| 00:09 | Artox | topi`: |
| 00:09 | Artox | only chromium |
| 00:09 | Artox | with patches by dv |
| 00:09 | Artox | which he sadly stopped updating |
| 00:10 | Artox | raspberyypi is using libopenmax |
| 00:10 | Artox | we dont have that |
| 00:10 | Artox | we have gstreamer |
| 00:10 | Artox | but somehow it doesnt work with firefox |
| 00:10 | Artox | or, lets say it didn |
| 00:11 | Artox | I figure it might work out now |
| 00:11 | Artox | but it possibly needs rebulilding firefox to use egl/gles instead of glx |
| 00:11 | Artox | I am really not sure |
| 00:11 | Artox | samsung is working on a gstreamer backend for chromium |
| 00:11 | Artox | but it is far from ready |
| 00:12 | Artox | so that will be a potential solution to this misery for chrome |
| 02:50 | _unreal_ | Artox, hi |
| 09:54 | topi` | Artox: I've also seen somewhere that samsung has a team assigned to work on Servo, the next-gen project from mozilla |
| 09:54 | topi` | and that's fundamentally all GL-driven |
| 09:58 | Artox | topi`: |
| 09:58 | Artox | for imx6, the danger is that its not egl/gles, but glx/gl |
| 09:59 | Artox | which is in really bad shape |
| 09:59 | topi` | OK |
| 09:59 | Artox | the fundamental problem that I ran across all over the place, is the choice to prefer GLX |
| 09:59 | Artox | everywhere |
| 09:59 | Artox | in accelerated gtk gui libraries |
| 09:59 | Artox | browsers |
| 09:59 | topi` | well, gtk is firmly rooted there |
| 09:59 | Artox | desktop environments |
| 09:59 | topi` | but luckily gtk is not the only game in town |
| 10:00 | topi` | I think someone built Servo with only GLES, disabling GL, and it still works so subsets of GL are just ok :) |
| 10:00 | Artox | well gtk isnt the problem, libclutter is |
| 10:00 | Artox | which many gnome applications like to use |
| 10:01 | topi` | GL is ancient anyways, Vulkan is coming :) |
| 10:01 | topi` | yes, I know libclutter |
| 10:01 | topi` | unfortunately :) |
| 10:01 | Artox | well I doubt Vulkan will come to imx6 |
| 10:01 | topi` | that's true |
| 10:01 | topi` | but I just wanted to point out that OpenGL has come to the end of its road |
| 10:01 | Artox | things might get better if the switch to wayland will ever be completed |
| 10:01 | Artox | yep, seen it |
| 10:02 | Artox | though I must admit, I thought gles was the future |
| 10:02 | topi` | yeah. I wish Ubuntu would have pushed Wayland |
| 10:02 | topi` | there is no future ;) just irrational decisions in big rooms. |
| 10:02 | Artox | damn, I am in a big room |
| 10:03 | topi` | also i.MX6 is getting old... |
| 10:03 | topi` | I wonder if there'll going to be a replacement for it from NXP |
| 10:04 | topi` | if they fix the broken things, and keep most of the subsystems intact, it'll be easier to upgrade to i.MX9 |
| 10:05 | topi` | when our team went from OMAP2 -> OMAP3, it wasn't all smooth sailing, but most of the stuff written for OMAP2 worked as-is, just changed the offsets |
| 10:05 | topi` | of course, what we got was a huge boost in application perf |
| 10:43 | Artox | omap3! |
| 10:44 | bencoh | they have imx7, but.... err... :) |
| 10:44 | bencoh | it's just supposed to be a "low-power" clone with a lot of changes |
| 16:57 | topi` | bencoh: I hope they'll have a better open source process for the kernel in the future |
| 16:57 | topi` | well i.MX6 is pretty well covered by mainline now, but it took quite a long time |
| 17:00 | bencoh | for "basic" use maybe, then... |
| 17:09 | topi` | soon the chinese will have better chips with better mainline status, just look at upcoming RK3399 |
| 17:10 | topi` | if they got the PCIe working properly, then that's one highly useful SoC to build entire platforms upon |
| 17:10 | topi` | fairly comparable to e.g. Braswell, but with faster CPU cores (cortex-a72) |