IRC Logs for #crux Wednesday, 2015-02-18

koribug spotted in the go package! godoc points to /usr/local/go when it should be /usr/lib/go00:01
*** toriso has quit IRC00:02
koriwait. I just read and the README00:02
koriapparently export GOROOT=/usr/lib/go is a thing.00:03
diversekori: plenty of languages have their own exclusive env vars that they use to make shit work00:04
korithis confused me because on arch this wasn't a thing00:04
koriI just realized I should probably not really talk about bugs here at all and just talk to an empty channel since I'll end up realizing I'm the bug00:05
diverseno it's fine00:05
prologicI maintain the go port00:05
prologicso if there's a bug let me know :)00:05
koriokay now this is weird00:06
kori$ godoc http00:06
kori2015/02/17 21:05:43 open /usr/lib/go/src/pkg/http: no such file or directory00:06
diversealso keep in mind kori, arch pkg maintainers suck at doing packaging right sometimes00:06
prologicwouldn't you have to00:07
prologicgo get http00:07
korioh, once again, it's pebkac00:07
koriit's net/hhtp00:07
prologicI know nothing about go really00:07
prologicyou're funny :)00:07
koridiverse had an interesting theory that everyone else should just talk about other things while I try to solve bugs and I'll solve them myself00:07
koritime and time again, that's proven00:08
diverseprologic: docker is written in go, since you have an affair with it all of the time00:08
*** dougl has quit IRC00:08
prologicwut? :)00:09
diverseI was being facetious about your love of docker00:09
koriI read this interesting interview with andrew gerrand, one of the main developers00:10
korihe said the key to learning any programming language is to write things in it, and only it00:10
korias in, if you were to write some script, use the programming language you want to learn instead!00:10
koriso I'm doing just that00:10
diversethat should be obvious, use the language you like most00:11
*** crash_ has quit IRC00:12
prologicPython ftw!00:12
diversekori: also another pattern I notice is that you need to have the problem written out by yourself in order to see the answer00:13
diverseyou need to talk about the problem out loud to solve it00:14
koridiverse: right after I posted that command I just went into a ls rampage to see what was in /usr/lib/go/src/pkg/ and there was no http, but all the other things from stdlib were there00:14
koriand so I looked at my code to double check and it said net/http00:14
diversesigh, don't beat yourself about it, just do what works for yourself, being able to solve your own problems is a valuable skill, everybody has their own quirks to do that.00:22
korioh I'm not beating myself over it, I just think it's silly every time it happens, and it seems to happen semi-frequently00:23
korialso, programming is hard.00:23
diverse(that's beating yourself up about it)00:23
diversekori: look, see it as a skill rather than as a problem and use it00:29
diverseit works00:29
diverseand programming is a complicated discipline, there are endless of views to it, but mostly importantly, learn what you want learn, let your interest be your drive.00:39
diverseprologic: wouldn't you agree?00:40
prologicsure why not :)00:42
diversesounds like you have more to offer, but I think I already know what that is.00:45
Workstergreduan, make on it's own calls make all00:45
*** fengshaun has joined #crux00:45
greduanWorkster: I know :)00:46
thetornainbowWorkster: that was me, i couldn't see the makefile - sometimes you can't assume :)00:49
Worksteri've never seen an instance of SRC being clobbered in a Makefile00:50
greduanWorkster: Me neither, I think I'll report it to the author00:51
Worksterbut i have seen instances of CFLAGS doing that. if you explicitly set it on make CFLAGS=00:51
Workstersed 's/SRC/MSRC/'00:52
Worksteror something00:52
Workstersince it's only happed just now on only 1 project i don't see a need for us to change our $SRC00:52
greduanWorkster: agree00:53
greduanactually it's abduco and dvtm but they go hand in hand and the same author00:54
diversehmm, well I guess running pkgmk runs a separate bash process and within that bash process, several instances within that process can influence the environment inside00:55
Workstershame we can't set SRC and PKG as a read only variable and not allow the shell to redefine those00:56
diversemaybe there is a way to make them immutable in bash?00:56
Worksterperhaps pkgmk needs to check for any variable that could get clobbered in the source but that's not KISS00:57
diverserather, there has to be a way00:57
Worksteri'd look at man set00:57
greduanall right that worked00:59
greduansed -i "s/SRC/SSRC/g" Makefile00:59
Worksterthrow it in the Pkgfile job done01:01
Worksteralso watch out for subdirectories wth Makefiles in them01:02
thetornainbowand watch out for the /usr/share/* and licensing fluff :)01:03
greduanthetornainbow: yep01:03
*** tilman has quit IRC01:03
diverseWorkster: alright I found it, both $PKG and $SRC need to be declared with `readonly` in pkgmk01:04
*** tilman has joined #crux01:05
diverseand perhaps the same can be done for $PKGMK_{SOURCE,WORK}_DIR01:09
diverseand so whenever a Makefile decides to try changing those vars, the build will fail warning the user, that way they know they need to sed it01:11
greduandiverse: nice :)01:19
greduanso pkgmk will get an update soon?01:31
greduanI'm learning that you really gotta study the Makefile with every port you make, cause they may or may not be using what you may consider common (like with the variables and such) so you gotta update the Pkgfile accordingly01:33
thetornainbowyeah stuff in the wild can be.... wild :)01:33
korithetornainbow: true facts01:33
greduansurprised tinywm isn't in the ports01:51
prologicgreduan> so pkgmk will get an update soon? <-- I seriously doubt it02:05
prologiconly update we've seriously talked about atm is adding support for anchors so we can give sources a filename02:05
prologice.g: github release tarballs02:06
diverseit's only something Romster and I came up with, it hasn't been discuss with the devs yet02:09
greduanI see02:20
greduanI was saying in order to do the readonly stuff02:20
greduanlooks like a painless change02:21
diverseit is a painless change, but the devs need to consider it first02:29
diverseimo, if a variable doesn't need to be changed at all, it should be readonly by default02:37
diversenot just for this scenario, but for programming in general, especially in large projects02:38
prologicsorry I still don't get it02:48
prologicwhat's this about readonly?02:48
diverseprologic: scroll up02:53
prologichmm I see02:55
diverselong story short, pkgmk's env vars can potentially be mutated from the Makefile. To prevent this from happening, make them 'readonly' as a safe guard and a way alert the user when you need to change something in the Makefile so it doesn't conflict02:58
diversesigh, see? This is why it's easier to have you read the logs than have to paraphrase 50 lines of conversation.02:59
greduanWhether this is put into pkgmk or not, I think a wiki entry should be made for this. I think I will03:04
greduanI'd really like to improve the CRUX wiki03:04
greduanwon't reach Arch Linux levels but still03:04
prologicI probably wouldn't do this necessarily03:10
prologicsince it could break builds03:10
prologicit's trivial for such broken upstream software/builds to preserve the mutated env in the Pkgfile's build()/03:10
*** kori has quit IRC03:13
*** kori has joined #crux03:15
*** kori has quit IRC03:15
*** kori has joined #crux03:15
diversethat's an alternative to sed'ing the Makefile03:16
greduanI reported the author of abduco and dvtm, btw03:23
greduanreported to the author03:23
greduanbig difference there03:23
prologicgood idea03:30
diversehopefully they won't reject the report03:32
greduanthe author probably hasn't read it so I can change anything if you guys want03:32
*** mavrick61 has quit IRC03:33
*** mavrick61 has joined #crux03:34
rmullgreduan: I don't really feel like this is upstream's problem03:34
rmullWhy should they tailor their vars to accommodate crux's unique needs?03:35
greduanrmull: good point03:35
greduanI changed the last line from "could you fix this?" (paraphrase) to "thought you should know"03:36
jaegerwe should possibly consider using a more unique var internally03:36
greduanjaeger: that's another solution, although that would break every port out there03:37
prologicI think this is a unique case03:38
prologicnormally auto* and cmake and othe rbuild tools don't pose a problem03:39
prologicbut this clearly hand crafted Makefile does03:39
diversejaeger: advocating $PKGMK_{SRC,PKG}03:39
prologicstill I *think* the port maintainer should just write a bit extra bash code :)03:39
greduanprologic: I don't mind doing that, myself03:39
greduanbut avoiding issues is always good. :)03:39
jaegerjust a thought, it would be a lot of work03:40
greduanwe could also just use $SRC and $PKGMK_{SRC,PKG}, for backwards compatibility03:40
greduanand any new ports or updated ports switch to $PKGMK_{SRC,PKG}03:40
prologicI'm with jeager03:40
prologiclot of work03:40
prologicfor little gain :)03:40
prologicalso is it not trivial to spwan "make" in a subprocess inside build()?03:41
prologice.g: (make)03:41
prologicpreserving the parent's env03:41
prologicit's called using a subshell03:42
prologicwarp make in ( )03:42
prologicno need to sed or OLD_SRC magic03:42
greduanprologic: also an option03:43
prologicnot just an option a very good damn point :)03:43
diverseif you can contain the make process, that's the best option imo03:43
prologicmakes any modification to pkgmk and your issue reported upstream kinda redundant :)03:43
prologicno offense :)03:43
prologicdiverse: yeah you can :)03:44
prologicnot sure why I *just* thought of it just then03:44
rmullWell, some Makefiles /depend/ on the ability to append env vars, usually done with CFLAGS for example03:44
prologicbut *meh*03:44
prologicneed more coffee :)03:44
diverseprologic: I ask myself that very often as well03:44
prologicrmull: good point -- but there's almost no use-case I can think of where modifications to CFLAGS needs to outlive the "build"/"make" process03:45
prologicso you'd just warp the offending command in a subshell (not everything)03:45
diversethe subshell basically copies everything over to begin with03:46
diverse(I assume)03:46
greduanImma go to sleep now03:49
greduanTomorrow somebody update me on the decision. ;P in any case I don't mind making the Pkgfile work around it so if no changes are made that's fine. readonly should totally be a thing though ;)03:50
*** greduan has quit IRC03:50
prologicdiverse: it would fokr()03:52
diversegood, then no worries03:52
prologicgreduan: use a subshell ( )03:52
diverseyou mean `$(make)`?03:55
*** SiFuh has quit IRC03:56
prologic( make )03:59
*** SiFuh has joined #crux04:16
*** SiFuh has quit IRC04:26
*** SiFuh has joined #crux04:31
Worksterisn't $() and `` also a subshell as ()04:34
*** SiFuh has quit IRC04:36
*** SiFuh has joined #crux04:38
prologicgood question04:48
prologic`foo`` defiantely is04:48
prologicand you're proably right about $()04:48
prologicwhat does the $ denoate?04:48
*** SiFuh has quit IRC04:51
prologicI guess is returns the value04:57
prologicwhich is only really useful if you want to capture the output I guess04:57
prologicso () `` or $() are probably all valid04:57
prologicbut ( ) makes more sense in this context I think04:57
diversedo we need to return anything at all?05:13
prologicin this case; no05:15
diverseas I thought05:15
*** SiFuh has joined #crux05:17
*** SiFuh has quit IRC05:22
*** SiFuh has joined #crux05:24
*** SiFuh has quit IRC05:31
*** SiFuh has joined #crux05:33
koridiverse: AHA06:05
koriI've come up with a name for this thing I do06:05
kori"rambling driven development"06:06
korigotta buy a rubber duck06:08
diverserubber duck debugging huh? That does seem very appropriate for you06:11
*** toriso has joined #crux06:13
diverseI suppose for now on, I'll ask: "Did you self review" or "Did you talk to the duck?" when you have problems06:25
koriI like "Did you talk to the duck?"06:26
*** arduo has joined #crux07:20
*** arduo has quit IRC07:23
*** arduo has joined #crux07:25
*** arduo has quit IRC07:39
Romsterduck debugging okay didn't now that.07:39
Romsterprologic, wouldn't a subshell also mess up output to stdout stderr for prt-get logging.07:41
Romstermaybe not07:44
*** arduo has joined #crux08:02
*** arduo has quit IRC08:02
*** arduo has joined #crux08:03
*** arduo has quit IRC08:09
*** arduo has joined #crux08:10
*** arduo has quit IRC08:16
*** arduo has joined #crux08:18
cruxbot[contrib.git/3.1]: torsocks: initial import09:03
*** Junka has joined #crux09:19
Junkai never tried crux but i would like to09:20
Junkais crux gnu/linux?09:22
koriit's a linux distribution, yes09:22
Junkabut it's not gnu right?09:22
koriit does have GNU09:23
koriso yes technically it's GNU + Linux09:23
Junkai see the website says its targeted to experienced users09:23
Junkamaybe it's not for me09:24
Junkawhat's package manager?09:24
koriit's a system similar to ports from the *BSDs09:24
JunkaI never tried BSD either09:26
*** doomicide has joined #crux09:26
koriJunka: I might not be the best person to give you a tour of CRUX09:26
Junkakori: thanks maybe it's not just for me :009:28
Junka;) *09:28
*** Junka has left #crux ()09:28
diversesigh, I think the word "experienced" should be changed to "enthusiast"09:31
diverseCRUX is an enthusiasts system, how "experienced" you are depends on your aptitude to learn09:32
koribut maybe more explanation on the core features could be good09:33
frinnstwhat features? :D09:33
kori0. lack of systemd09:33
kori1. systemd-less system09:33
frinnstcrux: now with less features!09:33
kori2. sysvinit + rc instead of systemd09:34
korisorry I just programmed for 8 hours straight09:34
korithe sun is rising....09:34
tilmanwho was the BASICS guy? was it arnuld?09:35
frinnstwas it the lack of lynx, or what was it?09:36
frinnstWITHOUT LYNX, WITHOUT BASICS.... Sorry, makes no fun.09:39
koriRegards Ronny09:40
koribest part of the email09:40
tilmanhow so? o_O09:40
koriit just is, ~just go with it~09:40
*** tilman has quit IRC09:41
frinnsthehe the lack of "," makes all the difference09:41
korifrinnst: exactly09:41
frinnstwe sure have had quite a few characters over the years09:45
koriI'd be surprised if that wasn't the case, haha09:45
diverseI find tilman's sudden okthxbye quit, after not accepting your no-reason excuse, the most hilarious part of the whole dicussion :P09:50
kori[✔] Same09:51
Romsteri started off deep in crux with little bits of knowledge. an i kept breaking my system with rejmerge before i realized my machine was i586 not i686 at the time09:51
*** tilman has joined #crux09:51
frinnstI migrated my slackware machine to crux on the fly09:52
frinnstthe only thing that broke stuff was replacing slackwares glibc with the crux package :)09:52
tilmanfrinnst: did i mention i run netbsd on my guruplug?09:52
frinnstno, you still use that old sucker?09:52
frinnstsorry, the need for a fan really made me lose interest09:53
tilmanwant to guess how much hotter the sucker is with netbsd instead of linux?09:53
tilmani didn't notice at first09:53
frinnsthaha really?09:53
Romsteri'd expect cooler09:53
frinnstfinger burning hot again?09:53
tilmanthe replacement cooler i put on the CPU is 20 degrees hotter w/ netbsd09:54
frinnstholy hell09:54
frinnstbut still stable?09:54
tilmani ran it in the original enclosure for 6 months without knowing about the temp difference09:54
tilmansome parts smelled a bit funny when i opened it09:54
tilmanyeah, it's stable09:54
tilmani hooked up a 60mm fan running at 5V09:54
tilmansilent and cool09:55
frinnstguru was arm5, right?09:55
Romsterwell correctly below ambient sound levels.09:55
frinnstah, right. sorry silly numbering09:55
tilmanRomster: most definitely below my tinitus sound level09:55
Romsteroh i hate tinitus  fan helps to make white noise09:56
tilmanand yeah, it's slow. but good enough for non-cpu intensive work09:56
frinnsti bet my rpi is quieter :p09:56
tilmani'd like to get a parallella09:56
RomsterI read the RPI wont boot without the binary blob09:57
Romsterafter reading that i was disapointed09:57
tilmanfrinnst: sure, but most of the single board computers these days only have one NIC :|09:57
tilmani'm using mine as a router, so i need two. and i'd rather avoid a usb nic ;p09:58
frinnstbit pricey and usually comes with a fan09:58
tilmani'm keeping mine. it's my hardware tinkering project :)09:58
frinnstI recently bought two ikea lack tables. I need something 19"!09:59
tilmani wanted to read the fan's RPM via the board's GPIOs. not sure that's feasible though:D09:59
tilmanlack rack09:59
frinnstyeah. crazy cheap. like 50SEK = ~5EUR09:59
tilmanmy lack has spare space beneath it as well10:00
frinnstwhat do you have in your lack?10:04
tilmani'm using it as a table only ;))10:05
diverseholy shit, 18 cores on a credit card sized board?10:06
frinnstNorco RPC-3216 mounted in a LackRack Enterprise Edition.10:07
frinnsthaha Enterprise Edition10:07
tilmanLevenstein distance of "Lack" and "Rack" is 1. Can this still be a coincidence?10:08
frinnstno, destiny10:09
frinnstwho's dotdev?10:23
frinnstunixporn regular it seems :)10:24
frinnstsorry, wrong link:
cruxbot[opt.git/3.1]: poppler: update to 0.31.010:53
cruxbot[opt.git/3.1]: poppler-glib: update to 0.31.010:53
cruxbot[opt.git/3.1]: poppler-qt4: update to 0.31.010:53
cruxbot[opt.git/3.1]: cups-filters: apply patch to build with poppler 0.3110:53
*** doomicide has quit IRC11:22
*** arduo has quit IRC11:52
*** arduo has joined #crux11:52
*** dougl has joined #crux12:13
*** leo-unglaub has joined #crux12:17
*** tnut has joined #crux12:31
*** tnut has left #crux ()12:32
*** dougl has quit IRC12:39
*** arduo has left #crux ("ERC Version 5.3 (IRC client for Emacs)")12:47
*** tnut has joined #crux12:59
*** tnut has left #crux ()13:01
*** sh[4]rm4 has joined #crux13:03
*** sh[4]rm4 has quit IRC13:03
*** sh[4]rm4 has joined #crux13:04
*** sh4rm4 has quit IRC13:04
*** leo-unglaub has quit IRC13:16
*** greduan has joined #crux13:27
*** jdolan has quit IRC13:57
*** jdolan has joined #crux13:57
thetornainbowfrinnst: :)14:06
*** leo-unglaub has joined #crux14:20
thetornainbowi've been thinking about setting up a crux binary repo at home, since my desktop and soon laptop will both run Crux14:34
thetornainbowjust compile on a third box and push to both14:34
thetornainbowi'm just brainstorming but i assume i'll need to mount a network drive and prt-get update from there?14:35
thetornainbowor pkgadd14:35
jaegeryou could have your package dir mounted on all hosts with NFS or use rsync or scp to copy packages from the build host to the others14:36
thetornainbowoh yeah, NFS! i couldn't remember the nix network share14:38
thetornainbowit would also make sense to comment out the other pkg locations in prt-get as well14:39
thetornainbowwish i was at home to start setting it up. i was hoping for a snow day but no dice14:40
thetornainbowonly 3 tickets open ;-;14:40
*** arduo has joined #crux14:47
*** jdolan has quit IRC15:34
*** jdolan has joined #crux15:57
*** dougl has joined #crux16:07
rmullSo I have a perl script that is invoked with the -P flag, which runs the C preprocessor on the script before interpreting it16:12
rmullIt does not appear that perl on crux offers the -P flag16:12
rmullWhat's up with that?16:12
*** dougl has quit IRC16:25
rmullI don't usually use perl so I don't know what is normal or expected16:28
jaegerIs that flag a patched-in feature or was it added in a speific version?16:30
rmullI couldn't say, I have no information about it - the script was part of the body of source code for the xombrero browser, which is developed by BSD guys16:31
rmullThere are linux man pages online that reference it, so I don't think it's like a weird BSD perl variant or something16:32
jaegerperhaps it needs a specific configure flag16:32
rmullCould be. Maybe I'll look into it if I get some spare time, but it's not that critical16:35
jaegerhaven't run into it, myself, but I don't use perl directly anymore, at all16:37
*** arduo has quit IRC16:45
*** kori has quit IRC16:48
*** Pingax has quit IRC16:49
*** kori has joined #crux16:50
*** kori has quit IRC16:50
*** kori has joined #crux16:50
*** Pingax has joined #crux16:53
*** sh[4]rm4 has quit IRC17:10
*** dougl has joined #crux17:10
*** sh4rm4 has joined #crux17:40
*** sh4rm4 has quit IRC17:40
*** sh4rm4 has joined #crux17:40
*** sh4rm4 has quit IRC17:45
*** dougl has quit IRC17:51
*** sh4rm4 has joined #crux17:56
*** sh4rm4 has quit IRC17:56
*** sh4rm4 has joined #crux17:56
*** lnds has joined #crux18:00
greduananyone play CSGO?18:00
*** doomicide has joined #crux18:05
*** hhhhhhhh has joined #crux18:09
greduanlet me change my question, anyone play CSGO on CRUX?18:13
tilmanno, why are you asking?18:20
tilmansteam works if that's your question :>18:21
greduanyeah that's my question18:21
tilmanglad i could help then :D18:21
greduanlol yep18:22
*** toriso has quit IRC18:37
*** sh4rm4 has quit IRC18:48
*** sh4rm4 has joined #crux18:54
*** arduo has joined #crux19:07
*** leo-unglaub has quit IRC19:12
*** doomicide has quit IRC19:20
*** dougl has joined #crux19:22
*** rmull_ has joined #crux19:30
*** arduo has quit IRC19:44
*** sh4rm4 has quit IRC19:44
*** mechanip1ter has quit IRC20:05
cruxbot[opt.git/3.1]: nss: added Alan's patch for nspr in nss.pc20:06
*** dougl has quit IRC20:09
*** leetspete1 has quit IRC20:09
greduanmorning prologic20:27
*** dougl has joined #crux21:28
cruxbot[core.git/3.1]: [notify] unzip: added patches for CVE-2014-{8139,8140,8141}21:28
*** lnds has quit IRC22:07
*** hhhhhhhh has quit IRC22:25
*** arduo has joined #crux22:30
*** arduo has left #crux ("ERC Version 5.3 (IRC client for Emacs)")22:54
*** jdolan has quit IRC23:40

Generated by 2.11.0 by Marius Gedminas - find it at!