[gentoo-user] Re: Problem matching latest kernel with latest Nvidia
On 01/10/2017 06:31 PM, Corbin Bird wrote: You might want to add "static-libs" to the use flags as well. Another useful utility gets built by it. There doesn't seem to be anything installed by it, except a static lib (libXNVCtrl.a) and some header files (/usr/include/NVCtrl). No executables.
[gentoo-user] Re: Problem matching latest kernel with latest Nvidia
On 01/10/2017 11:01 AM, Philip Webb wrote: 170109 Nikos Chantziaras wrote: On 01/09/2017 10:09 AM, Philip Webb wrote: I had a previous thread re Kernel 4.9.0 + Nvidia 375.26 (now stable). I tried recompiling that kernel with DRM disabled, & remerged Nvidia 375.26 , but X won't start. The nvidia driver isn't "automatic". It's not used by X.Org by default. X.Org only uses its own drivers by default. For the nvidia driver, you need a conf file. Something like this in /etc/X11/xorg.conf.d/nvidia.conf: http://pastebin.com/raw/0y3NMndp I didn't need this with earlier versions of Nvidia-drivers, but I've copied your template & am willing to give it a try. It looks as if there are several lines which wouldn't fit my machine. The file is generic. The only thing specific to me is "G2770PF", which my monitor's model name. Which is not parsed anyway.
[gentoo-user] Re: nvidia driver missing symbols
On 01/09/2017 10:15 PM, Daniel Frey wrote: On 01/09/2017 12:10 PM, Nikos Chantziaras wrote: On 01/09/2017 10:09 AM, Philip Webb wrote: 170105 Nikos Chantziaras wrote: On 01/05/2017 08:05 AM, wabe wrote: Make sure that you have also enabled CONFIG_DRM. CONFIG_DRM=y CONFIG_DRM_KMS_HELPER=y CONFIG_DRM_KMS_FB_HELPER=y CONFIG_DRM_FBDEV_EMULATION=y For nvidia, these need to be all disabled. Then it should work. I had a previous thread re Kernel 4.9.0 + Nvidia 375.26 (now stable). I tried recompiling that kernel with DRM disabled, & remerged Nvidia 375.26 , but X won't start. The nvidia driver isn't "automatic". It's not used by X.Org by default. X.Org only uses its own drivers by default. For the nvidia driver, you need a conf file. Something like this in /etc/X11/xorg.conf.d/nvidia.conf: http://pastebin.com/raw/0y3NMndp The nvidia driver has a tool to help with this too, it's called `nvidia-xconfig`. Which creates a global config IIRC. It's better to use a xorg.conf.d file instead of a global one. Let X.Org do auto-configuration for everything else, and just use xorg.conf.d/nvidia.conf for the nvidia driver only.
[gentoo-user] Re: nvidia driver missing symbols
On 01/09/2017 10:09 AM, Philip Webb wrote: 170105 Nikos Chantziaras wrote: On 01/05/2017 08:05 AM, wabe wrote: Make sure that you have also enabled CONFIG_DRM. CONFIG_DRM=y CONFIG_DRM_KMS_HELPER=y CONFIG_DRM_KMS_FB_HELPER=y CONFIG_DRM_FBDEV_EMULATION=y For nvidia, these need to be all disabled. Then it should work. I had a previous thread re Kernel 4.9.0 + Nvidia 375.26 (now stable). I tried recompiling that kernel with DRM disabled, & remerged Nvidia 375.26 , but X won't start. The nvidia driver isn't "automatic". It's not used by X.Org by default. X.Org only uses its own drivers by default. For the nvidia driver, you need a conf file. Something like this in /etc/X11/xorg.conf.d/nvidia.conf: http://pastebin.com/raw/0y3NMndp
[gentoo-user] Re: nvidia driver missing symbols
On 01/08/2017 07:49 AM, Dale wrote: P. S. I've had emerge -e world fix issues in the past too. Weird how that works. :/ Gentoo packagers are not wizards. They cannot know which packages need a rebuild when a dependency changes. They know some of them, but not all. So over time, packages start misbehaving because some of their deps have been upgraded and the maintainer of the ebuild doesn't know about the need for a rebuild.
[gentoo-user] Re: nvidia driver missing symbols
On 01/07/2017 07:49 AM, Daniel Frey wrote: So I just recompiled DRM/KMS from the kernel, recompiled, redid the initramfs (just in case) and rebooted. The errors are also gone but I now have this: [ 31.918334] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 375.26 Thu Dec 8 18:04:14 PST 2016 [ 31.918704] nvidia-modeset: Allocated GPU:0 (GPU-14e248cf-aecd-cf7a-31f4-113e6d075ece) @ PCI::01:00.0 ...which I didn't have before. Yep. The nvidia KMS module conflicts with the in-kernel KMS implementation. It doesn't get loaded if in-kernel KMS is enabled, and then you get errors because of that. Btw, if you pay attention to the initial emerge messages when emerging nvidia-drivers, they actually tell you to disable DRM/KMS in the kernel ;-) The ebuild checks your current kernel config, and if it sees that stuff enabled, it warns you that you will most probably run into issues. Now that all that crap is sorted out, the only couple annoying things left are alt+tab switching in plasma, and the slowness of dolphin. Task switching is slow as f*** and it's irritating. I get that too, but only the first time I press alt+tab. After the task switch effect has been displayed once, it seems it gets cached and then it's fast. But overall, KDE (and KWin in particular) doesn't play well with the nvidia driver. I was able to fix most of my issues by following some advice from a KWin developer: * Enable triple buffering in xorg. nvidia-drivers requires a conf file anyway to work correctly. I have it in /etc/X11/xorg.conf.d/nvidia.conf, and the contents are: http://pastebin.com/raw/0y3NMndp This enables triple buffering and disables twinview. * Set some KWin environment variables. Instead of setting them globally, use a script named "kwin_x11" in a location that appears before /usr/bin in PATH. /usr/local/bin does that, so I have a /usr/local/bin/kwin_x11 file (it must be executable: chmod +x kwin_x11) with this in it: KWIN_TRIPLE_BUFFER=1 __GL_YIELD="USLEEP" exec /usr/bin/kwin_x11 $@ (/usr/local/bin *must* be before /usr/bin in your PATH variable, otherwise this doesn't work.) * Configure kwin to think it must use a higher refresh rate than your monitor's refresh rate. For 60Hz, your ~/.config/kwinrc must contain: [Compositing] MaxFPS=70 RefreshRate=70 (There's other stuff in the [Compositing] section, don't delete those.) * See if disabling vsync in the nvidia-settings control panel helps. After doing all that, KDE is quite usable for me. However, it's far from perfect. But if you don't want to switch from KDE to some other desktop environment, and can't deal with the performance issues of the nouveau driver, then you have not much choice here.
[gentoo-user] Re: nvidia driver missing symbols
On 01/06/2017 05:51 AM, Daniel Frey wrote: I think I fixed it by accident. I couldn't find the KMS helpers to enable for an external module in the kernel, so I built the Intel driver with KMS support and it dragged in the KMS helper bits (confirmed by grepping .config). Why can't we just turn that option on directly in menuconfig for external modules? This is in gentoo-sources. I've rebooted and the problem has gone away, go figure. I don't know if you missed my other post, but for nvidia, it's recommended to disable DRM and KMS.
[gentoo-user] Re: nvidia driver missing symbols
On 01/05/2017 07:28 PM, J. Roeleveld wrote: Actually. Nvidia complains about all framebuffers. I do use the EFI framebuffer as I want to be able to fall back to the text consoles when necessary. I have not encountered any major issues with this. Might occur if I switch between X and text regularly. Which I generally don't. It only complains when the graphics card was initialized using BIOS. This is true even in UEFI mode. It's recommended to disable the CSM (Compatibility Support Module) in the BIOS setup. Not all mainboards have that option though. But if you disable it, the graphics card will be initialized by UEFI on startup, which is fully supported by the driver (and the warning message will disappear.)
[gentoo-user] Re: nvidia driver missing symbols
On 01/05/2017 08:05 AM, wabe wrote: Make sure that you have also enabled CONFIG_DRM. CONFIG_DRM=y CONFIG_DRM_KMS_HELPER=y CONFIG_DRM_KMS_FB_HELPER=y CONFIG_DRM_FBDEV_EMULATION=y Eh, no. For nvidia, these need to be all disabled. Disable all KMS and DRM options. Then it should work.
[gentoo-user] Re: alsamixer and pulseaudio - which is at fault?
On 12/30/2016 03:44 PM, Mick wrote: On Friday 30 Dec 2016 12:12:17 Nikos Chantziaras wrote: On 12/30/2016 12:04 PM, Nikos Chantziaras wrote: On 12/29/2016 03:21 PM, Mick wrote: Hi All, My sound has been behaving erratically for a while now, probably since pulseaudio started being shipped with various desktop applications. I had many similar issues years ago. I solved them by doing the following: [...] Oh, forgot to mention that I also deleted the ALSA custom configuration files: rm ~/.asoundrc rm /etc/asound.conf It's recommended to not have them, unless you actually need them. Thank you Nikos, I followed your advice above but the darn thing is still not working as it should. The interaction of pa with alsa is anything but aligned with the way my brain works and with how alsa used to work. Another thing I did was put this in the USE flags in make.conf: pulseaudio -alsa and then do a: emerge -auDN --with-bdeps=y @world That is, globally disable ALSA and enable PA. This makes sure that programs will use PulseAudio instead of ALSA. This prevents the issue where PA changes your ALSA mixer settings as the same time as an ALSA program tries to do the same, resulting in things getting FUBARed. Unfortunately, portage is missing a feature where you can disable one USE flag only if another is available. So it's not possible to disable "alsa" only if the package also has a "pulseaudio" flag. That means if you have packages that only support ALSA, you need to enable the "alsa" USE flag for that package in package.use. This includes media-sound/pulseaudio itself! So package.use needs at least: media-sound/pulseaudio alsa There's other packages where this is needed. My list in package.use: media-sound/audacity alsa media-libs/libcanberra alsa dev-java/oracle-jdk-bin alsa media-sound/audacity alsa In general, if a package doesn't support PA and "alsa" is needed to make it support sound output, then it needs an entry in package.use.
[gentoo-user] Re: alsamixer and pulseaudio - which is at fault?
On 12/30/2016 12:04 PM, Nikos Chantziaras wrote: On 12/29/2016 03:21 PM, Mick wrote: Hi All, My sound has been behaving erratically for a while now, probably since pulseaudio started being shipped with various desktop applications. I had many similar issues years ago. I solved them by doing the following: [...] Oh, forgot to mention that I also deleted the ALSA custom configuration files: rm ~/.asoundrc rm /etc/asound.conf It's recommended to not have them, unless you actually need them.
[gentoo-user] Re: alsamixer and pulseaudio - which is at fault?
On 12/29/2016 03:21 PM, Mick wrote: Hi All, My sound has been behaving erratically for a while now, probably since pulseaudio started being shipped with various desktop applications. I had many similar issues years ago. I solved them by doing the following: In /etc/pulse/daemon.conf, I've set: flat-volumes = no In /etc/pulse/default.pa, comment-out these entries: #load-module module-device-restore #load-module module-stream-restore #load-module module-card-restore Add the "alsasound" service to "boot" runlevel: rc-config add alsasound boot In /etc/conf.d/alsasound, set: RESTORE_ON_START="yes" SAVE_ON_STOP="yes" Stop the alsasound service: /etc/init.d/alsasound stop Delete the currently saved mixer settings: rm /var/lib/alsa/asound.state Use alsamixer to configure your sound card to your liking (hit F6 first and select the real device.) Start the alsasound service: /etc/init.d/alsasound start Stop it again to save the current mixer settings: /etc/init.d/alsasound stop Change /etc/conf.d/alsasound to: SAVE_ON_STOP="no" You're done. Reboot to check if everything is working as intended. What the above does is make PA not restore its own settings on boot, make ALSA restore your preferred settings on boot but not save alterations on shutdown (next reboot will restore your initial settings.) It also disables the "flat volumes" feature of PA, which for me at least resulted in many "RIP my ears" moments, and also makes the volume mixer (KMix and pavucontrol in my case) behave very weirdly (it seems I need to get a PhD from MIT first to figure out what the volume settings do when "flat volumes" is on.) Note that the above is only done once. Do it, and reboot. After that it should work forever. When upgrading PA or ALSA, make sure to not overwrite your custom /etc/ settings when running "dispatch-conf" or "etc-update" (or whatever you're using.) Hope this helps.
[gentoo-user] Weird warning message when emerging gcc
A world update emerged gcc-5.4.0-r2 (update from 5.4.0). At the end of the build, I got this: * Python seems to be broken, attempting to locate CHOST ourselves ... * Switching native-compiler to x86_64-pc-linux-gnu-5.4.0 ...PORTAGE_BZIP2_COMMAND setting is invalid: 'bzip2' PORTAGE_BZIP2_COMMAND setting from make.globals is invalid: 'bzip2' I'm not seeing how python is broken here (works fine), and why PORTAGE_BZIP2_COMMAND is invalid. Can someone explain what's going on here?
[gentoo-user] Re: How to stop ls from quoting output
On 03/05/16 20:07, Daniel Quinn wrote: Some time ago after an update |ls| started returning output that looked like this: |8hOk25T.jpg 'Janeway Wallpaper-iPhone.png' 'Screenshot from 2016-04-06 16-15-15.png' microsoft.png 'Away mission Wallpaper-iPhone.png' 'Screenshot from 2016-03-18 14-29-06.png' 'Screenshot from 2016-04-07 11-29-02.png' gcal.png | Note that some of the files have a single quote (‘) surrounding them, and others don’t. [...] I see that I can just write an alias: |alias ls="ls --quoting-style=literal" | But I’d hate to do that if the default is “literal” and there’s some config installed somewhere that’s changing this. [...] As mentioned already, it's an upstream default. However, on Gentoo, "ls" is already an alias for "ls --color=auto", because the upstream default is to now use colors. At least for bash anyway (the alias is set in /etc/bash/bashrc). So I'd say just do: alias ls="ls --quoting-style=literal --color=auto" in your ~/.bashrc and forget about it :-) Or, if you want it system-wide, just write that alias in a new file: /etc/bash/bashrc.d/aliases
[gentoo-user] Re: Thunderbird 45.0 unreadable fonts
On 23/04/16 03:33, Holger Hoffstätte wrote: On Sat, 23 Apr 2016 02:38:00 +0300, Nikos Chantziaras wrote: I upgraded thunderbird-bin from 38.7.0 to 45.0 and all window fonts became extremely washed-out, thin and generally unreadable. I then wiped the configuration files (mv ~/.thunderbird ~/tb_backup), but got the same issue. It has nothing to do with the configuration of Thunderbird; the reason is simply that TB 45 was ported to use GTK3. Gtk3 apps look fine here. So that can't be it?
[gentoo-user] Thunderbird 45.0 unreadable fonts
I upgraded thunderbird-bin from 38.7.0 to 45.0 and all window fonts became extremely washed-out, thin and generally unreadable. I then wiped the configuration files (mv ~/.thunderbird ~/tb_backup), but got the same issue. I then wiped it again and emerged normal thunderbird (sans "-bin"), same issue. I downgraded to 38.7.0 again, as the issue makes thunderbird virtually unusable. Anyone else with the same problem? I searched a lot, but I didn't find any reports for this, let alone a fix.
[gentoo-user] Re: kde 5 and portable media mounting
On 17/04/16 05:26, Francisco Ares wrote: On KDE 4, it was ok, it mounted removable media in /run/[username]/[medialabel] , as the users were already in "plugdev" group and, just in case, the "disk" group as well - and they are there up to now. When the pop-up notification comes up that says a device has been plugged in, did you click on the checkmark on that notification? If you click the device itself, you will get "permission denied". If you click the checkmark, it work fine. (Don't ask me why that is. This is KDE. I stopped asking such silly questions a very long time ago.)
[gentoo-user] Re: kde 5 and portable media mounting
On 16/04/16 19:30, Francisco Ares wrote: Hi, All. After some issues (some not yet solved, but on the way...), there is one pretty annoying: regular users have to provide root password in order to mount a flash drive, for instance. Am I missing something? Are the users in the "plugdev" group? See: https://wiki.gentoo.org/wiki/Udisks#Groups
[gentoo-user] Re: KDE Plasma 5 moved to stable
On 15/04/16 16:06, Alan McKinnon wrote: On 15/04/2016 15:00, Nikos Chantziaras wrote: On 15/04/16 01:10, Frank Steinmetzger wrote: We even had the choice of installing both KDE3 and 4 at the same time. That made transition far easier and it allowed me to get settled in due time. So I'm not particularly worried about the next few months, especially with the modularity and interchangeability of today's KDE components. You cannot install KDE5 without uninstalling KDE4. I was told this is an upstream decision, not Gentoo's fault. Partially true. It's plasma 4 and 5 that cannot co-exist, not KDE Yeah, KDE. ;-) I mean common, when people say KDE, they mean the desktop, not kcalc or koffice or kmail.
[gentoo-user] Re: KDE Plasma 5 moved to stable
On 15/04/16 01:10, Frank Steinmetzger wrote: We even had the choice of installing both KDE3 and 4 at the same time. That made transition far easier and it allowed me to get settled in due time. So I'm not particularly worried about the next few months, especially with the modularity and interchangeability of today's KDE components. You cannot install KDE5 without uninstalling KDE4. I was told this is an upstream decision, not Gentoo's fault.
[gentoo-user] Re: KDE Plasma 5 moved to stable
On 11/04/16 13:34, Peter Humphrey wrote: I saw that too, and it cast a gloom over the rest of my day. I intend to stick with KDE-4 until some more sensible themes become available in Plasma 5 The KDE 4 themes are still included. (Oxygen and Air.) Breeze is the default though.
[gentoo-user] Re: to anyone and everyone
On 09/04/16 07:54, Alan Grimes wrote: Decided the corsair psu was not worth rma-ing, at similar expense so decided to trash the company instead. =| You get what you pay for ;-) http://www.guru3d.com/news-story/corsair-extends-select-psu-warranties-from-7-years-to-10-years.html
[gentoo-user] Re: Blacklisting all packages from overlay except a specific group and version
On 31/03/16 09:11, Alan McKinnon wrote: On 31/03/2016 00:13, Nikos Chantziaras wrote: "Invalid atom in /etc/portage/package.unmask/qt: =dev-qt/*-5.6*::qt" (Doesn't work with or without the "::qt".) It's the leading "*" that's wrong there, because it's not a glob or a regex; it's more a placeholder metacharacter with it's own rules. Look at the relevant section in man 5 ebuild. It's not explicit but it does strongly hint that "*" goes at the end of a string It didn't look like it to me at all. From the man page: Examples: # match anything with a version containing , which can be used in # package.mask to prevent emerge --autounmask from selecting live # ebuilds =*/*-** So if that works, one would assume that "=dev-qt/*-*5.6*" should work just as well :-/
[gentoo-user] Re: Blacklisting all packages from overlay except a specific group and version
On 30/03/16 05:48, Bryan Gardiner wrote: On 30 March 2016 05:01:16 GMT+09:00, Nikos Chantziaras <rea...@gmail.com> wrote: So, I want to install packages from an overlay, but only from a specific group. In this case, the dev-qt/ group. The overlay name is "qt". It doesn't work: package.mask: */*::qt package.unmask: dev-qt/*-5.6*::qt With this, 5.6 versions are still masked. The syntax for "5.6*"-like versions has an equals at the start, for example =dev-qt/qtbase-5.6* That works. so if it's allowed, =dev-qt/*-5.6*::qt would be the correct package.unmask syntax. Does that work? Nope: "Invalid atom in /etc/portage/package.unmask/qt: =dev-qt/*-5.6*::qt" (Doesn't work with or without the "::qt".)
[gentoo-user] Re: Blacklisting all packages from overlay except a specific group and version
Because that is already covered by: */*::qt in package.mask. This masks everything from the qt overlay, including versions of Qt higher or equal to 5.7. I only masked >=5.7, then the other packages from the qt overlay would not be masked. What I need is mask ALL packages in the qt overlay EXCEPT dev-qt/*-5.6.*. On 30/03/16 00:14, Ian Bloss wrote: Why don't you mask versions higher than 5.7 instead? On Tue, Mar 29, 2016, 17:00 Nikos Chantziaras <rea...@gmail.com <mailto:rea...@gmail.com>> wrote: On 29/03/16 23:01, Nikos Chantziaras wrote: > So, I want to install packages from an overlay, but only from a specific > group. In this case, the dev-qt/ group. The overlay name is "qt". > > It doesn't work: > >package.mask: */*::qt >package.unmask: dev-qt/*-5.6*::qt > > With this, 5.6 versions are still masked. > > How do I do this? Well, I brute-forced it. I've use this mask: */*::qt and this unmask:
[gentoo-user] Re: Blacklisting all packages from overlay except a specific group and version
On 29/03/16 23:01, Nikos Chantziaras wrote: So, I want to install packages from an overlay, but only from a specific group. In this case, the dev-qt/ group. The overlay name is "qt". It doesn't work: package.mask: */*::qt package.unmask: dev-qt/*-5.6*::qt With this, 5.6 versions are still masked. How do I do this? Well, I brute-forced it. I've use this mask: */*::qt and this unmask: It works, but I would have expected that computers are good with wildcards :-/ The above should be equivalent to this:
[gentoo-user] Re: Blacklisting all packages from overlay except a specific group and version
On 29/03/16 23:21, R0b0t1 wrote: Not entirely sure what you wish. Maybe package sets? Not sure what *you* mean :-) How would package sets help me in masking all packages from an overlay and then whilelist only specific ones?
[gentoo-user] Blacklisting all packages from overlay except a specific group and version
So, I want to install packages from an overlay, but only from a specific group. In this case, the dev-qt/ group. The overlay name is "qt". It doesn't work: package.mask: */*::qt package.unmask: dev-qt/*-5.6*::qt With this, 5.6 versions are still masked. How do I do this?
[gentoo-user] Re: Keyboard is dead after emerge-update
On 21/03/16 13:52, Bertram Scharpf wrote: Hi, since an emerge-update-world on my notebook the keyboard does no longer respond in X. This is extremely annoying because when I have xdm in rc-update, X is started right at boot. I have no chance to get back to the console using Ctrl-Alt-F1, and the device in unusable. After xorg-server updates, it's recommended to run: emerge @x11-module-rebuild This will automatically rebuild xorg-modules.
[gentoo-user] Re: The war continues.
On 18/03/16 23:57, Alan McKinnon wrote: KDE 5 is absolutely nothing like KDE3. So by all means try it, but evaluate it on it's own terms. It's not a better KDE3, it's a whole different DE And full of bugs :-P Holy crap is it full of bugs. Like, seriously.
[gentoo-user] Re: The war continues.
On 17/03/16 18:00, Philip Webb wrote: I tried to fix it by setting wayland, gles2 and egl to default because they were breaking other packages. Why are you using Wayland ? -- it's still largely experimental, isn't it ? Some packages force the wayland USE flag on. Emerge that breaks if you try to disable wayland. (KDE 5 packages do this.)
[gentoo-user] Re: KDE's confusing versions
On 06/03/16 22:37, Philip Webb wrote: Eix tells me : root:505 ~> eix okular [I] kde-apps/okular Available versions: (4) 4.14.3(4/4.14)^t 15.08.3-r1(4/15.08)^t ~15.12.1(4/15.12)^t {aqua chm crypt debug djvu dpi ebook +handbook +jpeg kde mobi +pdf +postscript +tiff} Installed versions: 15.08.3-r1(4)^t([2016-03-05 18:13:44]) (crypt handbook jpeg pdf tiff -aqua -chm -debug -djvu -dpi -ebook -kde -mobi -postscript) So Okular 15.08.3-r1 is part of version (4), but Konsole 15.08.3-r1 is part of version (5). That's wrong. Okular 15.x is part of KDE Applications 15.08. I don't know why it's in the "4" slot. But it's wrong. Okular is NOT a KDE 4 app.
[gentoo-user] Re: [OT] Choice of graphics card
On 03/03/16 12:15, Peter Humphrey wrote: On Wednesday 02 March 2016 21:44:29 Nikos Chantziaras wrote: If you need 64-bit FP compute, AMD Radeon is your only choice. NVidia only supports 32-bit FP and cripples 64-bit FP on consumer cards, rendering them virtually useless; you'd have to buy a "professional" card (thousands of dollars). So you might want to check whether your applications need 64-bit FP or 32-bit. So for flexibility in the future it looks like AMD Firepro or similar. (I assume your comments on Radeon will also apply to Firepro.) Be careful not to buy a firepro that is as expensive (or more expensive) than a radeon card, but with lower performance. You need to make sure what the compute speed is compared to other card before you buy.
[gentoo-user] Re: [OT] Choice of graphics card
On 02/03/16 13:50, Peter Humphrey wrote: Hello list, It's approaching time to invest in a new system, and I'd like the panel's views on which make of graphics card to choose: AMD, nVidia or Radeon. I don't want to start a flame war, but which of those would give me the best performance in GPU calculations? I have some BOINC projects in mind, which apparently are greatly accelerated by using GPUs as well as CPUs. The nouveau driver for nVidia can't do CUDA (GPU) calculations so I'd have to use the nVidia driver; I don't know about the others which is why I'm asking. Does anyone here have an opinion to offer? If you need 64-bit FP compute, AMD Radeon is your only choice. NVidia only supports 32-bit FP and cripples 64-bit FP on consumer cards, rendering them virtually useless; you'd have to buy a "professional" card (thousands of dollars). So you might want to check whether your applications need 64-bit FP or 32-bit.
[gentoo-user] Re: openssl upgrade may miss some needed rebuilds
On 02/03/16 16:41, walt wrote: Today's upgrade of openssl to 1.0.2g-r1 may cause some necessary rebuilds to fail due to missing symbol errors. Example: libcurl was broken and caused the rebuilds of virtualbox and git to fail until I forced a rebuild of curl. Any installed package that is actually linked against openssl will be affected by this, notably curl or wget, which may prevent portage from fetching source files. Does that mean that the library name is the same and the "preserve-libs" FEATURE doesn't kick in in this case?
[gentoo-user] Re:
On 28/02/16 21:18, Dale wrote: Donahue Trevor wrote: What was that again? I can't hear you. ROFLMBO Just HTML-only spam :-)
[gentoo-user] Re: PYTHON_SINGLE_TARGET & libreoffice
On 27/02/16 22:04, Alan McKinnon wrote: There are still way too many packages requiring python-2.7 that don't support python-3 at all. Hm. I only have one such package: app-text/asciidoc. That's out of 1227 total packages installed.
[gentoo-user] Re: PYTHON_SINGLE_TARGET & libreoffice
On 27/02/16 21:40, Alan McKinnon wrote: I don't know much about PYTHON_SINGLE_TARGET and PYTHON_TARGET, it all seemed to just work so I never looked further. Until now. Of all packages, libreoffice seems to want only python3: I can set it in make.conf but as expected that causes all manner of other packages to fail emerge checks as they need python2.7. What's the preferred way to deal with this situation? PYTHON_TARGETS="python2_7 python3_4" PYTHON_SINGLE_TARGET="python3_4" For packages that require another PYTHON_SINGLE_TARGET version, you can set "python_single_target_python2_7" in package.use. For example: app-text/asciidoc python_single_target_python2_7 The reverse is also true. You can set: PYTHON_SINGLE_TARGET="python2_7" in make.conf, and then set "python_single_target_python3_4" in package.use for packages that require 3.4: app-office/libreoffice python_single_target_python3_4 I use the first method. Yes, it's a PITA :-) PS: Another syntax is also available for package.use: app-text/asciidoc PYTHON_SINGLE_TARGET: python2_7 Just saw a blog post about it; I think it's undocumented and works for all USE_EXPAND flags. The downside is that it has to appear at the end of the line.
[gentoo-user] Re: using package.provided
On 26/02/16 18:47, Harry Putnam wrote: First off, thanks to all posters for the excellent input Nikos Chantziaras <rea...@gmail.com> writes: On 25/02/16 05:55, Harry Putnam wrote: I'd like to stay on kernel-4.1.6, rather than keep installing the newest version at each upgrade. I'd instead recommend putting >=sys-kernel/gentoo-sources-4.2 in package.mask to stay with 4.1 (it's an LTS kernel) and always get the patches for that one. In /etc/portage/package-mask, using the line you suggest: >=sys-kernel/gentoo-sources-4.2 Makes emerge want to install 4.1.18: Yes, which is what I recommended. Don't block 4.1.x security/bugfix patches. Just block 4.2 and above.
[gentoo-user] Re: using package.provided
On 25/02/16 05:55, Harry Putnam wrote: I'd like to stay on kernel-4.1.6, rather than keep installing the newest version at each upgrade. I'd instead recommend putting >=sys-kernel/gentoo-sources-4.2 in package.mask to stay with 4.1 (it's an LTS kernel) and always get the patches for that one.
[gentoo-user] Re: [OT] thunderbird stop working
On 22/02/16 13:02, p...@xvalheru.org wrote: Hi, Sorry for such stupid question, but I'm lost :-( I'm using thunderbird and recently I'm not able to login to mail server. Server configuration doesn't changed. I've changed password only, but typing new password doesn't log me in :-( What makes me worry is the text in login dialogue: "Enter your password for @@:". Could someone help? Inspect the account details (Edit->Account Settings). Make sure everything is set up correctly there.
[gentoo-user] Don't know what portage wants here
Can someone decipher this one? As far as portage errors go, this is the most nonsensical one I've seen in a long time (before you ask, boost and boost-build are not masked anywhere): === emerge -auDN --with-bdeps=y @world [...] !!! The following update has been skipped due to unsatisfied dependencies: dev-libs/boost:0 !!! All ebuilds that could satisfy "=dev-util/boost-build-1.58*" have been masked. !!! One of the following masked packages is required to complete your request: - dev-util/boost-build-1.58.0::gentoo (masked by: ) (dependency required by "dev-libs/boost-1.58.0-r1::gentoo" [ebuild]) For more information, see the MASKED PACKAGES section in the emerge man page or refer to the Gentoo Handbook. === I've also tried --backtrack=20 (then 30, the 40, then 50), with the same result.
[gentoo-user] Re: ls config file?
On 02/02/16 10:54, Raffaele BELARDI wrote: After a recent update of coreutils to version 8.25, 'ls -l' started displaying names containing spaces enclosed in single quotes, e.g.: drwxr-xr-x 6 belardi users 4096 May 21 2012 'Audio Libraries' drwxr-xr-x 2 belardi users 4096 Jun 10 2014 Brochure The option that controls this is --quoting-style, so --quoting-style=literal returns to the old behaviour (which I prefer). I can alias 'ls' to include this option but was wondering if there is a global configuration file controlling such behaviour. This is done with aliases. Actually, the "ls" command should by default be an alias. If you just enter: $ alias you are shown current aliases. "ls" should actually be defined as: alias ls='ls --color=auto' If you actually enter "/bin/ls", you'll see that by default ls doesn't even show colors. So Gentoo's default install provides an alias for that. You can provide your own alias in /etc/bash/bashrc, which is sourced by all interactive shells. All you need to do is provide your own alias. You can do that in your ~/.bashrc file by adding this: alias ls="ls --color=auto --quoting-style=literal" Logout and in again and you're done.
[gentoo-user] Re: What happened to portage?
On 31/01/16 10:22, Pavel Volkov wrote: On четверг, 28 января 2016 г. 5:17:09 MSK, Nikos Chantziaras wrote: Syncing resulted in thousands of lines of: app-admin/python-updater/.~tmp~/ And the portage tree is full of ".~tmp~" directories. What's going on? Did you try a different mirror? I think there might be an issue with specific host. If so, contact its maintainer by email diplayed by rsync or file a bug to Gentoo infra... The problem was apparently fixed after I synced the next day.
[gentoo-user] What happened to portage?
Syncing resulted in thousands of lines of: app-admin/python-updater/.~tmp~/ app-admin/qpage/.~tmp~/ app-admin/qtpass/.~tmp~/ app-admin/quickswitch/.~tmp~/ app-admin/r10k/.~tmp~/ app-admin/radmind/.~tmp~/ app-admin/ranpwd/.~tmp~/ app-admin/recursos/.~tmp~/ app-admin/reportmagic/.~tmp~/ app-admin/restart_services/.~tmp~/ app-admin/rex/.~tmp~/ rsync: opendir "/app-arch/libarchive/.~tmp~" (in gentoo-portage) failed: Permission denied (13) rsync: opendir "/app-arch/libarchive/files/.~tmp~" (in gentoo-portage) failed: Permission denied (13) rsync: opendir "/app-arch/libpar2/.~tmp~" (in gentoo-portage) failed: Permission denied (13) rsync: opendir "/app-arch/libzpaq/.~tmp~" (in gentoo-portage) failed: Permission denied (13) And the portage tree is full of ".~tmp~" directories. What's going on?
[gentoo-user] What is forcing the qt4 USE flag?
I have a weird problem. I have an ebuild where either qt4 or qt5 can be enabled. They are both disabled by default and I have to choose which one I want. The ebuild does that with: IUSE="qt4 qt5" REQUIRED_USE="^^ ( qt4 qt5 )" I'm on the plasma profile which enabled qt5 automatically. However, portage complains: The following REQUIRED_USE flag constraints are unsatisfied: exactly-one-of ( qt4 qt5 ) The qt4 USE flag is enabled. I can't see where. I didn't enable it anywhere in any of my /etc/portage/* files. If I emerge with: USE="-qt4" emerge package then it works. So I have to explicitly disable the qt4 USE flag even though I didn't enable it in the first place. Can someone enlighten me?
[gentoo-user] Re: What is forcing the qt4 USE flag?
On 23/01/16 09:21, Michael Palimaka wrote: On 01/23/2016 06:04 AM, Nikos Chantziaras wrote: I have a weird problem. I have an ebuild where either qt4 or qt5 can be enabled. They are both disabled by default and I have to choose which one I want. The ebuild does that with: IUSE="qt4 qt5" REQUIRED_USE="^^ ( qt4 qt5 )" I'm on the plasma profile which enabled qt5 automatically. However, portage complains: The following REQUIRED_USE flag constraints are unsatisfied: exactly-one-of ( qt4 qt5 ) The qt4 USE flag is enabled. I can't see where. I didn't enable it anywhere in any of my /etc/portage/* files. If I emerge with: USE="-qt4" emerge package then it works. So I have to explicitly disable the qt4 USE flag even though I didn't enable it in the first place. Can someone enlighten me? Which package is it? We can add an entry to the Plasma profile's package.use to avoid the REQUIRED_USE. It's my own ebuilds (in an overlay.) I wasn't setting either USE flag in the ebuilds at all and this spooked me. Turns out it's the parent profile of the plasma profile doing this. What would be really nice to have in portage is if the profiles were able to say "if qt4 and qt5 are both supported but mutually exclusive, prefer qt5 unless the user or the ebuild itself has specified otherwise". But I guess it would be hairy to implement that. (It would help with other such USE flags as well, not just qt4 vs qt5.)
[gentoo-user] Re: What is forcing the qt4 USE flag?
On 22/01/16 21:45, Mick wrote: On Friday 22 Jan 2016 21:04:48 Nikos Chantziaras wrote: I have a weird problem. I have an ebuild where either qt4 or qt5 can be enabled. They are both disabled by default and I have to choose which one I want. The ebuild does that with: IUSE="qt4 qt5" REQUIRED_USE="^^ ( qt4 qt5 )" I'm on the plasma profile which enabled qt5 automatically. However, portage complains: The following REQUIRED_USE flag constraints are unsatisfied: exactly-one-of ( qt4 qt5 ) The qt4 USE flag is enabled. I can't see where. I didn't enable it anywhere in any of my /etc/portage/* files. If I emerge with: USE="-qt4" emerge package then it works. So I have to explicitly disable the qt4 USE flag even though I didn't enable it in the first place. Can someone enlighten me? euse -I qt4 should provide some pointers. It may be your desktop profile? Yep, it's the parent profile of the plasma profile. This results in both qt4 and qt5 being set.
[gentoo-user] Re: How to get rid of 32bits libraries
On 18/01/16 15:31, Helmut Jarausch wrote: If possible, I'd like to install any packages with abs_x86_32 which are required for dev-util/android-sdk-update-manager and app-text/acroread with a different PREFIX, such as /usr/local . Is that possible? There is exactly zero reason to do that, so no, there's no option for that. It might be possible by modifying the ebuild, but you're on your own. No one tried this, since it's not needed. Distro-maintained packages always go in /usr, user-maintained packages go in /usr/local. There is no reason to break that convention.
[gentoo-user] Re: {OT} Allow work from home?
On 16/01/16 06:17, Grant wrote: I'm considering allowing some employees to work from home but I'm concerned about the security implications. Currently everybody shows up and logs into their locked down Gentoo system and from there is able to access the company webapps which are restricted to the office IP address. I guess I would have to allow webapp access from any IP for those users and trust that their computer is secure? Should that not be scary? I've set up such systems using OpenVPN, as others have suggested. One thing to look out for, is to make sure that the setup only tunnels traffic to your servers, not ALL traffic. Otherwise, all traffic from your people is going to be tunneled through your network (Netflix, torrents, porn, everything else your people are doing at home.)
[gentoo-user] Re: How to get rid of 32bits libraries
On 18/01/16 12:00, Dale wrote: Neil Bothwick wrote: On Mon, 18 Jan 2016 09:12:23 +, Neil Bothwick wrote: equery hasuse checks which packages respect the given USE flag, it pays no attention to whether it is actually set. Try emerge -evp world | grep 'ABI_X86=32' Sorry, that should be emerge -evp world | grep 'ABI_X86="32' Yep. That one works better. How does it work "better" if all it does is expand to an "abi_x86_32" USE flag? :-/
[gentoo-user] Re: How to get rid of 32bits libraries
On 17/01/16 10:32, Helmut Jarausch wrote: Hi, I'd like to get rid of all 32bits libraries. There are only two packages which I'd like to keep and which need some 32bits libraries. That's dev-util/android-sdk-update-manager and app-text/acroread both of which I use only occasionally. Look in your make.conf and see if ABI_X86 is defined there. If yes, delete it. Then, in your package.use, at the very top, the first line should be: */* -abi_x86_32 This disables the "abi_x86_32" USE flag on all packages that have it. You can then add the "abi_x86_32" USE flag to individual packages on a per-needed basis. If your package.use is a directory, make sure to put that line on a file that's looked up first (for example "000-abi", the zeroes in front will make sure this file is parsed first). Then: emerge -auDN --with-bdeps=y @world should take care of doing the needed rebuilds. Is it possible to install this on /usr/local in such a way that it doesn't interfere with my standard Gentoo installation on /usr? Not sure what you mean with "interfere." Why would it interfere with anything? If there's file collisions, portage will point them out to you.
[gentoo-user] Re: Adobe flash warning and tree
On 16/01/16 14:54, Mick wrote: On Saturday 16 Jan 2016 12:49:30 you wrote: On Saturday 16 Jan 2016 04:15:33 Dale wrote: Neil Bothwick wrote: It's better to tell them you're using the Windows version of Firefox or Chrome. If you send an IE User_agent, some sites will start messing with ActiveX etc. That is true but in the cases I used that, it required not only M$ but also IE. Having it set to Firefox or something would be safer as you point out, if IE is not also required. Well I tried channel5 website with a changed FF useragent string, by adding a general.useragent.override key in about:config Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101 Firefox/38.0 and it still failed. Then tried a MSIE useragent string: Mozilla/5.0 (Windows NT 6.1; WOW64; Trident/7.0; rv:11.0) like Gecko and continue to get the same error: "To view this page ensure that Adobe Flash Player version 15.0.0 or greater is installed." :-( As already pointed out, you need Google Chrome which comes with recent Flash. You can override the User-agent there too and use a Windows Chrome string.
[gentoo-user] Re: The X11-trap: Once back on textconsole...and no one comes back ever...
On 15/01/16 18:56, meino.cra...@gmx.de wrote: Heiko Baums[16-01-15 17:24]: Am 15.01.2016 um 13:27 schrieb meino.cra...@gmx.de: This has worked for years and didn't changed anything from which I think it could be related. How can I fix this? https://bugs.gentoo.org/show_bug.cgi?id=564096 Hi Heiko, ...seems to be the same problem. How much do I have to pay (performancewise) for the downgrade, since my rendering graphics card (see previous mail for hardware setup) is a GTX 960 with maxwell architecture (quite new I think)... ? I use this card exclusively for Blender... Why can't you just use it for the desktop too? Is there an issue?
[gentoo-user] Re: The X11-trap: Once back on textconsole...and no one comes back ever...
On 15/01/16 21:50, meino.cra...@gmx.de wrote: regardless how powerful a payable graphics card may be... if one uses the same graphicsscards for rendering as for the desktop the desktop starts heavily lagging/freezing just in the moment blenders rendering engine Cycles start working. At least, these are my experiences. Thats why I use an old just-fast-enough graphics card for the desktop and a second, more current one for rendering only. Yeah, makes sense. Which drivers are you using? NVidia's changelogs almost always have a "fixed console switching issues" entry on every driver release. It seems they're still haven't figured this out completely, so the only suggestion I can give is to try more recent drivers. Another thing is to test whether a text-only console vs a framebuffer one makes any difference. Officially, nvidia only supports text mode consoles. At least on my system, the nvidia driver prints this on dmesg with my framebuffer console setup: NVRM: Your system is not currently configured to drive a VGA console NVRM: on the primary VGA device. The NVIDIA Linux graphics driver NVRM: requires the use of a text-mode VGA console. Use of other console NVRM: drivers including, but not limited to, vesafb, may result in NVRM: corruption and stability problems, and is not supported. Although I have no choice in the matter really, since text mode is not supported on my system. Fortunately it's been working without issues for me. But text mode might solve your issue (if your system supports that.) Or it might not :-P
[gentoo-user] Re: Adobe flash warning and tree
On 15/01/16 10:15, Peter Weilbacher wrote: On Thu, 14 Jan 2016, Nikos Chantziaras wrote: I thought these are only security fixes? Version 20 is actully feature-complete and in-par with the Windows version 20. v11.2.202.559 might be from December, but it's based on an ancient version from many years ago feature-wise. That may well be the case. But who cares about features in a tool that's basically dead and only installed on 99.9% of the machines because some stupid websites[1] need it? Peter. [1] That work with the feature set of v11.2 -- at least I haven't heard otherwise. You haven't visited enough porn sites then :-P
[gentoo-user] Re: Adobe flash warning and tree
On 14/01/16 21:30, Peter Weilbacher wrote: On Wed, 13 Jan 2016, Nikos Chantziaras wrote: On 12/01/16 23:10, Róbert Čerňanský wrote: https://get.adobe.com/flashplayer/ page: "Adobe Flash Player 11.2 will be the last version to target Linux as a supported platform. Adobe will continue to provide security backports to Flash Player 11.2 for Linux." That's ancient. The version that comes with Chrome is v20.0. No, Google just likes to increase their version numbers in big steps so that you think that they make quick progress. v11.2.202.559 on Linux is from 28th of December, see <https://helpx.adobe.com/security/products/flash-player/apsb16-01.html>. I thought these are only security fixes? Version 20 is actully feature-complete and in-par with the Windows version 20. v11.2.202.559 might be from December, but it's based on an ancient version from many years ago feature-wise.
[gentoo-user] Re: OT: No audio while playing HTML5-video (YouTube)
On 13/01/16 08:19, Mick wrote: On Wednesday 13 Jan 2016 08:00:11 Nikos Chantziaras wrote: On 13/01/16 07:53, Nikos Chantziaras wrote: default should be ~/.mpv/config Or "~/.config/mpv/config". I had the former file in which I added your suggested stanza, but youtube videos dragged from FF continue to refuse to play (mpv crashes). This is what is saved in the logs: Error Information: An exit code of 2 was returned from mpv --profile=pseudo-gui -- . Output Data: There was no output Error Logs: There was no error message Well, no idea :-P All I can say is that I have these: media-video/mpv-0.14.0-r1 lua and luajit USE flags are needed. media-video/ffmpeg-2.8.4 librtmp network openssl opus vorbis are needed net-misc/youtube-dl-2016.01.09 I would assume that if you're not on ~arch, you have too old versions. Newer versions are very important because YouTube changes their APIs and streaming formats quite often and thus stuff gets outdated quick.
[gentoo-user] Re: OT: No audio while playing HTML5-video (YouTube)
On 14/01/16 01:15, Mick wrote: I am on the stable tree, but have net-misc/youtube-dl-2016.01.0 at the moment. Our USE flags are the same. I will emerge the latest version of net- misc/youtube-dl tomorrow to see if this fixes it, but I somehow doubt it. I wonder if this problem is related to the desktop manager - I am using enlightenment-0.20.2 at present. I assume it's because of your mpv and/or ffmpeg versions not being able to stream from YouTube.
[gentoo-user] Re: OT: No audio while playing HTML5-video (YouTube)
On 13/01/16 07:53, Nikos Chantziaras wrote: default should be ~/.mpv/config Or "~/.config/mpv/config".
[gentoo-user] Re: OT: No audio while playing HTML5-video (YouTube)
On 13/01/16 00:40, Mick wrote: On Tuesday 12 Jan 2016 19:56:18 Nikos Chantziaras wrote: On 12/01/16 05:08, meino.cra...@gmx.de wrote: Stroller <strol...@stellar.eclipse.co.uk> [16-01-12 04:00]: On Mon, 11 January 2016, at 6:15 p.m., meino.cra...@gmx.de wrote: Oh, how I like music videos without sound ;) I trust you're aware you can use net-misc/youtube-dl until you get this sorted. It's not clear to me that _any_ HTML5 audio is playing in your browser. What version of Firefox do you have, what are it's USE flags? Stroller. ...yes, sour are right :) Until now no a single HTML5-video is playing its sound... A better way: Open mpv from your "start menu". Drag the youtube video into the mpv window. No need to mess with terminals :-) Hmm ... this won't work here. :( An exit code of 2 was returned. You need to have the youtube-dl script installed (net-misc/youtube-dl). Make sure you use the ~arch version, since YouTube can change their APIs very fast, and that script has almost daily updates. But it's safe using ~arch for this package, since it's just a python script. To make sure you get the highest quality youtube format (including getting 60FPS video when available), make sure this line: ytdl-format=bestvideo[ext=webm]+bestaudio/bestvideo+bestaudio/best Is in your mpv config file (default should be ~/.mpv/config). If you have a 1080p monitor, make sure you're excluding 4K. In that case, the above line should be: ytdl-format=bestvideo[width<=?1920][ext=webm]+bestaudio/bestvideo[width<=?1920][ext=mp4]+bestaudio/best In both cases, VP9 is given higher priority and H.264 is only used when no VP9 version of the video exists. (YouTube provides much higher quality video in the VP9 format.)
[gentoo-user] Re: OT: No audio while playing HTML5-video (YouTube)
On 12/01/16 05:08, meino.cra...@gmx.de wrote: Stroller[16-01-12 04:00]: On Mon, 11 January 2016, at 6:15 p.m., meino.cra...@gmx.de wrote: Oh, how I like music videos without sound ;) I trust you're aware you can use net-misc/youtube-dl until you get this sorted. It's not clear to me that _any_ HTML5 audio is playing in your browser. What version of Firefox do you have, what are it's USE flags? Stroller. ...yes, sour are right :) Until now no a single HTML5-video is playing its sound... A better way: Open mpv from your "start menu". Drag the youtube video into the mpv window. No need to mess with terminals :-)
[gentoo-user] Re: Adobe flash warning and tree
On 10/01/16 07:27, Dale wrote: Howdy, I keep getting a warning that Flash needs to be upgraded. I went to packages.g.o and there doesn't seem to be a newer version than what I have. What gives? Adobe does not update Flash for Linux themselves anymore. They gave that to Google. As a side effect, the only way to get the latest Flash version on Linux, is to use Google Chrome.
[gentoo-user] Re: Adobe flash warning and tree
On 12/01/16 23:10, Róbert Čerňanský wrote: On Tue, 12 Jan 2016 19:53:41 +0200 Nikos Chantziaras <rea...@gmail.com> wrote: Adobe does not update Flash for Linux themselves anymore. They gave that to Google. As a side effect, the only way to get the latest Flash version on Linux, is to use Google Chrome. Adobe still provides security fixes for version 11.2 though. From https://get.adobe.com/flashplayer/ page: "Adobe Flash Player 11.2 will be the last version to target Linux as a supported platform. Adobe will continue to provide security backports to Flash Player 11.2 for Linux." That's ancient. The version that comes with Chrome is v20.0.
[gentoo-user] Why does portage want to downgrade Grub?
emerge -auDN @world ... [ebuild UD ] sys-boot/grub-2.00_p5107-r2 [2.02_beta2-r7] USE="-custom-cflags%" GRUB_PLATFORMS="-yeeloong%" I currently have 2.02_beta2-r7 installed. It's not masked or anything. But portage wants to downgrade to 2.00_p5107-r2. I don't have any clue why that is.
[gentoo-user] Re: Why does portage want to downgrade Grub?
On 12/12/15 18:02, Mike Gilbert wrote: On Sat, Dec 12, 2015 at 10:39 AM, Nikos Chantziaras <rea...@gmail.com> wrote: emerge -auDN @world ... [ebuild UD ] sys-boot/grub-2.00_p5107-r2 [2.02_beta2-r7] USE="-custom-cflags%" GRUB_PLATFORMS="-yeeloong%" I currently have 2.02_beta2-r7 installed. It's not masked or anything. But portage wants to downgrade to 2.00_p5107-r2. I don't have any clue why that is. Possibly a bad Manifest entry. rsync has been pretty broken this week. Hm, indeed. grub-0.97-r16.ebuild was missing from the manifest, but it's weird that this would affect 2.02_beta2-r7. It shouldn't, right?
[gentoo-user] Re: Why does portage want to downgrade Grub?
On 13/12/15 00:56, Neil Bothwick wrote: Hm, indeed. grub-0.97-r16.ebuild was missing from the manifest, but it's weird that this would affect 2.02_beta2-r7. It shouldn't, right? If the manifest fails, all ebuilds in that directory are considered suspect. Is the version portage wants to downgrade to in an overlay? Nope. Portage-tree only.
[gentoo-user] Re: How is /etc/portage/env supposed to work?
On 20/10/15 17:26, Michael Orlitzky wrote: The whole thing is confusing. If you ever get it working please put a decent example on the wiki because I remember spending hours trying to do the same thing. OK, I just created an account, but it doesn't look that the wiki is actually editable: You do not have permission to edit this page, for the following reason: You do not have permission to edit pages in the Handbook namespace.
[gentoo-user] How is /etc/portage/env supposed to work?
I'm following the documentation here: https://wiki.gentoo.org/wiki/Handbook:AMD64/Portage/Advanced#Using_.2Fetc.2Fportage.2Fenv to hook into the postinst phase of an ebuild. The docs only provide an actual example for the global hook case (/etc/portage/bashrc), but it says it works the same way on a per-ebuild basis using /etc/portage/env. But it is totally silent on how to actually do that :-/ So I assume, since the docs don't mention anything special, that I should put the code in a conf file. So I created this file: /etc/portage/env/test.conf and in it: if [ "${EBUILD_PHASE}" == "postinst" ]; then echo ":: TEST" fi Then I added this line to /etc/portage/package.env: app-misc/mc test.conf But when I "emerge app-misc/mc", I get: !!! Problem in 'app-misc/mc' dependencies. !!! "/etc/portage/env/test.conf", line 1: Invalid token '[' (not '=') portage.exception ... done! "/etc/portage/env/test.conf", line 1: Invalid token '[' (not '=') I don't get it... Can't I have a per-package bashrc or something? The docs are suggesting that I can, but don't tell me how.
[gentoo-user] Re: How is /etc/portage/env supposed to work?
On 20/10/15 17:26, Michael Orlitzky wrote: On 10/20/2015 10:04 AM, Nikos Chantziaras wrote: I'm following the documentation here: https://wiki.gentoo.org/wiki/Handbook:AMD64/Portage/Advanced#Using_.2Fetc.2Fportage.2Fenv to hook into the postinst phase of an ebuild. Deja vu: http://comments.gmane.org/gmane.linux.gentoo.devhelp/151 https://bugs.gentoo.org/show_bug.cgi?id=388781 There's two separate uses for /etc/portage/env -- one of them will get you the make.conf parser (when you list the file in /etc/portage/package.env) and the other that will get you the bash parser (when you put the file at /etc/portage/env/${CATEGORY}/${PN}). The whole thing is confusing. If you ever get it working please put a decent example on the wiki because I remember spending hours trying to do the same thing. As suggested by Neil, putting the code in /etc/portage/env/app-misc/mc worked. Will later write an example in the Wiki.
[gentoo-user] Re: How is /etc/portage/env supposed to work?
On 20/10/15 18:37, Neil Bothwick wrote: On Tue, 20 Oct 2015 17:04:10 +0300, Nikos Chantziaras wrote: So I assume, since the docs don't mention anything special, that I should put the code in a conf file. So I created this file: /etc/portage/env/test.conf and in it: if [ "${EBUILD_PHASE}" == "postinst" ]; then echo ":: TEST" fi [...] But when I "emerge app-misc/mc", I get: !!! Problem in 'app-misc/mc' dependencies. !!! "/etc/portage/env/test.conf", line 1: Invalid token '[' (not '=') portage.exception ... done! "/etc/portage/env/test.conf", line 1: Invalid token '[' (not '=') I'm nut sure you can put that sort of thing in package.env, I thought it was only for settings variables, like make.conf but per-package. You can add function calls in /etc/portage/env/cat/pkg, for example I have used this for a couple of packages that didn't support epatch_user post_src_unpack() { cd "${S}" epatch_user } AFAIK you can do this for any of the standard ebuild functions listed in man 5 ebuild. Thanks! That works just fine. Btw, you can put your epatch_user in /etc/portage/bashrc instead and have it for all packages. It is safe to have it twice even in packages that do support epatch_user. That function is specifically written for this use case, so that even when using it multiple times, the patches are only applied once.
[gentoo-user] Re: clean-up root partition
On 01/10/15 15:41, the...@sys-concept.com wrote: How do you folks clean-up root partition, I have too much junk in there. df -h Filesystem Size Used Avail Use% Mounted on /dev/sda374G 61G 9.3G 87% / I've already removed all the files from: /usr/portage/distfiles You can maintain that with "eclean-dist -d". Also check portage/packages for old binary packages. ncdu also makes it easier to find where the space is consumed. "ncdu /" will scan and sort by size.
[gentoo-user] Re: clean-up root partition
On 01/10/15 23:14, Nikos Chantziaras wrote: ncdu also makes it easier to find where the space is consumed. "ncdu /" will scan and sort by size. Sorry, should have been: ncdu -x / This will exclude mounted filesystems.
[gentoo-user] Re: portage can not find local ebuild
On 18/09/15 06:15, the...@sys-concept.com wrote: I'm trying to emerge one of my local ebuild and portage can not find it. What am I missing? emerge -avq nxclient emerge: there are no ebuilds to satisfy "nxclient". My settings: make.conf. ... PORTDRI_OVERLAY="/usr/local/portage" cat /etc/portage/repos.conf/gentoo.conf [DEFAULT] main-repo = gentoo [gentoo] location = /usr/portage sync-type = rsync sync-uri = rsync://192.168.139.7/gentoo-portage Remove the PORTDIR_OVERLAY entry from make.conf. Then, create a new file in /etc/portage/repos.conf/, for example "local.conf", with these contents: [Local] priority = location = /usr/local/portage auto-sync = no The high value for priority is just to make sure that your local ebuilds override any other ebuilds of the same name that exist elsewhere.
[gentoo-user] Re: [solved] app-office/libreoffice-5.0.1.2
On 15/09/15 12:21, Stefan G. Weichinger wrote: On 15.09.2015 08:49, Stefan G. Weichinger wrote: I will retry a build and show you parts of the log. Something in my merging history seems to have solved it now. Merged OK today, sorry for the noise. I'd run a complete memtest just in case. Last time libreoffice failed to build for me, and the failure was at a different point in the build process, it turned out a memory module had developed an error.
[gentoo-user] Re: new machine : case + power
On 17/09/15 06:04, Philip Webb wrote: My previous machines' (2000/3/7/12) cases all contained a PSU, but it seems I now have to buy case/PSU separately, so my next question is whether I can safely use the case + power supply from the 2007 machine This should be fine, as long as the PSU has SATA power connectors. If not, you will need molex to SATA adapters. The PSU does not need to provide a PCI-e power connector, since the Asus GT610 does not take external power. It's powered exclusively by the bus. However, note that PSUs tend to go bad after years of operation. An 8 years old PSU might start losing power or voltage stability. If you see random machine resets or hangups, it's usually because the PSU is dying. With that GPU, you don't need more than 400W. And the amps on the 12V rail are also not important, as that only comes into play with higher-end GPUs. Other features can still be important though, like protection against surges and such (think lightning strikes that can potentially damage your PC, or some PSU malfunction that could do the same.)
[gentoo-user] Re: app-office/libreoffice-5.0.1.2 - fails
On 15/09/15 00:25, Stefan G. Weichinger wrote: can't emerge that new unstable libreoffice updated everything(?) possibly needed, used stable gcc ... no way. I don't want to spam the list with the build.log and I can't find a matching bug at b.g.o. Anyone successful here already? No problems here on ~amd64.
[gentoo-user] Re: /var/tmp/notmpfs - does not exist
On 05/09/15 08:52, Dale wrote: tmpfs /var/tmp/portage tmpfs noatime 0 0 Off-topic: forget "noatime". Use "lazytime" instead: https://lwn.net/Articles/621046
[gentoo-user] Re: git clone tricks
On 04/09/15 18:51, James wrote: hello, So I'm still learning the tricks of git.. I tried all sorts of things suggested on the net, but I cannot seem to find a way to clone this site: You cannot clone sites with git. You can only clone git repositories. What you have is an HTTP website, not a git repo.
[gentoo-user] Re: [SOLVED] media-libs/jpeg / media-libs/libjpeg-turbo - blocking each other
On 01/09/15 03:35, the...@sys-concept.com wrote: Found a solution. emerge -avC libjpeg-turbo emerge -av1 media-libs/jpeg:0 media-libs/jpeg:62 Solved the problem. That is not a solution. It's a workaround. The real solution is to make the package that depends on media-libs/jpeg not depend on it. I doubt that it really needs that package and can't work with libjpeg-turbo instead.
[gentoo-user] Re: UEFI booting
On 17/08/2015 03:59 μμ, Rod wrote: Hi list, I'm trying to figure out how to make my boot partition to boot from UEFI, I have grub2 installed, but I keep getting a error when I ask it to install the boot information. mount /dev/sdc1 201633156 201478 1% /boot/efi I have the /boot/efi part mounted ok.. # efibootmgr efibootmgr: EFI variables are not supported on this system. Boot in UEFI mode first. You can't access those variables if you boot in BIOS mode. To do that, just boot from a SysRescueCD USB stick in UEFI mode and install the bootmanager from there.
[gentoo-user] Re: [OT] Very recent change in behavior of gmail imap/smtp servers
On 22/07/2015 04:34 πμ, walt wrote: Google has just introduced a 120-second delay before allowing login to their email servers. Just in the last day or two, literally. No delay here with POP3. Login is instant.
[gentoo-user] Re: SDDM/KDE5: no sound card available?
On 22/07/2015 04:34 πμ, Jonathan Callen wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA512 On 2015-07-21 14:12, Nikos Chantziaras wrote: I upgraded to KDE 5 recently, and was using LightDM as the display manager. It seems that KDE 5 prefers SDDM though and offers a config module for it in System Settings. So I installed SDDM. However, when I log in with SDDM, I get no sound. My sound card just... disappears. alsamixer -c0 says: invalid card index: 0. [...] Are you using systemd? Nope. I'm on OpenRC. If not, did you build sddm with USE=consolekit and read the warning printed by portage? This display manager doesn't have native built-in ConsoleKit support. In order to use ConsoleKit pam module with this display manager, you should remove the nox11 parameter from pm_ck_connector.so line in /etc/pam.d/system-login Your issue is most likely that your X session is not being treated as a login session by logind/ConsoleKit, and therefore your user is not being added to the ACLs on the various devices under /dev, including all sound devices, certain input devices, any CD/DVD/BR devices you may have, and certain video devices. I had tried that already. Didn't mention it because I thought this only affects graphics. Anyway, it changes nothing. Still no sound card, even after reboot.
[gentoo-user] Re: [OT] Very recent change in behavior of gmail imap/smtp servers
On 23/07/2015 05:49 πμ, walt wrote: Nope. Wrong. I just changed my resolv.conf back to the IP address of the router that ATT forced me to upgrade to and the delay is *gone*. The delay I was seeing was apparently caused by something very local to me, and suddenly vanished after two days. The interwebz is a scary place :( A friend of mine had a problem where half the time he tried to browse to a URL, he would end up on a porn site. I thought it was some Windows malware. But when booting from a USB stick with SysRescueCd on it, even ping google.com would ping a porn site at first. His modem/router combo device was infected with something that hijacked the DNS setting. He was using a DSL-Modem/router from 2003. It *is* a scary place.
[gentoo-user] SDDM/KDE5: no sound card available?
I upgraded to KDE 5 recently, and was using LightDM as the display manager. It seems that KDE 5 prefers SDDM though and offers a config module for it in System Settings. So I installed SDDM. However, when I log in with SDDM, I get no sound. My sound card just... disappears. alsamixer -c0 says: invalid card index: 0. aplay -l says: aplay: device_list:268: no soundcards found... Even if I disable pulseaudio, there's still no ALSA cards found. When I ctrl+alt+F1 to a console and log in there, there's no problem. My sound card is found. But in a KDE session started with SDDM, nope. No sound card (and thus, no sound.) Works just fine with LightDM. I'm on ~amd64 with gentoo-sources-4.1.2, pulseaudio-6.0. KDE 5 is installed from portage (kde-plasma/plasma-meta). Any help?
[gentoo-user] Re: In the fear of getting hacked (WLAN setup)
On 18/07/2015 08:43 μμ, Andrew Savchenko wrote: On Sat, 18 Jul 2015 06:47:21 +0300 Nikos Chantziaras wrote: The problem I (possibly needless) see is: While I am tinkering and testing the configuration I may setup an open Wifi access point without noticing it in first glance and BANG! get hacked ... in the worst case: unrecognized... If you don't have any daemons running that provide network services (have opened listen ports), you can't get hacked. Yes and no. If user enabled network interface and has no network daemons running, kernel still listens to that interface (ARP, icmp and so on) and may be hacked using vulnerabilities in network stack, protocol handlers or even network device drivers. Which is not a realistic scenario. We can assume that for all intents and purposes, the OP is safe.
[gentoo-user] Re: In the fear of getting hacked (WLAN setup)
On 18/07/2015 06:34 πμ, meino.cra...@gmx.de wrote: Hi, in order to connect my ASUS Memp Pad 7 ME176CX to the internet I need a working WLAN (my DSL router/modem is of the copper area - no Wifi/WLAN). The hardware (an USB dongle) is already there...it needs only be configured and setup. The problem I (possibly needless) see is: While I am tinkering and testing the configuration I may setup an open Wifi access point without noticing it in first glance and BANG! get hacked ... in the worst case: unrecognized... What is the best practice here? Is there a certain independant configuration, which I can set, which prevents this scenario? Thank you very much in advance for any help! Best regards, Meino PS: If one knows the ASUS Memo Pad 7 ME176CX and knows a way to locally connect this tablet to the internet...this would be a way to go also. I would appreciate any hint in this case (Using Lollipop 5.0). If you don't have any daemons running that provide network services (have opened listen ports), you can't get hacked. This is usually a problem for Windows, which by default has a gazillion of services running (NetBIOS, printer/media/filesystem/everything sharing, messaging, remote desktop, etc.) On Gentoo, if *you* didn't set up a service, then nothing is listening on the network.
[gentoo-user] Re: How to avoid perl harbor (pun intended)
On 13/07/15 15:19, walt wrote: Today's update started as a disaster: perl wants to upgrade from 5.20.2 to 5.22.0, but all of my existing perl modules insist on having 5.20.2 so the perl update blocks and then emerge stopped with an error and left the whole mess for me to solve. (To me this appears to be a bug in the perl family of ebuilds.) Same here. I uninstalled perl and tried again. This always worked with other packages. But here: nope. Portage wants to then install both versions at the same time. Not good.
[gentoo-user] Re: How to avoid perl harbor (pun intended)
On 13/07/15 16:01, Simon Thelen wrote: On 15-07-13 at 05:19, walt wrote: Today's update started as a disaster: perl wants to upgrade from 5.20.2 to 5.22.0, but all of my existing perl modules insist on having 5.20.2 so the perl update blocks and then emerge stopped with an error [...] It worked for me after I added --backtrack=30 That was it. After using that option, it works. So I guess the default value is not good enough.
[gentoo-user] Re: [OT} CPU question
On 13/07/15 19:04, meino.cra...@gmx.de wrote: Are crosscompilers freely available for it (or is this CPU even i386 compatible) It's x86 and x86-64 compatible (it's a 64-bit CPU). With -march=native, GCC will use the most appropriate instruction sets for this CPU.
[gentoo-user] Re: Anyone else having a problem with bash?
On 11/07/15 23:56, Martin Vaeth wrote: Nikos Chantziaras rea...@gmail.com wrote: I really don't have time to learn arcane settings anymore. That's why it is good that you can adapt the shell completely to your needs: My opinion is that the computer must adapt to *my* habits and not vice versa. If it doesn't work out of the box I don't see the relation with the above: hobody recommended to use zsh without a reasonable configuration. You have been told several ways to obtain a reasonable configuration in a very quick way, and you are free to change it to any special wishes. It is with shells as with editors: Whatever is the default, quite a lot of people will not be satisfied with it. I disagree. It should work out of the box. People can change it later. Having it not work at all is just stupid. The good way to write software is having it work and let people customize it if they want. Shipping it in a broken state is not good.
[gentoo-user] Re: Anyone else having a problem with bash?
On 11/07/15 01:18, Marc Joliet wrote: Am Fri, 10 Jul 2015 20:39:05 +0300 schrieb Nikos Chantziaras rea...@gmail.com: On 10/07/15 18:00, Gevisz wrote: bindkey '^[[7~' beginning-of-line # Home (xterm) bindkey '^[[8~' end-of-line# End (xterm) lol... are these guys serious? It's 2015... What's wrong with C-a and C-e? On my keyboard they're not any more cumbersome than reaching for Home/End (probably a bit less cumbersome in fact). Pressing two keys at the same time with the same hand is not good. It just feels awkward. I have special keys for that. They should be used. Also, word-based navigation doesn't either. Ctrl+left or ctrl+right, for example. I guess the lack of good defaults kills this shell. I really don't have time to learn arcane settings anymore. If it doesn't work out of the box on what represents 99.% of machines, I'll have to pass and wonder what were they thinking?
[gentoo-user] Re: adobe-flash-11.2.202.481 available; patches exploit
On 10/07/15 14:00, Nikos Chantziaras wrote: On 09/07/15 17:41, Walter Dnes wrote: Adobe has released adobe-flash-11.2.202.481 for linux, and it is now available as an ebuild. Thanks to the devs for a quick turnaround with the ebuild. It is important to update, because a zero-day exploit was discovered in the 400 gigabytes of data from the Hacking Team hack. And yes, it can potentially hit linux too... http://venturebeat.com/2015/07/07/adobe-confirms-flash-vulnerability-found-via-hacking-team-leak-promises-patch-tomorrow/ Does this affect Chrome? It has its own version of Flash. Apparently I'm blind. It's already fixed for Chrome.
[gentoo-user] Re: adobe-flash-11.2.202.481 available; patches exploit
On 09/07/15 17:41, Walter Dnes wrote: Adobe has released adobe-flash-11.2.202.481 for linux, and it is now available as an ebuild. Thanks to the devs for a quick turnaround with the ebuild. It is important to update, because a zero-day exploit was discovered in the 400 gigabytes of data from the Hacking Team hack. And yes, it can potentially hit linux too... http://venturebeat.com/2015/07/07/adobe-confirms-flash-vulnerability-found-via-hacking-team-leak-promises-patch-tomorrow/ Does this affect Chrome? It has its own version of Flash.
[gentoo-user] Re: Anyone else having a problem with bash?
On 10/07/15 18:00, Gevisz wrote: bindkey '^[[7~' beginning-of-line # Home (xterm) bindkey '^[[8~' end-of-line# End (xterm) lol... are these guys serious? It's 2015...
[gentoo-user] Re: Anyone else having a problem with bash?
On 09/07/15 15:01, Gevisz wrote: On Thu, 9 Jul 2015 12:48:24 +0100 Neil Bothwick n...@digimed.co.uk wrote: On Thu, 09 Jul 2015 13:38:43 +0200, Stephan Müller wrote: As a wild guess into the blue, it could be related to readline. As I see gentoo's bash uses the standalone readline from coreutils, while the original bash source maintains an own trimmed version of readline.. just a thought In that case, re-emerging Bash with USE=-readline should get rid of the problem. smugI can't test it myself as a use a superior shell to Bash/smug Which one? And why is it superior to bash? Don't ask such questions ;-) This is in the same vein as the emacs vs vim argument.
[gentoo-user] Re: Anyone else having a problem with bash?
On 09/07/15 14:48, Neil Bothwick wrote: On Thu, 09 Jul 2015 13:38:43 +0200, Stephan Müller wrote: As a wild guess into the blue, it could be related to readline. As I see gentoo's bash uses the standalone readline from coreutils, while the original bash source maintains an own trimmed version of readline.. just a thought In that case, re-emerging Bash with USE=-readline should get rid of the problem. Doesn't seem possible. That USE flag seems to get ignored by portage: emerge --info bash [...] app-shells/bash-4.3_p39::gentoo was built with the following: USE=net (policykit) (readline) -afs -bashlogger -examples -mem-scramble -nls -plugins -vanilla ABI_X86=64 So readline is enabled. But: echo app-shells/bash -readline /etc/portage/package.use emerge -uDN --with-bdeps=y @world [...] Nothing to merge; quitting. That USE flag doesn't do anything.
[gentoo-user] Re: Anyone else having a problem with bash?
On 09/07/15 15:24, wraeth wrote: On Thu, Jul 09, 2015 at 03:19:19PM +0300, Nikos Chantziaras wrote: On 09/07/15 14:48, Neil Bothwick wrote: On Thu, 09 Jul 2015 13:38:43 +0200, Stephan Müller wrote: As a wild guess into the blue, it could be related to readline. As I see gentoo's bash uses the standalone readline from coreutils, while the original bash source maintains an own trimmed version of readline.. just a thought In that case, re-emerging Bash with USE=-readline should get rid of the problem. Doesn't seem possible. That USE flag seems to get ignored by portage: [...] echo app-shells/bash -readline /etc/portage/package.use emerge -uDN --with-bdeps=y @world [...] Nothing to merge; quitting. That USE flag doesn't do anything. Use the command `emerge -uav --changed-use app-shells/bash` - you need to identify that it's a changed use flag, otherwise it ignores because there are no new versions. Still nothing. But I was using -N to begin with (--newuse) which is stronger than --changed-use. It doesn't seem possible to change that USE flag.
[gentoo-user] Re: Anyone else having a problem with bash?
On 09/07/15 19:07, Neil Bothwick wrote: On Thu, 09 Jul 2015 15:07:40 +0300, Nikos Chantziaras wrote: Don't ask such questions ;-) This is in the same vein as the emacs vs vim argument. True, people persist with Bash and vim, but in the latter case it appears to be because they actually like it :-O Most people use Bash because it is the default and they have never tried anything else - just like the situation with most computer users and Windows. I'd been lectured on the wonders of zsh in the past, but it was only when I had to learn to use it (I wrote a comparative review of shells) that I realised what it offered over Bash as an interactive shell. I tried it, for exactly 10 seconds. My home/end keys didn't work. This gave me the impression of an unfinished project. Why on earth would anyone release a program after 1990 that doesn't know the home/end keys? :-/
[gentoo-user] Re: Anyone else having a problem with bash?
On 10/07/15 02:34, Nikos Chantziaras wrote: I tried it [zsh], for exactly 10 seconds. My home/end keys didn't work. This gave me the impression of an unfinished project. Why on earth would anyone release a program after 1990 that doesn't know the home/end keys? :-/ PS: The Del key doesn't work either.