IRC Logs for #crux-devel Sunday, 2015-01-04

Romstergcc: error: unrecognized command line option '-fdiagnostics-color=auto'02:10
Romsteron core/kmod02:10
Romsteri think it's a gcc 4.9.x option02:11
Romsterwhy is it now an error and before it was not.02:11
jaegerI'm not seeing any issues with kmod02:31
jaegerhave you tried it on a stock install?02:51
Romsterwell clean docker image yeah02:58
jaegerwithout distcc/ccache?02:59
Romstermaybe i can spin another one up without distcc ccache and try02:59
jaegerI can't reproduce it here03:29
*** mavrick61 has quit IRC03:49
*** mavrick61 has joined #crux-devel03:51
diverseah that's neat, maybe the next crux version can enable that on?04:08
RomsterGCC 4.9 added support for colorizing the warning and error output to make it easier to understand what went wrong during compilation of some source code. This probably wasn't added due to it being a good idea, but because clang has this feature. I'm of the opinion that it's a good idea, but the GCC team probably isn't, as it's not turned on by default, and in order to use this feature you must specify some command line parameters or use some04:13
Romsterenvironment variables.04:13
Romsterits for gcc 4.9.x04:13
diverseand we are stuck with 4.8.x until the next crux release04:14
Romsterso gcc 4.8.x has no idea what that option is... so yes lets add options and expect everyone to be running gcc 4.9.x than to use configure to test if gcc suports it04:14
jaegerwhat does it have to do with kmod?04:14
Romsterit failed to build. but then i have yet to test it in a clean docker image of no distcc/ccache.04:15
jaegerIt definitely isn't failing here with 4.8.4 for what that's worth04:15
Romsterwhy should that change the behavior of gcc 4.8.4 that's spitting out the error?04:15
diverseRomster: could you try with just a distcc configuration to see if the problem still occurs?04:16
diverse(without ccache)04:16
jaegerbefore the 4.8.4 push I bootstrapped 2 full ISOs without any issues04:17
jaegerwith 4.8.404:17
diverseI had no problems either04:18
Romsterof course it works without my ccache/distcc04:18
Romsterjust tested this04:18
diverseRomster: how about distcc only?04:18
jaegerAre you still surprised by that? hehe04:18
jaegerI truly don't mean to sound antatonistic, but it does seem to cause issues often04:19
Romsterits like i should jsut itch it and go buy a god damn workstation xeon dual cpu setup.04:20
jaegerI wouldn't go that far, I'd just always test it in a vanilla install when a problem like this crops up04:20
jaegerfor the things that work it's damn nice to have the speed boost04:20
Romsteri should just to make sure.04:20
Romsterbefore reporting.04:21
Romsterhmm perhaps i need to see if this is ccache or distcc and then submit a bug report to what causes it.04:21
diverseI do plan to play with distcc, but I would like to know if it's really ccache causing the problems04:21
jaegerI had to drop distcc and ccache both back when I was working on multilib, it caused problems so many times04:22
jaegerspent hours chasing them04:22
Romsterwaiting for wine to finish i don't have my cluster on atm actually... so i'm just compiling on my desktop04:22
Romsterhmm well i have no issues other than kmod to get to wine.04:22
diversejaeger: what about with just distcc?04:23
jaegerthey both caused lots of problems at the time04:23
Romsteras long as all the distcc hosts are running the same tool chain.04:23
Romsterthat's the only catch04:23
Romsteras for ccache04:23
Romsterexport CCACHE_COMPILERCHECK="%compiler% -dumpversion; crux"04:24
jaegerthis was back before we had official multilib, back when I was trying to figure it out again since danm and nipul both stopped theirs04:24
jaegerso pre 2.8 I think04:24
Romsteror some way to diffinerate between tool chains and gcc versions.04:24
jaegerI never used different toolchains, figured that was just asking for trouble04:24
jaegerI used a mix of fedora, gentoo, and LFS to test things but never mixed them for distcc usage04:25
Romsteri was trying for multilib then too i think... but i had all sorts of problems boostrapping so i left it to jaegers ports. then i converted my pure 64bit unofficial system to multilib.04:25
Romsterand jaeger mostly did the 32bit ports i think i stepped in and started taking over like per usual.04:30
jaegerYou use a lot more ports than I do in both architectures04:31
Romsterits mostly codecs an libs04:31
Romsteri've been trying to cut back.04:32
Romsterto be honest i haven't been using skype so much but who's gonna take over maintaining it.04:35
jaegerwhoever needs it04:39
diverseI think skype is going to fall on the waste side when tox becomes more dominate04:40
*** mechaniputer has quit IRC06:01
*** leo-unglaub has joined #crux-devel10:08
*** leo-unglaub has quit IRC11:42
*** diverse has quit IRC12:40
*** leo-unglaub has joined #crux-devel14:38
*** crash_ has joined #crux-devel16:31
jueRomster: we relased kmod 19 in midth of november last year and got no reports about any build issues, why do you think that we have a general build problem with kmod? Isn't it more obvious to suppose a local problem on your side?17:37
*** diverse has joined #crux-devel17:39
*** diverse has quit IRC17:39
*** diverse has joined #crux-devel17:39
jueif you look at you'll see that it's part of the tests to explore whether gcc supports the diagnostics-color option or not17:43
*** Feksclaus has joined #crux-devel17:45
*** crash_ has quit IRC17:49
*** leo-unglaub has quit IRC18:05
*** leo-unglaub has joined #crux-devel18:11
*** mechaniputer has joined #crux-devel20:06
*** leo-unglaub has quit IRC20:58
frinnstyeah I was wondering about that21:35
*** Workster has joined #crux-devel22:14
*** Romster has quit IRC22:42
*** Romster has joined #crux-devel22:43
diversehas anyone heard of miniz?22:43
diversecould this be a faster alternative to zlib?22:44
prologicdo they show benchmarks?22:50
diverseprologic: I guess there is this:
prologicso it'a faster? :)22:57
diverseprologic: yeah it shows compared to the common ones we use23:01
prologicfair enough23:01
prologicI always up for optimizations23:01
prologicbut you know...23:02
prologicnew things can break23:02
prologicsoftware developers don't write code nearly as well as they did once :)23:02
diversewhat do you mean "as they did"? Software developers never wrote code well ever23:03
diverseso yeah, lack of code maturity could be a problem23:05
diverseprologic: I think what you mean is, developers never write good designs as they once did23:07
diversegiven all this super-hip, let's store everything in binaries23:09
diverseand make stuff do everything, but not everything right23:10
diverseand be user-cryptic as possible23:18
prologicactually they died once23:27
prologicwhen it was important to23:27
prologicand resources were limited23:27
prologicyour funny :)23:28
prologicyes okay bad designs too23:28
diverseyou like what I did there, huh?23:45
*** Romster has quit IRC23:48
*** Workster has quit IRC23:55
*** Workster_ has joined #crux-devel23:55
*** Workster has joined #crux-devel23:55

Generated by 2.11.0 by Marius Gedminas - find it at!