Re: [gentoo-user] The war continues.
160317 Alan Grimes wrote: > My effort to update my syestem continues unabated. > One problem is that every time revdep-rebuild is run, > it always rebuilds all of libreoffice, an 8-hour build. When I upgraded to LO 5.1.0.3 , it took 1 h 4 m : do you have a very old machine ? > 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 ? > I don't even know how to read the current error message That's a common experience for most of us (smile) ! First advice : most of the "errors" are really advisories. > Current state of mind: > put a live hand grenade in to the computer and walk away. Better to put it all aside till tomorrow. > I've been using Gentoo every day now for ten years. > This is an entirely new level of bullshit. I've been using it since 2003 & have never had a problem this size. However in all that time, I've never done 'emerge world' without '-p' ; I update pkgs individually, typically with '-1'. Try 'emerge -Dvp world' & send the output to a file, then work thro' the list a few pkgs at a time ; if you run into a block, skip that pkg & see how far you can get ; repeat. The result wb to isolate what's going wrong, when you can test solutions, which are usually 1 level deep. HTH -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Re: The war continues.
160318 Nikos Chantziaras wrote: > On 17/03/16 18:00, Philip Webb wrote: AG> I tried to fix it by setting wayland, gles2 and egl to default > because they were breaking other packages. PW> Why are you using Wayland ? -- it's still largely experimental, isn't it ? NC> KDE 5 packages force the wayland USE flag on > & break 'emerge', if you try to disable wayland. So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). Everything works for me using Fluxbox + some KDE 4 apps. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 03/17/2016 11:31 PM, Mick wrote: On Friday 18 Mar 2016 06:01:17 Bill Kenworthy wrote: On 18/03/16 05:59, Bill Kenworthy wrote: On 18/03/16 05:14, Alan McKinnon wrote: On 17/03/2016 22:02, Håkon Alstadheim wrote: On 03/17/2016 02:03 PM, Bill Kenworthy wrote: On 17/03/16 20:26, Alan McKinnon wrote: On 17/03/2016 08:50, Håkon Alstadheim wrote: I have a server SUPPOSED to be running 24/7, but every once in a while during a prolonged absence the box will go down. The Real Time Clock will drift, and in the rush to get the box up again I let everything boot up automatically and get both wrong time on the main systems, and different times on the various systems. My setup has a main server which does NTP, but with no direct link to the outside. Router&firewall /have/ to be booted booted later (dumb setup, don't ask), after which I can finally get correct time from NTP. NTP initiates "11 minute mode", which makes /etc/adjtime useless as far as I understand. Anybody have a /correct/ way to account for RTC drift on a box running ntpd? Right now I have a ---file in /etc/cron.d/time-bad like so: * * * * * root adjtimex -S 5 >/dev/null 2>&1 Have you looked at adjtimex ... its in portage From the man page ... "For a standalone or intermittently connected machine, where it’s not ossible to run ntpd, you may use adjtimex instead to correct the sys-tem clock for systematic drift. There are several ways to estimate the drift rate. If your computer can be connected to the net, you might run ntpd for at least several hours and run "adjtimex --print" to learn what values of tick and freq it settled on. Alternately, you could estimate values using as a reference the CMOS clock (see the --compare and --adjust switches), another host (see --host and --review), or some other source of time (see --watch and --review). You could then add a line to rc.local invoking adjtimex, or configure /etc/init.d/adjtimex or /etc/default/adjtimex, to set those parameters each time you reboot." Used it at one time for dialup which approximates your condition. BillK forget it ... I forgot that's where you started from ... must be getting old :( Nobody mentioned net-misc/chrony. Would it be more appropriate for this use case? I see it also claims to contain an ntp server. I'll check it out.
[gentoo-user] Re: Confessional: how I generally use emerge.
On Thu, 17 Mar 2016 20:37:04 -0400 Alec Ten Harmsel wrote: > > emerge --update --newuse --deep --with-bdeps=y system --keep-going > > Add "--oneshot", same reasoning as above. When the target is a set (in this case @system), does portage ever add all of it to @world?
[gentoo-user] Re: The war continues.
On 2016-03-17, Alan Grimes wrote: > My effort to update my syestem continues unabated. =\ > > One problem is that every time revdep-rebuild is run, it always rebuilds > all of libreoffice, an 8-hour build. WTF, seriously, WTF? > > I made the mistake of syncing portage again and was thrown back into > useflag hell: Have you considered just re-installing from scratch? I know it's tough to admit defeat, but when a base install takes less than an hour[1]... [1] Plus the overnight "emerge" command that builds all the fancy stuff... -- Grant Edwards grant.b.edwardsYow! Oh, I get it!! at "The BEACH goes on", huh, gmail.comSONNY??
[gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
I have a server SUPPOSED to be running 24/7, but every once in a while during a prolonged absence the box will go down. The Real Time Clock will drift, and in the rush to get the box up again I let everything boot up automatically and get both wrong time on the main systems, and different times on the various systems. My setup has a main server which does NTP, but with no direct link to the outside. Router&firewall /have/ to be booted booted later (dumb setup, don't ask), after which I can finally get correct time from NTP. NTP initiates "11 minute mode", which makes /etc/adjtime useless as far as I understand. Anybody have a /correct/ way to account for RTC drift on a box running ntpd? Right now I have a ---file in /etc/cron.d/time-bad like so: * * * * * root adjtimex -S 5 >/dev/null 2>&1
Re: [gentoo-user] How to find which version of $package supports $USE ?
On 03/18/2016 12:06, Rich Freeman wrote: There is some debate about Gentoo having it enabled by default, since it is not part of upstream. However, the hpn patch is probably fairly popular. It is almost a fork, and I have no idea why it hasn't just been merged into openssh proper. Likely because several parts of it can also degrade performance, as stated on the patch's site. It also doesn't seem to have undergone any serious security audit, or at least that's the impression I got from a quick review.
OT: Re: [gentoo-user] Giving Gentoo Another Go
Peter Humphrey wrote: > On Friday 18 March 2016 12:03:46 Neil Bothwick wrote: > > > Top Oxymorons Number 33: American history > > Top Oxymorons Number 1a: atonal music. I don't agree to the second statement. ;-) -- Regards wabe
Re: [gentoo-user] The war continues.
Wow, sounds like you've been having a rough time. I think if you chunk things up into tiny pieces of logs, and take the issues on one at a time, you will be able to solve your problems. That said, if you would like to send us the last month or so of your emerge.log, we might be able to help focus your efforts. Until you have everything building cleanly and working as you hoped, I would avoid doing any new portage syncs just to keep from having an issue where the set of problems you're looking at might shift dramatically. This way you can look at the problems one at a time then eliminate them before new problems come up. Similarly, unless you see a problem where a program is complaining about not being able to load a shared object, I would just put off the revdep-rebuild until you've successfully updated. It looks like you have two distinct problems in the messages you posted here, a problem with asciidoc and another with kwin and mesa. On asciidoc I think a good first step would be setting PYTHON_SINGLE_TARGET="python2_7" if you haven't already, I see that in the make.conf you posted at the end of February[1] you have PYTHON_SINGLE_TARGET="python3_5". Could you give changing python_single_target a try and report back? I know it will not solve all your issues, but at least it should reduce the amount of negative feedback emerge is giving you. Also, is the script ./pretendupdate something like "emerge --pretend --verbose --all --newuse @world"? [1]: https://marc.info/?l=gentoo-user&m=145668323000774&q=p3 -- 0x7D964D3361142ACF On Thu, Mar 17, 2016, at 08:12, William Ernesto Cárdenas Gómez wrote: > El jue, 17-03-2016 a las 10:55 -0400, Alan Grimes escribió: > > My effort to update my syestem continues unabated. =\ > > > > One problem is that every time revdep-rebuild is run, it always > > rebuilds > > all of libreoffice, an 8-hour build. WTF, seriously, WTF? > > > > I made the mistake of syncing portage again and was thrown back into > > useflag hell: > > > > I tried to fix it by setting wayland, gles2 and egl to default > > because > > they were breaking other packages. > > > > > > I don't even know how to read the current error message: > > > > ## > > > > tortoise ~ # ./pretendupdate > > > > These are the packages that would be merged, in order: > > > > Calculating dependencies... done! > > > > !!! Multiple package instances within a single package slot have been > > pulled > > !!! into the dependency graph, resulting in a slot conflict: > > > > media-libs/mesa:0 > > > > (media-libs/mesa-11.1.2-r1:0/0::gentoo, ebuild scheduled for merge) > > pulled in by > > (no parents that aren't satisfied by other packages in this slot) > > > > (media-libs/mesa-11.1.2-r1:0/0::gentoo, installed) pulled in by > > media-libs/mesa[egl,gbm,gles2?,wayland] required by > > (kde-plasma/kwin-5.5.5:5/5::gentoo, ebuild scheduled for merge) > > > > ^^^ > > > > > > > > It might be possible to solve this slot collision > > by applying all of the following changes: > > - media-libs/mesa-11.1.2-r1 (Change USE: +wayland +gles2) > > > > > > !!! The ebuild selected to satisfy "app-text/asciidoc" has unmet > > requirements. > > - app-text/asciidoc-8.6.9-r2::gentoo USE="graphviz -examples > > -highlight > > -test" ABI_X86="64" PYTHON_SINGLE_TARGET="-pypy -python2_7" > > PYTHON_TARGETS="python2_7 -pypy" > > > > The following REQUIRED_USE flag constraints are unsatisfied: > > exactly-one-of ( python_single_target_pypy > > python_single_target_python2_7 ) > > > > The above constraints are a subset of the following complete > > expression: > > exactly-one-of ( python_single_target_pypy > > python_single_target_python2_7 ) python_single_target_pypy? ( > > python_targets_pypy ) python_single_target_python2_7? ( > > python_targets_python2_7 ) > > > > (dependency required by "net-misc/tor-0.2.8.1_alpha::gentoo" > > [installed]) > > (dependency required by "@selected" [set]) > > (dependency required by "@world" [argument]) > > tortoise ~ # > > > > # > > > > Current state of mind: put a live hand grenade in to the computer and > > walk away. > > > > I've been using Gentoo Every day now for ten years. This is an > > entirely > > New level of bullshit. =( > > > > > > Modest list of complete and utter FAIL: > > > > > > > > tortoise portage # tree -L 2 > > . > > ├── app-office > > │ └── texmacs-1.99.2-r1 > > ├── dev-libs > > │ ├── libcdio-0.93 > > │ └── libcdio-paranoia-0.93_p1 > > ├── dev-qt > > │ └── qtwebkit-5.5.1-r1 > > ├── kde-apps > > │ └── kdesdk-kioslaves-15.12.2 > > ├── kde-plasma > > │ ├── oxygen-5.5.5 > > │ └── oxygen-fonts-5.4.3 > > ├── media-gfx > > │ └── fontforge-20150824 > > ├── media-libs > > │ ├── opencv-3.1.0-r2 > > │ └── x264-0.0.20151011 > > └── media-video > > └── vcdimager-0.7.24 > > >
[gentoo-user] Giving Gentoo Another Go
Hello, After talking to a few diehard Gentoo fans at my local LUG, I decided I would like to give Gentoo another shot. Are there any good books that can supplement the Gentoo handbook as well as books that go more in depth than the Gentoo chapter on Portage? One of the main issues I faced with Gentoo when I first tried it is that I did not understand the power of package.use, and I put everything in to make.conf. However, I feel that given enough supplemental information, I can hopefully make Gentoo attempt 6 a more permanent thing, and, eventually, migrate my servers over to it. Any input is greatly appreciated. Thanks, Hunter
Re: [gentoo-user] Re: The war continues.
Philip Webb wrote: > 160318 waben...@gmail.com wrote: > > Alan Grimes wrote: > >> Philip Webb wrote: > >>> So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). > >>> Everything works for me using Fluxbox + some KDE 4 apps. > >> The last good version of KDE was 3.5.x. Then the flood came (Qt > >> 4) > > My conclusion at the time was that my current set-up is slightly > better : Fluxbox provides the same basics & is simpler than the KDE 3 > environment, while the KDE apps I use (mainly Konsole Gwenview > Okular) are better in 4 . Do you have installed the complete KDE4? If not, what packages do you have installed to configure the look and feel of your KDE apps? > However, there are 2 small items I really miss : Kmahjongg > Kworldclock . I had them in my previous machine (now stand-by), > but they need a library which I can't install anymore, > so I haven't tried to recreate them in my new machine. > (Yes, I know there is Kmahjongg-4 , but it doesn't do "removed > tiles"). > > > waben...@gmail.com wrote: > >> What I miss most of all is the fantastic konqueror. > >> It was way better than any other filemanager that I know. > > I use Krusader & recommend it very highly for heavy file-lifting : > have you tried it ? -- if not, do : you may well like it. Thanks for your suggestion. The dependencies are about the same as I would install konqueror. Over 20 new packages and also some blockers. And it seems that it is such a two panel thing, but that's usually not my thing. ;-) Is it possible to use tabs and just one panel? And does it have a tree mode? -- Regards wabe
Re: [gentoo-user] Confessional: how I generally use emerge.
On Thu, Mar 17, 2016 at 08:18:13PM -0400, Alan Grimes wrote: > I use two scripts for all emerge use, the goal is to run one command and > then walk away: > > Standard general update script: > ### > tortoise ~ # cat sysupdate > > #they must have moved or removed the logs, might have to track them down > again... > #rm /var/log/emerge* > > # cache /usr/portage > echo "caching /usr/portage. This will take a long time." > time ls -R /usr/portage > /dev/null > > emerge --sync > layman --sync ALL > > emerge --update --verbose portage You almost certainly want this to be: emerge --update --verbose --oneshot portage "--oneshot" will prevent portage from being added to your @world set. > emerge --update --newuse --deep --with-bdeps=y system --keep-going Add "--oneshot", same reasoning as above. > emerge --update --newuse --deep --with-bdeps=y world --keep-going > > rm -f /var/cache/revdep-rebuild/*.rr delete the above line > revdep-rebuild add "--ignore" to ignore any cached files and fully examine the system. > emerge --skipfirst --resume > emerge --skipfirst --resume > etc-update > eclean-dist > > > The eclean line was added just a few days ago from this thread... > > This one is intended to be a nice gentle update script. > It caches the portage tree, then syncs everything, then updates > everything starting with critical system packages, then all world > packages... > > Then it cleans stuff up, it jcakhammers the revdep-rebuild but not too > hard Seeing so much "--resume" and "--skip-first" and "--keep-going" frightens me. > This next script is what I use when emerge starts giving me shit: > > ## > tortoise ~ # cat keepgoing > emerge --update --newuse --deep --with-bdeps=y system > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > > emerge --update --newuse --deep --with-bdeps=y world > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > > rm /var/cache/revdep-rebuild/*.rr > revdep-rebuild > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > emerge --skipfirst --resume --nodeps > > etc-update > ### > > It's basically the same as the working section of the above but instead > of letting emerge do it's thing, it jackhammers that bitch as hard as > possible to get as much updated as possible, but it requires emerge to > do something and not error out for no good reason... I expect prune and > depclean to be useless but I kinda need update to basically work every > time. =\ > Whatever fails on this script, I just live with until next week/month. Wow. This script makes me cry. > ### > tortoise ~ # ./pretendupdate > > These are the packages that would be merged, in order: > > Calculating dependencies / > > !!! Problem resolving dependencies for sys-apps/util-linux from @system > ... done! > > !!! The ebuild selected to satisfy "sys-apps/util-linux" has unmet > requirements. > - sys-apps/util-linux-2.27.1::gentoo USE="caps cramfs ncurses nls pam > python readline suid udev unicode -build -fdformat -kill (-selinux) > -slang -static-libs -systemd -test -tty-helpers" ABI_X86="32 64 -x32" > PYTHON_SINGLE_TARGET="-python2_7 -python3_3 -python3_4" > PYTHON_TARGETS="python2_7 python3_4 -python3_3" > > The following REQUIRED_USE flag constraints are unsatisfied: > python? ( exactly-one-of ( python_single_target_python2_7 > python_single_target_python3_3 python_single_target_python3_4 ) ) > > The above constraints are a subset of the following complete expression: > python? ( exactly-one-of ( python_single_target_python2_7 > python_single_target_python3_3 python_single_target_python3_4 ) > python_single_target_python2_7? ( python_targets_python2_7 ) > python_single_target_python3_3? ( python_targets_python3_3 ) > python_single_target_python3_4? ( python_targets_python3_4 ) ) > > (dependency required by "@system" [set]) > (dependency required by "@world" [argument]) > > tortoise ~ # cat ./pretendupdate > emerge --update --newuse --deep --with-bdeps=y world --verbose --pretend > tortoise ~ # > > ### > > Google is not being helpful with this... =( What have you set PYTHON_SINGLE_TARGET to? It *must* be one of: * python2_7 * python3_4 Alec
[gentoo-user] The war continues.
My effort to update my syestem continues unabated. =\ One problem is that every time revdep-rebuild is run, it always rebuilds all of libreoffice, an 8-hour build. WTF, seriously, WTF? I made the mistake of syncing portage again and was thrown back into useflag hell: I tried to fix it by setting wayland, gles2 and egl to default because they were breaking other packages. I don't even know how to read the current error message: ## tortoise ~ # ./pretendupdate These are the packages that would be merged, in order: Calculating dependencies... done! !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: media-libs/mesa:0 (media-libs/mesa-11.1.2-r1:0/0::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (media-libs/mesa-11.1.2-r1:0/0::gentoo, installed) pulled in by media-libs/mesa[egl,gbm,gles2?,wayland] required by (kde-plasma/kwin-5.5.5:5/5::gentoo, ebuild scheduled for merge) ^^^ It might be possible to solve this slot collision by applying all of the following changes: - media-libs/mesa-11.1.2-r1 (Change USE: +wayland +gles2) !!! The ebuild selected to satisfy "app-text/asciidoc" has unmet requirements. - app-text/asciidoc-8.6.9-r2::gentoo USE="graphviz -examples -highlight -test" ABI_X86="64" PYTHON_SINGLE_TARGET="-pypy -python2_7" PYTHON_TARGETS="python2_7 -pypy" The following REQUIRED_USE flag constraints are unsatisfied: exactly-one-of ( python_single_target_pypy python_single_target_python2_7 ) The above constraints are a subset of the following complete expression: exactly-one-of ( python_single_target_pypy python_single_target_python2_7 ) python_single_target_pypy? ( python_targets_pypy ) python_single_target_python2_7? ( python_targets_python2_7 ) (dependency required by "net-misc/tor-0.2.8.1_alpha::gentoo" [installed]) (dependency required by "@selected" [set]) (dependency required by "@world" [argument]) tortoise ~ # # Current state of mind: put a live hand grenade in to the computer and walk away. I've been using Gentoo Every day now for ten years. This is an entirely New level of bullshit. =( Modest list of complete and utter FAIL: tortoise portage # tree -L 2 . ├── app-office │ └── texmacs-1.99.2-r1 ├── dev-libs │ ├── libcdio-0.93 │ └── libcdio-paranoia-0.93_p1 ├── dev-qt │ └── qtwebkit-5.5.1-r1 ├── kde-apps │ └── kdesdk-kioslaves-15.12.2 ├── kde-plasma │ ├── oxygen-5.5.5 │ └── oxygen-fonts-5.4.3 ├── media-gfx │ └── fontforge-20150824 ├── media-libs │ ├── opencv-3.1.0-r2 │ └── x264-0.0.20151011 └── media-video └── vcdimager-0.7.24 19 directories, 0 files tortoise portage # -- IQ is a measure of how stupid you feel. Powers are not rights.
Re: [gentoo-user] Confessional: how I generally use emerge.
Neil Bothwick wrote: > > > Why ask the same question again when you got an answer last time? > > Hint: look at the output for asciidoc. > -- > Sent from my Android phone with K-9 Mail. Please excuse my brevity. I have to add this. From what I understand about the scripts he is using, he is blindly letting emerge do updates without checking to see if the updates fall into line with what he *needs*. If I read it correctly, any USE flag change will be missed until it hits the fan and is broken. That would then mean taking a lot of time to go back through logs and figuring out just when it went wrong and most importantly, what caused it and how to fix it. Since the change could have happened several updates ago, that could involve some work and a lot of rebuilding. Doing the sync in a script and even getting a email or something with the -p output is fine but updating blindly is not a good idea. At this point, I think I understand why he is having so much trouble. At the very least, check to see what USE flags are changing before doing updates. There are other things that are important but that is one that changes a good bit. Dale :-) :-)
Re: [gentoo-user] Re: The war continues.
On Sat, 19 Mar 2016 04:08:15 +0100, waben...@gmail.com wrote: > What I also don't want is too much crap that I don't need, e.g. a > networkmanger. Although I set USE="-networkmanager" portage wants to > install it when I type "emerge -pv plasma-meta". It uses the geolocation code in the networkmanager library this is a hard dependency until someone decided to create a patch.. I too don't need NM (I use systemd-networkd) but I have to have NM installed, but not running. -- Neil Bothwick This is the day for firm decisions! Or is it? pgpW3r5uw7HP2.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Re: The war continues.
On Fri, 18 Mar 2016 21:51:24 -0500, Dale wrote: > >> The hard part, getting it to run as root. KDE doesn't like things > >> running as root so it took a hammer and some elbow grease. > > sudo konqueror works here. > This works here as a desktop shortcut. > > kfmclient openProfile filemanagement How does that run it as root? > Never cared much for sudo either. If you want to run things as root, it's the obvious choice. Even more so if you want to run some things as root without a password prompt (although doing that with a file manager s not a wise move). -- Neil Bothwick The thrill of victory, the agony of delete. pgpP5t8mE9pxm.pgp Description: OpenPGP digital signature
Re: [gentoo-user] local shared directory
On 17/03/2016 19:19, hw wrote: > > Hi, > > how can I make it so that multiple users on a system who create > files in a local, shared directory do have write access to files > created by other users within the shared directory? > > The directory is group-writeable, and the users belong to the group > which owns the directory. This enables them to create files within > the shared directory, yet the files they create belong to the user > who created it, and other users cannot modify them. The sticky bit > is set so that the files are owned by user:common-group. > > I would like to avoid changing the umask. If that cannot be avoided, > how do I change it? Users log in through x2goclient, and fvwm is > being executed on login. > Ooh, that's a horrible one, with no really obvious answer. First, you cannot do it with just regular Unix permissions. umask is just not viable either, as a) it's global and affects all files a user creates and b) by definition umask is modifiable by the user (it's a feature to help users out so they don't need to chmod every file every time) and c) you can't stop them doing it (by design). There is a way to do it with Posix ACLs, I figured it out once. It was ugly. It was horrible. It was impossible to describe to someone else. And it was invisible (you had to spot the tiny "+" in ls -al and know what it means to know to look further. The simplest way is to run chown -R g+w dir in a cron every few minutes. This works but it's inelegant. The best solution I have found yet is to use the inotify feature in the kernel. It's an event framework and really neat: tell the kernel to generate an event every time something specific happens on the filesystem, and write a small listener that run chmod. There are many examples in the man pages. In your case, the area to monitor is the shared directory itself, and the event to register is on_file_create and on_file_modify. The listener is a small script that launches chmod g+w Do read the man pages thoroughly, the above will become clearer. inotify is an amazing tool, I wish it were more in common use. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Giving Gentoo Another Go
Not sure of any books, but some documents on the Gentoo Wiki about the
various projects are helpful.
I use /etc/portage/package.{use,accept_keywords}/* in a somewhat unusual
way. For each package
that needs tuning the USE flags beyond the eselected profile and some
globals in make.conf, I have
a file named "." (e.g. media-libs.mesa) with a few handling
some multiples. This lets me
tune packages in a fine grained manner, and still see which ones I have
fixed quickly.
For example:
In eselect profile: default/linux/amd64/13.0/desktop/kde
In make.conf:
(may well repeat some from profile, try to group by functionality)
USE="smp \
cdr dvd dvdr css usb -bluetooth \
caps branding unicode ipv6 tcpd \
kde qt4 qt3support plasma gtk -gnome \
dbus udev udisks perl samba \
icu idn maildir -mbox \
alsa -pulseaudio ffmpeg -libav gstreamer \
X glamor opengl xattr xine xinerama xv xvid \
sdl wxwidgets \
ogg theora vorbis flac dirac schroedinger mp3 mp4 matroska mtp \
a52 aac audiofile cdda faac libsamplerate sndfile x264 \
vdpau xcb vaapi quvi \
lzma lzo lz4 zlib \
cvs git subversion \
sqlite mysql introspection \
bash-completion vim-syntax \
fontconfig truetype threads nptl \
ruby_targets_ruby21 -ruby_targets_ruby19 -ruby19 \
-nepomuk -semantic-desktop \
-ldap -evo -eds \
offensive"
and an ls(1) of
/etc/portage/package.accept_keywords
:
app-admin.mcelog
app-arch.lzma
app-cdr.k3b
app-emulation.qemu
app-emulation.virtualbox
app-forensics.unhide
app-portage.layman
app-text.yodl
dev-libs.libebml
dev-libs.nss
dev-libs.pugixml
dev-python.pyGPG
dev-python.pymountboot
...
virtual.ffmpeg
virtual.libusb
www-client.chromium
www-client.firefox
x11-libs.libva
x11-libs.libva-intel-driver
x11-libs.libva-vdpau-driver
x11-libs.wxGTK
zzzpackages.keywords
The last entry is an empty file to catch the "automagic" changes portage
may propose so that they can be
broken out into individual files. package.use/* is similar.
I have found some wedged cases where there isn't any way to solve it except
by using
some "unstable" arch packages (while waiting for them to be stabilized.)
It may seem unwieldy, but I fnid it much easier to deal with than one big
file with all the
packages isted in it.
For example, in /etc/portage/package.use/app-office.akonadi-server I have:
app-office/akonadi-server soprano -sqlite
which makes the emerge be quiet about having both mysql and sqlite enabled.
Don't be afraid to ask questions, it is an easy cure.
--
G.Wolfe Woodbury
redwo...@gmail.com
Re: [gentoo-user] Re: The war continues.
waben...@gmail.com wrote: > Dale wrote: > >> waben...@gmail.com wrote: >>> Alan Grimes wrote: >>> Philip Webb wrote: > So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). > Everything works for me using Fluxbox + some KDE 4 apps. The last good version of KDE was 3.5.x... then the flood came... (Qt 4) and all the developer started having Ideas about things that could go into the new version >>> Also IMHO KDE 3.5 was the best (K)DE. :-) >>> >>> But I can't envisage that KDE will ever reach that quality again. >>> >>> -- >>> Regards >>> wabe >>> >>> >> >> You two just said a mouth full and I agree. That last part of KDE3 >> was some good stuff. To this day, I still can't get my desktop slide >> show to *not* be random. I filed a bug way back when KDE3 forced >> folks to move to KDE4. I can see how making it random may require >> some work but disabling it shouldn't take to much work. Just beat >> the random thing until it dies. lol >> >> That's just one thing tho. KDE3 was much faster in my opinion. > What I miss most of all is the fantastic konqueror. It was way better > than any other filemanager that I know. Of course I've tested the KDE4 > konqueror and also dolphin but it was horrible compared to the old > konqueror. Now I'm using thunar. It's far away from being perfect, but > it seems to be the lesser of the evils. ;-) > > -- > Regards > wabe > > I still use Konqueror but I disabled some USE flags early on since I didn't want some of the bloat. I think I had to enable some since they were no longer a option but sort of forced. Anyway, this is the USE flags for mine but since it is a short list, USE flags for other packages may affect it more. [ebuild R] kde-apps/konqueror-15.08.3:4/15.08::gentoo USE="bookmarks handbook svg (-aqua) -debug" 0 KiB The hard part, getting it to run as root. KDE doesn't like things running as root so it took a hammer and some elbow grease. I use it to edit config files and it has to run as root to do that. I generally have it open only when I am doing a update tho. Obviously, I never do internet stuff with it either. < wags finger > Dale :-) :-)
Re: [gentoo-user] Confessional: how I generally use emerge.
On Fri, 18 Mar 2016 23:31:46 -0400, Alan Grimes wrote:
> The dependency graph that seems to be murdering me right now seems to
> be:
>
>
> kde -> wayland -> gles2 -> egl ->
I use KDE5 and don't have egl set anywhere in /etc/portage
> ..
>
> [ebuild N ] kde-plasma/plasma-meta-5.5.5:5::gentoo
> USE="display-manager gtk pam pulseaudio sddm sdk wallpapers -bluetooth
> -mediacenter -networkmanager" 0 KiB
> [ebuild NS] kde-apps/kde-meta-15.08.3:5::gentoo
> [4.14.3-r1:4::gentoo] 0 KiB
> [uninstall ] kde-apps/kde-meta-4.14.3-r1:4::gentoo USE="nls sdk
> -accessibility (-aqua) -kdepim -minimal"
> [blocks b ] kde-apps/kde-meta:4 ("kde-apps/kde-meta:4" is blocking
> kde-apps/kde-meta-15.08.3)
> [ebuild U ] dev-lang/mono-4.2.2.30::gentoo [4.2.2.10-r1::gentoo]
> USE="nls (-doc) -minimal -pax_kernel -xen" 0 KiB
> [ebuild N ] dev-dotnet/nuget-2.8.3::gentoo 0 KiB
> [ebuild U ] dev-dotnet/libgdiplus-4.2-r2::gentoo [4.2-r1::gentoo]
> USE="cairo" 0 KiB
> [ebuild U ] dev-util/monodevelop-5.9.5.9-r1::gentoo
> [3.0.2-r1::gentoo] USE="git gnome%* subversion -qtcurve%" 0 KiB
>
> Total: 119 packages (21 upgrades, 32 new, 54 in new slots, 12
> reinstalls, 80 uninstalls), Size of downloads: 395,848 KiB
> Conflict: 80 blocks
>
> WARNING: One or more updates/rebuilds have been skipped due to a
> dependency conflict:
>
> dev-qt/qtgui:5
>
> (dev-qt/qtgui-5.5.1-r1:5/5::gentoo, ebuild scheduled for merge)
> conflicts with
> ~dev-qt/qtgui-5.5.1[-egl] required by
> (dev-qt/qtmultimedia-5.5.1-r2:5/5::gentoo, installed)
I see no errors here, what is stopping the emerge from proceeding?
--
Neil Bothwick
I've found a solution to Fermat's Last Theorem but I see I've run out of
room o
pgpsVSilvFCVo.pgp
Description: OpenPGP digital signature
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 03/17/2016 10:59 PM, Bill Kenworthy wrote:
On 18/03/16 05:14, Alan McKinnon wrote:
On 17/03/2016 22:02, Håkon Alstadheim wrote:
On 03/17/2016 02:03 PM, Bill Kenworthy wrote:
On 17/03/16 20:26, Alan McKinnon wrote:
On 17/03/2016 08:50, Håkon Alstadheim wrote:
I have a server SUPPOSED to be running 24/7, but every once in a while
during a prolonged absence the box will go down. The Real Time Clock
will drift, and in the rush to get the box up again I let everything
boot up automatically and get both wrong time on the main systems, and
different times on the various systems.
My setup has a main server which does NTP, but with no direct link to
the outside. Router&firewall /have/ to be booted booted later (dumb
setup, don't ask), after which I can finally get correct time from NTP.
NTP initiates "11 minute mode", which makes /etc/adjtime useless as far
as I understand. Anybody have a /correct/ way to account for RTC drift
on a box running ntpd? Right now I have a ---file in
/etc/cron.d/time-bad like so:
* * * * * root adjtimex -S 5 >/dev/null 2>&1
Have you looked at adjtimex ... its in portage
From the man page ...
"For a standalone or intermittently connected machine, where it’s not
ossible to run ntpd,
I /can/ and do run ntpd once everything is up and running.
you may use adjtimex instead to correct the sys-tem
clock
System clock is fine, what I'm after is drift of the RTC clock ("bios
clock"). Ntp does nothing for that, as far as I understand. Now, I'd be
very happy if someone could tell me I've misunderstood.
for systematic drift.
There are several ways to estimate the drift rate. If your
computer can be connected to the net, you might run ntpd for at least
several hours and run "adjtimex --print" to learn what values of tick
and freq it settled on. Alternately, you could estimate values using as
a reference the CMOS clock (see the --compare and --adjust switches),
another host (see --host and --review), or some other source of time
(see --watch and --review). You could then add a line to rc.local
invoking adjtimex, or configure /etc/init.d/adjtimex or
/etc/default/adjtimex, to set those parameters each time you reboot."
Used it at one time for dialup which approximates your condition.
Re: [gentoo-user] Re: The war continues.
Neil Bothwick wrote: > On Fri, 18 Mar 2016 21:51:24 -0500, Dale wrote: > The hard part, getting it to run as root. KDE doesn't like things running as root so it took a hammer and some elbow grease. >>> sudo konqueror works here. >> This works here as a desktop shortcut. >> >> kfmclient openProfile filemanagement > How does that run it as root? When I click it, it asks for the root password and after that, it runs as root just like the old KDE3 did. It's actually how it worked in KDE3 since I stole it from there. > >> Never cared much for sudo either. > If you want to run things as root, it's the obvious choice. Even more so > if you want to run some things as root without a password prompt > (although doing that with a file manager s not a wise move). > > I've tried sudo and just don't like it. Same as I don't like Gnome either. It just isn't something I cared for. Besides, I don't mind typing in the root password. Dale :-) :-)
[gentoo-user] Re: dealing with distfiles bloat?
On 07/03/16 03:38, Alan Grimes wrote: I can't really read the stupid unformatted du output but it looks like I have 30 gb of bloat in some 3,600 files in my distfiles directory. is there any sane way to prune out some of the older versions, I am in no mood to spend all day hand-pruning these and the nuclear option is not too friendly to the portage servers that I want to respect. nuclear = rm * -> emerge --fetchonly --emptytree system Just cleaned 9GB distfiles using 'rm /usr/portage/distfiles/*'
Re: [gentoo-user] Re: The war continues.
Dale wrote: > waben...@gmail.com wrote: > > Alan Grimes wrote: > > > >> Philip Webb wrote: > >> > >>> So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). > >>> Everything works for me using Fluxbox + some KDE 4 apps. > >> The last good version of KDE was 3.5.x... then the flood came... > >> (Qt 4) and all the developer started having Ideas about things > >> that could go into the new version > > Also IMHO KDE 3.5 was the best (K)DE. :-) > > > > But I can't envisage that KDE will ever reach that quality again. > > > > -- > > Regards > > wabe > > > > > > > You two just said a mouth full and I agree. That last part of KDE3 > was some good stuff. To this day, I still can't get my desktop slide > show to *not* be random. I filed a bug way back when KDE3 forced > folks to move to KDE4. I can see how making it random may require > some work but disabling it shouldn't take to much work. Just beat > the random thing until it dies. lol > > That's just one thing tho. KDE3 was much faster in my opinion. What I miss most of all is the fantastic konqueror. It was way better than any other filemanager that I know. Of course I've tested the KDE4 konqueror and also dolphin but it was horrible compared to the old konqueror. Now I'm using thunar. It's far away from being perfect, but it seems to be the lesser of the evils. ;-) -- Regards wabe
Re: [gentoo-user] [Solved?] Getting a valid /etc/adjtime while using ntpd ?
On 03/17/2016 11:31 PM, Mick wrote: On Friday 18 Mar 2016 06:01:17 Bill Kenworthy wrote: On 18/03/16 05:59, Bill Kenworthy wrote: On 18/03/16 05:14, Alan McKinnon wrote: On 17/03/2016 22:02, Håkon Alstadheim wrote: On 03/17/2016 02:03 PM, Bill Kenworthy wrote: On 17/03/16 20:26, Alan McKinnon wrote: On 17/03/2016 08:50, Håkon Alstadheim wrote: I have a server SUPPOSED to be running 24/7, but every once in a while during a prolonged absence the box will go down. The Real Time Clock will drift, and in the rush to get the box up again I let everything boot up automatically and get both wrong time on the main systems, and different times on the various systems. My setup has a main server which does NTP, but with no direct link to the outside. Router&firewall /have/ to be booted booted later (dumb setup, don't ask), after which I can finally get correct time from NTP. NTP initiates "11 minute mode", which makes /etc/adjtime useless as far as I understand. Anybody have a /correct/ way to account for RTC drift on a box running ntpd? Right now I have a ---file in /etc/cron.d/time-bad like so: * * * * * root adjtimex -S 5 >/dev/null 2>&1 Have you looked at adjtimex ... its in portage From the man page ... "For a standalone or intermittently connected machine, where it’s not ossible to run ntpd, you may use adjtimex instead to correct the sys-tem clock for systematic drift. There are several ways to estimate the drift rate. If your computer can be connected to the net, you might run ntpd for at least several hours and run "adjtimex --print" to learn what values of tick and freq it settled on. Alternately, you could estimate values using as a reference the CMOS clock (see the --compare and --adjust switches), another host (see --host and --review), or some other source of time (see --watch and --review). You could then add a line to rc.local invoking adjtimex, or configure /etc/init.d/adjtimex or /etc/default/adjtimex, to set those parameters each time you reboot." Used it at one time for dialup which approximates your condition. BillK forget it ... I forgot that's where you started from ... must be getting old :( Nobody mentioned net-misc/chrony. Would it be more appropriate for this use case? This is looking really good. I never considered chrony since I scrapped my modems ~15 years ago, but chrony has these things going for it: - Good documentation in info format - It acknowleges the existence of the rtc - Allows turning OFF system -> rtc updates - Keeps its own drift files, which I believe watches both rtc and system clock drift. I'll run chrony with "-r -s" , with "rtcfile" and without "rtcsync". My hopes are high. :-D .
Re: [gentoo-user] The war continues.
On Thu, 17 Mar 2016 20:26:00 -0400, Alan Grimes wrote: > Then you should understand the need for detailed error reports and > reading the output from the various commands. > > I always set verbosity to very high or max... What's the point if you don't include the output is your posts/rants? Anyway, all verbose does is hide the important messages among a load of trivial messages about everything working as it should. > Back in the good old days there was this OS called DOS. All of it's > commands told you what they were and what they were doing by default, > and were, almost universally quite enjoyable to use. =\ > > These days you just get the next command prompt, and have no idea how > much your previous command might have deleted or otherwise screwed up. Succeed quietly, fail noisily, it's the Unix way. -- Neil Bothwick System halted - Press all keys at once to continue. pgpcohu5TPFgZ.pgp Description: OpenPGP digital signature
Re: [gentoo-user] The war continues.
El jue, 17-03-2016 a las 10:55 -0400, Alan Grimes escribió: > My effort to update my syestem continues unabated. =\ > > One problem is that every time revdep-rebuild is run, it always > rebuilds > all of libreoffice, an 8-hour build. WTF, seriously, WTF? > > I made the mistake of syncing portage again and was thrown back into > useflag hell: > > I tried to fix it by setting wayland, gles2 and egl to default > because > they were breaking other packages. > > > I don't even know how to read the current error message: > > ## > > tortoise ~ # ./pretendupdate > > These are the packages that would be merged, in order: > > Calculating dependencies... done! > > !!! Multiple package instances within a single package slot have been > pulled > !!! into the dependency graph, resulting in a slot conflict: > > media-libs/mesa:0 > > (media-libs/mesa-11.1.2-r1:0/0::gentoo, ebuild scheduled for merge) > pulled in by > (no parents that aren't satisfied by other packages in this slot) > > (media-libs/mesa-11.1.2-r1:0/0::gentoo, installed) pulled in by > media-libs/mesa[egl,gbm,gles2?,wayland] required by > (kde-plasma/kwin-5.5.5:5/5::gentoo, ebuild scheduled for merge) > > ^^^ > > > > It might be possible to solve this slot collision > by applying all of the following changes: > - media-libs/mesa-11.1.2-r1 (Change USE: +wayland +gles2) > > > !!! The ebuild selected to satisfy "app-text/asciidoc" has unmet > requirements. > - app-text/asciidoc-8.6.9-r2::gentoo USE="graphviz -examples > -highlight > -test" ABI_X86="64" PYTHON_SINGLE_TARGET="-pypy -python2_7" > PYTHON_TARGETS="python2_7 -pypy" > > The following REQUIRED_USE flag constraints are unsatisfied: > exactly-one-of ( python_single_target_pypy > python_single_target_python2_7 ) > > The above constraints are a subset of the following complete > expression: > exactly-one-of ( python_single_target_pypy > python_single_target_python2_7 ) python_single_target_pypy? ( > python_targets_pypy ) python_single_target_python2_7? ( > python_targets_python2_7 ) > > (dependency required by "net-misc/tor-0.2.8.1_alpha::gentoo" > [installed]) > (dependency required by "@selected" [set]) > (dependency required by "@world" [argument]) > tortoise ~ # > > # > > Current state of mind: put a live hand grenade in to the computer and > walk away. > > I've been using Gentoo Every day now for ten years. This is an > entirely > New level of bullshit. =( > > > Modest list of complete and utter FAIL: > > > > tortoise portage # tree -L 2 > . > ├── app-office > │ └── texmacs-1.99.2-r1 > ├── dev-libs > │ ├── libcdio-0.93 > │ └── libcdio-paranoia-0.93_p1 > ├── dev-qt > │ └── qtwebkit-5.5.1-r1 > ├── kde-apps > │ └── kdesdk-kioslaves-15.12.2 > ├── kde-plasma > │ ├── oxygen-5.5.5 > │ └── oxygen-fonts-5.4.3 > ├── media-gfx > │ └── fontforge-20150824 > ├── media-libs > │ ├── opencv-3.1.0-r2 > │ └── x264-0.0.20151011 > └── media-video > └── vcdimager-0.7.24 > > 19 directories, 0 files > tortoise portage # > > > Maybe you can install libreoffice-bin instead libreoffice package. And for you wayland problem only needs the gles2 use flag. About the PYTHON_SINGLE_TARGET you should read the portage news. -- William Ernesto Cárdenas Gómez signature.asc Description: This is a digitally signed message part
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 17/03/2016 22:02, Håkon Alstadheim wrote:
> On 03/17/2016 02:03 PM, Bill Kenworthy wrote:
>> On 17/03/16 20:26, Alan McKinnon wrote:
>>> On 17/03/2016 08:50, Håkon Alstadheim wrote:
I have a server SUPPOSED to be running 24/7, but every once in a while
during a prolonged absence the box will go down. The Real Time Clock
will drift, and in the rush to get the box up again I let everything
boot up automatically and get both wrong time on the main systems, and
different times on the various systems.
My setup has a main server which does NTP, but with no direct link to
the outside. Router&firewall /have/ to be booted booted later (dumb
setup, don't ask), after which I can finally get correct time from NTP.
NTP initiates "11 minute mode", which makes /etc/adjtime useless as far
as I understand. Anybody have a /correct/ way to account for RTC drift
on a box running ntpd? Right now I have a ---file in
/etc/cron.d/time-bad like so:
* * * * * root adjtimex -S 5 >/dev/null 2>&1 >>> ---
Combined with an old-fashioned setup for hwclock during boot and
shutdown. This feels really wrong, and I have no idea what I am doing.
TLDR: Anybody have a /correct/ way to account for RTC drift on a box
running ntpd?
>>>
>>> When the box was off, all questions of accurate ntp tracking are moot.
>>> ntp is designed around the idea that every second happens but from your
>>> machine's point of view they didn't happen since it was powered down.
>>>
>>> I would go the really simple route and force ntpdate to run once during
>>> boot up before ntpd is started, thereby avoiding the entire issue.
> Why can't I have proper drift information for my RTC ("bios clock") ?
> The old way ( where "system-time as set by ntp" minus "RTC time"
> gives "drift-value" written to /etc/adjtime ) used to work perfectly
> for me for several years. Is there no canonical way of getting that
> these days?
>
> My problem is that my WAN connection can not be brought up until well
> after the main server is up (stupid I know, but rearranging things
> entails a major overhaul). Thus a bios clock without drift information
> gives me a choice between ntpdate (which messes up my logs) and ntp with
> incremental adjustments (which might leave clocks wrong for several days).
>
> I really need the logs to be on the same clock for all systems. Don't
> ask, just assume I know why it's called bleeding edge . I also really
> need sub-minute accuracy on all clocks. I suppose I should try running
> ntpdate on everything once the WAN connection is up, just to see how bad
> the mess is.
I think your answer will be "the mess will be horribly bad".
Now that I understand your problem better, I don't know how to solve it.
If man pages don't provide a solution (and there must be a way to do it
surely) then James' idea of a cheap external time source sounds good.
On time sources, for the life of me I cannot understand why those in
computers suck so badly. The name-no brand one on my wrist, costing 30
bucks from a dodgy Chinese dude on the street corner, is easily accurate
to a second a month. It gets bumped, whacked a lot, immersed in water
frequently and subjected to temperatures sub-zero up to 40 deg C + in
direct African sunlight (hot enough to damage LCDs and make them bleed).
And it stays about a second a month, despite being cheap junk and
running off a shitty battery.
So why are the ones in my computer about as accurate as a sundial
without a reference? Always wondered about that.
--
Alan McKinnon
alan.mckin...@gmail.com
Re: [gentoo-user] Giving Gentoo Another Go
> On Fri, 18 March 2016, at 6:07 am, Alan McKinnon > wrote: > > … > USE flags enable and disable features of software at compile-time. Take > for example a music player. Maybe it can store the metadata about your > music in flat files, in sqlite, in mysql or postgres. Now you must make > a choice where to put the flag. Maybe your music collection is HUGE and > postgres is the best fit. > > If you add it to make.conf it becomes global and every piece of software > that supports postgres will now be rebuilt to give postgres support. > Maybe you don't need or want that. > > A flag like that is best put into package.use where it applies only to > the package you list there. So postgres gets installed, the music player > gets support and your MTA does not. To expand on this example, if `emerge -p` showed your music player had flags for mp3, mp4 and aac files, I would probably set those in /etc/make.conf, because I want all music and video players and converters to support these common file types. Stroller.
Re: [gentoo-user] Re: The war continues.
On Friday 18 Mar 2016 19:29:29 waben...@gmail.com wrote: > What I miss most of all is the fantastic konqueror. It was way better > than any other filemanager that I know. Of course I've tested the KDE4 > konqueror and also dolphin but it was horrible compared to the old > konqueror. Now I'm using thunar. It's far away from being perfect, but > it seems to be the lesser of the evils. ;-) > > -- > Regards > wabe I'm still using Konq rather than dolphin as a file manager and continue to be happy with it. :-) What's stopping you using it? -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Re: The war continues.
waben...@gmail.com wrote: > Alan Grimes wrote: > >> Philip Webb wrote: >> >>> So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). >>> Everything works for me using Fluxbox + some KDE 4 apps. >> The last good version of KDE was 3.5.x... then the flood came... (Qt >> 4) and all the developer started having Ideas about things that could >> go into the new version > Also IMHO KDE 3.5 was the best (K)DE. :-) > > But I can't envisage that KDE will ever reach that quality again. > > -- > Regards > wabe > > You two just said a mouth full and I agree. That last part of KDE3 was some good stuff. To this day, I still can't get my desktop slide show to *not* be random. I filed a bug way back when KDE3 forced folks to move to KDE4. I can see how making it random may require some work but disabling it shouldn't take to much work. Just beat the random thing until it dies. lol That's just one thing tho. KDE3 was much faster in my opinion. Dale :-) :-)
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 17/03/2016 08:50, Håkon Alstadheim wrote: I have a server SUPPOSED to be running 24/7, but every once in a while during a prolonged absence the box will go down. The Real Time Clock will drift, and in the rush to get the box up again I let everything boot up automatically and get both wrong time on the main systems, and different times on the various systems. My setup has a main server which does NTP, but with no direct link to the outside. Router&firewall /have/ to be booted booted later (dumb setup, don't ask), after which I can finally get correct time from NTP. NTP initiates "11 minute mode", which makes /etc/adjtime useless as far as I understand. Anybody have a /correct/ way to account for RTC drift on a box running ntpd? Right now I have a ---file in /etc/cron.d/time-bad like so: * * * * * root adjtimex -S 5 >/dev/null 2>&1 When the box was off, all questions of accurate ntp tracking are moot. ntp is designed around the idea that every second happens but from your machine's point of view they didn't happen since it was powered down. I would go the really simple route and force ntpdate to run once during boot up before ntpd is started, thereby avoiding the entire issue. Sometimes correctness really doesn't matter, this looks like one of those. alan
Re: [gentoo-user] The war continues.
On Thu, 17 Mar 2016 10:55:24 -0400, Alan Grimes wrote: > One problem is that every time revdep-rebuild is run, it always rebuilds > all of libreoffice, an 8-hour build. WTF, seriously, WTF? HTF can we tell you when you haven't provided the output. However, you may want to read the revdep-rebuild man page on the MASK options. > I tried to fix it by setting wayland, gles2 and egl to default because > they were breaking other packages. Wayland is still experimental, you should expect breakage when using it. > tortoise ~ # ./pretendupdate Which does what? > !!! The ebuild selected to satisfy "app-text/asciidoc" has unmet > requirements. > - app-text/asciidoc-8.6.9-r2::gentoo USE="graphviz -examples -highlight > -test" ABI_X86="64" PYTHON_SINGLE_TARGET="-pypy -python2_7" > PYTHON_TARGETS="python2_7 -pypy" > > The following REQUIRED_USE flag constraints are unsatisfied: > exactly-one-of ( python_single_target_pypy > python_single_target_python2_7 ) This one s quite clear, PYTHON_SINGLE_TARGET must contain one, and only one, of pypy or pthhon2_7. yours has neither. > I've been using Gentoo Every day now for ten years. Then you should understand the need for detailed error reports and reading the output from the various commands. > This is an entirely > New level of bullshit. =( Are you still referring to portage's output?... -- Neil Bothwick Would a fly without wings be called a walk? pgplAtMjpSrye.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 18/03/16 05:14, Alan McKinnon wrote: > On 17/03/2016 22:02, Håkon Alstadheim wrote: >> On 03/17/2016 02:03 PM, Bill Kenworthy wrote: >>> On 17/03/16 20:26, Alan McKinnon wrote: On 17/03/2016 08:50, Håkon Alstadheim wrote: > I have a server SUPPOSED to be running 24/7, but every once in a while > during a prolonged absence the box will go down. The Real Time Clock > will drift, and in the rush to get the box up again I let everything > boot up automatically and get both wrong time on the main systems, and > different times on the various systems. > > My setup has a main server which does NTP, but with no direct link to > the outside. Router&firewall /have/ to be booted booted later (dumb > setup, don't ask), after which I can finally get correct time from NTP. > > NTP initiates "11 minute mode", which makes /etc/adjtime useless as far > as I understand. Anybody have a /correct/ way to account for RTC drift > on a box running ntpd? Right now I have a ---file in > /etc/cron.d/time-bad like so: > * * * * * root adjtimex -S 5 >/dev/null 2>&1 --- > > Combined with an old-fashioned setup for hwclock during boot and > shutdown. This feels really wrong, and I have no idea what I am doing. > > TLDR: Anybody have a /correct/ way to account for RTC drift on a box > running ntpd? > > Have you looked at adjtimex ... its in portage >From the man page ... "For a standalone or intermittently connected machine, where it’s not ossible to run ntpd, you may use adjtimex instead to correct the sys-tem clock for systematic drift. There are several ways to estimate the drift rate. If your computer can be connected to the net, you might run ntpd for at least several hours and run "adjtimex --print" to learn what values of tick and freq it settled on. Alternately, you could estimate values using as a reference the CMOS clock (see the --compare and --adjust switches), another host (see --host and --review), or some other source of time (see --watch and --review). You could then add a line to rc.local invoking adjtimex, or configure /etc/init.d/adjtimex or /etc/default/adjtimex, to set those parameters each time you reboot." Used it at one time for dialup which approximates your condition. BillK
[gentoo-user] Re: local shared directory
Neil Bothwick digimed.co.uk> writes: > > On Thu, 17 Mar 2016 18:19:21 +0100, hw wrote: > > > how can I make it so that multiple users on a system who create > > files in a local, shared directory do have write access to files > > created by other users within the shared directory? > > ACLs. Hm. Perhaps one of our most knowledgable folks would like to share how to setup git locally on this machine and use it's features to 'skin the cat' vis a vis this problem. The more I learn about git the more I realize it is boundless on creative usages, so I'd be very surprise, if one of our smarter members does not know how to do this with git. The git advantage would be that if the work ever needs to span countries or continents, it should be then be trivial to migrate the repo(s) to the cloud, github or similar distributed/container environments. curiously, James
Re: [gentoo-user] Confessional: how I generally use emerge.
On 18 March 2016 01:08:26 GMT+00:00, Alan Grimes wrote: > > What have you set PYTHON_SINGLE_TARGET to? It *must* be one of: * > python2_7 * python3_4 Alec > > I had it on 3_4 but that's giving me: > > > > > tortoise ~ # ./pretendupdate > > These are the packages that would be merged, in order: > > Calculating dependencies... done! > > !!! Multiple package instances within a single package slot have been > pulled > !!! into the dependency graph, resulting in a slot conflict: > > media-libs/mesa:0 > > (media-libs/mesa-11.1.2-r1:0/0::gentoo, ebuild scheduled for merge) > pulled in by > (no parents that aren't satisfied by other packages in this slot) > > (media-libs/mesa-11.1.2-r1:0/0::gentoo, installed) pulled in by > media-libs/mesa[egl,gbm,gles2?,wayland] required by > (kde-plasma/kwin-5.5.5:5/5::gentoo, ebuild scheduled for merge) > > ^^^ > > > > > It might be possible to solve this slot collision > by applying all of the following changes: >- media-libs/mesa-11.1.2-r1 (Change USE: +gles2 +wayland) > > > !!! The ebuild selected to satisfy "app-text/asciidoc" has unmet > requirements. > - app-text/asciidoc-8.6.9-r2::gentoo USE="graphviz -examples > -highlight > -test" ABI_X86="64" PYTHON_SINGLE_TARGET="-pypy -python2_7" > PYTHON_TARGETS="python2_7 -pypy" > > The following REQUIRED_USE flag constraints are unsatisfied: > exactly-one-of ( python_single_target_pypy > python_single_target_python2_7 ) > > The above constraints are a subset of the following complete > expression: > exactly-one-of ( python_single_target_pypy > python_single_target_python2_7 ) python_single_target_pypy? ( > python_targets_pypy ) python_single_target_python2_7? ( > python_targets_python2_7 ) > > (dependency required by "net-misc/tor-0.2.8.1_alpha::gentoo" > [installed]) > (dependency required by "@selected" [set]) > (dependency required by "@world" [argument]) > tortoise ~ # > > > > > -- > IQ is a measure of how stupid you feel. > > Powers are not rights. Why ask the same question again when you got an answer last time? Hint: look at the output for asciidoc. -- Sent from my Android phone with K-9 Mail. Please excuse my brevity.
[gentoo-user] Re: The war continues.
On Sat, 19 Mar 2016 04:52:22 -0500 Dale wrote: > Neil Bothwick wrote: > > On Fri, 18 Mar 2016 21:51:24 -0500, Dale wrote: > > > The hard part, getting it to run as root. KDE doesn't like > things running as root so it took a hammer and some elbow > grease. > >>> sudo konqueror works here. > >> This works here as a desktop shortcut. > >> > >> kfmclient openProfile filemanagement > > How does that run it as root? > > When I click it, it asks for the root password and after that, it runs > as root just like the old KDE3 did. It's actually how it worked in > KDE3 since I stole it from there. That shouldn't work (and doesn't here) without some more "elbow grease and hammers" somewhere.
Re: [gentoo-user] Confessional: how I generally use emerge.
2016-03-18 14:41 GMT-03:00 Dale : > Neil Bothwick wrote: > > > > > > Why ask the same question again when you got an answer last time? > > > > Hint: look at the output for asciidoc. > > -- > > Sent from my Android phone with K-9 Mail. Please excuse my brevity. > > I have to add this. From what I understand about the scripts he is > using, he is blindly letting emerge do updates without checking to see > if the updates fall into line with what he *needs*. If I read it > correctly, any USE flag change will be missed until it hits the fan and > is broken. That would then mean taking a lot of time to go back through > logs and figuring out just when it went wrong and most importantly, what > caused it and how to fix it. Since the change could have happened > several updates ago, that could involve some work and a lot of rebuilding. > > Doing the sync in a script and even getting a email or something with > the -p output is fine but updating blindly is not a good idea. > > At this point, I think I understand why he is having so much trouble. > At the very least, check to see what USE flags are changing before doing > updates. There are other things that are important but that is one that > changes a good bit. > > Dale > > :-) :-) > > > My $0.02. Three scripts, at different cron entries: $ cat update #! /bin/bash if ! [ -e /root/.working ] then touch /root/.working LOG=/tmp/update.log LOG_TEMP=`/bin/tempfile` date > $LOG umount /usr/portage/distfiles 2> /dev/null mount /usr/portage/distfiles/ /usr/sbin/emaint sync -a 1>> $LOG 2>> $LOG && \ /usr/bin/layman -S 1>> $LOG 2>> $LOG && \ /usr/bin/emerge -fvuDN --with-bdeps=y --complete-graph=y --backtrac=100 world 1>> $LOG 2>> $LOG && \ /usr/bin/emerge -pvuDN --with-bdeps=y --complete-graph=y --backtrac=100 world 1>> $LOG_TEMP 2>> $LOG_TEMP && \ cat $LOG_TEMP >> $LOG date >> $LOG echo -e \\n >> $LOG rm -f /root/.working cat $LOG_TEMP | mail -b -c -s "update system" email_addr...@domain.com rm -f $LOG_TEMP fi $ cat emerg #! /bin/bash LOG=/tmp/update.log GREP_PATTERN="[<>=\*][<>=\*][<>=\*] [eEuUcCM(rN]" if ! [ -e /root/.working ] then date > $LOG touch /root/.working echo >> $LOG nice -n 10 emerge -vuDN --with-bdeps=y --complete-graph=y --backtrac=100 --keep-going --quiet-build world 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 revdep-rebuild -q -i -- --quiet-build --keep-going 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 emerge -vuDN --with-bdeps=y --complete-graph=y --backtrac=100 --keep-going --quiet-build world 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 revdep-rebuild -q -i -- --quiet-build --keep-going 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 /root/bin/xorg_rebuild 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 revdep-rebuild -q -i -- --quiet-build --keep-going 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 /root/bin/xorg_rebuild 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 revdep-rebuild -q -i -- --quiet-build --keep-going 1>> $LOG 2>> $LOG echo >> $LOG nice -n 10 /usr/sbin/rkhunter --propupd 1>> $LOG 2>> $LOG date >> $LOG echo -n ^D >> $LOG ; cat $LOG | mail -b -c -s "emerge altdimtzlgt002" francisco.ares.al...@gmail.com rm -f /root/.working fi if ! [ -e /root/.workhours ] then shutdown -h +1 fi $ cat bin/xorg_rebuild #! /bin/bash if ! [ -e /root/.working ] then touch /root/.working XORG_SERVER=`equery l xorg-server` if [ "$XORG_SERVER" != "`cat /root/xorg-server.txt`" ] then EMERGE_LIST=`equery l --format='$name' xf86*`" "`equery l --format='$name' "xorg*"`" "`equery l --format='$name' nvidia-drivers` # EMERGE_LIST=`equery l --format='$name' xf86*`" "`equery l --format='$name' "xorg*"` emerge -vD --with-bdeps=y --keep-going --quiet-build $EMERGE_LIST && \ echo $XORG_SERVER>/root/xorg-server.txt && \ echo $NVIDIA_DRVR>/root/nvidia-drvr.txt && \ rmmod nvidia 2>/dev/null && modprobe nvidia fi DRIVER="nvidia-drivers" MOD="nvidia" VIDEO_DRIVR=`equery l $DRIVER` if [ "$VIDEO_DRIVR" != "`cat /root/video-drv.txt`" ] then EMERGE_LIST=`equery l --format='$name' xf86*`" "`equery l --format='$name' "xorg*"`" "`equery l --format='$name' $DRIVER` emerge -vD --with-bdeps=y --keep-going --quiet-build $EMERGE_LIST && \ echo $XORG_SERVER>/root/xorg-server.txt && \ echo $VIDEO_DRIVR>/root/video-drv.txt && \ rmmod $MOD 2>/dev/null && modprobe $MOD fi rm -f /root/.working fi Cron entries: $ crontab -l # DO NOT EDIT THIS FILE - edit the master and reinstall. # (/tmp/cr
Re: [gentoo-user] Re: The war continues.
On 03/19/2016 07:56 AM, »Q« wrote: > On Sat, 19 Mar 2016 04:52:22 -0500 > Dale wrote: > >> Neil Bothwick wrote: >>> On Fri, 18 Mar 2016 21:51:24 -0500, Dale wrote: >>> >> The hard part, getting it to run as root. KDE doesn't like >> things running as root so it took a hammer and some elbow >> grease. > sudo konqueror works here. This works here as a desktop shortcut. kfmclient openProfile filemanagement >>> How does that run it as root? >> >> When I click it, it asks for the root password and after that, it runs >> as root just like the old KDE3 did. It's actually how it worked in >> KDE3 since I stole it from there. > > That shouldn't work (and doesn't here) without some more "elbow grease > and hammers" somewhere. > Using `kfmclient openProfile filemanagement` doesn't work for me (I can't browse /root), and `sudo konqueror` doesn't work (can't connect to X server.) However, what did work was `kdesu konqueror`, which asked me for the root password. When I entered it, I can browse /root and create/edit files. All were done in a terminal as a non-root user. Dan
Re: [gentoo-user] Re: dealing with distfiles bloat?
On 19/03/2016 12:47, Hans wrote: > On 07/03/16 03:38, Alan Grimes wrote: >> I can't really read the stupid unformatted du output but it looks like I >> have 30 gb of bloat in some 3,600 files in my distfiles directory. is >> there any sane way to prune out some of the older versions, I am in no >> mood to spend all day hand-pruning these and the nuclear option is not >> too friendly to the portage servers that I want to respect. >> >> >> nuclear = rm * -> emerge --fetchonly --emptytree system >> >> > > Just cleaned 9GB distfiles using 'rm /usr/portage/distfiles/*' > > > Or, you can use the gentoo-supplied script that lets you pick types of distfiles t remove. It's called eclean -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Re: Confessional: how I generally use emerge.
On 18/03/2016 20:43, »Q« wrote: > On Thu, 17 Mar 2016 20:37:04 -0400 > Alec Ten Harmsel wrote: > >>> emerge --update --newuse --deep --with-bdeps=y system --keep-going >> >> Add "--oneshot", same reasoning as above. > > When the target is a set (in this case @system), does portage ever add > all of it to @world? > > Effectively, yes. That's not what the code does of course (they go into world_sets) but the behaviour is as if the set was in world, and gets added/removed as a complete unit -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] Re: The war continues.
Daniel Frey wrote: > On 03/19/2016 07:56 AM, »Q« wrote: >> On Sat, 19 Mar 2016 04:52:22 -0500 >> Dale wrote: >> >>> Neil Bothwick wrote: On Fri, 18 Mar 2016 21:51:24 -0500, Dale wrote: >>> The hard part, getting it to run as root. KDE doesn't like >>> things running as root so it took a hammer and some elbow >>> grease. >> sudo konqueror works here. > This works here as a desktop shortcut. > > kfmclient openProfile filemanagement How does that run it as root? >>> When I click it, it asks for the root password and after that, it runs >>> as root just like the old KDE3 did. It's actually how it worked in >>> KDE3 since I stole it from there. >> That shouldn't work (and doesn't here) without some more "elbow grease >> and hammers" somewhere. >> > Using `kfmclient openProfile filemanagement` doesn't work for me (I > can't browse /root), and `sudo konqueror` doesn't work (can't connect to > X server.) > > However, what did work was `kdesu konqueror`, which asked me for the > root password. When I entered it, I can browse /root and create/edit files. > > All were done in a terminal as a non-root user. > > Dan > > > It has worked here ever since I switched to KDE4. It worked just yesterday and just for giggles, I just checked it again and it worked like it always has. As long as it works, I'll keep using it. Dale :-) :-)
[gentoo-user] Re: Confessional: how I generally use emerge.
On Fri, 18 Mar 2016 23:55:19 +0200 Alan McKinnon wrote: > On 18/03/2016 20:43, »Q« wrote: > > On Thu, 17 Mar 2016 20:37:04 -0400 > > Alec Ten Harmsel wrote: > > > >>> emerge --update --newuse --deep --with-bdeps=y system > >>> --keep-going > >> > >> Add "--oneshot", same reasoning as above. > > > > When the target is a set (in this case @system), does portage ever > > add all of it to @world? > > Effectively, yes. That's not what the code does of course (they go > into world_sets) but the behaviour is as if the set was in world, and > gets added/removed as a complete unit Hmm, that doesn't match my experience. I just tested with the smallest set I ever use, @module-rebuild : # emerge @module-rebuild Calculating dependencies... done! >>> Verifying ebuild manifests >>> Emerging (1 of 1) app-emulation/virtualbox-modules-4.3.32::gentoo >>> Installing (1 of 1) app-emulation/virtualbox-modules-4.3.32::gentoo >>> Jobs: 1 of 1 complete Load avg: 2.04, >>> 0.80, 0.52 Auto-cleaning packages... >>> No outdated packages were found on your system. * GNU info directory index is up-to-date. # cat /var/lib/portage/world_sets # file /var/lib/portage/world_sets /var/lib/portage/world_sets: empty I don't think I have anything in make.conf which would change the default behavior WRT world_sets, but here's what I have anyway: FEATURES="binpkg-logs buildsyspkg collision-protect downgrade-backup fail-clean fixlafiles news parallel-fetch parallel-install preserve-libs sandbox strict unknown-features-warn userfetch userpriv usersandbox usersync" EMERGE_DEFAULT_OPTS="--ask-enter-invalid --jobs=8 --load-average 11.2 --with-bdeps y"
Re: [gentoo-user] Re: The war continues.
On Sat, 19 Mar 2016 08:24:51 -0700, Daniel Frey wrote: > Using `kfmclient openProfile filemanagement` doesn't work for me (I > can't browse /root), and `sudo konqueror` doesn't work (can't connect to > X server.) Probably an environment variable not surviving the transition to root, try sudo -E konqueror. > However, what did work was `kdesu konqueror`, which asked me for the > root password. When I entered it, I can browse /root and create/edit > files. kdesu and kdesudo take care of the fiddly bits involved in running X programs as a different user (as does sux). -- Neil Bothwick DATA COMPRESSION: What You Get When You Squish An Android pgpB9nJVpmZnG.pgp Description: OpenPGP digital signature
Re: [gentoo-user] Re: Confessional: how I generally use emerge.
On 19/03/2016 18:43, »Q« wrote: > On Fri, 18 Mar 2016 23:55:19 +0200 > Alan McKinnon wrote: > >> On 18/03/2016 20:43, »Q« wrote: >>> On Thu, 17 Mar 2016 20:37:04 -0400 >>> Alec Ten Harmsel wrote: >>> > emerge --update --newuse --deep --with-bdeps=y system > --keep-going Add "--oneshot", same reasoning as above. >>> >>> When the target is a set (in this case @system), does portage ever >>> add all of it to @world? >> >> Effectively, yes. That's not what the code does of course (they go >> into world_sets) but the behaviour is as if the set was in world, and >> gets added/removed as a complete unit > > Hmm, that doesn't match my experience. I just tested with the smallest > set I ever use, @module-rebuild : > > # emerge @module-rebuild > Calculating dependencies... done! Verifying ebuild manifests Emerging (1 of 1) app-emulation/virtualbox-modules-4.3.32::gentoo Installing (1 of 1) app-emulation/virtualbox-modules-4.3.32::gentoo Jobs: 1 of 1 complete Load avg: 2.04, 0.80, 0.52 Auto-cleaning packages... > No outdated packages were found on your system. > > * GNU info directory index is up-to-date. > # cat /var/lib/portage/world_sets > # file /var/lib/portage/world_sets > /var/lib/portage/world_sets: empty > > I don't think I have anything in make.conf which would change the > default behavior WRT world_sets, but here's what I have anyway: > > FEATURES="binpkg-logs buildsyspkg collision-protect downgrade-backup > fail-clean fixlafiles news parallel-fetch parallel-install > preserve-libs sandbox strict unknown-features-warn userfetch > userpriv usersandbox usersync" > > EMERGE_DEFAULT_OPTS="--ask-enter-invalid --jobs=8 --load-average 11.2 > --with-bdeps y" > > > @module-rebuild is a dynamic set. It translates to "all the packages you have emerged that install out-of-tree kernel modules" So not really a fair comparison. Compare instead against a regular static set - "a bunch of packages defined by you that go together and live in /etc/portage/sets/" -- Alan McKinnon alan.mckin...@gmail.com
[gentoo-user] Simply approach:: Re: The war continues.
Alan Grimes verizon.net> writes: > My effort to update my syestem continues unabated. =\ I've had a few F'ers in my years of gentoo > One problem is that every time revdep-rebuild is run, it always rebuilds > all of libreoffice, an 8-hour build. WTF, seriously, WTF? No one tool or one single approach is fool_proof with gentoo. > I made the mistake of syncing portage again and was thrown back into > useflag hell: OK, not syncing may be the best option, but not if you've getting some corrupt packages. (checksums?) > I tried to fix it by setting wayland, gles2 and egl to default because > they were breaking other packages. OK, since you are the "major surgery point" here is what I suggest. Copy your /var/lib/portage/world file someplace for safe keeping. Eliminate all packages that cause you any problems, unless they are critical @system packages or fundamental to your desktop. If they bark, emerge -C those complainers, until you can get a stable system. If you have other gentoo systems, distcc can be your friend. If not, then it takes longer. > I don't even know how to read the current error message: use 'elogv' to see the error messages in addition to the log files. You do have make.conf configured to save log files, right? Also the 'q' applets, a collection of c admin code with vapier's paw_prints all over them, are excellent and more commonly know as portage-utils :: make sure you have them install and have tried out the ones that are useful for diagnosing your issues. (q --help) for a quick reference. So, by drastically pruning the system, WE can focus on those error messages, on your critical system needs. Once that is stable, then add back the complaining packages, one at a time, perhaps overnight on a long update. This is the semantics I use to fix severely borked systems. You run gentoo long enough, and experiment you are going to bork a few systems. It's kind of a 'badge of honor' with gentoo, so relax and enjoy the ride. Perhaps a tasty beverage, pleasing to your palette would allow you to calm down a bit? h
Re: [gentoo-user] Re: The war continues.
Alan Grimes wrote: > Philip Webb wrote: > > > So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). > > Everything works for me using Fluxbox + some KDE 4 apps. > > The last good version of KDE was 3.5.x... then the flood came... (Qt > 4) and all the developer started having Ideas about things that could > go into the new version Also IMHO KDE 3.5 was the best (K)DE. :-) But I can't envisage that KDE will ever reach that quality again. -- Regards wabe
[gentoo-user] Re: Giving Gentoo Another Go
Neil Bothwick digimed.co.uk> writes: > > On Fri, 18 Mar 2016 10:09:14 +, Peter Humphrey wrote: > > > > You will develop your way of doing things over time, and that way > > > could change as your needs do. Using your example of package.use, > > > moving USE flags from package.use to make.conf is an easy enough task > > > if you need to change. I tend to put them n package.use to start with > > > then migrate to make.conf if I find I am using the same flag on > > > several packages. The entire /etc/make.conf directory is parses, so you can take package.use and make a dir out of it and then logically organize your flags into several directories, once a system get's large and complex. > > A simple way to start off is to see whether the USE flag is listed in > > /usr/portage/profiles/use.desc or use.local.desc. If the former, it's > > likely to affect many packages in a typical system so put it in > > make.conf; if the latter, it's likely to affect only a few of your > > packages so put it in package.use. You can always move it later if you > > want to, as Neil says. > app-portage/euses is an easy way of looking up USE flags, give it the > name of a flag and it shows you the description. If it shows one or more > package names, the USE flag is defined in local.desc. All good information. The exciting thing happening in Gentoo right now, is some of the devs are promoting the concept of 'lazy flags'. This basically means some new and additional features will be added to portage or the Packaage Management system (portage, paludis, etc) where additional user defined logic will 'automagically' make default and necessary modifications to flag configurations, and the user just reviews those 'auto-enhancements' or something like that. Gentoo never stops innovating, but the caveat is you have to be patient and invest of yourself into learning Gentoo. Gentoo is an addiction, which most of us are quite happy with. Gentoo also has legendary status with many of the brightest minds in computer science, for a myriad of valid reasons. Gentoo is something that is wonderful to be a part of and is an 'honor_badge' of fortitude because one can deeply learn about linux, software and a host of relevant technologies quite readily in a Gentoo environment. Gentoo's future is very bright, unique and most rewarding. Gentoo is my pal and my best friend and what I use to earn money. hth, James
Re: [gentoo-user] Giving Gentoo Another Go
On Thu, 17 Mar 2016 23:03:47 -0400, Hunter Jozwiak wrote: > After talking to a few diehard Gentoo fans at my local LUG, I decided I > would like to give Gentoo another shot. Are there any good books that > can supplement the Gentoo handbook The Gentoo handbook really is the book. It's written by Gentoo devs so it is always up to date, something that cannot always be said of print books (a friend of mine wrote the Haynes Manual on Ubuntu some years ago, few months later they released Unity!). > as well as books that go more in > depth than the Gentoo chapter on Portage? The main mistake that people make with the Gentoo Handbook is that they follow it carefully through installation, then stop. They have only read chapter 1. The rest of the handbook, along with the Gentoo Wiki provide a lot more information. > One of the main issues I > faced with Gentoo when I first tried it is that I did not understand > the power of package.use, and I put everything in to make.conf. No one gets it right first time, there are many choices and many ways of doing things. No book can tell you which way is right for you, only experience can do that. Getting things like this wrong is a natural part of the Gentoo learning experience. You will develop your way of doing things over time, and that way could change as your needs do. Using your example of package.use, moving USE flags from package.use to make.conf is an easy enough task if you need to change. I tend to put them n package.use to start with then migrate to make.conf if I find I am using the same flag on several packages. There's also the choice of whether you make package.use (and its friends) and make.conf a single file or a directory of files. Ask in here and you will find proponents of both approaches, but it's an organisational choice, whichever works for you is best. -- Neil Bothwick "Mmmm, trouble with grammer have I, yes?" - Yoda pgpf4qJdxP9x5.pgp Description: OpenPGP digital signature
[gentoo-user] Re: Confessional: how I generally use emerge.
On Sat, 19 Mar 2016 19:03:24 +0200 Alan McKinnon wrote: > On 19/03/2016 18:43, »Q« wrote: > > On Fri, 18 Mar 2016 23:55:19 +0200 > > Alan McKinnon wrote: > > > >> On 18/03/2016 20:43, »Q« wrote: > >>> On Thu, 17 Mar 2016 20:37:04 -0400 > >>> Alec Ten Harmsel wrote: > >>> > > emerge --update --newuse --deep --with-bdeps=y system > > --keep-going > > Add "--oneshot", same reasoning as above. > >>> > >>> When the target is a set (in this case @system), does portage ever > >>> add all of it to @world? [big snip] > @module-rebuild is a dynamic set. It translates to "all the packages > you have emerged that install out-of-tree kernel modules" > > So not really a fair comparison. Compare instead against a regular > static set - "a bunch of packages defined by you that go together and > live in /etc/portage/sets/" Thanks much for the lesson. I may hijack more of the OP's threads to ask about trivia. :) $ sudo emerge --noreplace @testset Calculating dependencies... done! >>> Recording @testset in "world_sets" favorites file... I had also been under the mistaken impression that --update implied --oneshot, but I see that it's not so.
Re: [gentoo-user] Confessional: how I generally use emerge.
> What have you set PYTHON_SINGLE_TARGET to? It *must* be one of: * python2_7 * python3_4 Alec I had it on 3_4 but that's giving me: tortoise ~ # ./pretendupdate These are the packages that would be merged, in order: Calculating dependencies... done! !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: media-libs/mesa:0 (media-libs/mesa-11.1.2-r1:0/0::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (media-libs/mesa-11.1.2-r1:0/0::gentoo, installed) pulled in by media-libs/mesa[egl,gbm,gles2?,wayland] required by (kde-plasma/kwin-5.5.5:5/5::gentoo, ebuild scheduled for merge) ^^^ It might be possible to solve this slot collision by applying all of the following changes: - media-libs/mesa-11.1.2-r1 (Change USE: +gles2 +wayland) !!! The ebuild selected to satisfy "app-text/asciidoc" has unmet requirements. - app-text/asciidoc-8.6.9-r2::gentoo USE="graphviz -examples -highlight -test" ABI_X86="64" PYTHON_SINGLE_TARGET="-pypy -python2_7" PYTHON_TARGETS="python2_7 -pypy" The following REQUIRED_USE flag constraints are unsatisfied: exactly-one-of ( python_single_target_pypy python_single_target_python2_7 ) The above constraints are a subset of the following complete expression: exactly-one-of ( python_single_target_pypy python_single_target_python2_7 ) python_single_target_pypy? ( python_targets_pypy ) python_single_target_python2_7? ( python_targets_python2_7 ) (dependency required by "net-misc/tor-0.2.8.1_alpha::gentoo" [installed]) (dependency required by "@selected" [set]) (dependency required by "@world" [argument]) tortoise ~ # -- IQ is a measure of how stupid you feel. Powers are not rights.
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 17/03/16 20:26, Alan McKinnon wrote: > On 17/03/2016 08:50, Håkon Alstadheim wrote: >> I have a server SUPPOSED to be running 24/7, but every once in a while >> during a prolonged absence the box will go down. The Real Time Clock >> will drift, and in the rush to get the box up again I let everything >> boot up automatically and get both wrong time on the main systems, and >> different times on the various systems. >> >> My setup has a main server which does NTP, but with no direct link to >> the outside. Router&firewall /have/ to be booted booted later (dumb >> setup, don't ask), after which I can finally get correct time from NTP. >> >> NTP initiates "11 minute mode", which makes /etc/adjtime useless as far >> as I understand. Anybody have a /correct/ way to account for RTC drift >> on a box running ntpd? Right now I have a ---file in >> /etc/cron.d/time-bad like so: >> * * * * * root adjtimex -S 5 >/dev/null 2>&1 > --- >> >> Combined with an old-fashioned setup for hwclock during boot and >> shutdown. This feels really wrong, and I have no idea what I am doing. >> >> TLDR: Anybody have a /correct/ way to account for RTC drift on a box >> running ntpd? >> >> > > > When the box was off, all questions of accurate ntp tracking are moot. > ntp is designed around the idea that every second happens but from your > machine's point of view they didn't happen since it was powered down. > > I would go the really simple route and force ntpdate to run once during > boot up before ntpd is started, thereby avoiding the entire issue. > Sometimes correctness really doesn't matter, this looks like one of those. > > > alan > add a cheap gps setup as the reference clock to the server, or even better is a dedicated time server (either a real one or a raspberry pi/gps) on the network if you have internal connectivity. Going super cheap, but not quite as accurate for me was an arduino and rtc on a bluetooth pan for when the network was down but I needed a reference (to power up the real server :). google "arduino time server" for plenty of options :) BillK
Re: [gentoo-user] Re: The war continues.
On Fri, 18 Mar 2016 13:49:44 -0500, Dale wrote: > [ebuild R] kde-apps/konqueror-15.08.3:4/15.08::gentoo > USE="bookmarks handbook svg (-aqua) -debug" 0 KiB > > The hard part, getting it to run as root. KDE doesn't like things > running as root so it took a hammer and some elbow grease. sudo konqueror works here. -- Neil Bothwick "Theory and practice are the same in theory, but different in practice" pgp4U_qczQrzK.pgp Description: OpenPGP digital signature
Re: [gentoo-user] local shared directory
On 03/17/2016 06:38 PM, Rich Freeman wrote: > On Thu, Mar 17, 2016 at 4:59 PM, Alan McKinnon > wrote: > > Actually, this is completely viable... > > If users chmod a file then tell them not to. If you must, set up some > cron job to clean up after them. > > But, you can of course do this with ACLs as well. I haven't tried > setting those up personally. > I missed the beginning of this thread, but I just caught up on the archive. This has long been a pet peeve of mine. I don't think there's a way to make it work *at all* on Linux, which is stupid, since every somebody's-nephew can set it up in five minutes on a Windows server. You can very easily come up with a situation that umasks, group membership, and setgid can't handle. Suppose you want a public website directory to be, * Writable by the client (their developers) * Writable by your web developers * Readable by the Apache user You can't make Apache a member of the group that has write access, so while I haven't been real careful, I don't think you can make that extremely common situation work. Every law office (attorney/paralegal/secretary) and small business needs something similar and it just can't be done. ACLs also won't work, because nobody ever made default ACLs do the right thing. Everything in the "acl" directory should be rwx by the "apache" user below (that's what the setfacl does): $ mkdir acl $ cd acl $ setfacl -d -m user:apache:rwx . But, it's not! Just copy any file in, and see what happens: $ cp /etc/profile ./ $ getfacl profile # file: profile # owner: mjo # group: mjo user::rw- user:apache:rwx# effective:r-- group::r-x # effective:r-- mask::r-- other::r-- The write and execute bits are masked, so your website crashes, because Apache can't write that file (or traverse it, if we did the same experiment with a directory). The problem above is that most common tools will do something braindead in the presence of ACLs, and attempt to preserve the existing group bits. Even though, when there are ACLs around, those group bits don't signify group permissions. To make ACLs do the right thing, you need to run sys-apps/apply-default-acl on every file that the users create, so that the default ACLs get applied by default (craaazzzyyy). You can do that in a cron job like Alan suggested, or I've hacked tar, cp, mkdir, etc. to run it automatically on all of our servers. Why do I need to hack coreutils to share a directory between three people? The ACL/coreutils people don't really see this as a problem. They say, tell your paralegal to RTFM and set the permissions how he wants them. (It will take you about a week to read the man pages for ACLs.)
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On 18/03/16 05:59, Bill Kenworthy wrote: > On 18/03/16 05:14, Alan McKinnon wrote: >> On 17/03/2016 22:02, Håkon Alstadheim wrote: >>> On 03/17/2016 02:03 PM, Bill Kenworthy wrote: On 17/03/16 20:26, Alan McKinnon wrote: > On 17/03/2016 08:50, Håkon Alstadheim wrote: >> I have a server SUPPOSED to be running 24/7, but every once in a while >> during a prolonged absence the box will go down. The Real Time Clock >> will drift, and in the rush to get the box up again I let everything >> boot up automatically and get both wrong time on the main systems, and >> different times on the various systems. >> >> My setup has a main server which does NTP, but with no direct link to >> the outside. Router&firewall /have/ to be booted booted later (dumb >> setup, don't ask), after which I can finally get correct time from NTP. >> >> NTP initiates "11 minute mode", which makes /etc/adjtime useless as far >> as I understand. Anybody have a /correct/ way to account for RTC drift >> on a box running ntpd? Right now I have a ---file in >> /etc/cron.d/time-bad like so: >> * * * * * root adjtimex -S 5 >/dev/null 2>&1 > --- >> >> Combined with an old-fashioned setup for hwclock during boot and >> shutdown. This feels really wrong, and I have no idea what I am doing. >> >> TLDR: Anybody have a /correct/ way to account for RTC drift on a box >> running ntpd? >> >> > > > Have you looked at adjtimex ... its in portage > > > From the man page ... > "For a standalone or intermittently connected machine, where it’s not > ossible to run ntpd, you may use adjtimex instead to correct the sys-tem > clock for systematic drift. > >There are several ways to estimate the drift rate. If your > computer can be connected to the net, you might run ntpd for at least > several hours and run "adjtimex --print" to learn what values of tick > and freq it settled on. Alternately, you could estimate values using as > a reference the CMOS clock (see the --compare and --adjust switches), > another host (see --host and --review), or some other source of time > (see --watch and --review). You could then add a line to rc.local > invoking adjtimex, or configure /etc/init.d/adjtimex or > /etc/default/adjtimex, to set those parameters each time you reboot." > > Used it at one time for dialup which approximates your condition. > > BillK > > forget it ... I forgot that's where you started from ... must be getting old :(
Re: [gentoo-user] The war continues.
Then you should understand the need for detailed error reports and reading the output from the various commands. I always set verbosity to very high or max... Back in the good old days there was this OS called DOS. All of it's commands told you what they were and what they were doing by default, and were, almost universally quite enjoyable to use. =\ These days you just get the next command prompt, and have no idea how much your previous command might have deleted or otherwise screwed up. -- IQ is a measure of how stupid you feel. Powers are not rights.
Re: [gentoo-user] Getting a valid /etc/adjtime while using ntpd ?
On Friday 18 Mar 2016 06:01:17 Bill Kenworthy wrote: > On 18/03/16 05:59, Bill Kenworthy wrote: > > On 18/03/16 05:14, Alan McKinnon wrote: > >> On 17/03/2016 22:02, Håkon Alstadheim wrote: > >>> On 03/17/2016 02:03 PM, Bill Kenworthy wrote: > On 17/03/16 20:26, Alan McKinnon wrote: > > On 17/03/2016 08:50, Håkon Alstadheim wrote: > >> I have a server SUPPOSED to be running 24/7, but every once in a > >> while > >> during a prolonged absence the box will go down. The Real Time Clock > >> will drift, and in the rush to get the box up again I let everything > >> boot up automatically and get both wrong time on the main systems, > >> and > >> different times on the various systems. > >> > >> My setup has a main server which does NTP, but with no direct link to > >> the outside. Router&firewall /have/ to be booted booted later (dumb > >> setup, don't ask), after which I can finally get correct time from > >> NTP. > >> > >> NTP initiates "11 minute mode", which makes /etc/adjtime useless as > >> far > >> as I understand. Anybody have a /correct/ way to account for RTC > >> drift > >> on a box running ntpd? Right now I have a ---file in > >> /etc/cron.d/time-bad like so: > >> * * * * * root adjtimex -S 5 >/dev/null 2>&1 >> --- > >> > >> Combined with an old-fashioned setup for hwclock during boot and > >> shutdown. This feels really wrong, and I have no idea what I am > >> doing. > >> > >> TLDR: Anybody have a /correct/ way to account for RTC drift on a box > >> running ntpd? > > > > Have you looked at adjtimex ... its in portage > > > > > > From the man page ... > > "For a standalone or intermittently connected machine, where it’s not > > ossible to run ntpd, you may use adjtimex instead to correct the sys-tem > > clock for systematic drift. > > > >There are several ways to estimate the drift rate. If your > > > > computer can be connected to the net, you might run ntpd for at least > > several hours and run "adjtimex --print" to learn what values of tick > > and freq it settled on. Alternately, you could estimate values using as > > a reference the CMOS clock (see the --compare and --adjust switches), > > another host (see --host and --review), or some other source of time > > (see --watch and --review). You could then add a line to rc.local > > invoking adjtimex, or configure /etc/init.d/adjtimex or > > /etc/default/adjtimex, to set those parameters each time you reboot." > > > > Used it at one time for dialup which approximates your condition. > > > > BillK > > forget it ... I forgot that's where you started from ... must be getting > old :( Nobody mentioned net-misc/chrony. Would it be more appropriate for this use case? -- Regards, Mick signature.asc Description: This is a digitally signed message part.
[gentoo-user] local shared directory
Hi, how can I make it so that multiple users on a system who create files in a local, shared directory do have write access to files created by other users within the shared directory? The directory is group-writeable, and the users belong to the group which owns the directory. This enables them to create files within the shared directory, yet the files they create belong to the user who created it, and other users cannot modify them. The sticky bit is set so that the files are owned by user:common-group. I would like to avoid changing the umask. If that cannot be avoided, how do I change it? Users log in through x2goclient, and fvwm is being executed on login.
Re: [gentoo-user] Re: The war continues.
Mick wrote: > On Friday 18 Mar 2016 19:29:29 waben...@gmail.com wrote: > > > What I miss most of all is the fantastic konqueror. It was way > > better than any other filemanager that I know. Of course I've > > tested the KDE4 konqueror and also dolphin but it was horrible > > compared to the old konqueror. Now I'm using thunar. It's far away > > from being perfect, but it seems to be the lesser of the evils. ;-) > > > > -- > > Regards > > wabe > > I'm still using Konq rather than dolphin as a file manager and > continue to be happy with it. :-) > > What's stopping you using it? It's some years ago and when I'm honest, I can't remember exactly what displeased me. :-) IIRC one thing was that the detailed view mode was not as compact as it was with konqueror3. And there were some problems with the theme (missing icons and some other glitches). And IIRC I also missed some functions. Maybe I will try it again with KDE5. But this also depends on the dependencies. :-) I'm not sure if I'm willing to install the complete KDE environment for this test. -- Regards wabe
[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.
Re: [gentoo-user] Re: The war continues.
Philip Webb wrote: > So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). > Everything works for me using Fluxbox + some KDE 4 apps. The last good version of KDE was 3.5.x... then the flood came... (Qt 4) and all the developer started having Ideas about things that could go into the new version -- IQ is a measure of how stupid you feel. Powers are not rights.
Re: [gentoo-user] Re: The war continues.
160318 waben...@gmail.com wrote: > Alan Grimes wrote: >> Philip Webb wrote: >>> So that needs rephrasing : "Why is AG using KDE 5 ? " (smile). >>> Everything works for me using Fluxbox + some KDE 4 apps. >> The last good version of KDE was 3.5.x. Then the flood came (Qt 4) My conclusion at the time was that my current set-up is slightly better : Fluxbox provides the same basics & is simpler than the KDE 3 environment, while the KDE apps I use (mainly Konsole Gwenview Okular) are better in 4 . However, there are 2 small items I really miss : Kmahjongg Kworldclock . I had them in my previous machine (now stand-by), but they need a library which I can't install anymore, so I haven't tried to recreate them in my new machine. (Yes, I know there is Kmahjongg-4 , but it doesn't do "removed tiles"). > waben...@gmail.com wrote: >> What I miss most of all is the fantastic konqueror. >> It was way better than any other filemanager that I know. I use Krusader & recommend it very highly for heavy file-lifting : have you tried it ? -- if not, do : you may well like it. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Re: Confessional: how I generally use emerge.
»Q« wrote: > On Sat, 19 Mar 2016 19:03:24 +0200 > Alan McKinnon wrote: > Recording @testset in "world_sets" favorites file... > I had also been under the mistaken impression that --update > implied --oneshot, but I see that it's not so. > > > > Don't worry, a lot of us learned that the hard way. Dale :-) :-)
[gentoo-user] Fastest way to get an upstream kernel bug fixed?
I've done the easy part already: I git-bisected the guilty commit. I don't remember how to file a credible kernel bug report upstream so I hope to coax a gentoo dev into filing one for me :)
