Timestamps are in GMT/BST.
| 00:23:56 | | kaen Quit (Remote host closed the connection) |
| 00:29:05 | | raptor Quit (Ping timeout: 264 seconds) |
| 00:30:06 | | kaen has joined |
| 00:34:12 | | Watusimoto_ Quit (Ping timeout: 264 seconds) |
| 01:07:11 | | sam686 has joined |
| 01:07:11 | | ChanServ sets mode +v sam686 |
| 01:43:15 | | raptor has joined |
| 01:43:15 | | ChanServ sets mode +o raptor |
| 02:27:14 | | koda Quit (Quit: k thx bai) |
| 02:54:52 | raptor | kaen: poly2tri: http://sam6.25u.com/upload/6screenshot_12.png |
| 02:59:29 | kaen | PP refuses to even operate on levels like that |
| 03:00:04 | kaen | I'm working right now on doing a subtriangulation for circuitous barriers |
| 03:00:51 | kaen | and then passing that into the main Triangulate_EC call |
| 03:01:37 | kaen | also eclipse-cdt is the best IDE I've ever used on linux, from a functionality standpoint |
| 03:02:02 | kaen | Searching for declarations of a symbol consistently works... it's strange. |
| 03:03:04 | raptor | heh |
| 03:03:15 | raptor | and the RAM usage! it's so high! |
| 03:03:21 | raptor | (i use CDT..) |
| 03:03:36 | raptor | also CDT has gotten waaaay better with this last release just a month ago or so |
| 03:04:05 | raptor | also, poly2tri seems good - I remember we had this same problem with triangle.. |
| 03:04:32 | | fordcars_ Quit (Ping timeout: 245 seconds) |
| 03:04:41 | raptor | poly2tri would have failed when we first implemented Triangle because we weren't using the clipper lib to merge polys beforehand |
| 03:05:04 | raptor | but now that we are, I think that bug with multiple points on the same line won't effect us.. |
| 03:05:58 | raptor | also what on earth is a steiner point - i keep getting vague language in my research |
| 03:11:20 | raptor | sam686: do you remember when we had this problem with Triangle?: http://sam6.25u.com/upload/6screenshot_12.png |
| 03:11:26 | raptor | and do you remember how we solved it? |
| 03:24:53 | raptor | aww man |
| 03:25:13 | raptor | handling enclosed holes is something only Triangle might actually do... |
| 03:38:57 | raptor | ok, I'm beginning to thing that the Triangle library is just magic |
| 03:39:01 | raptor | *think |
| 03:40:58 | kaen | hmm... that would explain why it's written in archaic runes |
| 03:44:54 | raptor | so the output from merge polys of this map is two concentric polygons: http://sam6.25u.com/upload/5screenshot_13.png |
| 03:46:07 | raptor | which makes sense |
| 03:46:17 | raptor | but somehow Triangle works and nothing else does |
| 03:47:41 | raptor | mergePolys merges all the 'holes' |
| 03:48:20 | raptor | I don't even know how we could tell if something was a continuous hole |
| 04:08:16 | raptor | ok |
| 04:08:43 | raptor | it looks like the output polygons are wound in opposite directions |
| 04:19:03 | raptor | CGAL is commonly referenced... but my goodness, I don't think a library could be documented less for the layman than that one |
| 04:20:54 | raptor | I'm giving up until we hear back from Triangle dude |
| 04:21:02 | raptor | on to greater things! |
| 04:36:26 | raptor | sooo.. bed! |
| 04:36:29 | | raptor Quit () |
| 06:48:22 | | sam686 Quit (Ping timeout: 245 seconds) |
| 07:09:02 | | Watusimoto has joined |
| 07:22:22 | | Watusimoto Quit (Ping timeout: 264 seconds) |
| 07:22:52 | | koda has joined |
| 09:55:54 | | Platskies has joined |
| 10:00:02 | | Platskies Quit (Ping timeout: 240 seconds) |
| 10:07:55 | | Platskies has joined |
| 11:31:24 | | kodab has joined |
| 11:33:47 | | koda Quit (Ping timeout: 245 seconds) |
| 11:36:17 | | kodab Quit (Ping timeout: 245 seconds) |
| 12:21:07 | | koda has joined |
| 15:05:07 | | Platskies Quit (Remote host closed the connection) |
| 17:08:28 | | watusimoto1 Quit (Quit: Leaving.) |
| 17:13:11 | | raptor has joined |
| 17:13:11 | | ChanServ sets mode +o raptor |
| 17:25:38 | | Watusimoto has joined |
| 17:32:50 | raptor | lessons learned from last night: Triangle is magic |
| 17:33:03 | raptor | seems like I need to relearn that every year or so... |
| 17:34:11 | | koda Quit (Ping timeout: 245 seconds) |
| 17:35:52 | Watusimoto | no word from mr. triangle |
| 17:36:59 | raptor | I mean, it's magic! |
| 17:37:16 | raptor | I actually got poly2tri hooked up in under an hour: http://sam6.25u.com/upload/6screenshot_12.png |
| 17:37:34 | raptor | it was super fast and just as good for what it could do |
| 17:37:53 | raptor | but it can't seem to handle a 'continuous' hole |
| 17:38:02 | raptor | like the outside barrier of a map |
| 17:38:16 | raptor | in fact - *nothing* seems to be able to handle it, except for Triangle |
| 17:38:28 | raptor | from my research (and kaen's) |
| 17:49:14 | Watusimoto | I don't entirely follow -- what do you mean by a conitnuous hole? |
| 17:49:27 | raptor | http://sam6.25u.com/upload/5screenshot_13.png |
| 17:49:33 | raptor | the wall |
| 17:49:36 | raptor | is a hole |
| 17:49:40 | raptor | but a donut, too |
| 17:52:26 | raptor | maybe 'donut' is a better word |
| 17:52:39 | Watusimoto | ok |
| 17:52:47 | raptor | but Triangle is magic |
| 17:52:52 | Watusimoto | so they don't handle islands |
| 17:53:01 | Watusimoto | I think is the "technical" term |
| 17:53:05 | Watusimoto | technicalish |
| 17:53:09 | Watusimoto | island polygons |
| 17:53:19 | Watusimoto | or, rather, complex polytongs |
| 17:53:20 | raptor | poly2tri is blazing fast... maybe faster than Triangle - but not as robust, obviously |
| 17:53:23 | Watusimoto | they handle simple polygons |
| 17:53:26 | raptor | what? |
| 17:53:35 | Watusimoto | i think a complex poly is one with islands |
| 17:53:52 | Watusimoto | and a simple poly is one with just a border (of whatever complexity) |
| 17:54:22 | raptor | complex polys have lines that cross themselves |
| 17:54:35 | Watusimoto | so an outer area walled off from an inner area would create a complex polygon |
| 17:55:12 | raptor | i wonder if there'd be a way to detect things like this... |
| 17:55:19 | raptor | there's got to be |
| 17:55:56 | Watusimoto | your terminilogy seems to have some external support |
| 17:56:15 | Watusimoto | remmeber that we could have many levels of islands within islands |
| 17:56:20 | raptor | yep |
| 17:56:26 | raptor | and Triangle is magic |
| 17:57:20 | Watusimoto | so if you could identify the islands... then what... triangle those separately? |
| 17:57:38 | Watusimoto | then the islands within the islands... and so on |
| 17:57:53 | raptor | yeah, add the interriors as a new exterior polyline to triangulate |
| 17:59:15 | Watusimoto | so does poly2tri mention islands at all? |
| 17:59:19 | raptor | none |
| 17:59:30 | raptor | and I just posted this: http://code.google.com/p/poly2tri/issues/detail?id=74 |
| 18:00:39 | Watusimoto | so clipper handles making things wind in the proper order? |
| 18:00:44 | Watusimoto | that's actually a big advantage |
| 18:00:46 | raptor | yes |
| 18:00:50 | raptor | clipper is amazing |
| 18:00:58 | raptor | it sanitizes all of our inputs |
| 18:01:00 | Watusimoto | your questino is good |
| 18:01:14 | Watusimoto | then the solution seems simple |
| 18:01:17 | Watusimoto | conceptually |
| 18:01:25 | raptor | in fact since we've added clipper we could probably get rid of some of our own sanitization stuff.. |
| 18:01:27 | Watusimoto | triangluate every polygon wound cw |
| 18:01:29 | Watusimoto | or ccw |
| 18:01:42 | raptor | wait what? |
| 18:01:59 | Watusimoto | imagine a set of concentric polygons |
| 18:02:04 | raptor | ok |
| 18:02:08 | Watusimoto | that represent a worst case scenario |
| 18:02:13 | raptor | heh, yes |
| 18:02:15 | Watusimoto | the outer one will be wound cw |
| 18:02:18 | Watusimoto | the next ccw |
| 18:02:20 | Watusimoto | the next cw |
| 18:02:21 | Watusimoto | etc. |
| 18:02:24 | raptor | ok |
| 18:02:25 | Watusimoto | with me so far? |
| 18:02:28 | raptor | yep |
| 18:02:30 | Watusimoto | ok |
| 18:02:41 | raptor | ahhhh, i see |
| 18:02:43 | Watusimoto | so trianluate the outer one |
| 18:02:52 | Watusimoto | got it? |
| 18:02:54 | raptor | yep |
| 18:03:07 | raptor | so the trick would be how to find what is in what hole |
| 18:03:16 | Watusimoto | you don't need to know that |
| 18:03:25 | Watusimoto | I don't think |
| 18:03:44 | raptor | each triangulation of a ccw one requires it knows the holes of anything inside, right? |
| 18:03:59 | Watusimoto | what if you fed the triangluation holes that were not actually contained in the poly? |
| 18:04:07 | Watusimoto | would it work, or just die? |
| 18:04:13 | raptor | it would probably just work |
| 18:04:24 | Watusimoto | then feed it everything. it might work |
| 18:04:24 | raptor | actually, not sure - but we could always do bounds checking.. |
| 18:04:33 | Watusimoto | I'd try wihtout that first |
| 18:04:41 | raptor | hmmm.... |
| 18:04:42 | Watusimoto | take our concentric polys again |
| 18:04:50 | Watusimoto | the outer one is A, the next one is B, etc. |
| 18:04:54 | Watusimoto | A is easy |
| 18:04:57 | Watusimoto | feed it everything |
| 18:05:03 | Watusimoto | you get triangles btwn A and B |
| 18:05:12 | Watusimoto | B is wound wrong so you skip it |
| 18:05:17 | Watusimoto | next comes C |
| 18:05:23 | Watusimoto | you feed it everything |
| 18:05:29 | raptor | yes |
| 18:05:30 | Watusimoto | what would it do? |
| 18:05:33 | Watusimoto | A and B are completely outside C |
| 18:05:35 | Watusimoto | so might be ignored |
| 18:05:47 | Watusimoto | I would hope you'd get the space btwn C and D |
| 18:06:02 | raptor | ok yes |
| 18:06:09 | Watusimoto | we have no intersections |
| 18:06:14 | Watusimoto | so our input is clean in that sense |
| 18:06:18 | raptor | yay clipper |
| 18:06:22 | Watusimoto | :-) |
| 18:06:34 | Watusimoto | that would be the next thing to try, anyway |
| 18:06:46 | Watusimoto | and should be relatively easy |
| 18:06:50 | raptor | I guess my brain was thinking about performance too much :) |
| 18:07:15 | Watusimoto | idelaly, clipper would give us a way to distinguish the CW polys from the CCW ones |
| 18:07:28 | Watusimoto | so we don't need to examine each |
| 18:08:00 | raptor | there is a isWoundClockWise method or something in CLipper |
| 18:08:12 | raptor | because it just spits out points |
| 18:09:47 | Watusimoto | that should work |
| 18:09:54 | Watusimoto | so long as we don't need to visit each point |
| 18:10:56 | Watusimoto | on your question, you might want to clarify (if you can at this point) that you are hoping for triangles on the entire black area of the image |
| 18:11:02 | Watusimoto | at least the inner area |
| 18:11:14 | raptor | i didn't write that? let me check |
| 18:11:38 | raptor | ah, i guess not so explicitly |
| 18:11:51 | Watusimoto | you sort of did |
| 18:12:06 | Watusimoto | just thinking peopel don't read questions before answering! |
| 18:12:26 | Watusimoto | not a huge deal, just might make the image clearer |
| 18:12:54 | Watusimoto | because at a glance, it seems to ahve worked inversely... triangles outside the poly, no triangles in |
| 18:13:02 | Watusimoto | which is a completely wrong interpretation of the image |
| 18:15:33 | raptor | actually... |
| 18:15:40 | Watusimoto | I have a whole page of mapping out how our player names and passwords get set |
| 18:15:42 | Watusimoto | what a mess |
| 18:15:49 | Watusimoto | everything calls soemthign else |
| 18:15:51 | Watusimoto | ]blech |
| 18:15:58 | Watusimoto | deleting functions like crazy! |
| 18:15:59 | raptor | if we just always triangulate ccw walls... then maybe we don't need the outer rect! |
| 18:16:10 | raptor | actually no - many maps are open |
| 18:16:15 | Watusimoto | we would if there were an exit to the outside |
| 18:16:28 | raptor | oh yeah, the pasword/playername stuff is nuts |
| 18:16:30 | Watusimoto | which is why we have that outer rect in the first place |
| 18:16:37 | Watusimoto | trying to clarify it |
| 18:19:08 | raptor | oh good |
| 18:19:13 | raptor | also, yay for options menus! |
| 18:19:27 | raptor | that might be an old bug that you solved... |
| 18:23:51 | Watusimoto | with just a tiny bit of duplicated code, things become much clearer |
| 18:31:37 | | sam686 has joined |
| 18:31:37 | | ChanServ sets mode +v sam686 |
| 18:46:37 | Watusimoto | dinner... back later |
| 18:51:08 | | Watusimoto Quit (Ping timeout: 245 seconds) |
| 18:58:38 | | bobdaduck has joined |
| 19:40:16 | bobdaduck | Any progress on the death forcefield snapping bug? |
| 19:40:30 | raptor | http://code.google.com/p/bitfighter/source/detail?r=84ed46702cde5e0c253cfbb62bf8cedc19d56886 |
| 19:40:50 | bobdaduck | Whoo! |
| 19:40:59 | bobdaduck | What was the problem? How'd you fix? |
| 19:41:37 | raptor | i could explain - or I could defer to the code changes I made in which I document it... mut maybe it's best summed up by one word: fuzzy math |
| 19:41:42 | raptor | :) |
| 19:42:01 | bobdaduck | lol |
| 19:42:11 | raptor | also, I have a meeting - back later |
| 19:42:15 | bobdaduck | k |
| 19:50:50 | | Watusimoto has joined |
| 19:51:07 | bobdaduck | So what's all this I hear about triangles being magic? |
| 19:55:42 | Watusimoto | 3-4-5 |
| 19:57:51 | bobdaduck | pathagorean theorem |
| 19:58:46 | bobdaduck | ? |
| 19:59:01 | bobdaduck | Also serious question: How is /nextmap determined? |
| 20:06:47 | raptor | /next just pulls the next map, alphabetically by filename |
| 20:06:59 | raptor | unless it fails to meet certain criteria |
| 20:07:08 | raptor | like if it provides min/max players |
| 20:08:20 | bobdaduck | So we could potentially set up a map order... |
| 20:08:38 | bobdaduck | Oh by the way |
| 20:08:39 | raptor | actually yes - I did that somehow with the BBB |
| 20:08:43 | bobdaduck | BBBX map list complete |
| 20:08:52 | raptor | oops, gotta run - meeting! |
| 20:09:07 | raptor | yay, also I think we should not do endless war |
| 20:09:20 | bobdaduck | I will win that debate. |
| 20:09:23 | raptor | now we just need to start editing maps for scores/etc and set up an order |
| 20:09:33 | raptor | back later |
| 20:09:38 | raptor | and Triangle is magic! |
| 20:09:59 | bobdaduck | lol |
| 21:08:30 | | Watusimoto_ has joined |
| 21:09:16 | | Watusimoto Quit (Ping timeout: 264 seconds) |
| 21:16:13 | | BFLogBot Commit: 147dd6dabb50 | Author: watusimoto | Message: Rather sprawling update that simplifies the way usernames and passwords are stored/managed |
| 21:48:14 | bobdaduck | So wassup with triangles? |
| 22:35:25 | raptor | hi |
| 22:40:36 | kaen | hello |
| 22:40:49 | raptor | hey kaen |
| 22:41:02 | raptor | you've probably read my triangulation monologues |
| 22:41:10 | kaen | I believe so |
| 22:41:27 | kaen | I'm just about to give subtriangulation with PP a try |
| 22:41:47 | raptor | what does that mean? |
| 22:42:07 | raptor | a different algo, or finding the internal hole holes that need to be triangulated? |
| 22:42:44 | kaen | take the outside polygon of circuitous barriers as a polygon and in the inside polygon as hole, run it through PP, and the resulting triangles to the main triangulation call as holes |
| 22:43:15 | raptor | huh |
| 22:43:32 | raptor | 'circuit'! that's the word! |
| 22:43:40 | kaen | i just got off of a 12 hour shift so that came out kind of fuzzy... |
| 22:43:49 | raptor | no i get it |
| 22:44:06 | kaen | very useful that the metaholes are wound in the opposite direction |
| 22:44:09 | raptor | also, if you didn't see, I opened an issue with poly2tri about that: http://code.google.com/p/poly2tri/issues/detail?id=74 |
| 22:44:14 | raptor | ah, you did see :) |
| 22:49:12 | | SolumnMushroom has joined |
| 22:49:24 | SolumnMushroom | Hello all! |
| 22:50:38 | bobdaduck | sup mush |
| 22:53:24 | | SolumnMushroom Quit (Client Quit) |
| 22:55:35 | raptor | heading home! |
| 22:56:16 | raptor | kaen: we should compare notes later because I was planning on implementing your idea somehow with poly2tri, as well... |
| 22:56:52 | raptor | later! |
| 22:56:54 | | raptor Quit () |
| 22:58:21 | | SolumnMushroom has joined |
| 22:58:38 | SolumnMushroom | Sorry about the abrupt exit |
| 22:59:08 | SolumnMushroom | The library computer I was on was not making my left ear happy... |
| 23:40:25 | | BFLogBot Commit: a4e8db09b7eb | Author: watusimoto | Message: Spelling |
| 23:40:26 | | BFLogBot Commit: ca97ac38afbf | Author: watusimoto | Message: Shuffle things around a bit to be more coherent and more isolated |
| 23:40:28 | | BFLogBot Commit: ae027fdd5f97 | Author: watusimoto | Message: More shuffling and isolating... UIManager completely out of GameType, still a little UI stuff left in there, but not much |
| 23:43:37 | | BFLogBot Commit: 7f0e41c175f0 | Author: watusimoto | Message: Pass by ref |
| 23:44:06 | | Little_Apple has joined |
| 23:48:16 | | Little_Apple Quit (Ping timeout: 245 seconds) |
| 23:54:27 | SolumnMushroom | Goodbye |
| 23:54:30 | | SolumnMushroom Quit (Quit: Page closed) |
