On Mon, Dec 20, 2010 at 1:05 PM, Bernardo Barros <bernardobarros2(a)gmail.com> wrote: that doesn't make a whole heap of difference. people who use svn are at least 1 step closer to understanding that the first step after a crash is "svn update". people using arch build scripts ... not so much, i suspect. moreover, people using svn are probably (hopefully!) on the commit mailing list and can see that the version they got this morning is now 8 commits old by lunchtime. again, people using arch build scripts ... not so much

Em segunda-feira 20 dezembro 2010, às 15:30:48, você escreveu: ok, i was just asking, since the maintainer there comented it. no need to get angry, we are just curious..lol

On Mon, 20 Dec 2010 13:26:31 -0500 Paul Davis <paul(a)linuxaudiosystems.com> wrote: the ardour3-svn pgkbuild in AUR is basically nothing more then a simple script which saves users who want to run from svn some typing work. this is not an official package or whatever. users who use this know that it is not supported by Arch and obviously, that it builds from svn. what should probably be added is a clear disclaimer "do not ask upstream for support on this unless you know how to debug" or something like that. Removing the pkgbuild (and hence just making it harder to run from svn) doesn't solve the core problem. a warning like the one above does. Dieter

On Mon, Dec 20, 2010 at 1:39 PM, Dieter Plaetinck <dieter(a)plaetinck.be> wrote: but that's not the core problem. the core problem, from my perspective, is that svn evolves by the hour at the moment. when you release something that looks like a package (even if it doesn't walk like a package), people (users) tend to think that something significant has happened, when it has not. they don't realize that if they really want to try it out and then they run into problems that an update & a rebuild comes first. they thus easily conclude that its broken and that there is no way to know when it will be fixed, when the answer maybe that it was fixed 3 hours ago. this is the danger of "packaging" svn for any active project. people avoid SVN updates to avoid breakage, and thus they avoid the fixes too.

On Mon, 20 Dec 2010 19:58:51 +0100, Dieter Plaetinck wrote Was it understood by Fabio? The point you don't seem to get is: _users_ don't build from svn. And building from svn seems to be complicted enough to prevent them to do it. Easier for whom? Jes, as long as you tell folks clearly that _is_ fine. But that important step got lost ... Lukily Fabio only waisted developer time - but alpha software can easily erase your harddisk or go out and kill your grandmother - but wait, this isn't even alpha software! Get the pont ;-) Cheers, Ralf Mattes -- R. Mattes -- Systemeinheitsstreichler Hochschule fuer Musik Freiburg rm(a)inm.mh-freiburg.de

On Mon, Dec 20, 2010 at 2:26 PM, Bernardo Barros <bernardobarros2(a)gmail.com> wrote: I'm comfortable with that note. I just hope that people actually read it.

paul@parc:~$ yaourt -S ardour3-svn ==> Downloading ardour3-svn PKGBUILD from AUR... x PKGBUILD x ardour3-svn.desktop ... ... Comment by: smoge on Mon, 20 Dec 2010 19:45:59 +0000 This is not an official ardour or arch package, as most of AUR users ... ... The warning is done. First Submitted: Tue, 09 Feb 2010 23:46:26 +0000 ardour3-svn 8298-1 : A multichannel hard disk recorder and digital audio workstation. Version 3 with lv2 support (pre-release). ( Unsupported package: Potentially dangerous ! ) ==> Edit PKGBUILD ? [Y/n] ("A" to abort) ==> ------------------------------------ ==> (ellipses are editorial.)

On Mon, 2010-12-20 at 23:46 +0100, Renato wrote: Bauerbill does not show that warning, at least. Clyde may or may not, haven't touched it in a while, but doubt it does since it didn't last time. I'll go contribute to the AUR discussion on their mailing list. The problem is that most 'users' don't actually see the AUR as all that different from the repos, which is what Paul was alluding to.

On Mon, Dec 20, 2010 at 8:32 PM, Ken Restivo <ken(a)restivo.org> wrote: not knowing anything about ion3, i will mention that is my intention to provide distro-neutral binary packages of ardour3. i may have mentioned this before.

For the record, we aren't attacking you, but you are being used as an example, it isn't meant to be personal you just happened to bet he first one that mailed here about it. You did waste a little bit of time, but far less than those that DO come into IRC and say, I built this but why isn't this working, to which we go through a long and tedious debugging process to only find out they built via a build-script or similar and have no idea what actually happened in the build process(See my post above, I used to do this so don't think I am being mean:), and we have to tell them after having wasted sometimes hours of effort the exact same thing you got in the first response;) Yes this does happen, sadly more often than we would like. Seablade On Mon, Dec 20, 2010 at 2:27 PM, Fabio <capoeirista(a)arcor.de> wrote:

Greetings, Paul has *asked* distro maintainers to respect his wishes that A3 development remain tightly controlled by the team. Some distro maintainers obviously feel that this means they can ignore his wishes if they so please. Freedom raises its beautiful ugly head again, with its double-edged blade at the ready. Honestly, did the designers of the Arch script confer with PD before they made the decision to release the script ? It seems to me that it would have been the civil thing to do. They can't have been unaware of Paul's previous statements regarding A3's distribution, or am I wrong about that ? It just seems a bit slippery to release a buildscript in a repo. I think that if you want to build A3 you should demonstrably have the chops yourself and you should abide by the developer's intentions. Btw, is KXStudio still packaging A3 or a buildscript for it ? I asked the maintainer but he never responded. Best, dp

On 12/21/2010 03:22 PM, Fabio wrote: As soon as you create an AUR-script you help maintaining the distro. Best, Jeremy

Fabio wrote: Every one, from your first post to the latest. I understand them, too. And it makes no difference to the point. In effect the users become package maintainers, and they are still acting opposite to the *request* made by the developer. To me, the issue is one of civility. It's not like Paul's thoughts on the matter are unclear. At any time the script writer(s) could have directly asked Paul whether he could be supportive of the script. Perhaps then they could have worked out an acceptable scheme to preclude this kind of thread. Perhaps they would have been dissatisified with his response and proceeded to do as they did anyway, but at least there would have been some dialog. As it is, he finds out about this stuff by surprise, and then we have this thread. I build A3 daily from SVN sources. If I have a question or lament about it I go to IRC, not here and not on the Ardour forums. That is what the developer has asked testers to do and I feel I should honor his request. Yes, I'm certainly free to do otherwise, but I understand why Paul makes the request and I respect it. Best, dp

Excerpts from Dave Phillips's message of 2010-12-21 16:28:19 +0100: Lots of Arch user might have respected Pauls wish and not written a PKGBUILD, others may not have respected the wish or not have known it. You can't expect everyone to know it, it's easy to get the svn link etc. without knowing Pauls wishes. So yeah, the PKGBUILD could be deleted from AUR but someone else will upload another one, because that's the way we work, we write and share PKGBUILDs. So from my point of view this discussion is pointless. Adding messages to the PKGBUILD is a nice gesture of whoever maintains the PKGBUILD at the time and the only solution. I spare myself criticism regarding cooperation in the past and the increasing tendency to control, but one more thing to think about: 'advertising' A3 features on the one hand and trying to preventing 'casual' users is contradictory.

On Tue, Dec 21, 2010 at 11:05:18AM -0500, Paul Davis wrote: This is exactly the problem, and it is very fundamental. Developers can't control the user experience of an open source software package. Full stop. It is not possible. The users, and the distributions, have a lot more say in it. Unless you close the source, users on Linux *WILL* have varying uncontrollable levels of user experience-- it's the price of openness. I see that Ardour has a good following of users in the Mac world, where this kind of tight control of the user experience is rampant and was part of Steve Jobs's original closed-system vision in the early 1980s, and perhaps that kind of philosophy is important when dealing with that market. But that attitude doesn't fit in Linux land, although shipping a binary distro-independent package might get somewhere close, who knows. -ken

On Tue, Dec 21, 2010 at 11:29 AM, Paul Davis <paul(a)linuxaudiosystems.com> wrote: Unless you limit access to your repository how will you ever control that? It's Open Source. Good, bad, working, not working, if it's available then it's out there. As for me I don't even ask _questions_ about A3. Gentoo pro-audio has an ebuild. I build it and play with it. I won't depend on it until 6 months after you say it's ready. In the meantime it's just something for me to play with. As Ardour matures it certainly becomes important that you don't allow people to damage the brand you've created here. My 2 cents, Mark

Le 22/12/2010 01:15, drew Roberts a écrit : Oh yeah, and maybe Paul could rent some office in the pentagone, to be quiet ?

On Wed, Dec 22, 2010 at 4:25 PM, Louigi Verona <louigi.verona(a)gmail.com>wrote;wrote: The issue as I see it, if there is an issue, is not with the code being revealed. The code is open I would guess to encourage constructive contributions. The issue is whether any form of packaging will lead to users who will not provide constructive contributions, and in fact may potentially provide destructive contributions in the form of unwanted bug reports that create a public image of flawed software before it is officially released. I can see that it is perhaps a kind of pre-emptive damage control being performed by unequivocally taking a stand on packaging - if people are willing to read the build instructions they are also probably more likely to read instructions on how to properly report bugs in a way that will help the developers. So I accept the logic behind the stand, though perhaps not the vehemence of the stand to the point of using language like "appalling", a word I would reserve for the atrocities of war. Personally, knowing now that the alpha is so close, I'm inclined to let the Ardour team state whatever desires they like if it'll make the alpha testing phase come around faster. As such, I promise to say no more on this... -michael

On Wed, Dec 22, 2010 at 4:43 AM, Mark Knecht <markknecht(a)gmail.com> wrote: This seems to be one of the more sensible contributions here. The argument seems to be getting reduced to two kinds of people - those who are savvy enough or have enough time to SVN build with a full debugging toolkit and apparently morons who would use a bleeding edge package and be incapable of differentiating it from a finished product - which is at best a false dichotomy and at worst a not so veiled insult to those who fit neither category. I really dig ardour and I have faith that A3 will be a fantastic piece of software. I paid a subscription (though it recently lapsed) as much to use ardour2.x as in anticipation and support of ardour3.x. My anticipation comes equally from the fact that I've tried it out, and that Paul has been releasing teasers and responses to bugs in A2.x that it's all coming in A3 - some day! That's cool, I'm a patient guy, and really, I don't have an overpowering sense of entitlement in this situation. I get it that A3 is so far from finished (and was a little surprised at the CDM article that seemed to declare it almost done). But I, and I suspect many others, check in from time to time to see the state of things. I totally see Paul's concern, especially if mailing lists and forums get littered with a trail of bug reports that should be in mantis, potentially tainting the public image of a future product that is also potentially Paul's future livelihood. But the implication that my building A3 from svn will give some deeper understanding of the fluxness of things than if I choose to install a package because I have a spare fifteen minutes on a Sunday and want to check out the progress of software that I have a great desire to see succeed, is quite frankly, absurd. All that being said, merry Christmas or happy holidays or bah humbug, whichever suits. -Michael, from that subset of linux users that enjoys trying bleeding edge packages of many different software projects and doesn't cry if he gets cut from time to time.

On Tue, Dec 21, 2010 at 8:26 PM, michael noble <looplog(a)gmail.com> wrote: The CDM article was accurate in the sense that we are very close to an alpha release. If it wasn't December, it would probably happen by the end of this month.

Le 21/12/2010 20:29, Paul Davis a écrit : PLEASE PEOPLE ! Let Paul, and Team, work -or have holydays if they like to- they can't lose their time to reply to ~troll (yes, ~ MEANS some kind of) instead of coding.... The point is, what do you want ? Somebody to talk or A3....

On 12/21/2010 03:17 PM, Dave Phillips wrote: KXStudio doesn't have A3 packages. One of his other PPA's, Lucid (Bleeding-Edge) does have A3 packages that are updated frequently. Best, Jeremy

2010/12/21 Bernardo Barros <bernardobarros2(a)gmail.com>om>: actually the pkgbuild does just this: if [ -d $_svnmod ] ; then (cd $_svnmod && svn update) msg "The local files are updated." else svn co ${_svntrunk} ${_svnmod} fi

On Tue, Dec 21, 2010 at 09:17:24AM -0500, Dave Phillips wrote: The package contains 3 files: PKGBUILD ardour3-svn.desktop ardour3-svn.install 'ardour3-svn.desktop' is what you think it is. 'ardour3-svn.install' displays the warning messages. The file PKGBUILD is reproduced below for you to inspect. If you care to read it, you will discover that 1. it performs an 'svn update' when run for the second and later times it is run, 2. it builds a debug version. Running makepkg on it will also install dependencies. * It is just a convenient way to do what has to be done * in order to build ardour3 from svn, including updating your local copy as many times as required. If the intention is that _nobody_ should do this, then access to the code should be controlled. In the other case, it doesn't matter if people do it using the PKGBUILD or manually. If you want to bash Arch users for using this package, at least be consistent and bash everyone building ardour3. PKGBUILD ---8<------------------------------------------------------------ # Maintainer: Bernardo Barros &lt;bernardobarros(a)gmail.com&gt; # Contributor: Uli Armbruster &lt;uli_armbruster(a)web.de&gt; pkgname=ardour3-svn _realname=ardour3 pkgver=8319 pkgrel=2 pkgdesc="A multichannel hard disk recorder and digital audio workstation. Version 3 with lv2 support (pre-release)." arch=('i686' 'x86_64') url="http://www.ardour.org" license=('GPL') depends=("util-linux-ng" "liblrdf>=0.4.0" "liblo" "libgnomecanvasmm>=2.12.0" "slv2>=0.6.4" "aubio") makedepends=("boost>=1.39" "cppunit>=1.12.0" "soundtouch" "libvorbis>=1.1.2" "flac>=1.2.1" "python2>=2.3" "glib" "cppunit" "jack-audio-connection-kit") provides=('ardour3') conflicts=('ardour3') source=(${pkgname}.desktop) md5sums=('2736c648e452ae45c703e48cf64f75e7') install=ardour3-svn.install _svntrunk=http://subversion.ardour.org/svn/ardour2/branches/3.0 _svnmod=${_realname} build() { cd ${srcdir} rm -rf ${_svnmod}-build if [ -d $_svnmod ] ; then (cd $_svnmod && svn update) msg "The local files are updated." else svn co ${_svntrunk} ${_svnmod} fi cp -rf ${_svnmod} ${_svnmod}-build cd ${_svnmod}-build python2 waf configure --prefix=/usr --configdir=/etc --tranzport --freesound --freedesktop --lv2 -p --debug python2 waf build DESTDIR="${pkgdir}" python2 waf install # install some freedesktop.org compatibility install -Dm644 "${srcdir}/${pkgname}.desktop" "${pkgdir}/usr/share/applications/${_realname}.desktop" } ---8<------------------------------------------------------------ Ciao, -- FA There are three of them, and Alleline.

fons(a)kokkinizita.net wrote: I had read it, and I agree, it does what I do anyway. Fair enough. @Fabio : You wrote "I know that it will taste better if i wait. But,,,,,,,,,, I have to eat from it...." :)) You're right, appetite is one of the great stimulators. Best, dp

On 21 December 2010 02:26, Paul Davis &lt;paul(a)linuxaudiosystems.com&gt; wrote: IIRC we've always been telling each other not to upload any buildscript of ardour3 to AUR, but share them via other means, eg. IRC. It was an understanding all archlinux/ardour3 users had at that time and for as long as I know (until now..or more precisely February which I wasn't aware of) there was no such package in the AUR. However, I believe that's flawed and more importantly, stupid. What we end up with in the end is some script that looks almost the same and does exactly the same thing, duplicated on almost every system that matters to ardour3 development. Even one of your most useful/prominent ardour/jack users (he was featured in an article not too long ago) uses buildscripts because: 1) They take care of dependencies 2) They automate svn routines 3) They automate build routines And all that in a nice, unaltered way. You surely don't want to bother your useful users with those headaches, because surely, you want them to _use_ ardour3 and provide valuable feedback. Argueing that, someone who doesn't want to go through those headaches is not fit enough to contribute to ardour3 development, is absurd at best. Which brings us to: Our users are competent enough to understand the difference there - subconsciously. In fact, I believe that kind of thinking can only be attributed to new or casual users of FLOSS. Anyone who is capable of setting up an Arch Linux system is capable of using Subversion. Look at it this way: You would've received this query even if the user had resorted to manual methods of installation. A PKGBUILD is no different from a svn/wget script that prepares the local system, pulls upstream sources and does a build and an install run. Have a look: http://aur.archlinux.org/packages/ardour3-svn/ardour3-svn/PKGBUILD Surely, sharing that kind of script openly would not be against any kind of rule or ethic, nor would it be hampering ardour3 development any more than random support requests from uninformed users, regardless of platform. So just a warning pre/post-install would be good enough to prevent unwarranted queries like this. Which brings us to: On 21 December 2010 03:44, Paul Davis &lt;paul(a)linuxaudiosystems.com&gt; wrote: As long as it requests attention, like a prompt for a yes or no, it will be read in most cases. Else, it's no different from not reading about ardour3 from upstream sources. While navigating to the following page: http://ardour.org/download_full I was able to get the SVN command for 3.0 without bothering myself with the details. If, after installing (let's assume someone helped me with the dependencies on Fedora), I were to face a crash, there would be nothing preventing me from asking a question on the list, forums or IRC. You can rest assured that Arch Linux users don't bother about "distributions". Software is only "packaged" because it's a necessary step to maintain a record and system sanity. You can't really refer to our packaging and compare it to, say, Fedora's packaging. We "package" almost everything, even our own custom, personal scripts. -- GPG/PGP ID: B42DDCAD

On Tue, Dec 21, 2010 at 12:34 AM, Ray Rashif &lt;schivmeister(a)gmail.com&gt; wrote: You've said it there: Arch Linux *users*. In response to "warning dialogs": Am I the only one that tries to figure out if clicking "Yes", "No", or "OK" is going to make the program go on without reading the box? I don't think so. :-) I can see the use of the script, but I have to say that I think the damage of a new user not understanding "development" software, and thinking a program is broken due to them testing a pre-"pre-alpha" far outweighs the usefulness of the script. That's not to say that your not allowed to script the process, but don't share it online. It damages the reputation of the software you've packaged. Am I totally in the dark here? -Harry

i don't really want to argue this too much, but i do think that its pretty fallacious to propose that because Arch makes it relatively easy (like gentoo) to wrap the use of an svn repository in a package-related system that users of the result of this process will understand that there is really no difference between an explicit checkout of svn and management of one's repository, and the package-related tool version. the alpha testers for a3 to date have all been extremely clear that you don't bother checking it out unless you intend to (a) get on the commit mailing list (b) regularly svn update (possibly a few times a day on days when they are testing). i understand (now) that there is nothing about the pkgbuild system in Arch that prevents either of these two from still being the case. however, i don't believe that even a majority of people who use the method will understand the need for (a) and (b) or even just the general nature of what they're connected to. there were 16 commits to svn within the last 20 hours. some of these were deep changes to the way things work (though hopefully not resulting in any visible breakage, we don't know, which is why its not released yet). do you really think that arch linux users who are using a pkgbuild to try this out have any grasp of what is going on? maybe the answer, your answer, is "yes". i'm quite a bit more skeptical.

On Mon, Dec 20, 2010 at 10:44 PM, Ray Rashif &lt;schivmeister(a)gmail.com&gt; wrote: right. and that's the whole point. i don't want people who "just want to try it" to do so at this point because i don't believe its ready for them. and when they find its not ready, as they likely will, what will their reaction be? the idea is to limit the set of people "just trying it" to people who are ready for it to break on them and willing to get involved and help us fix the issue. what you're alluding to is to me more like a beta test. we haven't released an alpha yet.

Hi Paul, Paul Davis wrote: You are right, but I think also overall the Linux Users' community is grown up enough to know what an svn version is, and through this very discussion I think it is now clear to all that any ardour 3 version one 'gets hold of' it is to be considered such. I think - hope - the reaction of users building from svn is "hey it's coming along really well, thumbs up!". That was my reaction when I had a peak last time. Yes I publicly confess I also did a 'curiosity build' :) Finally, maybe the post on createdigitalmusic.co [1] did contribute to increase a certain curiosity... Lorenzo. [1] http://createdigitalmusic.com/2010/12/ardour-3-free-daw-is-nearly-done-and-…

Em terça-feira 21 dezembro 2010, às 06:50:35, Raffaele Morelli escreveu: look at that, an easy and practical solution....lol

On Tue, Dec 21, 2010 at 7:11 AM, Arve Barsnes &lt;arve.barsnes(a)gmail.com&gt; wrote: We have a winner! There are bound to be casual Ubuntu/Debian/Fedora users who have figured out how to get ardour3 via Subversion, simply because they're excited. There's no solution to this. This isn't the problem though. Those users that go through that process have demonstrated some dedication already in time invested in order to compile, and are that much closer to being a benefit to the community at large by providing useful feedback in debugging. In order for them to 'figure out' the process chances are at some point they did some reading and research, far more than is required by the buildscript process. If Ubuntu had a package of A3 in its repos, it would be just as much of a problem for that exact same reason. Once you put this time in once, doing so a second time is a cakewalk(And only needed if on a different machine or you have reinstalled most of your base system in many cases) and a much smaller investment in time. A third and fourth are even less. The people that provide useful debugging information, even after Ardour is released are the people that put in that effort to build a debug version. On Linux I was a Gentoo person, and heavy user of the portage ebuilds for SVN access to Ardour. It was nice, but I ran into problems on occasion that I couldn't answer only because I hadn't built it myself and didn't know every step of the relevant process what had happened. On OS X I build everything by hand(And if you think doing this on Linux is hard, you ain't seen nothing yet till you try to do it on OS X Snow Leopard). I have often considered writing scripts to automate this, but at this time have not for a variety of reasons, but primarily because this way I know exactly what version of what libs are involved and Paul can tell me exactly what version I need to be on for appropriate fixers not only in Ardour, but in other libs as well that he is far more familiar with than I am and has submitted patches to for problems discovered by Ardour. But here is the other half of the matter that Paul alluded to but did not go into detail on. A3 is in heavy development and for the testers being sought after, installation is not a common occurance. In fact it is recommended NOT to install A3 while testing, because often times you may need to test a patch on your codebase, your own local checkout, before it is committed to SVN. You can't do that extremely easily on an installed system, and A3's code is specifically set up to run from the build directory via scripts for this exact purpose, debugging and bug reporting. So the short of it is, consider the building the source by hand a filter to make sure you can and should be doing so. If you can and should be doing so, chances are you could write the buildscript yourself to check out A3's code and build it in a day at most(A few minutes if you are a coder really), with all testing. It isn't that difficult, the dependency handling is the most difficult and the benefit to distribution systems, but noone is saying you can't write your own package to do exactly that, just not to distribute it at this time is the request. Seablade

Le 21/12/2010 12:47, Jörn Nettingsmeier a écrit : +1 Fred

On Mon, 20 Dec 2010, Paul Davis wrote: Possibly the text on ardour.org/download_full should be edited from "If you are an active developer of Ardour, or a beta[sic] tester..." to something more like Paul's stronger statement above. ...or better yet, replace it with a link to the 'getting involved' section of the development page, then in the 'welcome' message of the ardour-dev mailing list, mention the requirement to also join the ardour-commit list, and in that list's welcome message, finally give the link to the svn checkout command. ...maybe then you'll have fewer distractions from drive-by construction-zone gawkers, as well as lower levels of 'buzz' and anticipation of the pleasures to come. It's a balancing act. To what extent are the developers motivated by the public attention of that 'buzz', balanced by the distractions of the occasional less-than-fully-formed bug report, and the drawn-out meta-discussions such as this one? -- Rick Green "Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety." -Benjamin Franklin "As for our common defense, we reject as false the choice between our safety and our ideals." -President Barack Obama 20 Jan 2009

Am Dienstag, den 21.12.2010, 15:19 -0200 schrieb Fabio: Well, I really start to get appetite . . .

On Monday 20 December 2010 21:55:20 Paul Davis wrote: Should such scripts add a cron job to do the updates several times a day? Or one that runs hourly and checks for some flag on the ardour site to know if it needs to update automatically? Asking from a level of ignorance here. all the best, drew

On Tue, Dec 21, 2010 at 11:51:36AM -0200, Bernardo Barros wrote: maybe add a dependency on gdb. the script already builds in debug mode as far as i can see. you need to get it to install the ardbg script though. -- torben Hohn