#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2015-01-17

Timestamps are in GMT/BST.

00:36:39raptorkaen: do you remember what triggered a shader in-game?
00:42:14kaenumm
00:42:19kaenoh
00:42:21kaenheh...
00:42:24kaenanother patch
00:42:36kaenthat's just the framework, I had a test shader
00:43:04raptori found the explosion one
00:43:15raptorbut somehow the code to use it was lost ...
00:45:59kaenah
00:46:43kaendoes this apply by chance?
00:46:43kaenhttp://pastie.org/pastes/6101893/text
00:47:09kaenshould be a fairly trivial merge if not
00:47:16raptorwhy yes!
00:47:21kaenalso shows the steps needed to create and use a shader
00:47:38kaenand how to set a uniform
00:47:43kaengreat example, kaen from the past!
00:48:03kaenI guess the ship should change colors
00:48:09kaenbased on how far away the mouse is
00:54:00kaenany luck?
00:54:19raptorit compiles... and i forgot to add the shader directory to where the exe could see it..
00:54:52kaenah yeah
00:55:28kaenshould be able to edit shaders on the fly, ctrl+0 should reload them
00:58:00raptorhmmm... i get no errors now, but nothing is altering the view of the ship
00:58:11kaenhmm
00:59:02kaenany error output in the console?
00:59:12kaenoh
00:59:14kaenguess not
00:59:16raptorthere was when i removed the NULL from the arrays..
00:59:30raptorso i know they're being loaded... once i added that back in
00:59:38kaenok
01:00:17raptoroh wait!
01:00:27raptor"fades between black and red"
01:00:36raptorthat may have been subtle enough that i didn't notice..
01:03:25raptornope, not working :(
01:03:31raptorthat's ok, i'll pick up tomorrow
01:03:37raptorneed to sleep
01:03:41raptorthanks for your help!
01:04:57Nothing_Muchwoo
01:05:26kaenyou too, night!
01:06:33raptor Quit ()
01:15:54Nothing_Muchgood evening, everyone
01:15:59Nothing_Muchor anyone
01:48:53kaenhuh
01:49:03kaenI think I've just built a debian package for 019d
02:33:23Nothing_Muchkaen: that's a good thing!
02:33:50Nothing_Much091d is the latest version!
02:38:46kaenjust one last little change is all it needs
02:39:59Nothing_Muchwhoa, 019d i meant
02:40:11Nothing_Muchwhat change?
02:40:48kaenjust needs to look ok with the debian fonts
02:41:37kaenI literally just need to change the size of two of the fonts
02:41:47Nothing_Muchoh cool
11:07:39kaen Quit (Quit: Leaving)
11:08:02kaen has joined
11:08:02kaen Quit (Changing host)
11:08:02kaen has joined
14:04:27kaenI DID IT!!!
14:04:43kaenI got the shader patch working
15:33:49fordcars has joined
16:36:41fordcarsIn Windows Visual Studio 2010, I had to add /NODEFAULTLIB for a bunch of libraries to make it compile
16:36:58fordcarsI also had to compile as release only
17:14:11sam686 has left #bitfighter
19:00:01raptor has joined
19:00:01ChanServ sets mode +o
19:00:03raptorgood day!
19:01:34fordcarsHi!\
19:01:56kaenhi!
19:02:05kaenraptor, http://imgbin.org/index.php?page=image&id=21857
19:02:17raptorWHAT
19:02:21raptorhow did you...
19:02:34kaenI got that patch to work
19:02:42kaenit required a fair amount of hacky hacky
19:02:43fordcarsWHOAAA
19:03:05raptorcan you please send me a diff?
19:03:08kaenyes
19:03:13kaenquestion
19:03:15raptori was working on that all last night and got stuck
19:03:17raptorsure
19:03:26kaenI messed up and actually applied that to tip
19:03:42kaenshould I just make a clone again?
19:03:45kaenor
19:03:52raptordid you commit?
19:03:54kaenhow do I make a diff that includes the untracked files
19:03:54kaenno
19:04:04raptoroh I do this:
19:04:15raptorhg add <untracked_files>
19:04:20raptorso then everything is tracked
19:04:23kaenoh derp
19:04:31raptorthen: hg diff > ../shaders_awesome.diff
19:04:36raptorthen: hg revert -a
19:04:47raptorthat's 'revert all'
19:05:14raptorso before you revert all, make sure you didn't have other changes you wanted to keep
19:05:47raptoroh, you could also do 'hg shelve' to stash away the changes internally
19:05:55raptorinstead of just reverting everything
19:05:57kaenhttp://pastie.org/9837609
19:06:10kaenit's not quite perfect
19:06:17kaenyou'll have to play in windowed mode
19:06:22kaenand set the scale factor
19:06:23kaento 1.0
19:06:46kaenfor any shader that uses worldToScreen() anyway
19:06:59kaen(which includes the phaser shader)
19:07:04fordcarsraptor: windows visual studio has a conflicting libraries
19:08:08kaenI also got this to 99% https://github.com/bitfighter/bitfighter-debian
19:08:23kaenonly thing left is to the fix the size we render debian's play at
19:08:30kaenother than weird looking text, it works perfectly
19:09:21kaenand I learned that bitfighter is no longer one of "my" packages on mentors.debian
19:09:23raptorcan you actually upload the diff file - pastebin places seem to corrupt diffs. I like to use: http://sam6.25u.com/upload3.php
19:09:26kaenso I can't upload there
19:09:30kaensure thing
19:09:52kaenhttp://sam6.25u.com/upload/shaders_new.diff
19:09:59kaenremember that this is based against tip
19:10:30raptorok, great
19:10:32raptorwow, thanks
19:10:53raptorfordcars: did you remove all cmake junk and regenerate?
19:11:07raptorlike in the build directory: rm -r CMake*
19:11:16fordcarsraptor: yeah, it was brand new, I added some NODEFAULT and had to compile as releae
19:11:16raptorkaen: i wonder if we can set multiple owners
19:11:27raptorwhat version of vc++?
19:11:31fordcars2010
19:11:50fordcarsJust letting you know though
19:13:26fordcarsOh and, I am getting this now, I added stuff to renderInterfaceOverlay(). Bitfighter crashes and spits this out in command line: http://pastie.org/9837610
19:13:54fordcars^^ I added this line:
19:13:54fordcarscolorLogitechTeamColor(mLevel->getTeamColor(mGame->getLocalPlayerShip()->getTeam()));
19:14:03fordcarsThat function works when I call it somewhere else
19:14:35raptorwhoa... fordcars, what have you done to your system??
19:14:53fordcarsIPlaying around with some dlls
19:15:03raptoris there any other output? none of that looks bitfighter related
19:15:15fordcarsYes, wait
19:15:50fordcarsOh wait, I get SIGSEV
19:15:58fordcars*SIGSEGV
19:16:07raptoroh... then that means null pointer!
19:16:18fordcarsWhole thing: http://pastie.org/9837623
19:16:38fordcarsWell, I don't think it's me, but some API thing for this keyboard
19:16:46fordcarsI don't use pointers :)
19:16:52fordcarsI must of messed up
19:17:27kaenthat sounds likely
19:17:29raptorsadly... that actually isn't a stack trace
19:17:37raptorso i can't help much
19:17:41fordcars:(
19:17:51kaencrashes on startup?
19:19:20kaenoh
19:19:24fordcarsNope, when I get into that function
19:19:27kaenthere's a good chance
19:19:27fordcarsSo uh
19:19:32kaenthat' localplayership is null
19:19:36kaenfor at least a few frames
19:19:42raptoroh yeah, when you die
19:19:50fordcarsOh man
19:19:56raptortry using the clientinfo instead
19:20:49fordcarsraptor: can I get the team color from ther?
19:21:04raptoryou can get the team
19:21:15raptorthen use that inside of getTeamColor
19:22:55fordcarsraptor: getLocalRemoteClient?
19:25:00kaenany trouble with the shader patch raptor?
19:25:30fordcarsgtg later!
19:26:00raptorkaen: sorry, got caught in debian research...
19:26:06raptorlet me ttry now
19:26:35kaenok, I have to do some chores but I'll be in and out
19:27:39raptorok... if i have to upload bitfighter, can i pull from your git thingy now?
19:28:01raptoralso - i say we leave the font size alone and use it to get them to fix the fonts-play in debian
19:28:10raptoror at least, leave it alone for now...
19:29:11kaenagreed
19:29:18kaenyep, you can pull
19:29:22raptorok
19:29:28kaenthen use git-buildpackage
19:29:50kaenI just used this guide: http://documentation.thebigrede.net/maintainers/howto-build-package.html
19:31:51raptorit compiled!
19:32:11kaendid you use git-pbuilder ?
19:32:17kaenor some other pbuilder?
19:32:49raptoroh no.. sorry, i'm on the shader patch (too many things at once)
19:32:58kaenoh haha
19:33:00kaenno problem
19:33:25kaenthis should get you a pbuild of bitfighter for sid after you clone: http://pastie.org/9837632
19:33:56kaenoh, are you doing the packaging on suse?
19:34:04kaenoh in a debian vm?
19:34:08raptori have a debian testing vm, yes
19:34:13kaenor in*
19:34:13kaenok
19:34:21kaencool, that works for me on ubuntu
19:35:00raptordo i need to explicitly enable shaders somehow?
19:35:27kaenno, should happen automatically
19:35:34kaendo you get any error output?
19:35:54raptoroh wait, i did: Unable to load "default.vert"
19:35:58kaenah
19:36:02raptornothing else... i wonder why
19:36:10kaenit will give up after that
19:36:15raptorlast night i was not getting error output yet no shaders
19:36:21raptorbut maybe i was just tired
19:36:25kaenit was likely not even trying to load them
19:36:33raptorso what would be common issues in this instance?
19:36:51kaendid you symlink resource/shaders ?
19:37:03raptor hides in shame
19:37:14kaenlol, I did exactly the same thing
19:37:28kaenin fact, I symlinked resource/shader after that (no s)
19:37:49kaenand went all the way into gdb to figure out my mistake
19:38:15raptorone shader works, the other not. and i got a GL ERROR: 1282
19:38:26kaenbummer :<
19:38:29kaenwhich does not?
19:39:15raptorhttps://i.imgur.com/cnPBMKm.png
19:39:16raptorbrb
19:39:19raptorthe phaser didn't
19:39:39kaendid you set windowscalingfactor to 1 in the ini?
19:40:18kaenthe glsl helper it uses doesn't understand that factor quite yet, so it has to be 1.0
19:49:19raptorohhh... then it works!
19:49:23raptorsweet!
19:49:30raptorok, dinner time
19:49:33raptorback later
20:07:22fordcarsOk I got it; mGame->getLocalRemoteClientInfo(); returns a NULL pointer, always
20:14:55amgine123 has joined
20:38:01raptorkaen: did you use your misc/*.sh scripts to build with pbuilder?
20:38:11raptorthey still say 019c in that repo..
20:39:31kaennope
20:39:36kaenI should actually nuke those
20:54:33raptoroh wow, you have several patches in debian/patches/
20:54:41raptor(i fell asleep on the floor)
21:01:53kaenhaha
21:02:12amgine123anything to test ?
21:02:53amgine123I think i found the cause of the gltich for walls not drwaing raptor il prusue it some more
21:03:18raptoramgine123: if you can duplicate it, i'm interested
21:03:46kaenhmm, have to reverse engineer bitfighter's camera system
21:04:27amgine123hmm is duplicatable if there is a specific set of steps to do it but it takes about 100 tries to get it to work ?
21:04:35amgine123seems random
21:04:41amgine123wonder if i can trace it
21:04:50raptorhaha - yeah, that's not quite duplicatable
21:05:08amgine123ah nuts dead end i was sure i corned the bug
21:05:13amgine123nevmind sorry
21:05:56amgine123btw ive made some slight progress on getting the ini settings to crash when hosting problem
21:06:11amgine123ive cornered it down to 1 section
21:06:24amgine123i jsut need to test the settings 1 by 1 and find the conflict
21:10:13sam686 has joined
21:10:13ChanServ sets mode +v
21:12:18amgine123for goodness sake i tested it and ive ruled out every INI setting cuasing a crash even though i know there is a crash somewhere
21:35:31fordcarsraptor: where can I get 019d hash?
21:35:36fordcarsFor hg, whatever you call it
21:35:41raptorit's a tag
21:35:46raptorbitfighter-019d
21:35:54fordcarsOhh right sorry
21:35:56fordcarsThanks
21:40:47sam686 has left #bitfighter
21:48:25sam686 has joined
21:48:25ChanServ sets mode +v
21:49:20raptorkaen: I was thinking we should do the non-native debian route (even though we are upstream) and just check in the 'debian' directory sources
21:49:36raptori'm re-reading that tutorial more-in-depth now
21:49:50kaenI must have missed something
21:50:25raptorthe distinction between native and non-native is in the shorter sections at the top
21:50:33kaenI'll reread it
21:50:39raptorwe're sort of in the middle
21:50:44kaenbut I think I did check in the debian directory
21:50:52kaenI think we're definitely non-native
21:50:54raptoryes, you did
21:51:05kaennative is for like debian-specific things right?
21:51:12raptoryes
21:51:45kaenok
21:51:46raptori'm only mentioning this because the workflow is slightly different for non-native, and it allows us to not need to check in the source into giyhub
21:51:48raptor*github
21:52:14raptorwhich i think i may prefer as having the source there can be... messy
21:53:43kaenhmm, I followed the steps under the subsection "non-native Debian" in the "with git" section
21:54:10kaen"Creating a repository and Debian package from scratch:"
21:54:19raptorok... i'm not quite there yet
21:54:32raptormaybe i misunderstood
21:54:36kaenoh wait
21:54:41kaenno, not starting with that quote
21:54:59kaen"In this case, the unmodified upstream sources are contained in the upstream branch (will be imported from an already existing .tar file, provided by the upstream authors) and all Debian modifications will be on the master branch.
21:54:59kaen"
21:55:03kaen^ that's the one
21:55:19raptorok, yes, that's where i am
21:55:27kaenok
21:55:37raptorthis flow lets us not check in all the bitfighter source, which i think may be cleaner for maintenance
21:56:01kaenit did have me import a source tarball
21:56:22kaenwhere does it mention not checking in the source?
21:56:25raptoroh? i'll just be quiet then until i finish reading... :)
21:57:01kaenI'll admit I didn't read it very thoroughly
21:57:10kaenso I'm genuinely not sure either :)
21:58:10raptorok first paragraph: "In this case, the unmodified upstream sources are contained in the upstream branch (will be imported from an already existing .tar file, provided by the upstream authors) and all Debian modifications will be on the master branch."
21:58:18raptoruhh... what you pasted above
21:58:19raptorok
21:58:39raptori see my confusion - did you commit all things to the 'master' branch? including upstream sources?
22:01:37kaenI committed everything required to build
22:01:42kaenwas that a mistake?
22:02:03raptornot sure yet. must read...
22:02:11kaenI then committed the different patchs, including the debian/patches filed generated by dpkg-source --commit
22:02:17kaenon master
22:03:35kaenoh, it sort of cuts off and has a stint about rubygems
22:03:41kaenbut picks up again at "Patching the source (only non-native Debian sources)"
22:04:16kaenah, and I also used the "Typical development-build workflow"
22:04:27kaenwhich describes what gets committed where
22:09:26kaenoops
22:09:28kaenman
22:09:36kaenhttps://wiki.debian.org/PackagingWithGit
22:09:45kaen"the tarball was repacked (to remove non-DFSG pieces, for instance)"
22:09:59kaenthere's a special process for repacking a tar to remove cruft like the windows libs
22:10:08kaenI missed it
22:10:31kaenbut, that's a separate commit in quilt
22:10:45kaenso we can repack, omit that commit, and import the history over top
22:11:05kaenor diff, squash, and blow away the old repo
22:12:25kaenhmm, maybe not
22:12:31kaengah, I can't tell if I did it right
22:12:48kaenthe only thing I'm unsure on is how I removed the "non DFSG" stuff
22:13:01kaeni.e. our libs and fonts that debian already carries
22:13:44kaenbut it builds happily with the git pbuilder workflow, and doesn't mention anything about it on lintian
22:13:45raptorman... i keep getting distracted
22:14:25raptorok, so i spoke to debian-mentors, they say I can just delete my bitfighter package on mentors.debian.net and then you can upload yours
22:14:27raptorwant to do that?
22:14:36kaenok
22:14:40raptorok, let me delete...
22:15:22raptordone
22:15:27raptoryou may attempt to upload :)
22:15:35kaencan you refresh me on the process?
22:15:48raptordput mentors package_blah.changes
22:16:33raptorread here under 'how to upload packages to mentors'
22:16:35raptorhttps://mentors.debian.net/intro-maintainers
22:16:56raptoroh oops, you need to set up a config file first; that section says how
22:17:15kaenok, thanks
22:20:54amgine123well that sucks bitfighter crashed in the middle of the level while i was playingwith some bots
22:22:47kaenhttp://pastie.org/9837782
22:22:54kaenlooks ok, don't see it under my account though
22:22:57kaendoes it take a while?
22:23:04raptorit takes about 15 min
22:23:31kaengroovy
22:23:35kaenfingers crossed then
22:24:00raptoris it just me or does all of debian feel like it is run by a bunch of cgi perl scripts
22:24:16kaenthat's absolutely what I expect
22:24:31kaenrunning on cron jobs
22:26:43kaenhuh
22:26:47kaenapparently it's python: https://alioth.debian.org/scm/browser.php?group_id=100127
22:27:18kaen(clicked on the "source code and bugs" link)
22:27:22raptorwell that's... unexpected
22:30:13raptoruhh: https://mentors.debian.net/package/bitfighter
22:30:24raptorthat looks suspiciously like the exact same data from my package
22:30:48raptorI suspect you need to change the build number to 019d-3, and reupload :-/
22:31:08kaenbah I think you're right
22:31:14fordcarsraptor: does this mean we will be able to apt-get bitfighter?
22:31:30raptorfordcars: if all goes well... maybe in a year or so :)
22:31:37fordcarsHaha good :)
22:34:53kaenwhat a silly thing to rebuild for .-.
22:35:22raptoryeah, it got me the first time...
22:39:54kaenoff-topic: do you know how I would translate a bfobject world position to a physical pixel coordinate?
22:40:05kaenas opposed to a "canvas" coordinate
22:42:01raptorget the ship coords
22:42:17raptorthe screen is the visual distance from the ship
22:42:25raptorthat might be set in Game.h?
22:42:29raptori can't remember...
22:42:58raptor*game.h
22:43:12raptorthere's PLAYER_VISUAL_DISTANCE_HORIZONTAL and VERTICAL
22:43:18sam686hint: the editor does the coordinates conversion for the mouse pointer
22:43:38raptoror just listen to sam686... he knows things
22:43:45raptor:)
22:44:58sam686there is convertCanvasToLevelCoord in editor
22:45:05kaenI need to go in the opposite direction though
22:45:15kaenworld coord -> physical pixel
22:45:27kaenwhich should just be the MVP matrix right?
22:45:44kaendon't we send world coords directly to OGL?
22:46:24sam686matrix is indirectly sent to opengl, as glScale/rotate/translate is used as editing the matrix
22:46:30kaenyes
22:46:38kaenand that matrix is available in the shaders I'm writing
22:47:08kaenbut I've tried just multiplying the combined matrix by a vector with the 2d world coords, and it doesn't line up
22:47:11kaenso the question
22:47:22kaendo we modify world coords before sending them to ogl?
22:48:01raptori know we scale the viewing area
22:48:07sam686yes for the most part,
22:48:16raptoraccording to Game::PLAYER_VISUAL_DISTANCE_HORIZONTAL
22:48:24kaenbut we do that with glScale right?
22:48:37kaenI mean, do we modify them excluding the matrix operations?
22:48:38raptoryes
22:48:46raptoryes to the first question
22:49:11raptoroh, there is the window scaling
22:49:15raptorafter that
22:49:23sam686GameUserInterface::renderGameNormal() basically does the entire matrix modification (modified by glTranslatef )
22:53:31sam686renderShip (the one with the renderPos as an arguments), and a few others, also does a glTranslate so it can just render at (0,0)
22:53:51sam686possibly also glRotate used too
22:54:03kaenyep, looks like I made a bunch of mistakes: http://mentors.debian.net/package/bitfighter
22:56:34kaenraptor, so it looks like I should repack the tarball without any of the windows libs?
22:56:55kaenand reimport to a new repo
22:57:03kaenI'll just import a diff of my other changes
22:57:07kaenand we'll start fresh :)
22:58:08raptorwell i was wondering
22:58:50raptoris there a way we could somehow leave the bitfighter source out of it... and just keep the debian directory? it seems like that would be more modularized, but i admit i'm still deficient in my reading..
23:02:49kaenI don't think so, I haven't read anything like that so far. also, I can't find a package on http://anonscm.debian.org/cgit/pkg-games/ with that structure
23:02:56kaenthey all seem to have committed the source
23:03:43raptorah, good idea to look there
23:04:26raptorit just seems incredibly space inefficient
23:06:52raptorthat hardening lintian check i could never solve
23:13:32amgine123 Quit (Ping timeout: 246 seconds)
23:13:45raptori need to sleep again... good night!
23:17:07raptor Quit ()
23:48:51fordcarskaen, do you know how to give ship->isModulePrimaryActive(ShipModule module) a module?
23:49:04fordcarsWell, how to create a ShipModule I guess
23:49:22fordcars(ShipModule)1 or something?
23:49:37kaenit's somewhat complicated
23:50:37fordcarsOr (ShipModule)ModuleShield;?
23:51:00fordcarsOh

Index Search ←Prev date Next date→

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