Re: [gentoo-user] Upgrading kernel
Grendel wrote:
People dont realise that this code leak is a blessing, we can analyse and
figure out previously hidden/unknown API's from the source code.
Actually, no, it isn't. Be prepared that MS will sue the crap out of any
kernel developer who they
think may have
a) incorporated windows source code into the Linux kernel
b) has violated MS's copyright by obtaining the windows source code
They could also try to (at least) create bad press for Linux by alleging
"theft of technologies", which
would probably be of no legal value (if they have a patent, it was
already public anyway), but
tell that to the average IT yellow press editor.
IMHO linux needs ntfs write support, with many people dual booting betweem
windows and linux.
IMHO, that is clearly the worse of two options. We don't need a full
NTFS driver for Linux, we need
full Windows drivers for filesystems typically used under Linux, e.g.
ext2/3 and reiserfs, later also
xfs and jfs. There are already utilities that allow file exchange with
such partitions under Windows,
however not transparently (with the same API as ordinary file access).
Why this way round, and not
the other?
- Most importantly: No hassle with Microsoft's "IP" department, as much
as they would like to harm Linux.
- Implementational difficulty: Though the NT filesystem API is reported
to be horrible to code drivers for, I guess
it is still easier than to reverse engineer NTFS further. (Mind you that
downloading the NT/2000 source
code is *illegal* and thus not an option.)
- Technical reasons: E.g. reiserfs has better performance and less
trouble with fragmentation than NTFS
(less features though, AFAIK, but the average dual booter will not miss
'em).
- The shared data goes on a ("real") Linux, not a Windows partition.
This means you have to move
less data should you finally decide to purge your Windows installation
entirely.
I can think of some advantages of the NTFS approach, but none that seem
to be able to outweigh
the disadvantages.
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] Update problem
Jani-Matti Hätinen wrote: Paulo da Silva kirjoitti (sunnuntai, 15. helmikuuta 2004 17:56): Now I would like to upgrade (emerge -Up world) but then emerge trys to revert to kde 3.1.5. How do I avoid this? emerge sync && emerge -Up world KDE-3.2.0 has been in the stable branch for some time. Mind you, the new qt will cause some problems (as you have probably noticed from reading the list). May cause some problems, it works fine on my machine. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] CPU Overload? IN KDE-3.2.0?
Thomas T. Veldhouse wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Is anybody getting the CPU overload warning strange audio problems with KDE 3.2.0? I am finding it quite annoying. I don't seem to have any problems with Gnome 2.4.2 and using esound. I have tried this with both the latest Gentoo sources and ALSA and with the mm-sources 2.6.3-rc1 kernel (also with ALSA). I used to have this problem too, when using OSS in combination with artsd. When one application already uses the OSS device (/dev/dsp), and artsd wants to play a sound, it seems to enter an infinite polling loop for a lock on /dev/dsp, or some sh** like that. CPU load skyrockets, the system freezes, until after half a minute or so artsd crashes with "CPU overload! Aborting." and the system becomes useable again. (Though, of course, arts sound doesn't work anymore until artsd is restarted). Very disturbing. I hope KDE makes artsd optional, and offers ALSA support in the future. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Which kernel version to select?
Grendel wrote: Thanks for the help, it is much appreciated. So currently the installation worked ok. It was hard to install it, but the end product is worth the sweat. ACK! Only a couple of more problems left, 1. The nvidia-kernel got compiled and installed but Xfree86 was unable to find the binary X module, a situation which I rectified by downloading the NVIDIAxxx5336.run from nvidia.com and installing it. You need both the nvidia-kernel and nvidia-glx packages. nvidia-kernel provides the kernel module, nvidia-glx the userspace libraries. Alternatively, the NVIDIA...run scripts provide both. 2. Surround sound (5.1 or greater)in linux, has anyone tried this on any of your sound cards? Audigy, sb live 5.1, or nforce2 (a7n8x and other motherboards based on the nforce2 chipset can have surround sound). To enjoy surround sound you will have to try to play a dvd with mplayer dvd:// --channels 6 and see if it works. Well, I have a card (Terratec DMX 1024) that supports double stereo (front and back), so it should be possible to configure it for 4 channels. However, I didn't try this yet. If you figure it out, could you share your experiences? -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Which kernel version to select?
Grendel wrote: Currently I got the 2.6 kernel installed, but genkernel asks me to pass the options "boot=/dev/ram0 real_boot=/dev/hda6", I tried but it dont work so I switched back to 2.4.20. Is there a way to get newer kernels ? I stick with www.kernel.org. Don't use genkernel, unless you configure your system exactly the way genkernel expects it, you will probably run into trouble. With kernel 2.6 they dumped "make dep" anyway, and "make" builds the bzImage and the modules - this makes kernel compilation so simple that there shouldn't be any need for a script doing obscure things. To install, I need one cp, one lilo, and one "make modules_install" command. Not overly difficult either! ;-) I have no need for a boot partition or initrd either, my sys is booting right off a 75GB reiserfs partition, with reiser support compiled into the kernel and just about everything else I need as a module. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Win NT & 2k
Greg wrote: Seriously the only real value I see in this is for someone to search for security holes by looking at the source. Otherwise I see no reason anyone should be interested in Mickey$oft code. Greg Doesn't "sheer curiosity" count as a reason then? Though I wouldn't want to risk a lawsuit just for curiosity, so I'll just wait for insider comments! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Win NT & 2k
Matt Garman wrote: On Fri, Feb 13, 2004 at 12:34:32AM +0100, Spider wrote: other technologies that can be seen as "infringing" on some of MS turf. Its an easy, blunt lega weapon to weild against whole projects, stating that all progress past this point is only because you copied MS sources, weilding a large batch of patents, tradesecrets and copyright infringement claims and slamming large and wide. What scares me, truly, is that it only takes one bonehead to do a copy and paste job from the MS code to a (previously) legitimate OSS application. And, as you say, that gives Microsoft, with their vast financial and legal resources, an opportunity to start doing a lot of damage to the OSS world. I'm probably overly paranoid, but part of me is thinking about a conspiracy. Even if it's not a conspiracy, you've got to admit that Microsoft is pretty good at taking advantage of circumstances, and working things to their favor. I'm sure someone there will figure out a way to use this "opportunity" to kill Linux---as many have suggested, Microsoft's (and SCO's) only way to beat OSS is through manipulation of the legal system. Yep. They will stop at nothing. Well, physical assaults, maybe. :-( I hope this isn't the first step towards the criminalization of Linux. It's the second step after SCO. At least it seems SCO are going to lose their poker game. I've often thought that the best way to get a lot of converts to Linux is for OSS to offer the next "killer app". My vision of the next killer app is highly integrated, "ubiquitous security"---systems that are by default completely secure, but the security implementation and infrastructure is transparent to the users; systems that are *by design* inherantly secure, so they can be open as well (kinda like "open DRM"). I think a component of such a system would be unmistakeable, impossible-to-forge digital audit trails and authentication systems. I'm thinking of *at least* making it a legal requirement to acquire Stamper-like* certificates for your source code if you want to enjoy copyright protection. Submitting hash codes would be enough, no need to hand in any actual source to a government agency or other organization. The beauty of the Stamper approach is, apart from its simplicity, also its inherent trustworthiness! If enough people regularily download stamper logs (this can be automated with wget, e.g.), there is no way for the key owner to produce backdated certificates without getting caught. *) see http://www.itconsult.co.uk/stamper.htm So then, if 2 people/projects claim the copyright to the same source code, the party that can provide an older certificate of a source archive containing said code, wins. Unless the other party has *very* convincing evidence to the contrary. Since it is *very* unlikely that a "code thief" will manage to register the code before the lawful owner, this would be highly secure. This would solve the company-puts-open-source-into-their-cvs-and-claims-its-theirs problem. This would serve a huge purpose for OSS: accountability, and and easy means to verify source code (who made it, where it came from, etc, etc). The intent is to help OSS "prove" that it is legitimate, to avoid SCO-like fiascos. What exactly are you trying to get at? Is it a system that allows CSS authors to check whether an OSS project incorporates their code? This would be IMO a good thing to have, especially since it can likely be very well automated, and wouldn't force any company to hand out source code of their own. Just some system would send new code snippets of all OSS projects that want to benefit from this protection to CSS subscribers who can then run them for pattern matching against their own archives. Hopefully this will not produce too many false alarms. If this works, I would make it legally clear that any copyright claims against these OSS projects must be made *immediately*, lets say in at most 4 weeks, or else it counts at least as a limited license to continue using the source code until a replacement can be provided. This may seem a bit harsh, but all it would take would be a 'that's our code' notification to the project maintainers. A stamper certificate could be used to have a lay assessor confirm (or deny) the claims, and the OSS project maintainers would then immediately be able to remove/replace the code. Chances are that it will not be in great use yet until then, and if, that the snippets are so few that they're easily replaced with legit code. Let me just point one thing out: This is *not* about protecting code thieves from legal action, both civil litigation and criminal prosecution. Instead, it is about protecting OSS project maintainers and users who, in goodwill, trusted the offending contributor to put legally waterproof code into their source repository when in effect he was not. These should not have to face damages for years of unwitting use of illegal code. Much of that in
Re: [gentoo-user] OT: Win NT & 2k
Jakub Krajcovic wrote: Hi guys, i know this is totally OT, but if you don't mind, check out slashdot concernig %subject GPG public keys available at pgp.mit.edu Old news, really. The Win2000 core source code has been on the net for years. -> http://www.databasejournal.com/features/mssql/article.php/10894_1488551_2 -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] unable to check root filesystem
Collins Richey wrote: On Fri, 13 Feb 2004 16:13:46 +0100 Christoph Gysin <[EMAIL PROTECTED]> wrote: my system doesn't boot correctly. it keeps hanging because checkroot complains not being able to check the root fs. [ rest snipped ] This won't help, but your experience only comfirms my decision never again to entrust data that I love to reiserfs. I know that almost everybody loves reiser, but I've experienced and read about too many failure scenarios. Good luck; maybe some of the reiser gurus can help you. In effect, this would have happened with any FS, see Mike's post. I'm using reiser since >3 years without any trouble. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] interesting... :")
raptor wrote: did u tried this : emerge -p info try it Yep, thanks! The -p option is superfluous, though. See 'man emerge': info This is a list of information to include in bug reports which aids the developers with fixing any problems you may report. Please include this information when submitting a bug report. Expanded output can be obtained with the --verbose option. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Opera-7.23 ebuild question
Alex Efros wrote: Hi! I's anybody knows why Opera require Openmotif? I've used different versions of Opera (including 7.23) a couple of years from 'opera*static*rpm' without any (Open|Less|...)tiff installed (on my LFS-based distro which I've used before hear about Gentoo). That's strange... Opera does not have any dependencies according to http://www.gentoo-portage.com/browse-program.php?program=3920 Gentoo uses the statically linked version of 7.23. Second question: why 'emerge -pv' say I still need to download 9MB Opera source AFTER I've downloaded and even merged Opera (and also Opera size is only about 4.5 MB, not 9.5): The >9MB download size looks like a bug. On my system, emerge -s opera says * net-www/opera Latest version available: 7.23 Latest version installed: 7.23 Size of downloaded files: 14,147 kB Homepage:http://www.opera.com/linux/ Description: Opera web browser. License: OPERA but when I "ls -l /usr/portage/distfiles/*opera*', I only get opera-7.23-20031119.1-static-qt.i386-en.tar.bz2, which is only about 4.8MB large. home / # qpkg -I opera net-www/opera * home / # emerge -pv opera These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild R ] net-www/opera-7.23 -gnome -kde 9,458 kB Total size of downloads: 9,458 kB unless you give the -u or -Uu options, portage will want to reemerge already installed packages, which is indicated by the 'R' marker. (An 'I' would stand for installing a new package, a 'U' for updating a package). E.g. "emerge -Uup opera" should output: These are the packages that I would merge, in order: Calculating dependencies ...done! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] I am stuck, please help
Radu Filip wrote: Hi, I am trying to install a Gentoo box from scratch and I cannot get over a large package (kdebase): [...] ./usr/share/pixmaps/kde2.xbm ./usr/share/pixmaps/splash2.png Done. extracting info extracting kdebase-3.2.0 tar: Skipping to next header tar: Archive contains obsolescent base-64 headers bzip2: Data integrity error when decompressing. tar: Error exit delayed from previous errors !!! Error extracting /usr/portage/packages/All/kdebase-3.2.0.tbz2 This errors appeared on other large packages like xfree86 or kdelibs but after a number of retries it worked. On this one, however, I am stuck :-( I suspect that the source package is defective. Delete (or rename) it in /usr/portage/distfiles, and try again - emerge will then download it anew. If emerge gets it from CD, remove the CD as a package source, at least for this package. (If you find it more convenient, "emerge -f kdebase" will only fetch the source, you can then compile and install it with "emerge kdebase" rsp. "emerge kde" later.) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] two Q about kernel-2.6.3-rc3-gentoo
Chris wrote: I finally figured out why I couldnt get kernel-2.6.X to work on my sys. It was a typo on my part, I could have sworn that I had put notail for my / - reiserfs partition when it was actually tail that I had put down. :( Q #1 When I chose the sound drivers I did not select alsa and when I booted into 2.6.3 there was no sound just and error message stating that /dev/dsp did not exist. How do I fix this? Did you choose the appropriate OSS module then? Did you check whether it's loaded at boot time? If not, put its name into /etc/modules.autoload.d/kernel-2.6. However, I'd suggest that you give ALSA a try - it has some nice features (OSS compatibility with alsa-oss, and realtime mixing of *all* sound output, so you can play MP3s while playing a game, for instance :-D ). Q #2 How do I get the hcfpcimodem-0.99lnxbetta03042700 for conexant chipset to work with kernel 2.6? Sorry... no idea. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Alsa & 2.6.1
Grendel wrote: On Fri, 13 Feb 2004, Arne Vogel uttered the following immortal words, Emerge alsa-oss, or emerge alsa-xmms and switch to the xmms alsa output plugin. I would recommend using alsa-oss anyway as a lot of audio utilities count on OSS support to be present. Yep. Apart from that, alsa-oss and alsa-xmms can be installed both, so you can have OSS emulation for the apps/games/whatever that need it, and use xmms with all features of the ALSA API (though I'm not aware of any great xmms-related benefits). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Alsa & 2.6.1
David Stevenson wrote: I started with 2.4.22, emerged alsa sound and it worked fine. Then I emerged 2.6.1-mm1 and lost alsa. I have now been through menuconfig adding alsa-oss modules, and redoing make modules modules_install and now it looks like it is working, but no music. The gnome volume mixer panel appears to be OK, and I get loud clicks as I unmute / mute Vol (and line 2 ? ) but when I play a .ogg xmms opens a looks like it is playing but no sound. I do have some sound from gaim, select sound events / test plays the bell sound quietly. The pcm fader on the mixer turns this off, but the main vol has no effect. If xmms is playing, then this gaim effect sound is masked - silent. Does this odd nearly working make any sense? David Emerge alsa-oss, or emerge alsa-xmms and switch to the xmms alsa output plugin. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] / partition full after emerge
Priit Külaots wrote: Tiago Lima wrote: Hi, I have my / 9Gb partition full after emerging a large number of packages... What can I safely remove? I think my /var/tmp is responsible... with 2.5Gb... and with openoffice I guess... Tiago Lima It's safe to remove anything in /var/tmp/portage and /usr/portage/distfiles . distfiles are tar.bz2 packages from what emerge does installing. Also, you can delete everything in /usr/portage/packages/All. That's where portage puts binary builds (in case they're enabled in /etc/make.conf or you explicitely requested them with the -b or -B emerge options). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] gcc-config
Mike wrote: I have three versions of gcc installed on my system: 2.95.3, 3.2.3, and 3.3.2. I'm using gcc-config to select a version. gcc-3.3.2 was recently installed. But, I'm still using gcc-3.2.3 most of the time and experimenting with gcc-3.3.2. Now, after ever package emerge portage wants to unmerge gcc-3.3.2. If I switch compilers to gcc-3.3.2 then portage will try to unmerge gcc-3.2.3. How can I tell portage to keep all three versions? Thanks in advance, Mike Try adding them to /var/cache/edb/world. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] udev & devfs - possible conflict?
Jakub Krajcovic wrote: Well, i couldn't wait for your reply so i learned the hard way... After unmerging devfsd and merging udev, i could not connect using ppp (ISDN) - and whoa: *user* panic. and since i know nothing about udev config, i came back to devfs. Oh... that's too bad! *supressing my glee ;-)* I'll definitely not emerge udev again until I'm sure I can spend some time figuring it out. Also, next time I'll just "ls -lR /dev >~/mydevices" when devfs is running, that way I know the majors and minors in case some device is not found without having to reboot. Then I can at least mknod until I'm through the udev documentation. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] perfomance 2.6.3 <= 2.4
Guy Van Sanden wrote: I tested the responsiveness of the GUI by starting the same DIVX rip on dvdrip under both kernels. In 2.4, I can still launch Mozilla and k3b, although it takes them a few seconds longer. (about 17-25 seconds to fully load) On 2.6, nothing on my desktop is visually responding to me. If I click a program, I see nothing happening for about 30-45 seconds, for k3b even longer. It takes up to a minute for the mozilla window to appear. I can access already opened programs though. Check whether you have preemption and DMA enabled. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] perfomance 2.6.3 <= 2.4 - A-FRIGGIN-Men!
TriKster Abacus wrote: Either way, these 2.6.X zealots are full of crap! I could probably write a small essay over the difference between speed and responsiveness, but I think I'll just plonk you. PS: "tar xjf linux-x.y.z.tbz2" made my system basically unusable for half a minute under 2.4.24. No longer under 2.6.1. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] udev & devfs - possible conflict?
Jakub Krajcovic wrote: Ok, thanks a lot, that really helped. But now the question is: do i get udev wortking by just emergeing it, unmerging devfs and rebooting the computer? Well, from my experience, I would say no... important device nodes will be missing (e.g. on my system /dev/ppp, which adsl requires, and /dev/nvidia*, even /dev/dsp). It would be easy for me to create them with mknod (possibly in an init script) with the same major/minor as used by devfs, but I don't think that's the preferred solution, adding them to the udev configuration would be much better. For now, I'll just use devfs since I don't use my PC in a way where udev would be really helpful (e.g. using a lot of USB mass storage devices simultaneously). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] udev & devfs - possible conflict?
Jakub Krajcovic wrote: Hi guys, i am still using devsf no 2.6.2 vanilla, but the other day i emerge udev, in order to give it a try, and this is what happened: after emerging i restarted my computer, but it refused to boot... the kernel boot was ok, the default init was also ok, but when it came to autoloading modules, the computer just froze and it stayed frozen for quite a long time until i did a hard reset. I then booted into 2.4.22-gentoo-r5 and everything was a-ok, since there is no sysfs - udev support in 2.4 kernels. Then i deleted all of the modules that i was loading at startup, booted into 2.6.2 again - and it froze again. Then in unmerged udev in 2.4 and rebooted into 2.6 again and everything was ok again. I also did an emerge udev yesterday, and found out the hard way that this not only installs the daemon on the hard drive, but also disables devfs and activates udev. I unemerged the bastard and switched back to devfs. The only problems I had though were missing devices in /dev. So the question is: is there some conflict between devfs and udev? Can they not co-exist together? Or, if this is not the case, does anyone have an idea as to what might be causing this? No. Devfs and udev are incompatible and cannot be used at the same time. And another question: is udev "100 %" able to replace devfs in terms of application compatibility? To be specific - most of the usb stuff i use relies on devfs entries - will this be ok with sysfs / udev? I'm not quite sure what you mean with that. Devfs and udev don't actually interact with the hardware, they just make the hardware devices (kernel major/minor) visible in the file system as /dev nodes. In principle, udev should be better for hotplugged devices, e.g. usb, since it can be made to "remember" a device by its serial number, and create a specific /dev node for it, whereas devfs would just count node numbers up in the order in which the devices were plugged in rsp. turned on. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] No /dev/dsp created with Alsa?
Grendel wrote: Alsa setup is also a bit of a PITA. All I did was to switch the kernel module from OSS to ALSA and emerge ALSA packages (alsa-lib, alsa-oss). My USE flags contain both alsa and oss and it works like a charm. My audio chipset is a Terratec Maestro (snd-es1968). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Has anyone actually gotten squidguard to compile?
Neal Lippman wrote: I cannot tell you what versions I have installed; I have yet to figure out how to do that under gentoo. Under debian I had apt-show-packages; is there something similar I should be using? etcat -v package. The installed version will be marked with a capital I. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] telnet setup
David Obwaller wrote: * lukas <[EMAIL PROTECTED]> [02/ 9/04]: I emerged netkit-telnetd along with xinetd and added xinetd to the default runlevel. in /etc/xinetd.d/telnetd I changed diabled = yes to = no, so telnet's enabled with xinetd. now, when I access my telnetd using telnet 127.0.0.1 I can login normally, without any problems, but when I try to log in from a remote computer on my local network it says 'connection refused by foreign host'. ssh works over lan, but as I said I need telnet to control my computer remotely. How are you connected to the internet? Are you using a router or a firewall(script) or something else? I`m using a router, which acts as a hub at the same time, to provide access internet. And I want to be able to access my computer from a remote host (outside the local network). I found the idea about making sshd listen on the telnet port quite interesting, how do I achieve this? David Do "man sshd_config" and look for ListenAddress. Adjust /etc/ssh/sshd_config accordingly and start/restart sshd. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] telnet setup
Norbert Kamenicky wrote: David Obwaller wrote: hi, my isp disables me in using ssh, so I'm forced to use telnet to remotely access my computer. I'll try to change this, but for now I want to set up telnet. It's absolutely unprobable ... until they are not diletants. Actually, some ISPs do impose a firewall on their customers. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Problem after upgrade to kde 3.2
Vanh Phom wrote: Hi All, After upgrade to kde 3.2, qt 3.3. I'm now having problem emerging kde apps like kdevelop, koffice... They all come up with the same error: hecking if UIC has KDE plugins available... no configure: error: you need to install kdelibs first. died running ./configure, kde_src_compile:configure * * Your KDE program installation died while running the configure script * !!! ERROR: app-office/koffice-1.3 failed. !!! Function configure_die, Line 170, Exitcode 0 !!! (no error message) I believe I have kdelibs install otherwise how can I run kde kicker,konqueror? Any idea. Vanh IIRC kdelibs have to be reemerged after upgrading qt, or was it the other way round? Sorry, I can't remember. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Graphics cards recommendations...
Spider wrote: No, xserver != XFree86. Same origins, different beast. (And, oops. the drivers would work if they'd been Open Source. ) So I can play America's Army with the nv driver? -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Graphics cards recommendations...
Brendan Sullivan wrote: nvidia drivers have worked with X since I started using linux about a bit over a year ago. Sure they're not perfect, but i haven't had any problems i haven't been able to find the solutions to and fix myself, and i'm not at all a linux guru. I use gentoo for daily email, chatting, web browsing, and play games such as quake3, UT2003, and RTCW. Other than the fact that my computer's specs aren't very good, i have no performance issues. I've used nvidia's drivers with 2.4 and 2.6 kernels with very little problems at all...is this the exception rather than the norm??? I'm using a GeForce 5200FX with the 5336 driver under kernel 2.6.1. I had one lockup when switching to console while America's Army was loading, otherwise it works fine. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Graphics cards recommendations...
Ciaran McCreesh wrote: On Mon, 9 Feb 2004 20:18:15 +0100 Michael Schreckenbauer <[EMAIL PROTECTED]> wrote: | Blah. Did you actually read their sources, or are you just flaming? I | am happy, that nvidia releases imho good drivers for linux. Please tell me where to find these good drivers. It'd be nice to get an ebuild made for them. *plonk* -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] telnet setup
David Obwaller wrote:
* Ric Messier <[EMAIL PROTECTED]> [02/ 9/04]:
Not sure why they would allow telnet and not ssh. However, since I
assume they are accomplishing this with port blocking, you can get
around it by getting ssh to listen on the telnet port. Or, if you are
using a cable modem/firewall, you can do a port forward. :-)
Ric
That's what I thought of, too -- they're probably just blocking the
port, and it's a very bad idea
to use an unencrypted remote shell over the Internet!
I'll try this, thanks a lot. And I really have to say that I don't have
the slightest idea why one should block ssh...
Probably this is some firewall feature ("feature"?), where only a
handful of standard ports may
be accessed from the outside. Ask your ISP whether they can permit this
kind of access
(or even disable the firewall for your account altogether, if you prefer
so).
If they can't, try to live with it or switch your ISP...
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] Fetch Restriction
Anthony Hoppe wrote: lol...hmm...maybe it makes no sense. Now for somebody to share how to disable it...plleeaaassseee? (if possible) :-D The problem is: Technically it would be possible to circumvent fetch restrictions, but it would typically be illegal (that's why the Gentoo folks enabled fetch restrictions in the first). For example, AFAIK the j2sdk may be downloaded for free from java.sun.com, however you are not allowed to redistribute it yourself on a server (at least not without Sun's permission), and thus they can't just put it on a Gentoo(-related) server where Portage can retrieve it via FTP. The reason why Sun does this is, among other things, so that they can "force" you to read their licensing conditions. If it bothers you much since it disrupts updates, just always do an emerge with the -p option first, and look for packages marked with a capital F. emerge -f will then tell you where to get the required file(s). Download them manually and put them into /usr/portage/distfiles. When you've got all those files, you should be able to run emerge without further troubles. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] how to make gentoo boot faster? (kernel 2.6.1)
Tianran Chen wrote: i notice that every time it startup, the dependency of services are scaned and checked, which take about 2~3 seconds. i wonder is this a step that can be cut off? This is done in /etc/init.d/depscan.sh. You could try to disable this, but you should rather have a bootdisk (or CD) ready, I have no idea what will happen. also, is /etc/init.d/modules neccesary? This loads the modules in /etc/modules.autoload.d. If you have no modules that need to be loaded this way, you could turn this off. You could also invoke modprobe manually from a boot script, this should be a little faster. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] how to test crond actions ??
raptor wrote: |> Is there a way to test crontab... what i mean is to run a script in the same way the crond did it, so I can check if there is some |> problems, permissions wrong etc... |> Otherwie I have to wait until crond trigger the action and still have to guess what is really happening by the external script actions... |> |> Is there such a way to simulate full crond.. and later to be sure that when the script executes after a day or so, everything will go fine... |> | |Why not just set it to run one time in five minutes? Then you can see |the result before having it run at its regular intervals. ]- not very helpfull, especialy when i can see maeningfull result after many invocations not just 2-3 ... such as rrd-driven scripts I'm not quite sure what you actually want this crond simulation to do. Maybe you want to take this to an RRD mailing list? http://people.ee.ethz.ch/~oetiker/webtools/rrdtool/mailinglists.html -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] apache2 gentoo windows partition
Claudinei Matos wrote: Hey guys, Well, I've tried to use a windows partition in DocumentRoot of my apache2.conf like the line below: DocumentRoot /mnt/dados/Jobs/Work/cjmatos.no-ip.com/www But when I try to view the page in the browser I get this error: "Forbidden You don't have permission to access / on this server. Additionally, a 403 Forbidden error was encountered while trying to use an ErrorDocument to handle the request." NTFS or FAT? Probably user apache lacks access rights to the windows partition. However, if it's FAT, you may want to consider using UMSDOS since FAT does not allow fine-tuning permissions. Otherwise, if you give apache the right to read your entire Windows partition, this will make the "chroot jail" less effective. Take a look at http://www.tldp.org/HOWTO/UMSDOS-HOWTO.html -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] New to Gentoo - general questions
Alexandre Aractingi wrote: Le lun 09/02/2004 à 16:36, Marius Mauch a écrit : No on binary. If you don't know yet, portage is a source-centered package manager. Sorry, I read about that right after I posted this mail :-) For some large packages, e.g. Openoffice.org, there are binary builds in the Portage tree though (in this example openoffice-bin). You can also compile binaries yourself (e.g. emerge -B whatever), but they will be configured for your system specifically (compiler options, USE flags etc.). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Backup Procedure?
Paul Stear wrote: Hi all, I have been working on a backup solution for my home system and have a full backup of /home and then incremental backups for the rest of the week. I know how to uncompress and and action the full backup, but have 2 questions regarding the incremental. The incremental only tar the files that have changed or are new, which is what I want, however, my script also saves empty directories, which I don't want. 1. How can I exclude empty directories? That's a tricky question since we can't look at your script, but if you're using the Bourne shell (sh or bash), you could use if ls "$DIR" | grep ^ >/dev/null; then # directory "$DIR" is not empty else # directory "$DIR" is empty fi This will not cover nested directories though, e.g. if directory x contains only directory y, and directory y is empty, x will not be regarded as empty, too. 2. What is the correct way to install the incremental backups after the full backup is installed so that my home directory is back to normal i.e. I don't want full directories over written with empty ones or end up with just a few files. You would need to store the information which files and directories were removed as well in your incremental backups. The way you would restore these would be the same way you created them. A full backup would be restored by deleting the previous directory and then installing the full tar file (e.g.) or, probably safer, to untar first and then remove files and directories not in the backup. An incremental backup stores only the differences between two system states. E.g. if your incremental backup stored the differences between the August 7 full backup and the August 8 FS state, the correct way to install it would be to first restore the system to the August 7 state and then replay the August 7 to August 8 incremental backup. I would strongly recommend against actually deleting files from home directories, moving them to a temporary directory where you can manually sift through them seems much wiser. Maybe however you'll find it simpler and safer to use BackupPC (http://backuppc.sourceforge.net) instead. I can't tell you much about it though since I'm looking forward to using it the first time, too. This is probably a dumb question, but I am learning. It's not a dumb question at all! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] libwww-5.4.0-r[12] compile fails
Harlan wrote: gcc -march=pentium3 -O3 -pipe -Wall -o .libs/libapp_1 libapp_1.o ../src/.libs/libwwwinit.so -L/usr/local/lib/mysql -L/usr/lib ../src/.libs/libwwwapp.so ../../Library/src/.libs/libwwwxml.so ../../modules/expat/xmlparse/.libs/libxmlparse.so ../../modules/expat/xmltok/.libs/libxmltok.so ../src/.libs/libwwwhtml.so ../src/.libs/libwwwtelnet.so ../src/.libs/libwwwnews.so ../src/.libs/libwwwhttp.so ../src/.libs/libwwwmime.so ../src/.libs/libwwwgopher.so ../src/.libs/libwwwftp.so ../src/.libs/libwwwdir.so ../src/.libs/libwwwcache.so ../src/.libs/libwwwstream.so ../src/.libs/libwwwfile.so ../src/.libs/libwwwmux.so ../src/.libs/libwwwtrans.so ../src/.libs/libwwwcore.so ../src/.libs/libwwwutils.so ../../Library/src/SSL/.libs/libwwwssl.so ../../Library/src/.libs/libwwwzip.so ../../Library/src/.libs/libwwwsql.so ../../modules/md5/.libs/libmd5.so -ldl /usr/local/lib/mysql/libmysqlclient.so -lz -lcrypt -lnsl -lm -lssl -lcrypto -Wl,--rpath -Wl,/usr/local/lib/mysql ../../Library/src/.libs/libwwwsql.so: undefined reference to `mysql_connect' ../../Library/src/.libs/libwwwsql.so: undefined reference to `mysql_create_db' collect2: ld returned 1 exit status make[3]: *** [libapp_1] Error 1 make[3]: Leaving directory `/var/tmp/portage/libwww-5.4.0-r2/work/w3c-libwww-5.4.0/Library/Examples' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/var/tmp/portage/libwww-5.4.0-r2/work/w3c-libwww-5.4.0/Library' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/libwww-5.4.0-r2/work/w3c-libwww-5.4.0' make: *** [all-recursive-am] Error 2 !!! ERROR: net-libs/libwww-5.4.0-r2 failed. !!! Function src_compile, Line 53, Exitcode 2 !!! (no error message) Maybe try doing an "emerge mysql" first? -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Don't update linux headers ! NEW Gentoo install broken
Ted Ozolins wrote: I use to look after quite a few win98 boxes for some of the eldery arround Edmonton Alberta. One of the gents had gotten a complete computer system as a bonus for buying a new car. He got so tired of having a window pop up requesting him to register his O/S that he decided to take matters in his own hand and deleted that dang "regestry" . He didn't think that a software company had the right to clutter his desktop with that crap every time he started his computer. He thought he had deleted whatever was asking him to register. I informed him that he had made the right choice I put Caldera 1.3 on it and he's been using linux since. Excellent! :-D But isn't Caldera now SCO? *shudder* I wouldn't want to run a kernel with the Darl McBride patches! ;-) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Don't update linux headers ! NEW Gentoo install broken
Shore wrote: On Sun, 8 Feb 2004, Arne Vogel wrote: Reminds me of my father who once killed my c:\windows\system dir because he thought it was superfluous... It isn't? :-) Yes, now that my PC doesn't have one Microsoft bit on its HD. *snicker* (ok, maybe somewhere in the browser cache from www.microsoft.com - reading those funny support messages like the one that tells you to copy links manually into the location bar instead of trusting IE! *g) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Init script problems ...
Aaron Walker wrote: On Sat, 2004-02-07 at 18:19, lukas wrote: You normaly can see these messages while your system is coming up. When you are logged in you can also type "dmesg | less" to see them. I get ... cant find module /dev/rtc cant find module /dev/misc/rtc Are you shure that devfsd is proper running? sys-fs/devfsd must be installed. If it isn't, emerge it. Your kernel must be compiled with "/dev file system support" and if you don't have it also compiled with the "Automatically mount at boot" option, you must add "devfs=mount" as a kernel bootparameter to your grub.conf (respectively lilo.conf). cu lukas Is there any reason my system works with devfs (and devfsd running) but not compiled with "Automatically mount at boot" nor the devfs=mount boot param? I used to always compile the kernel w/ auto mount at boot, but somewhere along the line I stopped, but all the gentoo boxen here still work.. Let's see... 1) Some Gentoo init script mounts it (grepped mine, but didn't find it). 2) You don't use any devices not already in Gentoo's stage1 /dev directory. 3) devfsd mounts it when started. 4-infinity) ? -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Another gcc/portage problem.
William Kenworthy wrote: env-update uses python so dies without the symlink. The paths are hard coded in the binary according to ldd, so the environment isnt the problem. I think I need a minimum of correctly linked binaries to kick off a system rebuild. BillK From your ldd output, I don't see why your /usr/bin/python should use hard coded paths; when I run ldd /usr/bin/python, I get: linux-gate.so.1 => (0xe000) libpython2.3.so.1.0 => /usr/lib/libpython2.3.so.1.0 (0x4002f000) libpthread.so.0 => /lib/libpthread.so.0 (0x40145000) libdl.so.2 => /lib/libdl.so.2 (0x40196000) libutil.so.1 => /lib/libutil.so.1 (0x40199000) libstdc++.so.5 => /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/libstdc++.so.5 (0x4019c000) libm.so.6 => /lib/libm.so.6 (0x40255000) libgcc_s.so.1 => /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/libgcc_s.so.1 (0x40277000) libc.so.6 => /lib/libc.so.6 (0x4028) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x4000) That doesn't look so different from your ldd's output, except for linux-gate.so.1, which is required by libgcc_s.so.1, the relocation addresses (I guess? ldd's docs don't mention this) and that your libgcc_s.so.1 resolves to the old gcc's path (and this is the only important difference). Maybe you can check whether there actually is a /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.2/libgcc_s.so.1 on your system. If so, the old gcc dir seems to have preference over the new one. You can correct this by editing the LDPATH variable in /etc/env.d/05gcc and running env-update. env-update should work if you run it with 'LD_LIBRARY_PATH="/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.2" env-update', since this variable takes precedence over /etc/ld.so.conf. Then do an ldconfig and things should work. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] autoconf version?
Daniel Drake wrote:
- How can I change the default to autoconf 2.58? Will my system
survive this? ;-} /usr/bin/autoconf is a symlink pointing to
../lib/autoconf/ac-wrapper.pl
# autoconf --version
Autoconf version 2.13
# export WANT_AUTOCONF="2.5"
# autoconf --version
autoconf (GNU Autoconf) 2.58
Written by David J. MacKenzie and Akim Demaille.
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There
is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE.
To make this default, add the line : export WANT_AUTOCONF="2.5"
to your ~/.bashrc or /etc/profile
Automake runs on a similar system. Check out the script (nano
/usr/bin/auto{conf,make}) to find out more about the variables which
are listened to.
Daniel
Ah, ok, thanks! That'll do.
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] kdepim-3.2.0-r2 failure
Tom Syroid wrote:
Morning all,
OK, I'm stuck. The problem is 24 hours old and my first instinct was
to assume it was a bug in one of the ebuilds I've merged in the last
36 hours. But after monitoring the list, it would appear it's a unique
localized problem.
I'm running two Gentoo systems here very similar to several others on
the list:
Full ~x86, one a single P3 the other a dual AthlonMP.
NPTL threading enabled on both systems.
phaedrus root # qpkg -I -v gcc
sys-devel/gcc-3.3.2-r6 *
sys-devel/gcc-config-1.3.4 *
All latest patches/fixes/updates applied.
The build for kdepim-3.2.0-r2 craps out with:
snip<<<
checking for xgettext... /usr/bin/xgettext
checking if C++ programs can be compiled... no
configure: error: Your Installation isn't able to compile simple C++
programs.
Check config.log for details - if you're using a Linux distribution
you might miss
a package named similiar to libstd++-dev.
died running ./configure, kde_src_compile:configure
*
* Your KDE program installation died while running the configure script
*
!!! ERROR: kde-base/kdepim-3.2.0-r2 failed.
!!! Function configure_die, Line 170, Exitcode 0
!!! (no error message)
ends<<<
This is NOT kdepim specific, I've come to find. I can't build anything
involving C++ code.
The weird part is, I built all the latest KDE 3.2.0 packages without
error or incident.
I upgraded to the kdepim-3.2.0-r1 release late Wednesday.
I've downgraded to gcc-r5 -- same results.
Where to now, folks?
Suggestions, pointers, and general direction welcomed with open arms.
Try ldconfig. Autoconf (via configure) can be a bit ambigous here IIRC -
it says "can't compile C++ programs",
but what it *really* means is "can't run the C++ program I just
compiled", and that may mean
that libstdc++.so.5 is not found after a GCC merge, the problem that
seems to already have
plagued every 2nd user or so on this list...
If this doesn't work, write a file test.cxx with e.g. the following content:
int main() {}
and do "g++ -o test test.cxx"
and, if this works, "./test"
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] Don't update linux headers ! NEW Gentoo install broken
Michele Alzetta wrote: Il Sun, 08 Feb 2004 10:21:00 +0100 Arne Vogel <[EMAIL PROTECTED]> scrisse: Reminds me of my father who once killed my c:\windows\system dir because he thought it was superfluous... How right he was ! Absolutely! It was Windows 98, which truly deserved it. You may have thought of something else, but well... -- [EMAIL PROTECTED] mailing list
[gentoo-user] autoconf version?
Hi, I've switched to Gentoo recently (bootstrap from stage1) and experienced problems with autoconf, which would reject a configure.in script of mine that worked before. It almost seemed to me as if things had changed back to the way they were in much earlier autoconf versions... now I checked if it may actually be the case that autoconf is outdated, so I did: $ etcat -v autoconf [ Results for search key : autoconf ] [ Candidate applications found : 4 ] Only printing found installed programs. * sys-devel/autoconf-2.57-r1 : [ ] 2.57-r1 (2.5) * sys-devel/autoconf-2.57a-r1 : [M~ ] 2.57a-r1 (2.5) * sys-devel/autoconf-2.58 : [ I] 2.58 (2.5) * sys-devel/autoconf-2.59 : [M~ ] 2.59 (2.5) Ok, so etcat says I have autoconf 2.58 installed, which so far is the latest unmasked version. However, when I do "autoconf --version", it says: Autoconf version 2.13 which is ancient! Also, the info pages seem to belong to an autoconf much older than 2.58. "qpkg -f `which autoconf`" says sys-devel/autoconf. I now found out that autoconf-2.58 actually installs *two* versions, namely 2.13 and 2.58, to be invoked by "autoconf-(versionno.)" (and the more recent info pages are accessed with "info autoconf25"), so I would like to know: - Why are the two versions not in two separate packages, and accordingly displayed by etcat? - Why does "autoconf" by default invoke the older version? - How can I change the default to autoconf 2.58? Will my system survive this? ;-} /usr/bin/autoconf is a symlink pointing to ../lib/autoconf/ac-wrapper.pl -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Another gcc/portage problem.
William Kenworthy wrote: Somehow I have mixed up libraries after updating world which went from gcc-3.2.3 to 3.3.2. gcc-3.1.1 has not been in use for ages, but there are some remnants on the system. Without a working python, most of gentoos tools are broke. Suggestions on how to fix? Try 'env-update && ldconfig'. But remove your symlinks before doing that. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Sensors and 2.6 kernels
Grendel wrote: If you have loaded them then you should get the inrromation, just type /usr/bin/sensors and tell us the output. Please tell your motherboard model as well. Also make sure that you have the latest 2.8.3 IIRC version of the lm_sensors userspace tools. It's lm-sensors (with hyphen, not underscore). At least in Portage! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ssh login problem
Stroller wrote: On Feb 6, 2004, at 5:38 pm, fisch wrote: hi, i have a problem when loggin in via ssh. workstation> ssh server -lroot -> logged in workstation> ssh server -luser -> access denied but server> ssh localhost -luser -> logged in (with warning, that no homedir exists - that's right) where is the differenz? Could you `grep ListenAddress /etc/ssh/sshd_config` please..? Stroller. Also check your keys. And use "ssh -vvv" for full debugging output. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Kernel 2.6 and desktop responsiveness under cpu load
Diego Zamboni wrote: I have been using the 2.6 kernel for a while in my system and noticed that the desktop is not responsive under cpu load. kernel gentoo-sources-2.4.22-r5 is better for me in this respect. I have the exact opposite experience - my desktop responsiveness has improved dramatically since switching to 2.6. One relevant setting is CONFIG_PREEMPT (under Processor type and features in menuconfig), which you want to be set for a desktop system. --Diego The same goes for my system, I turned the preemption option on, and the mouse cursor doesn't start to jump nearly as easily as with kernel 2.4.24 under heavy I/O load (e.g. when emerge executes a "tar xjf"). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Don't update linux headers ! NEW Gentoo install broken
Michele Alzetta wrote: Il Sat, 7 Feb 2004 15:10:43 +0100 [EMAIL PROTECTED] scrisse: try to execute ldconfig, it can help. Didn't think of that; I reformatted the partitions and started again from scratch. Almost through emerge system now. Hope I won't need to try out your suggestion ! Reminds me of my father who once killed my c:\windows\system dir because he thought it was superfluous... -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Installing Gentoo from a currently running distro
Shore wrote: On Fri, 6 Feb 2004, Arne Vogel wrote: - I would recommend using the vanilla kernel since they tend to be the most stable AFAICT. I tend to roll my own. Actually I always roll my own to trim the fat and enable things I need that aren't enabled. I'll have to test this carefully though. With vanilla I meant none of the patched versions that are popular with some Gentoo users. (gentoo-sources, love-sources...) For me, the original kernel is still the most trustworthy however (I never had a boot-up kernel panic with the vanilla version IIRC). - Be very careful where installing kernel modules, and not to overwrite your old kernel image. Install the new kernel only from within the chrooted environment. I plan on installing a 3rd SATS drive in this server for the Gentoo installation. My RH9 installation will remain intact until I'm satisfied that I'll never have to see it ever again. :) That's a good idea in case something goes wrong. - If your current kernel has the kexec patch, use it to boot into the Gentoo installation without changing the bootloader's default. That way, if anything goes wrong, the next reboot will reboot the previous distro. Now this sounds interesting. I don't know that I've ever heard of the kexec patch. It would be nice to be able to control where I boot after a failed boot. That would be most useful. I don't have a HW console server option yet (PCI card to control keyboard and video I/O for control from the BIOS to the actual login). I'll have to look into this further. Think it through very thoroughly... you have only one try to ensure that Gentoo either comes up remote-maintainable, or your previous installation is automatically rebooted into, everything else will require phys. access... Yeah, I figure I can always get someone over there to run down and kick the box for me if I have to. I'd rather not though since it's headless. It's doable though. I'd like to find a way to maintain a grub install that could boot from either RH9 on hda1 or Gentoo on hda3. I think it can do it but I'll have to play with it a bit to be sure. Thanks for all the info! Justin NP! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] 'cannot open shared object file' after emerge gcc
Marshal Newrock wrote: Today I did a sync and upgrade, which I haven't done for several days. There's quite a few packages to upgrade, but I started with portage. Then I upgraded gentoolkit (as indicated by the portage post-install messages). Then I proceeded to upgrade gcc. After that, I got the dreaded: /usr/bin/python: error while loading shared libraries: libstdc++.so.5: cannot open shared object file: No such file or directory Being a reader of this list, I proceeded to run 'ldconfig'. Now everything is working fine again. Hopefully this info will help someone else. I am running an 'x86' system. Yes, I and many others had the same problem, it happens after unmerging gcc (when installing a more recent version). The linker still looks for libstdc++.so.5 in the old directory. Running ldconfig fixes this, and points the linker to the new path. IMO it is a bug, ldconfig should be run automatically after changing /etc/ld.so.conf. I checked the log of the last gcc merge, and it didn't call ldconfig *after* the new compiler (and the libs) have been installed. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Need some advice on to switch from mandrake to gentoo.
Grendel wrote: Currently, emerge does not support simultaneous downloading and compiling (at least I know nothing about such an option), in the future it could download the next package while the current package is still compiling. For now, you can more or less emulate this by starting an emerge -f (options) (packages) first, waiting for one or more downloads to complete, and then starting another emerge (options) (packages) with the same arguments except "-f" (without terminating the first emerge). Though this may cause trouble should the second emerge ever catch up with the first... I'm not sure whether running 2 wget processes on the same file is a good idea. If the file to be downloaded exist, wget simply downloads the file as filename.1 But it seems emerge is using the -c option to continue downloads that have been interrupted. I have no idea what happens when two simultaneous wget -c's are run with the same output file... Thx, will be doing it tonight when no one is using the desktop :) I really appreciate all the help this list is giving me. This list is really good for sure! I'm learning a lot from the posts here. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Need some advice on to switch from mandrake to gentoo.
Grendel wrote: On Sat, 7 Feb 2004, Arne Vogel wrote: Sure you can! However, you may want to consider installing from your existing Mandrake installation. You should actually be able to install Gentoo on the same partition (into a subdirectory, using chroot - see the handbook), and later replace the Mandrak installation (except for /home) with it, though the transition may be a little tricky (mv, rm, cp etc. may break). I would recommend booting from the LiveCD before performing this last step. How much disk space do you have left on /? /dev/hda6 5.9G 3.7G 2.2G 63% / Should be enough. If you remove Mandrake before the install, it definitely should. When you run low on disk space with a Gentoo installation, check your /usr/portage/distfiles. That's where the downloads go, and Gentoo doesn't delete anything in it automatically. Here's a simple script that you might also find useful (I've installed this as /usr/bin/diskusage): du --max-depth=1 "$@" | sort -n -r | less When called without arguments, it displays the subdirectories of ".", sorted by size, descending. When called with argument "*", it displays files as well. And etcat -s tells you the installed size of a given package. Alternatively, install everything from binary, this will save lots and lots of time, you will be able to compile packages later. Yes, this is what I plan as i want to get the desktop up and running as fast as possible. Once I have a working desktop I will set about updating to the modern packages. kdegames: merge time: 44 minutes and 52 seconds etc. Does this time include downloading as well? I'm not sure... for broadband, download times should be much lower than compile times (my 768-kbps DSL line manages about 360 Megabytes per hour, others are still faster). The only notable exception are binaries, e.g. openoffice-bin or americas-army. Currently, emerge does not support simultaneous downloading and compiling (at least I know nothing about such an option), in the future it could download the next package while the current package is still compiling. For now, you can more or less emulate this by starting an emerge -f (options) (packages) first, waiting for one or more downloads to complete, and then starting another emerge (options) (packages) with the same arguments except "-f" (without terminating the first emerge). Though this may cause trouble should the second emerge ever catch up with the first... I'm not sure whether running 2 wget processes on the same file is a good idea. xfree: merge time: 1 hour, 9 minutes and 14 seconds. hmm.this was fastxfree86 needs a large volume of space and takes a lot of time to compile too usually. Maybe this is because the gcc was Athlon-XP-optimized with -O3? :-) Various other packages that will be needed for your installation add up to let's say 3 hours, that would be 12 1/2 hours in total. Deduct 20% for your faster processor, and it's still about 10 hours. Well I plan to use the live cd's and get a precompiled xfree86+kde setup ready then from the network upgrade the gcc, libc, kernel, xfree and KDE step by step as time permits, I suppose this is the best option. Make it so! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Problems after install
Grendel wrote: On Sat, 7 Feb 2004, Jani-Matti [iso-8859-1] Hätinen wrote: Arnold Lavier kirjoitti (lähetysaika lauantai, 7. helmikuuta 2004 11:06): edit /etc/fstab and add 'user' to your list of options like in this example : /dev/cdroms/cdrom0 /mnt/cdrom autoro,noauto,user 0 0 If you are sharing your computer with others, it might be a good idea to use 'users' instead. It allows other users to unmount a mounted device. Thus, if someone forgets a CD into the dirve, others can unmount and eject it. What I cant understand is that why isnt the exceelent supermount being used for removable media in gentoo, it makes using cd's easy, you simply inert the cd and the os auto mounts it, and you can eject the cd when you want, there is no need to unmount it manually before ejecting it. Its really a good thing on desktop system to have. But it does prevent ejecting if the CD is in use? To be honest, I find the way Windows does things (with "automounting") pretty f up, and don't see any need to have it under Linux. But oh, well, Gentoo is about choices... :-) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] possiable abcde ebuild bug?
Jayson Garrell wrote: While updating my system I got the folowing error on the 'abcde' (mp3 * ogg encoding) package. /bin/install -c -d -m 755 /var/tmp/portage/abcde-2.1.9/image//usr/bin make: /bin/install: Command not found make: *** [install] Error 127 So I did a 'whereis install' and it came up as /usr/bin/install not /bin/install. All I did was make a symbolic link from /usr/bin/install to /bin/install and it worked fine. This link should have existed. Should I post a bug at http://bugs.gentoo.orgn or ??? Please do that. "/bin/install link missing" or something like that. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Need some advice on to switch from mandrake to gentoo.
Grendel wrote: Hi all, I need some advice about switchin from mandrake 9.1 to gentoo. My setup is this, Athlon 3000xp with Asus a7n8x motherboard My 80gb disk is partioned in to / and /home, I would like to keep /home. I have downloaded the athlon-xp optiomised 2live cd's. My nvnet nforce2 ethernet is automatically detected by the live cd1, so I have everything working. Now since this is for my workstation, I would like to get it up and running fast, so I have some questions. 1. Can I install using the 2 live'cds and then upgrade later to the recent sources? Sure you can! However, you may want to consider installing from your existing Mandrake installation. You should actually be able to install Gentoo on the same partition (into a subdirectory, using chroot - see the handbook), and later replace the Mandrak installation (except for /home) with it, though the transition may be a little tricky (mv, rm, cp etc. may break). I would recommend booting from the LiveCD before performing this last step. How much disk space do you have left on /? Alternatively, install everything from binary, this will save lots and lots of time, you will be able to compile packages later. 2. How long will it take to get a running KDE setup with the two live cd's. The 6 hours I saw mentioned here seem like an underestimate for an installation from sources. On my Athlon XP 2400 , I have (displayed with genlop -t): kdelibs: merge time: 1 hour, 39 minutes and 10 seconds. kdebase: merge time: 1 hour, 50 minutes and 52 seconds. kdegraphics: merge time: 40 minutes and 19 seconds kdemultimedia: merge time: 1 hour, 4 minutes and 7 seconds. kdenetwork: merge time: 43 minutes and 24 seconds. kdegames: merge time: 44 minutes and 52 seconds etc. xfree: merge time: 1 hour, 9 minutes and 14 seconds. gcc: merge time: 26 minutes and 4 seconds. glibc: merge time: 55 minutes and 45 seconds Various other packages that will be needed for your installation add up to let's say 3 hours, that would be 12 1/2 hours in total. Deduct 20% for your faster processor, and it's still about 10 hours. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] mozilla 1.6 not compiling
Alberto Bert wrote: Hi all, I'm not an expert of gentoo at all, but I've never had compilation problems up to now (~ 1 year working), so I was very surpised today... I tryed to emerge mozilla 1.6, it requested several other updating and new emerging, among them xfree, gtk+, etc. The compilation stops with the following error message. I just compiled it successfully (using current non-masked packages). "etcat -u mozilla-1.6" says: U I [ Found these USE variables in : net-www/mozilla-1.6 ] + + java : Adds support for Java + + crypt: Add support for encryption -- using mcrypt or gpg where applicable - - ipv6 : Adds support for IP version 6 + + gtk2 : Use gtk+-2.0.0 over gtk+-1.2 in cases where a program supports both. + + ssl : Adds support for Secure Socket Layer connections - - ldap : Adds LDAP support (Lightweight Directory Access Protocol) + + gnome: Adds GNOME support - - debug: Tells configure and the makefiles to build for debugging. Effects vary acrosss packages, but generally it will at least add -g to CFLAGS. Remember to set FEATURES+=nostrip too. - - mozcalendar : blank - - mozaccess: blank - - mozxmlterm : blank - - moznoirc : If you do NOT want the IRC client built with mozilla - - moznomail: If you do NOT want the mail client built with mozilla - - moznocompose : If you do NOT want the web page composer built with mozilla - - moznoxft : blank "emerge -up mozilla" says mozilla and direct dependencies are up to date. Maybe this helps... :-| Plus, now something seems to be happened to emerge, since when I try to emerge centain packages it gives me errors like: snoopy root # emerge -pv xterm These are the packages that I would merge, in order: Calculating dependencies !!! all ebuilds that could satisfy "xterm" have been masked. !!! Error calculating dependencies. Please correct. This is no different from what I get. It seems all versions of xterm are currently masked. the error appear not for all packages... What am Isupposed to "correct"? Any help would be VERY appreciated. thanks, Alberto -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] kernel 2.6 and nvidia 5336
Hemmann, Volker Armin wrote: Hi, On Friday 06 February 2004 15:58, Arne Vogel wrote: The first problem was that KDE would refuse to start, complaining about a missing symbol _nv22 in /usr/lib/libGLcore.so.1 (which pointed to the new version of that NVIDIA library). GNOME would work fine. I could resolve this by removing the /usr/X11R6/lib/libGL.so.1 link (which pointed to a MESA library). Unfortunately, that link later popped up again, and I still don't quite know what's creating it. Have you did an opengl-update nvidia? No... I'll try that, thanks! (Danke sehr :-) The second problem I had yesterday - when starting America's Army (from KDE), I switched to console. After switching back to the virtual console on which X runs, the screen would be filled with garbage text, and I couldn't perform any inputs! I heard the music from AA playing, so the system wasn't totally hanging, but I could do nothing to get out of this "trap", and had to reset the machine. Usually, switching between console and X does work, when AA is running, though, Have you tried sys-rq-keys? I have a SiS746FX based board and with my gf fx 5200 I have to start X two times. Start X, kill X, start X again. Sometimes, this f* up the console, but after that X is working fine. Glück Auf Volker The SysReq option was disabled. To recover from a bad X failure I would need to press Alt+SysReq+K? -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ftpd doesn't show any contents
Gerhard W. Gruber wrote: On Fri, 06 Feb 2004 21:12:00 +0100, Arne Vogel <[EMAIL PROTECTED]> wrote: Also, it seems the linker requires these libraries to be referenced via symlinks just as in your ordinary /lib directory. You should be able to just do a "cp -dp /lib/libnss* lib" (from inside the chrroot-directory) to make all the NSS libraries available to the chrooted environment. Do *not* use ln, as that kind of defeats the purpose of the chroot environment (an attacker could open the hard-linked library for write access, and thus compromise your global /lib directory). Maybe one day Linux'll support copy-on-write for hard-linked files... :-) Now it works for anonymous. But I still have the problem of being blind when logging in as normal user. It seems that I have to createt his entire environment for all users I want to be accessible via ftp. I don't really like that, but having anonymous access is sufficient for now. Or is there some way to create a universal chroot environment? Hmm... I don't know. "info ftpd" may help. BTW: Doing the libs with ln doesn't work anyway, because when you do a chroot, then the root directory is set to the one you specified. So if you have this in your normal environment /lib/libc.so /home/ftp/lib/libc.so -> /lib/libc.so it will in truth point to /home/ftp/lib/libc.so as soon as you do the chroot. Symlinks will not work, of course, but hard links would. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ftpd doesn't show any contents
Gerhard W. Gruber wrote: On Fri, 6 Feb 2004 10:44:30 -0800 (PST), "Eric Paynter" <[EMAIL PROTECTED]> wrote: Is that 555 for files and 511 for directories? Try 555 for all, and if that fixes it we can work on getting more restrictive. I now tried 777 and I have still this problem. but I dont thin it is the permssion. I just tried manually to chroot /home/ftp and I get the errormessage chroot: /bin/bash : no such file or directory. Looks like one or more shared libraries are missing. Use "ldd /bin/bash" (rsp. /bin/ls) to see what bash/ls requires. Additionally, there are some libraries that are *not* listed by ldd - at least /lib/libnss* stuff. The glibc requires them, but bypasses "normal" dynamic linking so ldd doesn't know about them. Using something like "strace chroot . /bin/bash" should help figuring out which libraries are still missing. Also, it seems the linker requires these libraries to be referenced via symlinks just as in your ordinary /lib directory. You should be able to just do a "cp -dp /lib/libnss* lib" (from inside the chrroot-directory) to make all the NSS libraries available to the chrooted environment. Do *not* use ln, as that kind of defeats the purpose of the chroot environment (an attacker could open the hard-linked library for write access, and thus compromise your global /lib directory). Maybe one day Linux'll support copy-on-write for hard-linked files... :-) (snip) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Installing Gentoo from a currently running distro
Shore wrote: Howdy. I'd like to give Gentoo a try on my newest server. The server isn't doing anything critical at the moment so it's a good choice IMHO. Frankly I'm sick of RH9 and its library nightmare. The box is co-loed about 2 hours away so physical access isn't as easy as I'd like. This eliminates booting from a CD as well. I'm shipping a new SATA drive over to this place next week to install in one of the hot swap bays. I'd like to install Gentoo on it. Is it possible to install Gentoo onto that new drive remotely without booting from a CD and/or bringing down the machine as it is now? I installed from an existing SuSE installation using chroot onto the new partition, just following the manual. However, I did have phys. access. The things you should be aware of especially since you don't are (at least?) the following: - I would recommend using the vanilla kernel since they tend to be the most stable AFAICT. - Be very careful where installing kernel modules, and not to overwrite your old kernel image. Install the new kernel only from within the chrooted environment. - Be very careful with lilo/grub. Leave the existing options for booting your old system untouched, and, my recommendation (see below): Do not (yet) change the default boot image and partition! It is probably best to configure the bootloader from the *old* installation. - Install sshd and networking properly. - When rebooting to Gentoo, use the boot parameter "panic=N", with an integer N > 0 to tell the kernel to reboot after panicking. - Also, make sure Gentoo will reboot when anything goes wrong with upping the network or sshd. It's probably best to use an init-script to reboot after a fixed time, let's say 10 minutes. You then have 10 minutes to see whether ssh works, and whether you can maintain the Gentoo installation remotely. If you see that it works, you can then remove this safeguard (possibly with immediate effect in the Gentoo installation). For example, make an init-script that does "shutdown -r +10 &", which will start a reboot timer of 10 minutes. If you see that sshd works, you can remotely execute "killall shutdown" to "defuse the time bomb", and disable the script. Make sure the init script is installed properly and will actually execute! Copying it into /etc/init.d is not enough! Read the docs. - If your current kernel has the kexec patch, use it to boot into the Gentoo installation without changing the bootloader's default. That way, if anything goes wrong, the next reboot will reboot the previous distro. - If it doesn't, it may be possible to do something of this kind with lilo rsp. grub alone, but you'll have to ask someone else. Otherwise, you may be able to patch e.g. lilo so that it boots the old system in case the time is greater than let's say now + 1 hour. Don't blame me if you don't get this to work, though! :-) - If you can't boot to Gentoo without changing the default, boot without the panic option, or with a very long one to avoid putting the machine into a never-ending booting loop (e.g. panic=3600) if it does actually fail. And configure your Gentoo init scripts to restore the previous bootloader configuration after booting. This will require the bootloader config to work properly from the Gentoo installation (if it worked chrooted, it should work standalone too, however). Think it through very thoroughly... you have only one try to ensure that Gentoo either comes up remote-maintainable, or your previous installation is automatically rebooted into, everything else will require phys. access... HTH! :-) (Legal notice: These tips come without implicit or express warranty of any kind. ;-) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ftpd doesn't show any contents
Gerhard W. Gruber wrote: On Fri, 6 Feb 2004 09:56:04 -0800 (PST), "Eric Paynter" <[EMAIL PROTECTED]> wrote: Check the permissions of the directory. Also, as I discovered recently, some ftp servers won't list files > 2GB, but that's probably not the problem you have here. I'd focus on permissions first. I set the directories according to the help to 555 and 511. Set all the directories to 555. 511 will give only execute permissions to other users (e.g. the user as which ftpd runs), which means they can change into those directories, but not list the files in them. This seems to be exactly your problem. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] kernel 2.6 and nvidia 5336
Collins Richey wrote: FYI, I've read quite a few negative comments about the latest nvidia drivers, but I must say I have had the opposite experience. On my labrat box, I'm running a newly installed latest gentoop 2004 2.6.2 system with ~x86 everything and glibc with nptl activated, so the 5336 level nvidia drivers were selected. I noted: 1. On my nforce2 video card, I notice a 300fps speed improvement (glxgears) over prior 4496 level drivers. 2. I have a VIA chipset. The new nvidia driver is smart enough to disable AGP support even if requested instead of hanging the system as did the 4496 level driver. 3. I have no problem switching to-from an alternate console while X is active. No black screens as others have reported. Enjoy, Well, overall they run fine on my machine, too, and glxgears is really fast. There were two problems though... I'm using 2.6.1 since a couple of days as default kernel after I managed to solve the remaining upgrade-related issues (2.4.x to 2.6.x was definitely the most work-intensive transition I have experienced), and I installed the nvidia 5336 drivers. The first problem was that KDE would refuse to start, complaining about a missing symbol _nv22 in /usr/lib/libGLcore.so.1 (which pointed to the new version of that NVIDIA library). GNOME would work fine. I could resolve this by removing the /usr/X11R6/lib/libGL.so.1 link (which pointed to a MESA library). Unfortunately, that link later popped up again, and I still don't quite know what's creating it. The second problem I had yesterday - when starting America's Army (from KDE), I switched to console. After switching back to the virtual console on which X runs, the screen would be filled with garbage text, and I couldn't perform any inputs! I heard the music from AA playing, so the system wasn't totally hanging, but I could do nothing to get out of this "trap", and had to reset the machine. Usually, switching between console and X does work, when AA is running, though, so it was likely some spurious problem. It's just that I didn't have a hanging system (or sudden reboot or other major failure for all that matters) in the 4 months I have this PC, using kernels 2.4.21 and 2.4.24. So I would recommend saving all your work (and closing the mail application, especially if it is configured to fetch e-mails every x minutes) before starting a 3D game. The nvidia driver 5336 or kernel 2.6 *may* freeze the PC. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] gpg key expiration
Chris I wrote: When I started using gnupg, I had set my key to expire a year later. This day recently passed, and I'm curious if there is anything special I need to do about the old one (like revoke it, etc) and how I would go about doing that. After reading man gpg and the gentoo gnupg docs, i ended up making a certificate, as well as doing a revoke (although it didnt seem to invalidate anything afaik). It should warn anyone who's using the public key after it has expired that the key should no longer be used to encrypt messages to you, rsp. to check your signatures. There's no need for a revoke (which would trigger the same warning for a key that is not yet expired, but the revoke has to be distributed alongside a newer key rsp. over keyservers before it becomes effective). You can test this by encrypting a message to yourself (using the expired key). A similiar warning should be shown when you try to use your expired private key. You can test this by signing a message with it. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Spider - you crack me up!
Tom Wesley wrote: On Thu, 2004-02-05 at 21:05, Mike Wojcikiewicz wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Thursday 05 February 2004 14:26, Spider wrote: From M$ article Q265230 WORKAROUND To workaround this problem: Do not start messages with the word "begin" followed by two spaces. Use only one space between the word "begin" and the following data. Capitalize the word "begin" so that it is reads "Begin." Use a different word such as "start" or "commence." *G* Glad you found it. It just feels like a very bizarre "solution" by Microsoft. . sort of like their workaround for that URL bug in IE they had... went something along the lines of wanting their users to manually type URLs into the address bar instead of clicking them if they were concerned about security (rather than fixing the bug)... ahh micro$oft.. always good for a laugh:) I think you're being very unfair - they suggested you copy the shortcut to the clipboard and paste it into notepad before clicking it. ;) Maybe it should be renamed to URLpad. I still don't quite get it why "okay, our browser has very bad exploitable 'issues', but that's not a problem as long as you trust the owners of the websites you are surfing to" may legally be called a security policy. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ps2 mouse
Alex Nelson wrote: snip Another question is are you using a KVM? I had a similar problem with my optical scroll mouse going "nuts" under the 2.6.x kernels and it turned out to be my kvm. Not sure if there is a fix out for this yet or not. Good luck either way! -Alex Since I had to look up what a KVM is, I guess I don't use one. :-) (Google says it's a keyboard-video-mouse-switch) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] bug in kdelibs-3.2 configure ?
Andrej Kacian wrote: Following is an excerpt from kdelibs-3.2 configure script, lines from 32239 to 32255. Maybe I'm not reading this right, but is the second if statement correct? It seems to me that it builds with arts when configure option says"don't" (--without-arts). snip if test "$build_arts" != "no"; then include_ARTS_TRUE= include_ARTS_FALSE='#' else include_ARTS_TRUE='#' include_ARTS_FALSE= fi It seems these macros are passed on to the Makefile(s) where they either comment something out or not. Thus, when build_arts is anything but "no" (e.g. "absolutely not!" ;-) ), parts that have $(include_ARTS_TRUE) in front of them will be parsed, while parts that have $(include_ARTS_FALSE) in front of them will be commented out. Counterintuitive as it is, this seems to be correct. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ps2 mouse
Koala Gnu wrote: Hi all, I compiled 2.6.1 kernel on my gentoo distr and the behavior of my PS/2 mouse is changed. It seems to be too sensitive and then unusable. This will happen both in text mode (using gpm) and X. The mouse works fine in 2.4. I read from the linux mail list archive that this is not a new problem and several suggestions have been submitted. I tried using: psmouse_noext=1 (as suggested by Torvalds) as boot parameter but it does not work. Then I tried also with psmouse_rate=60 psmouse_resolution=200 but it does not work too. Any suggestion? PS let me know if you need more info. No idea why the boot params are not working, but I was too lazy to get informed and just adjusted the mouse sensitivity in KDE (and games...). That's not a pretty solution, but it will most definitely work for you. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Alsa ; System beep
Ralph Slooten wrote: Hello again, On Thu, 05 Feb 2004 12:50:48 + Daniel Drake <[EMAIL PROTECTED]> wrote: http://www.codemonkey.org.uk/post-halloween-2.5.txt From that page: - Users wanting support for the PC speaker need to enable CONFIG_INPUT_PCSPKR, or you won't get a single beep. Ahha, thank you ! Took me about 30 minutes to locate that option, hidden away under "misc", but I got it eventually. Initially I searched the .config file, but it had no mention of it, until I found it via menuconfig .. then an entry got added to the config. Err... you could've just added CONFIG_INPUT_PCSPKR=y to the .config file, I guess? :-} -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] g77 and g++ compilers
Valmor de Almeida wrote: Is there a reason for g77 be installed in /usr/i686-pc-linux-gnu/gcc-bin/3.2/g77 and not in say /usr/bin/g77 ? Also why g++ in /usr/bin/g++ is different from /usr/i686-pc-linux-gnu/gcc-bin/3.2/g++ ? I wonder whether this is particular for my installation. Thanks for any info. Looks the same on my system, though I have no idea why this is so. -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Problems emerging
Dave wrote: Dave wrote: Dennis Freise wrote: On Mon, 02 Feb 2004 17:49:29 + Dave <[EMAIL PROTECTED]> wrote: As belt and braces I upped my CPU fan speed and started compiling. 1 Hour later - compiler crash "recursive error". I looked at http://www.freehackers.org/gentoo/gccflags/flag_gcc3.html And selected ... Pentium III (Intel) CHOST="i686-pc-linux-gnu" CFLAGS="-march=pentium3 -O3 -pipe -fomit-frame-pointer" CXXFLAGS="-march=pentium3 -O3 -pipe -fomit-frame-pointer" Started compiling, three hours later - compiler crash ! I changed -o3 to -o2. Its been running for 24Hrs without a glitch. Although this seems to have solved the problem I'm unsure why the compiler could not handle -o3. I have PIII 700Mhz (coppermine) 256MB, 40GB Partition with DMA enabled. Any comments ? I've got gentoo on an P3 933Mhz/512MB/[EMAIL PROTECTED] completly compiled with the flags you got from freehackers with -O3 - the compiler never crashed for me. I'm @~x86. Maybe you should run some memtest86's... OK tracked dowm memtest86 - am going to run it overnight Dave Have now run memtest86 for 12 Hours straight - extended test. Zero errors ... mmm ... I'll limp along with -O2 for now ... sort the rest of Gentoo then come back to it. Dave If the memory is intact, it's most probably a gcc bug. If it persists in a more recent version, you should file a bug report to http://gcc.gnu.org/bugzilla/ (if it's not already listed). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] ~x86 emerge: broke emerge
Kurt Guenther wrote: tumbleweed root # emerge /usr/bin/python: error while loading shared libraries: libstdc++.so.5: cannot open shared object file: No such file or directory tumbleweed root # Try to 'locate libstdc++.so.5' and see if it's still there (the locate DB could be outdated, so verify with ls). If so, do export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:" and you should be able to use emerge (from the terminal(s) on which you executed the export command). Then I would suggest reverting to a stable emerge ASAP... If libstdc++.so.5 is indeed missing, gcc has to be reinstalled as Spider pointed out. Prepare for some work... And, as a side note, never ever use an emerge -d without emerge -dp first (and putting all you can't afford to lose in /var/cache/edb/world)! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] installation-problems with network-cards (sis900, 3c905)
martin morawetz wrote:
Arne Vogel wrote:
martin morawetz wrote:
hi,
I'm trying to install a stage1 gentoo linux from a livecd.
Problem No1:
Kernel-Panic while booting from the cd. The last line is: ''Starting
USB
and PCI hotplugging...'
The next time I booted with the parameters: 'gentoo nohotplug'
That worked, but I got the message 'No network device autodetected'.
Two network-devices are available. A SiS 900 and a 3Com (3c905c-TX)
card.
I searched for the drivers as described in the gentoo-installation
instruction, but
loading them leads to the same result -> kernel panic.
'modprobe sis900'
'modprobe 3c59x'
('modprobe ieee1394' for the firewire worked fine)
Not knowing how to resolve the problem I continued without
having the network configured..
I untared stage1, the portage-snapshot and the source-code from the
livecd, did
everything necessary as described in the documentation handbook, but
when it
comes to the bootstrap-part problem No 2 arises.
Which kernel sources did you choose? I also got a kernel panic
booting from the gentoo-sources kernel (2.4.22). I switched to a
"vanilla" 2.4.24 instead, and it works fine. I never had a boot-up
kernel panic with a vanilla kernel, there must be something wrong
with the patched version.
2.4.21
Sorry, I didn't quite get that "boot from CD" part! :-}
Yes, but is it an "official" vanilla kernel, e.g. as available from
www.kernel.org? Unfortunately, it seems the patched gentoo version has
some reliability problems, though I don't know what version is actually
included with the Gentoo LiveCD.
If it's a bad kernel, maybe try booting Knoppix instead. Then, if you
have broadband internet access, you can install from Knoppix (just
follow the instructions in the manual), and use your system at that time
already (I haven't used Knoppix yet, but I guess you can just create a
home directory on the computer's HD for personal files and settings).
I also have ideas for installing from Knoppix with packages taken from
the Gentoo CDs, but this would be getting too complicated right now
since you probably have broadband access and that's why you need to get
the network card running early anyway!
How did you switch? The live-cd that I downloaded
(pentium4-1.4-20030911-cd1.iso
<http://gentoo.inode.at/releases/x86/1.4/livecd/pentium4/pentium4-1.4-20030911-cd1.iso>)
a few days ago provides a 2.4.21 kernel and a 2.4.20 smp kernel, so I
actually had no
choice because I have a single processor-machine.
I was installing from SuSE and had a "vanilla" 2.4.24 lying around,
which I installed manually (only after the rest of the system was
already finished, including KDE, so I could boot right into a system
with a working GUI). When doing this, you have to be careful to install
the kernel modules, the kernel and lilo into the right directories rsp.
onto the right partitions (e.g. the right /lib/modules, the right /boot
and lilo using the right config file, by default /etc/lilo.conf, but
under chroot!). It's probably best to do that all in the chrooted
environment, everything else is too easy to mess up.
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] Possibly copy protected CDs?
Mark Knecht wrote: On Tue, 2004-02-03 at 17:28, Marshal Newrock wrote: On Tue, 3 Feb 2004, Mark Knecht wrote: How about just unmounting the disk and trying: dd if=/dev/cdrom of=EXILE_DISK_1.iso perhaps? That'll give you an exact (and burnable) copy of the disk. Not a happy dd process... Gentoo2 root # dd if=/dev/cdrom of=EXILE_DISK_1.iso dd: reading `/dev/cdrom':Input/oupuut error 3304+0 records in 3304+0 records out Gentoo2 root # That's correct. You get an I/O error when it hits the end of the disk. If you mount the .iso on loopback, you'll see it's complete. readcd (which I think comes with cdrtools) also does a similar thing. That's really interesting. Can you explain 'mounting on loopback'? What is that? I only have one Linux book (Linux in a Nutshell) and it doesn't have loopback in the index. It allows you to mount a file system from a file (it loops accesses to this nested file system back into VFS, hence the name). As already pointed out, you need kernel support and the "-o loop" option to mount. For example, you can mount an ISO image resting on a hard disk-based FS to one directory, and then mount a ROMFS image off the mounted ISO FS to another directory. I actually tried this, it really works! You just have to unmount in the reverse order of mounting these nested file systems (otherwise umount will simply tell you that the file system is still in use). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Using transparent terminal...
Diego Zamboni wrote: i m using Eterm and aterm, both can be transparent... (tested with: blackbox, fluxbox and KDE 3.x) There seems to be a confusion: Most modern terminal programs (aterm, Eterm, gnome-terminal, konsole, etc.) plus some others (Kopete, gdesklets, etc.) implement "pseudo-transparency". This is achieved by the program grabbing the chunk of background underneath its window, and redisplaying it as its own background. The effect is that of a transparent (or translucent) window, but it is not really transparent: if your "transparent" window is on top of some other window, it will still show the background image, and not the window underneath. Right, I just emerged Eterm, and it only supports the background image pseudo-transparency. Nice feature anyway, but not true alpha blending. It's also not updated immediately, so the image "hops" when moving the window. As for Fresco, you said it's not stable yet? For the moment I'll rather stick with a stable XFree86 4.3.0 than sacrifice stability for some rather superfluous special FX. Would be good for bragging to one's (Windoze-using) friends, though! ;-) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] What doesnt work with 2.6 ?
Alan wrote: On Wed, Feb 04, 2004 at 07:48:12AM +0800, William Kenworthy wrote: There's lots of "use 2.6" messages here, but does anyone have a list of what doesnt work/needs work list for general desktop machines before I take the leap? It's not really a leap as you can swap kernels back and forth with only a reboot, no need to recompile everything or any such sillyness :) Personally I have problems with my logitech mx700 mouse (ps2). There were speed issues (different sampling method or something) which are solved with a kernel parameter (search the archives) but with the same X config my extra buttons and scroll wheel have stopped working properly. This is the only reason I'm not back with 2.4 for my desktop :-\ alan Hmm, my mouse wheel (Fujitsu-Siemens mouse) does work. Here's the XF86Config section, maybe it helps (snipped all the comments). XF86 version is 4.3.0: Section "InputDevice" Identifier "Mouse1" Driver "mouse" Option "Protocol""imps/2" Option "Device" "/dev/mouse" Option "ZAxisMapping" "4 5" EndSection -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] What doesnt work with 2.6 ?
Alexander Klink wrote:
Hi,
On Wed, Feb 04, 2004 at 01:17:32AM +0100, Arne Vogel wrote:
Well, 3D-Hardware-Acceleration will not work until NVIDIA provides an
updated driver.
Hmmm, weird, I downloaded the latest driver from the NVIDIA site and
it seems to work quite well using 2.6.1
The newest version now works, thx!
Apart from that, the kernel-included "nv" driver seems
to work without problems (tested on GeForce 5200FX).
Not for me, somehow 2.6.1 and nv seemed to have some problems (messed up
graphics when starting X)
experience problems. The system responsitivity under heavy I/O load
("tar xjf"...) was greatly improved, so personally I hope these 2 things
that prevent me from switching, non-working Hardware OpenGL and ADSL,
are fixed soon...
Is kernel pppoe that much better performance-wise? I still run the
old system on my router, mainly out of lazyness ...
rp-pppoe is user space, and I don't mind. The problem was that ADSL
didn't run at all,
but adding the "ppp for sync tty" module now fixed it.
Another thing I noticed was broken in 2.6.1 on my Debian machine was
the IHC dhcp server - it complained because it was compiled with a 2.4
kernel, a simple recompile solved the problem, though.
Greetings,
Alex
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] What doesnt work with 2.6 ?
Alex Nelson wrote: Arne Vogel wrote: William Kenworthy wrote: There's lots of "use 2.6" messages here, but does anyone have a list of what doesnt work/needs work list for general desktop machines before I take the leap? e.g, is older hardware supported (PIIX - since they removed the tuning option from 2.4 this has been a disaster for me) nvidia keeps cropping up Well, 3D-Hardware-Acceleration will not work until NVIDIA provides an updated driver. Apart from that, the kernel-included "nv" driver seems to work without problems (tested on GeForce 5200FX). I cannot get rp-pppoe (ADSL), version 3.5, running on 2.6, too. Seems like a change in a kernel interface broke it. Strangely enough, ADSL via SuSE's smpppd did continue to work (I switched to Gentoo only recently). Update: I got it running, a kernel module for PPP over tty was missing (despite originally copying the 2.4.21 config to the 2.6.1 dir). Probably they renamed some config options which caused the module to "disappear". snip Ummm. I am running the latest NVIDIA drivers (5336-r1) on a FX5200 board with OpenGL and it runs just fine. In fact, it runs America's Army faster and with higher framerates than the same hardware under Windows. Yes it does taint the kernel, but I am fine with that for the performance. The biggest issue I have with 2.6.x is the problem with using KVM's. I have to plug my mouse directly into my desktop for the mouse to work correctly with the 2.6.x kernels. Sound and performance have been great otherwise. As with anything else, use caution when making the switch. Keep a copy of your current kernel around until you are sure that 2.6.x is for you. Good luck!! Ok thanks! I previously tried to install 5328 for kernel 2.6.1 and it refused to install the kernel module. I now got 5336, and it works, except that KDE doesn't start (undefined symbol in /usr/lib/libGLcore.so.1). GNOME and America's Army do work. Probably some library was linked against a previous version of libGLcore; I'm trying an emerge kdelibs now. If this works, I'm going to make 2.6.1 my default kernel! -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Using transparent terminal...
[EMAIL PROTECTED] wrote: Hi, I want to use a transparent terminal. I tried gnome terminal and eterm (Eterm -O). They support transparency feature but not actually transparent! I couldn't see the background and can just see the background picture of terminal or background of transparent tone colors. It seems the kde and gnome terminals only support blending in the background image (of the terminal or desktop). I haven't seen or heard of a really transparent terminal so far. Is there any actually transparent terminal??? How can I use that? I saw many guys using that... -- [EMAIL PROTECTED] mailing list -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] gentoo-sources-2.4.22-r5 BIG PROBLEMS
_JusSx_ wrote: Hi, i have installed gentoo-sources-2.4.22-r5 and i got a big problem. Everytime i shutdown my system my kernel start the process, so i can see from console, but my comp will never poweoff Thanx in advance You will probably need to enable APM (or ACPI?) support in the kernel. (in the kernel configuration, General Setup -> Power Management Support etc.) -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] What doesnt work with 2.6 ?
Brett I. Holcomb wrote: Is this renaming something that the ebuild does? I don't know - it should. I installed it manually (that was still under SuSE). And if so what happens when you go back to the 2.4 series - do the new tools work for that or do you have to rename them again? No, the new utils will check what kernel is running and transparently call the old ones if it's 2.4 or earlier. Arne Vogel wrote: Michael W. Holdeman wrote: I can't even get it to boot! Do you have up-to-date module-init-tools? Module loading will otherwise fail. The old tools (modprobe, insmod, rmmod, lsmod) must be renamed to xxx.old because they will be invoked when booting pre-2.6. Also, some modules have been renamed, so you will have to provide an adjusted /etc/modprobe.conf. Best check /var/log/messages to see *what* exactly fails (for me it was e.g. psmouse which was previously called mousedev). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] installation-problems with network-cards (sis900, 3c905)
martin morawetz wrote:
hi,
I'm trying to install a stage1 gentoo linux from a livecd.
Problem No1:
Kernel-Panic while booting from the cd. The last line is: ''Starting USB
and PCI hotplugging...'
The next time I booted with the parameters: 'gentoo nohotplug'
That worked, but I got the message 'No network device autodetected'.
Two network-devices are available. A SiS 900 and a 3Com (3c905c-TX) card.
I searched for the drivers as described in the gentoo-installation
instruction, but
loading them leads to the same result -> kernel panic.
'modprobe sis900'
'modprobe 3c59x'
('modprobe ieee1394' for the firewire worked fine)
Not knowing how to resolve the problem I continued without
having the network configured..
I untared stage1, the portage-snapshot and the source-code from the
livecd, did
everything necessary as described in the documentation handbook, but
when it
comes to the bootstrap-part problem No 2 arises.
Which kernel sources did you choose? I also got a kernel panic booting
from the gentoo-sources kernel (2.4.22). I switched to a "vanilla"
2.4.24 instead, and it works fine. I never had a boot-up kernel panic
with a vanilla kernel, there must be something wrong with the patched
version.
--
[EMAIL PROTECTED] mailing list
Re: [gentoo-user] What doesnt work with 2.6 ?
Michael W. Holdeman wrote: I can't even get it to boot! Do you have up-to-date module-init-tools? Module loading will otherwise fail. The old tools (modprobe, insmod, rmmod, lsmod) must be renamed to xxx.old because they will be invoked when booting pre-2.6. Also, some modules have been renamed, so you will have to provide an adjusted /etc/modprobe.conf. Best check /var/log/messages to see *what* exactly fails (for me it was e.g. psmouse which was previously called mousedev). -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] What doesnt work with 2.6 ?
William Kenworthy wrote:
There's lots of "use 2.6" messages here, but does anyone have a list of
what doesnt work/needs work list for general desktop machines before I
take the leap?
e.g, is older hardware supported (PIIX - since they removed the tuning
option from 2.4 this has been a disaster for me)
nvidia keeps cropping up
Well, 3D-Hardware-Acceleration will not work until NVIDIA provides an
updated driver. Apart from that, the kernel-included "nv" driver seems
to work without problems (tested on GeForce 5200FX).
I cannot get rp-pppoe (ADSL), version 3.5, running on 2.6, too. Seems
like a change in a kernel interface broke it. Strangely enough, ADSL via
SuSE's smpppd did continue to work (I switched to Gentoo only recently).
XFree looks like it has problems
Version 4.3.0 works fine under 2.6.1 on my machine. YMMV.
XFS and some other (unspecified) file systems have problems
Well, if you've heard these have problems and use those file systems, I
would advise caution until these things are worked out. Testing a
scanner is one thing, wasting an FS another...
In any case, if you want to try 2.6.x, make sure you install it
alongside the one you're using now so you can easily switch back if you
experience problems. The system responsitivity under heavy I/O load
("tar xjf"...) was greatly improved, so personally I hope these 2 things
that prevent me from switching, non-working Hardware OpenGL and ADSL,
are fixed soon...
--
[EMAIL PROTECTED] mailing list
