IRC Logs for #crux Tuesday, 2014-04-22

*** jdolan has joined #crux00:20
*** toriso has quit IRC00:38
*** arjovr has quit IRC01:50
*** arjovr has joined #crux01:51
*** tilman has quit IRC02:03
*** tilman has joined #crux02:05
*** BitPuffin has joined #crux02:26
openfbtdBy the way, about inits02:31's sinit seems to be something that would be very in line with CRUX' ideology02:31
cruxbot[contrib.git/3.0]: docker: 0.9.1 -> 0.10.002:31
prologicgot a link?02:32
openfbtdIt's even simpler than runit02:32
openfbtdIt's an init with three functions: be an init, poweroff and reboot02:34
prologicweb page is a pit sparse02:34
prologicif that's all it does02:34
prologicthat's practically all sysvinit does too02:34
BitPuffinwat they have an init system?02:34
openfbtdsysvinit is much more complex.02:34
prologicwell not really02:35
openfbtdBitPuffin, if you can call that a system02:35
prologicthere's also minit too I believe02:35
prologicthat buildroot uses02:35
openfbtdYes, really. Look at inittab and how complex that can become02:35
openfbtdsysvinit has a lot of functions compared to this one.02:35
prologicgive it a go on your system02:36
prologicand tell us what it's like :)02:36
openfbtdAlready have02:36
BitPuffin[core.git/3.1]: init -> switch to sinit02:36
openfbtdMy home desktop is currently booted with sinit02:36
openfbtdHad to push some of the functionality my own init has to my rc script02:36
openfbtdBut otherwise smooth sailing02:37
BitPuffinwell the less the init system does02:37
BitPuffinthe better kinda02:37
openfbtdMy reasons to switch exactly.02:37
openfbtdAnd my init is a bash script, so02:37
openfbtdThat's not ideal02:37
openfbtdBut I like sinit better02:38
openfbtdIt's dumb as bricks, so it pushes all the actual functionality to other entities02:38
openfbtdSo the init in my system is now truly dumb.02:39
openfbtdThe rc script does all the configuration and the service manager handles services02:39
prologicso you have sinit runnint on crux?02:39
prologicwhat rc changes were required if any?02:39
openfbtdNope, arch02:39
prologicyou suck :)02:39
openfbtdMy rc script is this:
openfbtdI've pushed some functionality from spark's example scripts to the rc02:40
openfbtdBut haven't pushed them to the repo yet02:40
openfbtdI'm considering making my own init dumber as a result02:40
BitPuffinhmm well02:41
BitPuffincrux rc.conf is nicer :D02:41
openfbtdThat is open to interpretation.02:41
openfbtdI personally like my rc.conf like it is.02:42
BitPuffinwell just that you do '' around the services02:42
openfbtdThat's just a matter of stylr02:43
BitPuffinI'm not sure using alternative initsystems are nice though, I didn't even consider changing the init system on arch. But even if I did it would suck with the package system because they package things with systemd scripts02:43
openfbtdI like my arrays quoted, it's not mandratory02:43
prologicstop talking and port sinit to crux for us02:43
BitPuffinyeah I guess02:43
prologiccomplete with a ported rc02:43
BitPuffinI've always wondered02:43
BitPuffinwhat does rc stand for02:43
openfbtdruntime configuration02:44
BitPuffinit all makes so much sense now02:44
openfbtdprologic, sinit is dumb as bricks, as I've said, writing a port would be quite easy. An rc... I've used my own set of tools on CRUX before02:45
*** _92AAAQLBQ has quit IRC02:45
BitPuffinopenfbtd: so you are saying you used systemd before it was cool? :P02:45
openfbtdThe only problem I've encountered is changing the service init scripts to conform to the /{bin,sbin} /usr/{bin,sbin} separation02:45
openfbtdBut my service init scripts are scripts for my own service manager02:46
openfbtdSo that's another tale02:46
*** _92AAAQLBQ has joined #crux02:46
BitPuffinwell arch doesn't even have sbin does it?02:46
openfbtdBitPuffin, I've tried using systemd when Arch switched to it, didn't like the way it does things, wrote my own set of tools to boot and mantain a system02:46
BitPuffinanymore at least02:46
openfbtdArch has it's system PATH merged into /usr/bin02:47
BitPuffinopenfbtd: then why are you still using arch xD02:47
openfbtdEverything else is symlinks02:47
openfbtdBecause otherwise I'm pretty comfortable with the distro02:47
BitPuffinwell pacman is pretty nice02:48
BitPuffinbut they have packaged some things really dumb02:48
openfbtdAUR also.02:48
BitPuffinwell crux is basically aur only :P02:48
prologicI think you miss my point :)02:48
prologicno-one else is going to try sinit02:48
openfbtdYou will never have ideally packaged packages02:48
prologicunless you create a port for it02:48
prologicand replacement rc02:48
prologicbecause I for one don't personally see the gain02:48
prologic"if it works don't fix it"02:49
BitPuffinopenfbtd: I know but on crux it would make sense to make my own port02:49
prologicfor instnace02:49
openfbtdmeh, I'm not ready to commit to mantaining ports for CRUX unless I use CRUX02:49
BitPuffinopenfbtd: making my own PKGBUILD for qt 4 doesn't really make sense02:49
prologicwhy would I waste a whole afternoon on a simpler init02:49
prologicthat gains me nothing?02:49
openfbtdYou probably wouldn't?02:49
prologicwell of course :)02:49
prologicit's a nice idea and all02:49
openfbtdI was just saying that it could be a nice addition to the next release maybe or something.02:49
openfbtdJust an idle thought02:50
prologicbut it requires someone to put the effort in02:50
prologicright now I'm putting my effort into docker+crux02:50
BitPuffinI guess the solution to system v is to replace it with something simpler, rather than something infinetely more complex (systemd)02:50
BitPuffinprologic: ah cool, I'm gonna run docker + crux on my server02:51
openfbtdYes. Maybe someone would look into it, try it on their system and write the necessary ports. I'm not a CRUX user currently, so I wouldn't do a good job of mantaining a port02:51
openfbtdor several02:51
BitPuffinso that I can isolate stuff02:51
openfbtdAlso the thing with CRUX and the port repos all over the place is pretty horrible02:51
BitPuffinhas anyone written a tool for managing that?02:52
prologicbbs - rebooting02:52
BitPuffinotherwise I might02:52
openfbtdBitPuffin, you've mentioned CRUX's port system being basically AUR — no. AUR is one place, a huge number of custom unmanaged unsigned repos from different people is a mess02:52
BitPuffinopenfbtd: I don't mind having multiple port repos02:52
openfbtdTo be fair, AUR is also unsigned.02:53
BitPuffinIt's just fetching the file that is pretty annoying02:53
BitPuffinand searching on the web interface02:53
BitPuffinopenfbtd: well yeah, crux ports is distributed aur02:53
BitPuffinopenfbtd: I can say that I trust this user, rather than I trust all of AUR02:53
BitPuffinmuch better02:53
BitPuffinstill think we should have signed pkgfiles though02:54
openfbtdDepends. I was still reading all the ports when I was using CRUX.02:54
BitPuffinopenfbtd: well you should do that on arch as well02:54
openfbtdI do read PKGBUILDS from AUR, yes02:54
openfbtdBut most of the packages are managed and signed02:54
openfbtdAUR is a last resort type of thing for an arch user02:54
BitPuffinplus, the PKGBUILDs in arch are messy02:55
BitPuffinthey used to be pretty much identical to crux02:55
openfbtdYeah, I think they went overboard with the feature set02:55
BitPuffinbut then they goofed up02:55
openfbtdYou can still write very simple PKGBUILDS02:56
openfbtdBut they support so much stuff that can be done in them02:56
openfbtdStill not deb-srcs though :D02:56
BitPuffinyeah well nobody really writes simple ones though haha02:57
prologicI run docker+crux servers02:58
prologicwell infrastructure that is02:58
prologicI'm a big fan of docker02:58
prologicand crux :)02:58
prologicso yeah02:58
BitPuffinopenfbtd: isn't that your repo though?02:59
openfbtdThat PKGBUILD is also in AUR03:00
BitPuffinso it doesn't count :P03:00
openfbtdI just have a git repo for my PKGBUILDs03:00
openfbtdDunno why.03:00
BitPuffintell me last time you saw a nice pkgbuild that you didn't write yourself03:01
BitPuffinwell that's by someone who likes suckless philosophy03:02
BitPuffinprologic: for what? :o03:04
prologicsorry what?03:05
BitPuffinprologic: what do you run the infrastructure for? :)03:05
prologicmostly all of my own stuff03:05
prologicwebsites, irc bots, services03:06
prologicname servers03:06
prologicI'm also using Docker for work more and more03:06
prologicbut it's slow going - red tape and bureaucracy et all03:06
prologicsearch for prologic03:06
prologicI have published quite a few images already03:06
prologicand I maintain the crux stadnard library image03:07
prologici.e: docker run -i -t crux /bin/bash03:07
*** jdolan has quit IRC03:07
*** jdolan has joined #crux03:07
BitPuffinprologic: and here I thought I was gonna have to create my own image :)03:09
BitPuffinwhat's the difference between crux and docker-crux03:10
*** jdolan has quit IRC03:11
prologicno no03:13
prologicdon't be silly :)03:13
prologici've done all the hard work03:13
prologiccrux is a fully supported Docker platform03:13
prologicand CRUX in Docker is well already there03:13
prologicas well as CRUX in Docker in CRUX03:13
prologicor the other way around03:13
prologicdocker-crux is just the name of my repo where I maintain the image03:14
prologicit's published to the Docker image std. lib as just "crux"03:14
prologiclike docker run -i -t centos /bin/bash03:14
BitPuffinprologic: question is though, is it really necessary to have a full crux installation if the host is already crux?03:19
jaegerwe have to crux deeper!03:19
prologicBitPuffin, no it isn't03:19
prologicbut that's not really the point now is it? :)03:19
BitPuffinnope :P03:20
prologicimages all share a base image nayway03:20
prologicso if you use crux as your base imag for your apps03:20
prologicyou won't be wasting disk space03:20
prologicbecause of the union file system that docker uses by default03:20
BitPuffinprologic: well it would still be 2 cruxes rather than one lol03:20
prologicbtrfs, devmapper or aufs backends03:20
prologicI use btrfs here on my desktop03:20
BitPuffinpfff btrfs :P03:20
prologicyeah that's still not the point03:21
prologicyou're not running a vm03:21
prologicit's just a rootfs in a container03:21
BitPuffinwell I mean03:21
prologicrun up a few apps that are Dockerized03:21
prologicand you'll get the idea :)03:21
BitPuffincan I have the containers use stuff from the host as well?03:21
prologice.g: get docker up on your crux host03:21
prologicand try:03:21
BitPuffinI mean some things that I specifically allow03:21
prologicdocker run -d prologic/kdb03:21
BitPuffinI dunno what that would be03:21
BitPuffinbut yea03:21
openfbtdThere is a very good reason source-based distros with rolling packages aren't used in production03:21
prologicthat's called volumes03:21
prologic-v host_path:container_path03:22
openfbtdOr just distros with rolling packages03:22
prologicI disagree with that but anyway :)03:22
BitPuffinI disagree03:22
prologicyou're obviously not a developer :)03:22
prologicif you've ever developed in say Python03:22
openfbtdI'm an administrator, yes.03:22
prologicand end up compiling Python from source on CentOS anyway03:22
openfbtdThat's desktops/dev stands03:23
BitPuffincrux is half rolling03:23
prologicthen binary/non-rolling distros are a completely waste of time03:23
openfbtdUsing CRUX on all of your infrastructure would be ill-advised.03:23
openfbtdBut hey, it's your infrastructure03:23
BitPuffinI don't think it would be03:23
prologicI doubt you could say that to me with a striaght face03:24
prologicyou're "ill-advised" is quite subjective at best03:24
prologicbased on your experience03:24
BitPuffinI wanna have the same thing on my server as on my personal computer03:24
prologicseeing as you don't use crux in production yourself03:24
prologichow would you know? :)03:24
BitPuffinso for a while I ran arch on my server03:24
prologicjust say'n03:24
BitPuffinbut that turned out to be retarded because they don't package things very well03:24
openfbtdYes, because I've never seen source-based distros in production. Obviously.03:24
prologicclassifying crux as a source-based sitro is well naive ihmo03:25
jaegerI used to work at a place that used gentoo for nearly all its infrastructure. It was quite smooth, actually. You just have to set things up correctly.03:25
prologiceverything comes/started from some kind of source03:25
BitPuffinso if I wanted to run murmur for example, it depends on qt4 CORE! but in arch there is only qt4, rather than splitting it up to separate packages like it really is. That means that in order to run murmur I would have to install ALL of qt 4 which in turn pulls in x11 and a bunch of input stuff and I don't even know03:25
BitPuffinso much shit03:25
prologicthanks jaeger :)03:25
openfbtdjaeger, and how fast and well are park-wide updates?03:25
prologicI personally have been running CRUX on the Desktop for over 12 years now03:25
prologic_as well_ as on all my servers03:25
prologicbut yeah03:25
prologicwho cares03:26
BitPuffinyou can do the upates on one box03:26
BitPuffinand then send the binary pkgs around03:26
jaegerI built updates on a single machine and pushed the built binaries out to the rest, it was pretty quick aside from that single host's build time03:26
BitPuffinno biggie03:26
prologicand that's usually what you do03:26
prologicif all the machines have identical architetures03:26
BitPuffineven if you couldn't taking a few minutes extra is no biggie03:26
prologicwatch out for crux 3.103:27
openfbtdSeriously. You are explaining local binary repos to me.03:27
prologicI'm planning on building a service to build binary packages03:27
prologicand serve this up as a service03:27
prologicvia pkg-get (I think)03:27
prologicbecause I want it for my own stuff too :)03:27
BitPuffinsounds nice03:27
jaegerI don't really get the point of this argument, was just commenting. Will butt back out now. :)03:27
prologicthis debate is pointless03:27
BitPuffinopenfbtd: well you sound like you need it explained to you kinda03:27
BitPuffinif you don't think a source based distro can work on a server03:28
openfbtdMy concerns are more in the area of rolling-release packages and the instability that poses. And the point of needing a building and testing infrastructure.03:28
openfbtdI _know_ a source-based distro can work as a server OS03:28
prologicyou clearly are _not_ a developer03:28
openfbtdI just don't think it's worth the time.03:28
prologicyou think in sysadmin terms03:28
BitPuffinwell, you still need to test updates on a "stable" release as well03:28
prologicof what'e easier for you as a sysadmin03:28
prologicstability doesn't mean the same thing for you as it does for me03:29
BitPuffinI trust upstream03:29
prologicbuilding and testing is far more impoartnt for me than it is for you03:29
BitPuffinand err03:29
prologicbut *meh*03:29
BitPuffinI'm a developer03:29
prologicI don't :)03:29
BitPuffinwell I don't mean I blindly trust it, I will test it first03:29
prologica yum upadte03:29
prologiccould just as easily break my app03:29
prologicas a prt-get sysup03:29
openfbtdWhere I work, we (mostly) trust debian's upstream these days. Testing is still domne, but at least we don't need to spend half a day rebuilding a new release.03:29
BitPuffinbut generally they test things before they call it a release you know03:30
prologicyeah the point though is03:30
prologicthat half a day you spend rebuilding03:30
BitPuffinprologic: well I meant upstream as in the projects themselves, thought debian like stuff was called downstream03:30
prologicyou're only saving because Debian devs have done it for you03:30
prologicthere really is no difference here03:30
prologicDebian or CRUX03:30
openfbtdBut yeah, there are different usecases. That's why I'm saying that rolling distros work well on desktops and development stands.03:30
prologicsomeone has to build packages03:30
prologicor "something"03:30
openfbtdNot on, say, a hosting03:30
jaegerThis is still a subjective argument03:30
jaegermaybe you guys should agree to disagree03:31
prologicI'm backing out03:31
prologiclunch time :)03:31
BitPuffinyou still need the same shit that you have on the development machine on the server03:31
prologicsorry openfbtd03:31
prologicyou are of course absolutely right03:31
BitPuffinlike if the development machine has python 3.x on it03:31
openfbtdI don't even see this as an argument really.03:31
jaegerprologic: no need to be petulant03:31
BitPuffinyou won't have fun putting it on a RHEL server03:31
prologiclike you said jaeger03:31
prologicpointless debate ;)03:31
prologicI think I said that too03:31
prologicit doesn't matter really03:31
openfbtdNow. Go to sleep or work on my init and rc.03:32
prologicI sometimes have these very debates with a sysadmin that travels on the bus to my workplace :)03:32
openfbtd(both of which I would never put on a production server for reasons)03:33
prologicwe're a RHEL enterprise at work03:33
prologicwhatever that means :)03:33
BitPuffinthe usecase for RHEL-like distros is basically companies with big tech debt03:33
openfbtdThere's also a question of security btw.03:33
BitPuffinie they run the same thing that they have been running for years and years03:34
openfbtdAll debates about anything are null and void till CRUX has signed repos03:34
openfbtdI will never allow my infrastructure to have anything that doesn't have basic security measures03:34
BitPuffinyeah absolutely, I mean malicious debian developer could modify the sources before building + signing it, and since you can't verify that the source is the same as the built package the security is worse03:34
BitPuffinso there is a big concern with a binary based distro03:35
openfbtdMy own personal infrastrucure can work because I can check thing for myself03:35
openfbtdOn a larger scale — NO NO NO NO NO03:35
openfbtdthings* infrastructure*03:35
openfbtdAnd even then my main servers are on debian.03:36
openfbtdI clench my teeth and deal with deb-srcs :<03:36
prologicI don't see how signing is any better than ssh keys tbh03:36
BitPuffinwell I think we should have signed pkgfiles but that's another story03:36
prologicI don't03:37
prologicthat's more work for no gain03:37
BitPuffinI still think it's way more secure to have unsigned pkgfiles and build from source than it is to blindly accept a signed binary from some dude03:37
openfbtdFrom some dude — maybe. But debian mantainers have enough good will to accept binaries from them.03:38
BitPuffinprologic: there is gain, someone who compromises a server that hosts a ports repo could modify the build script. With signing they would also have to compromise the key location of the port author03:38
prologicyou're talking about two things there03:38
prologicsomeone can't compromise the ports03:39
prologicbecause they are housed in git03:39
prologicyou'd see the new commits03:39
prologicand strip them03:39
prologicI do however see the point of signing binary packages03:39
prologicthe resulting build03:39
BitPuffinprologic: I'm not talking about core/xorg/opt etc03:39
prologicand keeping an equivilent of an .md5sum03:39
BitPuffinprologic: I'm talking about jaegers ports for example03:39
prologicbut for the "built" package03:39
prologicyes ok03:40
prologicbut that's 3rd-party03:40
BitPuffinand I want to say03:40
BitPuffinI trust this third party03:40
prologicI don't think we should impose such things on 3rd-aprties03:40
BitPuffinbut I don't trust that the third party can get compromised03:40
prologicyou should trust this guy03:40
prologicbecause you do03:40
prologicwell you can03:40
BitPuffinwell there is a difference03:40
prologicif you can trust my pports for example03:40
prologicI host them in a dvcs03:40
prologicyou can verify them against my repo03:41
prologicif you really want to03:41
openfbtdI can trust you, but I can't trust your system03:41
BitPuffinI can't say I trust his pkgfiles without saying I trust his server03:41
BitPuffinand that's the problem03:41
openfbtdEvery system is potentially vulnerable.03:41
prologicyou two are far too paranoid03:41
prologicif you can't trust a simple Pkgfile from me03:41
prologicthat's hosted in my bitbucket repo03:41
prologicof which I only have access to03:41
BitPuffinprologic: yeah but that's too manual, I think there should be a signature or something and it should be automated in prt-get03:41
prologicthen who can you trust?03:41
BitPuffinwhat does heartbleed have to do with anything we are talking about03:42
openfbtdI'll just hand a huge “heartbleed” banner over the conversation about paranoia03:42
prologicas soon as you make it automated03:42
prologicyou're moving thr trust to a system03:42
prologicand if that system is comprised03:42
prologicso is the trust03:42
jaegerhere's a thought: if you really want something like this, implement it and see how it goes03:42
openfbtdBitPuffin, my point is you can't be too careful03:42
BitPuffinprologic: well not really03:42
prologicI'm with jaeger on this one03:42
prologicdesign it03:43
prologicimplement it03:43
prologicuse it03:43
prologicand share it with us03:43
BitPuffinprologic: sure if my prt-get build is not authentic then maybe03:43
BitPuffinyeah I agree as well03:43
BitPuffinI might do it eventually03:43
BitPuffindon't give me that attitude :D03:43
prologicI think I can safely say that all crux users03:43
prologichave almost never had an issue with trust03:43
prologicor security when it comes to isntalling/updating software03:43
prologicbut hey03:43
BitPuffinI'm just arguing that it is a good idea03:44
prologicthat's because we either check things ourselves03:44
prologicor we trust one another03:44
BitPuffinand this would be a better way to say I trust you03:44
prologicbut perhaps that's also why many of us that do have 3rd-party repos and ports03:44
BitPuffinbut not your server03:44
BitPuffinwhich is good03:44
prologichave lots of duplicates03:44
BitPuffinthat way we probably wouldn't even have to check the pkgfiles03:44
prologicbecause we _don't_ trust the other person03:44
prologicso we clone the their port03:44
BitPuffinin fact I do trust crux users03:44
prologicedit it03:44
prologiccheck it03:44
prologicthen use it03:44
BitPuffinand usually don't check03:44
BitPuffinso don't call me paranoid03:44
prologicI know I'm stupid enough to do this03:44
openfbtdprologic, for two years people trusted a vulnerable library. Evrything is potentially vulnerable.03:45
openfbtdNever trust a single system completely.03:45
BitPuffina pkgfile is highly unlikely to be "vulnerable"03:45
openfbtdThat said, additional security layers help.03:45
prologicand with additional layers03:45
prologiccomes additional complexity03:45
prologicand effort and time03:45
openfbtdBitPuffin, and asking users to read every single pkgfile is silly03:45
prologicyou have to weight it up03:46
prologicbut by all means03:46
prologicbuild us something03:46
prologictry it out and let us know03:46
BitPuffinit's not entirely unlikely that it could be compromised.03:46
openfbtdIt's a way to do it, don't get me wrong. But it's silly.03:46
BitPuffinhowever crux is a pretty small target03:46
BitPuffinso I don't really worry03:46
prologicperhaps it could be as simple as03:46
prologicpkgsign Pkgfile03:46
prologicif you make it hard people won't do it03:46
BitPuffinopenfbtd: yeah and that's why signing the pkgfiles would put that pressure off if you trust that the user writes good pkgfiles03:47
openfbtdprologic, which is why I'm pretty sure people don't read updated pkgfiles03:47
openfbtdOr at least most of them don't.03:47
BitPuffinprologic: well then prt-get could go "Hey dawg this pkgfile is unsigned, would you like to check the pkgfile?"03:47
prologicyeah no I do03:47
openfbtdAdd that to the build times and you end up working for your distro, not the other way around.03:48
prologicI'm gonna go have lunch03:48
BitPuffinit could even go ahead and build in the background03:48
BitPuffinfor performance03:48
prologicthis is still all pointless without something to design, implement, test and use03:48
BitPuffinjust doesn't have to install03:48
BitPuffinWell no03:48
BitPuffinit's good to discuss03:48
BitPuffinbefore blindly going balls in03:48
openfbtdDiscussion is also healthy. I would have done so many things wrong in my own scrips if I wasn't discussing changes with my friends and colleagues03:49
BitPuffinit's like making a baby before discussing it with your partner03:49
BitPuffinnot a good idea03:49
openfbtdLike the time I've expanded my init's functionality03:50
openfbtdAnd right now I'm looking at the piece of code that is a watchdog function03:50
openfbtdAnd feel like I'm slowly becoming Lennart03:50
BitPuffinnobody wants to become Lennart03:51
openfbtdSo I'm thinking how should I approach the issue.03:51
BitPuffinIt's simple03:51
BitPuffinevery time I am faced with a decision I just ask myself "What would Lennart Pottering not do?"03:52
openfbtdRemove the function, push agetty spawning to the rc or even the service manager03:52
openfbtdI've already basically done that actually03:52
openfbtdBut now I need to properly write it, not hack it in03:52
openfbtdSo obviously I need to write my own agetty03:53
openfbtdAnd merge it into my init03:53
openfbtdBASH all the things!03:53
BitPuffinI wish there was a simple and beautiful bash we could use instead03:54
prologicAnyone got 32bit chrome port?03:56
BitPuffinfor what?03:56
prologicmy media box runs crux 2.7 i68603:57
BitPuffinprologic: why not just build crux 3.0 from source then03:58
BitPuffinor wait for 3.103:58
BitPuffinin the name of being a badass03:58
*** himynameisphil has joined #crux03:58
prologicbecause upgrading from 2.7 to 3.0 or 3.1 is all too hard04:01
prologicit's easier to rebuild04:01
prologicplus my media box's hw is a bit broken04:02
prologicit's bios hard freezes sometimes (once per month os so)04:02
prologicso I'd realyl like to rebuild the whole thing :)04:02
BitPuffindo it04:03
prologicno time atm :)04:06
prologicit's one of those things04:06
prologicwe have scheduling that menas there is very little windows to rebuild it04:06
prologiceventually I'll rebuild it with a new one whilst the old one is still runing04:06
prologicthen swap them over04:06
prologicall sorted04:22
prologicfor now04:22
prologicvery old version of chrome04:22
prologicbut oh well04:22
BitPuffinprologic: why do you have chrome on your media box anyway :P04:22
prologicwhy not?04:33
prologicit's used for Youtube, Hulu, ABC iView, Ten Play and SBS on Demand04:34
BitPuffinprologic: xbmc04:35
prologicxbmc doesn't support tv tuners04:36
prologicin fact it has to rely on a backend like mythtv anyway04:36
prologicso I don't see the point in all it's fanincess :)04:36
BitPuffinstill nicer to have it all in one place with a media box :P04:36
prologicI will however (one day soon) be replacing my frontend with a MiniX X5 mini android tv box04:36
prologicbut I will still use MythTV04:36
BitPuffinI haven't sorted my media needs yet (as I haven't moved out yet) but will probably do so by having a NAS, and have an xbmc cubox or something get the media from there04:37
BitPuffinand then intall the youtube apps etc04:38
BitPuffinshould suffice04:38
BitPuffinwould be cool to have it all running on crux04:38
BitPuffinbut perhaps freenas is better for the NAS. I dunno. Still though, crux all the things :P04:38
prologicyeah I'm effectivelly going to do the same thing04:38
prologicbut with Android04:38
prologicXBMC just doesn't cut it for me04:38
BitPuffineven gonna attempt making a crux router04:38
prologicfor several reasons04:38
BitPuffinprologic: well generally with xbmc there is a way04:39
BitPuffinfor example04:39
BitPuffinif you want chromecast like functionality04:39
prologicbut some of it I feel is quite hacky :)04:41
prologicso I'm going down the Android TV box route04:41
prologicI already have several Android devices and quite like them04:41
BitPuffinthat's fine04:41
BitPuffinbut you will always have this voice in the back of your head04:42
prologicbut in the end I'll also Dockerize my MythTV backend setup04:42
BitPuffinreminding you that the box isn't running crux04:42
prologicinto separate contains04:42
prologicI have a little 22RU rack cabinet int he server room04:42
prologiccrux isn't the end all and be all04:42
prologicbut I do use it exclusively on desktopa and servers04:43
BitPuffinI will probably never run it on my phone04:43
prologicand I may put it on an ASUS Transformer Book TRIO soon too04:43
BitPuffinor will I04:43
prologicplanning on buying one04:43
prologicthere just isn't good enough mobile phone software for Linux(es)04:43
BitPuffinunfortunately yeah04:43
prologicunless you rip off OpenMoko software04:44
prologicor similar04:44
prologicit'll come04:44
prologicbut it's not here yet AFAIK04:44
prologicwhen it does come04:44
prologicyou could in tehory run up CRUX on your phone04:44
BitPuffinthen hell yes04:44
prologicand install the "right" phone ports04:44
BitPuffinpretty nice04:44
BitPuffinwell tbh04:44
prologicthe cloest thing to this right now04:44
prologicis unlocked/rooted Android04:44
prologicwhich I have in tablet and phone04:44
BitPuffinit wouldn't be that hard to write a simple interface04:44
prologicNexus 5 and Nexus 704:44
BitPuffinI also have rooted android04:44
BitPuffina barely breathing htc desire hd04:45
prologicnot hard to write a simple interface04:45
prologicTouch UI isn't new to LInux(es)04:45
BitPuffinbut stuff like calling04:45
BitPuffinand texting04:45
BitPuffinand so on04:45
prologichw/drivers for GSM shipets04:45
prologicthat's the missing bit I think04:45
prologica lot of this crap is proprietary04:45
BitPuffindoes CM contain a lot of binary blobs as well?04:45
prologicwhen the general community and industry at large oces this to be more open04:45
prologicyou'll start to see more opprounities04:45
BitPuffinIt's sad how locked down the mobile industry is :/04:46
BitPuffinmy friend was dual booting ubuntu and cyanogenmod on his samsung though04:46
BitPuffinor was it debian04:47
BitPuffin(not ubuntu touch, like linux ubuntu)04:47
BitPuffin(and yes ubuntu touch is linux, but you know, desktop, think it was lubuntu)04:47
*** xvee has quit IRC04:52
*** jdolan has joined #crux04:58
*** darfo has quit IRC05:09
openfbtdYay. My init is completely compatable with sinit now05:20
*** darfo has joined #crux05:23
cruxbot[contrib.git/3.0]: c-ares: (NEW) Asynchronous C DNS library (required by mosquitto)05:24
cruxbot[contrib.git/3.0]: mosquitto (NEW): MQTT compatible server and client libraries05:24
BitPuffinwhat the hell is MQTT05:25
BitPuffinMQTT is a machine-to-machine (M2M)/"Internet of Things" connectivity protocol.05:25
BitPuffinwell dayum05:25
prologicsomething I'm maintaining05:26
prologicand also quite a fan of05:26
prologicI'm presently also building a Docker image for it05:26
prologicso you can: docker run -d -p 1883:1883 prologic/mosquitto05:26
*** jdolan has quit IRC05:45
*** heroux has quit IRC06:19
*** heroux has joined #crux06:20
frinnstjaeger: heads up:
BitPuffinfrinnst: security related?06:42
diverseBitPuffin: open the link, it says there is a problem with an update to the ESXi system (which jaeger uses)06:45
*** jdolan has joined #crux06:46
BitPuffindiverse: I did open the link, but I'm getting pretty sleepy06:47
BitPuffinso I'm derping06:47
diversethen sleep, don't fight it06:47
BitPuffinI'm at work06:47
BitPuffinI'm gonna go home in ~3 hours06:47
BitPuffinso it's aight06:47
diversego home early?06:49
diverseit would be a safety hazard if you drive home very tired06:50
*** jdolan has quit IRC06:51
cruxbot[contrib.git/3.0]: mosquitt: 1.3.1-1 -> 1.3.1-2 -- Forgot deps: cmake c-ares06:52
BitPuffindiverse: nah I'm trying to get in to a certain sleep schedule06:54
BitPuffinand I'm not driving06:54
BitPuffinI'm biking xD06:54
diversebest way to do that is chug some caffine pills or drink lots of energy drinks06:55
BitPuffinmight take a cup of coffe06:55
diversethat works too06:55
BitPuffincaffeine is very dangerous though06:55
BitPuffinplus it makes you tired in the long run06:55
diverseit's only dangerious if you take it in high amounts all of the time06:56
BitPuffinor well, it can probably still be dangerous to take it consistently06:56
BitPuffinat least coffee as some healthy aspects06:57
diversewell a little won't kill you, there are some healthy aspects to caffine06:57
diversejust don't overdose on it06:57
BitPuffinhaha yeah like I said :)06:57
diverseand like I said before you said06:57
BitPuffinyou did? hmm, must have been pretty delayed06:58
diverseit's alright, you are sleepy06:58
BitPuffinwell I mean that in my logs I said it first06:58
BitPuffinwell whatever :)06:59
BitPuffinwhat's going on today with you diverse06:59
diverseWell, I'm trying to think of a different WM to move to07:00
BitPuffinthe grass is always greener07:01
BitPuffindiverse: so you aren't using 2bwm?07:01
diverseI'm currently using pekwm07:01
diversehey z3bra07:02
BitPuffinlooks kinda strange07:02
BitPuffino/ z3bra07:02
diversewhat's strange?07:02
diversein what way?07:03
diverseyou mean the default theme?07:03
BitPuffinhmm, one image it looked like it was decorated regularly, and another screenshot looked like haiku/beos07:03
diversewell it uses pixmaps (like fluxbox) so you can literally make the decorations how you want it07:03
z3braquick question07:03
z3brahow do you find orpahn packages ?07:04
diversebe it old-school or new-school07:04
diversez3bra: prt-get listorphans07:04
z3branow that's a huge list xD07:04
BitPuffinis that when it gets installed as a dep and you remove what had it as a dependency?07:05
z3brapackages that are not a dependency of another package07:05
diverseit's leftover dependencies from other packages that don't require it or have been removed07:05
z3bra(but still installed)07:05
z3bradiverse, not really07:06
z3braeg: vim07:06
z3braor tmux07:06
frinnstprt-get listorphan07:06
frinnstlistorphans *07:06
z3brayeah diverse told me :)07:06
BitPuffinwell how does it keep track of if it was installed regularly or pulled in as a dep?07:06
z3bracheck the source :)07:07
diversez3bra: Well the idea is, listorphan list all the packages that have zero dependents07:07
frinnstbut be careful with that, it just goes by the "#Depends on" line in the pkgfile07:07
frinnstnot if something is actually linked07:07
diverseyou can remove that stuff without harming your system usually, but you have to be careful about what you remove07:07
diverseand always run revdep to make sure nothing broke07:07
z3braI tried installing cifs-utils07:08
z3braAnd it installed 2 dependencies07:08
z3braBut I closed the terminal and I want to know what's installed07:08
frinnstDependencies: samba,keyutils07:09
BitPuffinwell then it will say that a package with no deps is an orphan07:09
BitPuffinthat nothing depends on07:09
z3brayeah there was samba07:09
z3brabut samba pulled other dependencies, etc...07:09
BitPuffinmaybe if it only listed those that were listed as a dependency in some pkgbuild somewhere07:10
BitPuffinthat would make it really slow though07:10
BitPuffinunless you put it in like sqlite07:10
diversea zero dependent package can be remove, although you may not want to remove them because you want to keep some of them, after removing zero dependent packages, you can then remove their dependencies just as long as they are dependent for other packages. It's really that easy.07:16
diverse*they are not dependent for other packages07:17
diversehow I wish IRC had the ability to edit out your previous messages for mistakes07:19
diverseyeah to what?07:20
diversethe orphan thing?07:20
BitPuffinno the mistake thing07:21
diverseAh, well, wouldn't that be cool though?07:21
BitPuffinwell the problem is07:22
BitPuffinyou can't just do prt-get remove `prt-get listorphans` or whatever because as you said, you might not want to remove them07:22
frinnstyeah, dont fucking run that command07:22
frinnst(in case someone is in the habbit of running commands straight off the internetz)07:23
BitPuffinhowever if there was a way to filter based on that one of the orphans is listed as a dependency in a pkgfile on your system, chances are it's not a utility but a library that was installed as a dependency07:23
frinnsthave you seen what it lists? :)07:23
BitPuffinalthough you should still be careful07:23
frinnstgcc, glibc-32 etc07:23
frinnstbut go ahead, great excuse to test your backups :)07:24
BitPuffinfrinnst: well a better alternative to prt-get remove `prt-get listorphans` is "sudo rm -rf bin usr"07:24
BitPuffinor actually07:24
diverseBitPuffin: it's a matter of knowing what you want. The computer can't know what you really want to remove.07:24
BitPuffin"dd if=/dev/null of=/dev/sda"07:24
BitPuffindiverse: yeah07:25
BitPuffindiverse: well I might write a small cleanuputil07:25
BitPuffinthat builds a database of all the packages and all the dependencies etc. And then will list only orphans that are a dependency of some port (regardless if that port is installed or not)07:26
BitPuffinand then like I dunno, ask you yes or no for each orphan or something07:26
BitPuffinjust in case07:26
Romsterprologic, c-ares needs --mandir and the chroot a the end is not needed.07:55
diverseRomster: you mean chown?07:56
Romsterah chown doh07:56
Romstermosquitto as well we don't use /usr/share/man we use /usr/man07:57
Romstermosquitto is also at least missing cmake and probably other depends on ports07:58
Romsteroh next commit does deps.07:58
Romsterbut still mandir07:58
*** jdolan has joined #crux08:08
*** jdolan has quit IRC08:12
prologicokay :)08:13
prologicI normally don't really care about where the man pages are08:13
prologicdoesn't man look in /usr/man and /usr/sahre/man?08:13
prologicDo we officially in CRUX (someone point me to some wiki doc) use /usr/man and not /usr/share/man? :)08:13
prologicI forget :)08:14
prologicI cannot find any docs on this08:16
prologic4.6.2. Directories08:17
prologic/usr/man/Man pages08:17
prologicthere it is :)08:17
Romsterit's in the handbook08:17
prologicshows you how long it's been since I've read the handbook08:17
diverseRomster is the Crux master08:18
prologicAlthough it's exact working is: In general packages should install files in these directories. Exceptions are of course allowed if there is a good reason. But try to follow the following directory structure as close as possible.08:18
prologicSo :)08:18
Romsteryes but most programs put it in /usr/share/man just add a --mandir=/usr/man and bump the relase number and pkgmk -um the footprint after rebuilding it'll be good then.08:18
prologicby convention however since I do recall we like --mandir=/usr/man I'll fix the ports ;)08:18
prologicdiverse, Romster is the CRUX nazi :)08:19
prologicsorry for any Germans here :)08:19
Romsternah i just spotted it after diverse said about c-ares that i also use the chown is rarely ever needed on any port c-ares has it.08:19
Romsternazi i'm just following the guidlines.08:19
Romsterif i don't point it out who will?08:19
Romsterwhere will quality go?08:20
prologicit's okay :)08:20
prologicquality is good :)08:20
prologicthanks for picking my flaw up :)08:20
Romsteri don't wanna be the nazi police around here lol.08:21
Romsteri'm just observant.08:21
Romsterfeel free to pick on my ports if you can fault any.08:21
diversewhy would following protocol make someone a facist?08:21
Romsteri dunno. because i whinge and complain alot here in the past?08:22
diverseOh okay, the "bossy" part about you ;)08:22
diverseprologic: trust me, Romster is a good dictator08:25
diversenot an evil one08:25
Romsteri was looking at finit runit like but a single finit.conf file for setting up. maybe we could move to that instead of sysvinit.08:28
Romsterless obtrusive than runit itself.08:28
diverseRomster: being jokes aside, have you checked out s6?
Romsterseen it but this is even simpler
Romster1 conf file that's it08:32
diversepretty neat08:33
Romsternot sure how lvm and mdadm fit in but probably wouldn't be much effort.08:34
diversefinit looks pretty KISS to me08:35
*** openfbtd has joined #crux08:37
prologiclet us know what is' like :)08:42
diversewell I like the fact you can list the services in the finit.conf instead of making symlinks in a directory08:45
z3braHi Romster !08:45
z3braany link to finit ?08:45
Romsterhi it's at
z3bralooks nice08:48
prologicyeah I hate the whole symlink crap08:50
prologica bunch of scripts in /etc/rc.d08:50
prologicand a list of ones to start08:50
prologicplain and simple08:50
*** diverse has quit IRC08:51
Romsteri'm thinking of trying that out.08:51
z3brainit.c -> int main(void) { execv("/sbin/rc", NULL); return 0; }08:51
prologicsounds fine to me :)08:52
z3braand then, svc to manage daemons :)08:52
prologicin fact that sounds perfect08:52
z3brayou'll like sinit then08:52
Romsterso you take back the nazi comment now lol :D08:52
prologichonestly I'd even be happy with supervisord being my init of choice :)08:52
prologicit was a joke :)08:52
Romsteri know08:53
Romsteris that all that sinit is?08:53
prologicis that it?08:54
prologicthat's amazing08:54
prologiceven I could write that08:54
prologicnow that I"m writing more and more C tehse days08:54
prologicwith Arduino08:54
Romsterok so after that what do you do for starting services?08:54
Romsterand modprobe and stuff08:54
*** ardo has joined #crux08:54
prologicmodprobe should have nothing to do with init08:54
prologicnor should services08:54
prologicit's just a way of spawning something08:54
z3braIt's managed by other programs like svc and such08:55
prologicyou know the whole /etc/rc.d/<service> start|stop|restart we have in crux?08:55
prologicthat has little to do with sysvinit at all08:55
prologicit just starts or restarts the appropraite daemon08:55
prologicsysvinit actually spawns /etc/rc afaik in crux08:55
prologicat least that's what /etc/inittab tells it to tdo08:55
*** diverse has joined #crux08:55
prologicfrom there on it's all bash scripts08:55
prologicand some kind of structure we implemented08:55
z3bracheck this out08:56
prologichmm maybe later08:57
z3brathat's `svc`, a service manager08:58
z3brabut is actually a cli tool for sysV08:58
BitPuffinAmnesia: is that  for real :o08:59
prologicyeah I don't need or want a "service tool"08:59
prologicI'm quite okay with /etc/rc.d/<service> start|stop|restart|status08:59
z3braall you need is a function08:59
z3brastart() { /etc/rc.d/$1 start }08:59
prologicit's not that hard to abstract09:00
BitPuffinI'm not sure how I feel about it09:00
prologicbut otoh if I wanted such an init with service tool and "management" and maybe an API09:00
prologicI'd design and implement one from the ground up09:00
prologicbut without hard bound dependeicies on udev or any other silly nonsense09:00
prologicbut really IHMO09:00
prologicany API should be atop any init09:00
prologicand agnostic of any init09:00
BitPuffinAmnesia: would at least be good if they said who was doing it09:00
prologicso I wouldn't do that either09:00
z3brayou might like to follow this project:
prologicI'd just write a new init09:01
prologicand write an API for it separately09:01
prologicin fact I kinda do want to write an agnostic API for init systems -- at least sysvinit for starters09:01
z3braThey're building a distro using sbase, sinit and pkg management using `mk'09:01
AmnesiaBitPuffin:  LibreSSL is primarily developed by the OpenBSD Project, and its first inclusion into an operating system will be in OpenBSD 5.6.09:02
z3braYeah that's serious09:02
diverseAmnesia: ah, that's fantastic09:02
z3brabut the name is silly09:02
diverseWell OpenSSL is already taken09:03
z3braas well as OpenTLS, I know09:03
z3brabut meh :/09:03
BitPuffinAmnesia: I see, well still, some namedrops would be nice09:03
BitPuffinbut yeah OpenBSD reacted fairly strongly to the heartbleed thing09:03
BitPuffinso it's not entirely surprising09:04
diverseFinally OpenBSD is taking the stand09:04
BitPuffinopenbsd rocks09:05
*** jdolan has joined #crux09:08
diversewoohoo LibreSSL09:09
*** jdolan has quit IRC09:13
z3brawhy is it useless ?09:13
z3braoh, btw09:13
diverseBitPuffin: That's true, they could just fork a much better implementation as a base, but maybe they want to show that they can code it better than the OpenSSL team09:13
z3bradoes anybody uses fgetty ?09:13
BitPuffindiverse: well they probably don't wanna fork gnutls because gpl09:14
BitPuffindiverse: well I just mean why wouldn't they just work with openssl09:14
diversethere are tons out there, just not as well known09:14
BitPuffinhowever maybe they aren't too happy about openbsd coming in and taking over everything09:14
BitPuffinso maybe that's why it had to be through a fork09:14
BitPuffinthere may be indeed, but probably not very complete ones09:15
BitPuffinanyways one of the problems with OpenSSL is that it only has one fulltime guy on it09:15
BitPuffinwhich is weird09:15
BitPuffinbecause you'd think companies would at least try to invest in something that is so critical to their infrastructure09:15
diverseand OpenSSL has a very messy code base, which makes it very hard to audit the source09:16
BitPuffinindeed, but if more people worked on it it could be cleaned up09:16
diverseit's going to be a lot of work to clean up OpenSSL for the OpenBSD guys09:16
BitPuffinanyways I haven't checked the mailing lists to see discussions on the people who are making libressl even tried to work with openssl09:16
BitPuffinyeah openssl is a real mess09:17
BitPuffinhonestly it was well known that the codebase was messy, it's not surprising at all that there would be critical flaws in ti09:17
BitPuffinapparently eating one banana can make you non-sleepy for a pretty good timespan09:18
diversebut the more I think about this, I am starting to see this as also a marketing strategy, so they can say "My product is an *improved* fork version of OpenSSL, so switch to using my product, since it's *better*" because a lot of software use OpenSSL blindly and would give them incentive09:18
BitPuffinmaybe this should be my new coding fruit of choice09:18
diverseLike how with LibreOffice is to OpenOffice09:19
BitPuffinyeah perhaps09:19
BitPuffinI feel sorry for openoffice09:19
BitPuffinbecause I love that apache took over it09:19
BitPuffinbut since the GPL is such a shitty license they have to reinvent the actual useful things that libreoffice added if they are gonna add it09:19
BitPuffinbut I digress09:21
BitPuffindiverse: however right now it's only gonna be included in openbsd and I think by the time it gets available for other platforms heartbleed won't be as heated anymore09:22
diverseWell, who can't turn down what OpenBSD has to offer, I mean their OpenSSH implementation is freaking famous09:23
BitPuffinyou mean who can? :P09:23
cruxbot[contrib.git/3.0]: c-ares: 1.10.0-1 -> 1.10.0-2 -- Fixed man page location --mandir=/usr/man09:23
cruxbot[contrib.git/3.0]: mosquitto: 1.3.1-2 -> 1.3.1-3 -- Fixed man page location09:23
diversethere fixed09:24
BitPuffindiverse: well if that was how the world work we would all be running openbsd09:25
prologicFor those interested in MQTT:
prologicThis is ready to rock 'n roll with the ownTracks (previously known as MQTTitude) on Android for personal/private GPS Tracking09:26
diverseBitPuffin: well the OS is pretty niche, but the cross-platform tools they write work very well and doesn't matter which OS we use :P09:26
BitPuffindiverse: I wouldn't say it's that niche, it's a splendid all around general purpose os09:27
BitPuffinit just happens to be secure by default09:27
BitPuffinwhich is not a niche really09:27
BitPuffinit's something we should all want from an OS09:27
diverseWell it's niche in a sense, that all they do is focus on security, where as FreeBSD is more about being bleeding edge and trying to get more software to work on it09:29
diverseHowever if you have all the tools you want on OpenBSD, then it's perfect09:30
BitPuffinstrangely enough openbsd is like better at graphics drivers and stuff lol09:33
diverseoh, they got good open source drivers?09:34
BitPuffinI'd say FreeBSD is more about throughput and server performance etc but maybe that's just how I see it09:34
BitPuffindiverse: well generally when there is news about say FreeBSD or dragonfly BSD getting updated ati drivers or something it's always "ported the openbsd drivers"09:34
diverseYeah, for FreeBSD it's about performance and compatibility (like with the Linux emulator they have built into the kernel)09:35
BitPuffinpretty sure openbsd can do the linux compatability as well09:36
BitPuffinI know netbsd can09:36
BitPuffinheading home09:36
diverseto my knowledge, I'll check it out09:36
diverse*not to my knowledge09:36
BitPuffinI would be surprised if it didn't09:36
BitPuffinopenbsd is actually a much better regular everyday os than people think09:36
BitPuffinwill be back in about 4 hours09:37
diversehmm, than I wonder why FreeBSD is getting more attention?09:37
BitPuffinmindshare I guess09:37
BitPuffinmomentum with word of mouth09:37
BitPuffinsomething like that09:37
BitPuffincan be discussed lataz :D09:37
diverseyeah, go home already!09:38
*** BitPuffin has quit IRC09:42
diverseMaybe I will contribute to OpenBSD by buying their next release as a way to show my thanks for going out of their way to implement their own SSL/TSL10:08
diverseI'm serious :)10:12
prologicI know :)10:12
prologicbtw... is anyone else besides me using Docker on CRUX yet?10:13
prologicOr using my CRUX Image on some Docker platform? :)10:13
diversenot me, but I don't know what docker is10:13
diversesite is taking forever to load10:16
prologicheh you'er right10:29 is down too10:29
prologicmaybe some infrastructure has gone down somewhere in the world10:29
prologicor some routes10:29
diversewiki page would be all I need :)10:31
Amnesiayou folks might like aswell10:31
diverseoh okay, so docker is a wrapper to lxc giving it more features10:32
prologicsure that's one way of looking at it :)10:33
prologicbut that's only a very shallow view of it :)10:33
prologicI believe in the works (or already) are other execution engines such as libvirt, qemu, kvm, etc10:34
prologicit'a also now in OpenStack Icehouse as part of Heat10:34
prologicand you've got now support it as well and so on10:34
diverseI already know what lxc, containers, and jails are and what you could use them for. As I understand it now, docker gives you some app individuality, without the need to create a subsystem, say to jail to host services in case if hackers attack you and they can't go outside of the jail.10:44
diversewhat do they mean by portable though?10:46
diverseis it like process isolation and transferring it to another system?10:47
diversewhile it still runs?10:48
prologicDocker images are portable11:05
prologicwhat I buidl here you can run there11:05
diverseokay I got the full idea now11:06
prologicDocker also (if it wasn't clear before) provides a clean, consist, reliable way to package up your applications and services11:06
prologicfrom development to deployment11:06
prologicalso provides linking (discover ability) and volumes as well as ip management11:06
prologicas I've explained before:11:07
prologicdocker run <app>11:07
prologicplain and simple11:07
diversehmm, I wonder if this will replace a lot of the mainsteam existing package managers?11:08
prologicquite possibly in some way or another11:10
prologicbut it won't replace anything I don't think11:10
prologicDocker plays nice with others11:10
prologicIn fact it's quite normal to run Docker on top of existing virtual infrastructure11:11
diverseI'm seeing distros would go "goodbye rpms, sayonara debs! Hello Docker!"11:11
Romsteryour misisng the point docker is a complete image you can still use packages to install into a docker image.11:12
frinnstholy hell, this massive backlog spans 10 hrs11:14
diversebut what if someone decided to pack just the app files in the snapshots for application installation?11:14
diversealthough I guess uninstallation would be a different matter?11:16
diverseI guess docker is more or less a distributed system cloning tool?11:19
diverseand by distributed, I mean, you send chunks to clone in a variety of ways.11:20
diverseRomster: ^ ?11:21
Romsterdamn hg prologic your the hg expert how do you move to a specific commit in hg11:22
Romsteralready got the tree with hg clone ...11:22
Romsteror diverse if you use hg11:23
Romsteri cant' do everything at once11:23
diversedoes prologic do his ports in mercurial?11:25
diverseor something else?11:25
Romsterhg update -r d121a0b16ff4c1c1614f128d3e231ad3fc7a45ec11:25
Romsterfound it...11:25
*** jdolan has joined #crux11:31
Romsterfiles here i can't find what commit has them svg files11:32
diversewait with docker, apps can "run everywhere" but the container and host architecture must match? What a turn off.11:41
diversemight as well just compile things statically and share the one binary to others with the same host architecture11:42
*** jdolan has quit IRC11:45
*** jdolan has joined #crux11:45
prologic<Romster> damn hg prologic your the hg expert how do you move to a specific commit in hg <-- hg update <revision>11:47
prologicdiverse> does prologic do his ports in mercurial? <-- yes11:48
diversethat explains Romster's frustration11:48
Romsterdid that i'm not fidning crap tonight  givve up i'm just gonna go play a game11:48
prologicdiverse> wait with docker, apps can "run everywhere" but the container and host architecture must match? What a turn off. <-- that's only a limitation of lxc or openvz. the same would not be true if you used a libvirt, qemu or kvm driver for example (out of Dockers hands per say)11:49
prologicdiverse, so obviously you could think of Docker as an OS of sorts11:49
Romstertoo frustrated todo anything minds racing around at all the work i did today... so game time....11:50
prologicIf you had the Linux Kernel + Docker as your init and a simple CLI11:50
prologicYou'd basically have an entire OS complete with "package management"11:50
prologici.e: where you could run at the CLI: run <some_image>11:50
prologicMotor shield, Stepper motor, Transistors, Diodes, a Solenoid and a bunch o f servos and dc motors11:51
prologicyay :)11:51
Romsterat least i know my electronics stuff... damn computers.11:52
prologicCheck out:
prologicbunch of code samples I've put together for the TinyDuino paltform11:54
diverseif one day, docker allows apps compiled on one architecture and to run on a different one (assuming both hosts have Linux), then color me excited, because that could give Java a run for it's money11:54
prologicWell sure that's already possible :)11:56
prologicAs long as your App is written in Python11:56
prologicor Java11:56
prologicor Javascript11:56
prologicor Ruby11:56
prologicproblem is if you were referring to binaries11:56
prologicnow you're getting into some "HARD" stuff11:57
prologicthat is not easily solvable without some kind of abstraction (e.g: a virtual machine)11:57
prologicoh wait we already have that :)11:57
prologicPython, Ruby, Java, etc, etc11:57
Romsteryou could have universal binaries that have code blocks that pick one depending on the arch to run on.11:58
diversewell yeah, I'm talking about binaries. And LXC is a vm at the file system level. I'm mean, how cool will it be to have portable binaries that can run at almost hardware speed?11:58
Romsterbasicly compiled for all supported arches and some header is picked for the entry point.11:58
Romsteronly other way is by compiled and a interpreter11:59
prologicYeah this is called  JIT :)11:59
Romsteraware of that11:59
prologicyou guys make Computer Science sound easy11:59
prologicit's not :)11:59
Romstermaybe frustrated but not dumb <<11:59
prologicgo do your gaming :)11:59
prologicI have to do some reading for my PHD11:59
diverseI'm not saying anything is easy, I'm just saying if they manage that, that would rock12:00
diverseother than that, I don't see anything that interesting about docker, sorry.12:00
Romsterbut it's too hard too heck when t's eveither left or right indian or crap i get lost there.12:00
Romsterlsb msb i recall.12:01
Romsteryeah i'll just go game far eaiser.12:01
prologicdiverse, what aren't you getting?12:11
prologicfor me it solves all kinds of problems between developers, sysadmins and network admins12:12
prologicand bridges the gap between application developers and infrastructure12:12
*** dkoby has joined #crux12:33
*** jdolan has quit IRC12:44
*** jdolan has joined #crux12:57
Romsteri think he gets it now prologic13:01
*** jdolan has quit IRC13:01
diverseyeah, I didn't understand it's true purpose until I talked with Romster13:02
diverseprologic: ^13:02
*** linXea has quit IRC13:07
*** linXea has joined #crux13:07
prologicdiverse> other than that, I don't see anything that interesting about docker, sorry.13:12
prologicthat was your last statement :)13:12
prologicmakes me think you're missing something :)13:12
jaegerfrinnst: I like their workaround: "don't use it"13:13
diverseprologic: well I read what it did, but I didn't understood it's practicality, especially in a business pov13:15
diverse(considering I haven't done any real sysadmin or programming work yet)13:16
prologicthat's precisely where it really shines13:19
prologicin a business perspective13:19
prologicstreaming the whole development to production depployment13:19
prologicif you can just hand your infrastructure guys a single image13:19
prologicand they run "run it"13:19
prologicthat's quite invaluable13:19
prologicbut it's not just about the "image"13:20
prologicDocker and it's ecosystem is quite powerful and flexible13:20
prologicwhere you infrastructure guys can hook up all sorts of elaborate combinations of things knowing that your container/image (your app) will function no matter what13:20
prologicI guess I've done both sides of the life cycle of an app/service and still do :)13:21
prologictbh I don't trust sysadmins to do a job properly13:21
prologicno offense to any sysadmins here :)13:21
prologicso if I can just hand them a Docker image, that's +1 for me :)13:22
diverseit's less for you to worry about13:22
prologicwhat's more, there's none of this silliness where something breaks in production and you can't diagnose it because you can't reproduce production - now you can13:22
prologicor at least it's a lot easier to13:22
prologicthe value comes from incorporating it into your development workflow and life cycle13:23
prologicfrom development to full production13:23
prologicwhat you develop, and what ends up in production should not differ13:23
jaegerits use is also a subjective thing. I understand what it's for and the problem it's trying to solve but I still have zero use for it, myself13:23
jaegerFortunately I don't have developers to support... researchers are hard enough :P13:24
prologicanything always comes with some subjectivity :)13:24
prologicit's like anything really13:24
prologicif you have good tools you can do cool stuff13:24
prologicyeah that's the field I'm in these days too13:24
prologicthem pesky researchers13:25
prologicoh wait I'm going to become one13:25
prologicrats :)13:25
jaegerget out while you still can13:25
prologicnah I wanna get in for some weird reason13:25
prologicI guess maybe so I can have something to fall back on when I'm considered an old goose and can't program or keep up with the young nit wits :)13:26
prologicteach them instead :)13:26
jaegerthat's what retirement is for13:26
prologicI think I have a few decades before that13:26
*** jdolan has joined #crux13:27
diverseprologic: you may end up being either a project manager or a sysadmin after you aged13:28
Romsteror some old geaser like i used to run os/2 only it be some other system that's long since gone and moved computing into quantum mechanics or something13:30
Romsterwho knows where the world is heading.13:31
prologicdiverse, maybe13:33
prologicbut not any time soon13:33
prologicI care too much about my work13:33
prologiconce I stop caring, perhaps13:33
jaegeryou'll stop once you become a researcher :) especially if there's a doctoral degree involved13:34
prologicmaybe you're right13:35
prologicbut that'll be at least another 5-7 years ago :)13:35
diverses/ago/to go/13:35
diverseRomster: quantum computing is going to be like a whole new realm13:36
Romstermost likely.13:39
diverses/likely/definitely/ :P13:40
Romsterah jigsaw solving so relaxing13:41
*** nullmark has quit IRC13:41
Romsterdone like 5 jigsaws already..13:41
Romsterin the end you'll only wanna do what is useful to yourself.13:42
Romsteror pay someone to get it done13:42
*** nullmark has joined #crux13:42
prologicBtw, -- in the next coming weeks tehse guys will be offering $1 and $2 per month Docker hosting :)13:42
Romsterwhy so cheap?13:43
prologicbecause they can13:43
prologiccheap == more customers/users13:43
prologiclowers the barrier to entry13:43
Romsterprobably very little resources.13:43
prologicI've been in talks with the CTO and CEO for the past few months now13:43
diversemore customers == more demand13:43
prologicwith feedback and suggestions13:43
diversemore demand == more profit13:43
prologicRomster, it's not a vm :)13:43
Romsterthere is that too diverse get them in then bung up the price13:44
prologic64MB for an application container at $1/month13:44
prologicis more than enough to power an entire wordpress site for example13:44
prologicand then some13:44
Romsterno but they can restrict memory cpu time and bandwidth/quota per a user.13:44
prologicthat's kinda the point13:44
prologiceve at the resources of $1/month13:44
Romsterwordpress -_- that thing anyone hacks into.13:44
prologicyou could still spin up a CRUX Docker container on their infrastructure13:44
prologicand run sshd13:45
prologicand basically use it as a realyl really cheap vm13:45
Romsteri'll be using some resources when i start getting serious at spidering and testing mirrors.13:46
prologic1000 users at $1/month is more profit compared to 100 users $5/month :)13:46
prologicno :)13:46
prologicyou will scale up more resources on demand13:46
Romsterwant to distribute the load a bit on a few hosts to avoid downtime too.13:46
prologictake advantage of their auto scaling feature(s)13:46
prologicand scale down when your system is idle13:46
prologicTutum auto scales across AWS regions for you13:47
Romsterhmm i see13:47
prologicautomatically if you tell it to do so13:47
Romsterwont be all the time but those moments you need it.13:47
diversefrinnst: ^ another massive text wall :P13:49
frinnstim already ignoring it13:52
Romsterthe logs are over flowing13:52
*** nullmark has quit IRC13:53
prologicirclogger_ can handle it :)13:53
*** nullmark has joined #crux13:56
*** nullmark has quit IRC13:56
*** nullmark has joined #crux14:02
diversegeez, look at how sloppy the OpenSSL guys are and what the OpenBSD team is doing to clean up their mess:;r2=1.2315:50
diverseespecially that for loop with the /* increment */ comment15:51
prologicyes I saw15:54
prologic90,000 deleted lines of C15:54
prologic150,000 deleted lines of content15:54
prologicThat's what (unfortunately) happens to software projects over time15:55
prologicThey become large and un-maintainable15:55
prologicit's actually a real problem in complex software systems - one which I don't think we have a good answer for (good PHD topic IHMO)15:55
diverseyeah and that article pointed to the OpenSSL rampage site that tilman posted a couple of days ago and I started reading from the bottom and can't believe how f*cked up the code is. It's like OpenSSL was on drugs like what Theo said.15:56
diverseand they just do stuff which is not needed and lazy solving problems15:58
diverse*lazy at15:58
*** dkoby has quit IRC15:58
prologicwell call it what you want :)15:59
prologicsoftware development is by no means easy15:59
prologicor everybody would be doing it :)15:59
prologiclazy sure - lots of developers are lazy15:59
prologicoverly complex - sadly yes a lot of software gets that way15:59
diverseyeah, but there is no excuse for doing code in a sloppy style, with text editors to handle that for you.15:59
prologicmaybe not16:00
prologicbut tools aren't perfect16:00
prologicand you can't trust the tool to do the right thing anyway16:00
prologicjust look at any C# project :)16:00
prologicyou can practically know nothing about the langauge16:00
prologicwith all the fancy drop-downs and auto complete16:00
prologicdoes that mean you write good C#?16:00
diverseI'm not talking about autocompletion, I'm talking about the style of code16:03
diverseContent will be good or bad, but just to make your code look obfuscation on purpose to make auditing harder is the worst16:03
prologicSo in Python world16:03
prologicyou're referring to something like Pylama for Vim16:03
prologicwhich incorporates PEP8, McCabe and PyFlakes into your editor16:04
prologicensuring at least some level of quality code16:04
prologicstyle, complexity and basic errors16:04
prologicI agree16:04
prologicobfusation is never a good idea16:04
prologicand bad practice16:05
prologicit makes you look stupid16:05
diversedid you look at the code they wrote? It's like they must of been high or something16:05
diversewhatever their smoking16:05
diversethe code they write looks as if a 2 year old was mashing the keyboard to make code16:07
*** jdolan has quit IRC16:09
*** toriso has joined #crux16:12
*** ardo has quit IRC16:13
prologicI finally have CNAME hosted containers on my Desktop working16:19
diversemore power to you16:19
prologicCRUX + Docker (skydns + skydns + hipache) + * wildcard A record16:19
prologicwell it's kinda awesome16:19
prologicbecause now I can spin up any web app I want16:19
prologicand expose it to the web :)16:19
*** jdolan has joined #crux16:23
joacimhigh pitched whine coming from my old 440bx-based pc, 12V rail measures almost 14V17:04
joacimfun stuf17:04
*** jdolan has quit IRC17:44
*** jdolan_ has joined #crux17:46
*** jdolan_ has quit IRC17:55
*** jdolan_ has joined #crux17:59
*** jdolan_ has quit IRC18:15
*** jdolan has joined #crux18:57
diverseis there a way to clear all the join/quit messages in weechat?18:57
nogagplz/filter add joinquit * irc_join,irc_part,irc_quit *18:59
prologicX-Chat: 1. Right Click Channel 2. Settings 3. Hide Join/Part Messages19:00
nogagplzyeah but xchat is an ancient piece of junk, all the cool kids use weechat19:01
drijenfuckin lies19:01
drijenweechat is for you newfangled whippersnappers19:01
teK__/ignore #crux JOINS PARTS QUITS19:01
nogagplzalright alright I'm leaving gran torino, no need to kill yourself19:02
diversethanks nogagplz19:03
diverseah so much cleaner the channel is19:06
nogagplznow just ignore drijen and your training is complete19:06
diversedrijen who?19:07
nogagplzI have nothing left to teach you19:07
prologicyeah you whippersnappers19:11
diverseI was amused by drijen's words19:11
diversein the good way19:12
prologicit was funny as19:12
diversealways good to bring laughter to this channel19:16
joacimHello fellow kids19:22
joacimI use weechat too19:22
diversedrijen: got anymore remarks?19:23
drijenno just this19:27
nogagplzharder, faster, deeper19:28
nogagplzalso, do I still need to repair your tv19:28
nogagplzor are we done here19:28
diverseyou both get along well19:29
*** BitPuffin has joined #crux19:29
*** BurnZeZ has joined #crux19:30
BurnZeZ/usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory19:30
BurnZeZWhen trying to build gcc19:30
jaegeryou need glibc-32 installed19:31
jaegerit's the only -32 port that's required19:31
BurnZeZWhat the crap19:31
BurnZeZWhy isn't it in the dependency list?19:31
jaegerbecause it's in core19:32
diversehmm, didn't you use depinst?19:32
BurnZeZI wonder how it got removed19:32
BurnZeZYeah, depinst19:32
jaegerdepinst won't catch it since it's not listed but the assumption is that core is always installed19:32
BurnZeZMaybe I removed it without realizing19:32
BurnZeZOh god19:33
BurnZeZglibc-32 won't install for the same reason19:33
jaegeryeah, it breaks the toolchain when it's gone... your best bet is to reinstall it from the ISO19:33
diverseor just chroot and pkgadd it19:34
jaegerwell, the system should still be working, shouldn't need a chroot19:34
*** timcowchip has joined #crux19:34
jaegerjust pkgadd from the ISO should do it19:34
diversefair enough19:34
BurnZeZI still have an old package of it19:34
jaegeror that19:35
diverseboom there you go19:35
drijendiverse: i've known nogagplz and Romster since like crux 1 point something19:36
drijenlogn before that bastard child arch was born19:36
diverseah a Crux veteran, nice19:36
drijenno, just an old fart19:37
BurnZeZprt-get: updating glibc-32 from 2.16.0-6 to 2.19-119:37
BurnZeZtelinit: /dev/initctl: No such file or directory19:37
BurnZeZ-- Packages updated19:37
BurnZeZglibc-32 [post: failed]19:37
BurnZeZWhat's this /dev/initctl?19:38
jaegerdoes that happen if you run "telinit U" manually?19:38
BurnZeZOh, this is init crap19:39
diversedrijen: is there anyway you can convince nogagplz to stop using arch?19:40
BurnZeZjaeger: I don't use init19:40
jaegerthen you probably don't need to worry about the post-install script19:41
BurnZeZI just wrote some other script I use19:41
jaegerall it does is call "telinit U"19:41
BurnZeZAlright, cool19:41
drijendiverse: no19:48
nogagplzyou reach a point where you learn all you can19:49
nogagplzand just want shit to be fast and easy19:49
nogagplzrather than wondering why you have no /dev/initctl :P19:50
drijenif that was the case, then why aren't you using slack.19:50
nogagplzbecause they only accept the impotent19:50
drijenin that case, you'd be king in a month19:50
nogagplznot before dethroning you though19:50
drijensorry what19:50
drijeni can't hear you over the sound of my os not sucking19:51
BurnZeZI just use 9front19:51
nogagplzsays the microsoft shill19:51
BurnZeZPlan 9 master race19:51
drijenlies, openbsd19:51
BurnZeZ9front >>>>> OpenBSD19:51
nogagplzoh lord the theo de raadt cult19:51
drijeni won't deny that guys genius, but man, he traded it for a congenial attitude19:51
BurnZeZ9front has theo(1)19:52
BurnZeZ; theo19:53
BurnZeZGod bless the people who employ you, they need the blessing.19:53
diverseWell, I am glad they are doing the LibreSSL effort19:54
BurnZeZLinux would be far worse without OpenBSD picking up the slack19:54
BurnZeZUserspace tools and programs, that is19:54
nogagplzshouldn't the blame be placed firmly on gnu then19:55
nogagplzand linux be used with an openbsd userland19:55
BurnZeZI do something similar with my linux machine19:56
BurnZeZI use plan9port userspace19:56
BurnZeZSabotage is really cool, but butch is a mess19:56
BurnZeZSo what I do is use sabotage to compile programs statically linked with musl19:57
BurnZeZMy linux system is CRUX with heavy modifications, and then I use sabotage inside chroot for compiling some things statically19:57
diverseOh yeah, I think I remember you now19:58
diverseWanting to create your own variant of Crux with other userland tools19:59
diversebut you weren't sure back then20:00
diverseabout the userland part I mean20:01
*** timcowchip has quit IRC20:23
*** timcowchip has joined #crux20:31
timcowchipdoes crux have a ca-certificates.crt?20:35
diversei think it does20:35
timcowchipI see a cups.crt and a stunnel.crt20:35
timcowchipin /etc/ssl/certs/20:36
diverseI don't have it either, don't worry about it20:36
timcowchipI get 'SSL handshake failed' with uzbl20:37
jaeger /etc/ssl/cert.pem, provided by ca-certificates20:37
timcowchipthanks @jaeger20:38
*** jdolan has quit IRC20:51
*** jdolan_ has joined #crux20:54
*** jdolan_ has quit IRC20:58
*** timcowchip has quit IRC21:18
*** jdolan has joined #crux22:09
*** c0x` has joined #crux22:14
*** c0x has quit IRC22:19
*** darfo has quit IRC22:27
*** toriso has quit IRC22:33
*** jdolan has quit IRC22:48
*** jdolan has joined #crux22:49
*** jdolan has quit IRC22:54
*** darfo has joined #crux23:36
*** jdolan has joined #crux23:50

Generated by 2.11.0 by Marius Gedminas - find it at!