| Wednesday, October 13th, 2010, 01:30 UTC | ||
| [01:30:17] | Gibby (Gibby!~gibby@204.118.10.244) has quit (Ping timeout: 240 seconds) | |
| [02:49:38] | jpabq (jpabq!~jpabq@174-28-191-186.albq.qwest.net) has joined #mythtv-theming | |
| [02:50:07] | jpabq- (jpabq-!~jpabq@174-28-191-186.albq.qwest.net) has joined #mythtv-theming | |
| [07:02:20] | simonckenyon (simonckenyon!~simoncken@195.7.61.12) has joined #mythtv-theming | |
| [08:17:26] | simonckenyon (simonckenyon!~simoncken@195.7.61.12) has quit (Ping timeout: 264 seconds) | |
| [12:10:26] | gbee: | paul-h: that's a tricky question, as it currently stands there are some differences, screens being used in the OSD can't inherit from base and images _must_ be given explicit dimensions, this is to accommodate XV OSD rendering |
| [12:11:06] | gbee: | in all other respects the XML/behaviour should be identical |
| [12:12:11] | gbee: | the OSD doesn't currently use a screenstack, that's something I intend to fix, so you can't just add something to the OSD screenstack to have it appear over video, you'd need special handling for it in osd.cpp |
| [12:13:12] | gbee: | by 0.25 I hope to have most, if not all of the issues worked out so that there are no significant differences between the OSD and elsewhere |
| [12:16:29] | gbee: | the XV renderer is the biggest pain, hence why I want to see XV dropped entirely, unless someone comes up with a way of rendering the OSD at the screen resolution and not the video resolution |
| [13:28:54] | justinh: | doing away with Xv won't be good for me. opengl rendering just doesn't work on intel. I should get off my bum & try trunk |
| [14:51:48] | mrand (mrand!~mrand@ubuntu/member/mrand) has quit (Read error: Operation timed out) | |
| [14:59:41] | mrand (mrand!~mrand@ubuntu/member/mrand) has joined #mythtv-theming | |
| [16:18:08] | gbee: | someone should get Intel to sort out their act |
| [16:18:36] | iamlindoro: | Or we should stop letting Intel hamstring us |
| [16:18:50] | gbee: | poor performing 3D opengl is one thing, but 2D? |
| [16:19:29] | iamlindoro: | Myth has been a competent DVR for many releases, and will go on being such-- but we're only going to be able to move the UI forward if we demand that GL work, and that it be moderately performant |
| [16:19:41] | iamlindoro: | XBMC requires a fairly high level of GL performance, they are not hurting for users |
| [16:19:47] | gbee: | iamlindoro: well I certainly think we want to advise users to avoid Intel GPUs from now on, not just leave them off the list of recommended hardware but add them to the list of hardware to avoid |
| [16:20:21] | gbee: | that will cause us less grief when we do eventually drop XV etc |
| [16:21:16] | iamlindoro: | I know that this tends to fall on deaf ears (though maybe you agree, gbee) but I really think if we want to have a future with a flashy UI, we need to require GL, period |
| [16:22:25] | gbee: | iamlindoro: I don't think anyone really disgrees, but we can't make that switch overnight, we've got to give users a chance to come to terms with it and start saving for new hardware |
| [16:22:31] | iamlindoro: | I know that I sound like a bit of a broken record, but all of the arguments about why we have to keep Qt/Xv/etc. seem to come down to "But we need $AncientAPI to support $platformWeDon'tRunOn" |
| [16:22:33] | wagnerrp: | gbee: apparently the GMA HD stuff on the i3/5 line isnt half bad |
| [16:22:47] | iamlindoro: | gbee: Why? They can just *not upgrade* |
| [16:23:28] | natanojl (natanojl!~jonatan@c83-252-237-63.bredband.comhem.se) has joined #mythtv-theming | |
| [16:24:10] | iamlindoro: | Anyway, people are embedding XBMC in TVs, it runs admirably on ARM and beagleboard, and they (arguably) have more users than we have, with much more stringent hardware requirements |
| [16:24:17] | gbee: | iamlindoro: I wish we could, but I've seen too good projects/business lose an awful lot of goodwill in that way, if you give users plenty of warning we can avoid that |
| [16:24:44] | iamlindoro: | gbee: I'm not opposed to giving warning, but if we were going to drop Xv or require GL, we should make that determination now |
| [16:25:09] | iamlindoro: | gbee: I know from speaking privately to mark that he is willing to help improve the GL painter and add support for GL shaders and effects |
| [16:25:24] | gbee: | since we're not planning to drop XV for 0.24 we can still give that warning at the 0.24 release **0.24 will be the last release supporting XV, future releases will require good OpenGL performance** |
| [16:25:38] | iamlindoro: | yes, I think that would be totally okay |
| [16:26:02] | iamlindoro: | But that means starting the -developers discussion and getting the majority onboard, and being okay with overruling the inevitable few who will complain among us |
| [16:26:13] | iamlindoro: | we tend to give up on any idea that even one person objects to |
| [16:26:26] | iamlindoro: | Need to be willing to say "sorry, the majority rules" |
| [16:26:38] | gbee: | hell, even if we don't drop XV in 0.25, giving the warning now means that we can start to ignore XV without users complaining over the neglect |
| [16:26:38] | iamlindoro: | Then again, I may just be accustomed to being the bad guy/taking the heat now :) |
| [16:28:55] | gbee: | as I noted earlier, Xv as an API is nearly 20 years old, it's way past it's best |
| [16:29:08] | iamlindoro: | Yeah, totally concur |
| [16:29:15] | gbee: | of course OpenGL is 18 years old ... |
| [16:29:24] | iamlindoro: | But GL has been amended and updated many times |
| [16:29:34] | iamlindoro: | and we can require GL 2.0 with GLSL support |
| [16:29:37] | gbee: | but that just means there is no excuse for hardware/drivers which don't support OpenGL |
| [16:29:40] | iamlindoro: | which is basically any modern GPU |
| [16:31:50] | iamlindoro: | I think the big obstacle is that we always look for 100% consensus on a decision like this-- and we're just not gonna get it |
| [16:32:20] | iamlindoro: | Based on previous discussion, the vast majority of us are okay with dropping Xv, especially as mark would rework the GL renderer to improve performance-- we just need to be ready for a majority to be enough |
| [16:32:38] | gbee: | my suggestion would be to gently ease everyone, users and developers alike, into accepting that XV does not form part of our roadmap, we don't have to set a milestone for that just yet we let the idea sink in for a little while |
| [16:33:02] | iamlindoro: | which means another year of maintaining it, minimum--yuck |
| [16:33:19] | iamlindoro: | I feel like we have been violently easing the devs towards it for 4–6 months |
| [16:33:29] | iamlindoro: | writing should be on the wall, at least among us |
| [16:34:28] | gbee: | not necessarily, it could go for 0.25, that depends as much on Mark as anything |
| [16:35:32] | iamlindoro: | If it does go for .25, we need to have made that determination by the .24 release IMO |
| [16:35:35] | iamlindoro: | since we shoudl announce it |
| [16:36:20] | iamlindoro: | or at least announce "there is a real chance that .25 could require a GL capable card with GLSL support, details to follow" |
| [16:36:52] | gbee: | iamlindoro: we do the latter |
| [16:37:18] | iamlindoro: | gbee: If we do, we should at least have an agreed upon private understanding that that's what we want and are aiming for |
| [16:37:58] | iamlindoro: | So that at least no devs are surprised when they SVN up and their stupid vbox or Intel setup isn't adequate |
| [16:39:48] | iamlindoro: | Anyway, I know we are on the same side of the argument, just trying to translate talk into action :) |
| [16:40:31] | gbee: | although we say "future releases", it seems silly and I don't claim to be an expert, but it's psychology, something that politicians etc know well ... they announce that tax rises may be needed, then a month later they announce those rises (even when they knew they'd be doing so from the start) |
| [16:47:40] | gbee: | IMHO we make it a separate announcement, before the 0.24 release so that it doesn't get buried by that news, we say "0.24 is due to be released soon but while you wait we thought we'd bring you some news about the future. We have determined that in a future release (non-specific, it could be 0.25 but it could be 0.30 and that blunts the reactionary users) we will require GL support for the following reasons:" and then |
| [16:47:41] | gbee: | we list the reasons |
| [16:48:44] | iamlindoro: | Works for me |
| [16:48:57] | iamlindoro: | Want to broach it with -developers? |
| [16:49:06] | gbee: | halfway through the 0.25 cycle if we then announce that XV support is being dropped in favour of GL, no-one is surprised, they knew it was coming even if they'd hoped that it wouldn't be so soon |
| [16:49:15] | iamlindoro: | yeah, sounds sensible |
| [16:50:13] | iamlindoro: | I think it's a good plan and totally doable-- it just needs someone to get the discussion started, finished, and get the announcement out-- and I would really, really like it not to be me |
| [16:50:40] | iamlindoro: | I am happy to push on things, but sometimes I feel like I burn up goodwill trying to get things done faster than I earn it |
| [16:50:50] | gbee: | iamlindoro: I'll raise the idea of the announcement on -developers, but in the same indefinite terms |
| [16:52:56] | gbee: | i.e. "we WILL drop XV, we WILL require GL sometime in the near future, let's make an announcement now so that users will have fair warning" |
| [16:52:59] | wagnerrp: | one of the things that was discussed earlier was dropping the special renderers, instead pulling textures from VDPAU/VAAPI and using opengl from there |
| [16:53:07] | wagnerrp: | is that still on the table? |
| [16:53:35] | wagnerrp: | im just wondering as far as things like the GMA500 which (supposedly) will work fine with VAAPI, but has almost no opengl performance to speak of |
| [16:53:39] | iamlindoro: | That's what Mark had talked about, and it's how XBMC handles things-- whether it's still his idea of the best path I don't know |
| [16:53:40] | gbee: | no, not at the moment |
| [16:53:58] | gbee: | there's not really good reason to do that just yet |
| [16:54:07] | gbee: | not a |
| [16:54:20] | gbee: | and one hurdle at a time |
| [16:54:37] | iamlindoro: | Well, it's the only way to have a collapsable/deformable/animateable miniplayer |
| [16:55:05] | iamlindoro: | Still, I'll take a half step towards the goal over no step at all |
| [17:15:15] | gbee: | we can still use opengl in those circumstances, with vdpau/vaapi for fullscreen playback |
| [17:15:29] | gbee: | but let's worry about those details when we get there |
| [17:15:35] | iamlindoro: | only with an ugly handoff to a GL surface |
| [17:15:45] | iamlindoro: | versus it just being a usable surface to begin with |
| [17:15:55] | iamlindoro: | But yes, horse > cart |
| [17:21:46] | paul-h: | Switching to all OpenGL rendering would get my vote. The only way forward if you want funky effects. |
| [17:22:58] | iamlindoro: | Yeah-- I would be totally down to write another theme if I could get my excitement level back up-- and GL effects, transition, and animations would probably do the trick |
| [17:23:12] | iamlindoro: | and I suspect a great many themers would probably feel the same once that palette of tools was available |
| [17:27:34] | paul-h: | Yeah, I can't watch the news or sport without wondering how we would do those on screen graphic effects in Myth :) |
| [17:51:00] | gbee: | I haven't tried animation in the OSD yet ... wonder if Mark hooked up a pulse timer |
| [18:00:01] | iamlindoro: | he does fades, wouldn't he need a pulse for that? |
| [18:01:57] | sphery: | another vote for requiring OpenGL rendering--from a guy who would have to upgrade hardware to get good OpenGL rendering |
| [18:11:18] | gbee: | iamlindoro: fades? |
| [18:11:32] | iamlindoro: | gbee: the OSD windows have fades |
| [18:11:40] | gbee: | iamlindoro: mine don't ... |
| [18:11:50] | iamlindoro: | wrong renderer? |
| [18:12:05] | iamlindoro: | Working fine in VDPAU + GL for me |
| [18:12:15] | gbee: | using vdpau |
| [18:12:19] | iamlindoro: | *shrug* |
| [18:12:21] | iamlindoro: | work here |
| [18:18:35] | gbee: | hmm, seems fades are disabled in the Slim VDPAU profile, I wonder why |
| [18:19:04] | iamlindoro: | ah, yeah, I use Normal |
| [18:19:19] | iamlindoro: | anyway, I just figured maybe that indicated the presence of pulse |
| [18:20:57] | gbee: | probably does, I'd need to check the actual implementation |
| [19:08:03] | brfransen (brfransen!~brfransen@adrianDHCP-47.216-254-250.iw.net) has quit (Ping timeout: 245 seconds) | |
| [22:17:49] | natanojl (natanojl!~jonatan@c83-252-237-63.bredband.comhem.se) has quit (Ping timeout: 252 seconds) | |
| [22:26:03] | jannau (jannau!~janne@chybek.jannau.net) has quit (Read error: Connection reset by peer) | |
| [22:56:44] | abqjp (abqjp!~abqjp@97-119-165-177.albq.qwest.net) has joined #mythtv-theming | |
| [22:58:32] | abqjp (abqjp!~abqjp@97-119-165-177.albq.qwest.net) has quit (Client Quit) | |
| [23:59:09] | sphery (sphery!~mdean@user-0c6s50s.cable.mindspring.com) has quit (Quit: leaving) | |
IRC Logs collected by
BeirdoBot.
Please use the above link to report any bugs.