IRC Logs for #crux Thursday, 2017-09-21

*** vsteve has joined #crux00:09
vsteveif you folks do ldconfig -p do you see that you have a libX11?00:11
vsteveshouldn't installing xorg provide -lX11?00:12
abenzhi vsteve00:15
vstevehey abenz00:16
vsteveI just found the xorg-libx11 package00:16
vstevelooks like my application wants...a different -lX11, I'm guessing the 32 bit one00:16
*** itPuffin|osxB has joined #crux00:22
*** BitPuffin|osx has quit IRC00:23
abenzvsteve: there will be a vega nano!00:34
abenzI will start saving up some cash00:35
abenzhopefully miners wont rack up its price00:35
vsteveoh?  will it be faster?00:35
vstevethe fiji chip is insane00:36
abenzI was very surprised when I saw the OSS linux drivers almost match windows in performance00:36
vstevewow00:37
vstevecan mesa3d and mesa3d-32 be installed on the same machine and both work?00:37
vsteve...or am I just trashing a machine right now?00:38
abenzfury X is 8.6 TFLOPS, while Vega is 12.6 TFLOPS00:38
abenzso yes it will be significantly faster I reckon00:38
abenzvsteve: yes, can coexist00:38
vsteveoh thank god00:38
vsteveI'm trying to build this custom doom3 engine that needs a ton of 32 bit stuff00:38
vsteveoh....damn...I can't get llvm-32 to install00:41
vsteveit chokes on libffi - it says headers aren't installed but I installed libffi-3200:41
vsteveso I'm stuck00:41
vsteveso, Crux gods, if you could be so kind as to check the llvm-32 package, I think something's up00:46
xckoand libffi is installed? I have llvm-32, installed fine01:04
cruxbot[contrib.git/3.3]: wol: install man pages in /usr/share/man01:13
vstevelibffi is installed01:17
vsteveboth libffi-32 and libffi01:18
jaegertesting in a clean environment01:27
vsteveCMake Error at cmake/config-ix.cmake:287 (message):   libffi includes are not found. Call Stack (most recent call first):   CMakeLists.txt:582 (include)01:29
*** abenz has quit IRC01:46
*** xcko has quit IRC01:49
*** xor29ah has quit IRC01:57
*** xor29ah has joined #crux01:57
vsteveI wonder what, if anything, I broke to stop llvm-32 from being able to see libffi02:03
vsteveI really wish my application just didn't have 32-bit dependencies02:03
vsteveis there an easy way to remove all compat-32 installed items from crux?02:09
vstevelike prt-get remove compat-3202:10
jaegerprt-get remove $(prt-get listinst --regex '\-32' | grep -v glibc)02:11
vsteveyou're a saint02:11
jaegerI wouldn't go that far but thanks, heh02:11
vsteveI want to see why this application /needs/ 32 bit code to link02:12
vsteveit complied entirely without a single 32-bit  anything being installed02:12
jaegerwhat's the application?02:12
vstevefhDoom3, an updated doom3 engine02:13
vstevehttps://github.com/eXistence/fhDOOM02:13
vsteveI just find it odd that it compiled fine but the linker wants 32-bit things02:14
vstevemaybe that's normal for games?02:15
jaegerIt's normal for 32-bit apps, not games specifically02:15
vsteveoh ok02:16
vsteveI can't remember the last 32-bit thing I built that wasn't this02:16
jaegerbased on the "Building fhDOOM" comments at the bottom of that page, I assume it's a 32-bit program02:16
vsteveoh it is02:17
vsteveI just thought a 32-bit program would need 32-bit libraries at compile time02:17
vstevenot just linking time02:17
vsteveI don't get how that's not the case02:17
jaegerno idea there, sorry02:17
vsteve..the more I think about it....generic function calls being made..would sit on 64-bit code....so....how...*brain explodes*02:18
vsteveI've been pestering the author, I helped work out a few compile time bugs02:18
vstevenow i'm just stuck with llvm-3202:19
jaegermy test container has llvm and llvm-32 left, just takes a while to build02:19
vsteveah thank you02:19
ryu0vsteve: uh, the libraries aren't important until link time. you only need the headers to actually build it.02:20
ryu0compile*02:20
vstevehmm...ok02:20
ryu0not that it matters much. you still need both to produce something usable.02:21
vstevealso true02:21
ryu0only if the entire library is a header file can you avoid a library at link time.02:22
ryu0that's mostly a C++ thing though (templates).02:22
vsteveincidently this application is predominantly c++02:22
vsteveit's not as easy as rewriting the makefile to point to 64-bit libraries, is it?02:23
vstevethe whole thing would need to be rewritten...I'm assuming02:23
ryu0no idea. each distribution can have their own multilib or multiarch setup.02:24
vsteveoh I mean to get the application itself to be 64 instead of 32 bit02:24
ryu0if it's insisting on 32 bit, it's for a reason.02:24
vsteveI feared as much02:24
ryu0it may use 32 bit only ASM.02:25
ryu0or something else that makes it 32 bit only.02:25
vsteve...guh crap, that's probably what it is too02:25
vsteveparts of doom3 were famously written in ASM02:25
ryu0that's why zsnes will probably never be ported to another architecture.02:25
vsteveI'm not familiar02:25
vstevebut I think I see what you mean02:26
ryu0it's an older SNES emulator.02:26
ryu0it's 32 bit x86 only.02:26
vsteveah02:26
ryu0because it uses ASM.02:26
vsteveright02:26
*** mavrick61 has quit IRC02:26
vsteveI wonder what doom3 actually does in ASM, I thought the pixel shaders and things were some kind of weird gpu ASM that nvidia cooked up...but the game binary itself shouldn't...do that I wouldn't think02:27
vstevebut I recall hearing that it did02:27
*** mavrick61 has joined #crux02:27
vstevehuh, so someone forked it an un-ASM'd it02:28
vstevemade a 64-bit SDL port02:29
vstevebut of couse that engine doesn't have the features of this one02:29
vsteveI wonder then how hard it would be to tack these features onto that 64-bit engine02:29
*** abenz has joined #crux02:59
brian|lfshow hard would it be to run the first linux kernel on modern hardware03:29
ryu0brian|lfs: probably hard. it's unlikely any hardware would work.03:44
*** Dell has joined #crux04:19
*** maraku has joined #crux04:20
*** tilman_ has joined #crux04:23
vstevejaeger did those llvm builds ever finish?05:03
jaeger98% currently05:06
jaegerIt just finished, no problems.05:07
vstevedangit05:11
vsteveok, well thanks for checking that out05:11
vsteveI have no idea what's wrong with mine05:11
vstevebut it may actually be worth my while to avoid the 32-bit program, start with the 64-bit I found and try to add features to that one05:11
vsteveas much as i'd love having two of every library on my system..05:12
vstevethanks for checking it out though..sorry to have killed a few hours of cpu time05:13
*** vsteve has quit IRC05:17
jaegerno worries, had plenty of other stuff to do05:32
*** Workster has quit IRC07:18
*** Workster has joined #crux07:19
*** Workster has quit IRC07:19
*** Workster has joined #crux07:19
*** p4cman has joined #crux07:37
*** pedja has quit IRC08:24
*** abenz_ has joined #crux08:37
*** abenz has quit IRC08:40
cruxbot[opt.git/3.3]: hpcups: update to 3.17.909:18
*** onodera has joined #crux09:21
*** pedja has joined #crux09:42
*** abenz_ has joined #crux10:34
pedjaevery 14 days, or so, DNS breaks for almost all applications10:56
pedjadig shows that upstream DNS servers and DNS relay at the router work just fine10:57
pedjatcpdump shows that applications get the proper reply for DNS query10:58
pedjaand yet, almost none of the apps can resolve10:59
pedjaapparently, if the logs fill up root file system, DNS breaks :)11:02
pedja(root fs is small, 2Gb)11:03
pedjawhy 2 weeks? well, that is how long it takes the logs, for which I forgot to set up logrotate for, to grow just enough to fill up root11:05
*** itPuffin|osxB has quit IRC11:49
cruxbot[contrib.git/3.3]: ethtool: update to 4.11, closes FS#147412:27
joacimalways heard having different filesystems and partitions for the various directories under root was a good idea12:43
joacimalways been too lazy for that stuff. + i never really saw the point, besides a dedicated drive or partition for /home12:43
pedjathis mess is a result of me not thinking ahead :)12:49
pedjaI bought 1 drive, used it for a while, then bought another, and, rather poorly, 'integrated' it into the system12:51
pedjait will have to do, until I buy a 2Tb drive and reorganize it12:53
joacimwent through that with windows and partitoned the drive for both crux and windows12:55
pedjasystem and home on 2Tb drive, ZFS mirror or something with 2 1Tb drives for docker and VMs12:56
joacimi think my original winxp partition was 15GB, which ended up being too small for Vista, so I grew that, but that ended up being too small for windows 7 too once I've installed some updates12:56
pedja:)12:56
pedjamy usr partition is ~100G, 40Gb of it actually used13:00
pedjaso, in theory, I could use one of the smallish 120Gb SSDs for it without a care in the world13:02
frinnstwow, that much?13:34
frinnsti guess libvirt fills up ~50% of that?13:35
SitriI remember having Windows 98/XP on a 700MB drive, with a 1.5GB drive as a secondary13:38
SitriEventually got to the point where I had to fight with everything to make sure it didn't put anything on C:13:38
SitriAlso for some reason, I had decided to give this one system a / of 11GB, only other partition was /home13:39
joacimthe root partiton on my laptop is about 30 GB13:42
joacimi would feel better with a larger one13:42
joacimtempted to get a 64GB msata ssd, and use the current SSD for /home13:43
*** p4cman has quit IRC13:46
Sitrihttps://daniel.haxx.se/blog/2017/08/30/easier-http-requests-with-h2c/ <-- OMG13:51
*** JanC_ has joined #crux14:44
pedjafrinnst, nah, that's on another partition, bind mounted to var/lib/libvirt :)14:55
pedjapkgmk, usr/lib and usr/share are ~31Gb14:56
*** abenz_ has joined #crux15:03
SiFuhWhen I run xev or xbindkeys -k and check the keycode for my fn (function) key  nothing registers at all..15:23
*** p4cman has joined #crux15:41
darfomy laptop does that on some of the function keys too. hardware intercepts them and they never get to the keyboard controller.15:41
SiFuhOh well, I will just use the "windows" looking key with f1, instead of fn+f1 to toggle the touchpad15:42
*** chinarulezzz has quit IRC15:44
SiFuhdarfo: how weird is this... I bind fn+f1 it doesn't work but fn+f11 does work!16:06
*** frinnst2 has joined #crux16:54
*** tsaop has joined #crux17:38
*** chinarulezzz has joined #crux18:00
darfoe19:01
darfowrong window19:01
SitriFn is among the worst things to ever happen to keyboards19:07
*** tsaop has quit IRC19:55
*** p4cman has quit IRC21:16
*** jdolan has joined #crux21:56
*** Havde has joined #crux23:40

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!