/srv/irclogs.ubuntu.com/2012/11/06/#ubuntustudio-devel.txt

=== kubotu_ is now known as kubotu
zequenceHad the most wonderful memory leak. 8GB was flooded awesomely quickly06:59
zequenceSeems to be Java related07:00
falktxnice :S07:02
zequencehmm. upload to ppa seems to succeed, but nothing happens for hours, and no error mail09:59
zequence..nothing happens in the PPA I created10:00
falktxit takes time for things to build10:22
falktxhave you checked the ppa page?10:22
falktxif the package is built successfully, you won't get any email saying so10:23
zequencefalktx: Nothing in the ppa. No activity at all10:23
zequencefalktx: Do you have a nice workflow for patching packages and uploading to ppa?10:24
zequencefalktx: This is what I did https://wiki.ubuntu.com/ubuntustudio/UploadingToPPA10:24
falktxI usually do it like this:10:25
falktx1. create package10:25
falktx2. edit changelog as needed10:25
falktx3. create source package for upload -> "debuild -S -sa"10:25
falktx4. upload -> "dput ppa:kxstudio-team/ppa *.changes"10:26
zequencefalktx: But isn't that for when you create a new package?10:26
zequenceHow about patching an existing one?10:26
falktxon that case you skip 1.10:26
falktxand for 2, do:10:26
falktxdch -i10:26
* persia usually uses edit-patch or checks the patch system in use, and uses that directly10:26
falktxthen do 3. and 4. as usual10:27
falktxzequence: do you have your debian-sources somewhere? I can test the upload10:27
* persia also generally generates local binary packages with sbuild prior to testing, as testing source directly sometimes has unexpected consequences when the changes end up in the package10:28
persiaFor current PPA issues, also see Bug 107156210:29
ubottuLaunchpad bug 1071562 in Launchpad itself "UEFI signing failures cause binaries to be republished continuously" [Critical,Triaged] https://launchpad.net/bugs/107156210:29
persiaThis has apparently caused the PPAs to be out of space (see #launchpad /topic)10:29
zequencepersia: falktx: thanks for those tips. I'm writing all this stuff down for docs10:32
* persia thought LP had a PPA page with instructions to patch a package for a PPA10:37
persiaAlso, I generally don't use PPAs as part of development: I don't generally see the point: if the change is good, it ought go into the distro, and if it's not good, then it needs fixing.  Test-building locally is almost always faster than waiting for the PPA queue10:38
zequencepersia: Having it on a PPA does make it easier for multiple people to test it, if needed10:42
astraljavaOne more perk is that when you're without internet connection, your [s]chroot won't mind. :)10:42
zequencepersia: Actually, I'm thinking of what would be the right procedure for this. Patching jackd2, testing, and then doing an SRU10:43
astraljavazequence: If you intend to do SRUs, then PPA won't qualify as the testbed anyway. You'd be better off with creating the chroot with buildd option.10:44
persiaFOr multiple testers, if it's a SRU, I'd rather get their test reports in the SRU bug so the SRU team considers them, and if it's not, that's why they call it a development release :)10:44
persiaSo, for jackd2: I'd apply the patch to the development release first, and if that seems to work, port the patch to quantal, test locally, and push as an SRU candidate.  Once it gets published to -proposed, call for testers to the bug, and the SRU team can make an informed decision.10:45
zequencepersia: Ah, right. -proposed. In this case, the development release is the same as Quantal, so no change there. But, I'm doing SRU on both 12.10 and 12.04, and those are a bit different10:46
zequenceWell, naturally10:47
zequenceI should go to making the SRU requests, then10:48
persiazequence: The point of the upload to raring is twofold: firstly you get to test the patch in an environment where nobody minds a bit of bugginess, and secondly because you can't SRU somthing that isn'T fixed in the development release.10:48
persiaThe porting of the patch to precise will naturally involve some potential for it not working properly: that's why we stage SRUs in -proposed, so that they can be tested.10:48
persiaIf it turns out that your local testing of the precise version wasn't enough, that gets caught in -proposed, the SRU team removes it from the archive, and you get to try again.10:49
zequencepersia: Aha, I need to patch the development version first. How do I do that?10:49
zequenceI mean, how do I request to get the patch added?10:49
persiaYour wiki page has a good procedure, except change the dput line to a sponsor request :)10:49
persiaI'm not up on current sponsoring procedures, but the way we did it when I was last paying close attention to this was that one prepared a new candidate version (integrated patch, changelog, etc.) all set for dput.10:50
persiaThen one ran `debdiff current-revision candidate-revision` and generated a package patch (as opposed to the software patch)10:50
persiaOne then attached that to a bug in LP, and subscribed the ubuntu-sponsors LP team.10:51
persiaA sponsor would then catch it in the next few days (depending on sponsor availability) and upload it.10:51
persiaIf you want to get information about the current procedure, you can ask in #ubuntu-devel10:51
persiaAlternately, put the debdiff somewhere I can find it, and ask me to review/sponsor.10:52
zequencepersia: Thanks. I'll do that10:52
zequenceBut, now it's lunch time :)10:53
persiazequence: To be clear, my goal here is to get you enough sponsored uploads that you have access to upload to all Ubuntu Studio packages (not all the packages we use, but all the packages we use that other flavours don't use)10:53
persiaEnjoy your lunch :)10:53
zequencepersia: Ah. I didn't know that was a merit needed for that. 10:53
persiaI'm no longer one of the folk that decides who can upload to Ubuntu, but generally one needs to 1) have been working in development for at least one release cycle, 2) demonstrate familiarity with the set of packages one intends to upload, and 3) be suitable as an Ubuntu Member10:55
persiaAnd if one intends to work as a developer in the Ubuntu project, one may as well be able to upload: otherwise one is stuck with out-of-archive flavours or PPAs alone10:55
* persia looks at astraljava and falktx 10:56
astraljavapersia: Aww... making me feel guilty for real? :)10:57
astraljavaI am not even a Ubuntu member. Yeah, I guess I should do something about that.10:57
persiaastraljava: Only a little: I never know if it's that you don't have the time or if you just don't get around to it :)10:58
falktxpersia: I have some packages in universe10:58
falktxhttps://launchpad.net/ubuntu/+source/dssi-vst/0.9.2-1ubuntu310:59
falktx:)10:59
astraljavapersia: Mostly it's my instability. :) I lack persistence.10:59
falktxand this - https://launchpad.net/ubuntu/+source/zyn/1-0ubuntu210:59
persiafalktx: I know: I'm hoping you'll get enough to be confident applying for developer status at some point11:02
persiaastraljava: Given enough time, one is declared to persist, even if one is inconstant along the way: it all averages out11:03
persia(although I can't speak for the current DMB, so I can't be sure)11:03
falktxpersia: afaik I can fix packages as I am right now. just need to get a debdiff11:03
falktxpersia: I think I'm able to apply for maintainership of specific packages too. but i'm not sure...11:04
persiafalktx: Main differences are just 1) you need a sponsor and 2) you can't vote for the TB, support a flavour, or do other governance things11:04
persiafalktx: You could, but there should be a consistent theme for the packages you're maintaining, and it's preferred if you're one member of a team co-maintaining a set of packages.11:05
falktxwell, considering that I'm against some debian policies... I stop worrying about this long ago11:05
persiaWhich sorts of policies?  Do they also apply in Ubuntu?11:05
falktxbut now that my little cadence tools are getting ready, I'm starting to worry again11:05
falktxpersia: mostly about license issues11:06
persiaOh, yeah.  Ubuntu is a bit more flexible than Debian, but those hosting our mirrors only allow so much.11:06
falktxlinuxsampler example is the best one11:06
astraljavapersia: Right, I do understand that. I guess I'm also a bit insecure. It made me look other commitments, which in turn made me burn out a little. I suppose that maybe during 13.10 dev cycle, I could seriously get back to development duties again.11:06
astraljavaNow is the time to get my music-related skills polished. :)11:06
falktxpersia: but main reason is that I'm just too busy11:06
persiaTechnically, PPAs are supposed to have the same limits as Ubuntu, but they aren't as carefully policied (I think Canonical just removes any PPA that gets a compliant and doesn't do anything else)11:07
falktxpersia: I'm learning that I need to backoff sometimes, too much work is bad for the mind...11:07
persiaastraljava: My recommendation is just do a bit every week: don't worry about doing lots and lots: this prevents burnout.11:07
persiafalktx: Indeed :)11:07
astraljavapersia: True, this was indeed my problem.11:08
astraljavaI hogged all responsibility I could get. :) But I was in a very dark place in other ways, too, so I'm guessing I might be more balanced in the future.11:08
astraljavaI still haven't given up my dream of becoming a Ubuntu dev, some day. :)11:09
* falktx wants falktx@ubuntu.com :)11:09
astraljavaOh you're not a Ubuntu member, yet, either?11:11
persiaAs long as you're both motivated, I'm happy: that was the point of my look.11:11
persiaRemember that by getting the LP badges, you're also setting an example for those that follow, so they don't feel like it's impossible (Hey, that person has been doing this for *years* and still can't upload: I'll never get there)11:11
astraljavapersia: No way, I'm not losing that very easily. I still intend to get more active on debian-front, as well.11:11
persiaCool!11:12
astraljavaTrue, never thought of that.11:12
falktxpersia: I hate how debian wants to build everything in debug mode though, I still don't get that11:12
persiafalktx: The idea is that one builds in debug mode, and splits out the debug symbols later in the build process, stripping the binaries like if one hadn't built in debug mode.11:13
persiaOne then produces two packages: the regular package and the debug symbols package.11:13
persiaIf a user has an error, they can install the debug symbols package, and debug it directly, without a rebuild.11:14
persiaBut normal users don't waste space or execution speed on the debug symbols.11:14
persiaBest of both worlds, ideally.11:14
falktxpersia: that idea is plain wrong11:15
falktxit just shows they know don't how programming works11:15
falktxit makes me very sad11:15
falktxlike it's a toy/joke to them11:15
falktxdebug symbols != debug mode11:15
persiaIndeed, that sounds wrong, and I've never been told that is how it had to be done in Debian.11:17
persiaI suspect someone was confused.11:17
falktxthat is very bad for a pro-audio distro11:19
persiaAnd I suspect that if such an issue was brought up for discussion on the debian-multimedia list, others would agree that one only needs to expose debug symbols11:19
persiaIt's bad for a toy distro.  Bad for everything, if it requires extra cycles.11:19
persiaAre we inheriting anything from Debian built in such a way?11:20
falktxsome packages, yes11:20
falktxmost recent ones are like that11:21
persiaIs there a way to build them to expose only the debug symbols?  I think we want to patch them that way (and probably want to discuss with debian-multimedia about doing it that way)11:21
falktxfor example: https://launchpad.net/ubuntu/+source/lv2/1.2.0~dfsg1-111:21
falktxdebian/rules has there: '--debug'11:22
falktx:(11:22
falktxpersia: debug symbols are there by default when there's no stripping11:22
falktxthe sentence "debug symbols" sounds wrong though11:22
falktxthere's nothing about debug in the symbols11:22
falktxthey're just symbols11:23
persiaYeah, but folk mostly use them for debugging, and the phrase has stuck11:23
falktxa debug build always has symbols11:23
falktxotherwise it's useless11:23
persiaDepends on what else debug does11:24
falktxno optimizations11:24
persiaHrm.  waf is confusing.  Has anyone asked alessio why it's this way?11:32
falktxastraljava asked on debian-multimedia mailing list I believe11:34
falktxno response11:34
persiaNo response probably means everyone was too busy.  I haven't seen quadrispro around for some time, but he usually had reasons to do things.11:36
astraljavaAlessio is in Aberdeen as an Erasmus student, probably too much partying to pay attention to us mere mortals currently. :)11:37
ttoinefalktx, thanks for fixing the ardour lv2 problem in you repo for 12.1011:41
zequencettoine: What problem was that?11:42
zequenceIs there a bug report onit?11:42
falktxttoine: no prob11:42
falktxzequence: the 12.10 package doesn't show gui for some plugins11:42
falktxit might be fixed in 13.04, https://launchpad.net/ubuntu/+source/ardour/1:2.8.14-211:42
falktxI don't know how to request a backport for that... :S11:43
zequencefalktx: Ok. I'd like to fix that in the main repo too, if possible. I'll need to see what does not work, and would be greatful for a patch for this11:43
falktxzequence: that package should have the fix, you just need to request a backport11:44
falktxit's same package as in 12.1011:44
zequenceok11:44
falktxI can make a debdiff11:44
falktxbut I forgot how things work...11:45
astraljavafalktx: Backporting requires no debdiffs. It wants the package as-is, and needs to be tested by a few volunteers, forget how many, probably two.11:47
astraljavaThere's a tool for that, request-backport (I believe, not on an ubuntu machine now).11:47
falktxI'm not on 12.10, so I can't even test11:48
astraljavafalktx: Doesn't have to be you, though. :)11:48
persiaNeeds two testers.  Testing on VM counts, for those with kvm or virtualbox11:48
astraljavaI can do it, now that I still have 12.10 on the desktop. Won't be long, though, as I hear raring will be even more stable during the dev cycle than quantal was.11:49
ttoineI am on 12.1011:57
ttoinethe ardour package in the kxstudio repo works with lv2 gui11:58
zequencepersia: I was able to backport qjackctl to 12.04 with only me testing it12:06
zequencettoine: It doesn't help Ubuntu Studio to have working packages in ppa's 12:07
zequencettoine: If there's a bug, someone should file a bug report12:07
zequenceThen, someone should see about fixing it12:07
zequenceThis is something we need to get much better at. Probably more bugs out there, if we start looking12:07
astraljavazequence: Most definitely. This is actually the area where I want to focus nowadays; I stepped down from the "managerial" position, now I just need to roll up the sleeves in the "hands-dirty" position. :)12:09
zequenceastraljava: Sounds good. You having some experience in this. We definately don't have enough of that right now12:10
astraljavazequence: I'm a little out of touch, been for the past couple years. I used to do lots more a few years back.12:20
astraljavaBut maybe it isn't too big of a hurdle to get back in the game.12:20
astraljavaThe music hobby certainly does wonders to my mental health, so who knows, maybe I'll heal quicker than I thought. :) Yesterday was the best day in such a long while.12:21
zequenceastraljava: For me, exercise was the only thing to change my ship around12:25
zequenceSeems like http://developer.ubuntu.com will be a good resource to link to12:25
zequenceI know it's probably more tuned towards other developers than flavor devs atm, but it'll hopefully be a bit better formatted than the wiki12:26
astraljavazequence: Yeah, I'm doing that too. Badminton, walking/jogging and gym/hand-weights.12:30
astraljavaI think that'll be a good reference for anybody wanting to get into Ubuntu dev.12:31
smartboyhwHi scott-work13:35
scott-workmorning smartboyhw , how are you today?13:42
* smartboyhw wants to restart the dev of linux-rt kernel13:43
scott-worki'm curious to why you want to do that13:43
smartboyhwscott-work: good13:43
scott-worksmartboyhw: this is my view on the rt kernel:13:44
smartboyhwok13:44
scott-worki believe for most computers and applications the lowlatency kernel provides plenty of "oomph" (i.e. low enough latency and little-to-no xruns)13:44
scott-workbtw, "most computers" != 286 machines or atom processors13:45
smartboyhwyeah.13:45
zequencesmartboyhw: Any evidence of that -rt is better than -lowlatency? I think not13:45
scott-worksome 386 machines do well with onboard audio, some don't. but if you are serious about audio then you shouldn't be using onboard audio then13:45
zequenceAlso, -lowlatency is good enough, so no need for -rt13:46
scott-workeven within that subset (386 machines and above) you probably should have plenty of memory, say 2g minimum (making up a nice number), but even more if possible13:46
scott-workso, given that even reduced subset, there are going to be people who continue to have problems with xruns/latency....like ralph13:47
scott-worktwo things to mention about people like this: they are a minority most likely and it is probably hardware specific13:47
smartboyhwzequence, scott-work: call this a sideline project, main focus is testing (+ testing docs)13:48
scott-workrt kernel _might_ fix their problems, but then again it might not....especially if there are interrrupt problems13:48
smartboyhwmy sideline13:48
scott-workand then you start coupling the invasive nature of the patch, the need to recompile for video drivers....13:48
scott-workthe lowlatency kernel is an outstanding success compared to the -rt kernel13:48
scott-workit's just a matter of educating the public, IMO13:49
scott-worklike holstein says, "people just _think_ they need the -rt kernel because that is what they read/hear....but they don't"13:49
zequenceAlso, you might not get -rt patches for every kernel version, so some releases would either need an older kernel, or no -rt at all13:49
scott-workzequence: yes, yes. i forgot about that 13:49
scott-workgood points as well13:50
zequenceIt's just more work for no gain13:50
smartboyhwyeah. Just want to see whether it works for now....13:50
smartboyhwzequence: I don't gain anything:-P13:50
scott-worksmartboyhw: please, don't forget about the additional testing things i mentioned. if you want to brainstorm together, i would be gladly available to do so, even if all we do currently is establish topic that should be fleshed out13:51
smartboyhwscott-work: Add it to the blueprint then13:52
scott-worksmartboyhw: add which part to the blueprint, the brainstorming? the topics i think should be there?13:54
smartboyhwscott-work: yes13:54
ttoinezequence, for all my trials, the ardour package don't support the lv2 gui, and it has always been the case. That's why I am using the package in kxstudio, and was not thinking it was a bug in Ubuntu Studio 13:56
* smartboyhw is applying to become a #ubuntustudio IRC o13:56
smartboyhw*op13:57
smartboyhwchannel op I mean13:57
smartboyhwzequence: Interested?13:57
scott-workright, i will add items to the blueprint13:58
smartboyhw:-D13:59
zequencettoine: Still, it won't help Ubuntu Studio, if 1. no one is informed there is a bug 2. Therefor no one can do anything about the bug14:01
ubottuError: Launchpad bug 2 could not be found14:01
zequencettoine: The activity on solving bugs has not been very high, something we would like to change now14:02
zequencettoine: So, if you do find a bug, please report it14:02
smartboyhwlol it triggered the bot14:02
zequencettoine: And if you know of a solution, like knowing the kxstudio has this fixed, please do tell about that too (as you did)14:03
zequenceSomehow, we need to get the fix into the Debian package -> Ubuntu14:03
zequenceIt doesn't help us if the fix is in a PPA14:03
zequenceWe are not a PPA distribution14:03
zequenceThis is an official flavor of Ubuntu14:03
zequenceAnd Ubuntu relies on Debian14:05
falktxthe fix is there, just needs a backport14:05
falktxzequence: this was a different case since the package he was running was not from Ubuntu14:06
falktxby concidence the problem is present in Ubuntu too14:06
zequencefalktx: Your build of ardour is different, right?14:07
zequencefalktx: vst? What else?14:07
scott-workfalktx: where is the fix? just in your PPA? a patch from ardour?14:07
zequencefalktx: You mentioned no debug symbols14:07
scott-workif it is a patch from paul then we can just ask debian multimedia to update ardour, no?14:08
scott-workand then have ubuntu sync it14:08
falktxzequence: not different per say, but the libs it builds upon are not the same14:08
zequencescott-work: It seems the raring version should be ok14:08
micahgzequence: for backports, there's a requestbackport script in ubuntu-dev-tools14:08
zequencemicahg: Yep. I've done one backport already.14:08
falktxscott-work: it's already fixed in 13.04, and the package is the same. it just needs a backport14:08
falktxzequence: I have updated libs of several stuff (mostly lv2 related), and since this was an lv2 related issue it might not happen in Ubuntu itself14:10
zequencefalktx: Ok, so the bug was perhaps only in the kx version then..14:12
zequencettoine: Sorry about the noise.14:12
falktxzequence: yep14:12
zequenceI still haven't had the time to check this myself. 14:13
zequenceOf course, we should not be considering kx bugs Ubuntu bugs anyway14:15
zequenceAnd I have the feeling falktx is quite active on fixing and reporting bugs on that front, as it is14:15
zequenceAt least against the software developers (as in the case with jackd)14:15
falktxyes, things like these are easy to fix14:15
ttoinefalktx, it is not a coincidence, actually14:21
ttoineLV2 GUI don't works with the current ardour in Ubuntu14:22
ttoineAnd I reported to you because your package for 12.04 works with LV2 GUI, and I was disapointed with 12.10 don't working14:23
zequencettoine: Which Ubuntu?14:24
zequencettoine: I mean, which Ubuntu is it where ardour does not work (the stock package)14:24
falktxzequence: the bug is in 12.1014:31
ttoine12.1014:31
ttoinebut I am not sure that the package in 12.04 works better14:31
falktxthe 12.04 version uses old lv2 libs, so it's not affected14:32
falktxplus, it's an older ardour version14:33
zequencettoine: Thanks.14:36
ttoineyou are welcome14:37
zequencettoine: Could you give me an example plugin that does not work. I will try on 12.0414:38
ttoinefalktx, I use your repo for Ardour with my 12.04 workstation, to get LV2 support14:38
ttoinezequence, invada lv214:38
ttoinethe matter is there are not a lot of lv2 plugins in Ubuntu repo14:39
ttoinezequence, you may try too this14:40
ttoinehttp://www.linuxdsp.co.uk/download/lv2/download_dsr500/index.html14:40
falktxlinuxdsp work fine14:40
falktxthe issue is with gtk2 uis, linuxdsp use external-ui14:40
falktxso as lv2fil, so those will work even with the bug14:40
ttoinefalktx, even on 12.04 ubuntu ardour ?14:40
falktx12.04 shouldn't have this bug at all14:41
zequenceNope. I don't seem to have any problems with the guis14:43
zequencettoine: I have the linuxdsp plugins14:43
zequenceAlso mixbus14:43
zequenceHad I more time, I would be using it a lot more :)14:43
zequenceInvada was new to me (but then I haven't really been searching).14:43
zequenceCool guis (haven't yet heard them in action)14:43
ttoineit is not the better plugins I heard... but for free software it is not so bad15:02
scott-workwhat lv2 gui are we speaking about? i have used the calf stuff in ardour on 12.04 without any problems15:06
scott-workso i'm surprised that ttoine is saying that he is using kxstudio version of ardour for lv2 support15:07
zequencescott-work: I'm sure ttoine has other reasons for using kxstudio, but it does seem like only 12.10 was affected. Going to try now15:28
zequenceRight, so the lv2 guis do not show. Instead you get the standard guis that you always get15:37
scott-workoh, okay. i understand, i believe16:09
scott-workthe lv2 gui's are very much nicer than the standard ones :P16:09

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!