Debian broke too many times for me with package dependency circular hell that I've vowed to never use it again. Never had that problem with Ubuntus, which is odd considering it's background. To answer the previous question, I like lubuntu or xubuntu as a lightweight and power user version of ubuntu.
Wii U sucks, but my NNID is da66en. Steam is route66. 3DS is 2938-8099-8160.
Neo Geo Big Red owners club.
2009 PAX Puzzle Quest Champion
I have beat Rygar on the NES and many of you have not.
Weird! Changing /etc/systemd/logind.conf was the only fix that made my lid cause a hibernate.
logind is a shim-wrapper for the Gnome bits which are dependent on systemd. What happens though is that the power-management-daemon doesn't properly suppress logind handling the lid once the session has started, and your logind is still set to "do nothing" so it appears not to work.
I was under the impression this had all been fixed though - I know Linux Mint pulled a bunch of patches from gnome-power-manager explicitly to handle this.
I'm running 14.04, because I assumed the LTS would get all the fixes, so I couldn't say why it's still an issue.
Thanks for the breakdown; I'm having a hell of a time dealing with user quirks while also trying to manage sysadmin, arch and DBA!
So. Doing some digging today without much luck. Are there any "windows-like" solutions for Remote Desktop TO a Linux box? By which I mean:
1. Capture the existing desktop session.
2. But keep the local screen locked so no one can watch you work or just... you know... grab your local keyboard and fuck with shit.
I'm not seeing much to give me hope; I'd have thought someone would have figured this out by now.
X11/X.org uses a client-server protocol, even just running the desktop locally. You should be able to use an X11 client to connect to an existing session. I think?
Unless you're running a distro that has replaced X.org with Wayland or something.
Edit: um, never mind. Five minutes of research and everything is unbelievably awful. Have you looked at xpra? It miiiight do what you want? But ugh. I had no idea it was this bad.
X11/X.org uses a client-server protocol, even just running the desktop locally. You should be able to use an X11 client to connect to an existing session. I think?
Unless you're running a distro that has replaced X.org with Wayland or something.
Oh, there are plenty of ways to grab and forward the existing desktop; I tried both x11vnc and xrdp today.
The problem is that they just "mirror" what's on the local X instance, so anyone that walks the local monitor can watch whatever I'm doing, or even grab the keyboard and mouse and start "fighting" me to do stuff. When you use Remote Desktop to access a Windows box, the local screen locks and only displays a note that someone is logged on remotely.
From a security perspective, the latter is not just a good idea, it's absolutely necessary, and I can't understand how no one in the Linux world has figured this out yet.
normally I'd suggest using X forwarding (no, it's a different thing from what you're thinking) or Xvnc, but they are both completely separated from the remote's physical display, so don't fulfill the exact parameters of what you're asking for
However, as an experiment, in LightDM I was able to "user switch" to the login screen and still access the locked session using x11vnc (despite the overlay LightDM adds even when using ctrl-alt-f7 to forcibly switch back to it). I guess it'd be useful if you have something long running, but any time I have something long running, it's command line and thus is better to put inside screen or tmux anyway.
I wish that someway, somehow, that I could save every one of us
normally I'd suggest using X forwarding (no, it's a different thing from what you're thinking) or Xvnc, but they are both completely separated from the remote's physical display, so don't fulfill the exact parameters of what you're asking for
However, as an experiment, in LightDM I was able to "user switch" to the login screen and still access the locked session using x11vnc (despite the overlay LightDM adds even when using ctrl-alt-f7 to forcibly switch back to it). I guess it'd be useful if you have something long running, but any time I have something long running, it's command line and thus is better to put inside screen or tmux anyway.
Wait how did you accomplish this? Because it's been a constant concern of mine.
most desktop environments should have a switch user option, although admittedly xfce's stuff (which I use) doesn't know how to talk to lightdm (it expects gdm, apparently), so I used lightdm's dm-tool command to do it
I wish that someway, somehow, that I could save every one of us
0
Options
mightyjongyoSour CrrmEast Bay, CaliforniaRegistered Userregular
Anybody using Fedora 20 or 21? What's the maintenance/stability like on those?
Using it for work... I at least haven't noticed any major issues. As always using nvidia drivers on a laptop is a terrible idea. Other than that... pretty frequent updates, nothing broken, Wayland is iffy but it's an option at least.
I have a problem. My boot drive is a SSD (with windows), but linux is installed on another drive (a 1TB HDD). How do I install grub on the SSD so that it can boot from the HDD? For added kicks, (but not really) how do I do this from an ubuntu liveUSB session?
I run Linux Mint because it's easy to set up and stable. That said, I like running the latest version of the AMP part of the LAMP stack in order to keep up with bug fixes and get new features. What I'm wondering is:
Is there an actual reason why the AMP packages in repositories lag so far behind? I managed to get the latest Ubuntu packages for it, but the latest version of PHP it offers is 5.5.9, which came out nearly a year ago. I've tried dotdeb and Ondrej Sury's PPA, but both are bugged for me (which may be the answer to my question).
Why can't the software vendors have their own repositories one can tap into? Like, say you wanted the latest version of Apache. Point your package manager to their bleeding edge repo. Or you wanted the latest stable version of PHP. Point your package manager to their 5.6-stable repo. Or 5.6-stable-debian, or whatever.
I guess I just find it hard to believe that different distros are so different from one another that they require so much testing before release. I know I can compile from source, but that's a PITA that I'm unwilling to do.
I dunno... I come from Windows land where new versions of C# and whatnot either require purchase or an in place update, but the actual installation/update process is fairly mindless. I'm trying to figure out why it can't be that way on linux as well.
Why can't the software vendors have their own repositories one can tap into?
Because no vender wants to have to provide binaries for Ubuntu, Mint, Debian, RHEL, Centos, Slackware, Gentoo (you get the point).
I'm not a big Mint/Ubuntu guy, but you might look for either a PPA to grab them from (if you trust the compiler), or grabbing from SVN if you need bleeding edge.
Why can't the software vendors have their own repositories one can tap into?
Because no vender wants to have to provide binaries for Ubuntu, Mint, Debian, RHEL, Centos, Slackware, Gentoo (you get the point).
I'm not a big Mint/Ubuntu guy, but you might look for either a PPA to grab them from (if you trust the compiler), or grabbing from SVN if you need bleeding edge.
See, I don't get the appeal of that kind of fragmentation. Not in a "I don't like it" way (well... not entirely), but in a "Why the fuck would anyone want to deal with that kind of 'everything is linux, except when it's our flavor of linux" bullshit? What's the advantage? What's the appeal? Why isn't it more reasonable to have a single linux kernel, and build off that? Where productivity software (like programming, web server, and database software) simply targets that so it's universal across distros?
I'm not trying to start an argument or anything. I'm just trying to understand. Everyone that loves linux swears it's so much better than the more mainstream OSes, but I just don't see why, beyond the fact that they're not tied to for-profit entities.
There has to be a reason why that's considered better than what OSX or Windows does.
Why can't the software vendors have their own repositories one can tap into?
Because no vender wants to have to provide binaries for Ubuntu, Mint, Debian, RHEL, Centos, Slackware, Gentoo (you get the point).
I'm not a big Mint/Ubuntu guy, but you might look for either a PPA to grab them from (if you trust the compiler), or grabbing from SVN if you need bleeding edge.
See, I don't get the appeal of that kind of fragmentation. Not in a "I don't like it" way (well... not entirely), but in a "Why the fuck would anyone want to deal with that kind of 'everything is linux, except when it's our flavor of linux" bullshit? What's the advantage? What's the appeal? Why isn't it more reasonable to have a single linux kernel, and build off that? Where productivity software (like programming, web server, and database software) simply targets that so it's universal across distros?
I'm not trying to start an argument or anything. I'm just trying to understand. Everyone that loves linux swears it's so much better than the more mainstream OSes, but I just don't see why, beyond the fact that they're not tied to for-profit entities.
There has to be a reason why that's considered better than what OSX or Windows does.
Well, because it's Open Source. One person/group bundles up a bunch of software, everything they need for an OS. They include all their favorites. They share that with the world. The next guy decides they don't like package X or Y, so they change some things to suit their needs and share it. And so on.
Debian wanted to conform to the GNU Principles, and values stability over features.
Ubuntu wanted the base and package management of Debian, but software that was more current and relevant to desktop users.
Red Hat wanted a Linux that would be suitable for the Enterprise, something they could sell and charge to support.
CentOS wanted to be Red Hat, but without the price tag.
Gentoo mimics the BSD ports system and places an emphasis on compiling everything so it's optimized for your system.
Arch just wants to be bleeding-edge.
That's the point, though. Linux is, in fact, a kernel. A Distribution is what happens when you wrap that kernel with all the other software you need to actually have a functional computer.
Binary compatibility isn't everything. This allows distros to provide stable versions of software and backport security fixes, occasionally fixing the software before the upstream does.
You want software that runs for years without having to update any config files or do major version upgrades? This lets you have it, without any extra work for the vendor. You want the latest and greatest of everything? Source-centric software distribution lets you do that, too. Everyone is happy, except for people used to binary distribution who see this huge confusing mess of slightly incompatible systems. ;-)
I run Linux Mint because it's easy to set up and stable. That said, I like running the latest version of the AMP part of the LAMP stack in order to keep up with bug fixes and get new features. What I'm wondering is:
Is there an actual reason why the AMP packages in repositories lag so far behind? I managed to get the latest Ubuntu packages for it, but the latest version of PHP it offers is 5.5.9, which came out nearly a year ago. I've tried dotdeb and Ondrej Sury's PPA, but both are bugged for me (which may be the answer to my question).
Why can't the software vendors have their own repositories one can tap into? Like, say you wanted the latest version of Apache. Point your package manager to their bleeding edge repo. Or you wanted the latest stable version of PHP. Point your package manager to their 5.6-stable repo. Or 5.6-stable-debian, or whatever.
I guess I just find it hard to believe that different distros are so different from one another that they require so much testing before release. I know I can compile from source, but that's a PITA that I'm unwilling to do.
I dunno... I come from Windows land where new versions of C# and whatnot either require purchase or an in place update, but the actual installation/update process is fairly mindless. I'm trying to figure out why it can't be that way on linux as well.
What makes it magic is the VM system lets you just do "docker pull php/latest" or "docker pull php/5.2" and it'll download a prebuilt image containing exactly that for you built by the project maintainers (there's also lots of custom stuff - or you can roll your own).
EDIT: For example, I built a test version of one of my company's intranet sites by dumping the database onto a docker image, the website onto an apache one, and using the link commands (which automatically patches the hosts files) to recreate it in containers.
Then added a small script which booted all the daemons in foreground mode in separate tmux windows. The whole thing can be setup and torn down in seconds.
I run Linux Mint because it's easy to set up and stable. That said, I like running the latest version of the AMP part of the LAMP stack in order to keep up with bug fixes and get new features. What I'm wondering is:
Is there an actual reason why the AMP packages in repositories lag so far behind? I managed to get the latest Ubuntu packages for it, but the latest version of PHP it offers is 5.5.9, which came out nearly a year ago. I've tried dotdeb and Ondrej Sury's PPA, but both are bugged for me (which may be the answer to my question).
Why can't the software vendors have their own repositories one can tap into? Like, say you wanted the latest version of Apache. Point your package manager to their bleeding edge repo. Or you wanted the latest stable version of PHP. Point your package manager to their 5.6-stable repo. Or 5.6-stable-debian, or whatever.
I guess I just find it hard to believe that different distros are so different from one another that they require so much testing before release. I know I can compile from source, but that's a PITA that I'm unwilling to do.
I dunno... I come from Windows land where new versions of C# and whatnot either require purchase or an in place update, but the actual installation/update process is fairly mindless. I'm trying to figure out why it can't be that way on linux as well.
What makes it magic is the VM system lets you just do "docker pull php/latest" or "docker pull php/5.2" and it'll download a prebuilt image containing exactly that for you built by the project maintainers (there's also lots of custom stuff - or you can roll your own).
EDIT: For example, I built a test version of one of my company's intranet sites by dumping the database onto a docker image, the website onto an apache one, and using the link commands (which automatically patches the hosts files) to recreate it in containers.
Then added a small script which booted all the daemons in foreground mode in separate tmux windows. The whole thing can be setup and torn down in seconds.
Interesting. So, it's possible to have a dock(ed/ered?) version of a linux distro with the full LAMP stack?
I run Linux Mint because it's easy to set up and stable. That said, I like running the latest version of the AMP part of the LAMP stack in order to keep up with bug fixes and get new features. What I'm wondering is:
Is there an actual reason why the AMP packages in repositories lag so far behind? I managed to get the latest Ubuntu packages for it, but the latest version of PHP it offers is 5.5.9, which came out nearly a year ago. I've tried dotdeb and Ondrej Sury's PPA, but both are bugged for me (which may be the answer to my question).
Why can't the software vendors have their own repositories one can tap into? Like, say you wanted the latest version of Apache. Point your package manager to their bleeding edge repo. Or you wanted the latest stable version of PHP. Point your package manager to their 5.6-stable repo. Or 5.6-stable-debian, or whatever.
I guess I just find it hard to believe that different distros are so different from one another that they require so much testing before release. I know I can compile from source, but that's a PITA that I'm unwilling to do.
I dunno... I come from Windows land where new versions of C# and whatnot either require purchase or an in place update, but the actual installation/update process is fairly mindless. I'm trying to figure out why it can't be that way on linux as well.
What makes it magic is the VM system lets you just do "docker pull php/latest" or "docker pull php/5.2" and it'll download a prebuilt image containing exactly that for you built by the project maintainers (there's also lots of custom stuff - or you can roll your own).
EDIT: For example, I built a test version of one of my company's intranet sites by dumping the database onto a docker image, the website onto an apache one, and using the link commands (which automatically patches the hosts files) to recreate it in containers.
Then added a small script which booted all the daemons in foreground mode in separate tmux windows. The whole thing can be setup and torn down in seconds.
Interesting. So, it's possible to have a dock(ed/ered?) version of a linux distro with the full LAMP stack?
Yep. You can grab stuff people have made up like here: https://github.com/ushios/docker-lamp-ubuntu and change the docker file how you want. Or write your own - it's just a sequence of apt-get commands to set things up.
But another way to do it is just to pull down the individual application versions in their own containers, and link them together over the virtual network. This is actually the approach I'm starting to favor because it means I can have each apps console in the foreground while I work. So for instance I have this script for starting up the dev environment of a project I'm working on - it boots each container and attaches it to tmux, so I get the consoles all on one panel between the webserver and database:
This is a really neat way to do it, because it means you can go from docker to deploy without needing to change anything provided the hostname's for are the same (which was why I did it originally - because it was an existing project).
The other thing is that you can use the volume commands to bind files from your local drive into the container system - that's what I'm doing there with the -v switch. All the regular stuff like inotify and the like still works from host to guest too. So you can live edit stuff directly, but pull it into one (or several different) containers.
That's really cool. I could see myself making a generalized Docker template that had the latest stable LAMP and PhpStorm installed, and then spinning one up for each project I have. Then I could just install the project code via Composer and git pulls.
I'm assuming they can save state because they're really just fancy VMs.
That's really cool. I could see myself making a generalized Docker template that had the latest stable LAMP and PhpStorm installed, and then spinning one up for each project I have. Then I could just install the project code via Composer and git pulls.
I'm assuming they can save state because they're really just fancy VMs.
Yes and no - in the sense that it's optional. Docker containers save their state till you kill them (which is different to start/stop). That persists till you destroy the container.
But you can also define 'volumes' for specific folders which docker will persist when the same volume is re-attached to another container (i.e. they have their own name, it's kind of like having a virtual NAS).
Or you can just bind-mount the bits you want to folders on your host.
Or you can push the content to an image which then lets you duplicate it to many containers.
So, would I be able to do the following in Docker:
Define a LAMP development environment template, including IDE?
Work on a specific instance of that template, save it as a snapshot when I'm done working for the day, then reload where I left off?
I'm looking for more than a deployment environment. It'd be awesome if I could spin up an actual full featured work environment on the fly quickly.
EDIT: You may have answered my questions with your last statement about saving an image
IDE could be tricky, though there are ways to make X11 stuff run from within Docker environments - I've not tried it. That said, you would only need to do it once and I know it can be done.
But I'd probably just go with running my IDE on my normal OS, and using Docker to run the server side of things (since that tends to be the most intrusive normally).
Naturally, I'm thinking Bash is the easiest and cleanest solution for this, so I fire up Cygwin and catch myself doing this:
find . -name *.wav* -exec mv "{}" ../ \;
So that would move all the files to the upper directory of where I execute the script, but I need it to move it to the upper directory of each file. How do I work around this?
Naturally, I'm thinking Bash is the easiest and cleanest solution for this, so I fire up Cygwin and catch myself doing this:
find . -name *.wav* -exec mv "{}" ../ \;
So that would move all the files to the upper directory of where I execute the script, but I need it to move it to the upper directory of each file. How do I work around this?
I dunno what the right way to do it with bash is, but here's the horrible way I've solved similar problems in the past.
find . -name '*\.mp3' > mp3list.txt
Now you have a list of the mp3s. Next, you open it in Vim and make a macro. When the dust has settled, the first line looks like
So, it turned out that find has an -execdir flag so that the following command is run from the location of the found file. The resulting 1 line of Bash was this:
I'll be building a new computer tomorrow and I was wondering how well I'd do with SteamOS as the only operating system?
Edit: Looking over this list https://steamdb.info/linux/ I think the only games I'll be missing that I still want to play are the lego games, saints row series, and walking dead. Maybe a few others.
This'll be my first non-windows box since my old mac running OS8. I think I'll take the plunge and not setup a windows dual boot right away.
Veevee on
0
Options
augustwhere you come from is goneRegistered Userregular
I'm kind curious how well SteamOS can work as a primary os but I'd just install Ubuntu 14.10 and then download Steam.
Yeah, SteamOS is still a bit rough. If you want to do things other than games, get Ubuntu. Heck, Ubuntu is going to have a smoother setup process anyway. I don't think the SteamOS one has changed much at all since they first released it.
I personally prefer Fedora, but Steam gaming isn't as smooth as on Ubuntu; some less experienced developers do boneheaded things like dynamically linking to libraries and using Debian-specific library names. (Games ported by Valve or published in Humble Bundles are fine.) It's doable, just occasionally more tedious.
I really don't do anything other than casual computer use like google docs and music/video playing, and games of course. Which probably has a huge part of why I was able to wait as long as I did to upgrade (8 years). Can't remember the last time I opened paint, let alone had photoshop or similar installed. Hell, other than steam and a few other games I don't think there's anything left installed, or what is I haven't touched in years, and this windows install has been active since... 1/6/2010. Huh, Not bad for a windows install that hasn't seen an install of any antivirus software.
Re: music and videos, does SteamOS even ship with an MP3 codec? It must, for the Steam music player, right? Are they just eating the licensing fee? How does that work.
Steam supports mp3s through it's music player, yeah. As for the licensing fee, they're not saying, but it's valve and they never talk about stuff like that to the public.
I also don't have any mp3s and just stream all music from pandora and other services, so it's not too big of a worry for me and I don't think any self-respecting dev would use mp3s for their sound files anyway.
So, does anyone have any idea why my ls in Cygwin runs out of memory trying to dump an index of my work NAS to a file? I'm still working with this hellbeast of a server containing 10s of thousands of folders and I now need to copy a subset of those to another hdd. I made a bash script that FINDs the folders using a simple .txt database and copies them over, but goddamn FIND searches the entire NAS every time, resulting in a 10 minute down-time searching, 45s copying and repeat for 50.000 folders.. So, that's fun.
So I try using updatedb to make a database and then using locate to copy the files instead, but the command crashes.. out of memory, as far as I can tell. Huh. Looks like this
Error (Cygwin):
1 [main] find 8188 C:\cygwin\bin\find.exe: *** fatal error - crealld have returned NULL
/usr/bin/updatedb: line 411: 8188 Hangup $find $SEARCHPINDOPTIONS \( $prunefs_exp -type d -regex "$PRUNEREGEX" \) -prune -o $prion
Alrighty, then. That's weird, well, whatever.. I'll just dump the directories to a text file and use a script with grep to copy then, should still be faster then using FIND right now. So..
ls -RD1 > ~/index.txt
Error (Cygwin again):
0 [main] ls 5148 C:\cygwin\bin\ls.exe: *** fatal error - crealloc would have returned NULL
Hangup
Well, fuck. What do I do, I'm (again) looking at literally hours and hours of manual copy/paste work here.
Yeah, I'm on Windows. Maybe I should use Powershell instead? I've never used it, ever, but Cygwin seems to be behaving a lot differently than a regular unix setup..
Anyway, I jumped through a ton of hoops to export a file list of what I needed, so I'm getting stuff done. Fingers crossed.
Finally got a replacement laptop! It's used and refurbished, which means it's in pretty good condition and got it for cheap.
Of course I immediately added more RAM (because I am now, apparently, the sort of person who has spare DDR2 laptop RAM sticks just... lying around) and swapped out the hard drive for something bigger. Hooray for voiding warranties!
Went ahead and installed Kubuntu 14.04, since Gnome was driving me up a wall. KDE drove me up even MORE of a wall until I realized just how much I could customize. (Well, I STILL don't know how to make the scroll buttons move the view faster...) Consider me pleased and generally happy.
I like how, after a logoff or a restart, it is happy to restart my applications right where I left them. It does seem to take a bit longer than I'd like to go from "the desktop has appeared" to "you can actually do stuff". Also Kmail is pretty darn nice and the fact there is an IRC client included with the Kubuntu install is a good idea. Apparently installing Steam on this is gonna be a bit of a pain...
Overall I'd call my experience so far a sold "B." Certainly better than vanilla Ubuntu.
It pretty much is Ubuntu, they just swap a few of the default installed packaged around (KDE instead of Unity, etc). You can always apt-get Unity or Gnome or whatever.
And getting Steam running on Ubuntu was pretty easy last I tried.
Posts
Edit: it needs to run on some older hardware too so keep that in mind.
Or you could get some Arch going.
Or rice up a Gentoo.
Is Slackware still a thing?
You know what? Take it to the max. Linux from Scratch. You want to learn how a Linux distro really works? This will teach you.
Neo Geo Big Red owners club.
2009 PAX Puzzle Quest Champion
I have beat Rygar on the NES and many of you have not.
I'm running 14.04, because I assumed the LTS would get all the fixes, so I couldn't say why it's still an issue.
Thanks for the breakdown; I'm having a hell of a time dealing with user quirks while also trying to manage sysadmin, arch and DBA!
1. Capture the existing desktop session.
2. But keep the local screen locked so no one can watch you work or just... you know... grab your local keyboard and fuck with shit.
I'm not seeing much to give me hope; I'd have thought someone would have figured this out by now.
Unless you're running a distro that has replaced X.org with Wayland or something.
Edit: um, never mind. Five minutes of research and everything is unbelievably awful. Have you looked at xpra? It miiiight do what you want? But ugh. I had no idea it was this bad.
Oh, there are plenty of ways to grab and forward the existing desktop; I tried both x11vnc and xrdp today.
The problem is that they just "mirror" what's on the local X instance, so anyone that walks the local monitor can watch whatever I'm doing, or even grab the keyboard and mouse and start "fighting" me to do stuff. When you use Remote Desktop to access a Windows box, the local screen locks and only displays a note that someone is logged on remotely.
From a security perspective, the latter is not just a good idea, it's absolutely necessary, and I can't understand how no one in the Linux world has figured this out yet.
However, as an experiment, in LightDM I was able to "user switch" to the login screen and still access the locked session using x11vnc (despite the overlay LightDM adds even when using ctrl-alt-f7 to forcibly switch back to it). I guess it'd be useful if you have something long running, but any time I have something long running, it's command line and thus is better to put inside screen or tmux anyway.
Wait how did you accomplish this? Because it's been a constant concern of mine.
most desktop environments should have a switch user option, although admittedly xfce's stuff (which I use) doesn't know how to talk to lightdm (it expects gdm, apparently), so I used lightdm's dm-tool command to do it
Using it for work... I at least haven't noticed any major issues. As always using nvidia drivers on a laptop is a terrible idea. Other than that... pretty frequent updates, nothing broken, Wayland is iffy but it's an option at least.
I run Linux Mint because it's easy to set up and stable. That said, I like running the latest version of the AMP part of the LAMP stack in order to keep up with bug fixes and get new features. What I'm wondering is:
Is there an actual reason why the AMP packages in repositories lag so far behind? I managed to get the latest Ubuntu packages for it, but the latest version of PHP it offers is 5.5.9, which came out nearly a year ago. I've tried dotdeb and Ondrej Sury's PPA, but both are bugged for me (which may be the answer to my question).
Why can't the software vendors have their own repositories one can tap into? Like, say you wanted the latest version of Apache. Point your package manager to their bleeding edge repo. Or you wanted the latest stable version of PHP. Point your package manager to their 5.6-stable repo. Or 5.6-stable-debian, or whatever.
I guess I just find it hard to believe that different distros are so different from one another that they require so much testing before release. I know I can compile from source, but that's a PITA that I'm unwilling to do.
I dunno... I come from Windows land where new versions of C# and whatnot either require purchase or an in place update, but the actual installation/update process is fairly mindless. I'm trying to figure out why it can't be that way on linux as well.
Because no vender wants to have to provide binaries for Ubuntu, Mint, Debian, RHEL, Centos, Slackware, Gentoo (you get the point).
I'm not a big Mint/Ubuntu guy, but you might look for either a PPA to grab them from (if you trust the compiler), or grabbing from SVN if you need bleeding edge.
See, I don't get the appeal of that kind of fragmentation. Not in a "I don't like it" way (well... not entirely), but in a "Why the fuck would anyone want to deal with that kind of 'everything is linux, except when it's our flavor of linux" bullshit? What's the advantage? What's the appeal? Why isn't it more reasonable to have a single linux kernel, and build off that? Where productivity software (like programming, web server, and database software) simply targets that so it's universal across distros?
I'm not trying to start an argument or anything. I'm just trying to understand. Everyone that loves linux swears it's so much better than the more mainstream OSes, but I just don't see why, beyond the fact that they're not tied to for-profit entities.
There has to be a reason why that's considered better than what OSX or Windows does.
Well, because it's Open Source. One person/group bundles up a bunch of software, everything they need for an OS. They include all their favorites. They share that with the world. The next guy decides they don't like package X or Y, so they change some things to suit their needs and share it. And so on.
Debian wanted to conform to the GNU Principles, and values stability over features.
Ubuntu wanted the base and package management of Debian, but software that was more current and relevant to desktop users.
Red Hat wanted a Linux that would be suitable for the Enterprise, something they could sell and charge to support.
CentOS wanted to be Red Hat, but without the price tag.
Gentoo mimics the BSD ports system and places an emphasis on compiling everything so it's optimized for your system.
Arch just wants to be bleeding-edge.
That's the point, though. Linux is, in fact, a kernel. A Distribution is what happens when you wrap that kernel with all the other software you need to actually have a functional computer.
You want software that runs for years without having to update any config files or do major version upgrades? This lets you have it, without any extra work for the vendor. You want the latest and greatest of everything? Source-centric software distribution lets you do that, too. Everyone is happy, except for people used to binary distribution who see this huge confusing mess of slightly incompatible systems. ;-)
You need to discover the joys of Docker.
What makes it magic is the VM system lets you just do "docker pull php/latest" or "docker pull php/5.2" and it'll download a prebuilt image containing exactly that for you built by the project maintainers (there's also lots of custom stuff - or you can roll your own).
EDIT: For example, I built a test version of one of my company's intranet sites by dumping the database onto a docker image, the website onto an apache one, and using the link commands (which automatically patches the hosts files) to recreate it in containers.
Then added a small script which booted all the daemons in foreground mode in separate tmux windows. The whole thing can be setup and torn down in seconds.
Interesting. So, it's possible to have a dock(ed/ered?) version of a linux distro with the full LAMP stack?
Yep. You can grab stuff people have made up like here: https://github.com/ushios/docker-lamp-ubuntu and change the docker file how you want. Or write your own - it's just a sequence of apt-get commands to set things up.
But another way to do it is just to pull down the individual application versions in their own containers, and link them together over the virtual network. This is actually the approach I'm starting to favor because it means I can have each apps console in the foreground while I work. So for instance I have this script for starting up the dev environment of a project I'm working on - it boots each container and attaches it to tmux, so I get the consoles all on one panel between the webserver and database:
#!/bin/bash tmux new-session -d -s project-dev -n project-dev \ 'docker start -i running-projectdb' # Wait for running-projectdb to be running while [ -z "$(docker ps | grep running-projectdb)" ]; do sleep 0; done tmux split-window -t project-dev:0 -v \ "docker run -it --rm=true -p 127.0.0.1:3000:3000 \ -p 127.0.0.1:6868:6868 \ -v $(readlink -f .):/opt/project \ --name running-projectnode \ --link running-project:psql-int \ projectnode"This is a really neat way to do it, because it means you can go from docker to deploy without needing to change anything provided the hostname's for are the same (which was why I did it originally - because it was an existing project).
The other thing is that you can use the volume commands to bind files from your local drive into the container system - that's what I'm doing there with the -v switch. All the regular stuff like inotify and the like still works from host to guest too. So you can live edit stuff directly, but pull it into one (or several different) containers.
I'm assuming they can save state because they're really just fancy VMs.
Yes and no - in the sense that it's optional. Docker containers save their state till you kill them (which is different to start/stop). That persists till you destroy the container.
But you can also define 'volumes' for specific folders which docker will persist when the same volume is re-attached to another container (i.e. they have their own name, it's kind of like having a virtual NAS).
Or you can just bind-mount the bits you want to folders on your host.
Or you can push the content to an image which then lets you duplicate it to many containers.
So, would I be able to do the following in Docker:
Define a LAMP development environment template, including IDE?
Work on a specific instance of that template, save it as a snapshot when I'm done working for the day, then reload where I left off?
I'm looking for more than a deployment environment. It'd be awesome if I could spin up an actual full featured work environment on the fly quickly.
EDIT: You may have answered my questions with your last statement about saving an image
But I'd probably just go with running my IDE on my normal OS, and using Docker to run the server side of things (since that tends to be the most intrusive normally).
main_folder/2314/Album_name/blabla.wav main_folder/2314/Album_name/blabla2.wav main_folder/2714/Album_name/blabla.wav main_folder/2414/Album_name/blabla4.wavand.. eh.. there's about 100.000 of them. So I need to move all the .wav-files up one directory so that they're like this:
main_folder/2314/blabla.wav main_folder/2314/blabla2.wav main_folder/2714/blabla.wav main_folder/2414/blabla4.wavNaturally, I'm thinking Bash is the easiest and cleanest solution for this, so I fire up Cygwin and catch myself doing this:
find . -name *.wav* -exec mv "{}" ../ \;So that would move all the files to the upper directory of where I execute the script, but I need it to move it to the upper directory of each file. How do I work around this?
I dunno what the right way to do it with bash is, but here's the horrible way I've solved similar problems in the past. Now you have a list of the mp3s. Next, you open it in Vim and make a macro. When the dust has settled, the first line looks like Now you run that macro on every line of the file. Now you make the file executable...
(I'm really bad at bash)
find FOLDER -name '*.wav' -execdir mv {} .. \;TL;DR: Man files are pretty great.
Edit: Looking over this list https://steamdb.info/linux/ I think the only games I'll be missing that I still want to play are the lego games, saints row series, and walking dead. Maybe a few others.
This'll be my first non-windows box since my old mac running OS8. I think I'll take the plunge and not setup a windows dual boot right away.
I personally prefer Fedora, but Steam gaming isn't as smooth as on Ubuntu; some less experienced developers do boneheaded things like dynamically linking to libraries and using Debian-specific library names. (Games ported by Valve or published in Humble Bundles are fine.) It's doable, just occasionally more tedious.
I also don't have any mp3s and just stream all music from pandora and other services, so it's not too big of a worry for me and I don't think any self-respecting dev would use mp3s for their sound files anyway.
So I try using updatedb to make a database and then using locate to copy the files instead, but the command crashes.. out of memory, as far as I can tell. Huh. Looks like this
Alrighty, then. That's weird, well, whatever.. I'll just dump the directories to a text file and use a script with grep to copy then, should still be faster then using FIND right now. So..
Well, fuck. What do I do, I'm (again) looking at literally hours and hours of manual copy/paste work here.
Anyway, I jumped through a ton of hoops to export a file list of what I needed, so I'm getting stuff done. Fingers crossed.
Of course I immediately added more RAM (because I am now, apparently, the sort of person who has spare DDR2 laptop RAM sticks just... lying around) and swapped out the hard drive for something bigger. Hooray for voiding warranties!
Went ahead and installed Kubuntu 14.04, since Gnome was driving me up a wall. KDE drove me up even MORE of a wall until I realized just how much I could customize. (Well, I STILL don't know how to make the scroll buttons move the view faster...) Consider me pleased and generally happy.
I like how, after a logoff or a restart, it is happy to restart my applications right where I left them. It does seem to take a bit longer than I'd like to go from "the desktop has appeared" to "you can actually do stuff". Also Kmail is pretty darn nice and the fact there is an IRC client included with the Kubuntu install is a good idea. Apparently installing Steam on this is gonna be a bit of a pain...
Overall I'd call my experience so far a sold "B." Certainly better than vanilla Ubuntu.
And getting Steam running on Ubuntu was pretty easy last I tried.