uudruid74

I used to use screen back in the 90s so I could start a compile at work and then go home and dial in to the server to check on it. I think I even had it on my 3b2... yeah I had an AT&T 3b2 running Unix SysVr3.2 For new users that don't know the juice/tmux (or screen or byobu) combination, I forgot to mention you can swipe left/right to change terminal screens. Ill use any of the three - byobu just has cool status bar applets and I think tmux has more features for a couple weird things I do with it. I also use ES File manager mainly for the Sftp (usually over USB) but I'm looking to replace it. The author sends detailed information about everything you do with it to some Chinese server. I firewall off the two IPs it goes to. I have real issues with apps that immediately open a connection to the Internet when you open them. People need firewalls and packet sniffers on their phones! Yes, I have a GUI packet sniffer on my phone. We kinda need an OS that does a two-way data sanitize. Programs should scrub data before putting it into sql or a command line, but the reverse should happen. Programs shouldn't be able to open sockets to addresses that don't originate from the user (not practical on today's systems, but if things we're designed to empower users rather than application vendors may it could have happened). But, I'm wandering off topic.

NO, wait! WTF am I smoking? I must have been looking at the wrong screen or soemthing. I see a failure message. And it looks like its complaining about ruby21 now. And I try it again and it complains about ruby19!! What is IT smoking? navi ~ # ls -l /usr/bin/ruby lrwxrwxrwx 1 root root 6 Jun 29 19:30 /usr/bin/ruby -> ruby22 navi ~ # ls -l /usr/bin/gem lrwxrwxrwx 1 root root 5 Jun 29 19:30 /usr/bin/gem -> gem22 Emerge: >>> Failed to emerge www-servers/tengine-2.1.0-r1, Log file: >>> '/var/tmp/portage/www-servers/tengine-2.1.0-r1/temp/build.log' * Messages for package www-servers/tengine-2.1.0-r1: * tengine group already exist. * group creation step skipped. * tengine user already exist. * user creation step skipped. * ERROR: www-servers/tengine-2.1.0-r1 failed (prepare phase): * Passenger premake for /usr/bin/ruby19 failed! * * Call stack: * ebuild.sh, line 93: Called src_prepare * environment, line 5724: Called _ruby_each_implementation 'passenger_premake' * environment, line 581: Called _ruby_invoke_environment 'ruby19' 'passenger_premake' * environment, line 676: Called passenger_premake * environment, line 4588: Called die * The specific snippet of code: * rake -m nginx || die "Passenger premake for ${RUBY} failed!" * * If you need support, post the output of `emerge --info '=www-servers/tengine-2.1.0-r1'`, * the complete build log and the output of `emerge -pqv '=www-servers/tengine-2.1.0-r1'`. * The complete build log is located at '/var/tmp/portage/www-servers/tengine-2.1.0-r1/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/www-servers/tengine-2.1.0-r1/temp/environment'. * Working directory: '/var/tmp/portage/www-servers/tengine-2.1.0-r1/work/ruby19/tengine-2.1.0/passenger-release-5.0.6' * S: '/var/tmp/portage/www-servers/tengine-2.1.0-r1/work/ruby19/tengine-2.1.0' I am not permitted to upload this type of file (build.log). So here's a paste http://paste.ubuntu.com/11796523/ Why is the spellchecker underlining EVERYTHING?

I just wanted to start a thread that contains the most useful tools and tricks; stuff you end up installing because it makes your life so much easier. Here's a couple to get you started. VIM In your non-root .vimrc cmap w!! w !sudo tee "%" >/dev/null If :w! fails because you forgot to su/sudo, just use ':w!!' SSH the Byobu way! I generally use mosh since it can be forgiving of slow and interruptable links, but I also go one step further and install byobu (its a layer on top of tmux/screen, tmux is default). It will save your session from a network drop! If you've never use tmux or screen you are in for a treat. It multiplexes your terminal so that you can have multiple terminals through the same connection. But, its also a session manager and this is more important (IMHO). Start an emerge and then go ahead and detach (^A d) and you can drop your ssh connection and everything continues in the background. When you connect to byobu again, everything is there as if you never disconnected. Every terminal! You can even connect from multiple machines and share the open terminals if you wish. I use an Android APP called JuiceSSH (and the AnySoftKeyboard with the SSH extension keyboard, although Juice gives you the other keys you need). It manages your SSH connections and lets you connect from your phone. This means I can pull out my phone and tap a button and be logged in to a server, check on a compile or update or view a log, and then detach from byobu really fast. I keep byobu on the gnome "Drop Down Terminal" (desktop extension) so I have a terminal that I never lose track of and it multiplexes with byobu. It also stays active if I log out of Gnome!! Its session doesn't die when you log out of your desktop, its persistent. Also, if I "Juice" in, the terminal size is changed byobu to fit the phone screen - so I can see byobu on my laptop resize as I rotate my phone (yeah, not very useful, but it looks cool).

I think I see what is happening. The HOW-TO mentions to specify it to be external, like this: TENGINE_EXTERNAL_MODULES_HTTP="passenger" and I just realized that external could mean that the build is external to tengine, so I have to build it separately. Sure enough, passenger has a separate ebuild, which wants to pull in apache! But, it tried to build passenger when I emerged tengine, so I'll assume that was correct and the passenger ebuild is just for apache. Well ... it finished successfully! How odd. I could have sworn I tried ruby22. And BTW, 'eselect' is likely the preferred way to manage your Ruby version rather than a manual symlink. It lets you switch back and forth. Anyway, thanks for the help!

OK, I'm new to tengine and passenger and just kinda blindly following the How-To's at this point. I don't know if I really need passenger yet, but I'm thinking something is going to need it eventually. Here's where I'm at (and I tried to set Ruby to 19 via eselect, didn't help): >>> '/var/tmp/portage/www-servers/tengine-2.1.0-r1/temp/build.log' * Messages for package www-servers/tengine-2.1.0-r1: * tengine group already exist. * group creation step skipped. * tengine user already exist. * user creation step skipped. * ERROR: www-servers/tengine-2.1.0-r1 failed (prepare phase): * Passenger premake for /usr/bin/ruby19 failed! * * Call stack: * ebuild.sh, line 93: Called src_prepare * environment, line 5730: Called _ruby_each_implementation 'passenger_premake' * environment, line 581: Called _ruby_invoke_environment 'ruby19' 'passenger_premake' * environment, line 676: Called passenger_premake * environment, line 4594: Called die * The specific snippet of code: * rake -m nginx || die "Passenger premake for ${RUBY} failed!" * * If you need support, post the output of `emerge --info '=www-servers/tengine-2.1.0-r1'`, * the complete build log and the output of `emerge -pqv '=www-servers/tengine-2.1.0-r1'`. * The complete build log is located at '/var/tmp/portage/www-servers/tengine-2.1.0-r1/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/www-servers/tengine-2.1.0-r1/temp/environment'. * Working directory: '/var/tmp/portage/www-servers/tengine-2.1.0-r1/work/ruby19/tengine-2.1.0/passenger-release-5.0.6' * S: '/var/tmp/portage/www-servers/tengine-2.1.0-r1/work/ruby19/tengine-2.1.0' Should I post the contents of any files?

I probably shouldn't necrobump old threads, but school for me was mostly a waste of time. I was more or less teaching the teachers and making more than them, so I quit. Then again, I've been using computers since I was 3 or 4. I could do basic BASIC in kindergarten. My advice is to figure out what you are most passionate about and dive in deep. You'll end up being amazingly good at it and someone is going to come along and need those skills. Basically, find your niche. My niche seems to be abandoned, half-finished projects :) What do I do? Here's one of the more fun jobs I worked at :

Gcc supports a "native" target where it detects what cpu its running on and then uses that for almost code generation options. Unless you build on a host that is different from the target, why not use this feature?

I wish there was an automatic installer where it looked at your hardware, made some recommendations and then let you how to install from there using a simple set of meta-packages. Ex: You probably don't want to install the latest kde or something on a 512Mb system, and you likely won't use it as a workstation, so that leaves tiny desktop with all the options trimmed, or embedded server. I started something along those lines, but I had to put it on the back burner.

Odd - I had it set to 30 days. Maybe a bug in the uploader. I'll try it manually.. emerge --info http://pastebin.com/4xZVEBmV Xorg.log http://pastebin.com/KnAPZQrp

new install, xorg 1.16.1_r1 (reporting working with these drivers by drobbins on 11/6 https://bugs.funtoo.org/browse/FL-1638), Xorg.log file http://pastebin.com/NaArJWnD , emerge --info http://pastebin.com/hvztpAkb , the at-drivers is 14.9.-r1 (same bug report). kernel is ck-sources 3.14.4 (configured by me) Yes, I read the install guide for Funtoo. And the one for Arch which is much more informative https://wiki.archlinux.org/index.php/AMD_Catalyst#Troubleshooting Before you ask, radeon is configured as module, but listed in /etc/modprobe.d/blacklist.conf so it shouldn't be loading. At least I know I can ignore that error and try something else.

Thats what I needed to know. I figured it was just checking there and would check other directories next. Thanks

OK - here is the only line in my logs that might explain: Taro ~ # grep '(EE)' /var/log/Xorg.0.log (WW) warning, (EE) error, (NI) not implemented, (??) unknown. [ 478.015] (EE) AIGLX error: failed to open /usr/X11R6/lib64/modules/dri/fglrx_dri.so, error[/usr/X11R6/lib64/modules/dri/fglrx_dri.so: cannot open shared object file: No such file or directory] This is odd because the file is in /usr/lib64/modules/dri. Why is it looking in the wrong place and how do I tell it to look in the right place? but later, it says .. [ 478.266] ukiOpenByBusid: ukiGetBusid reports PCI:0:1:0 [ 478.904] (II) AIGLX: Loaded and initialized OpenGL driver(II) GLX: Initialized DRI GL provider for screen 0 So .. did it initialize or not? Did it finally look in the right place and find it? Am I chasing shadows? The X server starts up, initializes the monitors (no picture though), and then shuts down. It never says WHY it shuts down. I can pastebin the full log if anyone wants to read it. However, it leaves my screen in an unusable state. Likely, because of: [ 481.169] (II) fglrx(0): Restoring Recent Mode via PCS is not supported in RANDR 1.2 capable environments So ... once I try starting it, I have to reboot! That makes debugging this myself time consuming. Anyone have any suggestions, or seen this before?

486s? My 8Mhz Atari ST can open a window faster than my multi-Ghz Linux box. Application load speed was about the same and considering how slow that old hard drive was .... Vi mode inside emacs ? Wow. My head exploded. Emacs really DOES do everything. Thats like having an Amiga mode inside your Atari. Oh wait .... that was called Spectrum 512. We did have that, we just didn't admit it. It might have ended the war. And the MCIBTYC wars spurned years of healthy innovation. Bringing back the Desktop wars is tempting, but these days its more of a foot fight - no one rallies behind a cause, they just sling food at targets all over the place. I use the evil mouse mode in vim. Latency I can kind of understand. Linux wasn't designed for low-latency, and it has a lot going on under the hood that has to be protected from concurrent access - which means a lot of nasty locking. Most of the design choices have been cleaned up and latency issues are getting better, but there is still work to be done. Thats why I don't believe that the same installation should be used for servers and desktops both. There are a lot of ways I'd configure them differently, and the kernel at the very least should be different. But there is also a difference between a hard real-time application that needs absolute latency at the expense of all else (RT patch set), and a desktop that should simply be low-latency and not let your video player skip when you multitask (CK patch set). And of course, on a server your video player be damned - we're gonna do it the fast way and serve as many database connections as possible. Most desktop linux systems only have partial low-latency patches - being stuck in the middle. I just sold my old 3B2 - the ad said to bring a truck and come get it before I take it to the landfill. He actually paid me to ship it to him. I forget if its 4 or 8MB of RAM, dual 30MB drives I think. Not sure. Don't think that model even had a hardware MMU, but it would run a dozen users on it at once without a noticeable slowdown. All text based via serial ports, but still impressive multitasking for the day.

I run ati-drivers which build against the kernel headers. I happen to know that enough changed when 3.16 was released the ati-drivers would fail to build its kernel module, and that was with kernel headers that actually matched the running kernel. So, knowing the API changed that much in 3.16, running a 3.13 kernel and headers for 3.17 just sounds like suicide to me, especially since I need the ATI drivers because they magically fix some LyX issues and I can't live without LyX. But - you are guys are telling me that it won't hurt, and that I can't stop it from doing its thing anyway. I just don't see why I would need 2 copies in the first place. Or don't the kernel sources come with headers? *sigh* If they don't have their own copy, then I'd be afraid to rebuild the kernel with mis-matched headers knowing how much changed from release to release of the kernel. I know libc stuff won't break, but when I next get a video driver update that refuses to install because the kernel got updated ... and most modern distros now give you a black screen if the drivers don't install and they have cntrl-alt-backspace turned off. Elementary even has terminal switching turned off. But thats okay, this is like Windows now - we can reboot into "Safe Mode". Thats why I'm switching over to Funtoo - I'm tired of being told what I can and can't do with my own computer.

It didn't used to be that way. Today, we are integrating search engines into desktops (a bad interface on top of a security and privacy violation) rather than fixing bugs. I was just told that the inconsistent interface items in Elementary (some applications can't change themes and stuck on "dark", and if you change the rest of the windows to dark to match than your notifications are black on black - *doh* - idiots!) The reply (after about 3 or 4 months) is that is an intentional developer feature. The devs can't change there own theme that they have to be locked in to one that is non-functional!?? I was trying to figure out why I could get an X connection over tcp the other day, in-spite of the fact that I changed my /etc/X11/xinit/xserverrc. I did some research and lightdm intentionally has it hard-coded to ignore xserverrc (which worked fine) and refuse tcp connections for security purposes. My house is behind a double-nat (no incoming connections possible) AND I have a Cisco ASA - I don't care about security. There was a patch made so that you could turn on listening over TCP with a cryptic configuration parameter (what happened to my shell scripts?), but this feature seems to be broken - it doesn't work. Why are they removing functionality from my apps? There are so many cool things you can do with X, but I suppose they'll all be gone once we're forced into Wayland. And in the old days, the developers were fewer, but they were rock-solid. Now its a huge ecosystem - too many cooks spoil the soup. Its like Gentoo vs Funtoo - a single BDFL keeps the direction focused. Pulseaudio is okay, but it seems to be kinda resource heavy. I like being able to reroute audio at will, have different apps default to different audio devices (I have many), or broadcast to multiple devices ... and while it works with jack with just an extra click, they should be integrated. Consolekit I kinda understand. Its a really ancient security problem that was never addressed really well and every distro did things differently and it made using pluggable devices problematic sometimes. I don't like emacs, but I hear you on latex! I cheat and use LyX. Awesome program, but I tend to have slow display issues - thats the reason I use the ati binary drivers - they make my LyX performance problems go away. I don't care about games. Gstreamer makes a lot of sense. Its better to consolidate the APIs so that each program doesn't have a mess of libraries bloating the system. Then you can work on optimization instead of each program trying to deal with compatibility and optimization on its own. The new Gnome is a bit resource heavy, but I after using it, I kinda like it. I hear the old ones weren't as good. And I don't like KDE at all, but at least we get a choice. Elementary OS won't let me install Gnome :/ Thats why I'm moving over to Funtoo (that and I hate systemd and like the direction Funtoo is going).

Can you explain to me what that means? And I just got the old sandbox violation back. Its not consistent with what error it gives me :/

Did you have the vala USE flag on? Masking that flag for this build will fix it, but its kinda pointless to have the flag if it doesn't work. And I just did a sync a few hours ago and it didn't fix it. I just did ANOTHER sync (I'm sure glad funtoo uses git and not rsync) and it looks like it might work this time! I always update dependencies and such, so its pulling in a few updates before it does gedit - and its done! SOLVED I know that packages are sometimes masked for testing. Is it possible to mask a package by USE flags and version and not just version? If this were possible, you could get the latest version where a particular set of USE flags was tested and working. If you need a newer version, turn off a broken flag to unmask the newer one. If you needed to add a USE flag, it would automatically mask the package version that has problems and warn the user you have to downgrade. But that would depend on a lot of community feedback as to what works and what doesn't. Anyway - it was just an idea.

The sandbox violation went away!!! But, it still fails ... but the log looks clean. Look: Emerge: checking whether to build static libraries... yes checking how to run the C++ preprocessor... x86_64-pc-linux-gnu-g++ -E config.status: creating ext/phar/phar.1 config.status: creating ext/phar/phar.phar.1 config.status: creating main/php_config.h config.status: executing libtool commands config.status: executing default commands >>> Source configured. >>> Failed to emerge dev-lang/php-5.5.19, Log file: >>> '/var/tmp/portage/dev-lang/php-5.5.19/temp/build.log' The log ends with "Source configured". Yes, I have bash for /bin/sh ... its not the issue. Pretty much the entire system has been rebuilt. However, I found one thing in the log that looks suspect - what do you think? ESC[33;01m*ESC[0m This package has a configure.in file which has long been deprecated. Please ESC[33;01m*ESC[0m update it to use configure.ac instead as newer versions of autotools will die ESC[33;01m*ESC[0m when it finds this file. See https://bugs.gentoo.org/426262 for details. ESC[32;01m*ESC[0m Running autoconf ... So .. do I have to downgrade autotools, emerge php and then upgrade autotools again?!?

It's not that I don't know how. I'm busy USING the system. I've decided to use ShellInABox for a project and it hasn't been updated in about 2 years, so I'm picking up all the patches and forks people have dropped all over the net, plus fixing a bunch of stuff myself and reconfiguring how it handles colors, adding a color reducer for programs that think your terminal can handle 256 color ... *sigh* ... and its all in Javascript. Haven't used that in ages. I just assumed that after a decade this stuff would be more mature. However, it seems like a regression. I used to use Gentoo a long time ago. In fact, a company I worked for (back in the dark ages) claims that their network security product successfully protected Spamhaus.org from a massive DDOS attack. Halloween ... 2004? Nope! The applicance failed and was taken out of the loop. What kept the site up was me sticking a nice little filter line into the Squid proxy .. running GENTOO!! Very nice set-up, too. I had everything served out of LDAP with a couple scripts for the few things that couldn't read the tree directly, with a nice web interface. Very smooth. And I think I was running ~arch back then. Now that seems to be common. I'd really like to see a bit more stability where things are tested a bit more. The solution seems to be "sync and try again", which might fix problem or pick up another ebuild that will break something else. I know Drobbins was working on a "stable tree" idea, but I haven't seen any completion announcements.

But .. I have the whole kernel sources including the headers. Installing headers for a different version that may have a different set of patches doesn't seem right. I guess I'll let it. :/ Can't break it any more than it already is.

OK, I have my kernel in @system, not in @world, and my kernel sources don't support 3.17 yet because its ck-sources, nor do I want to upgrade my kernel. I'm tired of compiling and tired of stuff being broken. WHY is it trying to install kernel-headers 3.17-r1? My kernel is 3.13, so this sounds like a really BAD idea. Shouldn't portage be smart enough not to do this? Maybe funtoo was a bad idea. Weeks now and still not finished installing! This shouldn't be this difficult. I only did a sync because it was supposed to fix gedit's vapigen thing, and its still broken just giving me different error (see other thread), php is still not building due to sandbox violations, and now I have a new problem I didn't have before. Next step is a hammer! Anyone know how to fix this??

Fresh sync emerge: there are no ebuilds built with USE flags to satisfy "dev-lang/vala:0.26[vapigen]". !!! One of the following packages is required to complete your request: - dev-lang/vala-0.26.1::gentoo (Missing IUSE: vapigen) (dependency required by "app-editors/gedit-3.14.0[vala]" [ebuild]) (dependency required by "gedit" [argument]) I'm too tired to look at it

Musl on the other hand, I've heard of and it looks promising. There are people using it on Gentoo, but it looks like the support is pretty preliminary. I would actually like to see it get better support for bloat-reduction. Linux is getting FAT instead of FAST.

Different problem, different topic, sorry. I just thought you may have remembered it from earlier. Still not sure if its a bug I need to report and waste people's time with, or something else ... http://forums.funtoo.org/index.php?/topic/270-cant-emerge-php-sandbox-violation/

I did a similar trick having the Documents, Pictures, Videos, etc linked to the Windows versions of those directories on my Dad's laptop. That shares Linux and Windows files pretty seamlessly. I also did a trick where I can boot other versions of Linux and then boot the other in a virtual machine. The one in the virtual machine won't be able to see the shared partition and won't mount it, but it fails over to a network mount (shared by the host OS that does have it mounted) in the same location. You won't be able to share the base of your home directory, just subdirectories, but this is probably more organized that way. You could even make symlinks for ~/.cache and such.