| --- Log opened Wed Nov 30 00:00:30 2016 |
| 00:04 | <@celticminstrel> | Okay so hopefully this works. |
| 00:04 | <@celticminstrel> | ...oh right, I forgot to account for the case where there are no stairs. |
| 00:12 | <@celticminstrel> | Okay, so clearly something didn't work. |
| 00:21 | | [R] [rstamer@genoce.org] has quit [Operation timed out] |
| 00:23 | | [R] [rstamer@Nightstar-d7h8ki.org] has joined #code |
| 00:23 | | mode/#code [+ao [R] [R]] by ChanServ |
| 00:28 | | Kindamoody is now known as Kindamoody[zZz] |
| 00:36 | <@celticminstrel> | Since my log message doesn't show up, I guess the problem probably lies in the DFS to enumerate the continents... |
| 00:57 | | Derakon[AFK] is now known as Derakon |
| 01:29 | | Vash [Vash@Nightstar-uhn82m.ct.comcast.net] has joined #code |
| 02:26 | <@celticminstrel> | So my DFS is running too long. |
| 02:44 | <@celticminstrel> | I think I probably should've done a BFS instead. |
| 02:44 | <~Vornicus> | both should cover everything in the same time |
| 02:45 | <@celticminstrel> | Well okay. |
| 02:47 | <@celticminstrel> | It just seems like BFS would result in fewer spaces being pushed multiple times. |
| 02:47 | <@celticminstrel> | But maybe that doesn't really matter. |
| 02:48 | <~Vornicus> | You shouldn't have spaces getting pushed multiple times in the first place |
| 02:48 | <~Vornicus> | and if you are then you shouldn't *look* at them multiple times, and these both you can do by a set membership test |
| 02:49 | <@celticminstrel> | Yeah, I have a set membership test going on here. |
| 02:49 | <@celticminstrel> | Hmm, I think I'm pushing neighbours even for impassable tiles... |
| 02:50 | <@celticminstrel> | I can't actually tell easily whether a tile is passable from stepping through in the debugger. |
| 02:51 | <@celticminstrel> | Or wait, yes I can, because it won't enter that if statement if it's not. |
| 03:05 | <@celticminstrel> | Wait, when did x become negative. o.O |
| 03:06 | <@celticminstrel> | Ohhhh... |
| 03:10 | <@celticminstrel> | So it was spilling off the edge of the map because I added the neighbours of impassable tiles and didn't bound-check the coordinates. |
| 03:13 | <@celticminstrel> | Fixed that, and it's now fast enough that jumping between breakpoints around the DFS is nigh-instantaneous. |
| 03:20 | <&Derakon> | For some reason I read "BFS" as "Big Fucking Sword" instead of Breadth-First Search. |
| 03:20 | <@celticminstrel> | Heh. |
| 03:20 | <&Derakon> | And yeah, you almost always want to use BFS instead of DFS. |
| 03:20 | <@celticminstrel> | Well, it should be easy to change. |
| 03:20 | <@celticminstrel> | Mostly just change "pop" to "shift", right? |
| 03:21 | <@celticminstrel> | Maybe move the if(visited[next]) check. |
| 03:21 | <@celticminstrel> | Into the push-neighbours loop. |
| 03:21 | <&Derakon> | The general idea for a BFS is something like: |
| 03:21 | <&Derakon> | queue = [start] |
| 03:21 | <&Derakon> | visited = [] |
| 03:21 | <&Derakon> | while queue: |
| 03:21 | <&Derakon> | node = queue.pop() |
| 03:21 | <&Derakon> | for neighbor in node: |
| 03:22 | <&Derakon> | if neighbor not in visited: queue.push(neighbor) |
| 03:22 | <&Derakon> | visited.push(node) |
| 03:22 | <&Derakon> | Of course you need a termination condition, and you need to track the actual path somewhere, but that's the rough idea. |
| 03:22 | <@celticminstrel> | My DFS is basically like that except a stack and check visited after popping instead of before pusing. |
| 03:23 | <@celticminstrel> | So changing it to BFS should be trivial. |
| 03:24 | <@celticminstrel> | Okay, so my A* is not removing spaces from the discovered array,.,. |
| 03:25 | <@celticminstrel> | Which will almost certainly cause it to never terminate... |
| 03:27 | <@celticminstrel> | Hmm, maybe I'm wrong on that, but it's still probably not a good thing. |
| 03:27 | <@celticminstrel> | BTW Firefox, why is there no stop button. |
| 03:27 | <&Derakon> | Hit Escape. |
| 03:27 | <@celticminstrel> | Eh? Really? |
| 03:27 | <&Derakon> | I think so. |
| 03:28 | <@celticminstrel> | That just opens the console in split-screen mode... |
| 03:28 | <@celticminstrel> | Basically if I realize it's an infinite loop I'm forced to wait until Firefox decides my script is not responding. |
| 03:28 | <&Derakon> | Ahh. |
| 03:29 | <&Derakon> | Yeah, I think Escape just stops it from loading, not from executing Javascript. |
| 04:01 | <@celticminstrel> | Not sure how to tell if the A* is working... |
| 04:06 | <&Derakon> | Make a unit test. |
| 04:06 | <&Derakon> | Set up a simple environment with some obstructions and ensure that it returns the correct path. |
| 04:08 | <@celticminstrel> | Okay, factored out the A* into a separate function. That should make this easier. |
| 04:09 | <@celticminstrel> | (Also factored out the DFS for connected components into a separate function.) |
| 04:10 | <@celticminstrel> | (Which may or may not become a BFS later.) |
| 04:40 | <@celticminstrel> | I should've made the first test a clear path instead of around a lake. >_> |
| 04:41 | <&Derakon> | Heh. |
| 04:42 | <&Derakon> | Remember to test the situation where there is no path, as well. |
| 04:42 | <@celticminstrel> | That shouldn't happen with the cost function I used though. |
| 04:42 | <&Derakon> | Ah, it's just really expensive to go through water? |
| 04:42 | <@celticminstrel> | Yeah, currently 100. |
| 04:43 | <@celticminstrel> | Other tiles are no more than 3 IIRC. |
| 04:43 | <&Derakon> | Fairynuff. |
| 04:44 | <@celticminstrel> | I feel like it should've finished by now... |
| 04:51 | <@celticminstrel> | Oh. I forgot I'd hard-coded it to pathfinding through the underground area instead of the surface. |
| 04:55 | <@celticminstrel> | (I guess in theory it could be made to pathfind across layers. Maybe I'll add that later.) |
| 04:58 | <@celticminstrel> | It does seem like it's working for very simple paths, but I'm not seeing the output (it's supposed to draw it on the map). |
| 05:01 | <@celticminstrel> | ...ah, I see why it's not. Duh. |
| 05:01 | <@celticminstrel> | Since I originally wrote this for digging passages I had it draw only on impassable tiles. |
| 05:01 | | macdjord|wurk is now known as macdjord |
| 05:07 | | Derakon is now known as Derakon[AFK] |
| 05:37 | <@celticminstrel> | Well it seems like it's working. |
| 05:39 | <@celticminstrel> | At least when I run it from the command console for short paths. |
| 05:39 | <@celticminstrel> | Even for somewhat long paths actually. |
| 05:39 | <@celticminstrel> | But now I'm running it in the original context and it's not terminating. |
| 05:40 | <@celticminstrel> | I wouldn't be surprised if it needs to visit nearly every clear space before realizing it needs to tunnel... |
| 05:42 | | gnolam [lenin@Nightstar-t1tbf0.cust.bahnhof.se] has quit [[NS] Quit: Quit.] |
| 05:46 | <@celticminstrel> | Pretty sure it's in an infinite loop. |
| 06:15 | | Vash [Vash@Nightstar-uhn82m.ct.comcast.net] has quit [Connection closed] |
| 07:15 | | Reiver [quassel@Nightstar-ksqup0.co.uk] has quit [Connection closed] |
| 07:15 | | Orthia [quassel@Nightstar-ksqup0.co.uk] has quit [Connection closed] |
| 07:25 | | Kindamoody[zZz] is now known as Kindamoody |
| 07:30 | | Kindamoody is now known as Kindamoody|afk |
| 07:39 | | celticminstrel [celticminst@Nightstar-h4m24u.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 07:43 | | Orthia [quassel@Nightstar-ksqup0.co.uk] has joined #code |
| 07:43 | | mode/#code [+o Orthia] by ChanServ |
| 07:43 | | Reiver [quassel@Nightstar-ksqup0.co.uk] has joined #code |
| 07:43 | | mode/#code [+ao Reiver Reiver] by ChanServ |
| 08:53 | | macdjord is now known as macdjord|slep |
| 09:44 | | crystalclaw [crystalclaw@Nightstar-12q9ui.xyz] has quit [Ping timeout: 121 seconds] |
| 09:45 | | crystalclaw [crystalclaw@Nightstar-12q9ui.xyz] has joined #code |
| 09:45 | | mode/#code [+o crystalclaw] by ChanServ |
| 12:47 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 13:28 | <@TheWatcher> | https://twitter.com/DJSnM/status/803659072218099713 - I wouldn't put it past some to try... |
| 13:29 | <@Tamber> | hee |
| 13:44 | <&VirusJTG> | I'd try |
| 13:44 | <&VirusJTG> | I enjoy basterdized configurations. |
| 14:52 | | You're now known as TheWatcher[d00m] |
| 15:41 | | macdjord|slep is now known as macdjord|wurk |
| 15:43 | | You're now known as TheWatcher |
| 16:03 | | celticminstrel [celticminst@Nightstar-h4m24u.dsl.bell.ca] has joined #code |
| 16:03 | | mode/#code [+o celticminstrel] by ChanServ |
| 16:35 | | Emmy-AFK is now known as Emmy |
| 16:53 | | Alek [Alek@Nightstar-cltq0r.il.comcast.net] has quit [Operation timed out] |
| 16:56 | | Alek [Alek@Nightstar-cltq0r.il.comcast.net] has joined #code |
| 16:56 | | mode/#code [+o Alek] by ChanServ |
| 17:35 | | Kindamoody|afk is now known as Kindamoody |
| 17:56 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Connection closed] |
| 17:57 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 17:57 | | mode/#code [+o Emmy] by ChanServ |
| 18:02 | <@celticminstrel> | So I realized the taxi-cab distance isn't quite admissible, because negative x wraps to positive. I don't think this is going to cause any problems though with respect to whether it terminates with a possible path... |
| 18:03 | <@celticminstrel> | I wonder if using a proper priority queue rather than pushing into an array and linear-searching would be more efficient. |
| 18:04 | <@celticminstrel> | Proper priority queue would still involve search though... but it could be a binary search I guess... |
| 18:23 | <@celticminstrel> | Seems like failed binary search yields the index of the predecessor... (or maybe it depends on whether < or > is made non-strict) |
| 18:26 | <@celticminstrel> | (No, that doesn't look like it'd have any effect.) |
| 18:34 | | Vash [Vash@Nightstar-uhn82m.ct.comcast.net] has joined #code |
| 18:40 | <@celticminstrel> | For now I guess I'll ignore wrapping. |
| 18:41 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 18:41 | <@celticminstrel> | I might want to support it once I get to monster pathfinding though... as well as movement between coordinate patches... |
| 19:05 | | Vash [Vash@Nightstar-uhn82m.ct.comcast.net] has quit [Connection closed] |
| 19:19 | | ErikMesoy [Erik@Nightstar-hq72t5.customer.cdi.no] has joined #code |
| 19:19 | | mode/#code [+o ErikMesoy] by ChanServ |
| 19:19 | <@ErikMesoy> | Adventures In Budgeting. |
| 19:19 | <@ErikMesoy> | Consulting at a new workplace this month, I was supposed to submit my end-of-month timesheets. |
| 19:19 | <@ErikMesoy> | The timesheet submission software has a limited number of licenses. |
| 19:20 | <@ErikMesoy> | Every employee logged in to submit hours counts as a license user. |
| 19:20 | <@ErikMesoy> | The number of licenses the company has is substantially lower than the number of employees(+contractors). |
| 19:20 | <@ErikMesoy> | Result: complaining all around the table and all around the floor about the "you cannot log into timesheet submission right now" messages that persisted for various people ALL BLOODY DAY. |
| 19:21 | <@ErikMesoy> | Wonderful example of penny wise, pound foolish to buy fewer licenses for the timesheet submission software so that your workers have to queue for hours to use it. |
| 19:22 | <&jerith> | Yay timesheets. |
| 19:22 | <&jerith> | I always made timesheets go away by accurately tracking the amount of time I spent tracking time. |
| 19:23 | <&jerith> | These days, for the things I need to log hours on, I use powers of two and round up. |
| 19:23 | <@ErikMesoy> | Pfffft. Interesting approach. |
| 19:24 | <&jerith> | Minimum one hour (because if it's worth tracking at all, it's interrupty enough to cost me at least that), then two hours, then four hours, then eight hours. |
| 19:24 | <@ErikMesoy> | For that matter, I wonder why they have a license-limited thing in the first place. |
| 19:25 | <&jerith> | Fibonacci numbers also work well, but I like the four and eight for half and full days. |
| 19:25 | <@ErikMesoy> | Unless there are Great Hidden Depths in the timetracker, it's a glorified spreadsheet, not an MMO, and should be able to handle everyone in the building logging in if it's running on a server from this generation, right? |
| 19:26 | <&jerith> | ErikMesoy: Sure, and that means the vendor can extract money for every person using it.~ |
| 19:27 | <@ErikMesoy> | jerith: yeah, but I'm thinking why buy from this kind of vendor in the first place? eldritch quality assurance policies somewhere? |
| 19:28 | <@ErikMesoy> | A horrendous spreadsheet market where every single vendor wants per-simultaneous-user license fees? |
| 19:30 | <@ErikMesoy> | I mean, it was bizarre in the first place to learn that employees have to *queue* for hour filing. Learning the why of "license limit" was just pushing the issue back. I wonder how many layers of bizarre why-ness lurk below here. |
| 19:33 | <@Tamber> | "...well, we're not going to have everyone try to do their timesheets at the same time, right? So we don't need that many licenses!" |
| 19:34 | <@ErikMesoy> | Tamber: Yeah, they might have had different expectations. |
| 19:34 | <@ErikMesoy> | They have definitely realized the problem now, though, as while I was logged in, I got an alert from the system requesting me to please finish fast and log out so other people can log in. |
| 19:34 | <@ErikMesoy> | Signed with a name of some mucky-muck rather than just "system". |
| 19:35 | <@ErikMesoy> | I have no great hopes for this impelling them to actually fix it though >_> |
| 19:39 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 19:39 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 19:51 | | abudhabi is now known as Logar |
| 21:35 | | gnolam [quassel@Nightstar-t2vo1j.tbcn.telia.com] has joined #code |
| 21:35 | | mode/#code [+o gnolam] by ChanServ |
| 21:44 | <@celticminstrel> | I wonder if there's an easy way to get min distance between two points on a given surface with a given metric... |
| 21:47 | <@celticminstrel> | Gasp, "PriorityQueue.prototype.clear = PriorityQueue;" |
| 21:47 | <&[R]> | Nice, the clear function just calls the constructor again. |
| 21:49 | | JustLurk [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 21:49 | | JustBob [justbob@Nightstar.Customer.Dissatisfaction.Administrator] has quit [NickServ (RECOVER command used by JustLurk)] |
| 21:49 | | mode/#code [+o JustLurk] by ChanServ |
| 21:49 | | JustLurk is now known as JustBob |
| 21:59 | | gnolam [quassel@Nightstar-t2vo1j.tbcn.telia.com] has quit [Connection closed] |
| 22:02 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:17 | | Logar is now known as abudhabi |
| 22:39 | <@celticminstrel> | If I'm going to use this A* for monster pathfinding at some point, I really should probably make it able to cross levels and coordinate patches... and even pass teleporters (which don't exist yet but will later). |
| 22:44 | <@celticminstrel> | (Crossing layers is the easy part of that.) |
| 22:46 | <@celticminstrel> | (Teleporters are also easy if they're always an "isthmus", which was what I'd been planning.) |
| 22:53 | <@celticminstrel> | Hmm, my priority queue inserted 3 after 5. Not good. |
| 22:53 | <@celticminstrel> | Oh, wait. |
| 22:53 | <@celticminstrel> | It then inserted 8 before 5. |
| 22:53 | <@celticminstrel> | Maybe I just have the comparison backwards. |
| 22:56 | <@celticminstrel> | No, that's not it. Maybe I want to insert at mid instead of mid+1. |
| 22:57 | <@celticminstrel> | :| |
| 23:05 | <@celticminstrel> | ... |
| 23:06 | <@celticminstrel> | "right = this.priorities - 1" |
| 23:07 | | JustBob [justbob@Nightstar.Customer.Dissatisfaction.Administrator] has quit [Ping timeout: 121 seconds] |
| 23:16 | <@celticminstrel> | Okay, aside from that it seems I can't assume either mid or mid+1 but rather need an additional test to determine which. |
| 23:16 | <@celticminstrel> | Works now. |
| 23:26 | | JustBob [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 23:26 | | mode/#code [+o JustBob] by ChanServ |
| --- Log closed Thu Dec 01 00:00:31 2016 |