| maxiaojun | https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/1066599 | 03:24 |
|---|---|---|
| maxiaojun | the solution is pretty clear now | 03:24 |
| ubottu | Launchpad bug 1066599 in mesa (Ubuntu) "Wine is unable to detect OSMesa correctly when compiling from source" [Undecided,Confirmed] | 03:24 |
| maxiaojun | Mesa needs to build twice: one for OSMesa without --enable-shared-glapi, one for the rest with --enable-shared-glapi | 03:25 |
| maxiaojun | separate build is done in Mageia and openSUSE, not done in Arch, Debian, Ubuntu | 03:28 |
| maxiaojun | anyone? | 03:58 |
| mlankhorst | maxiaojun: worksforme? | 03:59 |
| maxiaojun | how do you test it? | 04:00 |
| mlankhorst | just checking if it's found in configure | 04:00 |
| maxiaojun | using wine? | 04:01 |
| maxiaojun | wine's test is simple | 04:01 |
| maxiaojun | it checks whether libOSMesa has symbol glAccum | 04:01 |
| mlankhorst | yeah and the ubuntu-wine packages have libosmesa, afaict... | 04:01 |
| maxiaojun | maybe it just bypass the check | 04:02 |
| mlankhorst | no they don't.. | 04:02 |
| maxiaojun | can you show me source package | 04:02 |
| maxiaojun | i don't know how to find source package about a ppa | 04:03 |
| maxiaojun | http://paste.ubuntu.com/5748753/ | 04:05 |
| maxiaojun | gcc foo.c -lOSMesa -lSM -lICE -lXext -lX11 -lm | 04:05 |
| maxiaojun | works for you? | 04:05 |
| mlankhorst | https://launchpad.net/~ubuntu-wine/+archive/ppa/+sourcepub/3232603/+listing-archive-extra | 04:06 |
| mlankhorst | checking for -lOSMesa... libOSMesa.so.6 | 04:08 |
| maxiaojun | no | 04:09 |
| maxiaojun | there is a waring at the end of configure | 04:09 |
| maxiaojun | because Ubuntu's libOSMesa is broken according to wine devs | 04:10 |
| maxiaojun | can you run ldd on ubuntu-wine's wine? | 04:10 |
| maxiaojun | i'm still downloading the deb | 04:10 |
| mlankhorst | it dynamically links against it.. | 04:10 |
| mlankhorst | but it shows up in strings libgdi32.dll.so | 04:11 |
| maxiaojun | ? | 04:13 |
| maxiaojun | which file(s) in wine packages links libOSMesa ? | 04:14 |
| maxiaojun | i cannot find one in ubuntu-wine's amd64 package | 04:16 |
| maxiaojun | and can you pass the test i mentioned before, the compile a file test? | 04:17 |
| maxiaojun | broken libOSMesa can break native program also as shown in https://bugs.gentoo.org/show_bug.cgi?id=399813 | 04:22 |
| ubottu | bugs.gentoo.org bug 399813 in Applications "media-libs/mesa-7.11.2: libOSMesa.so undef. ref to_glapi_*" [Normal,Resolved: fixed] | 04:22 |
| maxiaojun | and intuitive for OSMesa -- Off Screen Mesa to use pure software GL implementation | 04:25 |
| maxiaojun | rather share GL API with GPU (can GPU access memory?) | 04:28 |
| mlankhorst | maxiaojun: are you using binary drivers by any chance? | 04:30 |
| maxiaojun | not on this box, but gamers are likely to use binary drivers, right? | 04:31 |
| maxiaojun | anyway, can you acknowledge the problem now? | 04:35 |
| maxiaojun | ? | 04:38 |
| maxiaojun | mlankhorst: ping? | 04:42 |
| maxiaojun | mlankhorst: i have to leave now, hope that you can take care of this issue and make Ubuntu better | 04:49 |
| seb128 | tjaalton, mlankhorst: could you one you review the patch on https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-mga/+bug/1180986 ? | 13:09 |
| ubottu | Launchpad bug 1180986 in xserver-xorg-video-mga (Ubuntu) "X Segmentation fault with dual-head config on Matrox G45FMDVP32DB /32MB /DVI /VGA" [Undecided,Confirmed] | 13:09 |
| mlankhorst | seb128: I'll look at it on monday, today is a free day for me | 14:50 |
| tjaalton | same here :) | 14:50 |
| mlankhorst | (and probably commit it upstream if its good) | 14:52 |
| seb128 | tjaalton, mlankhorst: ok, no hurry, thanks | 14:52 |
| seb128 | enjoy your w.e! | 14:53 |
| mlankhorst | i will, i have a barbecue planned and a horseback ride with my sister in the woods here :D | 15:05 |
| === ubott2 is now known as ubottu | ||
| === ubottu is now known as 17SAC2T0E | ||
| === 17SAC2T0E is now known as ubottu | ||
| === ubottu is now known as 17SAC2T0E | ||
| === 17SAC2T0E is now known as ubottu | ||
| === ubottu is now known as 17SAC2T0E | ||
| === 17SAC2T0E is now known as ubottu | ||
| Sarvatt | ScottK: I know you don't like the aggressive mesa updates, but you do realize the only real changes in mesa at the moment were to fix kwin specifically right? the aggressive mesa updates are due to waiting for stable releases and new hardware support that is absolutely needed, would it be preferred to get git snapshots in a few months in advance? | 16:12 |
| Sarvatt | aka https://bugs.freedesktop.org/show_bug.cgi?id=61554 | 16:12 |
| ubottu | Freedesktop bug 61554 in Drivers/DRI/i965 "[ivb] Mesa 9.1 performance regression on KWin's Lanczos shader" [Normal,Resolved: fixed] | 16:12 |
| ScottK | Sarvatt: I don't like having kwin developers yelling at us. Honestly, I know it's not all on one side. | 16:13 |
| * Sarvatt wishes mesa did real release candidates | 16:13 | |
| ScottK | My fundamental problem is that whatever the situation is now, when I hear people say, "Trust us, it'll be fine", I don't. | 16:14 |
| ScottK | Nothing personal with you guys. I know you're trying hard to please everyone. | 16:15 |
| Sarvatt | I don't see anything that will change personally but that's probably not much of a reassurance, wayland will always be in main, we do maintain it, mesa needs it, mir is just a plugin the same wayland support in it is and won't affect the wayland side | 16:15 |
| Sarvatt | I doubt mlankhorst would ever want to break something he uses primarily also (kubuntu) :) | 16:33 |
| ScottK | That's good to know. | 16:36 |
| mlankhorst | ScottK: but in those cases it's usually upstream, we are just the packagers, is doing everything at the last possible moment better? it reduces the testing matrix immensely just because 1 specific chip of 1 specific manufacturer has problems | 16:56 |
| seb128 | ScottK, your email seems to lack a bit of fact checking before posting (or it would be good to include details about the issues)... | 16:59 |
| mlankhorst | (I am convinced that if we pushed mesa in raring earlier despite performance issues upstream might have tried harder to fix it, now it looks like upstream doesn't care any more, all solutions suck) | 16:59 |
| ScottK | seb128: I don't fact check everything I hear from everyone. | 17:00 |
| seb128 | ScottK, well, you should maybe hold from direct accusations then | 17:01 |
| seb128 | ScottK, it's not like it was hard to check if an ubuntu source contains patches from Canonical | 17:01 |
| seb128 | ScottK, easier to listen the haters I guess... | 17:01 |
| ScottK | It's easier to trust in people that are working in the broader FOSS community over groups going and doing their own thing without worrying overmuch about the broader impacts. | 17:02 |
| ScottK | It may be 3 years, it may be longer, but I'm pretty convinced nothing not driven by Canonical will survive in Ubuntu. | 17:03 |
| seb128 | ScottK, easier but maybe not better, there is no Mir patch in that source | 17:03 |
| ScottK | Not yet. | 17:03 |
| seb128 | that's a different topic | 17:03 |
| seb128 | your email suggest that they are already patches in there breaking things | 17:03 |
| ScottK | When quantal was released, kwin was in fact broken. | 17:04 |
| seb128 | not by mir patches for sure | 17:05 |
| seb128 | there was no mir at this point of time | 17:05 |
| ScottK | No, by Ubuntu being really aggressive about what version of mesa being shipped. | 17:05 |
| ScottK | Now the good news about that one is we all worked together and fixed it. | 17:05 |
| seb128 | right, that's a different topic than "your mir patches breaks our compositor" | 17:05 |
| seb128 | which is pure invention | 17:05 |
| ScottK | Anticipation. | 17:06 |
| seb128 | ... | 17:06 |
| mlankhorst | anyway I'm off, this can wait until monday :-) | 17:06 |
| seb128 | very constructive | 17:06 |
| ScottK | mlankhorst: Have a nice weekend. | 17:06 |
| seb128 | I'm out of it, no point trolling | 17:06 |
| ScottK | I'm not the one that decided that Ubuntu had to go off on it's own on common infrastructure. | 17:06 |
| ScottK | What's next, a kernel fork? | 17:07 |
| seb128 | ScottK, I understand you have concerns, but speculation on what could broke and how, worded in a way that suggests it already happened is a bit disappointing from you | 17:07 |
| seb128 | that's all I've to say | 17:07 |
| ScottK | OK | 17:07 |
| seb128 | I don't plant to reply/argue/feed the troll | 17:07 |
| seb128 | ScottK, have a nice w.e anyway | 17:08 |
| ScottK | I don't think I'm trolling. I think I'm expressing reasonable concerns that are (while still in the future) a pretty straight line extrapolation from recent history. | 17:08 |
| * seb128 calls it a week as well | 17:08 | |
| seb128 | ScottK, I agree with the concerns, I don't agree with the wrong accusations " | 17:09 |
| seb128 | Upstream kwin tells us they already see bug reports from Kubuntu users due to" ... changes that don't exist | 17:09 |
| seb128 | I get that the kwin upstreams hate us | 17:09 |
| seb128 | and decided to fud us | 17:10 |
| seb128 | it's just a bit sad that you rely those fud in the discussion | 17:10 |
| seb128 | you could made your point about the concerns you have without fud | 17:10 |
| bjsnider | ScottK, it's the CCLA that causes canonical to be so isolated and "not invented here"-ish | 17:13 |
| bjsnider | in my view anyway | 17:13 |
| ScottK | bjsnider: It's a huge factor. | 17:13 |
| ScottK | KDE was looking for a KDM replacement and although LightDM was the best technical solution available, it got rejected specifically because of that. | 17:14 |
| seb128 | Qt has a CLA and it doesn't block KDE to use it... | 17:17 |
| ScottK | Different situation. | 17:18 |
| ScottK | To start with, I don't think any of the various owners of Qt have ever solicited code from KDE developers and then retroactively tried to get a copyright assignment and then thrown the code away when the authors wouldn't agree to what was at the time a VERY poorly written document. | 17:20 |
| ScottK | In order not to have regressions, we had to add all the thrown away code back as a distro patch. | 17:22 |
| ScottK | Talk about pointless waste of effort. | 17:22 |
| seb128_ | yeah, not saying that Canonical is perfect, maybe there is a bit too much of history there ... and most people in opensource seems rather happy to behave like hates anyway on those topics | 17:24 |
| === seb128_ is now known as seb128 | ||
| ScottK | Personally, I won't sign a CLA/copyright assignment for anyone unless it's paid work. I just avoid contributing to things that require it. | 17:25 |
| ogra_ | so you never contribute to anything owned by the FSF ? | 17:26 |
| ScottK | Or once I wrote a bug report that, instead of including a patch (which would have needed a copyright assignment), I said in text, change line xxx from foo to bar. | 17:26 |
| ScottK | ogra_: Not so far. | 17:26 |
| JanC | copyright assignment for non-paid work is legally dubious in many countries anyway | 17:28 |
| JanC | (and, yes, I know that'snot really required anymore by Canonical) | 17:28 |
| seb128 | JanC, right, that issue is deprecated with the contributor agreement | 17:31 |
| seb128 | you just give the right to use your code | 17:31 |
| seb128 | that doesn't change the ownership of the code or limit your rights in any way | 17:31 |
| JanC | well, it doesn't limit you rights, but improves Canonical's rights ;) | 17:32 |
| ScottK | seb128: It changes the issue, but doesn't resolve it. You still end up having less rights to the code base that you contribute to than Canonical. Some people think that's problematic. | 17:34 |
| ScottK | Personally, I'm not signing any contractual agreements that aren't reviewed by my lawyer and I'm not doing that for stuff I do for free. | 17:35 |
| JanC | many countries have contract law provisions that protect individuals against companies by requiring proportionality in such a contract | 17:37 |
| JanC | I'm not sure they apply to this sort of contract though | 17:37 |
| JanC | in any case, it's a fact that a lot of people object to contributing code to projects that require a CLA or similar contract | 17:39 |
| seb128 | JanC, ScottK: I've no real issue with whoever does 99% of the work having more work that a side contributor doing 1% | 17:39 |
| seb128 | it is | 17:40 |
| seb128 | but, well, companies need to make money in some way, there is only so much you can give away | 17:40 |
| JanC | seb128, when puttig it black/white, maybe the 99% contributor is just employees working well-paid 9-5, while the 1% contributor is a volunteer spending all his/her free time on it unpaid; that should put some perspective to who really contributes more ;) | 17:44 |
| JanC | (and I know a lot of you work more hours than you really have too!) | 17:44 |
| seb128 | JanC, right, I'm just thinking that as somebody owning a business you can't afford to get screwed because you gave contributors rights to block change in your projects | 17:46 |
| JanC | contributors can never really block a project | 17:46 |
| JanC | as long as it stays open source | 17:46 |
| seb128 | JanC, not really true, if you need to change the license at some point you need the ack of all the copyright holders | 17:47 |
| seb128 | or to throw out the code of those who refuse | 17:48 |
| JanC | and what's also important: you certainly block contributions by other companies | 17:48 |
| seb128 | well, you have to decide what line you are walking and what you value most | 17:48 |
| seb128 | I can understand both side | 17:48 |
| seb128 | but I can understand why somebody putting millions of dollars in a projects want to keep the ability to change the license if needed | 17:49 |
| bjsnider | of course | 17:49 |
| bjsnider | that argument is not going to move RMS though | 17:50 |
| JanC | I can understand why somebody thinks more control is important when spending money, but I'm not sure that's always the best decision ;) | 17:51 |
| seb128 | well, I've no strong feeling either way, I understand some people have and I'm fine with Canonical has a company defining the rules that work for them on the projects they invest in | 17:51 |
| ogra_ | bjsnider, i wouldnt even want to touch RMS ... why do you think anyone would want to move him ? | 17:51 |
| seb128 | at this end of the day I still think that Canonical does more that most closed sources companies | 17:51 |
| seb128 | even if they are not perfect | 17:51 |
| seb128 | it's "funny" how people blame you for "not giving enough" when you are giving time and money away contributing to opensource projects | 17:52 |
| seb128 | even if Canonical was closed source company and Mir was closed source | 17:52 |
| ogra_ | get along with it or don't ... raving about it all the time wont change anything | 17:52 |
| seb128 | there would still be value in the testing, patches, etc we do on other opensource projects | 17:52 |
| ogra_ | yeah | 17:53 |
| seb128 | yet internet comment makes it like Canonical was just stealing and being evil | 17:53 |
| bjsnider | that's eseentially what RMS is telling them | 17:54 |
| ogra_ | yeah, which is nonsense ... and most half way intelligent people i talk to grasp that | 17:55 |
| JanC | Canonical isn't stealing (nobody really has to sign the CLA or whatever other contract), but Canonical is losing contributions because they require it, and that is hurting Ubuntu users | 17:55 |
| ogra_ | the CLA exists since years ... isnt not like it is something new ... but loud complaints only started to happen very recently | 17:57 |
| ScottK | OTOH, everytime someone suggest I fix Launchpad, I say "Sorry, CLA" and they understand, so it's not like it's all bad. | 17:57 |
| seb128 | it would hurt Ubuntu users more if Canonical went out of business and stopped funding working on Ubuntu | 17:58 |
| ogra_ | it didnt hurt the ubuntu users for all these years, why does it now ? | 17:58 |
| ScottK | ogra_: It's been highly controversial since it was started. | 17:58 |
| JanC | ogra_, that sounds like you were out f toh for years... | 17:58 |
| JanC | out of touch | 17:58 |
| seb128 | seems like users are just dreamer | 17:58 |
| ScottK | ogra_: Because it applies to more and more of what's in Ubuntu as Canonical increasingly goes it's own way on things. | 17:58 |
| ogra_ | ScottK, oh, definitely ... i just dont get why it suddely should be more hurtful than it was 3 years ago | 17:58 |
| seb128 | "would be good if Canonical was just paying people to do free work even if they don't make money" | 17:58 |
| ScottK | ogra_: It sucked the whole time. | 17:59 |
| bjsnider | yes, i think RMS is basically against profit | 17:59 |
| bjsnider | not that canonical has ever made any profit, but it could happen in the future | 17:59 |
| bjsnider | depending on the phone market | 18:00 |
| JanC | ogra_, if you are only now hearing it, you have been out of touch with the community for years (sorry if that sounds harsh) | 18:00 |
| ogra_ | ScottK, yes, but it wasnt a widley promoted topic across news sites etc | 18:00 |
| ogra_ | JanC, i dont say i am only now hearing it ... i just say it didnt hurt the users 3 years ago more than it does today | 18:01 |
| ogra_ | yet people behave like it would | 18:01 |
| JanC | it has been hurting people for years | 18:01 |
| ScottK | Well, the whole topic of Canonical doing it's own stuff, requiring CLA, etc plays very nicely into the "Canonical doesn't contribute" meme that Red Hat pushes. | 18:01 |
| ogra_ | JanC, i dont think it has hurt my mom in any way, sorry | 18:02 |
| ogra_ | and she is a happy ubuntu user | 18:02 |
| ScottK | They've been on that for awhile, this stuff just makes it easy for them. | 18:02 |
| ogra_ | neither three years ago, nor today | 18:02 |
| ogra_ | sure | 18:02 |
| ogra_ | we surely feed enoough trolls with it | 18:02 |
| seb128 | well, for sure Canonical doesn't contribute much to RedHat project :p | 18:03 |
| ScottK | Also, stuff like sending search results even for local searches (AIUI anyway) to Canonical is troubling. | 18:03 |
| seb128 | they contribute a lot to Canonical projects though | 18:03 |
| ogra_ | well, people know about it, it is not like it happens secretly | 18:03 |
| ogra_ | and it is an essential bit of unity (not sure you have looked at the smart scopes) | 18:04 |
| ScottK | There's just a lot of things that make it easy to say bad things about Canonical and overshadow the good parts. | 18:04 |
| ogra_ | definitely ... | 18:04 |
| ScottK | I haven't, but if I can't do a local search without network access, I'd call that a fundamental design flaw. | 18:04 |
| ScottK | I'd breach confidentiality agreements in multiple consulting contracts just to use it, in all likelihood. | 18:05 |
| ogra_ | the new unity largely integrates the internet in everything ... | 18:05 |
| ScottK | Well, I guess if I wanted Chrome OS, I'd install it. | 18:06 |
| JanC | I love internet integration, but only when *I* want it :p | 18:06 |
| ScottK | Maybe it's because I'm old, but I think I will always want a clear distinction between what's local and what's remote. | 18:06 |
| JanC | and only with the sites I want it | 18:07 |
| seb128 | you can do local search, just turn off the option in system settings | 18:07 |
| ogra_ | JanC, well, its part of the concept ... and its not a hidden fact | 18:07 |
| seb128 | JanC, in saucy you can select the sources (sites) in the filters | 18:07 |
| JanC | not to mention that the default internet integration is utterly useless for most people | 18:07 |
| seb128 | that's in the ui | 18:07 |
| ogra_ | if you dont like the concept, reconfigure it or use something else ... its not like there wouldnt be choice enough | 18:08 |
| ScottK | seb128: Only because people screamed. | 18:08 |
| ogra_ | JanC, not on a tablet or phone ;) | 18:08 |
| seb128 | ScottK, well, they were right, and we got that option at this end | 18:09 |
| JanC | ogra_, my phone doesn't do anything I don't want and I have no tablet | 18:09 |
| seb128 | so not perfect but not the end of the world either | 18:09 |
| ScottK | Actually, for me, Ubuntu Phone would be an interesting option if it did give me more control over what went off the device and what didn't compared to Andriod. | 18:09 |
| ogra_ | JanC, so you dont have an iphone/android phone then, ok | 18:09 |
| ScottK | I do have an Andriod phone and am not happy about this kind of thing. | 18:10 |
| seb128 | reality is that where 95% of the world is going | 18:10 |
| ogra_ | well, you surely have ful control about the ubuntu phone as you have about an ubuntu desktop with the free images | 18:10 |
| seb128 | e.g new xbox announce this week | 18:10 |
| JanC | dn't (although with an Android phone you could configure it properly, I guess) | 18:10 |
| seb128 | it needs to be always connected | 18:10 |
| ogra_ | i doubt you will have that much contol on a shipped vendor image | 18:10 |
| seb128 | the mics are always on | 18:11 |
| ogra_ | and the kinect too :) | 18:11 |
| ogra_ | you can power it on with a gesture ... so it needs to constantly monitor the room | 18:11 |
| JanC | my dad is a regular ordinary computer/phone/internet user, and he objects against those things too | 18:12 |
| ogra_ | what does he use on his phone/computer ? | 18:12 |
| JanC | (maybe because he once worked/lived/travelled in countries where that sort of information got abused to hurt & kill people) | 18:13 |
| seb128 | well, I don't like those either, but reality is that the world is moving there and Ubuntu is still a lot better than Google Apple or Microsoft on those topics | 18:16 |
| JanC | sure | 18:16 |
| JanC | in any case, that has nothing to do with the CLA issue ;-) | 18:17 |
| tjaalton | flowers, rainbows, ponies, kittens and all that.. :) | 20:26 |
| tjaalton | will check the thred next week.. | 20:27 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!