Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
On Tue, Apr 7, 2009 at 11:56 AM, Allan McRae al...@archlinux.org wrote: My summary of this: 1) Maintainer tag: It is a comment - makepkg does not care so nor should you 2) $foo vs ${foo} : they do the same thing (except in rare cases where brackets are needed...) - makepkg does not care so nor should you. 3) $srcdr vs $srcdir. The quotes are good for people who build in directories with spaces in their name - so it may be good to use quotes. But I don't care about people who use spaces in directory names, so I am not going to do this. Note the prototype does include quotes... 4) $startdir/src vs $srcdir. $srcdir is correct. As stated in the PKGBUILD man page, there is no guarantee that these will continue to point at the same directory in future pacman versions. Agreed.
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
On Mon, 6 Apr 2009 08:53:06 +0300 Evangelos Foutras foutre...@gmail.com wrote: I'm not sure if this has been mentioned before, but here's my take: - Maintainer: The person who currently maintains a package. - Contributor: The person who first submitted the package. If a package is so badly constructed that it needs to be rewritten from scratch, the contributor tag would only list the person who did the rewrite. I know we've agreed on multiple contributor tags, but I believe the method detailed above is cleaner, more maintainable and more straight-forward. I'll most likely adopt it for my own packages, but I'm not saying that anyone else should. I would say that we should hold a voting to make a final decision. However, it's not an important issue at all, so the current way of doing things (multiple contributor lines) is sufficient. As I already mentioned someplace else it's still mixed up here: http://wiki.archlinux.org/index.php/Arch_Packaging_Standards I'd love it if someone knowledgeable could go through that page and could correct any other errors before I try to adhere to those standards. There are also other points that aren't mentioned there, like whether one should use $pkgver or ${pkgver}. All the PKGBUILd.proto files use $pkgver, I've seen both versions in wildlife. /me is now off to edit contributor tags.. Philipp
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
Scripting/coding style has been discussed before. To use ${pkgver} instead of $pkgver is due to consistency. Technically, the braces enable one to append to variables eg. ${pkgver}alpha. Another camp would like double-quotes as well. If you remember, the reason for quotes on ${src,pkg}dir is the fact that no one knows which bloke keeps his local build dir named with spaces. To KIS, # Maintainer: for binary maintainer and # Contributor: for the following: * A submitter of a package. * A _significant_ contributor to the content of the buildscripts including the PKGBUILD -- so how would one define significant? well, it's up to you. If a person does not own a package but has contributed a good deal to it via other means (comments on AUR), the owner _may_ choose to add that person to the list as well. It's not supposed to be a big deal, so long as there's a way to contact someone.
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
On Tue, 7 Apr 2009 05:35:15 +0800 Ray Rashif schivmeis...@gmail.com wrote: Scripting/coding style has been discussed before. To use ${pkgver} instead of $pkgver is due to consistency. Technically, the braces enable one to append to variables eg. ${pkgver}alpha. Another camp would like double-quotes as well. If you remember, the reason for quotes on ${src,pkg}dir is the fact that no one knows which bloke keeps his local build dir named with spaces. Quotes on ${src,pkg}dir ? I believe I haven't seen this, can you give an example? To KIS, # Maintainer: for binary maintainer and # Contributor: for the following: * A submitter of a package. * A _significant_ contributor to the content of the buildscripts including the PKGBUILD -- so how would one define significant? well, it's up to you. If a person does not own a package but has contributed a good deal to it via other means (comments on AUR), the owner _may_ choose to add that person to the list as well. It's not supposed to be a big deal, so long as there's a way to contact someone. I believe it was agreed and agreed again today that maintainer is not only for maintainers of binary packages but also for the maintainers of AUR packages. --Philipp
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
I meant srcdir or pkgdir (: Right..If that's the way it's enforced if at all. On 06/04/2009, hollun...@gmx.at hollun...@gmx.at wrote: On Tue, 7 Apr 2009 05:35:15 +0800 Ray Rashif schivmeis...@gmail.com wrote: Scripting/coding style has been discussed before. To use ${pkgver} instead of $pkgver is due to consistency. Technically, the braces enable one to append to variables eg. ${pkgver}alpha. Another camp would like double-quotes as well. If you remember, the reason for quotes on ${src,pkg}dir is the fact that no one knows which bloke keeps his local build dir named with spaces. Quotes on ${src,pkg}dir ? I believe I haven't seen this, can you give an example? To KIS, # Maintainer: for binary maintainer and # Contributor: for the following: * A submitter of a package. * A _significant_ contributor to the content of the buildscripts including the PKGBUILD -- so how would one define significant? well, it's up to you. If a person does not own a package but has contributed a good deal to it via other means (comments on AUR), the owner _may_ choose to add that person to the list as well. It's not supposed to be a big deal, so long as there's a way to contact someone. I believe it was agreed and agreed again today that maintainer is not only for maintainers of binary packages but also for the maintainers of AUR packages. --Philipp
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
On Tue, Apr 07, 2009 at 10:56:33AM +1000, Allan McRae wrote: My summary of this: 1) Maintainer tag: It is a comment - makepkg does not care so nor should you 2) $foo vs ${foo} : they do the same thing (except in rare cases where brackets are needed...) - makepkg does not care so nor should you. 3) $srcdr vs $srcdir. The quotes are good for people who build in directories with spaces in their name - so it may be good to use quotes. But I don't care about people who use spaces in directory names, so I am not going to do this. Note the prototype does include quotes... 4) $startdir/src vs $srcdir. $srcdir is correct. As stated in the PKGBUILD man page, there is no guarantee that these will continue to point at the same directory in future pacman versions. Case closed!
[aur-general] Maintainer vs Contributor tag let's find a solution ; )
Hi, this thread were discussed in the history, so I think is time to clarify and put the correct information to the wiki. (Actually on the recent TU application and sunjdk package). IIRC: a) Maintainer tag in PKGBUILD is just use for people who maintain the binary package generated by this PKGBUILD on official repositories (core, extra, community*), and the people with abilities to do this are TU and Devs. b) Contributor tag is for people who upload the PKGBUILD for first time or is maintaining it at this time. But sometimes this is hard to apply, for example I adopted an orphan PKGBUILD on AUR and I decided to update it and maintain it, what I should have to do: 1.- Should I remove the past contributor and add myself as a Contributor? 2.- Should I keep all the contributor list even if they are more than 4 different people (4 lines more to the PKGBUILD) 3.- Add myself to the maintainer tag? IMO I will use the second option, to keep all the contributor list, maybe tomorrow I won't be able to contribute with this PKGBUILD but it will be nice, to the future owners of these PKGBUILD to know who maintained before them. But maybe we will have a long list of contributors, so, it would be nice to discuss an idea to have tags for maintainers of PKGBUILD with have a binary and contributors of PKGBUILDs, as I said, i would like to improve a method to apply the second idea. Please try to don't flame the ideas, try to propose new ones, or improve the exposed by me. Regards Note: * Many people should disagree with the idea about community and official repo, IMO if community it's enabled by default on pacman, community became official, no matter what the history was... -- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
On Mon, Apr 06, 2009 at 08:20:41PM +1930, Angel Velásquez wrote: Hi, this thread were discussed in the history, so I think is time to clarify and put the correct information to the wiki. (Actually on the recent TU application and sunjdk package). IIRC: a) Maintainer tag in PKGBUILD is just use for people who maintain the binary package generated by this PKGBUILD on official repositories (core, extra, community*), and the people with abilities to do this are TU and Devs. b) Contributor tag is for people who upload the PKGBUILD for first time or is maintaining it at this time. But sometimes this is hard to apply, for example I adopted an orphan PKGBUILD on AUR and I decided to update it and maintain it, what I should have to do: 1.- Should I remove the past contributor and add myself as a Contributor? 2.- Should I keep all the contributor list even if they are more than 4 different people (4 lines more to the PKGBUILD) 3.- Add myself to the maintainer tag? The accepted practice is to keep a contributor comment for all significant contributors of a PKGBUILD. I don't think it really matters if there's a maintainer comment or not. Maintainers for all packages are tracked by other means. That comment doesn't carry much weight. There's really no reason why maintainers of unsupported packages couldn't use it if they really wanted to.
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
On Sun, Apr 5, 2009 at 21:06, Loui Chang louipc@gmail.com wrote: The accepted practice is to keep a contributor comment for all significant contributors of a PKGBUILD. I don't think it really matters if there's a maintainer comment or not. Maintainers for all packages are tracked by other means. That comment doesn't carry much weight. There's really no reason why maintainers of unsupported packages couldn't use it if they really wanted to. I agree with this to a degree, but it is somewhat useful to have that information stored in the PKGBUILD, so that you don't have to go through the AUR interface, IMO
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
5. Change the previous maintainer tag to a contributor tag and add yourself as maintainer I don't quite follow... you say that you want to improve the method, but you insist that we don't change it and use the old one? Please correct me if I'm wrong This should be 4.- and it's more like than my 2nd point .. then that point about Maintainer is just because exist a binary package in official repos and it's maintained by will be lost, so the concept will change to Maintainer is the people who actually owns the PKGBUILD in any repo or AUR.. (just to clarify how to use the tags). Please try to don't flame the ideas, try to propose new ones, or improve the exposed by me. Uh.. what? Disregarding this... Don't know, sometimes I just generate flames, maybe is my bad english :D. So resuming: there is a new point on the list! by Daenyth, who will decide will be the valid point is the question that I have right now. Cheers! -- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
2009/4/5 Angel Velásquez an...@archlinux.com.ve: This should be 4.- and it's more like than my 2nd point .. then that point about Maintainer is just because exist a binary package in official repos and it's maintained by will be lost, so the concept will change to Maintainer is the people who actually owns the PKGBUILD in any repo or AUR.. (just to clarify how to use the tags). I don't really have any comment to add here, but I'm not quite sure if I understand... you're saying that the intention of the maintainer tag is to store the data because the binary repos don't? Don't know, sometimes I just generate flames, maybe is my bad english :D. Ah, ok. I was confused :P So resuming: there is a new point on the list! by Daenyth, who will decide will be the valid point is the question that I have right now. As I said before, it seems like the general consensus was in favor of changing it. http://www.archlinux.org/pipermail/aur-general/2008-October/002502.html
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
I think it makes the most sense to designate the person currently maintaining the package/PKGBUILD as the maintainer irrespective of that person's status in the community or the destination of the package/PKGBUILD. It immediately indicates to anyone looking at the PKGBUILD whom they should contact about updating it. I think previous maintainers should be listed as contributors along with anyone who's contributed signficant changes to the package. Telling people that they can't claim to be a maintainer of a package because they're not a dev or TU comes across the wrong way too. Just because the binary isn't hosted in the AUR doesn't mean that the work of maintaining a package (updating, responding to comments, etc) is any different than if the binary were uploaded. Just my 2¢.
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
As I said before, it seems like the general consensus was in favor of changing it. http://www.archlinux.org/pipermail/aur-general/2008-October/002502.html But this was the last reply [1] by foutrelis, and confused me.. seems that even Aaron Griffin agree with having the maintainer tag for the actual maintainer and the past maintainers became contributors. P.S: I knew that this wasn't a Deja-vu and we discussed this before (for the record I didn't participated in the last discussion) [1] http://www.archlinux.org/pipermail/aur-general/2008-October/002514.html -- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
Here goes an section from the the PKGBUILD man page: EXAMPLE The following is an example PKGBUILD for the patch package. For more examples, look through the build files of your distribution’s packages. For those using Arch Linux, consult the ABS tree. # Maintainer: Joe User joe.u...@example.com pkgname=patch pkgver=2.5.4 pkgrel=3 Note the use of Maintainer... In the end, it is a comment and nothing more so who really cares about this. Allan
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
I like option 1.- Should I remove the past contributor and add myself as a Contributor?. This is exactly like it is in /usr/share/pacman/PKGBUILD.proto example. -- José Valecillos
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
2009/4/6 José Valecillos valecillo...@gmail.com: I like option 1.- Should I remove the past contributor and add myself as a Contributor?. This is exactly like it is in /usr/share/pacman/PKGBUILD.proto example. -- José Valecillos The problem with this is that it's essentially claiming all the work in the package as your own, which it clearly isn't
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
Then you must add all the past Contributors, even if they are 10 or 100?. On the other hand, in the web interface or when you install the package it don't show anything about the contributor, this should be there I think, dont' you?. I mean, you only can know who is the contributor if you open the PKBUILD directly. However, there are things more importants to discuss. This is really irrelevant if you think about it.
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
2009/4/6 José Valecillos valecillo...@gmail.com: Then you must add all the past Contributors, even if they are 10 or 100?. On the other hand, in the web interface or when you install the package it don't show anything about the contributor, this should be there I think, dont' you?. I mean, you only can know who is the contributor if you open the PKBUILD directly. However, there are things more importants to discuss. This is really irrelevant if you think about it. Agreed. Let's stick to the one which was mostly agreed upon last time; that is any one who maintains the package anywhere can add a Maintainer tag and past contributors be listed in the PKGBUILD. I've filed a bug report to remove this info from namcap: http://bugs.archlinux.org/task/14114 -- Abhishek
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
2009/4/6 José Valecillos valecillo...@gmail.com: Then you must add all the past Contributors, even if they are 10 or 100?. On the other hand, in the web interface or when you install the package it don't show anything about the contributor, this should be there I think, dont' you?. I mean, you only can know who is the contributor if you open the PKBUILD directly. However, there are things more importants to discuss. This is really irrelevant if you think about it. Excuse me, And who are you to call the things irrelevant?, if is irrelevant for you, then don't reply and it's enough. The fact is that we will take the last discussion were several people agreed to use the Maintainer tag for the current maintainer of the PKGBUILD and past maintainers or the original contributors will be on the contributor tag (even if there are many contributors). So this thread at least was useful to remember the way to handle these tags. (at least for me, and to remember to several people who delete past maintainers). -- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
Re: [aur-general] Maintainer vs Contributor tag let's find a solution ; )
I'm not sure if this has been mentioned before, but here's my take: - Maintainer: The person who currently maintains a package. - Contributor: The person who first submitted the package. If a package is so badly constructed that it needs to be rewritten from scratch, the contributor tag would only list the person who did the rewrite. I know we've agreed on multiple contributor tags, but I believe the method detailed above is cleaner, more maintainable and more straight-forward. I'll most likely adopt it for my own packages, but I'm not saying that anyone else should. I would say that we should hold a voting to make a final decision. However, it's not an important issue at all, so the current way of doing things (multiple contributor lines) is sufficient.