Re: [aur-general] Somebody using freetype2-infinality, too?
I believe infinality is dead now. I saw someone posted the other week about this new fonts meta package [1] that allows similar things, but all my remaining arch systems are headless these days, but I would give it a go rather than infinality now. [1] https://aur.archlinux.org/packages/fonts-meta-base/
Re: [aur-general] [review-request] python-intelhex
This would probably be best posted to the package page for the python2-intelhex package. The mailing list isn't really for issues with individual packages. FYI, removing the bin on python2 will be so that you can install both packages side-by-side without conflicts, but the python3 will be the default one used as a binary on the path. On Tue, 24 Jan 2017 at 11:49 Quentin Bourgeoiswrote: > Hi, > > I am currently looking for building python2-intelhex[0] from > AUR. Unfortunately the build failed, while I am using aursync[1] and > require build into a clean (created for this build) chroot: > > $ aursync -c python2-intelhex > [...] > ==> Starting package_python-intelhex()... > /startdir/PKGBUILD: line 17: python: command not found > ==> ERROR: A failure occurred in package_python-intelhex(). > Aborting... > ==> ERROR: Build failed, check /var/lib/aurbuild/x86_64/user/build > > Note that with the python3 package version I come up with the same > bug. > Thus, I decide to have a look at the PKGBUILD and suggest some > modifications I guess the maintainer will be glad to merge any > proposal, but did the community see any thing more to change ? > Basically I made use a lot more of its internal variable __pkgname, > define python{,2}-setuptools as makedepends and add the license file > to the package. > > In another way I don't really understand the rm on the bin directory? > The problem I saw is that only installing python2-intelhex wont allow > the use of the provided scripts[2]. However, I conducted small > tests and they seems to works even with python2. So my guess is to > create an other AUR PKGBUILD that will perform pretty the same things > but only packaging the provided scripts, let call this package > python-intelhex-scripts. Then, one need to enforce a dependencies of > python{,2}-intelhex with python-intelhex-script at the same upstream > release version. > > I propose some modification in the python-intelhex[3] PKGBUILD and a > new PKGBUILD for python-intelhex-scripts[4]. > > What do you think ? > > Thanks in advance for your feedback! > > [0] https://aur.archlinux.org/packages/python2-intelhex/ > [1] https://github.com/AladW/aurutils > [2] https://github.com/bialix/intelhex/tree/master/scripts > [3] > https://git.bourgeois.eu/aur_python_intelhex.git/tree/PKGBUILD?id=e6187787c3743e63c5830f78dd18d58f5bc40f79 > [4] > https://git.bourgeois.eu/aur_python_intelhex_scripts.git/tree/PKGBUILD?id=b346e34509a53cd0ed094f686f635d7e3873a372 >
Re: [aur-general] PKGBUILD for GnuSocialShell
Still no attachment
Re: [aur-general] PKGBUILDs for monkeysphere feedback
For creating users, there is now sysusers.d which makes system user creation a bit more standardised, take a look here: https://www.freedesktop.org/software/systemd/man/sysusers.d.html Regards, Justin
Re: [aur-general] PKGBUILD review request: understand-bin
You can specify a general source array as well as architecture specific ones, it adds them together. So you should put desktop and the /usr/bin binary file in there instead of repeating it. Otherwise it looks good to me now. Regards, Justin
Re: [aur-general] My first PKGBUILD
You'll need to post a link, attachments are stripped from the mailing list. Regards, Justin
Re: [aur-general] afraiddns-git PKGBUILD: review and criticism welcome
You can add an empty file in the package for the key, and set it in the backup array, that way it won't overwrite existing files and will work without doing it via a shell script and not tracking the file in pacman. Your pkgver function should probably be below the list of static variables at the top, not in the middle of them. The license should be 'custom' unless it's a more standard one. See [1] for more info on that. If you don't use a build function, just don't include it. No point making a no-op one in there. Otherwise it looks good to me [1] https://wiki.archlinux.org/index.php/PKGBUILD#license Regards, Justin
Re: [aur-general] Can someone get from old version of AUR4 repo a package removed?
Just use git to clone it. Packages don't get removed from the repo, just hidden from the web interface.
Re: [aur-general] pkgbuild - suggestion
Sometimes there may be a new version out, but the maintainer needs to test and make sure it's working as expected. It may need a patch to work correctly, or it may have a major flaw and every other distro is also avoiding updating to the latest version. The flagging system, I feel at least, is more so there for users to nudge a maintainer if they haven't updated it. Often people can wait a few days after the upstream release for the new version, and if they can't, they should be using the -git version anyway.
Re: [aur-general] Continuous integration of AUR packages
I was using the chroot tools for a year and a half or so before that, but had some issues occasionally with things like updating it, or using package caches inside the chroot, or other various things. I do still have the shell script I used at the time, it was called by jenkins and cloned a chroot based off a clean copy and numbered by the jenkins executor to support multiple builds at once. But I managed a few tens of thousands of docker containers for work, and figured I could use a similar build system for the arch packages, and it's been reliable for me since then. As Doug said, you need to add the packages you build to a repository for the next build to be able to install them, but that part is pretty simple and explained on the arch wiki already. So, in Jenkins I use the build generator plugin, and created a parameterized job (config.xml here: http://hastebin.com/oyuqiwokem.xml) that basically just calls `dmakepkg` on the git checkout for the aur package; you just give the build an aur package name and it will generate the job to track that package and any changes pushed to the AUR and build it. (it polls git once an hour). `dmakepkg` is in the AUR ( https://aur.archlinux.org/packages/dmakepkg-git/ ) and is what I ended up making to build the packages inside a docker container. It supports the same flags as makepkg (just passes them through) but defaults to '--force --syncdeps --noconfirm' unless other flags are specified (no point removing packages since the container is destroyed after use). Uses a slightly modified version of the official docker arch image, and is updated much more often (monthly at least instead of the 3-6 months they take). It auto-detects makepkg configs like SRCDEST/etc and copies it in to the instance and bind mounts the folders specified, as well as the pacman package cache, so all the builds follow the makepkg config on the host, but using everything else from a clean environment. If you make your own repo and add it to your host, you can use `-x` flag to use the host's pacman.conf inside the container.
Re: [aur-general] Continuous integration of AUR packages
I actually have a Jenkins setup for my packages that does almost exactly that, I have a job generator build that I give an AUR package name to and that's all the config for new packages, and it builds them in a docker container. You can obviously set up dependencies between the jobs to build in order, but that part isn't automatic yet. Its been pretty reliable for me over the past year or so since I last really changed anything. If you're interested, let me know and I can share the scripts and stuff. Just on my phone at the moment or I would link them here. Cheers, Justin On Thu, Feb 18, 2016, 9:51 PM Baptiste Jonglezwrote: > Hi, > > I would like to run some kind of continuous integration of my AUR > packages. The goal is to know when a package I maintain fails to build > because either: > > 1/ its dependencies have been updated (new API, new incompatible version >of GCC, ...) > > 2/ for -git packages, changes made upstream broke something (new >dependency needed, new build system...) > > The scripts in devtools [1] look like they should work just fine to > automate these kind of builds. After all, they are used to build the > official Archlinux packages. > > However, I found that the build scripts do not really handle dependencies. > When building a given package, they just install deps and makedeps from > the existing Archlinux repositories. This is an issue when AUR packages > depend on each other, because then dependencies cannot be installed from > the Archlinux repositories... > > It *is* possible to manually pass packages to install in the chroot before > building, but this is far from convenient. For instance, ring-daemon > depends on opendht, so I would need to do this: > > cd opendht > extra-x86_64-build > cd ../ring-daemon > extra-x86_64-build -- -I ../opendht/opendht-0.5.1-1-x86_64.pkg.tar.xz > > When there are multiple dependencies, it quickly becomes a nightmare to > automate (especially because dependencies need to be passed in the right > order). > > Is there any script that automates dependency handling when building? > Basically, it would probably need to perform a topological sort, build > packages in this order, and push them to a local repository to be able to > build later packages. > Or did I take the wrong approach entirely? > > Thanks, > Baptiste > > [1] https://projects.archlinux.org/devtools.git/ > > PS : Some existing efforts I found about CI with Arch: > > - > https://lists.archlinux.org/pipermail/arch-dev-public/2014-November/026757.html > https://jenkins.arch-ci.org/ > No script provided, site appears to be down > > - http://openbuildservice.org/2012/09/10/arch-linux-support/ > Source code is unreadable (enormous Perl scripts, no modularity) > >
Re: [aur-general] PKGBUILD review
You've still got references to OS X and windows installs Also instead of if test _arch == '32'; then scons -j $cores platform=server target=release_debug tools=no bits=32 else scons -j $cores platform=server target=release_debug tools=no bits=64 fi you can just do scons -j $cores platform=server target=release_debug tools=no bits=$_arch You're setting _arch in 3 places even though it can only have 2 values; it's also in an array for no apparent reason, you can shorten it to _arch=32 [[ $CARCH == x86_64 ]] && _arch=64 Or some such variation. The same goes in the install, you have a bunch of 'if _arch == 32' and then using 32 twice in a string; just substitute that with the variable. Regards, Justin
Re: [aur-general] [aur-dev] [PRQ#4382] Deletion Request for b43-firmware
I didn't want to reply all, but good job :) - Justin
Re: [aur-general] [aur-dev] [PRQ#4382] Deletion Request for b43-firmware
Can we just make a poll or something and post it on the feature request for this? There is a lot of noise and talk and generally shitty behaviour all around; I (and I assume many others) are signed up to the mailing list for AUR related news; not petty arguments over english semantics. - Justin
Re: [aur-general] [aur-dev] [PRQ#4382] Deletion Request for b43-firmware
I can understand someone getting confused when they first click on it; and any of the suggested name changes sound pretty logical. But, how does someone click on that, then select deletion or merge to some other request type, enter text and then click submit without realizing 'This probably wasn't a download link' - Justin
Re: [aur-general] Out-of-date notifications
FYI, you can use things like IFTTT and Zapier; github allows you to get atom RSS feeds for things like tags and commits. I have a check that sends me an email when a new tag is pushed to a project I package on the AUR. It's super helpful. - Justin
Re: [aur-general] namcap - games group & non-standard directory
On Tue, 22 Sep 2015 at 21:16 respiranto <respira...@icloud.com> wrote: > Hi, > I have just finished packaging freesweep, a minesweeper-game, after > having had many problems to do so. > > Anyways, the current problem is, that namcap complains about: > /var/games/freesweep/{,sweeptimes} being owned by root:games (E) and > /var/games/{,freesweep/{,sweeptimes}} existing in a non-standard > directory (W) > > The point is, that freesweep needs a file, that any user can write to. > The way this would be done by `make install' is exactly how I did it in > the PKGBUILD - make the file (sweeptimes) and its directory owned by > root:games and give the games group write permissions. > > Some days ago, Justin Dray proposed to do exactly this to solve a more > or less identical problem [1], however I would like to avoid pushing a > PKGBUILD whose package does not pass namcap. > So is there any better way to achieve what I want to? > > Furthermore I don't know, how strict the policy to only use standard > directories, as listed here [2], is. > And why does the /var/games directory exist if a package may not use it? > May it only be populated by installed programs? > > [1] > > https://lists.archlinux.org/pipermail/aur-general/2015-September/031535.html > [2] > https://wiki.archlinux.org/index.php/Arch_packaging_standards#Directories /var/lib/freesweep would probably be the right path and why it is complaining. (see man file-hierarchy for more info, which does not list /var/games at all). I'm not sure why it would complain though, since /var/games is created by the filesystem package, and as you said, that sounds like an endorsement for it's existence to me. Ownership I'm not sure on, but someone else on this list should be able to advise. Cheers, Justin
Re: [aur-general] How to grant all users permission to write to a subdirectory of /opt?
On Sat, 19 Sep 2015 at 04:05 Storm Dragonwrote: > Howdy, > I'm working on a package. The package will need to download sounds to a > subdirectory of itself. I have placed the package in /opt. > The package is /opt/tintin-alteraeon. It downloads the msp sounds from the > Alter Aeon mud as it encounters them. But my current setup gives me, for > example: > sounds/spell//tensor.opus: Permission denied > So, how can I give normal users permission to write to the sounds > directory? Also, is that a bad thing to do? If so, I'll have to come up > with a different way to handle sounds. > -- > Powered by Arch Linux! I am registered Linux user number 508465: > https://linuxcounter.net/user/508465.html > My blog, Thoughts of a Dragon: http://www.stormdragon.tk/ > get my public PGP key: gpg --keyserver wwwkeys.pgp.net --recv-key 43DDC193 > Twitter and Facebook are so ... yesteryear. Get your 2MB Social account > TODAY! http://2mb.social/main/register > The great thing about Object Oriented code is that it can make small, > simple problems look like large, complex ones. > "As the moon is rising, give us the sign. Now let us rise up in awe, rock > 'n' roll angels bring thy hard rock hallelujah, demons and angels all in > one have arrived." > Lordi - Hard Rock Hallelujah > Ideally you would either download the files beforehand, or per-user. But if you want a shared folder, create a group for your application, and put in the post_install a message to add users who want to use the application to that group. There is a built in games group that you could probably just use for this? If you set the directory to g+s it will inherit group ownership when a user creates a new file; but you need ACLs to inherit actual group write permissions, and 99% of systems will have a umask of 022; so no group writes. - Justin
Re: [aur-general] Registering, misspelling email, losing account
On Tue, 25 Aug 2015 at 01:42 Bruno Pagani bruno.pag...@ens-lyon.org wrote: Le 26/07/2015 22:29, Daniel Micay a écrit : On 26/07/15 04:01 PM, Igor Morozov wrote: That's right, I messed up. Instead of typing fastmail.com, I typed fastmai.com. And now there is no way I can access my account. The only option is to send an email to this mailing list describing my problem and hope that somebody will help me out. Basically, that's what I'm doing right now. Okay, so it can ask the user to provide the same email in two fields. It could treat an unconfirmed account as a temporary placeholder and replace it if registration is done again for the same username. It shouldn't be possible to log in without confirming the email unless all of the actions (voting, submitting packages, commenting, etc.) beyond editing account information are gated on whether the account is registered. People tend to make mistakes. I'm not the only one who messed up during registration. And there is no easy way to get our account back. Mailing list is not the best option for account recovery. What if the misspelled email exists and the owner decides to proceed and register? What if the owner decides to do nasty things using my username, full name and email that looks alike? That would affect my reputation in the community since it's difficult to prove that I was not the bad guy. The usual account activation prevents this stuff. A lot of web sites do not automatically log user in after account confirmation, so it kind of prevents malicious activity (the bad guy doesn't know the password, you see). Someone could have just created a fake account before you did, so it's really not an issue related to the confirmation design. And by the way, the fact that you can use an unused (not registered) email in account recovery and not get any errors is frustrating. Took me 8 hours to realize that it says okay, even though the email is not in use. Please, do something about it! Emails aren't received instantly, so there's no error to report during registration. Sorry to respond so late, but I had a little idea (but maybe it’s not a good one) to enhance things here. OP was concerned about the owner of the false adress being capable to do nasty things. Since it seems we ask for PGP key, would it be possible for the server to encrypt the account confirmation mail ? And check that (one of) the email(s) on the key correspond to the one provided during registration (to help once again avoiding typos, but see also below)? Thus, even if the malicious bad guy registers a false account using your nickname and your full name, they are two possibilities: – he registers with an email he owns, using a false GPG key. You may then prove this and show it wasn’t you (which was the concern). – he registers with your key, but then the email verification step blocks him. Any thoughts? Why don't we do what every other site does and have a confirm email field? Or a way to change passwords over ssh, since putting in a public key is a field on registration as well? - Justin
Re: [aur-general] AUR activity log?
On Thu, 13 Aug 2015 at 04:20 Doug Newgard scim...@archlinux.info wrote: On Wed, 12 Aug 2015 15:07:57 -0300 Giancarlo Razzolini grazzol...@gmail.com wrote: Em 12-08-2015 15:01, Doug Newgard escreveu: On the contrary, this is exactly the mechanism for that. You disown a package so that someone else can adopt it. Why else would you disown a package? Let me rephrase it. Disowning a package isn't the mechanism for allowing others to maintain a package, if you still need/use/care for it. Co-maintainer functionality is for that. There were people on the old AUR that would disown a package so that someone else could update it, and then disown it again, and so on. This should end. You aren't getting it. If you don't want to maintain a package and want to make it available to others, you disown it. This doesn't mean you want it deleted, it simply means you want someone else to maintain it. If someone approaches you and want to help, you make them a co-maintainer. Two completely different things. Sure they can, why wouldn't they be? If someone adopt it. When they are in orphaned status, they can't. But, then again, if someone adopt it, then it wouldn't be deleted, and we wouldn't be having this discussion. Anyone can push to the repo of an orphaned package. That person then automatically becomes the maintainer, but will often simply disown it again. Doug There's definitely some discrepancies in how we're all thinking about how it should work (for the record I'm totally aligned with Doug in this regard), but I have to say: The metric here should be based on relevance (actual PKGBUILD downloads) and time since it become orphan. Sounds perfect. But we currently don't have a way (or not that I'm aware of anyway) to do this without opening each package manually. Having even a weekly/monthly script run through that data and present a list of old/possible unused orphans would be pretty helpful. - Justin
Re: [aur-general] Deletion of orphaned packages on AUR4
On Thu, 13 Aug 2015 14:01 Sam S. sml...@gmail.com wrote: Note that we do keep the Git repositories of deleted packages, so if anybody wants to maintain the package later, he can always clone the repository of the deleted package, fix the package and simply push it afterwards Can you give some details on that? For example the libtiff4 package (which one of my packages depends on) was deleted, but when I do git clone ssh://a...@aur.archlinux.org/libtiff4.git all I get is an empty repository. I cloned one I had deleted and it was still there, I made a comment change, committed and repushed, it lost the comments and votes but the package was there. I'm not sure why yours might have failed. - Justin
Re: [aur-general] AUR activity log?
On Thu, 13 Aug 2015 at 03:17 Giancarlo Razzolini grazzol...@gmail.com wrote: Em 12-08-2015 14:10, Jens Adam escreveu: That wasn't what happened. Read the thread(s) again - a single TU went around and deleted a bunch of freshly orphaned packages, nothing to do with the migration. I really don't see the issue here, even if is this that happened. They were orphaned, weren't they? And, they aren't deleted, they are hidden. Unless the TU also removed the git repository, which I believe didn't happened. Truth is, the disown functionality has been misused and you can't really complain if you orphaned a package and it was deleted. Even more now that AUR has a co-maintainer functionality. Cheers, Giancarlo Razzolini But by 'hidden' it also deletes all comments and votes, and stops people being able to search for the package, see that it isn't maintained and picking it up. Almost all of my packages have become mine by trying to install something, finding it useful, and when it became an orphan, just taking it over and fixing it up. If we wanted to delete packages we would have asked for deletion, not orphaned it. What is the point of orphaning packages if they are just going to get deleted anyway? - Justin
Re: [aur-general] AUR activity log?
On Thu, 13 Aug 2015 03:36 Giancarlo Razzolini grazzol...@gmail.com wrote: Em 12-08-2015 14:21, Justin Dray escreveu: But by 'hidden' it also deletes all comments and votes, and stops people being able to search for the package, see that it isn't maintained and picking it up. Well, the TU could have waited, I give you that. Almost all of my packages have become mine by trying to install something, finding it useful, and when it became an orphan, just taking it over and fixing it up. This is how I ended up maintaining quite a few packages. But I didn't waited for them to become orphan, in most cases. If we wanted to delete packages we would have asked for deletion, not orphaned it. What is the point of orphaning packages if they are just going to get deleted anyway? Now this is were I fail to see the point. If you still wanted/needed the package, why orphan it? I'm perfectly happy with the way AUR works today. But, if you guys really want orphaned packages to stay around for some time, I suggest you guys implement it and send a diff. Perhaps something that prevents even a TU from deleting (hiding) a orphaned package that isn't orphan long enough, lets say, a couple of months. Cheers, Giancarlo Razzolini Perhaps seeing active comments or that the packages had to have been updated within month since everything was cleared for AUR4? We already have a mechanism for disowning a package and allowing others to maintain it without deleting it. It's called orphaning. The problem here is that how they are treated has apparently changed with no community involvement or even a warning that orphan packages will be deleted at random. Perhaps if TUs are able to view the last updated time from a search fable, they could see an orphaned package with no updates for X months. But as has been said before, orphaned does not mean useless or broken. - Justin
Re: [aur-general] Deletion of orphaned packages on AUR4
On Tue, 11 Aug 2015 17:09 Antonio Rojas aro...@archlinux.org wrote: Rob McCathie korr...@gmail.com Wrote in message: Just a query: Why were packages i added to AUR4, ensured were in good working order (and made an enhancement to one of the packages compared to the last release on AUR3), know are used by at least some users, and then orphaned so some other interested party can take over maintenance, were deleted from AUR4? compiz-gtk-standalone was actually the ONLY package on AUR4 that provided the Compiz 0.8 series core component. Since it's deletion there is now at least one package on AUR4 that has unresolvable dependencies. (eg. https://aur.archlinux.org/packages/ccsm/ ) Regards, Rob McCathie I was wondering the same thing... Many of the kde-applications git packages that I uploaded to AUR4 and then disowned ( but were working fine) have been deleted, and I can't find any related request. -- I also disowned an LSI raid card utility a few days ago, and it got deleted within 2 hours. What's going on with this? - Justin
Re: [aur-general] Deletion of orphaned packages on AUR4
On Wed, 12 Aug 2015 at 15:37 Rob McCathie korr...@gmail.com wrote: On 12/08/15 13:49, Doug Newgard wrote: In my case, I have some that I'm actively trying to get maintainers for; in the mean time, I'm looking after them even though they are listed as being orphaned. Is this not to be allowed now? Should all orphan packages in the official repos be deleted, just assume nobody is looking after them? I updated one package just a few days before it was randomly deleted. There's other stories further up in this thread about them being deleted after only a few hours, all with no notice. If a time limit is to be implemented, it needs to be limit long enough that the package is both unlikely to be being used and unlikely to work anymore. A month or two wouldn't cut it. A notice should also be sent out to anyone set to get notifications for that package with enough lead time for someone to pick it up. Doug Same here. I was still monitoring the couple of packages i'd orphaned, i was hoping someone would take over maintenance. For a time at least, i'd have addressed any issues with them. Anyways, i've re-added the packages and will stay maintainer of them until things settle down a bit. -- Regards, Rob McCathie I've had to do the same thing. The problem is, if it isn't orphaned, and you try to update it when you get a chance it is hard to find a new maintainer. I've never seen someone ask for maintainership of a maintained and up-to-date package before. From the reports I'm seeing as well it's a single TU deleting them all. - Justin
Re: [aur-general] Deletion of orphaned packages on AUR4
On Wed, 12 Aug 2015 07:24 David Phillips dbphillip...@gmail.com wrote: I suppose some may subscribe to the view that if someone wants it badly enough, they'll submit, maintain and stick with it. In my case, I uploaded a perfectly working package for LSI raid controllers, but someone commented that a newer version was available, I no longer use any LSI controllers and can't test that it is working correctly with the new version, I said as such, and orphaned the package only for it to be deleted within a couple hours. If orphaned packages are going to be deleted straight away I would have hung on to it. But then what is the point of their being an orphan button? There's already a delete one. It's seems really poor form to be just deleting any random orphaned packages off the AUR. Perhaps TUs that are doing this should no longer be TUs since they are clearly abusing that privilege to do things however they want instead of within the guidelines set by the community. - Justin
Re: [aur-general] Can't upload packages to aur now
On Sun, 9 Aug 2015 08:33 Pablo Lezaeta Reyes prfl...@gmail.com wrote: 2015-08-08 19:14 GMT-03:00 Simon Hanna simon.ha...@serve-me.info: On Sat, Aug 08, 2015 at 06:59:48PM -0300, Pablo Lezaeta Reyes wrote: I have 51 packages with the user Jristz but I decide change the .git/config and change sed s|aur4.|aur.|g but now I can't upload packages and fixes. I attempt changing my ssh-key and I test everyone (ecdsa,ed25519,rsa1,dsa,rsa) but without succes, then I attempt to reclone a package for test but aur refuse to allow me to do it. so far my .ssh/config is Host aur.archlinux.org aur4.archlinux.org Hostname aur.archlinux.org IdentityFile ~/.ssh/id_rsa.pub User aur Port 22 and I have aur.archlinux.org in the know hosts list, so What could happened that I cant even clone my packages?? -- *Pablo Lezaeta* AFAIK IdentityFile should be the private key not the public key. Regardless, it's hard to help when you don't provide exact commands and their output. give us the clone command you use and the output you get. You can try ssh aur.archlinux.org help just to check if you can reach aur. next stop would be the urls you use for cloning/pulling/pushing the aur article, the last time I check not have info about the config file, so is need add. but you are right the .pub was the problem. thanks and I hope this will be not repeat; but the lack of info about the propper config of the config file could has prevented this because I first check wiki/AUR -- *Pablo Lezaeta* But that's in the man page for ssh config, its not really an AUR issue, I feel it would be out of place to explain ssh config on the AUR wiki page. - Justin
[aur-general] Update notifications
Hi, I was thinking it could be useful to show in comments when a new version is pushed by the maintainer of a package; just a line break and maybe v1 - v2 or the git commit message or something. It would be easier to make sense of comments on many of the packages I've seen and allow us to see if comments pertain to the current version or not without losing comment history. Cheers, Justin
Re: [aur-general] Update notifications
Why not just check the last updated date? -- Eli Schwartz Because that would only work for the current version, and back one, this way you could see the relevant change happening after an issue is reported and such. It seems like a much easier way to view the differences to me at least. - Justin
[aur-general] Git over HTTPS
If your network admins don't know the difference between incoming and outgoing ports, or not opening things like ssh ports to the internet that really isn't an Arch problem... - Justin -- Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] AUR4 feature request
I t's neither provided by the old, nor by the new AUR. I never downloaded the PKGBUILD, since I build with yaourt. The old git mirror for AUR 3 never removed old packages, so you can still retrieve it from there. But it would be good to maintain access to old packages on the new AUR as well. - Justin
Re: [aur-general] [AUR4] Is there a tool for using a single git repo for several packages?
How will that work for new packages you take maintainership of? I'm not a git wizard, but taking a look at the stuff I can't see any way it merges in an external git repo in to yours when you take over an existing orphaned package. I tried the submodule thing as well and found it horrible inadequate. I was going to just query the AUR to find my package list and clone down/update all the returned repositories, and just keep the scripts in a repo by itself. - Justin Dray
Re: [aur-general] [AUR4] Is there a tool for using a single git repo for several packages?
Ah, sounds good! Thanks for that. - Justin
Re: [aur-general] Ho to update .SRCINFO without creating a tarball?
Generating such a file manually is error-prone, using mksrcinfo ensures that the file is exactly what AURweb needs. I believe that is what he is saying should be in the wiki. And since there are so many questions around it this past week on the mailing list, it might be a good idea to have it on record for new maintainers when they go to submit their first packages as well. - Justin
[aur-general] Duplicate packages
Hi, I was just wondering what the process is when someone finds duplicate pacakges? Assuming I'm not the owner of either package, but someone has uploaded foo and foobar and both provide the same thing, what would be the best we to get that resolved? It's quite annoying to have to read 2+ pkgbuilds to see which one is actually decent when all I want to be able to do is install the package. Regards, Justin
Re: [aur-general] Duplicate packages
Thanks for that.
Re: [aur-general] Cannot push to aur-dev, was: Re: [aur-dev] AUR 4.0.0 pre-alpha
If you just cloned a repo; master will have upstream set to origin/master. You should only have to set upstream when you create a new local branch. and if you do, you can use 'git push -u origin branch-name' to set it and push in one go. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] AUR 4.0.0 pre-alpha
On Tue, Dec 30, 2014 at 11:19 AM, Pablo Lezaeta Reyes prfl...@gmail.com wrote: but what if the file not contain any extention? or if a different file is need (like a .svg or a .sh or a .webm? We will ended adding all the posibilities for images? or there will be an alternat8but is need to installive if a user need update a package with anything that is not in the whitelist withut ended renaming it at makepkg time. Maybe if a user need add a exeption they can contact the mailist for ask about an exception to they blacklist in they git-sub-tree-whatever. 2014-12-29 22:14 GMT-03:00 SpinFlo sl1pk...@gmail.com: 2014-12-30 0:57 GMT+01:00 SpinFlo sl1pk...@gmail.com: 2014-12-30 0:52 GMT+01:00 SpinFlo sl1pk...@gmail.com: or exist a method to use white list (only files/directory allowed) instead of blacklist (list of exclude dirs/files) greetings oh, i fount it: http://stackoverflow.com/questions/15288712/gitignore-whitelist-on-directory-and-its-contents greetings ok i tried test with a .gitignore (do'h, 'git add *' don't work again. but use 'git add .gitignore' works :/) with this content: * !.gitignore !.SRCINFO !PKGBUILD !somefile.desktop !somefile.patch !someimage.png !sometextfile.ext then added the package sources and dozens of files. 'git add *' says that files are ignored, included the src/pkg dirs(and contents) i'm sure this is a best chice to avoid the duplicate local repos (one to work, and other to clean push) greetings -- *Pablo Lezaeta* You can still force add them with 'git add -f foo.bar' But most smart shell auto completions won't work with it either which is a bit annoying. I use a blacklist, and it works for all of my packages, but I don't have anything pulling things like .deb sources or a million other possibilities; I think using a whitelist instead of a blacklist may be the way to go. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Delete ipycli-git
On Mon, Dec 29, 2014 at 12:26 PM, Ralf Mardorf info.mard...@rocketmail.com wrote: On Mon, 29 Dec 2014 13:37:20 +1300, David Phillips wrote: From https://aur.archlinux.org/ : Package orphan requests, merge requests, and deletion requests should be filed in the Package Actions box. This point has also been brought up at least twice a week for the last six months on this list. The last two mails sent before the OP's mail, are exactly about this issue too. They came through the list 12 hours, before the OP's mail was sent. Whatever happened to the idea of sending a once off email to everyone who has a valid email on the AUR? These constant emails are getting to be quite annoying, but I don't want to unsubscribe from the mailing lists to get rid of them. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Email Notifications of new AUR orphaning/deletion/merging method
I would very much like to get one more email, instead of 2-5 a week. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Sun, Dec 7, 2014 at 5:57 PM, Ralf Mardorf info.mard...@rocketmail.com wrote: On Sat, 6 Dec 2014 17:44:03 -0500, Brent Saner wrote: On Dec 6, 2014 5:36 PM, Ralf Mardorf wrote: On Sun, 7 Dec 2014 11:25:59 +1300, David Phillips wrote: What do you all think? I guess your idea is very good +1 Perhaps it's better to repeat sending this information at the end or beginning of each month for the next quarter.
Re: [aur-general] Unsupported architectures in the AUR
Good question. One of my packages got deleted last year because it was arm only. I no longer used any arm systems, and was just maintaining it, so I didn't bother chasing it up. But I'm also interested in the answer to that. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] AUR Best Practice for New Package Upload
That sounds like the most reasonable way to build it. I would also like to suggest that Florian should read the man page for file-hierarchy as it is quite helpful if you are unsure of where files should be going. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Requests
Perhaps just the package name without a link? To be honest, the package name alone would satisfy me. Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Requests
Hi everyone, I think it would be a good idea to link to the package for a request when it has been fulfilled. The messages do not show up in the same thread and there is no link, just 'request #1234 accepted' and which TU has accepted it. I then have to go and search for that number to find out what the request was for. Thoughts? Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Java name guideliness
The fact that one naming convention or the other being chosen is still not the issue, it's the fact that in the interim there are *both* packages in there. There should only be duplicates of the same package for as long as it takes for a merge request to be processed. This is clearly not the case, and I think one of the more annoying points that has been brought up in this thread. In fact it probably never would have come up on this mailing list if there weren't 2 subtly divergent jdk/jre packages in the aur simultaneously. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Fri, Sep 12, 2014 at 4:02 AM, Pablo Lezaeta Reyes prfl...@gmail.com wrote: If look to the pther packages (not java) the fewers who need a version they apend the versión at the endm so for consistency whit the all others packages I think is better keep the version at the end: vendor-jre/jdk-version: oracle/openjdk-jre-7/8 2014-09-11 13:41 GMT-03:00 Det nimetonma...@gmail.com: More input. On 09/09/14 23:08, Justin Dray wrote: Part of the issue here however is that now there are both jre7 and jre7-oracle and so on duplicate packages in the AUR. Yes, but why are you bringing that up as an issue, as we are trying to decide exactly which one to keep before just removing the other. I think you know neither I nor the maintainers were never for leaving both - we just don't, at this point, know which one. I mean, isn't that like saying: A: We need to figure out the right type of fuel for this car. B: Yes, but the issue is the car doesn't move. A: Well. Yeah.. On 10/09/14 01:20, Pablo Lezaeta Reyes wrote: Refusal is what happend when two or more not agree in something I never mention who is refusing who cause both side from the vewpoint of the other is refusing the other side of view. Who is refusing whom? One not want use the other guidelines, so using the bare meaning of refusal that mean not accdept the other. But you see you hit the problem right there. We don't _have_ a guideline for the naming. That's what the debate is about: we are trying to establish one. or this is sick. Maybe you are overreacting (or I not expresed it corretly), I mean that is no sane (synonimous of ill synonimos of sick) Ok. Just use makes no sense. I thing that is bvous that all are java. so why not something like providerjre/jdk-version: openjdk-9 or oraclejdk-7. The names in the official repos are jre/jdk8-openjdk, so that's why the previous suggestion was jre/jdk8-oracle. However, I believe it should be pretty obvious which one you're dealing with, if a package is named just jre or jdk (isn't that the ultimate KISS?). The java- prefix is used for archlinux-java (extra/java-common), and already decided upstream. That I would refuse to divert from. On 10/09/14 08:38, P. A. López-Valencia wrote: My opinion is that the AUR should follow the example set by the Arch Linux developers in the extra repository and everything else must go, starting with the jdk/jre pair as clarity trumps over brevity in naming Explained above. As far as I know in all the years I maintained these things, nobody ever confused them with OpenJDK, because that's always mentioned. On 10/09/14 07:43, Rafael Ferreira wrote: +1 for 'java-8-jdk' and 'java-8-jre' is a good name. Just would be nice to have the word Oracle in the description, so a yaourt -Ss oracle could easily track your package. I agree. Added. Again, to summarize the Java guidelines: Package name: project nameversion-vendor (e.g. jdk8-openjdk, jre7-openjdk) archlinux-java name: java-version-vendor(/jre) (e.g. java-8-openjdk, java-7-openjdk/jre) And what I support for AUR (same as what we had before): Package name: project name(version) (e.g. jdk, jre7) archlinux-java name: java-version-project name(/jre) (e.g. java-8-jdk, java-7-jre/jre) (just java-7-jre unsupported by archlinux-java) Det -- *Pablo Lezaeta*
Re: [aur-general] Java name guideliness
Agreed, someone complained on one of my java packages that it was using the incorrect path for oracle java from the aur, and upon inspection I noticed these multiple packages providing the same thing, none of which work as they should be. I assumed they would have fixed it since that was only a short time after the new java common stuff, but it appears that they may need some assistance with that task. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Wed, Sep 10, 2014 at 8:18 AM, Pablo Lezaeta Reyes prfl...@gmail.com wrote: Since the new java-common come to the repo Is now possible have multiple java, but this bring and open another issue, java naming scheme the guy in jre/jdk[1] and jre-devel and jdk-devel refuse to follow a convention non generic name and the other maintaining jre7/jdk7 [2] and jre7-oracle/jdk7-oracle that do the same [3] refuse to accept or merge jre7-oracle into jre7 for the same reason even if the jre-oracle was merged into jre, this is a chaos. Many packages doing the same in different verion having different name conventions and ALL arguin bein correct. There is need to a conventional standar name scheme or this will be worst, instead to be kiss this is sick. There is a name scheme or name convention to follow? [1] https://aur.archlinux.org/packages/jdk/ [2] https://aur.archlinux.org/packages/jdk7/ [3] https://aur.archlinux.org/packages/jdk7-oracle/ There is also a jdk5-compat. Also mergins as ben posponed until name convention come. -- *Pablo Lezaeta*
Re: [aur-general] Java name guideliness
Part of the issue here however is that now there are both jre7 and jre7-oracle and so on duplicate packages in the AUR. If someone says 'oh, i need oracle jdk, I can search on the AUR for that.' Well now they have to go and read all of the comments and look around on the wiki/mailing lists/forums to figure out which one they actually want. And it's not even a dispute between different maintainers, 'joschi' is the maintainer for both packages and are seemingly totally different; different groups, different upstream urls, different dependencies, different provides/conflicts. It also appears that jre8-oracle was merged in to jre package recently, so there is another discrepancy in the naming there. I'm not fussed one way or another on the naming, but by having both, I've really got to agree with Pablo; it's far from KISS. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Wed, Sep 10, 2014 at 12:22 PM, Det nimetonma...@gmail.com wrote: On Wed, Sep 10, 2014 at 8:18 AM, Pablo Lezaeta Reyes prflr88 at gmail.com wrote: Since the new java-common come to the repo Is now possible have multiple java, but this bring and open another issue, java naming scheme the guy in jre/jdk[1] and jre-devel and jdk-devel refuse to follow a convention non generic name and the other maintaining jre7/jdk7 [2] and jre7-oracle/jdk7-oracle that do the same [3] refuse to accept or merge jre7-oracle into jre7 for the same reason even if the jre-oracle was merged into jre, this is a chaos. Many packages doing the same in different verion having different name conventions and ALL arguin bein correct. There is need to a conventional standar name scheme or this will be worst, instead to be kiss this is sick. There is a name scheme or name convention to follow? First of all, I really *really* urge you to stop using phrases like refusal or this is sick. If that really was the case, it would only split all parties further. It's not refusal to talk something through before doing it. In fact, _nowhere_ do I see anybody refusing to do _anything_. The talk in jdk7[1] is discussion on the appropriate name, and what I told everybody both in there and jdk[2] was my view on things and why I did what I had done (use jdk/java-8-jdk as the name, rather than jdk8-oracle/java-8-oracle). You realise how unbelievably easy it is for me to revert to the jdk8-oracle approach, if that winds up being the consensus? Or if I somehow wouldn't, then how easy would it be to kick me off from maintaining that thing? Enough of that already. Why I chose the java-8-jdk naming comes from the fact that java-8-openjdk sounds like we're trying to do java-major version-project name. The project name of JDK is not Oracle JDK, and that's why I chose it. Now, OpenJDK apparently still calls these projects by their base name[3], but _I_ would still prefer (read: I don't refuse; I prefer) having packages called jdk8-openjdk and jdk that install in /usr/lib/jvm/java-8-openjdk/ and /usr/lib/jvm/java-8-jdk/, respectively. This also means we can currently do: $ man java-openjdk8 $ man java-jdk8 To access the man pages. I really didn't like the following at all: $ man java-openjdk8 $ man java8-oracle [1] = https://aur.archlinux.org/packages/jdk7/ [2] = https://aur.archlinux.org/packages/jdk/ [3] = http://openjdk.java.net/projects/jdk8/ Det
[aur-general] Package maintenance
Hi, The issue I've got at the moment is in regards to a poorly packaged package that is also providing files made by other packages? The one in question I've also suggested the fixes to the maintainer, and requested it be orphaned when it was not updated for several months. Within a day the maintainer bumped the version up but didn't fix any other issues. I'm not fussed about being the maintainer or not, I just want the packages that I use to work correctly. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Package maintenance
Sure, it was the google-cloud-sdk. It includes all of the appengine SDKs as well, which are already in google-appengine-* packages, making it several hundred MB instead of 21MB, and it doesn't link a single binary to /usr/bin, so you may as well have just untar'd the source in /opt. https://aur.archlinux.org/packages/google-cloud-sdk/ Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Wed, Aug 6, 2014 at 10:00 PM, Johannes Löthberg johan...@kyriasis.com wrote: Might help if you actually name the package in question. -- Sincerely, Johannes Löthberg PGP Key ID: 3A9D0BB5
Re: [aur-general] virtualbox-extension-pack vs. virtualbox-ext-oracle
Even if the script is the recommended way, shouldn't the package be designed so that it will track all of its files? Having a post install script that places files through your filesystem feels like a dirty hack at the best of times. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On 05/08/2014 7:13 am, Laurent Carlier lordhea...@gmail.com wrote: Le lundi 4 août 2014, 18:15:32 Christian Hesse a écrit : Hello everybody, following a lengthy discussion sublu just deleted my package virtualbox-extension-pack from AUR. This is what his package (virtualbox-ext-oracle) does: * Install an archive file. * Use install script to copy a number of files to /usr without pacman knowing about it. I think this is the wrong way, so I created my own package (virtualbox-extension-pack) that tries to get it right: * Just install the files required, ready to use for virtualbox. * No crappy install script required! My package had about 75 votes IIRC, probably there would have been more if more people knew about the details. The discussing had a number of comments that agreed about my opinion regarding installing/coping files to /usr with pacman or the install script. Although I think it is wrong seblu is free to provide his package via AUR. But is there any good reason mine is not allowed to reside there? I've checked both packages, then i've also checked virtualbox documentation. Documentation is available at http://www.virtualbox.org/manual/ * Ch8.36. VBoxManage extpack * The extpack command allows you to add or remove VirtualBox extension packs, as described in Section 1.5, “Installing VirtualBox and extension packs”. In regards of VirtualBox docs, Seblu is installing extension pack the proper way. Your package isn't following upstream way to install extension package and you are not sure it will keep working, you are just lucky. Your package is only a duplicated package of seblu's one, only differing on the way to install extension pack files isn't a good reason enough. Seblu was right to remove your package, there was an explanation before suppressing, rules were followed. Nothing more to say. -- Laurent Carlier http://www.archlinux.org
Re: [aur-general] Dependancies and Split Packages
On Tue, Jun 24, 2014 at 8:31 AM, Storm Dragon stormdragon2...@gmail.com wrote: Hi, I'm back again lol. Ok, so when I made p-git depend on the package pypump2-git that I made, and when I installed python2-click after modifying the python-click PKGBUILD everything works as expected. When, however, I make p-git require python2-click and python2-pypump-git it fails to build. It installs the package, for example python2-click, then at the end, I get something like installing python-click 100% Then it fails to install p-git saying dependancies are not met, missing python2-click and now also same error with python2-pypump-git. So, how can I fix it? I really want to do this stuff the right way, but I also want it to work at the end of the day. There's nothing worse than finding a nice shiny new package in the AUR and trying to build it only to have it fail. Should I switch back to the old way? Or, should I just put python setup.py and python2 setup.py in the same package() function? But, even if I do put them in the same function, it still won't fix my problems with python2-click. Thanks for the help Storm -- -- Registered Linux user number 508465: https://linuxcounter.net/user/ 508465.html My blog, Thoughts of a Dragon: http://www.stormdragon.us/ get my public PGP key: gpg --keyserver wwwkeys.pgp.net --recv-key 43DDC193 How many Internet mail list subscribers does it take to change a lightbulb? http://goo.gl/eO4PJ One more time we stare into the blackened sky, for tonight, in our hearts now we feel. One last time see our destiny reveal. DragonForce - The Last Journey Home Definitely don't put both in the same pacakge. Make seperate python 2/3 packages, otherwise if someone only has python3 installed, they will be pulling in python2 dependencies as well when they don't need it. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Merges
No, but the original email asking about cleaning them up was 10 days ago. Which the other maintainers responded to. I also said this in my original email to the AUR mailing list: (this was only around 10 days ago however. I'm not sure if that means this will still need to wait the full 14 days to be able to merge these or not though). I thought leaving it up to a TU to decide would be enough? If you want the packages you can have ownership of them, I just found it frustrating to have to read 7 pkgbuilds to figure out which one actually worked, when it should have been 2 packages. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Fri, Jun 13, 2014 at 11:41 PM, Timothy Redaelli tredae...@archlinux.info wrote: On Thursday 12 June 2014 23:11:18 Justin Dray wrote: Hey, I've recently cleaned up the big pile of skype4pidgin packages in the AUR. Of the previous maintainers I could contact, they have orphaned the packages and I have taken them up prior to deletion/merging or have agreed to the merge. The maintainer for skype4pidgin-svn-dbus has not responded to my any attempts to sort out these packages so far (this was only around 10 days ago however. I'm not sure if that means this will still need to wait the full 14 days to be able to merge these or not though) Anyway, as a result, can someone please merge the following: skype4empathy-svn [1] - skype4pidgin-svn [3] skype4pidgin-svn-dbus [2] - skype4pidgin-svn [3] skype4empathy [4] - skype4pidgin [6] skype4finch [5] - skype4pidgin [6] I'm the maintainer of skype4pidgin-svn-dbus, you last mail was from 3 days ago and you wrote: Hi again everyone, If there is no response in the next 14 days regarding this stuff, I'll ask for orphanage on the packages in question so that I can fix it up. 3 days are NOT 14 days... -- Timothy M. Redaelli Arch Linux Trusted User
[aur-general] Merges
Hey, I've recently cleaned up the big pile of skype4pidgin packages in the AUR. Of the previous maintainers I could contact, they have orphaned the packages and I have taken them up prior to deletion/merging or have agreed to the merge. The maintainer for skype4pidgin-svn-dbus has not responded to my any attempts to sort out these packages so far (this was only around 10 days ago however. I'm not sure if that means this will still need to wait the full 14 days to be able to merge these or not though) Anyway, as a result, can someone please merge the following: skype4empathy-svn [1] - skype4pidgin-svn [3] skype4pidgin-svn-dbus [2] - skype4pidgin-svn [3] skype4empathy [4] - skype4pidgin [6] skype4finch [5] - skype4pidgin [6] [1] https://aur.archlinux.org/packages/skype4empathy-svn/ [2] https://aur.archlinux.org/packages/skype4pidgin-svn-dbus/ [3] https://aur.archlinux.org/packages/skype4pidgin-svn/ [4] https://aur.archlinux.org/packages/skype4empathy/ [5] https://aur.archlinux.org/packages/skype4finch/ [6] https://aur.archlinux.org/packages/skype4pidgin/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Deletion request
Please remove the following pkgbuild: https://aur.archlinux.org/packages/xf86-input-evtouch/ Upstream is dead, package hasn't been able to be built in *years* and the maintainer has said as much on the package comments, and someone has also put this on the archwiki touchscreen page. Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] ttf-droid-sans duplicate of community/ttf-droid?
I thought he was being sarcastic, and was implying that we should delete it, because what is the point of saving a couple MB on removing a small portion of a font at the expense of more packages to maintain. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Thu, May 29, 2014 at 9:43 AM, Pedro Alejandro López-Valencia palop...@gmail.com wrote: El may 28, 2014 4:47 PM, Bartłomiej Piotrowski b...@bpiotrowski.pl escribió: On Wed, 28 May 2014 16:19:16 -0500 Pedro Alejandro López-Valencia palop...@gmail.com wrote: El may 28, 2014 1:31 PM, Bartłomiej Piotrowski b...@bpiotrowski.pl escribió: On Sun, 25 May 2014 12:57:43 +0200 SanskritFritz sanskritfr...@gmail.com wrote: On Sat, May 24, 2014 at 9:35 AM, Jerome Leclanche adys...@gmail.com wrote: I don't really understand the difference between community/ttf-droid and any of these: https://aur.archlinux.org/packages/?O=0K=ttf-droid Any idea about duplicates in there? ttf-droid-sans Comment by graysky 2012-10-16 21:14 This is only the sans version; the package in [community] contains more flavors. I'm a bit unsure if it's worth a standalone package. Unwanted files can be ignored via NoExtract in pacman.conf… Not in this case. The packege in community should really be split into several different ones that reflect the style and width variants, namely normal vs codensed and sans vs serif. This way we should also split ttf-bitstream-vera, ttf-dejavu, ttf-freefont, ttf-liberation and ttf-ubuntu-family. Why limit ourselves to only normal, condensed and so on? Let's split bold and monoscape variants too, because it can save 15MB of our incredibly small hard drives. I'll wait a week with any further actions to see if fellow TUs have different opinion. I can do it, if you don't mind having someone give you a hand with the task. I'm not a TU, obviusly.
[aur-general] Merge request
Hi, Can I please get [1] 'sickbeard-tvrage-git' merged in to [2] 'sickrage-git'. Upstream changed names, just following suit. [1] https://aur.archlinux.org/packages/sickbeard-tvrage-git/ [2] https://aur.archlinux.org/packages/sickrage-git/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Disown request
Hi, Can I please get shellcheck disowned? I would like to take over for it, and have gotten no response from the current maintainer. https://aur.archlinux.org/packages/shellcheck/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] feed the beast
Hey everyone, [1] has been out of date for many years and provided the same application as [2]. Please merge [1] in to [2]. Please merge [3] in to [4] as I have renamed the package to be more accurate. [1] https://aur.archlinux.org/packages/ftb-launcher/ [2] https://aur.archlinux.org/packages/feedthebeast/ [3] https://aur.archlinux.org/packages/feedthebeast-server/ [4] https://aur.archlinux.org/packages/ftb-ultimate-server/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Code review request: new PKGBUILD omodoro
For the git pkgver functions, you're best off checking the wiki: https://wiki.archlinux.org/index.php/VCS_PKGBUILD_Guidelines The first one you posted is the only one listed there. The function you are using at the moment will return a random version as it's just the fingerprint of the latest commit; That means it may or may not get seen as an update when people go to build the next version. You need to make sure it is something that makes sense and increments as new commits are made. You should rarely have to deviate from what is on that VCS PKGBUILD guidelines page. Regards, Justin Dray
Re: [aur-general] Promoting use of .AURINFO
On Mon, Jan 13, 2014 at 6:17 AM, Jerome Leclanche adys...@gmail.com wrote: On Sun, Jan 12, 2014 at 3:11 PM, Lukas Fleischer archli...@cryptocrack.de wrote: On Sun, 12 Jan 2014 at 15:51:48, Anton Larionov wrote: Hello, I was under the impression that .AURINFO was introduced to override some fields in PKGBUILD when they are written in format which can't be properly displayed by AUR (or maybe I've missed something). But why do you want to force it's usage for all packages? In most cases AURINFO will just duplicate same fields from PKGBUILD. The long-term plan is to use it for all AUR packages, improve the format and finally make it an official feature of makepkg(8) (the file will probably be called .SRCINFO then but we're far from there). See my other reply to Sebastien for some reasons on why it should be used. So the official goal is to have it generated as part of makepkg -S? Because I see that as the only way the format will get popular: if it's nobody's problem. J. Leclanche Also I have some questions about it's format: 1) If package has different dependencies for 86_64 and 686, what should I put in depend array? Good question. This cannot be answered properly, though, since dependencies actually are a property of the binary package and not a property of the source package. Maybe we should loosen the format for dependencies of source packages and allow optdep-like comments? Something like: depends = foo depends = bar depends = foobar: x86_64 only Just an idea. Comments welcome. 2) Which 'pkgname' will be unique - from PKGBUILD or AURINFO? E.g if I upload package with name 'foo' and overriden name 'bar' will someone be able to upload new package with name 'foo'? Or 'bar'? Only the information from .AURINFO will be used. You can already trick the AUR into reading a completely different name from the PKGBUILD than it actually produces (and that problem is unavoidable), so that isn't a (new) issue. -- Regards, Anton Larionov It might also be a good idea to write out what fields are available and their purpose on the wiki similar to the PKGBUILD page ( https://wiki.archlinux.org/index.php/PKGBUILD) and perhaps link to it from the AUR user guidelines page? It will be forgotten by most packagers if the only information about it is a commit message and a mailing list thread. Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] linux-bcache-git
Hey everyone, Can we get linux-bcache-git package removed? All the changes have been merged upstream several kernel versions ago and even it's maintainer has said as much in the comments. https://aur.archlinux.org/packages/linux-bcache-git/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Removal of repo-update
Hey, the repo-update package is out of date, hasn't been updated in several years and upstream source is gone, can we get this one deleted? Thanks. https://aur.archlinux.org/packages/repo-update/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] 'provides' info in AUR
On Thu, Dec 12, 2013 at 7:11 AM, Dave Reisner d...@falconindy.com wrote: On Wed, Dec 11, 2013 at 08:21:44PM +, Jerome Leclanche wrote: On Wed, Dec 11, 2013 at 6:42 PM, Dave Reisner d...@falconindy.com wrote: On Wed, Dec 11, 2013 at 06:12:59PM +, Jerome Leclanche wrote: On Wed, Dec 11, 2013 at 5:52 PM, Sam S. sml...@gmail.com wrote: On Tue, Dec 10, 2013 at 4:21 PM, Jerome Leclanche adys...@gmail.com wrote: I don't understand why makepkg -S doesn't include the .PKGINFO file from makepkg (and subsequently the AUR would use that if present instead of the grep system which fails as soon as variables/expansions are involved which is every other package). All the implementation is there. It would probably be better then what we have now, but a perfect solution would also account for PKGBUILDs that use Bash conditionals to set different variables on i386 systems than on x86_64 systems (which is pretty common among AUR packages that package upstream binaries rather than compiling from source). Reading values from .PKGINFO would populate the AUR with the values for whichever architecture the package uploader happened to use. (So if the maintainer changes, or the same maintainer works on different computers, a simple re-upload of an AUR package could suddenly change the package's meta-info, i.e. the AUR would become more fragile.) Of course, the perfect solution would be pretty difficult to implement. Gentoo had a GSoC project last year [1], to implement an efficient and safe (side-effect free) limited Bash parser / pseudo-interpreter in C++, sufficient to extract all necessary values from Genoo's equivalent of PKGBUILDs. Surely, this could have been useful for the AUR as well. But I can find no evidence of continued project activity after the GSoC period ended, so it appears they have given up... :( --- [1] http://dev.gentoo.org/~qiaomuf/libbash.html I love the fact someone could be working on a bash parser but that solution is *insane*. It's designed to be incomplete, and the project appears very dead. This is a solved problem: use a metafile compiled by whatever tools you use in your distro/domain that can be parsed safely and easily. For Arch, those are PKGINFOs. No, go see historical conversations about a mythical .SRCINFO -- this is what .AURINFO is based on. .SRCINFO is vaporware right now, and I again refer you to unresolved discussions about how it would handle split packages and package-specific overrides. Good point on the differences per arch. I guess the obvious solution that comes to mind here is to have makepkg -S generate the source files for each arch value (eg. PKGINFO.x86_64, etc) but that's not necessarily good and is the subject of another discussion imo. To be clear, .PKGINFO is not the solution here. This file describes a built package (something the AUR explicitly does not support). d Care to explain the reasoning? I'm looking at a few example PKGINFOs and they contain nothing that can exclusively be generated at package() or build() time (other than pkgver but that's already the case currently). J. Leclanche Here's a few reasons that come to mind: 1) As you've already discovered, you'd need a .PKGINFO file for every potential architecture, rather than just a .SRCINFO which might describe what architectures are known available. A .SRCINFO could express all package and architecture specific overrides. The fact that an override exists might be a useful bit of information to convey. 2) .PKGINFO doesn't contain things that are useful for source packages, like, say... the source URL? checksums? 3) You can't possibly express split packages well. Having pkgname fields in a .SRCINFO would mean you could describe all packages created by a PKGBUILD rather than having some loose association between a PKGBUILD and some possible .PKGINFO files which it might have generated. This is about *source* packages. The metadata needs for source packages are not the same as the needs for binary packages. What is to stop us using both? I don't think anyone has said we should stop parsing the PKGBUILD for the AUR and exclusively parse a .PKGINFO instead. It would just provide some supplementary information that isn't currently available in an easy to parse method from the PKGBUILD without running arbitrary code on the server side. Reading both files would solve that, right? Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Multiple removals
Hey, There are a ton of zeromq packages on the AUR even though there is zeromq v4.xx in community [1] zeromq-pgm (OpenPGM support is already compiled in to the one in community) [2] zeromq3-git [3] zeromq-dev (there is already a zeromq-git (which is also out of date) but this 'dev' version is based on a beta of version 3...) I'm not sure about zeromq2 as it is a dependency for zeromqgw, so maybe it should be left alone for now. [1] https://aur.archlinux.org/packages/zeromq-pgm/ [2] https://aur.archlinux.org/packages/zeromq3-git/ [3] https://aur.archlinux.org/packages/zeromq-dev/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Deletion Request
Hi, Can I please get unrar4[1] removed? Unrar5 is in extra already; v4 was kept until pyunrar2 was updated to support the new version, which it was the other day. [1] https://aur.archlinux.org/packages/unrar4/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Remove gmailreader
Hi, The gmailreader [1] package hasn't been updated in 5 years. It uses an unmaintained library that is incompatible with gmail and has been for a long time. Can we please remove this package from the AUR? [1] https://aur.archlinux.org/packages/gmailreader/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Delete sixpair-bluez
Can we please get sixpair-bluez removed: https://aur.archlinux.org/packages/sixpair-bluez/ It has been out of date for many years and there is already a sixpair package that is updated. Regards, Justin Dray E: jus...@dray.be M: 0433348284
[aur-general] Deletion request ffmpeg-full-git-fixed
Trying to navigate the pile of different versions of ffmpeg on AUR is quite a hassle currently. Please either delete ffmpeg-full-git-fixed or merge it in to ffmpeg-full-git. It appears to have only been made due to the original maintainer of ffmpeg-full-git not responding quickly, and the maintainer of ffmpeg-full-git-fixed has stated in the comments that he no longer maintains it. It has also been marked as out of date for 2+ months. https://aur.archlinux.org/packages/ffmpeg-full-git-fixed/ Regards, Justin Dray E: jus...@dray.be M: 0433348284
Re: [aur-general] Deletion Request - rEFIt
Slim had a decent sized update about 2 weeks ago, it started supporting the normal xsession methods rather than a specific list and a variety of other things, the version before that was less than a week old. It's stable, but certainly not dead. Regards, Justin Dray E: jus...@dray.be M: 0433348284 On Wed, Oct 16, 2013 at 11:31 AM, Sam Stuewe halosgh...@archlinux.infowrote: On 2013-10-15 13:56, Keshav Padram Amburay wrote: Hi, rEFIt is no longer maintained upstream and its author recommends everyone to switch to rEFInd instead, at http://refit.sourceforge.net/ . So please delete AUR refit https://aur.archlinux.org/**packages/refit/https://aur.archlinux.org/packages/refit/package. With Best Regards, Keshav I am not a TU, so this is not an official set of logic, but if the sources still exist (even if it has been deprecated or is no longer maintained), I see no reason for the package to be removed unless there is a specific and blatant security flaw. For example, SLiM is still in the official repos despite maintenance largely being dead and many other packages being considered preferable. If people would like to continue using rEFIt, despite rEFInd's superiority, I don't see the inherent issue. All the best, -Sam
[aur-general] Orphan request: lsiutil
I've tried emailing the current maintainer of lsiutil as it is out-of-date and have not had any response. I'd like to update it/maintain it. Regards, Justin Dray E: jus...@dray.be M: 0433348284