ripps | Does anybody have any clue how well the integrated intel graphics on an core i5 2400 is compared to an nvidia gt 240? | 07:39 |
---|---|---|
tjaalton | what do you mean? | 07:48 |
tjaalton | it's perfectly fine for a "normal" linux desktop | 07:49 |
tseliot | Sarvatt, tjaalton: didn't we use to strip the OS ABI tag from Mesa's libGL? | 10:41 |
tseliot | this patch: http://sarvatt.com/downloads/patches/100_no_abi_tag.patch | 10:47 |
jcristau | wasn't this made unnecessary by your alternatives stuff? | 10:47 |
tseliot | yes but I'm now reconsidering this as the alternatives system kind of breaks dlopen() | 10:58 |
tseliot | I think we can get both things to work | 10:58 |
jcristau | dlopen("libGL.so.1") should still work... | 11:00 |
tseliot | i.e. libGL.so.1 will be a slave link in /usr/lib and point to either the lib in /usr/lib/mesa/ or the one in /usr/lib/$driver | 11:00 |
tseliot | right now we don't have any libGL.so.1 in /usr/lib | 11:01 |
jcristau | i know. | 11:01 |
tseliot | only libGL.so is there | 11:01 |
tseliot | I think someone reported a bug about it | 11:01 |
jcristau | so dlopen("/usr/lib/libGL.so.1") won't work. dlopen("libGL.so.1"), otoh... | 11:02 |
tseliot | ah, I guess the bug report was about apps which specified the full path | 11:03 |
tseliot | yes, that was the case. Thanks jcristau | 11:09 |
jcristau | (the linux libGL ABI spec says it should be in /usr/lib/libGL.so.1 so in theory it should work. i don't know what apps rely on that part of the spec though) | 11:14 |
tseliot | right | 11:23 |
=== highvolt1ge is now known as highvoltage | ||
Amaranth | hrm, I see a mesa update in oneiric, scary | 16:09 |
Sarvatt | very :) | 16:43 |
tjaalton | oh right, now that llvm is in main | 16:47 |
Amaranth | oh, it's not anything mesa 7.11, i'm probably ok | 16:59 |
Amaranth | last time I checked 7.11 stuff still breaks my gles | 16:59 |
Prf_Jakob | Amaranth: bug filed? | 17:00 |
Amaranth | I believe one was filed | 17:00 |
Amaranth | It was something in the build system disabling GLES for intel when using the shared glapi stuff | 17:00 |
Prf_Jakob | Hmm ok | 17:01 |
Amaranth | I spoke to RAOF about it at UDS and he said a bug was filed and he was going to look at it, I believe I found the bug that same day | 17:01 |
Amaranth | Trying to find it again now | 17:01 |
Amaranth | hrm, now I can't find the bug or the gles code in mesa to find the problem again | 17:05 |
Prf_Jakob | Things have changed in that area I think. | 17:06 |
Amaranth | iirc -DFEATURE_ES2 was getting lost when building with --enable-shared-glapi, but only for intel | 17:07 |
Amaranth | I'll have to see if xorg-edgers has a recent looking mesa snapshot to see if the problem is still happening | 17:07 |
Amaranth | Prf_Jakob: Looks like that bug was fixed, I'm on a mesa snapshot from the 16th and have gles2 | 19:31 |
=== yofel_ is now known as yofel | ||
Sarvatt | https://launchpad.net/ubuntu/oneiric/+localpackagediffs allowing ia32-libs diff computation and doing it on the fly every time? incoming DoS | 23:04 |
Sarvatt | xserver-xorg-video-ati has taken about 5 minutes to compute so far :( | 23:05 |
bryceh | Sarvatt, I assume it caches the result after doing it once | 23:15 |
bryceh | Sarvatt, out of curiousity what are you looking at in -ati? | 23:16 |
Sarvatt | bryceh: what happened was i clicked calculate diff for ia32-libs which will take a year and probably be 1GB+, switched pages to cancel and tried to get a diff on -ati. turned out i'll have to wait until ia32-libs is done :) | 23:17 |
bryceh | ah | 23:17 |
bryceh | hm, I've several natty -intel bugs marked now as "fixed" due to ickle's dropping of uxa 3d pipeline | 23:18 |
Sarvatt | dont think its worth picking up that commit to SRU? | 23:21 |
bryceh | well I'm thinking about it | 23:21 |
Sarvatt | people using xterm will complain no doubt :) | 23:22 |
bryceh | Sarvatt, what do you think? I've got it packaged and am going to slam it in a ppa at least | 23:22 |
bryceh | Sarvatt, why's that, does this optimization path solve a problem for them? | 23:22 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!