IRC Logs for #crux Friday, 2013-11-01

*** joacim has quit IRC00:09
*** joacim has joined #crux00:09
*** sudobaal has joined #crux00:40
*** prologic has joined #crux00:41
*** Rotwang has quit IRC01:20
*** Romster has joined #crux01:32
*** Romster has quit IRC01:32
*** Romster has joined #crux01:32
*** AndChat150689 has joined #crux02:17
*** vee__ has quit IRC02:18
*** tilman has quit IRC03:04
*** tilman has joined #crux03:05
*** __mavrick61 has quit IRC03:46
*** __mavrick61 has joined #crux03:48
*** AndChat150689 has quit IRC03:57
tilmanfrinnst: do you use lilo with your btrfs system, or extlinux?07:06
diverselast I recall, frinnst used syslinux, so I'd say he uses extlinux.07:11
diversefrinnst: how is syslinux recently?07:33
*** Faze has joined #crux07:38
FazeI need some help :) would you help me?07:38
diverseSince this channel is usually quiet and I got nothing better to do, sure.07:39
FazeHaha! okay :)07:39
FazeSo I need to build a crux image.07:40
FazeI am a noob at this.07:40
diverseSorry, this is out of my expertise. I think frinnst, Romster, jaeger, etc are the ones that can help you with that.07:43
FazeOh that's okay, thank you :)07:45
diverseIf you don't mind hanging out, those guys will come around eventually.07:46
FazeOkay, so what do you do for a living?07:47
diversecurrently a student07:48
FazeOh okay, me too by the way.07:48
FazeHow old are ya?07:48
tilmanFaze: hello do you want to chat07:52
frinnstwhat do you mean by "image"? install image/iso or a kernel?07:52
FazeYes yes!07:55
Fazeinstall image07:56
frinnstthe sources are available here:;a=summary07:57
FazeIs there a good tutorial?07:58
frinnstdont think so, but you pretty much just run "make"07:58
frinnstcheck the Makefile for more info07:58
*** DaViruz has quit IRC08:03
FazeWhat is the best host os for building crux?08:04
frinnstcrux probably08:04
frinnstbut you should be able to do it with most distros i think08:04
FazeNice :)08:07
FazeWhere should I start?08:08
frinnstclone the repo, checkout the branch (3.0), copy/clone the core, opt, xorg repos and place them in the iso-tree's ports path08:10
frinnstrun make bootstrap and hope for the best08:10
diversefrinnst: now that's funny08:11
*** DaViruz has joined #crux08:15
*** Faze has quit IRC08:27
cruxbot[core.git/3.0]: dhcpcd: temporary switch source URL to crux.nu08:32
*** frinnst_ has joined #crux08:52
diverseis there a git command that gets the latest commit checksum from a branch?08:54
*** frinnst has quit IRC08:58
diverseso far I found this: git ls-remote <url> HEAD08:59
Romstergit branch08:59
Romstershould show stuff.08:59
Romsteroh you wanna do it remotely.09:00
diverseyeah I want to remotely grab it and use it to update the Pkgfile09:00
diverseunfortunately that output shows "HEAD" right after the checksum09:01
Romsteri usually use git rev-list HEAD -n 1 | head -c 709:01
Romsterbut that's after git fetch ; git reset --hard head09:02
diverseI guess I will need to use some sed magic09:06
Romsteryou could do make a maintainer script and sed the to the current git version.09:07
diverseis that how I get the revision replaced in the Pkgfile?09:09
Romsteris one way you can do it.09:09
Romstersee how it's done with mozconfig in firefox.09:09
diverseAlso, I found this, thanks to awk: git ls-remote <url> HEAD | awk '{ print $1 }'09:10
Romsteryeah that works or cut -d' ' -f109:12
diverseRomster: looking at mozconfig, I don't really understand what I am suppose to do. Use mk_add_options?09:20
Romsterno see the sed line in Pkgfile09:20
Romstersee how it uses a and sedes the #CFLAGS# to mozconfig out09:21
Romstersame idea only different files and say version=#VERSION# on pkgfile.in09:21
Romsterlet me know if your still confused and i'll write an example09:27
diversesorry man, I am still confused. I don't see how it works and what I am really doing.09:28
Romsterdo you have a Pkgfile already? pastebin09:28
Romsterare you trying to just bump the verison= automaticly or actaully generate a tarball and/or bump the version?09:33
diversebump the version= automatically only09:34
diverseand cversion=09:35
diverseversion= would only be taking the first 7 chars of sversion= and suffixing it with "git-"09:39
Romster or some such name set it to executable.09:44
Romsterthen mv Pkgfile Pkgfile.in09:44
Romsterand edit version and sversion to be version=#VERSION# sversion=#SVERSION#09:44
Romsterevery time you run generate it'll fetch the new master commit and write it to Pkgfile09:45
Romsteri haven't tested this but it should work. i copied code out of my tarball generating script for git snapshots.09:45
*** Rotwang has joined #crux09:46
diverseI see how this works now, I will give it a shot, thanks Romster.09:47
Romstershould be trivial to fix if i messed up.09:47
Romsternow that you can see how it works.09:47
Romsterversion="$(cd work/${name}-repo && git rev-list HEAD -n 1 | head -c 7)"09:48
Romsteris what i use locally.09:48
Romsterafter git fetch and git rebase09:48
diverseah I gotcha09:48
Romsterbut then i tarball it up. and give the filename like x264 has fakeroot tar --remove-files -cf ${name}-${DATE}-${version}.tar ${name}/09:49
Romsterdate being DATE="$(date +%Y%m%d)"09:49
*** sudobaal has quit IRC09:50
*** sudobaal has joined #crux09:53
*** sudobaal has quit IRC09:55
*** Pingax has joined #crux09:59
diverseRomster: weirdest thing, I get an error saying: sed: -e expression #2, char 32: unknown option to `s'10:01
diverseit didn't recognize the s command10:01
diverseor wait, I guess that means, its not happy with the 2nd -e option10:04
diversebut I have #SVERSION# written correctly in the Pkgfile.in10:04
Romsterwonder why sed does not like more than one expression...10:06
*** sudobaal has joined #crux10:08
diverseI guess its limited to only one -e expression?10:10
Romsterhmp my string line got messed up10:12
Romsterstring="$(git ls-remote master |awk '{print $1}')"10:12
diversenow its happy10:13
Romsteryep and it works here now.10:14
diversesweet, now I can do this for my other git ports and be lazy \o/10:14
Romsterbeing that your using > and not >> on the sed line means it'll wipe Pkgfile and cat though them regex's before writing Pkgfile.10:14
Romsteri've been thinking of doing something similar10:15
diverseyeah I know that > overwrites and that >> appends to the existing file10:16
Romsterser -e for good measure10:16
Romsterexits on error.10:16
Romsterbut being $() is a sub shell. i check the result before continuing.10:17
Romsteri'm sure i make mistakes somewhere with all the stuff i write -_-10:17
diverseso does everybody10:18
diverseI didn't blame for the missing 'i' in "/bn/bin" nor that you had to connect the10:19
diverse"=" to the string var10:19
diversebesides you're awesome10:20
diverseOh I got an even lazier idea! :)10:21
*** sudobaal has quit IRC10:22
diverseRomster: why not also make a "giturl" variable which holds url and use that to replace the #GITURL# in the pkgfile. That way if the location changes, I just make one change :P10:24
Romsteryou can do that.10:25
Romsteri'm being lazy <<10:25
diversewell you are either lazy in the less productive or most productive way.10:26
Romsteri wanna make a patch to pkgmk so i can do a unpack() function for odd stuff like git and double tarballed archives. so they have the source in $SRC on pkgmk -eo10:27
Romsterwouldn't take much todo that.10:28
*** Pingax has quit IRC10:29
diversewould unpack() automatically pull the latest git commit?10:30
Romsterno but you can throw the git commands but out of build() into unpack()10:30
Romsteror maybe a download function or something or adding git support into pkgmk dunno yet i'll think of some way.10:31
diverseI see, you want to separate the building processes into separate functions in the Pkgfile10:33
Romsterwas really just looking at unpack.10:33
Romstersome times you do not want to unpack a file and other times you got to unpack something differently.10:34
Romsterdon't want feature creap but it could also be used to git svn hg cvs for unpacking into $SRC10:35
Romsterif it's even worth doing that.10:35
diversearch linux does the same thing, but they call it "prepare()" and they have a separate function called "package()"
Romsteryeah that might be better keep build() as is and make a prepare()10:37
diverseif you use prepare() it would make it easier to fork their pkgbuilds to our pkgfiles ;)10:41
Romsterthat was not the intention.10:42
Romsterdo we want that...10:42
diverseI'm just saying as an "added bonus"10:42
Romsteri'm not even sure i like my ideas yet.10:42
Romsterlots of my ideas are too far off KISS10:43
diversewell their nothing wrong with just build()10:44
diverse*there is10:44
*** sudobaal has joined #crux10:47
diversebut yeah, you can also see that arch is no longer following the KISS way anymore with their packaging :P10:47
Romsteryeah arch is bloat.10:50
diverseso I am glad I left it a year ago10:51
diversealthough there is one change I would like to see in pkgmk10:54
Romsteryeah i am anoied at no pkgmk -eo on git but really i should just do it how your doing it now but i'd do it to make the tarball first.10:54
*** Pingax has joined #crux10:55
diverseoh I see10:56
diversebut would there be a failsafe way to download directly from the git branch, just in case if your server is down and the tarball can't be download?11:00
diversejust thinking out of the box here.11:01
diverseor maybe that's not the point here with the source()11:03
Romsterwell if i did put the generate scripts wit the Pkgfile then anyone can remake the tarball if need be.11:03
Romstermy host is rarely down.11:04
Romsterand even fi it is it's also on my home machine.11:04
*** sudobaal has quit IRC11:05
diverseRomster: sounds good11:10
diverseand thanks for the generate script :)11:10
Romsterno problem.11:12
*** sudobaal has joined #crux11:12
*** sudobaal has quit IRC11:23
diverseRomster: maybe I should go with the tarball generating too11:23
Romsterit would save the user having to install git.11:25
diversewell not just that, but save time too. Everytime I make changes to the pkgfile and try to update the package on the same git commit, it has to pull the same files again and have to wait for the downloading process to finish, arg!11:28
Romsteryour doing it wrong if that's the case.11:29
Romstershould just update and take less time.11:29
diversecan you check to see what is wrong with the pkgfile I pastebin you earlier?11:30
Romsterlooks ok at a glance compare it too
diverselooks the same... I wonder what I am doing wrong, hmm11:36
diverseI guess I have a habit of doing prt-get update -fr most of the time11:37
Romsteroh that you got the source so downloading isn't a isue11:38
Romsterit's having to rebuild as pkgmk removes everything and unpacks again.11:38
Romsterinstead of running make again or what ever.11:39
Romstersometimes it's easier to do the stuff manually in the source.11:39
diversewell whenever i rebuild (on the same git commit) it keeps pulling the source from the git repositories11:39
Romsteryes but it takes far less time...11:40
Romsteras it'll see it's already upto date.11:40
Romsterand only downloads new deltas.11:40
diverseI guess the main branch is, but not the submodules that are included in the branch11:40
Romstershould never downlaod everything over again.11:40
Romsteroh sub modules always get pulled in again.11:41
diverseIt's because of the submodules, that's why11:41
diverseits the projects Makefile's fault11:41
diverseso yeah, tarballs ftw11:41
Romsteri once tired too fool a project to downlaod the files for the submodules myself and stuck them in $SRC but then the build system was too stupid to extract them...11:42
diverseyeah, lets keep it KISS and just rely on extracting the premade git tarballs11:44
diverseoh wow I just said "keep it keep it simple stupid"11:45
diversenote to self, just say "keep it simple" in the future, because saying "lets KISS" would be just too awkward to say.11:48
Romsteri once asked what the other S stood for :D11:48
Romsterwhen i had forgot11:48
diverseWell I don't really like the "stupid" part, maybe "keep it seriously simple" is better :)11:50
diverseor "keep it stupid simple" and use it as an adjective11:51
*** dkoby has joined #crux11:54
diverseunless you want to call someone stupid, by all means11:56
Romsterwhere is vee when ya need him :D joking.11:57
diverseyeah I have a tendency to tease the stupid ones.11:57
Romsteri should stop doing that11:57
diverseits kind of hard to stop though11:58
diversethe urge just grows with every second of their stupidity11:59
Romsterhonestly some of them repeat the same problems over and over.11:59
diverseI guess vee caught a serious case of noobitis12:01
diverseand yes that word does exist (on
Romsterit could be heredity.12:02
Romstereverything exists on the internet.12:03
diverseoh man, we should get timcowchip back here, he was the easiest to tease12:06
Romsteri think he gave up.12:06
Romsterhonestly i wasn't too good with crux when i first started either. and i try to give everyone the benefit of the doubt they'll improve.12:07
diversefair enough12:08
Romsterwhat i should do is just ignore those that fail to progress.12:08
Romsterlike soemone told me i wont name who. if i keep helping them they'll never learn for themself. and i'll also add or give up and do something they are more capable of doing.12:09
Romsteri'm no where near the best either. at least i try.12:09
diverseYou know the saying "give a man a fish and he eats for a day; teach a man how to fish and he eats for a lifetime." I think part of the problem is when helping people like vee they just take the fish you gave them instead, while you were showing them how to fish. As a metaphor.12:13
*** Rotwang has quit IRC12:13
Romsteryeah i think i need to stop helping vee12:13
Romsteri don't mind helping but god damn it.12:14
Romstergets frustrating when i have to repeat myself.12:14
diverselets change the topic12:20
diverseso yeah, I want to go with your idea for the ports that rely on git12:20
diverseRomster: so basically take the git extraction from the Pkgfile and put it in the generate script and create the tarball from the $PKGMK_SOURCE_DIR/$name directory?12:28
diversewithout those vars of course, you get my point12:29
*** jdolan has quit IRC12:29
*** joacim has quit IRC12:30
*** joacim has joined #crux12:30
cruxbot[contrib.git/3.0]: opus: force removal of html pages if doxygen is installed12:32
Romsterlike that.12:35
Romsteredit it to suit your needs.12:35
Romsterbut you'll have to do something for the sub-modules.12:36
diverseI got two questions12:47
diverse1. I noticed you used paranthesis as a way to create blocks (like in C with curly brackets) but why is that needed? (I haven't done bash scripting before)12:49
diverse2. You created the tarball inside the work directory, doesn't the work directory along with everything inside it get deleted after the build process finished?12:52
diverseRomster: ^12:53
Romster() are a sub shell.12:56
Romsterlike $() but multiline12:56
Romsterfeel free to remove those.12:56
diverseah, interesting12:56
joacimso i can do stuff in a different working directory without leaving the one for the main script?12:57
Romsterthe tarball is made in ../ of the exported git tree.12:57
Romstercorrect joacim12:58
joacimi was looking for something like that. but by now i've forgotten what for.12:58
Romsterbut error checking is somewhat more of a pain.12:58
cruxbot[contrib.git/3.0]: qca: fixed gcc47 compilation (Thanks to Danny)12:58
cruxbot[contrib.git/3.0]: qca-tls: removed port (obsolete and there are no dependent ports)12:58
joacimi dont know how to make bash scripts, but i still manage to slap some together.12:59
diversewell if you programmed before, its not too hard to figure out, except for some unusual stuff13:00
Romstershell sucks though i need to learn more python and C and stuff. iv'e hit the limitations of shell.13:01
joacimit is more like logic that i've picked up from using gnu/linux for 10 years13:01
Romsterlearned from doing. and hacking other examples up.13:02
diversethats how you learn programming13:03
Romsteri'll never be as good as a real programmer but i'll keep plodding along.13:03
diverseah I didn't pay close attention to line 38, that's why, naruhodo13:05
Romsterthat'll likely fail for you <<13:06
Romsterbut i left it there as an example13:06
Romsterupload tarball to server then pkgmk -um after editing the version-13:07
Romsteror have the script do that trick from Pkgfile.in13:07
*** Pingax has quit IRC13:17
diverseRomster: thanks for sharing that13:18
diverseIf there is one language I want to learn right now, that would Rust.13:23
diverse*would be'13:23
joacimyes. that language would rust13:27
*** jdolan has joined #crux13:29
diverseI don't know why Mozilla called it Rust, but so far it kicks the ass out of C++13:29
tilman"close to (the) metal" maybe13:39
diverseIt is a systems language, so very possible. I had also thought that the name might have been named because the language is inspired by really old but proven paradigms of programming.13:44
diverseYou got the old C and C++ languages which are close to the metal but unsafe, tendency to create bugs, and deminish security at the cost of performance. While you have old (if not older) languages like Lisp, Scheme, ML, which are functional programming languages, and these languages are proven to be perfect (which means less to no bugs, issues, security problems) but at the cost of (extremely) low performance.13:48
Romsternext someone will write a upper later wrapper for rust called oxide.13:48
diverseSo by combining the best of both worlds, you get a language like no other and it will change how we produce code.13:49
diverseRomster: I wouldn't be surprised if that happened13:52
Romsteri was joking.13:56
diversejust by reading the url I was like: "F U micosoft, F U apple!"14:02
Romstercompanies can't play fairly.14:02
diverseyeah... its all about getting more money14:03
Romsterall for the share holders no doubt14:03
diverseand their company14:05
joacimwell. I can't really feel bad for google14:07
joacimthey're playing the same damn game14:07
Romsterbuy patantes use others.14:07
diversethe problem with google is that they are like an NSA214:08
diverseso you got the NSA spying on you and Google on top of that, and who knows what else.14:10
diverseit got quiet of a sudden?14:17
diverseI'm going to pass out now. Thank you Romster for the scripts, it was fun chatting with you and everybody. Night.14:21
Romsternini diverse14:24
Guest39544Romster: cool uh:)14:57
*** Amnesia has quit IRC14:57
*** Amnesia has joined #crux14:57
Romstermore like a major pain14:57
Romsteri have to disinfect computers with this....14:57
Romstergonna have to reflash more BIOS's if it even lets me do that.14:58
Romsteror i'll have to unsolder BIOS chips and reflash them that way.14:58
Romsterseriously not every machine will have a firmware bug.14:59
AmnesiaRomster: communication over ultrasound is pretty awesome:D14:59
Amnesia"if" it exists15:00
Amnesiathe badbios malware that is15:00
Romsterhow is it ultra i doubt many speakers work near 20Khz15:00
Romsterand i can hear to 18KHz at my age. yes i have rare high frequency hearing ability. so if it was doing that i will probably hear it.15:01
Romsterthis could all be a hoax too.15:01
joacimyou can hear the matrix15:01
joacimi doubt a normal pc microphone will be able to pick frequencies like that15:02
joacimi assume they're tuned for the vocal range15:02
Romsterwhich anyone could hear15:04
*** dkoby has quit IRC15:11
*** nrxtx has joined #crux16:18
*** twyf has joined #crux17:06
*** doomicide has joined #crux17:20
*** nthwyatt has joined #crux18:48
*** darfo has quit IRC18:49
*** frinnst has joined #crux18:54
*** doomicide has quit IRC19:09
*** doomicide has joined #crux19:10
*** nthwyatt has quit IRC19:28
*** vee_ has joined #crux19:28
vee_is the temaspeak client being worked on?19:29
*** Rotwang has joined #crux19:59
diverse*yawn* I wake up and come to vee_ asking another indirect question. Use your brain and figure out who maintains teamspeak and them ask directly.20:06
diverseand since teamspeak-client already exists, better to say, "teamspeak needs an update, version 3 is out"20:11
jaegerthat reminds me, I should remove teamspeak2-client from contrib20:12
jaegerHaven't used it in years.20:13
diversejaeger: I'm sure vee_ would want you to replace that with teamspeak3-client20:13
jaegerI'd be happy to pass maintenance of it to someone else, I no longer use it20:14
jaegerThough it would make more sense to simply replace it with a teamspeak 3 port, yes20:15
vee_maybe ill pick that up as my learn how to compile and maintain a port, project20:18
*** frinnst has quit IRC20:19
*** frinnst has joined #crux20:20
teK_I think I'm going to be reviving my full disk encryption initrd for my shiny samsung 840 pro20:23
vee_out of curiosity, why do you encrypt?20:38
tilmanyou don't have news in murrica?20:38
teK_because of vee spying on me!20:39
jaegerMake sure you do some research on that. Unless the SSD has specific support for FDE you'll trash its performance20:42
*** joacim has quit IRC20:43
teK_will do20:43
teK_until now I populated encrypted /home20:43
teK_and apparently I'm too stupid to use sopcast20:44
diversevee, because of the same reasons why you don't want anyone to know you watch ferret porn.20:44
horrorStruckFWIW, on my ssd i just use an encrypted loop device for the things i really dont want anyone to see20:48
teK_that's the point encryption should be done for /var, for example, too20:49
*** jdolan has quit IRC20:57
*** joacim has joined #crux20:59
vee_maybe because i reinstall crux a trillion times i never get a chance to encrypt anything21:05
joacimthinking about doing encryption, but that is almost pointless when my backups arent :p21:15
joacimi guess i should encrypt my backups too21:15
joacimmy uploads on tpb21:21
*** darfo has joined #crux21:26
teK_jaeger: I just read about ATA master PW + Hardware AES crypto in the disk.. not sure to trust that combo :\21:27
teK_but I have yet to discover the performance thingy you mentioned.. do you mean performance degradation because of TRIM-issues on encrypted dsisk>21:28
jaegerwell, a fully-encrypted disk appears to the SSD's controller as 100% filled data21:28
jaegerwhich means no TRIM and no free cells/pages21:28
teK_yeah thought so21:29
teK_but using trim seems risky security-wise, too21:29
teK_meh :-)21:29
joacimwhat if the encrypted data is stored in a sparse bundle?21:30
jaegerIf security is your priority don 't use an SSD21:30
joacimor sparse image21:30
jaegeror use the image suggestion from earlier21:30
jaegeror find an SSD whose controller supports encryption21:30
teK_the 840 Pro apparently does21:30
teK_yet I give control over encryption to _some_ firmware which sucks (yeah I won't be captured by some shady three letter agency anyway...)21:31
jaegerworth a try, then... worst that happens is you get shit performance and kill some write cycles21:31
teK_I have more quality than security issues wrt firmware :P21:31
teK_issues = objections21:32
joacimI guess encryption is more important on a laptop than on a desktop21:32
teK_this is for a laptop21:33
teK_although I do encrypt my /backup (on a HDD), too ;)21:34
teK_quoting OCZ forums:21:35
teK_The ATA password is broke in firmware v1.5. It will possibly be fixed in the next firmware release.21:35
frinnstomg i laughed21:40
jaegerI counter with
teK_dont you dare making fun on game of thrones21:45
jaegerIt's worth it21:45
teK_no time, gotta watch 'the hunt'21:47
teK_btw frinnst do you know when californication will continue?21:47
frinnstno clue, stopped watching after a season or two21:47
teK_last season (5?) was T-H-E BOMB21:47
diverse#include <stdio.h> void dunovny() { printf("mothaaafuckaaa"); }21:54
teK_in falsetto voice :D21:56
teK_and you are missing a newline, and use formatted printing without using formatting =)21:56
*** jdolan has joined #crux21:57
diverse1. fine, sed s/a"/a\n"/ 2. I don't care, at least it prints.21:58
teK_software bloat!21:59
tilmanyeah, no21:59
teK_use write!22:00
tilmangcc is smart about printf22:00
diverseoh what by like 10 kilobytes of bloat, hah.22:00
tilmanlet it optimize and have a look at the machine code22:00
tilmangcc will turn printf("x\n") into puts("x") -- so you'll still use stdio22:02
teK_quoting dietlibc:22:02
teK_lib/__v_printf.c:link_warning("__v_printf","warning: the printf functions add several kilobytes of bloat.")22:02
diverselike I said "10 kilobytes, hah!"22:03
diverseand thats pretty minimal compared to some languages with a bloated runtime. (Yes, I am looking you, the D programming language!)22:04
nrxtx -> Static hello22:04
teK_what have I done :>22:05
teK_the tables was compiled by the musl authors, right?22:05
nrxtxsure that's why musl shows up at best22:06
teK_dangerously nonconformant22:06
teK_fuck utf-8!122:06
teK_sounds like a marketing pitch22:07
teK_Attention to corner cases22:07
diversemusl: its dangeriously nonconformant ;) /slogan22:08
diversepepsi meth: its dangeriously good!22:11
nrxtxat least the static hello world size seems to be true:
nrxtxwoah, gcc static hello world is 800kb22:12
diverseis that because of musl or how bloated glibc is in general?22:14
viabut who cares when the libc is going to be dynamically loaded and the same copy of code shared across all programs that use libc22:14
diversetrue, why static compile and bring whole C runtime in, when you can dynamically link to shared objects and shrink the binary size22:17
nrxtxvia: speed and maintainable clean code, is the only reason why people are using other libs22:18
viamaintainable clean code sure, but i would very hesitant to say speed when you might be speeding up some cases but really slowing down others22:18
viamy line is mostly towards the style static-linking-is-better crowd that seems completely asinine to me22:19
tilmannrxtx: hum? i think embedded folks care about the size =)22:19
nrxtxtilman: sure22:19
viaembedded deinetely makes sense22:19
nrxtxbut on desktop most don't care22:19
tilmanyou said 'only'22:19
nrxtxsed 's/only/possible/'22:20
nrxtxone big problem with glibc i see is that they maintain buggy behaviour with all it's use cases22:21
nrxtxjust to not break progrums relying on glibc bugs22:21
nrxtxand that ends up in a lot overhead22:21
viadiversity is always a good thing22:22
viait promotes the use of portable code22:22
viaso it would seem22:23
nrxtxvia: hopefully, the people around the musl-based distributions patch up everything22:23
nrxtxquestion is if the patches will be accepted upstream22:23
nrxtxThere is crux Oo22:25
nrxtxdon't know if it seems to be that easy :D22:25
tilmanseeing bedrock linux in that list...22:26
tilmanthat shit is crazy22:26
tilmanchek it out22:26
diverseI guess would it be good to replace glibc with musl? I don't think any sane project would use glibcisms in their project.22:28
nrxtxdiverse: replacing it with any other c-library can be interesting, but it will break a lot of packages/ports :D22:30
diversedamn you RMS!22:31
tilmanyou know nothing22:31
teK_diverse: you like RMS? :)22:32
teK_ <<<22:33
diverseteK_: I give him credit for starting the "free software" initiative, but sometimes he gets out of control with this ideas.22:34
nrxtxtilman: did you already try bedrock?22:36
tilmannrxtx: nah22:38
diverseteK_: plus I said it out as tongue-and-cheek, because the humongous influence he has.22:39
nrxtxtilman: are you working with embedded linux?22:44
diversemusl benchmarks look pretty good22:44
nrxtxdiverse: except it is not known how much you can trust the results from the etalabs table22:45
nrxtxsince they are the ones developing musl22:45
diverseso it could be bias perhaps22:46
nrxtxi made some tests on my own but not that detailed22:47
nrxtxcreated a port for testing it without interfering with the default c-library installed22:49
nrxtxbut they miss some functionality which i need for preempt_rt, so did not use it further22:50
diverseare you using lasers for something?22:55
nrxtxdiverse: no :D22:57
diverseI guess Real-Time has a lot of uses23:00
nrxtxdiverse: worked a lot with embedded linux in the last years, especially in hard-realtime systems23:00
nrxtxtried to use musl there because it was a lot smaller than glibc23:01
nrxtxbut musl prevents scheduler usage, so is useless in this case23:01
diversethat sucks...23:02
nrxtxaccording to the developers they first want to test it a lot more, before they enable it23:03
nrxtxhm that reminds me of creating a rt-tests ports23:05
diversepreempt_rt is looking pretty badass23:07
diversethe more I read about it23:08
nrxtxone interesting project which has its origin there is kernelshark23:10
nrxtxso in case your crux systems seems to be slow you can exactly see why and who uses your cpu: :)23:12
diverseI can see why they call it kernelshark.23:15
tilmannrxtx: i'm not23:18
diversenrxtx: do you use the preempt_rt patch on your desktop machine?23:20
nrxtxdiverse: no, it will result to problems with nvidia binary drivers23:21
nrxtxthose are not preemptive23:21
tilmannrxtx: wonder how kernelshark/trace-cmd compares or relates to perf tools23:23
*** twyf has quit IRC23:23
nrxtxtilman: perf just collects internal counters23:26
nrxtxtrace-cmd uses the kernel ftrace framework and can track any internal function23:26
nrxtxwith timestamps and details23:26
diverseso kernelshark is like wireshark expect you look into kernel tasks in more depth, instead of network packets?23:27
nrxtxdiverse: yes23:27
diversethat would be another great tool for cyber security23:30
tilmannrxtx: i thought perf let you install your own trace points too. but i might have gotten it wrong; never tried that23:30
nrxtx the introduction describes it best23:31
nrxtxcorrect mandir on crux is /usr/man ?23:38
diversewe keep it simple like that23:39
cruxbot[contrib.git/3.0]: qca-ossl: initial import23:46
cruxbot[contrib.git/3.0]: qjson: romster -> contrib23:46
cruxbot[contrib.git/3.0]: protobuf: romster -> contrib23:46
cruxbot[contrib.git/3.0]: clementine: romster -> contrib23:46
diverseRomster: giving your repo a good cleaning? :)23:47
Romsterstarting too.23:47
Romsterbeen using clementine for awhile now ad quite like it.23:48
nrxtxnow i can remove at least one port from my local ones :D23:51
Romsterwhich one is that?23:51
Romsterah been in my rosmter repo for ages.23:52
Romsterhonestly think i'm nuts when it comes to packages. -_-23:52
diverseRomster's repo is like contrib2, so you can't go wrong adding his repo to your collection23:53
nrxtxRomster: but old version, did you update when moving to contrib?23:53
Romsternot everything works in romster23:53
Romsternah 2.5.0 brakes clementine23:54
Romsteri tried.23:54
Romsterfound no patches as of yet.23:54
Romsterif you want to help and find a patch for clementine. i'm working this weekend.23:54
*** doomicide has quit IRC23:55
Romsterat least i think it did. i do so much it's hard to keep track of everything.23:56
nrxtxRomster: i'm not in hurry on this one23:56
Romsteri would prefer stable working over bleeding edge and possibly broken.23:56
nrxtxso it should work at least23:58
nrxtxanyway interesting player23:58
Romsterhas an android remote app too23:58
Romsterbeen using that as well.23:58
Romsterpretty heavy on dependencies though.23:59
*** Rotwang has quit IRC23:59

Generated by 2.11.0 by Marius Gedminas - find it at!