#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2011-05-31

Timestamps are in GMT/BST.

00:01:00watusimoto has joined
00:49:00raptorthe mountains are covered in snow!!
00:50:00raptorit's not like it's almost summer or anything..
01:04:00watusimotodon't rub it in
01:05:00raptorhot weather?
01:05:00watusimotono, cool and rainy
01:05:00watusimotoas always
01:12:00sam686Sound needs air to travel, can't hear anything without air in space. So that could make Bitfighter sound-less?
01:12:00raptorpretend that you are in the ship when it fires... :]
01:13:00sam686you won't be able to hear explosions and other ships firing?
01:18:00raptormaybe there is more matter in space than we think...
01:20:00watusimoto Quit (Ping timeout: 276 seconds)
01:40:00karamazovapyit was hot like carl here
01:41:00karamazovapyI got a sunburn while I was reclining in the back yard eating hot dogs from grill to bun
01:42:00karamazovapygranted, I had to hide in the basement last night because of a tornado warning, but I'm fine
01:42:00raptor Quit (Ping timeout: 244 seconds)
01:47:00raptor_ has joined
01:47:00ChanServ sets mode +o raptor_
01:48:00raptor_ is now known as raptor
02:01:00Zoomber has joined
02:01:00Zoomber Quit (Remote host closed the connection)
04:11:00watusimoto has joined
04:35:00raptorhi watusimoto, what are you up to?
04:47:00watusimotoreading my employment contract in french
04:47:00watusimotoor rather having it read to me
04:47:00raptorgonig to learn another language?
04:47:00watusimotoI'm going to try to stick to German
04:48:00watusimotomy kids will have to learn french
04:48:00raptorwhat do they think about moving back to EU?
04:48:00watusimotoI think if we could take their guinea pigs it would be easier
04:49:00raptorno rodents?
04:49:00watusimotoit would be awfully expensive, I think
04:49:00watusimotoI think we've foudn them a very good home
04:50:00watusimotook, back to bitfighter
04:50:00raptormy sister brought her rabbot cross-country in her backpack, once
04:50:00raptorrabbit
04:50:00watusimotohmmm... should try that! trying to get the editor to load properly
04:50:00watusimotoaside from walls, I think it's mostly working now
04:51:00watusimotoor at least not obvikously not working
04:51:00raptorhaha
04:51:00watusimotoI'm going to try to get something checked in tonight
04:51:00watusimotoif I don't fall asleep first
04:52:00raptorok cool
04:52:00raptorthen maybe i'll rebase my shared_ptr work
04:52:00raptorand try to get linux at least compiling again...
05:02:00raptornow this is cool: http://www.bbc.co.uk/news/world-asia-pacific-13593871
05:05:00watusimotowow
05:17:00sam686Current repository version in UIEditor have loadLevelFromFile(fileBuffer); but EditorUserInterfase::processLevelLoadLine is blank so nothing loads. Because of class EditorUserInterface : public UserInterface, public LevelLoader, it calls levelLoader, then levelloader calls multiple processLevelLoadLine..
05:18:00sam686there is this line void processLevelLoadLine(U32 argc, U32 id, const char **argv) { /* TODO: Delete this! */}; should only be deleted if you are not using LevelLoader
05:19:00sam686maybe you could fill in most of code from ServerGame::ProcessLevelLoadLine..
05:19:00raptori think watusimoto is actually working on that code
05:20:00sam686i was looking at it..
05:34:00raptorgood night folks
05:35:00raptor Quit (Remote host closed the connection)
08:02:00watusimoto Quit (Ping timeout: 260 seconds)
13:27:00kodax has joined
14:44:00raptor has joined
14:44:00ChanServ sets mode +o raptor
15:20:00watusimoto has joined
15:50:00watusimotook, after reading comments above, I decided to commit what I got working before I fell asleep last night
15:50:00watusimotoloading seems to work with the exception of walls
15:51:00watusimotoloading in the editor, that is. Loading in game should be fully working
15:51:00raptorok
15:51:00watusimotoI should get the walls working tonight.
15:51:00raptordid you push?
15:51:00watusimotoyes
15:51:00watusimotoso you and sam can at least see the loading framework now
15:52:00watusimotowe use the same load routine in game, rather than having separate versions in uiEditor and servergame
15:52:00raptori don't see thpush..
15:52:00watusimotoI just need to override the part that handles the walls because they still need to be handled differently in the editor and in -game
15:53:00watusimotook, shoul dbe there now
15:53:00raptorWOW
15:53:00raptorloads of changed files
15:53:00watusimotonot that many.... did they?
15:53:00watusimotoshould have only been 8-10
15:53:00raptorabout 30..
15:54:00watusimotoreally?
15:54:00raptorhaha, yep
15:54:00watusimotoI don't see that
15:55:00watusimotoah, ok
15:55:00watusimotoa lot of it was adding lines like this:
15:55:00watusimotoParent::onAddedToGame(theGame);
15:56:00BitfighterLogBot - Commit 23435a5456b1 | Author: watusim...@bitfighter.org | Log: Loading in editor works for many items -- but not yet walls
15:56:00BitfighterLogBot - Commit 9e83197ad335 | Author: watusim...@bitfighter.org | Log: Merge
15:56:00watusimotoi.e. not major changes, just small tweaks that affected many objects
15:57:00sam686one thing i may be finding is it may take up more memory hosting and joining in-game then before the major editor changes.
16:05:00watusimotoit might
16:05:00watusimotothough it shouldn't be to omuch
16:06:00watusimotobecause most of what's changing is adding more functions to the objects, so the memory increace would mainly be a bigger vftable
16:06:00watusimotoand that should be measured in the 10s of bytes per object
16:06:00watusimotobut there may also be some memory leaks going on
16:07:00raptorwatusimoto: can you enumerate exactly what isn't working int he editor? So I know what I break when I try to get shared_ptr working again..
16:07:00watusimotoloading walls
16:07:00watusimotoany undo that involves walls
16:07:00raptorhow about editor functionality?
16:07:00raptormoving walls/polywalls works?
16:07:00watusimotoI think so
16:07:00watusimotoI would suggest keeping a reference copy so you can compare
16:08:00watusimotoif you have any questions
16:08:00raptori can't - doesn't compile in linux...
16:08:00watusimotooh, great.... building fails after last merge
16:08:00raptormake clean!
16:09:00watusimotono, that's not the problem
16:09:00watusimotoi may have clobbered a sharedPtr instance
16:11:00watusimotook, looks like that fixed it... will check in
16:11:00raptorhooray!
16:16:00BitfighterLogBot - Commit 6154cf7d73ef | Author: watusim...@bitfighter.org | Log: Fix compile errors
16:18:00sam686in game.h, should mGameObjects be a Vector<GameObject *> or should it be Vector<SavePtr<GameObject> > ? Thats one reason why you might have to remove .getPointer() as merge failed to include my changes to Vector<SavePtr<GameObject> >?
16:18:00raptorwe need to make sure we aren't getting SaftPtr and shared_ptr mixed up
16:18:00raptorSafePtr
16:19:00raptorshared_ptr has the memory clean-up
16:19:00raptorbe back in a bit..
16:24:00watusimotoI think we want GameObjects to be a shared_ptr
16:24:00watusimotoI thinnk that's what I dropped during the merge
16:25:00watusimotobut... I think we'll be getting rid of mGameObjects
16:25:00watusimotoI want to encapsulate all that in the object database
16:26:00watusimotoback in a bit
16:30:00watusimoto Quit (Ping timeout: 240 seconds)
16:52:00sam686 has left
16:52:00Zoomber has joined
16:55:00Zoomberraptor:
16:55:00Zoombermake command now works on my mac, and I have ordered a usb sound card for 3 dollars which can be usefull for other things as well
16:55:00raptorhello
16:56:00Zoomberbluetooth seems to be working better each time I use it
16:56:00Zoomberoh no, i forgot to say hi
16:56:00raptor'make' works? that is surprising
16:56:00raptordid you have to use 'mac ports'?
16:56:00Zoomberyes, im going to see how far i can get with bitfighter
16:57:00raptoryou should know that the codebase right now is broken for anything but windows
16:57:00Zoomberyes, but i have gone back to the working tarball
16:58:00raptorso i would compile at revision bitfighter-015a
16:58:00raptoroh ok
16:58:00watusimoto has joined
16:58:00raptorthat's fine then
16:59:00watusimotoright -- why shouldn't the database keep a list of all objects that can be returned without a spatial query?
16:59:00watusimotoand why should we have to maintain such a list elsewher?
16:59:00watusimotoI want the database to be the goto for all objects
17:00:00raptorsoudns good to me
17:00:00raptorthe database is just a custom hash pointing to all the objects, correct?
17:00:00watusimotothough we may have a couple of different ones -- one for the bot zones, and two others for special wall snapping purposes in the editor
17:00:00watusimotommmmm
17:00:00raptorhash based off of position/type
17:00:00watusimotosort of
17:00:00watusimotoit's a collection of 256 buckets correspoinding to spatial position
17:01:00raptoryes
17:01:00watusimotoan entity will be in one or more buckets depending on where it falls
17:01:00raptorthat's what i mean by 'custom hash'
17:01:00watusimotoI propose making the database also maintain a nonspatial list of all objects
17:01:00raptornot quite direct lookup, but much faster
17:01:00watusimotoright
17:02:00watusimotoso when you make a non-spatial query, it doesn't need to go through the rigamarole of searching all the buckets in the database, as it does now
17:02:00raptoryou mean a bucketless constant-time retreival?
17:02:00watusimotoI mean a list of all objects
17:02:00watusimotoso if you do a spatial query, you use the buckets
17:02:00watusimotoif you do a non-spatial query, you use the list, which will be faster
17:02:00raptorlist = unordered_map?
17:02:00watusimotolist = vector
17:03:00watusimotobut there may be ways to make things even more efficient for specific queries
17:03:00Zoomberhmm
17:04:00Zoomberit seems if you edit your repository summary to have two || before and after a word, google draws a box around the word when you save changes
17:05:00raptorwould we have derivative collections built for specific object types?
17:08:00watusimotoif that made sense
17:08:00watusimotoif we're doing the same query over and over again, and we can optimize it, then we can optimize it
17:16:00Zoomber Quit (Quit: Zoomber)
17:25:00kodax Quit (Ping timeout: 276 seconds)
17:38:00sam686 has joined
18:55:00raptoris there any function that SafePtr does that shared_ptr does not?
18:56:00watusimotosafePtr nicely returns null when the underlying object has been deleted. I don't think shared_ptr does that
18:57:00watusimotoin fact, I think shared-ptr keeps the underlying object from being deleted, which may be good in some cases, or may lead to pseudo-memory leaks in others
18:57:00raptoryes, that
18:57:00raptorthat makes sense
18:59:00watusimotoI think we should treat the database as the ultimate owner of each object
18:59:00watusimotowhen it's deleted from the database, it's deleted
18:59:00watusimotowhile it's in the database, it's valid
18:59:00watusimotoso most things should be weak_ptrs
18:59:00watusimotobut the database itself should be a shared-ptr
19:00:00watusimoto(the shift key on this machine barely works!!)
19:00:00watusimotoat least that's my current thought
19:00:00raptorweak_ptr, i'm unfamiliar with
19:00:00watusimotoit's like a shared_ptr but doesn't count as a reference for keeping the object alive
19:00:00watusimotoit's designed to work with shared_ptr
19:01:00watusimotoso, for example, we have a lot of places where we fill fillVector with pointers from the database
19:01:00watusimotothose pointers may linger until fillVector gets cleared again
19:01:00watusimotobut they shouldn't keep an object alive that would otherwise be deleted
19:02:00watusimotoso those would be weak_ptrs
19:02:00raptorso shared_ptr can destroy the object and we'd still have the weak_ptr around?
19:02:00watusimotoyes
19:04:00raptorok, let me check in the classes for weak_ptr and shared_from_this
19:04:00raptorin preparation
19:04:00watusimotoI forget what shared_from_this is
19:05:00watusimotoas for the weak_ptr, I'm not sure what advantage it gives over just a plain old c pointer
19:05:00watusimotohttp://www.boost.org/doc/libs/1_46_0/libs/smart_ptr/weak_ptr.htm
19:05:00raptorhttp://stackoverflow.com/questions/142391/getting-a-boostshared-ptr-for-this
19:07:00raptorthere are several methods that use 'this' in them, but won't work if it is a shared_ptr
19:07:00raptorthe solution is the 'shared_from_this()' boost helper
19:08:00watusimotook
19:09:00raptori noticed a lot of inherited methods in EditorObject that act on the calling object
19:09:00watusimotothough see the 8pt answer, the second one
19:10:00raptoryes
19:10:00raptori guess we'd have to juggle shared_ptr and '*' a little
19:11:00watusimotoI hope it will be clear based on context.
19:27:00raptorboost also provides unordered_map
19:27:00raptorequivalent to java HashMap
19:52:00watusimotogood
20:24:00koda has joined
23:26:00karamazovapy';s;'s;'s;ahfdhjksda
23:26:00raptorno speaking in regex
23:26:00raptorcan't parse..
23:27:00karamazovapyhttp://www.bbc.co.uk/news/world-asia-pacific-13593871 ...the itsy bitsy spider...
23:27:00karamazovapyI'm in an odd mood
23:32:00raptor Quit (Remote host closed the connection)
23:48:00Zoomber has joined
23:48:00Zoomberkaramazovapy: are you there?
23:48:00karamazovapyI am in the heezy
23:48:00Zoomberzoomber:
23:49:00Zoombermaybe I should create a post that has all the games listed on it, so people can subscribe to new posts there, and therefore get updated when there is a new game?
23:49:00Zoombercracatoa made a post saying he checks email more often than forums
23:51:00karamazovapyor he could just decide to check the forums if he wants to play...
23:52:00koda Quit (Quit: k thx bai)
23:52:00Zoombertrue
23:53:00Zoomberi do think it would be a good idea to have 1 post with all the upcoming games though. as sometimes the team captains forget to post in their team category when they will play next
23:58:00Zoomber�l
23:59:00Zoomberwith that i must go
23:59:00Zoomberi will see you later _K
23:59:00Zoomber Quit (Quit: Zoomber)

Index Search ←Prev date Next date→

These logs were automatically created by BFLogBot on irc.freenode.net.