IRC Logs for #crux Sunday, 2018-08-19

jaegerwe've talked about it many times but not yet implemented it00:03
pedjaone of those things that keeps coming back :)00:11
crash_pedja: it seems like you have the weekend figured out :)00:12
pedjaI couldn't code myself out of wafer thin wet paper bag :)00:13
pedjareuse existing code and glue a few things together? maybe, with a lot of swearing. create something from scratch? nope.00:17
pedjaI think that 'run Furmark, your rx580 will either survive it or...not' is a solid advice00:29
pedjaas bug reports go, 'I just *feel* that my gfx card isn't working properly' is pretty useless00:34
j_vI had a patch pkgmk for file renaming, but I think it was concidered too invasive, which I understand. I seem to remember that fun had something, also, but can't remember the details right now. I probably saved it somewhere, but too much other stuff right now to focus on that.01:06
joacimjaeger: you mean you havent found one shove it all on yet? =)01:33
jaegersorry, not understanding what you meant there01:51
joacimthe job of implementing the renaming01:52
jaegerAh, heh. No, not yet01:59
ryuoOne idea is to use a separate variable to define alternate names.02:17
ryuoFor the downloaded file.02:17
jaegerWhat I would do is just some separator like := or :: or similar02:18
ryuoI suppose :: would work. Can use % escapes for URLs that happen to use ::.02:20
ryuoOr even backslash.02:20
joacimi think arch does it nicely02:39
joacimat least in my eyes02:39
joacimnot really looked that deep tho02:39
joacimidk. maybe i saw it somewhere else02:40
j_vthe patch I made used ::02:56
][_R_][# a source entry can have two forms :03:06
][_R_][# 1) "filename::http://path/to/file"03:06
][_R_][# 2) "http://path/to/file"03:06
][_R_][get_filename() {        local filename="${1%%::*}";        echo "${filename##*/}"; }03:06
][_R_][get_url() {        echo "${1#*::}"; }03:06
][_R_][(From pacman's makepkg)03:06
*** john_cephalopoda has quit IRC03:12
*** tilman_ has joined #crux03:24
Romsterwe were talking about it but then git hub had this method to download04:24
brian|lfshey all04:29
brian|lfseveryone ok04:29
*** saptech has joined #crux04:43
*** saptech has left #crux ("Leaving")04:55
*** abenz has joined #crux04:56
j_vRomster: the subject came up again because that method currently does not work05:09
j_vis anyone using extlinux? i'm wondering if anyone else is having issues with booting newer kernels with it. so far, every kernel >= 4.16.17 that i try to boot with extlinux, throws a file not found error. the same kernels boot fine from grub2 boot iso.05:16
ryuoj_v: what FS?05:17
ryuoj_v: ext4? did you format it with newer features enabled?05:18
j_vhad to disable one of them, forget which one... been a while since I installed05:19
j_vyeah, that was it05:19
ryuothat's one thing that can cause read errors.05:19
ryuootherwise no idea. the kernel version shouldn't matter.05:19
j_vright, that is what i thought... version shouldn't matter... though I am thinking that there could be a fs bug in the kernel i am running when i install the kernel05:21
j_vhmmm, since I am running 4.18.2 booted via grub boot iso, i should try reinstalling those kernels then try booting them again05:22
ryuohonestly i got fed up with extlinux...05:23
ryuoit's not keeping up with emerging trends.05:23
j_vyeah, up til now, it did the trick fine... was already thinking about installing grub boot blocks, just might and save myself some trouble05:24
abenzcheck this out05:49
abenz1. (threadripper 2990wx)
abenz2. DUAL epyc 7601:
abenzthe single 2990wx beats the dual 760105:50
][_R_][j_v: I'm using Extlinux with a 4.17 kernel, /boot is its own partition and Ext2.  Works fine without issues.05:53
j_v][_R_][: thanks for the info. makes me think a fs bug might be worth looking into.05:56
ryuoTIL. Here-docs can be used in shell pipelines.06:46
ryuocat << 'EOF"06:46
ryuocat << 'EOF' | wc -l06:46
cruxbot[compat-32.git/3.4]: glib-32: 2.56.1 -> 2.56.206:58
*** slek has joined #crux07:26
Romsterj_v, i see you fixed your certificate, do you have in the works for openjdk10 10.0.2.u13? and update for openjdk8 8.u181?07:46
*** chinarulezzz has quit IRC08:16
*** chinarulezzz has joined #crux08:16
*** abenz has quit IRC09:17
cruxbot[opt.git/3.4]: qpdf: update to 8.2.109:22
*** abenz has joined #crux09:37
*** znh has joined #crux09:57
cruxbot[contrib.git/3.4]: ffmpeg: --enable-avresample10:27
*** abenz has quit IRC10:33
*** onodera has joined #crux10:45
*** abenz has joined #crux10:45
onoderagood morning10:45
onoderadoes anyone run into this issue: pkgfoster can't find libglvnd10:45
onoderait should even look for it because it is in core though10:45
onoderaoh its not in core acutally hehe11:02
onoderastill weird that it compains though because the port IS in the ports tree11:02
onoderaprt-get cache fixed it :311:06
*** abenz has quit IRC11:56
*** groovy2shoes has joined #crux12:56
*** frinnst has quit IRC13:30
cruxbot[opt.git/3.4]: freerdp2: initial import13:39
Romsterspent to much time trying to get freerdp to work but i got the new version built.13:39
*** john_cephalopoda has joined #crux14:01
pedjathis new GH behaviour is annoying af14:37
pedjaRomster, shouldn't ccache/distcc be used only for massive sw, like qt5, rust, ff etc? freerdp2 compiles fast even on my potatoputer14:40
pedjaadding that to the freerdp2 port seems a bit overkill, imho :)14:42
*** frinnst has joined #crux14:48
*** abenz has joined #crux14:50
*** abenz has quit IRC15:32
*** abenz has joined #crux16:13
*** ileach has joined #crux16:24
*** kInOzAwA has joined #crux16:25
*** ileach has quit IRC16:44
*** ileach has joined #crux16:45
*** xor29ah has joined #crux17:14
*** abenz has quit IRC17:19
*** ileach has quit IRC18:28
*** Kruppt has joined #crux19:07
*** Kruppt has quit IRC19:16
j_vRomster: I might have some time to look into it in the next two weeks, though I am hesitant to promise anything as I will be travelling for work and back to doing heavy overtime at job again very soon. I will try to get to it. It is on my list, though.19:28
frinnsteveryone are soooooo screwed20:14
frinnstwell, if anything you touch run on intel hardware which is 100% of the population?20:14
*** j_v has quit IRC20:19
ryuofrinnst: this whole year keeps getting worse and worse for Intel.20:32
ryuoat this rate the next vulnerability will be called WTF.20:32
*** onodera has quit IRC20:41
pedjaand yet, their stock price is at all time high, apparently. stock markets are weird20:43
frinnsta colleague just patched up a new esxi host (not in prod) to the latest 6.5 version20:44
frinnstand unless you toggle a setting it will psod20:44
pedja'non-trivial performance impact' ouch20:45
pedjapink screen of death?20:45
frinnstwill do some benchmarking tomorrow :(20:45
pedjahow bad should the performance impact be before companies take a closer look at AMD CPUs?20:49
pedjaiow, when does it make business sense to switch?20:50
pedjait's a bit of a chicken-and-egg problem. companies might buy more AMD servers, if server vendors made more, and they would made more if more companies would buy them20:54
pedjaor something like that20:54
ryuopedja: i imagine smaller companies could get away with custom builds.20:55
ryuoBut... yea...20:55
ryuoI've only seen HPE offer AMD servers.20:56
pedjadid Vmware certify AMD stuff yet?20:56
ryuoNo idea.20:56
ryuoI'll look.20:57
pedjaepyc 7xx1 series is compatible with esxi-6.5u1+21:01
ryuofrankly, I suspect HP and HPE are the #1 PC manufacturers that use AMD.21:01
ryuoDell doesn't use AMD in anything business last I checked.21:02
ryuoLenovo... i think only in a select few thinkpads.21:02
ryuopedja: I recall HPE being a major company that the AMD CEO creditted with EPYC's success or so.21:03
*** j_v has joined #crux21:03
ryuoI mean... AMD isn't an OEM so... it seems HP and HPE have become their main OEM customer.21:03
pedjafrinnst, whatever happened with that nvidia gpu blade thingie you were testing? did you keep it?21:08
pedjagoogle provides free access to tesla k8021:10
j_vryuo: thanks for mentioning the 64bit option for ext4. I was certain I had turned that feature off for my boot partition. Glad you mentioned it and I finally decided to double check with dumpe2fs. Sure enough, it was on. Turned it off. Now booting any kernel is fine. I was so certain that I had turned that feature off. Oh well, all good now.21:10
*** john_cephalopoda has quit IRC21:30
*** wh0pie has joined #crux21:42
frinnstpedja: its powered off, we're returning it soon21:46
frinnstwe needed a nvidia grid to test what we wanted to test21:47
frinnstand even so, never nvidia again21:47
pedjayou're not on a ML bandwagon :) ?21:47
pedjamachine learning21:48
*** slek has quit IRC21:48
frinnstah no, we dont have any use for that21:48
*** slek has joined #crux21:49
pedja:). why 'never nvidia again'? software/hardware issues?21:51
pedjaor the company?21:51
frinnstconsidering they charge a license fee monthly/yearly to use their drivers. fuck em21:51
pedjaI really wish openCL usage picks up, they are, what's the word, abusing? their de facto CUDA monopoly21:54
pedjamaybe next year, with openCL/Vulkan221:55
pedjaspeaking of Vulkan, wth am I going to do with my Vulkan ports now GH broke the sane d/l21:56
frinnstgithub broke?21:56
frinnstjoy, 40421:57
pedjaI wish :) no, just that workaround to give files sane names instead of v1.2.3.tar.gz21:57
pedjaaccording to their status page, all is A-OK21:58
jaegerrehost them elsewhere, I guess21:58
pedjaI'll put them in my Bitbucket repo, I already have a couple of things there21:59
pedjanot all repos are broken, the ones that already had sane tarball names still work, afaik22:01
pedjawhat's interesting, not a peep about it in tech press, or reddit, or HackerNews22:02
jaegerThey just removed the feature or whatever, I guess22:08
jaegerand not many used it22:08
ryuoi think i'll write a system for managing my iodd2531. it's a bit of a hassle to install new ISOs by hand.22:08
pedjaI am not familiar with Witcher 3, but can rx580 push 80-100fps in it on a 144hz monitor?22:12
pedjacan't be bothered to look it up, gaming is just not that interesting to me. unless it's something casual like Moorhuhn22:19
pedjaor 'wholly shit, I am too stupid to play this' like Go22:20
pedjaI wish I had the resources to make two self-learning Go bots play each other, that would be cool to watch22:22
joacimpedja: sure. depends on res and settings22:45
joacimmight be dips below 60 on higher settings22:45
joacimi get dips to 50 i think on my r9 fury, but i play with maxed out settings at 2460x144022:46
*** john_cephalopoda has joined #crux23:27
*** wh0pie has quit IRC23:28
*** slek has quit IRC23:29
*** wh0pie has joined #crux23:33

Generated by 2.14.0 by Marius Gedminas - find it at!