Re: [gentoo-user] Re: per-package CFLAGS
Holger Hoffstätte wrote: On Tue, 27 Jan 2015 14:35:35 +0100, Raffaele BELARDI wrote: I have a package failing build (media-tv/mythtv). One resource on the net suggests using the -fno-devirtualize gcc flag. Google tells me that the way to do that would be something like $ echo CFLAGS=\${CFLAGS} -fno-devirtualize\ /etc/portage/env/media-tv/mythtv but the references are pretty old. Is this still the preferred way? This alone won't work and might also be more confusing than necessary, as the package-hierarchy-like naming is not required. It's all documented here: http://wiki.gentoo.org/wiki//etc/portage/env -h Perfect, and very simple! thanks, raffaele
[gentoo-user] ffmpeg-2: when getting stable?
Hi all, when can we expect ffmpeg-2 to get stable? I found some information that it's because of building issues but I found none. I looked into /usr/portage/profiles/package.mask and there's no information about the reason. Thanks in advance for your answer, Jan Sever
[gentoo-user] Re: per-package CFLAGS
On Tue, 27 Jan 2015 14:35:35 +0100, Raffaele BELARDI wrote: I have a package failing build (media-tv/mythtv). One resource on the net suggests using the -fno-devirtualize gcc flag. Google tells me that the way to do that would be something like $ echo CFLAGS=\${CFLAGS} -fno-devirtualize\ /etc/portage/env/media-tv/mythtv but the references are pretty old. Is this still the preferred way? This alone won't work and might also be more confusing than necessary, as the package-hierarchy-like naming is not required. It's all documented here: http://wiki.gentoo.org/wiki//etc/portage/env -h
Re: [gentoo-user] Re: per-package CFLAGS
On Tue, Jan 27, 2015 at 8:54 AM, Holger Hoffstätte holger.hoffstae...@googlemail.com wrote: On Tue, 27 Jan 2015 14:35:35 +0100, Raffaele BELARDI wrote: I have a package failing build (media-tv/mythtv). One resource on the net suggests using the -fno-devirtualize gcc flag. Google tells me that the way to do that would be something like $ echo CFLAGS=\${CFLAGS} -fno-devirtualize\ /etc/portage/env/media-tv/mythtv but the references are pretty old. Is this still the preferred way? This alone won't work and might also be more confusing than necessary, as the package-hierarchy-like naming is not required. This alone works fine for me, and I don't find the naming confusing. I'm sure the way you referenced would also work, though it ends up dumping the configuration for every package into a single file. cat /etc/portage/env/media-tv/mythtv CFLAGS=-march=amdfam10 -Os -pipe -frename-registers -fweb CXXFLAGS=-march=amdfam10 -Os -pipe -frename-registers -fweb I would recommend that instead of just putting flags for each package in a separate file that you instead create a per-package config and then symlink it from the package name. As in: ls -l /etc/portage/env/media-tv/mythtv lrwxrwxrwx 1 root root 13 Oct 5 2012 /etc/portage/env/media-tv/mythtv - ../noparallel -- Rich
Re: [gentoo-user] [Extremely OT] Ansible/Puppet replacement
On 27/01/2015 19:49, Alec Ten Harmsel wrote: On 01/27/2015 11:33 AM, Alan McKinnon wrote: On 27/01/2015 10:49, Tomas Mozes wrote: I haven't tested it yet, however I like the minimalistic syntax. As an ansible user - do you plan to allow using default values for modules and/or variables? +1 for that. I'm also a happy ansible user with zero plans to change, but I can't imagine a deployment tool without sane rational explicit defaults. A whole host of problems simply stop being problems if that feature is available. I'm curious, what exactly do you mean about default values? Is there a small example you can give me? The tutorial on Ansible's website is a little confusing. When I saw your other reply to Tomas, I thought you might ask that question :-) What I was thinking of is defaults like you find in roles on galaxy. In a sub-dir called defaults you find a file main.yml containing variables defaults as key-value pairs. A good example is Stouts.openvpn, it has eg openvpn_port: 1194 openvpn_proto: udp If you don't define those vars yourself in the playbook, the role uses those defaults, which is exactly what you want - they match upstream default. The location is very explicit, the defaults are in a named file in an obvious location and there's no mysterious automagic. I definitely wasn't talking about crazy magic permission modes like the suggestion you mentioned in the other mail. A default mode is fine, but make it a variable called mode or umask that we can see in a file and make it what we want. I think we do agree on this :-) -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] [Extremely OT] Ansible/Puppet replacement
On 2015-01-26 16:30, Alec Ten Harmsel wrote: Hi, I've been working on my own replacement for Bcfg2 - bossman[1] - over the past few months, and it's finally ready to be released in the wild. I would be honored if anyone on this list who's thinking of trying puppet, chef, ansible, bcfg2, etc. would try out bossman instead. bossman has an incredibly simple syntax; no ruby DSLs or XML. My main motivation for writing it was dealing with bcfg2's XML config on a daily basis. Additionally, bossman has a (hopefully) great 'pretend' mode and checks for a lot of errors. If you are already using another solution and have some time to check out bossman, I would love feedback. The only config manager I've used in practice is bcfg2, so getting perspectives from those using other solutions would be fantastic. bossman is written in C99 and is built with CMake. I don't recommend it for production deployments quite yet, but I plan on actively working on it. It currently only supports pulling configuration from a mounted filesystem (i.e. local disk, NFS, etc.), but HTTP support (and a deployment tutorial/guide) will be added in v0.2. I'm sorry to spam gentoo-user, but I'm not sure who else would be interested in something like this. Also, feel free to email me with bugs in the code or documentation, or open something in GitHub's issue tracker. Alec [1] https://github.com/trozamon/bossman/archive/v0.1.0.tar.gz [2] https://github.com/trozamon/bossman-roles I haven't tested it yet, however I like the minimalistic syntax. As an ansible user - do you plan to allow using default values for modules and/or variables?
Re: [gentoo-user] ffmpeg-2: when getting stable? [SOLVED]
I found the bugzilla: https://bugs.gentoo.org/show_bug.cgi?id=510340 So marking as SOLVED. Thank you, Jan Sever P.S. Btw. poli in politics is from latin word polis = a place, not poly = many. But it's a nice parallel. -- Původní zpráva -- Od: Neil Bothwick Komu: gentoo-user@lists.gentoo.org Datum: 27. 1. 2015 15:15:51 Předmět: Re: [gentoo-user] ffmpeg-2: when getting stable? On Tue, 27 Jan 2015 14:41:48 +0100 (CET), Jan Sever wrote: when can we expect ffmpeg-2 to get stable? I found some information that it's because of building issues but I found none. I looked into /usr/portage/profiles/package.mask and there's no information about the reason. package.mask files control masking of packages, they have nothing to do with keywording, which is what stable/testing, arch/~arch is. I'd look on bug.gentoo.org for a stabilisation request for ffmpeg-2 and file one if it hasn't been done already. -- Neil Bothwick Politics: Poli (many) - tics (blood sucking parasites)
[gentoo-user] virtualbox installation problem
Hello everyone. I want to install virtualbox, and for some reasons I need to download the distfiles before the emerging process, so the emerge command wont try to download anything. I have download the required files, but extension pack ebuild ignores the Oracle_VM_VirtualBox_Extension_Pack-4.3.18-96516.vbox-extpack file that I have downloaded and tries to fetch it again. Is there anything that I am missing? Thanks for your help.
Re: [gentoo-user] grub - gummiboot: good
On Tuesday 27 Jan 2015 21:21:05 Stefan G. Weichinger wrote: Just wanted to share that I switched from booting via grub2 to gummiboot ... UEFI only now, sure. Back then it was quite a hassle to get my grub2-setup working, back then I had software raid and LVM and stuff ... today with my btrfs-only installations it was a very quick learning process. Just one boot from sysresccd to get the paths right ... on the 2nd machine I got it right at the first time. Removed grub already and cleaned up my setup ... nice. Additionally I adjusted my kerninst.conf (one more pointer to Canek's helpful tool kerninst at [1]) to write the correct loader-entries ... works as well already. No actual experience with dual-booting windows or so ... but according to the docs that should work out as well. .. I am now into checking [2] : just curious how minimal an fstab could get. Just some recommendation ... if someone is installing gentoo onto a clean UEFI-system, gummiboot might be easier to use than grub2 imo. Stefan - [1] https://github.com/canek-pelaez/kerninst [2] http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec/ ... and if you have no need for multi-booting then the kernel efi stub is very simple to configure and still works reliably (with openrc and no initrd here). No need to install a separate boot manager. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] virtualbox installation problem
On Wed, 28 Jan 2015 01:39:59 +0330, behrouz khosravi wrote: I want to install virtualbox, and for some reasons I need to download the distfiles before the emerging process, so the emerge command wont try to download anything. I have download the required files, but extension pack ebuild ignores the Oracle_VM_VirtualBox_Extension_Pack-4.3.18-96516.vbox-extpack file that I have downloaded and tries to fetch it again. Is there anything that I am missing? The output from the emerge command? Chances are the file has changed recently and does not match the size or checksum known to portage. Try re-syncing. -- Neil Bothwick Top Oxymorons Number 16: Peace force pgpaxKoUneLOW.pgp Description: OpenPGP digital signature
Re: [gentoo-user] virtualbox installation problem
The output from the emerge command? Chances are the file has changed recently and does not match the size or checksum known to portage. Try re-syncing. Thanks. It just got fixed using the hints in the IRC channel. The problem was that ebuild renames the file: Oracle_VM_VirtualBox_Extension_Pack-4.3.18-96516.vbox-extpack to Oracle_VM_VirtualBox_Extension_Pack-4.3.18-96516.tar.gz but I had though that it compresses the file. A simple renaming solved the problem. Thanks.
[gentoo-user] Ghost cyber threat
Does anybody know more about this security flaw in the open-source Linux GNU C Library http://www.theglobeandmail.com/technology/linux-makers-release-patch-to-thwart-new-ghost-cyber-threat/article22662060/?cmpid=rss1 -- Joseph
Re: [gentoo-user] grub - gummiboot: good
On Wed, 28 Jan 2015 00:40:32 +0100, Stefan G. Weichinger wrote: No need for chainloading with UEFI. Set Gummiboot as the default boot option and hold down Esc (or whichever key your motherboard uses, it would have been nice if UEFI had standardised that too) when booting if you want Grub instead So I would have to have grub-2.x and gummiboot installed in parallel? Both as UEFI-boot-entries, right? Yes. When I first tried Gummiboot, I left Grub as the default until I was happy with Gummiboot, then I changed the default with efibootmgr. -- Neil Bothwick What if there were no hypothetical situations? pgp8FAgeF_z0f.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Ghost cyber threat
150127 Joseph wrote: Does anybody know more about this security flaw in the open-source Linux GNU C Library : http://www.theglobeandmail.com/technology/linux-makers-release-patch-to-thwart-new-ghost-cyber-threat/article22662060/?cmpid=rss1 Acc to this, it was patched 2013 today threatens only long-term systems : http://threatpost.com/ghost-glibc-remote-code-execution-vulnerability-affects-all-linux-systems/110679 I'm running 2.19-r1 , installed 140802 ; vulnerable are 2.18 . Linux systems are at risk only when admins don't keep versions upto-date. My netbook was running 2.15-r3 (121216) till I updated it this week (I haven't used it since 130127 ordinarily don't use it for I/net). -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
[gentoo-user] Ghost cyber threat
https://www.qualys.com/research/security-advisories/GHOST-CVE-2015-0235.txt
[gentoo-user] dependancy poppler and xorg-server
I'm getting two dependancy errors during upgrade: !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: x11-base/xorg-server:0 (x11-base/xorg-server-1.15.2-r1:0/1.15.2::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (x11-base/xorg-server-1.13.4-r1:0/1.13.4::gentoo, installed) pulled in by x11-base/xorg-server:0/1.13.4= required by (x11-drivers/xf86-input-mouse-1.9.0:0/0::gentoo, installed) ^^ (and 5 more with the same problem) app-text/poppler:0 (app-text/poppler-0.26.5:0/46::gentoo, installed) pulled in by app-text/poppler:0/46=[cxx,jpeg,lcms,tiff,xpdf-headers(+)] required by (net-print/cups-filters-1.0.53:0/0::gentoo, installed) ^^ (and 1 more with the same problem) (app-text/poppler-0.24.5:0/44::gentoo, ebuild scheduled for merge) pulled in by =app-text/poppler-0.12.3-r3:0/44= required by (app-text/texlive-core-2012-r1:0/0::gentoo, installed) 1.) I've unmerge poppler-0.24.5 rebuild: net-print/cups-filters dev-tex/luatex 2.) Do I unmerge x11-base/xorg-server-1.13.4-r1? -- Joseph
Re: [gentoo-user] Rkhunter now showing Warnings for two files: /bin/egrep fgrep
On 1/26/2015 5:53 PM, Neil Bothwick n...@digimed.co.uk wrote: On Mon, 26 Jan 2015 11:27:05 -0500, Alec Ten Harmsel wrote: script: /bin/fgrep: POSIX shell script, ASCII text executable Anyone know if this is due to something changing in Gentoo? Upstream changed egrep and fgrep from binaries to shell scripts. This happened a while ago on testing portage but the version with the change only hit stable at the weekend. You can tell rkhunter to ignore them. % grep grep /etc/rkhunter.conf.local SCRIPTWHITELIST=/bin/egrep SCRIPTWHITELIST=/bin/fgrep Perfect, thanks Alec/Neil, problem solved... :)
Re: [gentoo-user] ffmpeg-2: when getting stable?
On Tue, 27 Jan 2015 14:41:48 +0100 (CET), Jan Sever wrote: when can we expect ffmpeg-2 to get stable? I found some information that it's because of building issues but I found none. I looked into /usr/portage/profiles/package.mask and there's no information about the reason. package.mask files control masking of packages, they have nothing to do with keywording, which is what stable/testing, arch/~arch is. I'd look on bug.gentoo.org for a stabilisation request for ffmpeg-2 and file one if it hasn't been done already. -- Neil Bothwick Politics: Poli (many) - tics (blood sucking parasites) pgpkvlPOr20pr.pgp Description: OpenPGP digital signature
Re: [gentoo-user] grub - gummiboot: good
On Tue, 27 Jan 2015 22:31:21 +, Mick wrote: On Tuesday 27 Jan 2015 21:21:05 Stefan G. Weichinger wrote: Just wanted to share that I switched from booting via grub2 to gummiboot ... UEFI only now, sure. I switched to Gummiboot a while ago, after Canek mentioned it I think, and I am glad I did. I still keep GRUB around because Gummiboot cannot boot from an ISO image so Grub stays for my sysresccd emergency boot option. ... and if you have no need for multi-booting then the kernel efi stub is very simple to configure and still works reliably (with openrc and no initrd here). No need to install a separate boot manager. It's not only for multi-booting. AFAIR you need to build the kernel arguments into the kernel when booting like this, which means you have only one option for each kernel. I prefer to have a No X option as well as a desktop boot. Gummiboot just makes things easy without getting in the way. -- Neil Bothwick To boldly go where I surely don't belong. pgpW9cMvru0Sy.pgp Description: OpenPGP digital signature
Re: [gentoo-user] ffmpeg-2: when getting stable? [SOLVED]
On Tue, 27 Jan 2015 23:07:15 +0100 (CET), Jan Sever wrote: P.S. Btw. poli in politics is from latin word polis = a place, not poly = many. But it's a nice parallel. Yes, and a tic is a nervous twitch, the blood sucking parasite is a tick, but when did jokes have to be linguistically accurate? -- Neil Bothwick Sacred cows make great hamburgers. pgpHjeugYUQ2H.pgp Description: OpenPGP digital signature
Re: [gentoo-user] grub - gummiboot: good
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 27.01.2015 23:47, Neil Bothwick wrote: On Tue, 27 Jan 2015 22:31:21 +, Mick wrote: On Tuesday 27 Jan 2015 21:21:05 Stefan G. Weichinger wrote: Just wanted to share that I switched from booting via grub2 to gummiboot ... UEFI only now, sure. I switched to Gummiboot a while ago, after Canek mentioned it I think, and I am glad I did. I still keep GRUB around because Gummiboot cannot boot from an ISO image so Grub stays for my sysresccd emergency boot option. Yep, correct. That was *nice to have* afaik you could chainload grub from gummiboot to do that . ? ;) I think I get on with plugging in that usb-stick when it really is necessary booting from the ISO wasn't that common for me in the last year or so. For now I prefer the simplicity of the whole boot-environment. Gummiboot just makes things easy without getting in the way. ACK! -BEGIN PGP SIGNATURE- Version: GnuPG v2 iQIcBAEBAgAGBQJUyBcPAAoJEClcuD1V0Pzm2w8P/RtWCeIJIRSDRVhZYvI1ufj2 7/IKCgreXYXLBGRbP/llNdQKIMyPw7Z2vl/lWFmab10257fKXBTUVglNo7nHCz4K OoqzYHzF0FyZPgsnMp9vb/CjT7UOU3E86aZMUKSSeiY4RfaNsN1dztS1si1QHpAk t9/K3YeumLqhm3b+YyCokqNjDKPllmWfwCxnv6yYsQ6UvZjSI1/JNnFMYrosanFo g+fpt9YhbVhIotBbfiWHOEQ7vj+W9HnLZlvQx/+a9zdnh9x0FjWM2tDJJmj6yOBz ZWKPAqONJ9uARBfnoEM4os4xjxObyQS+ukfOXG7N7ZJEITFGLIKLd1+5Os0SVds7 LjV5XG7UywNZ6dLVZ+Ww4/g3lwnOeGjZBzDCqxAQydJHQlFMS+k5/Lwu9pja2TAo 7GPkyDmH2JoBDdkn8GeENg3nq0qj/eCeaqntd9w8RuvF1sUD/JGG1o+l/iQnNILJ jYlWVPOVV0rVKsFefajrQCvR523dT8Y8x2iaQ2coYqpIc9UvszuWF++j25GYpcna hPnV0zPGq8nqjtl45KUfEJsvCxOWjoEtmgxhGYQ/W5twFFqtk/gFM+Mt3gY2+F+N 0SIpCCHJYkuRCc8sHOd+UCDf8ipJEV9Ss56EoPcFSB1TSoIHV5LsGKk1DM9WJR+p 1Z8Bzf9I0/45VRyoGS+x =/q4J -END PGP SIGNATURE-
Re: [gentoo-user] grub - gummiboot: good
On Tue, 27 Jan 2015 23:54:07 +0100, Stefan G. Weichinger wrote: I switched to Gummiboot a while ago, after Canek mentioned it I think, and I am glad I did. I still keep GRUB around because Gummiboot cannot boot from an ISO image so Grub stays for my sysresccd emergency boot option. Yep, correct. That was *nice to have* afaik you could chainload grub from gummiboot to do that . ? ;) No need for chainloading with UEFI. Set Gummiboot as the default boot option and hold down Esc (or whichever key your motherboard uses, it would have been nice if UEFI had standardised that too) when booting if you want Grub instead I think I get on with plugging in that usb-stick when it really is necessary booting from the ISO wasn't that common for me in the last year or so. Which makes it even more unlikely you'll be able to lay your hands on the stick when you really need it :( -- Neil Bothwick The cow is nothing but a machine which makes grass fit for us people to eat. pgpZIGsrkDHni.pgp Description: OpenPGP digital signature
Re: [gentoo-user] grub - gummiboot: good
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 28.01.2015 00:35, Neil Bothwick wrote: afaik you could chainload grub from gummiboot to do that . ? ;) No need for chainloading with UEFI. Set Gummiboot as the default boot option and hold down Esc (or whichever key your motherboard uses, it would have been nice if UEFI had standardised that too) when booting if you want Grub instead So I would have to have grub-2.x and gummiboot installed in parallel? Both as UEFI-boot-entries, right? Maybe I look into this in the next days, sure ... I think I get on with plugging in that usb-stick when it really is necessary booting from the ISO wasn't that common for me in the last year or so. Which makes it even more unlikely you'll be able to lay your hands on the stick when you really need it :( ;-) -BEGIN PGP SIGNATURE- Version: GnuPG v2 iQIcBAEBAgAGBQJUyCHwAAoJEClcuD1V0PzmeuAP/AsjlJZsHX8+0/nOpgpQA5EP EriMN9jSY/JkcSNF6XWVp8+UbqKrKqqGlh5as3d+RbXHrH6B/OMw8bHVT0PbY+/3 zjpF7PimSMD4iCLSkecHpqBS3fgWr6k/u331ajd1P7jEHv/ZbJVg/CcG/aQpM8cZ k8XSZpbO2oZoHkCufptHGPb9J6EOm1i0pJzVCN8XaHdvb/lIV5K/bhEEL51DlFWY aGjq6Jm7F3oKxnwk14JyLhp/151jZZ22t+71F+foO/A22RU+cS9UxTOjCkVBbIUP 4hwNCK/wRzPela0KNe/jigIRqzEnX84WZ0uleLB5LtjsCYJWmq7BbkG7cnhOjOJ1 Waj0RIvywlNf0XfhMHXYzeql9zSuUGJ/hkHT9RXWuUzxzy+xRUbMQ1f1Xw/AiNNP Vz4cqsr6AbZn2Cyij0qPSd14++/FK2w7Mc+6T6ADbVazRwexmJJBlXvuhNL/MhbF UCBGGmt5quspQa1ZZMFeReorrquZxRNRIYI5ZjRyE/M6Ofarxv/sFwlB0a8gLHxL z/PQNQWMCAGB8p74NOpmMp+kSbygwbECXADN2w7l3MoV1Bvb5z240Z1uu7rVEMaW aq9LPpgfWTQYCDcxZiV9bc4MWfdqcOz7RoX5oGr+sxFD504PxKgWTuTupqKl2/AQ IznBCpVxBXLbeXnmrmZY =hNbM -END PGP SIGNATURE-
Re: [gentoo-user] [Extremely OT] Ansible/Puppet replacement
On 01/27/2015 03:49 AM, Tomas Mozes wrote: I haven't tested it yet, however I like the minimalistic syntax. Thanks. Testing is kind of a high barrier; it took me hours to write a role for jenkins. As an ansible user - do you plan to allow using default values for modules and/or variables? I am not really a fan of default values; I prefer explicitness over forcing someone to have a complete knowledge of bossman. For example, one of my coworkers pointed out to me that file /etc/ssh/sshd_config should install `/etc/ssh/sshd_config` with whatever owner/permissions it has on disk in the configuration files directory. In other words, he wanted the implicit defaults to make it easier. Personally, I think this is harder because now I have to care about permissions of a file that is being copied and writing a role; instead I like having explicit roles in which it is mandatory to type file /etc/ssh/sshd_config root:root 600 This makes it easier to see everything on the same terminal, don't have to care about local permissions, etc. I'm trying to write a config manager that has the fewest amount of gotchas possible, and I think that means that a role should explicitly describe the state a machine should end up in. I do not have variables, but I might add them if someone gives me a compelling use case that cannot be done without them. If this didn't answer your question, my bad. Thanks for the feedback. I know a lot of the people on this list work full-time and have families; there's always time to procrastinate on homework though ;) Alec
[gentoo-user] Managing a Build Server (as chroot per server or single dedicated)
Hello @all... I am currently in the process of revamping my server infrastructure and two things on my list are to finally use Salt (CM system) and to reduce the update time for each server and have a small staging area before I emerge packages/updates to the live system. So, I wanted to ask the community for advice on how they deal with the whole build server issue -- especially if you go for a build chroot on the same machine? Before I go out and reinvent the wheel for all sorts of problems and tools, I thought this is definitely something a lot of people have run into before. :) Why a build chroot on each server instead of a dedicated build server? For one thing, scaling up for me means, I will maintain more servers for my clients (I am self-employed) which results in a server park that is generally diverse and not all that homogeneous. I only have very few servers I own and use myself. Secondly, if the client ever decides to cancel my services, he can still use the build/staging system as-is which is imho a nice to have feature. I generally don't believe in locking the customer in -- at all. And last but not least: A dedicated build machine is also a matter of costs and since there would be quite a few build environments running on it (containers, chroots, whatever) and factoring in other things like traffic and all, this can easily get rather expensive soonish once scaled up which is something I would like to avoid. From my perspective, a few areas are problematic: Automation of the build process for example is one area where I think it is difficult to just schedule a cron job since due to dependency and/or other conflicts, builds might not even run. Or packages change their default USE flags and might need adjustments. Or revdep-rebuilds. So, I guess there is still some manual intervention necessary here? I am just purely looking at the build server part here. The actually updating to the system is a totally different matter. Are there any tools out there to help manage a build server? I have read about people using catalyst for that -- which is not quite what I am looking for. I guess Gentoo Infra might have developed their own set of tools -- and maybe released those to the wild? :) Is a build server chroot per server really feasible and advisable or are there any good reasons against it? I would appreciate any help, input or feedback that I can get. Thanks a lot in advance. With Kind Regards from Germany, Matthias -- Dipl.-Inf. (FH) Matthias Dahl | Software Engineer | binary-island.eu services: custom software [desktop, mobile, web], server administration
Re: [gentoo-user] dependancy poppler and xorg-server
(x11-base/xorg-server-1.15.2-r1:0/1.15.2::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (x11-base/xorg-server-1.13.4-r1:0/1.13.4::gentoo, installed) pulled in by x11-base/xorg-server:0/1.13.4= required by (x11-drivers/xf86-input-mouse-1.9.0:0/0::gentoo, installed) ^^ (and 5 more I would just unmerge xf86-input-mouse, update xorg-server, merge xf86-input-mouse.
[gentoo-user] Re: ffmpeg-2: when getting stable?
Jan Sever n32 at email.cz writes: when can we expect ffmpeg-2 to get stable? I found some information that it's because of building issues but I found none. I looked into /usr/portage/profiles/package.mask and there's no information about the reason. There is currently a gentoo-dev thread that you might find informative: [gentoo-dev] RFC: USE=libav as replacement for broken || ( libav:= ffmpeg:= ) hth, James
[gentoo-user] per-package CFLAGS
I have a package failing build (media-tv/mythtv). One resource on the net suggests using the -fno-devirtualize gcc flag. Google tells me that the way to do that would be something like $ echo CFLAGS=\${CFLAGS} -fno-devirtualize\ /etc/portage/env/media-tv/mythtv but the references are pretty old. Is this still the preferred way? thanks, raffaele
[gentoo-user] Re: [Extremely OT] Ansible/Puppet replacement
Alec Ten Harmsel alec at alectenharmsel.com writes: I'm sorry to spam gentoo-user, but I'm not sure who else would be interested in something like this. Also, feel free to email me with bugs in the code or documentation, or open something in GitHub's issue tracker. One man's spam generates maps for another. So my map of todo on ansible is all about common gentoo installs. [1] Let's take the first and most easy example the clone. I have a gentoo workstation install that I want to replicated onto identical hardware (sort of like a disk to disk dd install). So how would I impress the bossman by actually saving admin time on how to use the bossman to create (install from scratch + pxe?) a clone. Gotta recipe for that using bossman? Or is that an invalid direction for bossman? curiously, James [1] http://blog.jameskyle.org/2014/08/automated-stage3-gentoo-install-using-ansible/
Re: [gentoo-user] Re: [Extremely OT] Ansible/Puppet replacement
On 01/27/2015 10:34 AM, James wrote: Alec Ten Harmsel alec at alectenharmsel.com writes: I'm sorry to spam gentoo-user, but I'm not sure who else would be interested in something like this. Also, feel free to email me with bugs in the code or documentation, or open something in GitHub's issue tracker. One man's spam generates maps for another. So my map of todo on ansible is all about common gentoo installs. [1] Let's take the first and most easy example the clone. I have a gentoo workstation install that I want to replicated onto identical hardware (sort of like a disk to disk dd install). So how would I impress the bossman by actually saving admin time on how to use the bossman to create (install from scratch + pxe?) a clone. Assuming that disks are formatted, a stage3 has been freshly extracted, bossman is installed, and the role/config files are on a mounted filesystem, it should be similar to the role below: file /etc/portage/make.conf root:root 644 ! emerge-webrsync ! emerge --sync file /etc/locale.gen root:root 600 ! locale-gen pkg sys-kernel/gentoo-sources file /usr/src/linux/.config root:root 644 ! make -C /usr/src/linux all modules_install install pkg sys-boot/grub ! grub-install /dev/sda # I can't remember all the options needed here file /etc/default/grub ! grub-mkconfig -o /boot/grub/grub.cfg # Generating /etc/fstab using something similar to Arch's `genfstab` would be much better file /etc/fstab root:root 644 # Root password file /etc/shadow root:root 640 # Logger pkg app-admin/syslog-ng # Network pkg net-misc/dhcpcd enable dhcpcd # For remote access pkg net-misc/openssh file /etc/ssh/sshd_config root:root 600 file /etc/ssh/known_hosts root:root 600 # Other sshd files... enable sshd There are a ton of assumptions that make this work; if installing manually, the installer is responsible, and if installing from PXE, this stuff would have to be baked into the ISO. Gotta recipe for that using bossman? Or is that an invalid direction for bossman? curiously, James [1] http://blog.jameskyle.org/2014/08/automated-stage3-gentoo-install-using-ansible/ Automating the bootstrapping of a node is reasonably complicted, even harder on Gentoo than on RHEL. This is the type of thinking I want to do, and I'm working on doing this with my CentOS box that runs ssh, Jenkins, postgres, and Redmine. Alec
Re: [gentoo-user] [Extremely OT] Ansible/Puppet replacement
On 27/01/2015 10:49, Tomas Mozes wrote: I haven't tested it yet, however I like the minimalistic syntax. As an ansible user - do you plan to allow using default values for modules and/or variables? +1 for that. I'm also a happy ansible user with zero plans to change, but I can't imagine a deployment tool without sane rational explicit defaults. A whole host of problems simply stop being problems if that feature is available. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] [Extremely OT] Ansible/Puppet replacement
On 01/27/2015 11:33 AM, Alan McKinnon wrote: On 27/01/2015 10:49, Tomas Mozes wrote: I haven't tested it yet, however I like the minimalistic syntax. As an ansible user - do you plan to allow using default values for modules and/or variables? +1 for that. I'm also a happy ansible user with zero plans to change, but I can't imagine a deployment tool without sane rational explicit defaults. A whole host of problems simply stop being problems if that feature is available. I'm curious, what exactly do you mean about default values? Is there a small example you can give me? The tutorial on Ansible's website is a little confusing. Alec
Re: [gentoo-user] grub - gummiboot: good
On Tuesday 27 January 2015 23:35:15 Neil Bothwick wrote: The cow is nothing but a machine which makes grass fit for us people to eat. ...while gassing us with methane :-) -- Rgds Peter.
Re: [gentoo-user] dependancy poppler and xorg-server
On 01/27/15 18:43, Joseph wrote: I'm getting two dependancy errors during upgrade: !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: x11-base/xorg-server:0 (x11-base/xorg-server-1.15.2-r1:0/1.15.2::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (x11-base/xorg-server-1.13.4-r1:0/1.13.4::gentoo, installed) pulled in by x11-base/xorg-server:0/1.13.4= required by (x11-drivers/xf86-input-mouse-1.9.0:0/0::gentoo, installed) ^^ (and 5 more with the same problem) [snip] 2.) Do I unmerge x11-base/xorg-server-1.13.4-r1? OK I've solved the problem with poppler by re-emerging app-text/texlive-core but I still get this error: x11-base/xorg-server:0 (x11-base/xorg-server-1.15.2-r1:0/1.15.2::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (x11-base/xorg-server-1.13.4-r1:0/1.13.4::gentoo, installed) pulled in by x11-base/xorg-server:0/1.13.4= required by (x11-drivers/xf86-input-mouse-1.9.0:0/0::gentoo, installed) ^^ (and 5 more with the same problem) -- Joseph