IRC Logs for #crux Thursday, 2014-01-30

*** deus_ex has quit IRC00:02
pshevtsovHello! Is there any way to prevent pkgmk to compress man pages?00:17
*** deus_ex has joined #crux00:18
*** deus_ex has quit IRC00:22
*** deus_ex has joined #crux00:26
prologicpshevtsov: I don't believe so -- Is it even pkgmk that's doing it?00:48
prologicpkgmk afaik is a bash script00:48
prologicso you could just modify it to suit your needs00:48
prologicperhaps if you do, provide a cli option and send us a patch00:48
*** BitPuffin has quit IRC00:51
*** BitPuffin has joined #crux00:53
pshevtsovprologic: Hi! Yes, patching pkgmk is good, but that's not what I actualy want as a result. I want to make some port with uncompressed man pages00:54
prologiclike I said01:09
prologicpatch pkgmk :)01:09
prologicadd in the option to do so01:09
prologicI'm not 100% certain what compresses the resulting man pages myself01:09
prologicit's never really been an issue for any of us01:10
prologicwhat's your use-case for an uncompress man page?01:10
*** jdolan_ has joined #crux01:13
*** jdolan has quit IRC01:13
prologicThere ya go :)01:14
prologicThanks jaeger !!!01:14
prologicno way to configure this currently01:14
prologicsend us an appropriate patch to pkgmk01:15
prologicmaybe --no-compressed-man-pages01:15
prologicseems a bit long :)01:15
pshevtsovprologic: send where? flyspry?01:15
jaegerI'm curious about the use case as well, something for really slow hardware?01:16
prologicyes flyspray is appropriate01:16
prologicyes what's the use-case here?01:17
prologicI find this hardly a problem (if at all)01:17
pshevtsovnode package manager (npm) uses man pages for its internal help (npm help something), but it doesn't know how to work with compressed man pages. And I don't want to patch npm, it's easier to patch pkgmk at all01:19
*** deus_ex has quit IRC01:20
pshevtsovs/at all/et al/ :)01:20
prologicopps wrong window :)01:21
prologicpshevtsov:  can I make a bold suggestion?01:21
prologicplease don't be offended01:21
prologicI will anyway :)01:22
pshevtsovsure! go on01:22
prologicThis is kind of out of scope for CRUX's pkgutils01:22
prologicReally a feature request should be filed with nom upstream01:22
prologicnom should understand how to decompress files it's meant to read01:22
jaegerIt's not a bad idea for us to consider, though... similar to nostrip or choice of package compression01:22
prologicI probably (I'm not a CRUX core dev) would not accept such a patch in this case01:22
prologicjaeger: I agree01:23
prologicit's a small enough patch/change ihmo01:23
prologicassuming we can find a nice option for it that's shorter and saner than what I proposed ;)01:23
prologicIn general though I still do think it's "npm's problem"01:24
prologicAlso pshevtsov ...01:24
prologicyou could just as easily01:24
prologicfind /usr/man -type f -name "*.gz" -exec gunzip {} +01:24
prologicJust saying :)01:24
jaegersomething like PKGMK_COMPRESS_MANPAGES="" and -ncm01:25
prologicthat could work01:27
prologicread the option out of /etc/pkgmk.conf instead01:27
prologicmaybe you're right01:27
prologicthat sounds not too bad01:27
*** sepen has quit IRC01:29
*** doomicide has quit IRC01:32
prologicSo I guess we're all +1 for a patch :)01:35
*** deus_ex has joined #crux01:35
prologicat least two of us :)01:35
*** deus_ex has quit IRC01:54
*** deus_ex has joined #crux02:08
*** deus_ex has quit IRC02:13
*** deus_ex has joined #crux02:28
*** deus_ex has quit IRC02:34
*** Workster has joined #crux02:43
Worksterlooking at the log and i decided to add my 2 cents worth. to compressed man pages.02:44
WorksterI would onlu say +1 if we had a .nocompressman file wile on content just having the file causes pkgmk to not compress the man pages02:45
jaegerWhy that over the pkgmk.conf option?02:46
Worksterbecause it can be set only in the ports that require it like node package manager (npm)02:47
Worksterand not system wide as in pkgmk.conf02:47
pshevtsovI agree with Workster02:48
*** deus_ex has joined #crux02:48
Worksteri would prefer if node package manager (npm) got smarter with handling itself or use a post-install script for that port with what prologic suggested to decompress the man pages02:48
Worksterand my second point is we should probably look at xz for man pages for crux 3.1 i'll add a feature request when i get home on that.02:49
Workstergroff has had support for xz for quite some time now02:50
Worksterreally though i'd be seeing if node package manager (npm) can be patched.02:51
Worksternever seen this to be a problem before.02:51
Workstertroll gentoo/arch/slackware for node package manager (npm) and see how they deal with it pshevtsov02:51
Worksteron there package pages02:52
Worksteri'd prefer fixing the port sources some how than some extra option for one port to use.02:53
Worksterbtw pshevtsov i'm Romster just at work though :)02:55
*** deus_ex has quit IRC02:58
prologicI'm -1 on the idea of .nocompress in the port's directory02:59
prologicthat adds extra complexity to pkgmk that doesn't buy us much when a simple optional -ncm/--no-compress-man would suffice03:00
Worksterthat would require a README for npm port then03:00
prologicbut upstream npm* __shoudl__ still be responsible fox fixing this and making npm deal with "compressed man pages"03:00
prologicit's not like it's an unusual thing03:00
prologicWorkster: that's perfectly acceptable03:01
Worksterand if ya going that far why do not add that option at all and have the user decompress the required man pages in post-install03:01
prologica lot of ports do contain README(s) for that very reason03:01
prologicthat's kinda what I said as well03:01
prologicbasically in short - it's kinda not our problem (crux)03:02
prologicbut I think we'd be okay with an option to pkgmk03:02
Worksterbut before all this i would lke to look at the source and see if there is a way to do this without another pkgmk option.03:02
prologicthe source of nom?03:02
Worksterthis is called feature creep. one option for one port seems hardly worthwhile.03:03
Worksterpshevtsov, this is it?
*** tilman has quit IRC03:03
*** tilman has joined #crux03:05
prologicI couldn't agree more03:07
prologicfeature creep is what kills a lot of projects03:07
Worksteri honestly don't like the idea of a no compress man page option at all and deal with this in the npm source03:07
prologicneither do i03:08
prologicI think it's silly not to compress them03:08
prologicwe've done so for years and years03:08
prologicprobably ever since gruff had support for reading compresses man pages03:08
Worksternpm looks like java and shell code to me i was expecting C or C++03:09
prologicit's written in javascript03:10
prologicnom runs on nodejs03:10
prologicthat (nodes) is written in C/C++03:10
prologicand uses the v8 javascript engine03:10
Worksterthat's all ugly stuff.03:10
prologiceverything nodejs is written in javascript03:10
prologicwith possible C/C++ extnesions03:10
Worksterbut surely patchable to allow it to decompress man pages03:10
prologicjust like in the Python land :)03:10
Worksteryeah but python is far more elegant03:11
prologicI would suspect so03:11
prologicthere are compressions libraries for Javascript/NodeJS03:11
prologicjust be a matter of installing, importing03:11
prologicand patching npm appropareily03:11
prologicyes I agree Pythoni is far more elegant :)03:11
*** deus_ex has joined #crux03:13
WorksterI'm hardly a java or javascript or python dev but I've hacked at code on all these systems and I prefer python03:14
Worksteranyways i'd look at the npm source even submit a feature request to there bug tracker03:15
Worksterpshevtsov, is it npm itself or nodejs that can't handle compressed man pages?03:17
*** deus_ex has quit IRC03:19
*** cryptorchild has quit IRC03:19
*** cryptorchild has joined #crux03:20
*** Feigr has joined #crux03:26
*** deus_ex has joined #crux03:33
prologicWorkster: it would be npm -- nodes is just the interpreter/compiler ala python/cpython03:40
*** mavrick61 has quit IRC03:41
*** mavrick61 has joined #crux03:42
*** horrorStruck has joined #crux03:42
*** horrorSt1uck has quit IRC03:45
Worksterso byte code compiler JIT03:47
prologicthat's one of the reasons why nodejs is quite popular lately04:01
prologicit performs really well for web applications and I/O bound work04:01
prologicnamely because of it's inherent async I/O builtin to the core langauge/runtime features04:01
prologicand v8 jit compiler04:01
prologicPython has this in the form of (3rd-party libraires and an upcoming new async I/O standard) and PyPy (JIT Compiler / Interpreter)04:02
*** jdolan_ has quit IRC04:02
*** jdolan has joined #crux04:02
*** deus_ex has quit IRC04:03
*** jdolan has quit IRC04:07
*** deus_ex has joined #crux04:20
*** jdolan has joined #crux04:32
*** SiFuh has quit IRC04:53
*** SiFuh has joined #crux04:53
Worksterprologic, your circuits framework would greatly improve with async i/o?05:14
Worksteror is it pretty much omgoptimised now05:15
prologicit's almost as optimized as it gets - yes05:17
prologicstandard async I/O in Python will not change circuits05:17
prologicwe will at some stage just rewrite or write additional Async I/O components that wrap the standard python ones05:17
prologicAsync I/O in circuits is not what circuits is :)05:17
prologicthey are just "components"05:18
prologicyou can very easily plug them in/out with any other compatible async I/O (we call them pollers) component(s)05:18
prologice.g: wrap libevent, libev, threading, blah blah05:18
prologicthe ones we have circuits.core.pollers are a directly borrowed implementation from Twisted to suit the circuits message bus05:19
Worksterso in theory if you got bored you could write other async components and test the performance with some benchmark tool05:20
Worksternot saying any one needs too05:20
prologicyes absoltuely05:21
prologicnot in theory - in practice05:21
prologicI guess we never have because we find the performance "good enough"05:21
prologicI run many sites on circuits.web05:21
prologicin production05:21
prologic-and- circuits.web has been (RAW HTTP/1.1) been benchmarked at 7k req/s at a concurrency of 10k clients05:22
prologicon a 8-core BSD server with kernel tweaks05:22
prologicI'm quite surprised you know about it :)05:23
prologicwhere did you hear about my little framework? :)05:23
Worksterlol ages ago...05:23
prologicdo you use it at all?05:23
prologic <-- known community and subset of known projects and applications using circuits05:24
*** jdolan has quit IRC05:25
Worksteri've toyed with it in the past and i did have it though mod_python then mod_wsgi on apache ages ago. but not currently no. but i want to use it on versionsort.com05:46
Worksteri'messing with the sql database for it and getting the back end to store results currently.05:47
Worksterbut at some point circuits probably mako templating and either lighttpd or apache for the httpd server05:48
Worksterdunno on sqlite3 if it's good enough but i'll start on that befoe moving to postgresql or something if needed05:48
Worksterneed a example index page for quering a sql database and showing results using circuits05:49
*** xvee has joined #crux05:52
xveegood day everyone05:52
Worksterprologic, maybe unrelated to my current version sort project, i'll have a use for circuits for automation datalogging a off the grid home that my friends is doing and wants a small computer to control heating/cooling, generator, and other such stuff.05:53
Worksterhi xvee pshevtsov05:53
pshevtsovSo, I'll try patch npm and push my changes upstream, or just provide patch with my port06:00
Worksterwell geez i don't think anyone uses uncompressed man pages anymore06:10
Worksterjust patch the port and add a # comment as upstream wont fix this issue06:11
Worksteri'm sure one day some other distro will see it or some user and pester them to fix it.06:11
Worksterif you cna't decompress you could add a /usr/share/<foo>/help/ that's not compressed and change the path to there for help.06:12
Worksterand leave the man pages as is.06:12
*** BitPuffin has quit IRC06:27
*** Workster has quit IRC06:31
*** BitPuffin has joined #crux06:53
*** BitPuffin has quit IRC06:58
*** Rotwang has joined #crux07:54
*** xvee has quit IRC07:59
*** Romster has joined #crux08:05
*** Rotwang has quit IRC08:34
*** BitPuffin has joined #crux08:50
*** horrorSt1uck has joined #crux08:57
*** horrorStruck has quit IRC08:58
Romsterprologic, how dows this database work? I am following
Romstersince the manual page on doens't seem to give a basic example.09:41
Romsteri have to learn joins after i get this working, starting with 2 tables.09:42
*** deus_ex has quit IRC09:44
Romsteropps curly brackets start missing..09:44
*** BitPuffin has quit IRC09:45
Romstermetadata = BoundMetaData(db)09:45
RomsterNameError: name 'BoundMetaData' is not defined09:45
Romsteri can't see how that is not defined.09:45
*** BitPuffin has joined #crux09:55
Romsterdamn it get tired when i mess with my own stuff.09:59
Romstermore coffee time09:59
*** deus_ex has joined #crux09:59
*** BitPuffin has quit IRC10:04
*** BitPuffin has joined #crux10:31
RomsterBoundMetaData is deprecated and replaced with MetaData10:34
Romsterworks now10:38
frinnstwhat to eat, what to eat..10:48
Romsterflaming cows?10:48
frinnstaustralian speciality?10:48
frinnstdont think we have that one here :p10:49
Romsteri thought for loops in python made stuff on new lines. this is all on one line.10:49
*** BitPuffin has quit IRC10:57
*** BitPuffin has joined #crux10:59
Romsterprint 'File:', row.file, 'Version:', row.version,11:29
Romsteri get many lines11:29
Romsterprint 'File:', row.file, 'Version:', row.version, '\r',11:29
Romsteri only get one line.11:29
Romsterthis does not make any sense to me.11:29
Romster\r is new lie right?11:30
Romster]n works better.11:34
Romster\n for that matter...11:34
Romstergawd i'm so rusty at this.11:35
*** BitPuffin has quit IRC11:38
*** BitPuffin has joined #crux12:10
prologicanyone have a clue as to why my ipv6 (over a 6-to-4 tunnel to isn't working from my desktop with a bridge? it works from my router and used to work just fine without a bridge -- I'm briding so qemu vms on my desktop can gain access to my local network without nat12:42
*** BitPuffin has quit IRC12:44
*** jdolan has joined #crux12:46
*** deus_ex has quit IRC12:55
*** BitPuffin has joined #crux12:57
*** deus_ex has joined #crux13:10
*** c0ck4m0u53 has joined #crux13:12
*** jdolan has quit IRC13:34
*** jdolan has joined #crux13:35
*** BitPuffin has quit IRC13:35
*** BitPuffin has joined #crux13:36
*** jdolan has quit IRC13:39
Romsterno idea sorry13:41
*** jdolan has joined #crux13:43
*** Romster has quit IRC14:00
prologiccould I get anyone here interested in trying out my docker-bin port?14:04
prologicyou will have to quite possibly reconfiure your kernel hwoever14:04
prologic$ docker run -i -t prologic/crux crux14:15
prologicCRUX version 3.0 x86_64-multilib14:15
prologiccrux within crux via docker container14:15
*** nrxtx has quit IRC14:22
*** nullmark has quit IRC14:23
*** cybin has quit IRC14:23
*** nrxtx has joined #crux14:23
*** teK_ has joined #crux14:24
*** nullmark has joined #crux14:24
*** cybin has joined #crux14:25
rmullprologic: hmm, what is docker? does it use kvm?14:52
prologicno lxc14:56
*** c0x` has joined #crux15:05
*** c0x has quit IRC15:06
*** deus_ex has quit IRC15:18
*** deus_ex has joined #crux15:34
*** c0x has joined #crux15:37
*** c0x` has quit IRC15:38
*** deus_ex has quit IRC15:49
*** c0x has quit IRC15:50
*** doomicide has joined #crux15:56
*** c0x has joined #crux15:56
*** deus_ex has joined #crux16:04
*** c0x` has joined #crux16:05
*** c0x has quit IRC16:06
*** syncn has joined #crux16:14
*** pshevtsov has quit IRC16:22
*** deus_ex has quit IRC16:30
*** deus_ex has joined #crux16:45
*** doomicide has quit IRC16:58
*** doomicide has joined #crux17:00
cruxbot[core.git/3.0]: openssh: updated to 6.5p117:47
cruxbot[xorg.git/3.0]: xkeyboard-config: updated to 2.1118:38
*** lasso has joined #crux19:21
*** deus_ex has quit IRC19:30
*** BitPuffin has quit IRC19:37
*** c0ck4m0u53 has quit IRC19:39
*** chris2 has quit IRC19:41
*** deus_ex has joined #crux19:47
*** chris2 has joined #crux19:47
*** Rotwang has joined #crux20:23
*** lasso has quit IRC20:24
teK_btw.. anybody in here using openvswitchd?21:04
teK_I created a port and am thinking about putting it to life in contrib21:05
jaegerI've not used it myself but it's a neat project, certainly21:07
teK_I needed it for port mirroring in a virtualised full packet capture scenario.. despite the strange syntax it worked fine21:08
*** BitPuffin has joined #crux21:24
prologicteK_:  I'd like to have a play with it21:41
prologicI wouldn't mind playing with CRUX as a Router/Switch21:41
teK_I can push it later21:41
prologicsweet ping me when you do21:42
prologicI'll try it out21:42
teK_last thing to decide is for the name of the script in rc.d ;)21:43
Rotwanghow is it going "crux"?21:49
RotwangI'm going on a 3 day weekend, and I'm supper happy about that21:50
teK_have a great time21:51
cruxbot[contrib.git/3.0]: openvswitch: initial import21:53
teK_prologic: have fun ---------------^21:53
*** rexich has joined #crux21:55
prologicteK_:  thanks!21:59
teK_caring is sharing :>22:00
*** BitPuffin has quit IRC22:19
rexichhello guys, what's up?22:22
*** pshevtsov has joined #crux22:39
*** jdolan has quit IRC22:51
*** deus_ex has quit IRC22:54
*** doomicide has quit IRC22:56
cruxbot[contrib.git/3.0]: openvswitch: put the start script in the correct folder23:13
teK_prologic: have even more fun --^23:13
*** Rotwang has quit IRC23:15
prologicteK_:  haha thanks :)23:28

Generated by 2.11.0 by Marius Gedminas - find it at!