm |
| 01:53 | <&[R]> | I have FreeBSD running on one though, that might be a good test |
| 03:06 | <&[R]> | Wait, what's the difference between "bitcode" and "bytecode" |
| 03:06 | <&[R]> | ? |
| 03:07 | <&McMartin> | Sarcastic answer: "bytecode" has Java cooties |
| 03:08 | <&McMartin> | More serious answer unrelated to what LLVM may or may not be doing: I would assert that a "bitcode" doesn't care about 8-bit alignment when it's doing its thing. |
| 03:08 | <&McMartin> | This would comport with the use of 'bitstream' in video and audio codecs. |
| 03:09 | <&McMartin> | And by this standard, ARM instructions are bitcoded, 6502 instructions are bytecoded, and Z80 depends on how you squint at it. |
| 03:17 | | Kindamoody is now known as Kindamoody[zZz] |
| 04:21 | | NuClEuS` [Muhammad@Nightstar-9tc.16g.64.196.IP] has quit [Ping timeout: 121 seconds] |
| 04:45 | | Vorntastic [uid293981@Nightstar-2dc.p8m.184.192.IP] has joined #code |
| 04:45 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 06:55 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:35 | | Degi [Degi@Nightstar-0snccr.dyn.telefonica.de] has joined #code |
| 08:53 | | Kindamoody[zZz] is now known as Kindamoody |
| 09:27 | | * McMartin ! |
| 09:28 | | * McMartin has had an insight |
| 09:28 | <&McMartin> | Alternately, I just saw Solid Snake, and had better sound the alarm before he notic |
| 09:31 | <~Vorntastic> | ?? |
| 09:34 | <&McMartin> | Some of the Atari madness |
| 09:35 | <&McMartin> | When I tried to write Lights Out for the Atari 800, I couldn't get the hardware to do a display I liked, and once I got close, I ended up realizing I'd designed it not for the 800, but rather the 2600 |
| 09:35 | <&McMartin> | I think I've just found a way to make a Very Good Display for the 800. |
| 09:36 | <&McMartin> | It involves, in part, lumping the missiles together to make them pretend to be displaying letters in a text mode. |
| 09:37 | <&McMartin> | (The big problem I've been having here is that in order to write letters in the cells at a size I actually am OK with, I have to use a strictly monochrome graphics mode.) |
| 09:37 | <&McMartin> | But I can mix sprites and bitmaps to get a much more workable result. |
| 09:37 | <&McMartin> | I'll just have to do hella rasters. |
| 09:38 | <&McMartin> | Fortunately, Hella Rasters is what the Atari is all about, and that will give me an excuse to play with parts of the 800 that the 2600 did not have, and that even the C64 works hard to match. |
| 09:40 | <~Vorntastic> | Oh I was wondering: what makes BBC BASIC so good? |
| 09:40 | <&McMartin> | What I don't know is whether or not I can do it without running into its (hugely chaotic and intractable) badlines. |
| 09:40 | <&McMartin> | BBC BASIC had three things going for it |
| 09:41 | <&McMartin> | (a) incredible longevity and official support, like the BBC says |
| 09:41 | <&McMartin> | (b) About as good support for structured programming as it is possible for BASIC to have until QBASIC restructured the language to look like Pascal |
| 09:42 | <&McMartin> | (b.1) This includes proper procedures and functions, with local variables, and the invocation of which were significantly faster than GOSUB |
| 09:42 | <&McMartin> | (b.2) This also meant that once BASIC environments got more graphical line numbers became an afterthought that your editor would often hide |
| 09:43 | <&McMartin> | (c) BBC BASIC included a full-featured symbolic assembler, and BASIC-side support for pointers and byte buffers to back it up |
| 09:43 | <&McMartin> | So you could have assembly language routines, in proper assembly language, with labels, in the middle of your BASIC code as part of the listing |
| 09:44 | <&McMartin> | This in turn meant that while in the US we got pushed to Pascal and C very early on just to get decent speed out of our programs or to get more reliable access to the underlying hardware, the only reason a lot of home computing enthusiasts would bother leaving BBC BASIC is because they'd like access to proper nested data structures |
| 09:44 | <&McMartin> | Because, like, that was pretty much all you were missing |
| 09:45 | <&McMartin> | It is nevertheless true that once you get some C in you you can't go back |
| 09:46 | <~Vorntastic> | Ok that's pretty impressive |
| 09:46 | <&McMartin> | *As a language* I'd say QuickBasic is better. |
| 09:47 | <&McMartin> | But QuickBasic also came, like, ten years later. |
| 09:47 | <&McMartin> | And it doesn't have pointers or inline asm~ |
| 09:47 | <&McMartin> | (Strictly speaking the BBC BASIC stuff isn't *inline* asm; when you "ran" the lines of assembly language code you were filling up a bytebuffer to be invoked with USR at a later time) |
| 09:48 | <&McMartin> | (And depending on the code you might have to put it in a loop, one iteration for each pass the assembler needed, which is hilarious but also appropriate) |
| 09:49 | <~Vorntastic> | (pffff) |
| 09:52 | | * Vorntastic has a thought, wonders why he's prefetching kickout exclusions in the first place he can just examine the grid live |
| 09:55 | <~Vorntastic> | Would need to make the grid respond appropriately to out-of-bounds selections |
| 09:56 | | Kindamoody is now known as Kindamoody|afk |
| 09:57 | <~Vorntastic> | But I kind of need that anyway |
| 10:11 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 10:11 | <&McMartin> | ? |
| 10:13 | <~Vorntastic> | Making collision, grid vs aabb. Question is which way do you move the aabb to kick it out of solid block collisions. Typically the answer is "whichever direction you are colliding less in", but |
| 10:14 | <~Vorntastic> | In some situations you can crash into a wall in such a way that you collide a lot with the wall but not much with one of the grid tiles vertically |
| 10:14 | <&McMartin> | oh this isn't crossroads |
| 10:15 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Connection closed] |
| 10:15 | <~Vorntastic> | This kicks you vertically, and may incorrectly count you as hitting a floor or ceiling, activating other mechanics, which is bad. |
| 10:16 | <~Vorntastic> | So to prevent that, one must exclude kickouts that push towards other full tiles |
| 10:16 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 10:16 | | mode/#code [+o himi] by ChanServ |
| 10:18 | | NuClEuS` [Muhammad@Nightstar-8eu.eog.179.160.IP] has joined #code |
| 10:22 | | NuClEuS` [Muhammad@Nightstar-8eu.eog.179.160.IP] has quit [Ping timeout: 121 seconds] |
| 10:26 | <~Vorntastic> | (this is the thing that gets done in smb1 speedruns - hit the wall the right way and you get kicked upward, grounding you so you can jump) |
| 12:31 | | * ToxicFrog gnaws on ttymor |
| 12:31 | <&ToxicFrog> | Multi-tile entities suuuuuuuuuck |
| 12:32 | <&ToxicFrog> | The current implementation-in-progress uses multiple single-tile entities linked together, so that when something happens to one it sends messages to the others |
| 12:32 | <&ToxicFrog> | This means a lot of bookkeeping |
| 12:33 | <&ToxicFrog> | An alternate implementation would be support for true multi-tile entities, with a width and height, that exist in multiple map tiles at once |
| 12:33 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 12:33 | <&ToxicFrog> | This makes the map/render code more complicated, but makes it impossible for different parts of the entity to get out of sync |
| 12:34 | <&ToxicFrog> | A third possibility would be to have a MultiTile component that handles all the linkage and message-passing, but I'm not sure if this can actually work in a general way |
| 12:34 | <&ToxicFrog> | E.g. when you <verbs> a door, if it's open, it returns #{"Close"} iff all tiles of the door are unobstructed |
| 12:35 | <&ToxicFrog> | And the MultiTile would somehow have to know that |
| 13:24 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 13:24 | | mode/#code [+o Reiv] by ChanServ |
| 13:29 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 13:46 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 13:46 | | mode/#code [+o Reiv] by ChanServ |
| 14:02 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 14:03 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 14:03 | | mode/#code [+o Reiv] by ChanServ |
| 14:08 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 14:08 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 14:08 | | mode/#code [+o Reiv] by ChanServ |
| 15:06 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has joined #code |
| 15:06 | | mode/#code [+o celticminstrel] by ChanServ |
| 15:20 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:20 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:34 | | Degi [Degi@Nightstar-0snccr.dyn.telefonica.de] has quit [[NS] Quit: Leaving] |
| 15:34 | | Degi [Degi@Nightstar-0snccr.dyn.telefonica.de] has joined #code |
| 17:24 | | Vorntastic [uid293981@Nightstar-2dc.p8m.184.192.IP] has quit [[NS] Quit: Connection closed for inactivity] |
| 17:30 | | ErikMesoy [Bruker@Nightstar-r88323.bb.online.no] has left #code [] |
| 17:39 | | ErikMesoy [Bruker@Nightstar-r88323.bb.online.no] has joined #code |
| 17:39 | | mode/#code [+o ErikMesoy] by ChanServ |
| 19:00 | <~Vornicus> | Okay. load up two bytes with random numbers... that have to be different from the previous random numbers that occupied those spaces. one of them is random between 0x6b and 0x3e inclusive, the other is picked from a list. ...hm. Well this is ...simultaneously surprising and a hunch that panned out. It turns out that all 8 entries in this list are... codes for colors that work well on a white background? |
| 19:01 | <~Vornicus> | (the list is 2 4 5 6 8 9 11 14; there are three other colors that are quoted as working well on white: 0 12 and 15) |
| 19:01 | <~Vornicus> | I'm mostly surprised it's *white* |
| 19:25 | <~Vornicus> | ok. cpx #$05 bcc (blah) . we skip if... x < 5 |
| 19:30 | <~Vornicus> | wait no, it's accumulator - target, so if x >= 5. |
| 19:32 | <~Vornicus> | this is the part of 6510 assembly I don't like; these different ways of doing comparisons are subtle in ways I keep getting backwards |
| 19:48 | < Degi> | Oh crap, smh ATMega328 -> C# over serial reads a different number of bytes each time despite the same number being sent and always reads less than being sent |
| 19:54 | < Degi> | Heh oops |
| 20:09 | | Kindamoody|afk is now known as Kindamoody |
| 20:20 | < Degi> | Hmh now I thought I read all bytes but it still doesnt work... Maybe the UART misses a few bytes |
| 20:20 | < Degi> | Though it did finish once. |
| 21:37 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 21:37 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 21:37 | | mode/#code [+o Reiv] by ChanServ |
| 21:56 | <&McMartin> | Vornicus: Worse, Carry has opposite meanings on 6502 and Z80 :( |
| 21:56 | <~Vornicus> | srs |
| 21:56 | <&McMartin> | C is a True Carry Bit on Z80, and it's a Borrow bit on 6502. |
| 21:56 | <&McMartin> | You borrow from the carry |
| 21:57 | <~Vornicus> | I still don't get that at all |
| 21:57 | <~Vornicus> | Most proper subtractions it's done it's set C first |
| 21:57 | <&McMartin> | The 6502 one is the one that, to me, makes sense |
| 21:57 | <&McMartin> | Yeah |
| 21:58 | <&McMartin> | That's the rule |
| 21:58 | <&McMartin> | So, like, with Carry set it you go in with something that looks like, (three bit example here) 1001 - 010 = 0111 |
| 21:58 | <~Vornicus> | and the only way that makes sense is if having C set means the subtraction turns out the way I expect it to if there's nothing affecting it |
| 21:58 | <&McMartin> | Carry clear, there's a borrow |
| 21:59 | <&McMartin> | If carry is clear, SBC subtracts an additional 1 from the result. |
| 21:59 | <&McMartin> | Er, if carry is clear *going in* |
| 21:59 | <&McMartin> | That's how it ripples. |
| 22:00 | <&McMartin> | This parallels addition, where 010 + 111 = 1001 |
| 22:01 | <~Vornicus> | This so far has made zero sense by the way |
| 22:01 | <&McMartin> | Um, hm, OK, let's step back a bit |
| 22:01 | <&McMartin> | Though it's 2PM and I haven't had lunch yet so this will be brief |
| 22:02 | <&McMartin> | You get how carries work in *addition*, right, so as to do 16-bit adds? |
| 22:02 | <~Vornicus> | Yeah, you've got numbers big enough they add up to somewhere from 256 to 510, and the 2^8 bit shows up in C |
| 22:03 | <~Vornicus> | then you can use C (and your high bytes) to get the correct high byte in the result |
| 22:03 | <&McMartin> | And then that sets C the same way |
| 22:03 | <~Vornicus> | yes, etc etc |
| 22:03 | <&McMartin> | So think of that as a secret ninth bit in the arguments, where both are zero. |
| 22:04 | <&McMartin> | For subtraction, think of the secret ninth bit as being one in the first operand and zero in the second, and then C holds the ninth bit of the result |
| 22:04 | <&McMartin> | With an extra step of "if C is zero going in, that means that the previous step had to borrow from you, so decrement an additional amount". |
| 22:05 | <&McMartin> | http://michaelcmartin.github.io/Ophis/book/x641.html |
| 22:06 | <&McMartin> | Now foods |
| 22:08 | <&McMartin> | Well, first |
| 22:08 | <&McMartin> | The easy answer is "have an index card somewhere that says BCC = Unsigned <; BCS = Unsigned >=" |
| 22:08 | <~Vornicus> | which side does the thing in the accumulator go |
| 22:25 | | NuClEuS` [Muhammad@Nightstar-9o0.qp0.154.105.IP] has joined #code |
| 22:46 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:50 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 23:12 | | mac [macdjord@Nightstar-rslo4b.mc.videotron.ca] has joined #code |
| 23:12 | | mode/#code [+o mac] by ChanServ |
| 23:14 | | macdjord [macdjord@Nightstar-rslo4b.mc.videotron.ca] has quit [Operation timed out] |
| 23:23 | <&McMartin> | Left side |
| 23:23 | <&McMartin> | Same as the SBC instruction. |
| 23:23 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| 23:23 | <~Vornicus> | ok perfect *writes this down somewhere* |
| 23:51 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| --- Log closed Mon Aug 19 00:00:52 2019 |