#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2015-03-29

Timestamps are in GMT/BST.

00:00:20Flynnn has joined
00:17:35Invisible has joined
00:17:59Invisible is now known as Guest51464
00:44:56Watusimoto has joined
01:02:16Guest51464 Quit (Ping timeout: 264 seconds)
03:03:07Watusimoto Quit (Ping timeout: 256 seconds)
03:10:56YoshiSmb has joined
03:10:58YoshiSmbevening!
05:29:08Flynnn Quit (Quit: This computer has gone to sleep)
05:32:42Flynnn has joined
05:36:35YoshiSmb Quit (Ping timeout: 246 seconds)
09:13:56Flynnn Quit (Quit: This computer has gone to sleep)
10:18:07Watusimoto has joined
10:35:12raptor has joined
10:35:12ChanServ sets mode +o
10:36:04raptor Quit (Client Quit)
12:09:20Watusimoto Quit (Ping timeout: 244 seconds)
12:14:59Watusimoto has joined
12:17:33Invisible has joined
12:17:57Invisible is now known as Guest73846
13:04:20Watusimoto Quit (Ping timeout: 255 seconds)
14:23:24Watusimoto has joined
15:11:41Guest73846 Quit (Ping timeout: 255 seconds)
16:41:03Watusimoto Quit (Ping timeout: 246 seconds)
16:58:32Watusimoto has joined
17:12:25fordcars has joined
17:34:11Darrel Quit (Read error: Connection reset by peer)
17:34:30Darrel has joined
18:27:20Flynnn has joined
18:33:26Watusimoto Quit (Ping timeout: 264 seconds)
18:47:39Watusimoto has joined
19:10:00Flynnn Quit (Quit: This computer has gone to sleep)
19:12:26fordcars Quit (Ping timeout: 246 seconds)
19:44:38raptor has joined
19:44:38ChanServ sets mode +o
19:44:43raptorgood day
19:46:42raptorWatusimoto: how did you get that feedback about clip2tri?
19:46:55raptori'm surprised it didn't come to me...
19:48:17Flynnn has joined
19:50:15raptoralso, i have a question regarding this rendering refactor -
19:50:43raptorI'm attempting to avoid including the OpenGL headers in any class except for the new RenderManager class.
19:51:21raptorThe consequence is that calls that use GL defines in other classes don't work; like: mGL->renderVertexArray(vertices, 3, GL_LINE_STRIP);
19:51:32raptorthe GL_LINE_STRIP is then undefined
19:52:46raptorthe reason for fully abstracting is so that we can always catch rendering API issues
19:53:18raptorso i'm considering even abstracting the GL_LINE_STRIP/GL_LINE_LOOP, etc., values, as well
19:53:36Watusimotohowdy
19:53:37raptorbut... is this too crazy? Am I going overboard? Is there a better solution?
19:54:00Watusimotoi was just sitting down to do my taxes!
19:54:03Watusimotohooray!
19:54:20Watusimotothe clip2tri came to devs@bf.org, which I think directs to me
19:54:39raptorTAXES!
19:54:48Watusimotoit's the most fun you can have in late march
19:55:07raptorhaha
19:55:27Watusimotoregarding your rendermgr class
19:55:31Watusimoto...
19:55:59WatusimotoI'm not sure how you would meaningfuly abstract gl_line_x
19:56:05raptoryeah..
19:56:13Watusimotobut you could create our own version that calls gl_line_x
19:56:20Watusimotolike render_line_strip
19:56:24Watusimotoor something
19:56:24raptornot without unneeded complexity...
19:56:36Watusimotoand just make that a passthrough that is implemented in rendermanager
19:56:47Watusimotoand if we need to make it fancier later, we can
19:56:58Watusimotobut for the moment, that seems to achieve your goal with a minimum of work
19:57:10raptorhmmm... maybe an enum {} with values that exactly match the GL values?
19:57:35Watusimotooh, for the constants? yes, i think that would make sense
19:57:53raptorwith GLES2 we get to reimplement the matrix stack!
19:58:01Watusimotoyeah, I known
19:58:03Watusimotoknow
19:58:09Watusimotoit won't be too bad though
19:58:49raptorI have one last major refactor - that is to catch all the other gl* calls into RenderManager
19:59:18raptorand i'm sure i thoroughly broke tests/dedicated build
19:59:23raptorbut the client compiles!
19:59:38Watusimotook; I have done nothing on bf since I left; I may do a little tonight, but nothing that will interfere with your work
20:00:02Watusimotothe tests might be ok... very few have anything to do with rendering
20:00:39raptorwell... i touched a *lot* of classes
20:01:34raptoralso, abstracting fontstash and oglconsole will be a challenge - i'm not too familiar with texture-based GL code
20:01:42raptori may just leave those be for now
20:03:43Watusimotowell, definitely don't touch oglconsole
20:04:15raptorthe original author of fontstash rewrote it to allow for graphics abstractions: https://github.com/memononen/fontstash
20:04:34raptor^^ that's not the version we use, we use this: https://github.com/akrinke/Font-Stash/
20:04:59raptorours was originally more popular than the original because of the extra features, like multiple stashes, etc.
20:06:14Watusimotois a "stash" a single font/size combo?
20:06:41raptori think so, yes
20:09:19raptorbut a new size it generated when going fullscreen or resizing the window
20:09:44raptoractually, now i'm not sure - it may try to stash the new stuff into the same stash object
20:09:51raptoruntil it fills up
20:13:48Watusimotoyeah, that sounds like it imight be right
20:18:06Watusimotobut if that's the case, what's the attraction of mulitple stashes?
20:19:13raptora single stash must fit within the max texture size of the graphics card
20:22:08Invisible has joined
20:22:32Invisible is now known as Guest62476
20:34:03raptorso, since we have more than 1 font, at potentially many sizes, i'd expect to fill the stash up... but maybe not?
20:37:55WatusimotoI see... so you can't just make it bigger; you need to create a new one... and switch them in and out as needed?
20:38:19WatusimotoI think for our needs, one is probably enough as we don't go too crazy with fonts
20:38:21raptoryes, texture size is requested from graphics layer before anything is added to it
20:39:57raptorlooks like we use a 512x512 stash
20:40:17raptorwhich i think will end up being too small for everyletter of the 2+ fonts, and possibly different sizes
20:47:02raptorso we can do a 1024x1024 - but i'm not sure what generation of graphics cards woudl support that...
21:03:19Flynnn Quit (Quit: This computer has gone to sleep)
21:09:56Flynnn has joined
21:58:18Guest62476 Quit (Ping timeout: 256 seconds)
22:03:39Watusimoto Quit (Ping timeout: 265 seconds)
22:22:38raptor Quit ()
22:32:31fordcars has joined
22:59:25Flynnn Quit (Quit: This computer has gone to sleep)
23:05:15Watusimoto has joined
23:06:21Invisible has joined
23:06:45Invisible is now known as Guest6461
23:31:05Guest6461 Quit (Ping timeout: 256 seconds)

Index Search ←Prev date Next date→

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