IRC Logs for #crux Wednesday, 2019-02-20

john_cephalopodaOooh, we'll be at Linux 5.0 soon.00:01
john_cephalopodaNothing groundbreaking, it seems. Apparently the only reason for that bump is that Linus doesn't like big numbers after the comma.00:02
dlcusaYeah, 5.1's going to introduce new async I/O improvements that could be really efficient.00:03
john_cephalopodaThe only kernel changes that I really noticed are bugfixes in the nouveau driver.00:27
john_cephalopodaOther changes are more subtle.00:27
john_cephalopodaLike, in general.00:28
john_cephalopodaWhen I update my kernel I usually have no indication of the change except for a new version number in uname and maybe that openGL stuff glitches out less or in different ways :þ00:28
*** crux_user has joined #crux00:34
crux_userGimp don't show icon's, only red X for every botton on tools thing and stuff.00:35
crux_userI'm thinking seriously to never update again my crux.00:37
*** john_cephalopoda has quit IRC00:39
*** john_cephalopoda has joined #crux00:41
rmullcrux_user: prt-get depends gimp shows all deps are installed?00:53
crux_userrmull: was a gimp problem, I had to go edit, then preferences, it seems have issues with resolution / icon theme pre-configured.01:03
crux_userI forked the ports, now I will try to learn how to maintain them, I have master branch where I will try just to add fix's ...01:05
crux_userfor my litle experience and for what I had learn, core is ok to update without looking to much, same I can't say about opt and contrib, each time I update from upstream I'm sure I will have many problems.01:07
crux_userI need to learn how to pick just specific commit's, so I can update x or y package without updating all repository.01:07
*** xor29ah has quit IRC01:17
*** xor29ah has joined #crux01:19
*** xor29ah has quit IRC01:25
*** xor29ah has joined #crux01:27
*** crux_user has quit IRC02:02
*** crux_user has joined #crux02:17
*** _________mavric6 has quit IRC03:27
*** _________mavric6 has joined #crux03:29
*** abenz has joined #crux04:02
*** mayfrost has quit IRC04:03
*** abenz has quit IRC05:04
*** abenz has joined #crux05:18
Abdullahis there a solution for package which I have been compiling for hours and after compilation has done, and in the last line I had an error so it didn't install package, so can I skip recompliation and install it?05:52
joacimwhat kind of error?05:56
joacimif the package was built, but install failed, you can typically just pkgadd / pkgmk -i the package05:57
joacimif it was a build error, there is no other way to restart the build, after fixing what might have been wrong05:58
Abdullahit was only find error finding a file06:00
Abdullahlemme check if package is there or not06:00
Abdullahnah, package isn't there in packages06:01
joacimif it was a footpring mismatch, the package should be there, missing files should be checked out06:04
Abdullahjoacim: it was other error06:04
joacimbuild logs should be in /var/log/pkgbuild06:05
Abdullahmv $PKG/usr/${D}/libstdc++.so.*-gdb.py $PKG/usr/share/gdb/auto-load/usr/${D}06:19
Abdullaherror is here in this line06:19
Abdullahdoes the pkgmk save the pkgfile in ram? if I edit the Pkgfile after running the pkgmk, will it apply?06:23
AbdullahI commented out those lines now. I hope it will work now ;-) or maybe I get a broken system ;-)06:25
Abdullahas long as I understand, these are 32 libs which I don't need06:26
Abdullahlets see what happens ;-)06:26
joacimhave to restart the build after editing i think06:34
Abdullahokahy06:36
Abdullahit took one hour when I did with j 8 and with default settings it took 3 hours06:37
AbdullahI wish we could use ccache06:37
joacimyou can if you want to06:39
joacimrequires some configuration tho06:39
Abdullahhmm06:44
*** admincom1lex has joined #crux07:35
*** admincom1lex has quit IRC07:35
*** admincom1lex has joined #crux07:54
*** admincom1lex has quit IRC07:54
*** admincom1lex has joined #crux08:08
*** admincom1lex has quit IRC08:09
*** admincomplex has quit IRC08:22
*** admincomplex has joined #crux08:23
frinnstAbdullah: romster is a heavy user of ccache and distcc i think09:21
Abdullahhmm09:22
*** admincomplex has quit IRC09:57
*** admincomplex has joined #crux09:57
*** guido_rokepo has joined #crux10:26
*** xor29ah has quit IRC11:26
*** xor29ah has joined #crux11:28
*** xor29ah has quit IRC11:33
*** xor29ah has joined #crux11:35
*** marwellous has joined #crux11:36
marwellousHey.11:37
marwellousOne thing bothers me: since anyone can contribute to this project, how do I now if some malicious hacker didn't infect my computer with some keylogger or other malware?11:39
frinnstWell first off: You have similar troubles with all software. It all comes down to trust11:49
frinnstWe use signify to "sign" upstream tarballs. But that doesnt really matter all that much. since a upstream project could already be compromised11:50
frinnstthen everything is stored in git - that's dificult to manipulate11:52
frinnstbut yeah, a developer/contributor could push a port with a "rm -rf /*" script in it11:53
frinnstall I can say is that our process is open: https://crux.nu/gitweb/11:54
*** marwellous has quit IRC11:56
*** marwellous has joined #crux12:50
marwellousSorry, I was disconnected12:51
marwellousIs it possible to submit bugs anonymously?12:58
frinnstYou can always email us or the maintainer of a specific port13:03
frinnstBut such reports can be easily lost so we recommend bugs be filed against our bugtracker13:04
frinnstYou can always also share a bug in this channel. But like with email, such reports can get lost easily13:04
marwellousAnd in order to do that you have to register, in both cases.13:05
marwellousRight?13:05
frinnstto file bugs, yes13:05
frinnstwe cant have nice things on the internet. If we leave it open we get swamped in spam13:06
marwellousI understand.13:06
frinnstBut all that's required is an email address and that's pretty anonymous13:06
marwellous(I'm being a little bit offtopic right now) And once again we can't have nice features because of people. Like you said: "We can't have nice things on the internet."13:08
frinnstif you missed anything while you were disconnected you can read it here: http://irclogs.shortcircuit.net.au/%23crux/latest.log.html13:10
frinnst(we log irc)13:10
marwellousIs there a reason for doing that?13:10
frinnstyes, its useful :)13:10
pedjaTLA agencies made us do it13:11
marwellousWhat's TLA?13:11
frinnstsince you missed some of what I said you can read it again13:11
frinnstafk, lunch13:11
pedjaThreeLetterAgency13:11
marwellousYeah, it's fucked up to live in the USA.13:12
pedjaevery country has one13:12
pedjaand the reach of that particular tla is hardly just the us-of-a13:14
marwellousOk.13:19
marwellousI checked the wiki; it says that a first installation is 200-500MB. Is that true?13:19
marwellous*may be depending on your set up, of course.13:20
john_cephalopodaWhen you do a minimal install, that could be right.13:21
pedjaiirc, docker image with just the core is ~600Mb, fwiv13:22
john_cephalopodaThe iso (with a lot of compressed files) is ~700 MB, so an installation is likely less than 2GB.13:22
pedjatoolchain is the majority of that13:25
marwellousBye.13:29
*** marwellous has quit IRC13:29
cruxbot[opt.git/3.4]: gnutls: 3.5.18 -> 3.6.614:07
cruxbot[opt.git/3.4]: p11-kit: 0.23.10 -> 0.23.1514:07
cruxbot[opt.git/3.4]: [notify] make-ca: new dependency for p11-kit14:08
*** Stealth has joined #crux15:28
*** guido_rokepo has quit IRC15:40
cruxbot[xorg.git/3.4]: xorg-libfontenc: updated to 1.1.416:34
cruxbot[xorg.git/3.4]: xorg-xvinfo: updated to 1.1.416:34
cruxbot[xorg.git/3.4]: xorg-xrdb: updated to 1.2.016:34
cruxbot[xorg.git/3.4]: xorg-xmodmap: updated to 1.0.1016:34
cruxbot[xorg.git/3.4]: xorg-xhost: updated to 1.0.816:34
cruxbot[xorg.git/3.4]: xorg-xev: updated to 1.2.316:34
*** marwellous has joined #crux16:47
*** marwellous has quit IRC16:49
*** SiFuh_ has joined #crux17:04
*** SiFuh has quit IRC17:05
*** marwellous has joined #crux17:10
marwellousWhat about the kernel? It's outdated; the last stable release is 4.20.11.17:11
john_cephalopodamarwellous: You are responsible for keeping the kernel up-to-date.17:13
john_cephalopodaThe kernel isn't packaged.17:14
marwellousOk17:14
*** saptech has joined #crux17:15
*** marwellous has quit IRC17:16
*** saptech has left #crux ()17:17
*** denres has joined #crux17:31
*** denres has quit IRC18:01
*** denres has joined #crux18:47
*** sammi` has quit IRC19:13
*** sammi` has joined #crux19:14
*** crux_user has quit IRC19:35
frinnstreal men roll their own kernel!19:38
frinnst(also women)19:39
joacimr9 290 while rendering a desktop eats up about 50W19:47
joacimsame as my rx 550 under 100% load19:47
denresNah, there are no girls on the internet20:09
*** timcowchip has joined #crux20:22
abenzjoacim: yes the power consumption drop in recent years is significant20:24
ryuothat was some work.20:27
ryuoi implemented heapsort.20:27
joacimtheir performance is very different tho. r9 290 is a lot faster20:28
joacimi see people getting rid of some used rx 480s now, might get one of those20:29
*** timcowchip has quit IRC20:30
*** dlcusa has quit IRC20:46
*** dlcusa has joined #crux20:49
abenzjoacim: I'd be careful, those are likely miner cards21:04
abenzran 100% non-stop for very long21:04
joacimyeah21:10
denres@google crux linux21:15
denres@helo21:16
denres@help21:17
john_cephalopodaabenz: When you can get them 90% off, you can buy 10 and use them one after another, in case one breaks down :þ21:17
abenzhah. if thats the case I stand corrected21:18
joacimif the card looks in good shape, and they have the original recipe, i'm ok with it21:26
joacimstores have to cover issues that werent caused by the end user within the first 5 years21:26
*** denres has quit IRC21:31
abenz5 years?22:10
abenzping Romster Workster22:14
joacimhttps://lovdata.no/dokument/NLE/lov/2002-06-21-34#KAPITTEL_622:15
abenzRomster: https://pastebin.com/VezUijJM22:17
abenzjoacim: I think AMD put something in their own terms with regards to using cards in mining22:17
joacimsomething they cannot prove have been done to the card22:18
joacimas far as they know, the card might as well have been used to play minesweeper22:18
joacimnever tell the people handing your support case that you've overclocked, overvolted, or mined =)22:19
joacimi think overclockers do a lot more damage to their stuff in general, than miners22:21
joacimunless the miners placed their rig right next to an AC unit22:21
joacimAC is rare in norway, so unlikely ot have happened22:21
joacimjust gotta stay away from the gamer overclocking types that put liquid metal on everything22:21
joacimalso the warranty doesn't really matter in norway, unless it is longer than 5 years22:24
joacimseen some consumer routers with massive warranties22:24
abenzsafe to buy in normay I suppose :)22:27
Abdullahhttps://ptpb.pw/ZPJy I get this error after compiling gcc, this is the pkgfile https://ptpb.pw/fXWx22:49
jaegerDoesn't look like that output and that Pkgfile match22:53
jaegerMake sure you're actually using the one you think you are22:53
jaegeralso, if you just want gcc 8.2.0, take a look at https://crux.nu/gitweb/?p=ports/core.git;a=blob;f=gcc/Pkgfile;h=99224dca98198b101af3bfb7e922e8815c2cb994;hb=refs/heads/3.522:54
Abdullahoh yeah I think I commented out last 3 lines22:54
Abdullahhttps://ptpb.pw/OrJs jaeger  this is the pkgfile22:55
AbdullahI forgot to uncomment the last line which were problematic22:55
jaegerThat -j100 isn't going to do anything useful for you22:58
jaegerAlso, you might want to set up ccache to be used globally instead of having to modify every Pkgfile where you want it used22:58
jaegersee the "Using ccache to cache builds" section of https://crux.nu/Wiki/HowToSpeedUpBuildingPackages for an example22:59
john_cephalopodaYou can specify the job number in /etc/pkgmk.conf22:59
john_cephalopodaexport JOBS=$(nproc)22:59
john_cephalopodaexport MAKEFLAGS="-j $JOBS"22:59
Abdullahjaeger: I think ccache won't take effect in this case for now?23:01
jaegerNot sure what you mean23:03
AbdullahI mean as if I enable ccache now, it won't take effect for this gcc compile for now?23:04
jaegerCorrect, not until you do a compile with it enabled23:04
AbdullahAs this will be the first compile after enabling ccache.23:04
Abdullahjaeger: thanks that was what I meant.23:05
jaegerIf you think it might fail, add '-kw' to pkgmk so that you can investigate if a file exists or doesn't, if a path is wrong, etc.23:05
jaeger(after the build fails)23:05
jaegeralso, regarding the -j100 - the threshold for diminishing returns varies with each package or piece of software but it's definitely there. As an example, see http://jaeger.morpheus.net/misc/llvm-32.png23:06
AbdullahI think -kw is better option as most of the compile we do doesn't take longer23:06
*** john_cephalopoda has quit IRC23:15
*** john_cephalopoda has joined #crux23:17
Abdullahfrom where I can download these patches for gcc. giving error for no URL specified23:18
jaegerwhich patches?23:22
jaegerIf you mean the ones in the Pkgfile they're shipped with the port itself23:22
Abdullahthe ones in gcc pkgfile.23:23
jaegerIf you want to get the one from the 3.5 branch, 'rsync -a crux.nu::ports/core/3.5/gcc/ gcc/'. Just make sure to do it somewhere other than where your current one is if you want to keep it23:24
Abdullahin my homedir23:25
AbdullahThanks got them23:26
jaegerNot that crux 3.5 isn't released yet and gcc 8.2 on crux 3.4 isn't heavily tested23:26
jaegers/Not/note/23:26
Abdullahjaeger: I love taking risks ;-)23:27
jaegerfair enough23:27
AbdullahI'd go with 3.5 if someone guide me properly23:27
*** timcowchip has joined #crux23:47
AbdullahHow can I do this for 3.5 all ports so if I do ports -u it wont get again 3.4 ports ?23:48
*** timcowchip has quit IRC23:54
TimB_change the files in /etc/ports I guess23:55
TimB_Abdullah: I tried termite btw, I'll keep st ;)23:55
AbdullahTimB_: were you able to install it?23:56
TimB_Abdullah: yes, you need to initialize another repo inside the source dir.. and you need to install vte3-ng23:56
AbdullahI have vte3-ng. can I have your pkgfile?23:57
Abdullahfor termite23:57
TimB_I didn't use one for this quick test23:57
TimB_maybe I can provide you one (dirty-ish) Pkgfile23:57
TimB_you just need to initialize another repo inside the sources before building23:57
Abdullahplease create one. I were unable to do that.23:57
Abdullahyeah I tried that. but I was unable to understand how arch and slackbuild were doing that. I was unable to uderstand their variables23:58
*** cruxbot has joined #crux23:58
TimB_he, I too, look at what other packagers do. I understand your struggle23:59
TimB_let me have a look in some minutes23:59
TimB_both machines are running some jobs right now23:59
Abdullahokay I'm waiting23:59

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