Re: [gentoo-user] NFS Server Tuning
On 1/24/08, Stefan Onken [EMAIL PROTECTED] wrote: Any ideas how to improve the speed ? noatime? -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] emerge -avC cdrkit emerge -av cdrtools
On 7/7/08, Joerg Schilling [EMAIL PROTECTED] wrote: Let me quote _this_ file to verify that there is a 4 clause BSDL. Jörg, there are indeed four asterisks/clauses to count. But which clause represents the original GPL-incompatible advertising clause? IANAL, but I cannot see that clause in there. What I think they've done is they've made their yet another own fork of the BSDL by slicing the last clause of the 3-clause BSDL into two and sprinkling some Apache-licensisch do not smear original name stuff in there. Stupid, possibly, confusing, certainly, but most likely *not* the original 4-clause BSDL. -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
RE: [gentoo-user] shared portage tree
In this particular case I want to use a machine with arc=amd64 as a nfs server and share its tree with several x86 machines. I ran a similar NFS setup couple years back when breaking my teeth with Gentoo. Had intermittent problems with the NFS locks, and speed was slow too, so gave up. Though, it might've been just me and my poor NFS setup skills, or some settings in my kernel. From other people's comments and links I noticed they now have some helpful scripts. Wish I had seen them earlier or been able to type out some of my own. :) I switched the server over to providing local Gentoo rsync mirroring for the LAN instead: http://gentoo-wiki.com/HOWTO_Local_Rsync_Mirror I combine this with Squid (proxy) and a few gigs of proxy disk cache on the server end to have the server perform as a complete buffer/cache between my LAN boxen and the external Interweb world. And it doubles as the proxy for regular browsing, too. HTH. YMMV. HAND. -- arttu v. _ Connect to the next generation of MSN Messenger http://imagine-msn.com/messenger/launch80/default.aspx?locale=en-ussource=wlmailtagline-- [EMAIL PROTECTED] mailing list
RE: [gentoo-user] Re: Emerge firefox 2.0.0.8 and now printing is hosed
I'm using mozilla-firefox-2.0.0.8, not -bin, and printing works fine here also. Just another data point; sorry I don't know how to help I might have missed it, but was revdep-rebuild tried already? :) Also, maybe it could be something with gtk+ or some printing component (old cups? cups ebuild was just updated) instead of firefox? So, how about other apps and their print dialogs? Are they broken too? -- arttuv _ Connect to the next generation of MSN Messenger http://imagine-msn.com/messenger/launch80/default.aspx?locale=en-ussource=wlmailtagline-- [EMAIL PROTECTED] mailing list
RE: [gentoo-user] brlcad users here?
From sourceforge.net I downloaded brlcad_7.10.4_ia32.tar.bz2 and untarred it. I'm not a BRL-CAD (or any CAD for that matter ;) ) user, but take a look at the file name you're downloading. It's probably a pre-compiled binary, so that's probably why there is no configure either? The source is likely to be in another tar.bz2 file on sourceforge download page if that is what you're looking for. In fact there are no configure scripts anywhere under /usr/brlcad Anybody else had this problem and was able to overcome it? Have you tried the ebuild they have been working on over here: http://bugs.gentoo.org/show_bug.cgi?id=77197 (It barely compiles on my amd64 box, but doesn't draw anything except the menus -- and that only after I manually tell it where the ITCL_LIBRARY and ITK_LIBRARY are. On x86 they seem to have it working better.) -- arttuv _ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/-- [EMAIL PROTECTED] mailing list
RE: [gentoo-user] brlcad users here?--RESOLVED
Date: Tue, 18 Dec 2007 11:33:52 -0800 From: [EMAIL PROTECTED] Just to clarify: What does a 'pre-compiled binary' do? I meant just regular binary files, which have been compiled -- as opposed to the source files (non-compiled ;) ). Sorry for being unclear. :) -- Arttu V. _ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/-- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] mysql 5.1
On 1/14/09, Marcin Niśkiewicz mniskiew...@gmail.com wrote: Hello Is it known when mysql version 5.1 will be available in portage for amd64 (stable)? I don't know if we on the *user* list really can know about the devs' decisions. Maybe gentoo-dev list or irc channel could give better answers? AFAIK the devs rarely reply on this list. I found one comment which appears to be quite insightful of the situation (I don't use mysql that much, so I have to take this at face value), although it doesn't give exact time frames: http://bugs.gentoo.org/show_bug.cgi?id=194561#c25 Eating your data is bad for a database and if that was the situation a bit over a month ago, I wouldn't hold my breath for a stable marked ebuild in portage for a while. -- Arttu V.
Re: [gentoo-user] shoutcast problem 64bit gentoo
On 1/23/09, Xav' x...@linuxant.fr wrote: This is a joke, isn't it ??? the error message coming from rc script is a bit clear IMHO... Well, the error message is clear and that listing from someone's home directory has me really puzzled -- what's the connection, why present it? But maybe he has an /opt he has unmounted since he did emerge shoutcast-server-bin? I do installs the way that I only boot at the very end of compilations -- when it's really necessary to do so. So maybe he installed everything -- and forgot /opt from fstab or marked it not to be mounted automagically? Or maybe he didn't emerge shoutcast at all, but did something weird with instructions for some non-gentoo installations? Anyway, to be of any assistance (and I don't even know or run shoutcast myself) Mr Daletka should give some more details on how exactly he got into the situation he is in? How did he install shoutcast, etc? Why are there binaries in someone's home directory? (And why is he looking at them instead of the ones that should be where the init script is looking for?) -- Arttu V.
Re: [gentoo-user] Apply a patch to an emerge from the command line
On 1/25/09, Grant emailgr...@gmail.com wrote: I could have sworn there was a way to specify a patch to be applied to an emerge from the command line, something like: EPATCH=file.patch emerge packagename I've been searching Google and the mailing list but I can't find mention of it anywhere. Was it a figment of my imagination? This would be nice, but I don't think you can do this, at least not this way. You can grep for EPATCH_* under /usr/portage to see some samples of how EPATCH_* stuff seems to work, especially some php eclasses give nice examples on one line. Still, I think in practice it is for ebuilds' and eclasses' internal use only -- and even then only for ebuilds/eclasses which inherit the eutils.class which in turn implements epatch() which actually does anything with the values. Inside an ebuild you can set up those variables in the right spot (e.g., right in the middle of src_unpack() after unpacking the source) and immediately call epatch(). Coming from command line there probably won't be that dangling epatch() call waiting for your parameters in the right spot -- and other epatch calls or setting those variables within the ebuild may interfere or override the settings you provided depending on the case. Still, I'd be happy to be mistaken, this would make my life easier as well. -- Arttu V.
Re: [gentoo-user] Sonic-Visualiser compile error
On 1/31/09, meino.cra...@gmx.de meino.cra...@gmx.de wrote: I tried to emerge sonic-visualiser (1.4) and it fails to compile From the logfile (attached): Project MESSAGE: WARNING: Failed to find pkg-config package redland Project MESSAGE: Using pkg-config package rasqal with version 0.9.10 Project ERROR: Redland RDF datastore required But emerge -pv librdf gaves me: Calculating dependencies... done! [ebuild R ] media-libs/liblrdf-0.4.0 0 kB and emerge -pv redland [ebuild R ] dev-libs/redland-1.0.4 USE=berkdb ssl -mysql -sqlite -threads 0 kB What is missing here, what fails and how can I fix it ??? :) # pkg-config --libs redland Package 'Redland' requires 'rasqal = 0.9.12' but version of Rasqal RDF Query Library is 0.9.10 Something eats that error message? It looks like you need the unstable rasqal-0.9.15 or .16 to get it compiling. -- Arttu V.
Re: [gentoo-user] Sonic-Visualiser compile error
Mr Cramer, you replied in private email, but your blacklist bounced/dropped my consequent reply. Replying on-list: On Saturday 31 January 2009 18:07:02 you wrote: But emerge tolds me: emerge -pv rasqual These are the packages that would be merged, in order: Calculating dependencies... done! emerge: there are no ebuilds to satisfy rasqual. There is a typo: just drop the extra 'u'. It's just rasqal, like you used it correctly here below: while qsearch says: qsearch rasqal dev-libs/rasqal library that handles Resource Description Framework (RDF) Before I screw up my system: Is there something more fundamental damaged? This is the spot where I think FEATURES=buildpkg should be mentioned. It will help you roll back the old binaries should something bad happen. It is mentioned on the list regularly by many posters, most of them more experienced with it than I am. :) Still, I don't think upgrading rasqal should cause major havoc. Worst thing that happened to this box was that openoffice and soprano wanted to be re-emerged. So just some automated extra compiling, no big deal for me. YMMV Or does rasqal only exists as unstable ebuild (reading your posting I would tend to answer this with NO ... 0.9.10 has been stable on Gentoo for nearly three years according to /usr/portage/dev-libs/rasqal/ChangeLog The newer versions have had some bugs reported, but they have been fixed. You can look them up in bugs.gentoo.org if you're worried. -- Arttu V.
Re: [gentoo-user] Sonic-Visualiser compile error
meino.cra...@gmx.de wrote: Now the build succeeds. Nevertheless, starting sonic-visualisers complains: sonic-visualiser: error while loading shared libraries: libvamp-hostsdk.so.2: cannot open shared object file: No such file or directory I cannot find libvamp-hostsdk. Neither emergeing libvamp-hostsdk succeeds (no such package/ebuild) nor qsearching libvamp-hostsdk or libvamp or hostsdk produces anything which I can use to detect the correct package to emerge... What you are looking for is media-libs/vamp-plugin-sdk-2.0, but I wonder why it cannot find it since it's supposedly a dependency in the sonic-visualiser ebuild. What version of vamp-plugin-sdk do you have installed? (emerge -pvC vamp-plugin-sdk) What use flags are you using for sonic-visualiser? (emerge -pv sonic-visualiser) Is your system otherwise sane and up-to-date? (E.g., have emerge @preserved-libs revdep-rebuild -i been run successfully after upgrading the libraries?) Just as an example from a system without sonic-visualiser: when I try emerging sonic-visualiser on an amd64 system without many of the libraries it needs, I have to allow at least following packages that are marked unstable on amd64 (i.e., ~amd64): dev-libs/rasqal-0.9.16 (or 0.9.15) dev-libs/redland-1.0.8 (an earlier one might suffice, but I went with the latest available in portage) media-libs/liboggz-0.9.8 media-libs/vamp-plugin-sdk-2.0 media-libs/libfishsound-0.9.1 And also naturally the sonic-visualiser-1.4 itself, which is also unstable. -- Arttu V.
Re: [gentoo-user] KDE4.2 compile problem
On Tuesday 03 February 2009 07:53:22 Dirk Uys wrote: Hi I'm trying to emerge kde-4.2, but the kde-base/systemsettings-4.2.0 ebuild fails: snip/ Have anyone else successfully built kde4.2? Three systems so far, one x86, two amd64. Every single one failed to emerge kde 4.2 cleanly in a single run, but subsequently now has it. What I mostly did was just restarted emerge -- and it would have the packages in different order and pass the problem spot a few packages later without the problems showing up any more. Having kde 4.1 stuff on the background cluttering up my system might have had an effect on this. I don't know if I ran into your specific problem, just restarting emerge fixed most stuff for me. I did run into a hell of a problem with Xorg after successfully emerging kde 4.2, but that was due to the upgraded unstable xorg- server 1.5.3 evdev not liking my xorg.conf -- and me not having cared to explore the evdev-stuff before being thus forced to. No keyboard and no mouse makes using kde 4.2 even worse than kde 4.1 *with* keyboard and mouse. ;) -- Arttu V.
[gentoo-user] Ctrl+c kills KOrganizer 4.2 when it should copy text
Hello, I'm currently googling around for some insight into an extremely annoying feature (or a bug?) which appeared with KDE 4.2, and/or the xorg-server upgrade to 1.5.3-r2 that I did along with the kde upgrade. The problem reveals itself when I press ctrl+c in some KDE applications and that it turn kills/exits kde/X. For example, KOrganizer edit menus still show shortcuts for copy, paste, etc as ctrl+something (ctrl+c for copy). However, when I'm adding a new event or todo and press ctrl+c to copy some text around -- boom, X closes itself and I'm back to the text console (I use startx from the text console, not graphical login screens). Last time that happened was 15 minutes ago. Sometimes the crash takes keyboard and mouse with it so that even the text console doesn't have them any more (wtf?). So, maybe this has something to do with the new evdev? Or klipper the copy-paster utility? Anyway, system messages and Xorg.logs don't indicate any error there, latter just reports that evdev modules got cleanly unloaded. I think the same has been happening at least with Konqueror and Kmail as well, not sure about other apps, since I haven't done any rigorous tests across kde apps, just tried to use the ones I'm used to. Any insight on where to start looking for the problem (and especially solution) is most welcome, including pointing out stupid newbie mistakes. I hadn't had to modify xorg.conf for nearly three years (IIRC) until the new, shiny and better evdev-stuff in new xorg-server required me to re-hack the conf file to get my keyboard and USB-mouse to be recognized at all. Thanks in advance! -- Arttu V.
Re: [gentoo-user] Ctrl+c kills KOrganizer 4.2 when it should copy text
On 2/6/09, Sebastian Günther sam...@guenther-roetgen.de wrote: Some application does not capture ctrl+c, which is normaly the interupt shortcut in bash. So it surely is passed down to bash which intrupts startx, since it is the active job. I bet it won't happen with any Xsession started from any *dm. Thanks for the tip. I tried adding xdm to default runlevel and rebooted. Logging in from the kde(?) graphical login screen and pressing ctrl+c in korganizer then resulted in screen flickering via text mode back into the kde(?) login screen. So the same seems to happen there. Well, actually a bit worse: no user can login after that any more, later login attempts will flicker straight back into the login screen. If it's possible then please test this, anyway you should pinpoint down, where else it is possible to kill your Xsession, so that you can file a bugreport to the right people. Will do. Deep inside I was hoping that it would be just some configuration issue I had missed. I'm currently browsing through all the ctrl+something bugs in bugs.kde.org and will either find something or squeeze a dump with debug info for a new one. -- Arttu V.
Re: [gentoo-user] Ctrl+c kills KOrganizer 4.2 when it should copy text
On 2/6/09, Stroller strol...@stellar.eclipse.co.uk wrote: On 6 Feb 2009, at 12:36, Arttu V. wrote: ... Any insight on where to start looking for the problem (and especially solution) is most welcome, including pointing out stupid newbie mistakes. Have you run revdep-rebuild? Yes to revdep-rebuild -i and emerge --depclean and emerge @preserved-rebuild, with revdep having run last. But yes, it's always good to check that as we are talking about Gentoo systems. :) -- Arttu V.
Re: [gentoo-user] Ctrl+c kills KOrganizer 4.2 when it should copy text [Maybe solved: downgrading]
On 2/6/09, Willie Wong ww...@princeton.edu wrote: I lost your original mail, so sorry if this is not threaded properly. Ok, I have to be sorry about my slow replies as well. My excuse is that I'm so used to pressing ctrl+c, ctrl+v etc when composing mails that I've already killed X twice just while trying to type in this reply (not to mention the earlier ones where I was trying to copy and paste outputs from the terminal windows ...). I must have killed X more than 50 times during the last 24 hours, sometimes on purpose while trying out things, sometimes accidentally and then ending up cursing out loud. :) Question: does the bug manifest itself only in KDE apps or in others? Ctrl+c should also copy to clipboard in, say, firefox, if you have it installed. Does it do the same thing? If yes: do you have another WM installed and can you try running firefox under the other WM to see if it behaves similarly? This should help figure out whether it is a KDE issue or an X issue. It happens under seamonkey and firefox as well. And it's not limited to ctrl+c, also ctrl+v kills/crashes X (sometimes) and some other ctrl+something combos as well, but they're not so usual keypresses by me, so I cannot say for sure how often they would cause a crash. But that WM was a good idea, I tried with a bare(?) Xsession as the XSESSION in rc.conf. It gave just a plain terminal from which I started firefox and seamonkey. Result: ctrl+c while trying to copy-paste text in a form field on a web page results in X killing itself as well. So it would really seem to be about xorg-server 1.5.3 and that darned evdev 2.1.1 which already gave nearly me two evenings without a keyboard and a mouse due to configuration issues. I'll probably finally get to downgrade it over next night. Also the other, earlier replies had good ideas, like the kxkb. Just clicking on the KDE keyboard layout manager's country flags and switching the keyboard from our local gibberish (fi) to us and back (and then running extra setxkbmap by checking the xkb settings CTRL in lower left corner?) had me smiling for a while -- I could pound on ctrl+c with impunity in all programs. But then I restarted X and retried, and the trick no longer seemed to work, not sure why. If you open an xterm, and hit Ctrl+c, does it reboot X (a long shot, but if it does happens, it means something else is grabbing the event and passing it up before the xterm sees it and passes it to the shell in it)? Ctrl+c in xterm/gterm/kterm == X gone, just like in all other programs. Also, what is in your ~/.xinitrc? I don't have the file in my ~, only the system file in /etc/X11/xinit/xinitrc -- not modified by me. But anyway, since it seems to be focused on xorg-server and/or evdev, I'll do the reasonable thing: downgrade back to 1.5.2(or even lower?) and normal keyboard and mouse drivers. Maybe the new xorg-server is only supposed to work reliably on a fully ~amd64? This box has mostly stable amd64 and only select packages, mostly end-used apps like seamonkey, OOo and firefox are allowed to be ~amd64 along with some of their more obscure requirements (like xulrunner). Anyway, thanks for everyone for the helpful comments and ideas! I'll post a short reply about the success with the downgrade hopefully tomorrow and then try to leave this for now. -- Arttu V.
Re: [gentoo-user] Ctrl+c kills KOrganizer 4.2 when it should copy text [Solved]
On 2/7/09, Arttu V. arttu...@gmail.com wrote: Maybe the new xorg-server is only supposed to work reliably on a fully ~amd64? This box has mostly stable amd64 and only select packages, mostly end-used apps like seamonkey, OOo and firefox are allowed to be ~amd64 along with some of their more obscure requirements (like xulrunner). Anyway, thanks for everyone for the helpful comments and ideas! I'll post a short reply about the success with the downgrade hopefully tomorrow and then try to leave this for now. After a week of frustrating nightly recompilations with emerge -e failing at various points (gcc-43 not supported for *stable* packages?! o.O [1]) and futile downgrade-attempts not changing the situation, I've finally figured out the culprit for the keyboard trouble: acpid. I had mistakenly added it to default runlevel. (IIRC, due to something hinting to do it with the latest xorg-server.) Anyway, after dropping acpid the keyboard is back again, functioning as expected, ctrl+c copying stuff, cursor keys working (I even lost them at some point), etc. (Not to forget some keyboard related conf settings that appeared in /etc/hal/fdi during the emerge -e kde-meta.) That should teach me to remember *all* the changes I do, not just suspect the unstable packages. :) [1] http://bugs.gentoo.org/show_bug.cgi?id=229269#c4 -- Arttu V.
Re: [gentoo-user] hal-hell - please help
On 2/21/09, Helmut Jarausch jarau...@igpm.rwth-aachen.de wrote: To make it even work I had to put Option AutoAddDevices no to my xorg.conf file What am I missing? (Sorry if this came through already, gmail's draft saving via IMAP and a spotty wlan really mix up threads and messages in gmail's view.) During my short-lived and generally moderately clueless experimentation with the latest xorg-server, evdev and a hal-enabled PS/2 keyboard and a hal-enabled Logitech USB mouse, the mouse was not the problem, but the keyboard layouts were the killer which prompted me to disable hal altogether (ref: earlier CTRL+C kills korganizer-thread). Mouse worked ok with following changes to my earlier xorg.conf and I had no need for plugging cables in and out, it Just Worked: Section Module: Loadevdev Section ServerFlags: Option AllowEmptyInput false Section for the mouse InputDevice needed to change driver to evdev. Section ServerLayout: Option AutoAddDevices false Option AutoEnableDevices true (But I ended up commenting them out and the mouse still worked ok, so not sure if you need to toggle the defaults values for these at all.) Those changes gave me a functional USB mouse pointer with xorg-server 1.5.x, but my keyboard problems went away only after I disabled acpid and hal, and re-emerged xorg-server with USE=-hal. Wasted nearly three good weeks' nights and weekends there with kde 4.2.0 upgrade, so you can understand my above-average grumpiness about hal -- just disable it unless you really really need it. :( -- Arttu V.
Re: [gentoo-user] Mutually exclusive ebuilds?
On 2/21/09, Grant emailgr...@gmail.com wrote: I'm getting this and I'm wondering if it means I can't have miro and wicd installed simultaneously: !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: dev-python/pyrex:0 ('installed', '/', 'dev-python/pyrex-0.9.8.5', 'nomerge') pulled in by =dev-python/pyrex-0.9.6.4 required by ('installed', '/', 'media-tv/miro-2.0.1', 'nomerge') (and 1 more) ('ebuild', '/', 'dev-python/pyrex-0.9.4.1', 'merge') pulled in by =dev-python/pyrex-0.9.3-r2 required by ('installed', '/', 'dev-python/dbus-python-0.82.4', 'nomerge') It may be possible to solve this problem by using package.mask to prevent one of those packages from being selected. However, it is also possible that conflicting dependencies exist such that they are impossible to satisfy simultaneously. The gurus remain silent, but I have a theory. A bug in portage. I've had several similar-looking nonsensical blockers lately -- after portage learned to solve some blockers automatically, actually. I think at the same time portage lost its ability to figure out that if you have an unstable version package-3.4.5 installed and you're installing another package, which requires, e.g., =package-1.2.3 and the stable for that slot (and oldest available in portage tree) is between the two, e.g., package-2.3.4 then portage seems to produce these kinds of situations: wants to downgrade to 2.3.4, but simultaneously to keep the 3.4.5 in the same slot. I just had dev-libs/glib do the same two-three days ago. I have the unstable 2.18.4 installed, package comes along that requires =glib-2.14 and the latest stable in portage is 2.16.4 or something for that slot. Boom, similar nonsensical blocker. I've been solving them by unmerging the blocker and retrying the emerge. Portage has invariably proceeded without a slightest hint of blockers -- and re-emerged the unstable package version it just complained was causing a blocker. So, there, I think it could very well be a bug in portage. Maybe it has a bug report? -- Arttu V.
Re: [gentoo-user] Re: upgrade all [SOLVED]
On 2/25/09, Neil Bothwick n...@digimed.co.uk wrote: On Wed, 25 Feb 2009 16:24:21 +0200, Nikos Chantziaras wrote: Why not simply emerge -a1 `qlist -I` ? You should add a -u to emerge as he wants to upgrade, otherwise it re-installs everything that qlist -I spits out. Neither this nor the eix --only-names method take account of slots. But you should get that by changing the qlist part above to: qlist -CISL Not that it matters. I just like qlist over these newfangled eix things. :) -- Arttu V.
Re: [gentoo-user] gcc: error: 'nothrow' was not declared in this scope
On 3/7/09, Michael Sullivan msulli1...@gmail.com wrote: mich...@camille OurRPG $ make g++ -O2 -W -Wall -pedantic `sdl-config --cflags` -c draw.cpp enemyparty.o allyparty.o /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4/memory: In function 'std::pair_Tp*, int std::__get_temporary_buffer(ptrdiff_t, _Tp*)': /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4/memory:83: error: 'nothrow' was not declared in this scope Have you tried to help the compiler find nothrow declaration by explicitly doing: #include new (Disclaimer: I'm no c++ guru, just googled for that.) -- Arttu V.
Re: [gentoo-user] gcc: error: 'nothrow' was not declared in this scope
On 3/8/09, Michael Sullivan msulli1...@gmail.com wrote: On Sun, 2009-03-08 at 17:27 +0200, Arttu V. wrote: On 3/7/09, Michael Sullivan msulli1...@gmail.com wrote: mich...@camille OurRPG $ make g++ -O2 -W -Wall -pedantic `sdl-config --cflags` -c draw.cpp enemyparty.o allyparty.o /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4/memory: In function 'std::pair_Tp*, int std::__get_temporary_buffer(ptrdiff_t, _Tp*)': /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4/memory:83: error: 'nothrow' was not declared in this scope Have you tried to help the compiler find nothrow declaration by explicitly doing: #include new (Disclaimer: I'm no c++ guru, just googled for that.) I put it in there, but it didn't help. Thanks for trying. Another thing might be to try the -E option for gcc, and check the results to see if it actually correctly includes everything necessary when preprocessing. After that, I'm out of ideas, it could be your code or it could be the compiler. -- Arttu V.
Re: [gentoo-user] What annoys you?
Wyatt Epp wrote: So I was curious...what have people that are /not/ myself and my mate noticed that is mildly irritating and disruptive to the Gentoo experience? Mildly irritating, disruptive, etc: - Managing USE flags is sometimes quite irritating, having to fidget around /etc/make.conf, /etc/portage/package.use and such with everyone's favourite text editors. (Maybe I'm micromanaging them too much?) - Keeping the crud out of /etc/portage/package.keywords and other similar portage files (much like with the case of USE flags above) for us who don't run bleeding edge ACCEPT_KEYWORDS=arch ~arch. - Difficulty of predicting how long some new package compilations (with dependencies, upgrades and revdep-rebuilds etc) will actually take (genlop -t only knows about individual packages that have been emerged before) - Portage 2.2 stopping dead with you should re-emerge foo with USE=bar with the new cat/foo[bar]-style dependencies. -- Arttu V.
Re: [gentoo-user] What annoys you?
On 4/4/09, Neil Bothwick n...@digimed.co.uk wrote: On Sat, 04 Apr 2009 02:24:58 +0300, Arttu V. wrote: - Managing USE flags is sometimes quite irritating, having to fidget around /etc/make.conf, /etc/portage/package.use and such with everyone's favourite text editors. (Maybe I'm micromanaging them too much?) emerge flagedit Yes, flagedit sure helps on a single box, but when running several Gentoo boxes, with slightly differing USE settings, arches and whatnot (firewall, server, old box for light browsing/office work, new-ish multimedia/gaming workstation, a laptop, one with radeon-drivers, another with nvidia-drivers etc), it all tends to get a bit hairy. What I'm maybe thinking about is some kind of Microsoft AD / SUS style or similar pushing parts of policies or changes across all boxes in a domain (or whatever they call them there). But then again, if that is a priority, then running Windows or CentOS Linux with Red Hat's stuff might be better for that than Gentoo in the first place? :) - Keeping the crud out of /etc/portage/package.keywords and other similar portage files (much like with the case of USE flags above) for us who don't run bleeding edge ACCEPT_KEYWORDS=arch ~arch. eix-test-obsolete That was a good call, thank you! I have been avoiding eix for various reasons, but now I emerged it and tried it. Blindfolded and both hands tied behind its back eix-test-obselete beat my own embarrassingly simplistic helper perl-scripts. Very good indeed, thanks! Would you rather portage simply re-emerged installed packages with different USE flags without consulting you? I suppose a package X needs to be remerged with USE=Y, proceed Y/n message could be useful, maybe with an option to do accept this automatically. Yes, I think this is more of a default settings issue. I'm betting 999/1000 cases or even worse, the user/admin just goes and emerges the dep with the required USE anyway. So why waste the time (or introduce compulsory interactivity) from most people and not have it as a default? Ok, I admit, it might create some really hairy situations for deep trees when some critical system package would get silently re-emerged with a new flag -- but maybe the automagic could only work for packages not in system/@system? Then there would be hope of having a reasonably sane box (command line) remaining even if the automagic in the emerge would end up shooting yourself in the foot ... Gentoo is about providing ultimate control to the admin, so you can't really complain about having to make those choices :) Hrpmft, I thought by now -- after about a decade or so of development of the brightest young minds of planet Earth -- there would be a package available in Gentoo with a command or script like eix-do-what-i-mean(t). ;) -- Arttu V.
Re: [gentoo-user] What annoys you?
On 4/4/09, Stroller strol...@stellar.eclipse.co.uk wrote: Do you really need to know EXACTLY how many? Or can you just accept a little patience as the cost of using Gentoo? I'd be happy for a modest, rough guesstimator. A Gentoo emerge weather forecast-gauge/meter if you will. :) I'm currently running boxes from ~1Ghz up to this four-year-old single-core AthlonXP 3500+ (and thinking about getting one of 'em blisteringly fast Phenom IIs or Core2Duos), so I may be an anachronism with my systems. I think in theory one could probably use the kernel's bogomips value, CPU count, source package language and size, and maybe some known few key system package's emerge times to guesstimate the emerge time of the rest. It could provide, e.g., a guessed worst and best scenario. Naturally it couldn't do much about the potential revdep-rebuild part, though. Darn, maybe I have to allocate some time to try to build one myself? Just don't hold your breath while waiting nor keep hopes too high ... ;) -- Arttu V.
Re: [gentoo-user] GCC-4.3.2
On 4/7/09, Joseph syscon...@gmail.com wrote: I boot from CD, chroot and tried to recompile the gcc but it has been compiling for over two hours and it can not finish, I think it is in some kind of loop mode? That loop is the multilib profile you're probably using, which will compile you both 64-bit and 32-bit version of gcc? You also stated in an earlier email that your system has an Athlon 3000+ processor. As a comparison, here with an Athlon 3500+ and 3GB of RAM the gcc:4.3 compilation takes around two hours if the ccache is empty and an hour and 20 minutes if it hits mostly the cache, so the duration doesn't appear surprising for a first compilation (varying by the USE flags, naturally). -- Arttu V.
Re: [gentoo-user] Reproducible Firefox crash
On 4/7/09, Volker Armin Hemmann volkerar...@googlemail.com wrote: I have flash and only the flashgot extension - and it doesn't crash. Did you enable the flash piece? With Firefox 3.0.8 on amd64, Adobe Flash 64-bit plugin and flashblock here, problems only appear once I enable the blocked flash piece. I can tell NoScript to allow the scripts etc, and the tab will close cleanly, without crashing the whole browser. However, the moment the flash object is enabled, the close tab button becomes the self-destruct button. -- Arttu V.
Re: [gentoo-user] Kbackup emerge error.
On 4/8/09, Dale rdalek1...@gmail.com wrote: Anybody ran into this: snip/ Selector.cxx: In constructor 'Selector::Selector(QWidget*)': Selector.cxx:125: error: '::getenv' has not been declared snip/ I searched the forums but it returned very little and most of that was really really old. I went back a good piece and I think all the error should be there but if I missed something, let me know. I can go back a little further if needed. Ok, the gurus won't touch this (at least on list) so I'll throw out a wild guess: maybe a gcc 4.3 related issue in kbackup 0.5.3's sources? If kbackup 0.5.4[-r1] (currently ~arch) has the same then you probably want to file a bug about it at b.g.o. Otherwise a bug report will just be a waste of time, it'll get either ignored or closed with a suggestion to use the newer one. (I.e., then one must try to out-clever the devs and file a stabilization bug for the newer one? ;) ) -- Arttu V.
Re: [gentoo-user] Kbackup emerge error.
On 4/13/09, Dale rdalek1...@gmail.com wrote: I have the same thing for both Kbackup Apparently, it was a gcc-4.3 issue and I tried to provide a fix for it in bug #266026. and hwinfo. It looks like it is a boo boo between the kernel and gcc. Which version of hwinfo and what does it complain about? sys-apps/hwinfo 14.19 and 13.28 emerge just fine on my two trusty, old amd64 boxes (I don't have x86 unless I reboot), while 8.38 doesn't even have any amd64 keywording (so didn't even try) and 13.0 squeals about a thing that I guess might also be fixable by a one- or two-line patch: gcc -c -Wall -pipe -g -O2 -march=athlon64-sse3 -pipe -I/var/tmp/portage/sys-apps/hwinfo-13.0/work/hwinfo-13.0/src/hd hwinfo.c gcc hwinfo.o -Lsrc -lhd -o hwinfo src/libhd.so: undefined reference to `dbus_connection_disconnect' collect2: ld returned 1 exit status make: *** [hwinfo] Error 1 -- Arttu V.
Re: [gentoo-user] CMake and empty directories
On 4/15/09, Patrick Holthaus patrick.holth...@uni-bielefeld.de wrote: INSTALL(CODE FILE(MAKE_DIRECTORY ${INCLUDEDIR_B}/iwPlugins)) snip/ It seems for me that the variable INCLUDEDIR_B is not expanded to the sandbox path but /usr/include within the CODE statement. Just a wild guess from a perl-fan: would it work if you escaped that dollar sign with a backslash? It might move the expansion around a bit, maybe even to the right place if you're lucky. -- Arttu V.
Re: [gentoo-user] gentoo installation - emerge problem
On 4/15/09, Christoph Schrauth schrauth-christ...@gmx.de wrote: Hi all, I'm installing a new gentoo system. After installing the minimal system I tried to emerge applications like kde, samba, cups or something like that, but everytime I get the following message: emerge -av kde-meta These are the packages that would be merged, in order: Calculating dependencies... done! emerge: there are no ebuilds built with USE flags to satisfy app-text/docbook-sgml-utils[jadetex]. !!! One of the following packages is required to complete your request: - app-text/docbook-sgml-utils-0.6.14 (Change USE: +jadetex) (dependency required by media-libs/fontconfig-2.6.0-r2 [ebuild]) Could you relax your USE-flags, i.e., drop USE=doc? Fontconfig's ebuild only needs the sgml stuff for extraneous docs and I doubt whether you really need the pdf documents created by that USE flag -- especially given that you're apparently still on the command line where viewing them might req ... -- Arttu V.
Re: [gentoo-user] Re: Mysql and utf8: Can't initialize character set utf-8
On 5/2/09, Galevsky galev...@gmail.com wrote: Well can someone tell me if a utf-8 gentoo box shoud have the file /usr/share/mysql/charsets/utf-8.xml ? AFAICT no and AFAICT nor should any other MySQL instance on any platform. According to sql/share/charsets/README those files are only for simple charsets, e.g., with single-byte characters. UTF-8 is multi-byte and thus should be compiled in the strings library. -- Arttu V.
Re: [gentoo-user] nanosleep broken on ~amd64?
walt wrote: Could someone else compile the test and confirm that it returns 119 on ~amd64 instead of 0? It returns 119 on an semi-ancient Athlon64 3200+ box here as well. Could kernel HZ-settings affect the outcome? This box has CONFIG_HZ=250, but tomorrow I can try on another amd64 which runs a 1000HZ kernel IIRC. -- Arttu V.
[gentoo-user] What's the deal with CAMERAS anyway?
Hello, Inspired by the other digicam thread I took a quick look at an older problem of mine, namely getting libgphoto2 to compile drivers for all cameras. Anyone got a clue why *not* setting CAMERAS won't build drivers for all cameras as claimed several times in the libgphoto2 ebuild? All I see is that CAMERAS variable is not handled directly by the ebuild (but maybe libgphoto2's Makefiles or build scripts do something with it?), and when I comment out CAMERAS=canon line from /etc/make.conf, then *poof* libgphoto2 plans to drop even that only driver I currently have compiled: [ebuild R ] media-libs/libgphoto2-2.4.3 USE=exif hal nls -bonjour -doc CAMERAS=-adc65 -agfa_cl20 -aox -barbie -canon* -casio_qv -clicksmart310 -digigr8 -digita -dimagev -dimera3500 -directory -enigma13 -fuji -gsmart300 -hp215 -iclick -jamcam -jd11 -jl2005a -kodak_dc120 -kodak_dc210 -kodak_dc240 -kodak_dc3200 -kodak_ez200 -konica -konica_qm150 -largan -lg_gsm -mars -mustek -panasonic_coolshot -panasonic_dc1000 -panasonic_dc1580 -panasonic_l859 -pccam300 -pccam600 -polaroid_pdc320 -polaroid_pdc640 -polaroid_pdc700 -ptp2 -ricoh -ricoh_g3 -samsung -sierra -sipix_blink -sipix_blink2 -sipix_web2 -smal -sonix -sony_dscf1 -sony_dscf55 -soundvision -spca50x -sq905 -stv0674 -stv0680 -sx330z -template -topfield -toshiba_pdrm11 5,052 kB Something's fishy with this and I just haven't had the time to investigate further. A quick glance-comparison shows libgphoto2 ebuild missing the critical lines actually doing something with the CAMERAS variable -- while for similar variable settings lines exists for, e.g., lirc (LIRC_DEVICES) and alsa-driver (ALSA_CARDS) in their respective ebuilds. After taking a two-minute look at the ebuilds I just have the question: is it supposed to work, does leaving it empty work for someone? -- Arttu V.
Re: [gentoo-user] Resizing ext3 Partition
On 5/4/09, Marco listwo...@gmail.com wrote: On Mon, May 4, 2009 at 1:44 PM, Andrew MacKenzie amack...@edespot.com wrote: +++ Marco [gentoo-user] [Mon, May 04, 2009 at 12:56:05PM +]: [...] Just to be sure you checked - Gentoo keeps temporary files in /var/tmp/portage/ (build temp location, sometimes things get left here) and /usr/portage/distfiles/ (download location). /usr/portage/distfiles can get pretty large over time. I cleaned this directory frequently, but still running low on disk space... I may have missed something of your configuration or partitions (did you have a separate /usr?), but check old and useless kernel sources from under /usr/src (and under /lib/modules if you've compiled and installed modules). Clean out old and unused ones. Compiled kernel sources directories can be over 800 MB *each*. For example for my current /usr/src/linux (which points to ./linux-2.6.28-gentoo-r5) du says 818MB. You don't need too many of these to fill up a 10GB partition. -- Arttu V.
Re: [gentoo-user] What's the deal with CAMERAS anyway?
On 5/4/09, Paul Hartman paul.hartman+gen...@gmail.com wrote: On Mon, May 4, 2009 at 6:43 AM, Arttu V. arttu...@gmail.com wrote: Anyone got a clue why *not* setting CAMERAS won't build drivers for all cameras as claimed several times in the libgphoto2 ebuild? src_configure() in the libgphoto2 ebuild looks like it handles the none=all logic. That was the same part I glimpsed over earlier in wonder, and the ebuild didn't really answer my question as it doesn't even touch the CAMERAS variable (only the local lower case cameras and also IUSE_CAMERAS). But after a fair bit of grepping around portage's sources I have now come up with a theory that satisfies my curiosity for now. I think the problem is my poor brain not understanding the somewhat misleading printout from emerge -p libgphoto2. It has a minus sign in front of all CAMERAS drivers listed, so trying to be logical I assumed it wouldn't emerge any of them. emerge -p lirc does something similar with LIRC_DEVICES. It is another example of these USE_EXPANDed (learned a new word of gentoo-eeze while grepping :) ) variables, *does* show the string all amongst its list of LIRC_DEVICES while the actual device driver names are also still preceded by minuses. But now I've found out that that's just due to all being an extra option specifically listed in /usr/portage/desc/lirc_devices.desc. /usr/portage/desc/cameras.desc doesn't have such an all line, so emerge won't print all for CAMERAS, only all the drivers' names with minuses in front. Ergo, my confusion follows from the premises? Sorry about the confusion and thanks for the replies. But is this a bug or a feature? Am I the only one who has fallen for this? -- Arttu V.
Re: [gentoo-user] kde:3.5 how to phase out
On 5/12/09, Helmut Jarausch jarau...@igpm.rwth-aachen.de wrote: Hi, I'd like to phase out KDE:3.5. One reason is, that I cannot install both, 3.5 and 4.2 without using 'prefix'. For doing so, I need to find out which (installed) packages use which parts of KDE:3.5. Is there something better than unmerge all of KDE:3.5 and let revdep-rebuild find out. Furthermore, how to unmerge all of KDE:3.5, e.g. emerge --unmerge kde-base/kde-meta:3.5 does not work. Some of your questions are answered here, right on the top (I hope I link to the up-to-date version and not the outdated one): http://www.gentoo.org/proj/en/desktop/kde/kde4-guide.xml http://www.gentoo.org/proj/en/desktop/kde/ I don't know if my advice beyond those links is of any use as I did it all the wrong way -- didn't unmerge previous kde, just unmasked the new kde versions, didn't use sets, still have USE=-kdeprefix etc. Still, sky didn't fall on me, at least not until I unmasked the then-unstable xorg 1.5 ... -- Arttu V.
Re: [gentoo-user] emerge spamassassin: re2c failed...
On 5/21/09, Jarry mr.ja...@gmail.com wrote: configure:2460: checking for C++ compiler default output file name configure:2487: x86_64-pc-linux-gnu-g++ -march=athlon64 - O2 -pipe -Wl,-O1 conftest.cpp 5 x86_64-pc-linux-gnu-g++: O2: No such file or directory x86_64-pc-linux-gnu-g++: -E or -x required when input is from standard input configure:2490: $? = 1 CFLAGS='-march=athlon64 -O2 -pipe' CXXFLAGS='-march=athlon64 - O2 -pipe' Nah, let's make a better guess: most unfortunate typo in /etc/make.conf? That single space in CXXFLAGS really makes the maximum difference? :) (Sorry about the earlier mails, their formats, etc., I was fighting with gmail on a non-regular box again.) -- Arttu V.
Re: [gentoo-user] ARGH I uninstalled python
Dale wrote: You may want to look into that setting for next time. I did however notice that although python is in the system set, it is not saving a copy for some reason. Anybody know why this setting is not working? FEATURES=--keep-going buildsyspkg sandbox fixpackages Is the buildsyspkg option not valid anymore? http://sources.gentoo.org/viewcvs.py/gentoo-x86/profiles/base/packages?r1=1.40r2=1.41 Python appears to be gone from the base system? Or maybe I'm looking at the wrong file? Anyway, 'emerge -pv @system | grep python' would seem to agree here. Now I don't have a PhD in Gentoo Package Manglement (barely passed the kindergarten level so far?), but that commit looks rather unsettling for other parts as well. Are the devs switching over to paludis in droves? Or why is sys-apps/portage commented out as well? Why is it replaced by the hard-coded (old-style?) virtual/portage? -- Arttu V.
Re: [gentoo-user] ARGH I uninstalled python
On 5/23/09, Dale rdalek1...@gmail.com wrote: This is what I get with this command: r...@smoker / # emerge -ep system | grep python [ebuild R ] app-admin/python-updater-0.5 [ebuild R ] dev-lang/python-2.5.4-r2 r...@smoker / # So python is there if you do it emptytree. emptytree pulls in *all* dependencies, regardless of whether they are in the system set or even if they are just regular dependencies (random packages) from an external overlay. So you're not comparing apples to apples there with -e, you're getting lot of false positives from outside the real system set. It is still a good question why one would have any system set packages having critical dependencies to non-system set packages, but reasoning for that has to be asked from the portage devs ... But backing up a little, try this instead for an amusement (remember to --pretend or --ask): ~ # emerge -pvC python These are the packages that would be unmerged: dev-lang/python selected: 2.5.4-r2 protected: none omitted: none 'Selected' packages are slated for removal. 'Protected' and 'omitted' packages will not be removed. ~ # emerge -pvC binutils These are the packages that would be unmerged: !!! 'sys-devel/binutils' is part of your system profile. !!! Unmerging it may be damaging to your system. sys-devel/binutils selected: 2.18-r3 protected: none omitted: none 'Selected' packages are slated for removal. 'Protected' and 'omitted' packages will not be removed. ~ # Note how binutils still is within the system set and gets the extra warning, while python doesn't get the special system set warning with the three exclamation marks? I.e., python *is no longer* in the system set, probably removed by the very change I gave a link for in previous email. AFAIK (haven't checked in portage code), buildsyspkg and such only work with the packages explicitly mentioned in the system set, so binutils is in, python is (nowadays) out -- and that probably explains also what is happening on your smoker. :) Anyone finding flaws in my theories or facts or observations that outright nullify them? :) -- Arttu V.
Re: [gentoo-user] Excessive digest failures in portage
On 5/25/09, Saphirus Sage saphirus...@gmail.com wrote: So, is this just the crying screams of a failing drive, or is there some other problem at hand here? If the drive is SMART-capable, you could try smartmontools, enable smart and see what the tests say. It might help to diagnose whether it is the drive or some other part. -- Arttu V.
Re: [gentoo-user] USE=mmx mmxext sse sse2 ssse3 3dnow 3dnowext
On 5/27/09, Wyatt Epp wyatt@gmail.com wrote: Reading this thread, though, it seems like it would be useful to have a FEATURES=cpudetection for Portage. Congratulations, it's already there! Sort of. -march=native ;) -- Arttu V.
Re: [gentoo-user] Whats better for crossplatform applications?
Volker Armin Hemmann wrote: On Sonntag 31 Mai 2009, Roy Wright wrote: java - portability was one of the original design goals and it is still ugly. Also very funny and old bugs. Given the OP's question (identical look and feel across platforms a very high priority) he should really check out Java and Swing: http://java.sun.com/docs/books/tutorial/uiswing/lookandfeel/plaf.html qt4 - some kde apps are starting to be portable kde always was portable. Also developed as a cross plattform lib. Lots of windows apps already use qt. Great documentation. Lots of language bindings, nicer licensing. Yes, I'd say qt is the other one to check. gtk+ might mean trouble, IIRC different gtk+ versions didn't use to live so nicely on a Windows box. GIMP and some other program required different versions of gtk+ on Windows when I last tried -- and the Highlander paid a visit: there can be only one ... (Maybe they've fixed that? Or was it about Windows not being too tolerant about different versions of libraries?) ... but anyway I'm mostly classificable as a server-side java dude, and don't know the more obscure choices with GUIs that well. So, my opinions are worth their weight in uranium (hazardous material, need proper handling, not suitable for most uses, keep away from the reach of children ;) ). -- Arttu V.
Re: [gentoo-user] Re: What's wrong with emerge
On 6/3/09, alain.didierj...@free.fr alain.didierj...@free.fr wrote: After three days, emerge --sync still returns the same error message. Is there anything I can do on my computer to fix it ? I don't know the portage internals well enough even to know if my following suggestion is good or bad (and I wonder why isn't your profiles directory updated by the sync?), but the troublemaking change is probably the one unmade by this: http://sources.gentoo.org/viewcvs.py/gentoo-x86/profiles/updates/2Q-2009?r1=1.24r2=1.25 So, if you're feeling lucky (and nothing else works nor you get any better advice), you can try removing the reddened line out of your /usr/portage/profiles/updates/2Q-2009 (I'd make a backup copy before editing). -- Arttu V.
Re: [gentoo-user] Where to find 32 bit emul libs?
On 6/5/09, Alan McKinnon alan.mckin...@gmail.com wrote: I have all emul-linux packages installed anyway (except -qtlibs), but now I'm stumped. How would I find which packages provide these 32 bit libs? Possibly none, but maybe this page helps: http://www.gentoo.org/proj/en/base/amd64/emul/content.xml -- Arttu V.
Re: [gentoo-user] Failed to (re)build kaffeine 0.8.6.
On 6/7/09, meino.cra...@gmx.de meino.cra...@gmx.de wrote: Meanwhile I compiled kaffeine ... but now it does not find xine_part.desktop ... on the other hand: ls /usr/share/services/xine_part.desktop /usr/share/services/xine_part.desktop Could you try copying it over to /usr/kde/3.5/share/services/ and see if the problem persists? image # find . -iname 'xine*.desktop' ./usr/kde/3.5/share/services/xine_part.desktop It could be another similar, possibly kde-3.5 path related(?) issue like your first one probably was (if you checked the comment dropped into bug referred in an earlier mail in this thread and also couple minutes later bug #273030). -- Arttu V.
Re: [gentoo-user] Failed to (re)build kaffeine 0.8.6.
On 6/7/09, Arttu V. arttu...@gmail.com wrote: Could you try copying it over to /usr/kde/3.5/share/services/ and see if the problem persists? image # find . -iname 'xine*.desktop' ./usr/kde/3.5/share/services/xine_part.desktop Naturally meant if it is *there* and you copy it to the other location. ;) (Typing faster than I'm thinking again :D ) -- Arttu V.
Re: [gentoo-user] Failed to (re)build kaffeine 0.8.6.
On 6/7/09, meino.cra...@gmx.de meino.cra...@gmx.de wrote: Arttu V. arttu...@gmail.com [09-06-07 18:01]: On 6/7/09, Arttu V. arttu...@gmail.com wrote: Could you try copying it over to /usr/kde/3.5/share/services/ and see if the problem persists? image # find . -iname 'xine*.desktop' ./usr/kde/3.5/share/services/xine_part.desktop Naturally meant if it is *there* and you copy it to the other location. ;) (Typing faster than I'm thinking again :D ) -- Arttu V. THAT was helpful! Thanks a lot ... it simply WORKKS! Great to hear! At least sometimes my blind guessing works. :) Anyway, you might want to file a bug for it, if it prevents running kaffeine. Also, there is already one (by some other kind person) possibly fixing your original problem with kaffeine-0.8.6: http://bugs.gentoo.org/show_bug.cgi?id=273030 But there is also this one: http://bugs.gentoo.org/show_bug.cgi?id=272736 I'm wondering if this latter one is due to similar path problems? It would seem as if all kaffeine (or kde?) versions could be suffering from looking for various things like .desktop entries, icons etc from under similarly wrong paths? -- Arttu V.
Re: [gentoo-user] Trying SRC_URI first during fetching
On 6/9/09, Nikos Chantziaras rea...@arcor.de wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? As others have pointed out, caches, proxies et cetera are often Your Friends. Thus, I wonder if there is something specific you are trying to achieve here? Maybe these wise men from odd countries (like Mr McKinnon's weird country on an unknown planet, with companies with actual clueful management! ;) ) can help you even more if you reveal a bit of the *why* behind the question. -- Arttu V.
Re: [gentoo-user] How to list every installed package?
On 6/10/09, Johannes Geiss johannes.ge...@web.de wrote: Hi there, how can I list every installed package in Gentoo? The file /var/lib/portage/world shows some installed packages, but how can I see all other packages which have been installed automatically? qlist And I often use it with -CISL, but your needs may differ. -- Arttu V.
Re: [gentoo-user] How to list every installed package?
On 6/10/09, Dale rdalek1...@gmail.com wrote: P. S. Ain't it funny that everybody has a different way of doing the same thing? And a different tool for each one. No, the suggested commands aren't all even doing the same thing. Especially emerge -e world is an outlier. On a box with a lazy or a first time install, learning Gentoo admin it's likely to give a very different result from most of the others. As the OP hasn't given a reason or motivation why he needs the package list, what it is for, we can only make assumptions and throw out wild guesses. After all, he might even be after emerge -pv --depclean, as it, too, sort of prints out pretty much all packages on the system. Not to mention a few thousand lines of dependency information. :) -- Arttu V.
Re: [gentoo-user] Re: Introduce Manual manipulation during an emerge
On Fri, 2009-06-12 at 16:22 -0500, Harry Putnam wrote: i486-pc-linux-gnu-gcc -O2 -march=i486 -pipe -fno-inline-functions2 -march=i486 -pipe -fno-inline-functions _autotst.c -o _autotst -Wl,-O1 cc1: error: unrecognized command line option -fno-inline-functions2 What do put in your CFLAGS (e.g., in /etc/make.conf)? Typos there? -- Arttu V.
Re: [gentoo-user] Re: Introduce Manual manipulation during an emerge
On Fri, 2009-06-12 at 18:39 -0500, Harry Putnam wrote: Here is what is in there right now.. apparently from the stage3 pulled down during install a few days ago. CFLAGS=-O2 -march=i486 -pipe CXXFLAGS=-O2 -march=i486 -pipe CHOST=i486-pc-linux-gnu Is there a reason to use i486 stage3? I think an i686 one might have been available and a better hit if your system is/was set up as an i686 before this? Well, not that it counts now, gotta go with what you have unpacked. Looking at make.conf from an old backup I see: CFLAGS=-O2 -march=i686 -pipe CHOST=i686-pc-linux-gnu CXXFLAGS=${CFLAGS} So trying that now. Haa finished up smoothly with those flags changed. Do you think I do something like `emerge -vuDN world' since everthing was compiled up to now with the old flags shown above? The friendly gentoo devs have whole guide for this: http://www.gentoo.org/doc/en/change-chost.xml And, as you can see, it starts with the soothing words Changing the CHOST is a big issue that can seriously screw up your system ... I'd go through that guide first to get to a (hopefully) sane system, and continue with other emergings only then. -- Arttu V.
Re: [gentoo-user] tar and huge tarballs used for back-ups
On 7/12/08, Dirk Heinrichs [EMAIL PROTECTED] wrote: Am Freitag, 11. Juli 2008 schrieb Dirk Heinrichs: Am Freitag, 11. Juli 2008 schrieb Daniel Iliev: Any help will be much appreciated. [1] http://bugs.gentoo.org/show_bug.cgi?id=230813 Will take a look this weekend. Don't even get it compiled manually :-( I took Daniel's ebuild from the bug and scrubbed it slightly. I'm no pro in ebuilds either (they always get rewritten by some real dev :) ), so I dare not claim that I would have made it any better either, only different. :) Actually, splitpipe compiled pretty much as such on amd64 for me and my only problems using the thing (writing files to media) were figuring out the right device and command. Ok, something was written correctly, but something wasn't. Upon restoring attempt, splitpipe/joinpipe choked at the end of the first disc: snip UUID of this session is '8fb1a646f13a1d02 204d506d196884ab' joinpipe: volume was started on Sat Jul 12 14:51:50 EEST 2008 joinpipe: found volume 1, as expected Fatal: during read of a stretch of input: Input/output error gzip: stdin: unexpected end of file tar: Unexpected EOF in archive tar: Unexpected EOF in archive tar: Error is not recoverable: exiting now /snip Files before that last one on the first disc were apparently restored correctly, I checked by comparing with diff -r to the originals. However, the last file that is split across discs and any later ones were unrecoverable. I'll see where to dump the couple coasters I just became the happy owner of and then see if I get around later to try with cdrtools instead of wodim (cdrkit). I'm afraid that if someone won't try that, it'll be all FUD on cdrkit and bugs all over again (without us actually knowing whether it really was, e.g., a PEBKAC/ID-10T-problem at my end). Ebuild is attached in the bug (#230813) if anyone wants to try this further, e.g., on x86 or with dvds. I only tried on amd64 and with CDR-80s. -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] tar and huge tarballs used for back-ups
On 7/12/08, Dirk Heinrichs [EMAIL PROTECTED] wrote: Am Samstag, 12. Juli 2008 schrieb Arttu V.: Actually, splitpipe compiled pretty much as such on amd64 for me Hmm, maybe a compiler issue. Which gcc version are you using? I'm using 4.3.1. gcc version 4.1.2 (Gentoo 4.1.2 p1.1) -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Simplistic bash-script to manage package.use
On 7/13/08, Mikko Husari [EMAIL PROTECTED] wrote: i did not found (nor did i look for long enough) any tools/scripts to easily manage package.use file, so i did my own. now i started to think that, if there really is not any, my creation would be sufficient (although possibly buggy) for most users who wish to have easy control over package.use. How does that compare to things like flagedit (available in portage)? -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] tar and huge tarballs used for back-ups
On 7/12/08, Arttu V. [EMAIL PROTECTED] wrote: then see if I get around later to try with cdrtools instead of wodim (cdrkit). Ok, my final conclusion on splitpipe: it compiles but won't work for me. I've produced about a dozen coasters while having tried some combinations of two different machines/arches (x86, amd64, both with stable core packages), two different burners (trusty last millennium HP CD-Writer+ 9100 series and newish el cheapo LG branded do-everything hyper-duper-combo thingie), two different tars (GNU tar and star) and two different pieces of burning software (cdrkit and cdrtools). Consistently, the failure is at the end of the first volume (disc), where splitpipe must have been writing some broken bytes at the end of the volume (possibly miscalculating something first?) -- which it realizes when it does the reading (as joinpipe) and then there's the barf (Fatal: probably from splitpipe) and then bad input propagates to tar/zip/others which join in in the mass-barfing contest: UUID of this session is '35a141883a219349 0c2b7efad2ffdaf7' joinpipe: volume was started on Sun Jul 13 16:38:38 EEST 2008 joinpipe: found volume 1, as expected star: WARNING: skipping leading '/' on filenames. Fatal: during read of a stretch of input: Input/output error star: Tar file too small (amount: 101 bytes). star: Unexpected EOF on input. star: Cannot recover from error - exiting. star: 68552 blocks + 5733 bytes (total of 701978213 bytes = 685525.59k). -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] tar and huge tarballs used for back-ups
On 7/14/08, Daniel Iliev [EMAIL PROTECTED] wrote: Arttu: - Can you, please, post the command you used to make the test backups? During weekend I used pretty much the spells given on splitpipe's examples page [1], only changing the directory given to tar and the drive device (and speed for the older drive, I'm not sure if it can even *read* at this speed): tar clz /test | splitpipe -s cdr-80 -o 'cdrecord dev=/dev/cdrw speed=24 -tao -v -gracetime 2 driveropts=burnfree -data -' Last run yesterday was using star instead of tar, otherwise using that same line. Still no luck. Today I tried switching off burnfree and dramatically dropping speed for the newer drive as well: tar clz /test | splitpipe -s cdr-80 -o 'cdrecord dev=/dev/cdrom speed=4 -tao -v -gracetime 2 driveropts=noburnfree -data -' But the result was the same, otherwise ok contents, but some bytes crapped at the end of the first disc. Hmm, if this seems to be consistent, then maybe there is something calculated/const'ed wrong for sizes of CDRs in splitpipe? I think I caught a glimpse of calculate correct sizes somewhere in its TODOs. [1] http://ds9a.nl/splitpipe/examples.html - Did you try splitpipe on DVDs or on CDs only? CDR-80s only, Samsung Pleomax CD-R 52X 700 MB/80 min as it says on top of the discs still in the spindle. I don't even have DVD media to try with, only couple 100 piece spindles of those same CD-Rs. Actually, my motivation with splitpipe is/was getting multivolume CDs working so I wouldn't have to move to DVDs (or tapes, the horrors!) for a while. :) Dirk, Arttu: - Should we get off-list on this subject or at least open a new thread? If you move, I'll follow -- if I even have any reasonable input for the discussion (of which I'm not too sure). -- Arttu V. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Still kde problems with monolithic-split
On 10/4/08, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: [blocks B ] gnome-base/gail-1000 (is blocking x11-libs/gtk+-2.14.3-r2) I'm no expert on this, but to me this looks like you're running some unstable gnome/gtk packages? This particular line reminds me of this quite fresh bug at bugs.gentoo.org: http://bugs.gentoo.org/show_bug.cgi?id=239490 Maybe some of your unstable gnome stuff is causing some weird side-effects? What if this gnome's unstable gail- and gtk+-related issue is first fixed by that hint in that bug's last comment? -- Arttu V.
Re: [gentoo-user] Still kde problems with monolithic-split
On 10/4/08, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: No, it does not help. The current state (excerpt) is: snip/ Total: 23 packages (3 upgrades, 4 downgrades, 15 new, 1 in new slot, 18 blocks), Size of downloads: 63,220 kB There is one blocker less, which is slightly better but certainly not a full solution. What worries me is if your portage doesn't find the kdegraphics package to remove -- as you wrote in your later email to Alan. Does it find any other installed monolithic kde packages to remove at all, i.e., what does this command print out (should be all on one line/command, probably gets wrapped in email client): emerge -pvC kdebase kdeaddons kdeadmin kdeartwork kdeedu kdegames kdegraphics kdemultimedia kdenetwork kdepim kdetoys kdeutils kdewebdev kdeaccessibility The printout should tell us if you have any remnants of monolithic kde packages left. -- Arttu V.
Re: [gentoo-user] Re: [gentoo-user] Продажа квартир.Скидки становятся меньше...
On 10/9/08, Nikolay Mikheev [EMAIL PROTECTED] wrote: 9 октября 2008 г. 22:53 пользователь Andrew Tchernoivanov [EMAIL PROTECTED] написал: С какой херовой стати? is much better. I don't think so ;) What the fuck means, sorry, Какого х*я :) All these weird characters and unpronouncable symbols have made me curious. How would the ancient Egyptians have printed that exclamation with their hieroglyphs? O.o -- Arttu V.
Re: [gentoo-user] AbiWord 2.6.4 build failure
On 10/18/08, Liviu Andronic [EMAIL PROTECTED] wrote: PORTDIR_OVERLAY=/usr/portage PORTDIR_OVERLAY=/usr/local/portage I don't really know an answer, but I'm curious about what has been attempted with these two lines. Would it work better without overlays, i.e., by commenting these two out? -- Arttu V.
Re: [gentoo-user] emerge world failing with No module name libxml2
On 10/20/08, Brian Wince [EMAIL PROTECTED] wrote: I have tried running python-updater -p and it complains about blocked packages which are older packages that I have not yet been able to get updated via the emerge world. Would python-updater -i be of any assistance? It looks at slots rather than versions. -- Arttu V.
Re: [gentoo-user] emerge world failing with No module name libxml2
Sorry for the slow reply. I was expecting more people interested in sharing their insights into updating old, ancient-grade dinosaur systems. :) On 10/20/08, Brian Wince [EMAIL PROTECTED] wrote: So the -I ignores the versions of packages installed and just fixes packages broken by a python upgrade, independent of the version? Yes, I'd think for example in terms of the difference between emerge foo and emerge =foo-1.2.3. AFAICT, python-updater does the latter by default, but with the -i option it is given a permission to do the former within a slot. Since most packages only have a single slot, the within a slot becomes irrelevant for them. I understand that slots can be used to install different versions of an app but not sure how that relates to python-updater. Sorry if I lead you too much into this slot territory. The slots are more of a red herring here, or just a small, distracting detail. For most packages slots won't matter. But there are the few, like qt, for which it is important to keep within a slot. Therefore it is nice that helpful scripts, like python-updater, try to provide automagic support for the slotting related stuff where it is needed. Still, python-updater -i is not guaranteed to fix your original problem, it was just a suggestion. The problem might not even lay with libxml2, but one of its dependencies, their dependencies, or dependencies of python itself. If you haven't already tried it, then revdep-rebuild from the gentoolkit package is your other friend. You might need try running it first and then retry with the python-updater. -- Arttu V.
Re: [gentoo-user] Openttd
On 10/21/08, András Csányi [EMAIL PROTECTED] wrote: Few weeks ago I reinstalled my desktop and now I want to play with openttd but I wonder this package is masked. I saw the version of this package int the portage is 0.5.3 but on the openttd's website the latest stable version is 0.6.3. So I want to ask: what is the problem whit this package? Hi, Did you read these bugs? http://bugs.gentoo.org/show_bug.cgi?id=215776 http://bugs.gentoo.org/show_bug.cgi?id=233929 They answer at least partially to many of your questions. Old one is blocked due to security reasons, new one has been under work and a version is available from that bug, they took extra time since openttd developers use some funky multi-platform configuration stuff of their own making, etc. Maybe you could settle for dosbox with the old TTD MS-DOS version until they get everything sorted out with openttd's new version? :) -- Arttu V.
Re: [gentoo-user] python-updater broken
On 10/23/08, James [EMAIL PROTECTED] wrote: Hello, One of my systems in an old Pentium III (Coppermine). python-updater will not complete. It echoes one line and then sits indefinately (for a day or more). ... Here's what else happens: # eselect python list !!! Error: Can't load module python thoughts and ideas are most welcome. Some thoughts, not really informed, but trying to lure some better experts and ideas out: - Maybe you should run python-updater with several -v options? Maybe it will become more talkative? (See man page and/or the script itself) - Maybe the disks or filesystems have some problems or bit-rot after having possibly been powered on 24/7 for a long time? Has there been power outages? Voluntary fsck time! - Has something happened to your scanelf? emerge -1 pax-utils might fix some issues - I'd guess eselect python complaining like above is probably just normal, unless you're supposed to have eselect-python package installed (many systems probably don't and you say your system is sort of minimal anyway?) - Does anything unusual appear in /var/log/messages or other logs? -- Arttu V.
Re: [gentoo-user] Masking for Educational Purposes only ?
On 11/24/08, b.n. [EMAIL PROTECTED] wrote: Well, why? If you want to test 2.2 instead, you can do it. It's not blocking you from anything. Yes, before the change the only thing blocking portage 2.2 was ~arch. Now it requires the bigger tool, Thor's Hammer of Unmasking +5 (+15 against portage devs and other undead? ;) ). He wants users to test *more* a given version, and that in turn will probably benefit us all. It's a psychological trick, maybe, but an absolutely innocent one. In fact, once you know that, you *should* jump the bandwagon and help test the previous version. Let's think for a moment about the logic and psychology underlying this situation. Why have these people upgraded to unstable portage 2.2 in the first place? Just to test it out? To show off ricing to the debian and ubuntu kids running their ancient stable stuff? No, it's the package manager for crying out loud! My guess is that most have packages explicitly requiring them to run the newer portage, packages like anything related to kde4. Will such people help with testing of 2.1 on their normal system? No. So how many new testers did this actually acquire? And how much work was wasted unmasking the packages and ranting on mailing lists? My guess is zero new testers and a lot of time wasted with hundreds or thousands of people whipping out their text editors and unmasking the package, then going off ranting on mailing lists, forums and irc channels. So, yes, I'd tend to agree with Mr McKinnon and Mr Jarausch on this issue. Something was bad in this execution of this need more testers -- regardless of the possible original good intentions. /semi-serious-rant -- Arttu V.
Re: [gentoo-user] Re: Portage wants KDE4, I don't
On 11/25/08, Nikos Chantziaras [EMAIL PROTECTED] wrote: Zsitvai János wrote: On Tue, Nov 25, 2008 at 3:18 PM, Nikos Chantziaras [EMAIL PROTECTED] wrote: So you're saying I must spend 1 hour to mask all the stuff? (Not joking, the list is big, I have to do emerge -av world each time and see the package, mask it, emerge again, mask it, emerge again, mask it, ad infinitum. This is not funny :P) eix -C kde-|grep \[[IU]\]|awk -F '{print $2 :4.1}' /etc/portage/package.mask/kde4 package.mask is a file here, not a directory. Maybe you should use existing, old portage features: slots. Add :3.5 to the end of all those kde-base/-starting lines in package.keywords, and you'll accept unstable from slot 3.5 but not from 4.1 (current kde4 slot). You probably can use match/search-and-replace within your favourite text editor to get it to the end of all those lines without doing it manually. Your lines will look like this (but with slot 3.5 instead of 4.1): kde-base/superkaramba:4.1 Yes, that's the way they suggest unmasking stuff when testing out kde4 ... :) Slots 3 3 3 -- Arttu V.
Re: [gentoo-user] Update the clock using internet servers: recommendations
On 11/26/08, Dale [EMAIL PROTECTED] wrote: damian wrote: I agree. I been using ntp here and it works fine. If you need help configuring it, let me know. Off list if needed, just put Gentoo in the subject line. Ok, thanks Dale. But I can you tell me if there is any difference among ntp and htpdate? This is the first time I have heard of htpdate so I can't say anything good or bad. I know ntp works well here and the setup was not to hard. Run a couple commands and put the results in the ntp.conf and start the service. It just works here. Heck, I'm on dial-up and it works pretty good. If I had a always on broadband, oh boy!! Dale :-) :-)
Re: [gentoo-user] Update the clock using internet servers: recommendations
On 11/26/08, Arttu V. [EMAIL PROTECTED] wrote: And I beg your pardon for once again clicking on Send instead of Archive in gmail ... . -- Arttu V.
Re: [gentoo-user] qtiplot
On 11/26/08, b.n. [EMAIL PROTECTED] wrote: b.n. ha scritto: Hi, I have an x86 gentoo system, and I would like to install qtiplot. Unfortunately: - qtiplot 0.8.x requires qwt-4. I have both qwt-4 and qwt-5 installed, and when compilng qtiplot seems to pick invariably the qwt-5. How do I force qtiplot to build with qwt-4 ? Wild guesses, semi-tested on amd64 only: - in the current qtiplot-0.8.5-qmake.patch replace -lqwt with -l:libqwt.so.4 - check that you have qt:3 built with USE=opengl (- gcc-4.3 patching, which were a sort of a detour since I just didn't want to downgrade gcc in the chroot in which I made my testing in) After that I got qtiplot-0.8.5 emerged in a simplistic chrooted amd64 environment, which had little more than the basic qt stuff and explicitly emerged qwt 4 and 5: x11-libs/qwt selected: 4.2.0-r3 5.0.2-r1 However, I don't know this piece of software and I don't think I'm going to learn it right now, so that's as far as I planned on testing it right now. :) Maybe you would like to file a bug in b.g.o? I can dump my crude, couple-line patches/changes there and whoever needs/maintains qtiplot can then refine them or reject them as needed. -- Arttu V.
Re: [gentoo-user] Open Source Family Tree software?
On 11/29/08, Mark Knecht [EMAIL PROTECTED] wrote: Anyone have any recommendations of something in portage. (Or not in portage?) Wikipedia has some, they've even tried to categorize a bit and made some tables of the basic features: http://en.wikipedia.org/wiki/Comparison_of_genealogy_software I found gramps in portage. I did not find Lifelines. I haven't uncovered any other project names as yet. I set up my sister's old Ubuntu box to have gramps installed and (what I've heard) she has been using it successfully for her intermittent lightweight hobby genealogy studies of our relatives from previous centuries. I don't actually know much about these genealogy systems, so I don't know if such light-weight use is what you have in mind -- and whether gramps is a major beast or only usable for this kind of playtoy-use. -- Arttu V.
Re: [gentoo-user] Buying a low-cost printer for Linux
On 12/4/08, KH [EMAIL PROTECTED] wrote: Some weeks ago I bought Samsung clp-300 color laser printer for less than 130 Euros. I use cups and I don't have any problems. Did not have to by new color jet. I have been told I can print 7000 pages before I have to :-) You can go cheap and good. The model is from 2006 or so. I have bad experiences with Samsung, though not with that exact model. Samsung proprietary driver may be some sort of world record of ugly hacks and hairy stuff. See, e.g., Gentoo bug #139715 and all the complaints from Samsung-trying Linux-users in the blogosphere / Intertubes. But I'm glad if they've fixed their stuff recently. -- Arttu V.
Re: [gentoo-user] if I unmerge XYZ...
On 12/9/08, Andrew Gaydenko [EMAIL PROTECTED] wrote: Hi! Is there a way to predict which dependecies will be unsatisfied at case of unmerging some package without real unmerging? Not being quite sure what without real unmerging here means, I'd recommend checking out the man page of equery, especially the depends option. Maybe equery --indirect depends foo is what you're looking for? -- Arttu V.
Re: [gentoo-user] mplayer-1.0_rc2_p28058-r1 USE dvdnav
On 12/20/08, Willie Wong ww...@princeton.edu wrote: I think you can try use unmasking the flag: add a line to /etc/portage/package.use.mask It's profiles related stuff, so I think /etc/portage/profiles/package.use.mask will be the right place. At least if one trusts portage's man page. -- Arttu V.
Re: [gentoo-user] Hard to find netiquette, enculturation bug. (Was: Re: [gentoo-user] GNOME: Cant logout and Lock Screen is showing different background from GNOME screensaver)
On 12/22/08, Mark David Dumlao madum...@gmail.com wrote: The bug report is not about the clarity of an already-late reminder. The problem is that it was late for me, it will be late for a few hundred other posters after me, and it will be late for a few hundred posters after them unless something is done about it. Hence the bug report. Hopefully a presubscription FAQ or guide could at least curb such problems. But the confusing part here is that most people *will* get the warning from the oldtimers even when they've already transgressed against the netiquette. Even you got one now, once you didn't have a Gnome-specific subject like your emails in the archives. So, I have another theory, related to the cry Gentoo is dying! ${HERD} is down to one dev only!, i.e, low manpower. My guess is that Gentoo (or at least this gentoo-user list) is low on Gnome-proficient commenters. That would make the E17, Fluxbox, ratpoison and KDE-running oldtimers skip your (correctly Gnome-subjected) emails altogether and then there would be no Gnome-using-oldtimer to catch the ball. Does this sound plausible? -- Arttu V.
Re: [gentoo-user] mplayer-1.0_rc2_p28058-r1 USE dvdnav
On 12/22/08, Willie Wong ww...@princeton.edu wrote: HAHA! Arttu, you also have a typo! ;) The file should be in /etc/portage/profile/package.use.mask Note that 'profile' is singular, not plural. Sorry, must've been the fever and cold I've been having for the last couple of days. But true, fever or no fever, always check my typing, someone off the Intertubes might offer you some clever equivalent of rm -rf /! :D One thing to remember is that portage devs are probably ingenious and lazy (in the positive sense), so they probably do things in a similar way under /usr/portage and /etc/portage. This gives hints of the correct paths and file names, maybe even the syntaxes (although in this case the -dvdnav was a bit unexpected and illogical IMHO). Of course, /usr/portage has those different sub-directories for, e.g., base and many different platforms, but you can keep /etc/portage flat. Or this is just my deduction after running Gentoo for 5-6 years, I haven't actually touched the python code much and probably not pushed the configurations too far either. I'm very happy if someone points out a typo in my logic here -- I'm sure a little embarrassment early on will save plenty of my time in the future! :D -- Arttu V.
Re: [gentoo-user] Hard to find netiquette, enculturation bug. (Was: Re: [gentoo-user] GNOME: Cant logout and Lock Screen is showing different background from GNOME screensaver)
On 12/22/08, Volker Armin Hemmann volker.armin.hemm...@tu-clausthal.de wrote: no. His mails were ignored because nobody had an answer. Shown by the fact that nobody complained about his triple posting or html mails. This is my favourite conspiracy theory as well, not the no-html-theories. We've all been through situations where there just aren't any people with an answer. We've already collected all the html/no-html flamewar-stuff in this thread, so we may as well collect the All-Time Favourite Gentoo Complaints. First one I and (Mr. Wong?) already threw out: lack of manpower / knowledge base. Second favourite one is that the devs are reluctant to mix us proles (users). Well, it's a users' list with some volume, so I can understand that they don't. The Gentoo Gnome dev team is 7 or 8 people (same size as Gentoo KDE team) according to the dev list, so it's not as Gnome is dead on Gentoo. Third one is the confusion of where should I ask question? -- mailing lists, forums or just file bugs in b.g.o? Mr. Dimlao apparently only used this one mailing list -- and I can understand why, the forums are ugly and terrible to use (IMHO phpbb or whatever it is, is terrible to use). Still, unlike this list, the forums seem to be quite actively perused by the devs, so given how clear he is in titling his messages etc, he could've probably gotten a response there. I still don't buy this it's the HTML attachments-theory. I think many people have jumped on this html-attachment theory for no purpose (no other flamewars going on? ;) ). *If you're so sure it's the html in his emails, then why don't you answer his questions about Gnome?* Unless you can then I'd place a large probability on the theory of currently missing the required knowledge on *this* list. The answer may be available by asking in, e.g., the forums. -- Arttu V.
Re: [gentoo-user] Hard to find netiquette, enculturation bug. (Was: Re: [gentoo-user] GNOME: Cant logout and Lock Screen is showing different background from GNOME screensaver)
On 12/23/08, Volker Armin Hemmann volker.armin.hemm...@tu-clausthal.de wrote: On Montag 22 Dezember 2008, Arttu V. wrote: *If you're so sure it's the html in his emails, then why don't you answer his questions about Gnome?* well, I saw his mails - and I did not react to them, because I don't use gnome Sorry, that was meant for the subscribers of the html-attachment conspiracy theory, not you or your text personally. I agree, it was unnclearly typed and should have started with To all html-theory believers:, but I plead for my current slight fever and cough for minor typos and misthoughts. :) since the 2.0 disaster. Since then I am a happy KDE-only user. Not using gnome makes me skip gnome threads automatically - the few things I remember back from the 1.4 days won't help anyway. I run KDE as well (so maybe I'm biased to comment about gnome-related stuff?) and recently fixed a very trivial and minor bug in esound (part of Gnome, possibly outdated?). Happily filed a bug upstream. A month counting so far with no changes to the bug at all, while the Gentoo Gnome team threw the one-line patch in the tree within hours, I think. That's my recent experience with helping with gnome issues. Not really exciting, so I'm skipping the gnome-threads when busy, which is quite often. Gnome-help is under the rock someplace. KDE is known to break constantly (double that for KDE4), so perhaps it has more fixing knowledge spread out there? This still doesn't say that KDE is better than Gnome. When busy I just love using Ubuntu -- it shows what Gnome can be: no fuss, pleasant etc (but the OK-buttons are backwards! ;) ), while Kubuntu is ... a bit clunky feeling. -- Arttu V.
Re: [gentoo-user] Looking for an old ebuild: courier-authlib 0.58
On 12/30/08, Stefan Onken supp...@stonki.de wrote: Hello, for updating an old system I am looking for courier-authlib 0.58 (which still has vpopmail support). Unfortutely I don't find any ebuilds for it. Can anyone help ? Perhaps one of these will do: http://sources.gentoo.org/viewcvs.py/gentoo-x86/net-libs/courier-authlib/?hideattic=0 -- Arttu V.
Re: [gentoo-user] [cifs] permissions of mounted share
On 1/11/09, Harry Putnam rea...@newsguy.com wrote: Any regular files I create on the mounted share end up 744. Can I set a umask in the mount syntax or what do I need to do here to have the files keep the standard permissions? force create mode = 644 force directory mode = 755 Wouldn't that do it on the server so you wouldn't have to worry about the clients? Or have you already tried and it still fudges the access permissions? -- Arttu V.
Re: [gentoo-user] Re: package.keywords
On 6/22/09, James wirel...@tampabay.rr.com wrote: Where do I read more and find more of the latest example for syntax with portage and the different files therein? snip / It should have read: 'man portage' does not show the use of the (=) syntax within the package.keywords file ? No, but it instructs one to peek into man 5 ebuild for the info. If you look again at the portage manpage, the section on package.keywords states of the file format: one DEPEND atom per line followed by additional KEYWORDS And at the top (Glossary) it says about these atoms: DEPEND atom A string which matches a package. It is of the form category/package. It may also contain optional logical operators and versions. More reading: ebuild(5) Follow this more reading-hint by entering man 5 ebuild. The resulting manpage tells you quite extensively (even everything?) about the atoms, with many examples, and more. It takes a meticulous, technically inclined (or otherwise anal-retentive? :) ) mind to read these manpage-thingies which were structured and invented before the modern hypertext-intarweb hit the streets two decades ago. And sure, some of them are written badly, but IMHO the gentoo ones are pretty ok from a user's POV. So, as a conclusion, you probably want to use ~ instead of = in there as you apparently are running a mostly stable box (arch) instead of testing (~arch)? -- Arttu V.
Re: [gentoo-user] Re: package.keywords
On 6/22/09, James wirel...@tampabay.rr.com wrote: Arttu V. arttuv69 at gmail.com writes: More reading: ebuild(5) Ah, ok so there is not restriction on using any of the the boolean operators in any config file underneath /etc/portage? as section 5 does not mention any Well, for those files that are actually supposed to contain DEPEND atoms that is probably true. I haven't checked the whole man page and all files it lists to see if it places any restrictions on some specific files therein, so I'm a bit timid to claim full no limits either. But for package.keywords it didn't mention any limits and I'd believe the others don't have them either. :) Still, the best identifier of the allowed syntax is probably portage itself. Make a typo in files under /etc/portage and portage will tell you next time you emerge -pv something or emerge -av something. And it also breaks in a quite predictable fashion when it reads your typos in those files: it will just drop the lines it doesn't understand (while also printing warning messages) and then dutifully proceed to next line. So if ever in doubt, you can try it and see what portage's parser thinks about your fuzz-test. :) This is my (mis)conception, although, as you have suggest, there are (gentoo) cultural norms that do suggest certain boolean operations should not be used, in say for example, package.keywords? Not sure if by cultural norms you're referring to something I'd rather label as best practices? :) But then again, there are best practices for bleeding edge folks and best practices for stability is paramount folks and they may overlap, but most likely they aren't quite the same. I'm mostly running stable with exceptions being enabled via the /etc/portage file structure. Usually it's small, but now with kde4, BLOAT is my modus operandi, not by choice.. Then, again, you probably don't want the = as it will unmask the latest testing version currently available, plus in future any new testing grade versions as they get added to portage tree. What you want is either = or ~. (And do note that ~ at the front of a DEPEND atom means a different thing from the testing keyword in ~arch). But it is all up to you to decide, you're the commander'n'chief of your own boxen. -- Arttu V.
Re: [gentoo-user] [OT] Use of sed
On 6/24/09, Neil Bothwick n...@digimed.co.uk wrote: On Wed, 24 Jun 2009 00:48:07 +0100, Peter Humphrey wrote: And while I'm at it, how do I change the field separator from / to enable me to search on that character? By using something else, you don't need to tell sed, it works it out for itself, just use something that isn't in your search string, : is a good candidate. If I read his question right, he asked about just the simple matchers: //. Perl solves this problem with the optional m in front (m//), so you can do m:/foo: or m+/foo+, but I don't know of a similar toggle for sed (well, I'm a sed newbie, so there might still be one). I don't even think substituting the string with itself (s+/foo+/foo+) would work as I think s/// will succeed every time, even when it doesn't actually substitute anything, so maybe it cannot be used for an if-then in sed either? -- Arttu V.
Re: [gentoo-user] problem reading zip files
On 6/25/09, Doug ONeal on...@dbi.udel.edu wrote: I am guessing that all of the data is actually present in the zip archives but I cannot get to it. Trying version 3.0 of info-zip or any of the other linux windows unzip programs has not been successful in extracting any of the later files in the archives. I'm at a loss; does anybody have suggestions on how to retrieve this data? Thanks. Does any of the other unzip programs here contain java's jar as well? -- Arttu V.
Re: [gentoo-user] plainTeX instead of LaTeX
On 6/28/09, meino.cra...@gmx.de meino.cra...@gmx.de wrote: All my *.tex-files are handled as they would be written in LaTeX. But they are good old plainTeX. or in other words: tex texfile.tex ends up in errors due to the fact, that texfile.tex contains syntax with LaTeX dont understand. Fo example a missing documentstyle. man tex When called as virtex it will use the plain format. Not sure if this is what is called plainTeX, I've only used LaTeX. -- Arttu V.
Re: [gentoo-user] plainTeX instead of LaTeX
On 6/28/09, meino.cra...@gmx.de meino.cra...@gmx.de wrote: Arttu V. arttu...@gmail.com [09-06-28 13:27]: man tex When called as virtex it will use the plain format. Not sure if this is what is called plainTeX, I've only used LaTeX. Yes, it is... tetex leaves the names of the executables untouched... Ok, here is what I got: /home/mccramervirtex texfile.tex This is TeXk, Version 3.1415926 (Web2C 7.5.7) %-line parsing enabled. kpathsea: Running mktexfmt plain.fmt I can't find the format file `plain.fmt'! Could they be hidden in the USE=extra as Debian seems to have something similar sounding in such a package: http://packages.debian.org/fi/lenny/texlive-formats-extra texsis -- Plain TeX macros for Physicists. Unfortunately I have USE=-extra as they caused some trouble on Gentoo and I didn't really need them at that moment. -- Arttu V.
Re: [gentoo-user] Kbackup and dvd sized slices
On 6/28/09, Dale rdalek1...@gmail.com wrote: A box should pop up and if you click the little button under Maximum Archive Size, it should show the available options. Mine has unlimited, 650Mb CD, 700Mb CD and custom. The custom will not let me enter anything in Gb. If by Gb you mean the spinner to the right of the dropdown stating custom then yes, over here it lets me type in numbers (2000), and automatically adds MB once that widget loses focus. But I just installed kbackup-0.94-r1 from scratch on this amd64 box two minutes ago. So, e.g., my ~/.kde/share/config/kbackuprc only has the window dimensions saved. Maybe yours has some more settings saved in there? There is really not a USE flags to change on this one. Ideas? Someone confirm it is just me or not just me? kbackup uses kde 3.5 and I've noticed that on some of my boxes they've been rebuilt quite often over last few weeks, alongside the underlying qt libraries (which I presume provide the base of those widgets as the underlying toolkit). So, maybe you can try re-emerging kbackup and/or check the sanity state of the whole thing with revdep-rebuild? -- Arttu V.
Re: [gentoo-user] Kbackup and dvd sized slices
On 6/28/09, Dale rdalek1...@gmail.com wrote: Mine will only let me type in 3 numbers. Examples: 300 or 600 It will not accept Gb or anything over 999 Try something starting with a '3', you should be able to feed in four numbers. Anything starting with a four or greater will only get three numbers here as well. Weird, there must a limiter somewhere in the source code itself ... (Moments later; suggested background soundtrack: Homer Simpson saying repeatedly d'oh!) Yes, it is actually limited in the SettingsDialog.cxx code itself: maxSliceSize = new QSpinBox( groupBox2, maxSliceSize ); maxSliceSize-setEnabled( TRUE ); maxSliceSize-setMaxValue( 4096 ); maxSliceSize-setMinValue( 0 ); Can you patch that yourself to have a suitably big number where it says 4096? I have no idea what might break, but that's where the limit comes from. ebuild unpack and see for yourself in /var/tmp/portage/app-backup/kbackup-0.5.4-r1/work/kbackup-0.5.4/src/SettingsDialog.cxx, line 62 ... :) -- Arttu V.
Re: [gentoo-user] [WAY OT] GUI programming for Linux (and Windows possibly)
On 6/28/09, Mark Knecht markkne...@gmail.com wrote: Actually I'm liking the suggest to try using R. I have already managed to read my data files using the read.csv function. When I understand headers and tables better I'll likely be able to make my plots from that data pretty easily. It's cross platform so it solves that problem and keeps me focused on where I might add value - evaluating the market data - and not worrying about how to program in C or Python. Open to other ideas but this one is looking pretty good to me so far. If you have your data in a text-like, tabular format that R likes then you might also try feeding them into a data mining package like Weka. Depending on what kind of analysis you are looking for, it might be even better (and easier) than R -- or entirely unsuitable. :) -- Arttu V.
Re: [gentoo-user] Kbackup and dvd sized slices
On Sun, 2009-06-28 at 15:58 -0500, Dale wrote: Arttu V. wrote: maxSliceSize-setMaxValue( 4096 ); Ahhh, so it isn't just me. That's cool. I'm not sure about editing anything tho. I tend to break stuff. I'm just wanting to really fill up a DVD is all. I wonder why they changed that? Confusing. They seem to have an explanation on the webpage: http://www.kde-apps.org/content/show.php?content=44998 The 4GB limit is due to Qt3 limits on 32-bit architectures. They don't mention if it affects also 64-bit systems. But in other news there, they seem to have released kbackup 0.6 just a month ago, which incidentally among other things removes the 4GB limit: - Port to cmake and KDE4/Qt4, which also removes the 4GB limit per file -- Arttu V.
Re: [gentoo-user] Kbackup and dvd sized slices
On 6/29/09, Arttu V. arttu...@gmail.com wrote: But in other news there, they seem to have released kbackup 0.6 just a month ago, which incidentally among other things removes the 4GB limit: If someone is brave and stupid enough to try, I tried to modify the ebuild from 0.5.4-r1 to compile 0.6. Bug #275824 Unfortunately, kbackup-0.6 still seems to have the 4GB limit for the custom size setting, but it also has some standard DVD sizes to choose from which go well beyond in size (17.1GB is the largest -- and then there's always the unlimited). -- Arttu V.
Re: [gentoo-user] Persistent conflict between readline 5 and 6
fe...@crowfix.com wrote: This has been going on now for months. Is there some reason these two can't co-exist? Or is there some better fix than just ignoring it or adding an entry to package.mask? !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: sys-libs/readline:0 ('ebuild', '/', 'sys-libs/readline-5.2_p13', 'merge') pulled in by =sys-libs/readline-5* required by ('installed', '/', 'dev-lang/ghc-6.8.2', 'nomerge') (and 54 more) ('installed', '/', 'sys-libs/readline-6.0_p3', 'nomerge') pulled in by =sys-libs/readline-4.1 required by ('installed', '/', 'dev-db/unixODBC-2.2.12', 'nomerge') =sys-libs/readline-4.1 required by ('installed', '/', 'net-misc/socat-1.7.1.1', 'nomerge') sys-libs/readline required by world (and 51 more) I don't use haskell, but ghc's v6.8.2 ebuild seems to have been silently bumped [1], which changed among other things the dependency on readline to be more lenient upwards. You probably have the old one installed and as the bump is silent, portage doesn't re-emerge ghc automatically. My suggested fix (hopefully fixes this, but I'm not 100% certain): emerge -1 ghc, so that portage will realize the changes from the silent bump. [1] http://sources.gentoo.org/viewcvs.py/gentoo-x86/dev-lang/ghc/ghc-6.8.2.ebuild?rev=1.11view=log HTH -- Arttu V.
Re: [gentoo-user] [OT] Help tweaking an ebuild
On 7/12/09, Mick michaelkintz...@gmail.com wrote: a)How should I specify the URI for the source file if it is already copied into my /usr/portage/distfiles? Add RESTRICT=fetch, plus add an informational function pkg_nofetch. Search for them for example here: http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2chap=1 or on the man page (man 5 ebuild). b)Trying to create a manifest gives me errors: # ebuild rtmpdump-1.2.ebuild manifest /usr/local/portage/media-libs/rtmpdump/rtmpdump-1.2.ebuild: line 34: unexpected EOF while looking for matching `' I believe the problem spot could be anywhere before that line. That line might just be the last one with a quote char, so there won't be any later quotes in the file to match? Thus, check all lines before that one. -- Arttu V.
Re: [gentoo-user] Hard masked package
On 7/17/09, Mick michaelkintz...@gmail.com wrote: Hi All, I didn't have to do this before, so will need help to manage portage: A particular package has been hard masked for removal. I would like to either move it in my local portage, or even better emerge a more recent version. So, I found an ebuild, modified it a bit and now I am trying to emerge it, but it seems that the package.mask won't let me: = !!! The following installed packages are masked: - app-pda/plptools-0.13 (masked by: package.mask) /usr/portage/profiles/package.mask: # Jeremy Olexa darks...@gentoo.org (30 May 2009) # Masked for removal in 60 days. Doesn't build, version bump pending. Designed # for really old PDA hardware (circa '90s) - will need a maintainer that has # hardware. bug 227601 = I am trying to emerge v1.0.8, not 0.13 as shown above. Should the above mask stop me emerging it from my local overlay? How do I allow a local overlay emerge, while respecting portage's hard mask for ebuilds in the main tree only? man portage: package.unmask Just like package.mask above, except here you list packages you want to unmask. Useful for overriding the global package.mask file (see below). Note that this does not override packages that are masked via KEYWORDS. So, add app-pda/plptools to /etc/portage/package.unmask to remove package.mask blocks from affecting the emerge calculation for that package -- and if you're running stable arch and the package only has ~arch, also add ~arch for your package in /etc/portage/package.keywords (or edit your own ebuild, as in your own overlay you can decide if it is stable or not). If portage still doesn't offer to emerge v1.0.8 after doing those you still may have something wrong with your local overlay / ebuild. Did you digest/manifest the ebuild you copied to the local overlay? Is the local overlay added properly in /etc/make.conf (no typos in path etc)? -- Arttu V.
Re: [gentoo-user] Multiple instances of PyQt4
On 7/21/09, Alan McKinnon alan.mckin...@gmail.com wrote: On Tuesday 21 July 2009 17:11:33 Alex Schuster wrote: Hi there! My @world update shows this problem: !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: dev-python/PyQt4:0 ('ebuild', '/', 'dev-python/PyQt4-4.4.4-r5', 'merge') pulled in by =dev-python/PyQt4-4.4.0[X] required by ('installed', '/', 'kde- base/plasma-workspace-4.2.4', 'nomerge') dev-python/PyQt4[X] required by ('installed', '/', 'media- sound/picard-0.11', 'nomerge') =dev-python/PyQt4-4.4 required by ('ebuild', '/', 'dev- python/qscintilla-python-2.3.2-r2', 'merge') (and 1 more) ('ebuild', '/', 'dev-python/PyQt4-4.4.4-r2', 'merge') pulled in by =dev-python/PyQt4-4.4.4-r1[dbus,qt3support,svg,webkit,X] required by ('ebuild', '/', 'kde-base/pykde4-4.2.4', 'merge') (and 4 more) I would like to understand where the problem is. The bottom says that pykde4-4.2.4 needs at least PyQt4-4.4.4-r2. This is fine, I can see the corresponding DEPEND line in the pykde4-4.2.4 ebuild. The top list tells that: plasma-workspace-4.2.4 needs PyQt4-4.4.0, picard-0.11needs PyQt4 and qscintilla-python-2.3.2-r2 needs PyQt4-4.4 Why then is PyQt4-4.4.4-r5 needed? -r2 should satisfy these needs, too, doesn't it? Or is the problem the 'and 1 more' line? How would I find out which other package would need -r5 then? And why would the other line be listed then? I don't get this. I grepped for PyQt4-4.4.4-r5 through all ebuilds, it appeared only in the PyQt4-4.4.4-r5.ebuild itself. This is not an important problem - I think I can unmask some stuff and it will work. But wanted to cut the amount of masking down a little, and I'm just curious why this happens. So, thanks for any insights, if want to practice your portage-fu, but don't bother if you have more important things to do, I'm fine :) There was issues with pykde4 recently resulting in these blockers. Easiest way out is to remove all local masks you might have for PyQt and pykde4, unmerge both and emerge -uND world. Let portage sort it out. IIRC you can get by by just unmerging one of them and it works. But, I'm too lazy now to figure out which way round it goes, and memory ain't what it used to be. PyQt4-4.5.1 ebuild was removed from the portage tree over the weekend. I had only that specific version unmasked (IIRC kde4 dependencies?), and ran into a similar(?) situation yesterday. Maybe you hit the same? (Similar situation == portage surprises and wants to downgrade PyQt4 to 4.4.4-something.) No worries, IIRC I just unmasked 4.5.2 and everything has been smiles and sunshine again afterward. -- Arttu V.
Re: [gentoo-user] emerge kdirstat fails
On 7/27/09, Nicolai Beuermann nicolai.beuerm...@gmx.de wrote: Hello, emerging kde-misc/kdirstat-2.5.3-r1 doesn't compile due to this error: This bug has someone fighting with this package with moderate success (uninstalling : http://bugs.gentoo.org/show_bug.cgi?id=248883 -- Arttu V.
Re: [gentoo-user] emerge kdirstat fails
On 7/27/09, Arttu V. arttu...@gmail.com wrote: On 7/27/09, Nicolai Beuermann nicolai.beuerm...@gmx.de wrote: Hello, emerging kde-misc/kdirstat-2.5.3-r1 doesn't compile due to this error: This bug has someone fighting with this package with moderate success (uninstalling : http://bugs.gentoo.org/show_bug.cgi?id=248883 Grumble, grumble, accidental send-click again in gmail (sorry). Meant to type uninstalling qt 4, then emerging kdirstat, then re-emerging qt 4 in the parenthesis. -- Arttu V.
Re: [gentoo-user] Re: Strange issue with YouTube not working at all in some browsers
On 8/1/09, Remy Blank remy.bl...@pobox.com wrote: Paul Hartman wrote: My ISP's DNS servers have 3 strikes against them: 1. It is slow, slow, slow, slow, slow... and did I mention slow? :) 2. They have previously sold user's DNS/browsing history to advertisers. They claim to have stopped, but... 3. They hijack DNS, making every invalid address resolve to an address anyway, which when viewed in a web browser goes to an error page (full of advertisements and sponsored links). You never know if a hostname is really invalid or not, which makes troubleshooting non-HTTP connections interesting. Time to switch to http://www.opendns.com/ ? (I haven't used them, but they seem to be recommended a lot theses days) Which part of the problems he lists are not associated to OpenDNS as well? :) http://en.wikipedia.org/wiki/OpenDNS#Privacy_issues.2C_conflicts_and_covert_redirection -- Arttu V.