IRC Logs for #crux Tuesday, 2018-03-06

*** xor29ah has joined #crux00:01
*** Romster has quit IRC00:02
cruxbot[xorg.git/3.3]: [notify] xorg-libxcb: updated to 1.13. Now depends on libinput00:06
cruxbot[xorg.git/3.3]: xorg-xcb-proto: updated to 1.1300:06
cruxbot[xorg.git/3.3]: xorg-iceauth: updated to 1.0.800:06
xckospeaking of libinput, is synaptics still needed when libinput is installed/00:07
*** xor29ah has quit IRC00:10
*** xor29ah has joined #crux00:10
j_vnot sure, maybe someone else knows more about that... sorry, will be afk for a while00:10
jaegerI believe it's not needed, but mine is an elantouch or whatever, not synaptics00:11
xckookay I'll just remove it and see what happens00:11
joacimits annoying when manufacturers have 2-3 different touchpad drivers on their support pages00:15
joacimhave to guess which one i need the driver for00:15
*** xor29ah has quit IRC00:16
*** xor29ah has joined #crux00:17
xckothose of you running crux on laptops, whats your strategy for power management?00:20
xckotlp, laptop-mode-tools, custom scripts in acpid?00:21
*** Roomster has joined #crux00:22
joacimi let linux run the show for controlling the cpu00:22
joacimi use acpid to detect certain events related to suspend and hibernate00:22
jaegerI haven't spent any time on power management yet00:24
jaegerI used to use powertop in the past00:24
jaegerusually powersave or userspace cpufreq governor00:24
xckojoacim: use the conservative cpufreq governor then?00:25
xckojaeger: is powersave included with powertop?00:25
*** Roomster has quit IRC00:26
joacimi made powersave the default in my kernel ocnfig00:28
*** Romster has joined #crux00:33
*** Romster has quit IRC00:37
jaegerIt's one of the standard cpufreq options in the kernel00:38
xckogotcha thanks00:40
*** [EaX] has joined #crux00:49
*** xor29ah has quit IRC00:50
*** Romster has joined #crux00:51
*** [EaX] has quit IRC00:54
*** Romster has quit IRC00:55
*** xor29ah has joined #crux00:57
*** Romster has joined #crux00:58
*** Romster has quit IRC01:02
*** Romster has joined #crux01:06
xckoI've noticed programs like dwm and st have a config.h included to configure the program, yet if config.h is edited, then neither the md5sum or signature succeed. also updating the ports tree resets config.h. Is there a way to do something like gentoo's USE=savedconfig so the file is preserved and the port gets updates?01:06
joacimit could be named config.h.default instead01:10
joacimthen the user could copy it to config.h01:10
*** Romster has quit IRC01:10
xckoseems sensible. the default config in the src is named config.def.h -- maybe follow the same convention?01:12
xcko(not trying to be rude - sorry)01:12
joacimif the user didnt make a config.h, the Pkgfile could instead use the config.h.default01:12
*** Romster has joined #crux01:13
joacimoh dont worry. i dont get offended by better suggestions =)01:13
joacimnever used dwm, so i dont know their default naming schemes01:13
joacimi guess you dont even need to use the externall bundled config.h. since dwm will use its own config.def.h if config.h is missing01:15
joacimbuilding dwm was surprisingly quick =)01:16
*** Romster has quit IRC01:18
joacimj_v: is dwm your port?01:19
xckomaybe no external config.h is included with the port but the Pkgfile checks for one, and if it exists, then it copies it into place01:20
xckoit'd make it harder to configure, as the default config.h would only be in the source01:21
xckowhoops sorry for the blinky - idk how I did that - butter fingers01:21
joacimmy irc client ignores text formating too =)01:22
xckonice what do you use?01:22
joacimi think you have to say something mean about the sami to offend me01:22
joacimweechat right now. i had a similar config when i used irssi01:22
*** xor29ah has quit IRC01:23
xckoI gotta be afk for a bit, will be back later01:24
*** xcko has quit IRC01:24
*** Romster has joined #crux01:27
*** Romster has quit IRC01:32
joacimi think the current config should be bundled, but with a different name. make the Pkgfile copy the config if the user made one.01:32
joacimif [ -e $SRC/config.h ] ; then cp $SRC/config.h . ; fi01:34
*** Romster has joined #crux01:42
*** SiFuh has quit IRC01:44
*** SiFuh has joined #crux01:45
*** Romster has quit IRC01:48
*** Romster has joined #crux01:56
v0latileAfter searching all my backups I found my application to deal with package dependencies when removing them :-)01:59
*** tilman_ has joined #crux02:04
*** Romster has quit IRC02:04
*** Romster has joined #crux02:06
*** Romster has quit IRC02:10
*** xor29ah has joined #crux02:15
*** Romster has joined #crux02:18
*** Romster has quit IRC02:23
*** Romster has joined #crux02:25
*** chinarulezzz_ has quit IRC02:29
*** Romster has quit IRC02:32
j_vjoacim: Are you saying, have the one in the port directory named something like 'config.def.h', and then only copy if it finds 'config.h'?02:53
*** xor29ah has quit IRC02:54
j_vSince the config.h in the port directory is an exact copy of the config.def.h in the dwm source, I'm not clear I see the benefit here. The dwm build routine copies the config.h, whether changed or not, into the dwm source directory. So if it is not changed by the user, it is merely copying in a verbatim copy of config.def.h, which exactly what the Makefile will do if it doesn't find a config.h.02:57
*** xor29ah has joined #crux02:59
j_vSo, if the user alters the config.h, it gets copied into the source directory anyways, and dwm will be built with the altered config.h.02:59
j_vAm I missing some finer point with this?03:00
j_vWhat I think I should add to the dwm port README is that if the config.h is edited, then the md5sum will need to be ignored or updated to match the config.h.03:02
*** Romster has joined #crux03:22
*** _________mavric6 has quit IRC03:53
*** _________mavric6 has joined #crux03:54
jaegerj_v: at a guess, I suspect the maintainer did that to give the user an easy choice to make their own local copy of the port if they want to customize it04:01
j_vjaeger: Yep. I'm just trying to understand how joacim is suggesting I change it. I'm willing to improve it. I'm just not clear how how using a 'if..then' to copy the file in conditionally makes it better. Maybe I'm being dense. Wouldn't be the first time.04:09
j_vNot sure what a 'how how' is, slightly embarrassed that I typed it.04:10
jaegerI don't really see it either, for what that's worth04:18
j_vOk, thanks.04:25
*** Romster has quit IRC05:12
*** Romster has joined #crux05:16
*** Romster has quit IRC05:21
*** Romster has joined #crux05:23
*** brian|lfs has joined #crux05:32
*** v0latile has quit IRC06:13
*** v0latile has joined #crux06:13
*** abenz has joined #crux06:29
*** abenz has quit IRC06:30
*** abenz has joined #crux06:32
*** abenz has quit IRC07:36
*** SiFuh_ has joined #crux07:37
*** abenz has joined #crux07:37
*** SiFuh has quit IRC07:39
*** xcko has joined #crux07:42
*** workodera has joined #crux08:06
*** SiFuh_ has quit IRC08:09
*** SiFuh has joined #crux08:10
*** xcko has quit IRC08:17
joacimj_v: the problem is, when you edit the config.h, you start getting md5sum and signature errors08:27
joacimso it would be better to have the config.h not tracked by the signature and md508:27
joacimeg. rename it to config.def.h, and let the user copy it to config.h if they want apply their own config08:28
joacimits a suggestion to solve a problem that someone had08:29
joacimi dont think it is good to tell people that they need to work around errors in a readme. i think it is better if the user get no errors at all08:30
*** SiFuh has quit IRC08:40
*** SiFuh has joined #crux08:41
*** abenz_ has joined #crux09:22
*** abenz has quit IRC09:26
*** abenz__ has joined #crux09:49
*** abenz_ has quit IRC09:52
*** v0latile has quit IRC10:42
*** chinarulezzz_ has joined #crux10:56
*** Evoken has joined #crux11:16
*** frinnst has quit IRC11:16
*** frinnst has joined #crux11:18
*** frinnst has quit IRC11:18
*** frinnst has joined #crux11:18
*** wan7m0re has joined #crux11:22
frinnstman I dont like the lwn subscriber view11:27
*** workodera_ has joined #crux12:06
*** workodera has quit IRC12:06
*** onodera has joined #crux12:09
*** onodera has quit IRC12:19
*** ][_R_][ has quit IRC12:30
*** ][_R_][ has joined #crux12:32
*** SiFuh_ has joined #crux13:02
*** SiFuh has quit IRC13:02
cruxbot[compat-32.git/3.3]: [notify] libinput-32: initial import for xorg-libxcb-3213:03
cruxbot[compat-32.git/3.3]: [notify] xorg-libevdev-32: initial import for libinput-3213:03
Romster5 commits and the bot only shows 213:09
Romsteri have no idea what i missed in this time my net has been a yoyo frinnst13:14
Romsterdoesn't look like a lot13:14
Romsterthey say it's not fully fixed yet. but at least my internet is useable now13:15
frinnstwonder if libinput really is needed for compat-3214:30
cruxbot[contrib.git/3.3]: opus: remove doc pages if docbook-xsl is installed14:32
cruxbot[compat-32.git/3.3]: mpg123-32: 1.25.8 -> 1.25.1014:38
cruxbot[compat-32.git/3.3]: libdrm-32: 2.4.90 -> 2.4.9114:38
Romsteri was wondering that too frinnst with all the stuff going on and my internet acting up, i don't know anymore.14:46
Romstercan someene test for me14:47
frinnstDont have any 32 bit stuff here, so cant test easily im afraid14:52
frinnstbut if everything worked fine before it should be safe to turn off. the changelog for libxcb just mentioned it's now enabled by default14:55
Romsteri'll look into it more on the weekend15:00
joacimsomeone shrieked at me for using evdev a while ago. libinput is the new thing that everyone should use?15:07
frinnstwell it works in wayland so its more future proof. but both should work without issue15:17
*** g0relike has joined #crux15:22
*** g0relike has quit IRC15:36
*** g0relike has joined #crux15:48
*** abenz_ has joined #crux15:59
*** hiei has joined #crux16:33
*** john_cephalopoda has joined #crux16:35
*** workodera_ has quit IRC16:35
*** JanC has joined #crux16:42
j_vjoacim: A good point.17:15
j_vjoacim: I think there is a chicken and egg problem with the solution.17:17
joacimnot sure if i see it17:18
j_vIf the file is not in the source array, it won't be available to build()17:19
j_vIf the file is in the source array, but missing, md5sum and signature errors also.17:21
j_vA bit of a conundrum17:22
joacimdrop the port for being annoying =)17:22
j_vWhat I'm leaning toward is to remove the config.h stuff from the Pkgfile, etc, but explain how to fork the port in the README17:23
j_vor maybe drop... idk yet17:23
*** v0latile has joined #crux17:24
joacimmm. i guess there is no way around these errors17:52
fun-j_v: use PKGMK_ROOT17:54
fun-It is the port dir17:55
j_vAh, nice, that should make joacim's method work.17:56
fun-What method? To drop it?17:57
j_vthe 'only copy config.h in if it exists' thing. I like it, and using the PKGMK_ROOT was the one thing standing in the way of it.18:01
j_vor, well, not using it was standing in the way, i guess18:02
joacimrm -rf $PKGMK_ROOT18:02
j_vwould be some interesting bug reports ;)18:04
*** v0latile has quit IRC18:06
fun-One disadvantage is that won't work for those using your repo with the official git driver, which deletes local files,  a18:07
fun-rmull created a bug report.18:09
j_vhmmm. ok. something to consider.18:11
j_vI wonder if something with 'git stash' in the git driver would help that18:16
j_vmight get complicated, because that would probably cause merge conflicts on 'git stash pop' or 'git stash apply'18:19
*** onodera has joined #crux18:22
*** v0latile has joined #crux18:30
rmullI vote in favor of just dropping the dwm port. I've been using it for years outside of package management. It sees very few updates that it's not worth the heartburn18:36
j_vrmull: Do you think it could be considered a starting point for forking it as a private port? Like, for someone new to CRUX?18:40
j_vThough tbh, most people who use software for any length of time will usually be aware of how to configure and compile it.18:41
rmullSure, I don't see why not18:48
j_vI currently use bspwm, so no heart break for me if I drop it. I'm going to think about for a bit before I act one way or the other. It's very little work for me to maintain, so in that respect, not a big deal to keep as a reference either.18:48
j_vI do appreciate all the input18:49
j_vWould be nice to find a solution the git driver18:51
rmullI personally feel that the git driver blowing away all downloaded and built files is a harmful bug19:03
rmullBut I haven't submitted a patch19:04
*** leah2 has joined #crux20:02
*** pedja has quit IRC20:11
dbrookeI build dwm outside of package management too; in addition to the build time config I also patch it for some alternative/additional behaviour20:40
*** onodera has quit IRC20:44
*** onodera has joined #crux20:45
j_vyeah, that makes sense to me. I know when I used dwm, one of my beefs with it was layout and I vaguely remember patching it for that, though I don't remember specifics. Been using bspwm for 3-4 years at least, maybe more.20:46
dbrookegiven that I apparently last built it in Oct 2015 I can't remember specifics either 8-)20:50
j_vFor now, I think I'm going to leave dwm. It does no harm, though I think joacim has a good point about the config.h file, so even if I keep maintaining dwm port, I will probably adjust the Pkgfile to copy in a config.h file if found in PKGMK_ROOT. It will still run into issues with the git driver cleaning it out, but that I should probably make the dwm/README more verbose anyways and can cover that there.20:54
dbrookedebian provide a package with default config and note that it's just a taster and you probably want to build it yourself21:03
j_vThat is probably how the dwm/README should explain it, also21:04
j_vI'm thinking that this same discussion could be had for many of the ports that are based (dmenu, st, tabbed, ...)21:05
dbrookeprobably, though for dmenu I just use a custom dmenu_run script in ~/bin to get what I want21:08
j_vyep. I've been learning to use dmenu for other things, also. It can be pretty handy, especially if you don't go in for all that DE stuff.21:09
dbrookeI have a ~/.dmenu directory with files which each contain the command I want executed and the script is just exe=`ls ~/.dmenu  | dmenu  ${1+"$@"}` && exec `cat ~/.dmenu/$exe`21:10
j_vOne thing I'd like to try to ensure, is that I don't want to force specific configs on anyone who uses these types of ports.21:10
j_vNice. That sounds very useful. I will probably encorporate that idea in my own stuff.21:12
dbrookeso all the things I use often I just drop a file in there and anything infrequent I just start from cli in the ever present screen session where I can leave it running and just open another shell21:17
j_vFirst script I wrote that utilizes dmenu is for starting qemu instances base on a conf file which sets image files, vga options, vnc stuff, etc. Brings up a dmenu with all available conf's, runs qemu on that with appropriate viewer, usually vnc.21:21
j_vIt's pretty hackish, but generally does what I need.21:23
dbrookeneat, it's quite flexible really, being essentially a selector in a pipeline21:24
j_vyeah. it's pretty cool that way. I've looked at a few spinoffs, but so far what I've seen doesn't seem any better. I'm currently using a modified patch for lineheights and xyw, which is pretty handy for placement.21:28
*** wan7m0re has quit IRC21:31
darfomaybe install the dwm config in /etc so pkgadd will put it through rejmerge21:35
darfoconditionally symlink the /etc copy in build() if it is present21:35
darfoand note in the readme where to put user updates so they don't lose their21:35
darfonot sure this will work (tm)21:36
j_vdarfo: that's definitely something to consider / play with21:37
darfoi don't use dwm any more but when I did I just put a fork in my local repo to keep my config21:38
j_vI think that is probably the most common. And I'm leaning toward the contrib/dwm port (if it's to remain) being suggested as a reference.21:39
j_vI could play with stuff like suggesting an exported $DWM_CONFIG, but I think that may be getting more complicated than needed.21:40
darfoyes a reference with a fat warning in the readme about what happens on the next update21:49
darfowhich reminds me, it's time to do another full system backup :)21:50
j_vI've been getting away from that and mainly keeping backups of etc, home, and built pkgs.21:52
darfoI do that regularly. But about twice a year I do a full backup just in case21:52
darfosomething important got put in an unlikely place21:52
j_vHmmm, I have the disk space. I think I will take that as advice, since an unused backup is better than one that is missed when needed.21:54
darfowell, the corollary is "did I check that I can actually do a restore?"21:54
darfothat one always makes me laugh in sick sort of way21:55
j_vtoo true. i finally packaged my preferred backup method (dirvish). it's really easy to configure, but it's been a long time since i tested restoring. bad habits so easily return.21:56
darfoI use an external drive for my semi-annual backup which is just an rsync with a lot of excludes21:57
j_vdirvish is just perl controlling rsync, using hardlink method, so it's not that special, just really easy to setup/configure.21:58
darfoSounds kinda like BackupPC21:58
darfohow is dirvish on doing restores?21:59
j_vit doesn't, it's up to the user to do that part22:00
darfofair enough. i might look into to do cross-system backups for the important, fast-changing stuff you mentioned.22:01
darfolook into dirvish i mean22:02
j_vi have a port in my j_v repo if you want a reference22:02
darfok, thanks22:03
j_vi'd be interested in improvements if you see room for it, have time for it, etc22:03
darfodo you use the dependency p5-time-period from alan repo?22:05
darfolooking at the Pkgfile. Not an easy one to package. You might want to add dev/ to exclude:22:09
darfoor maybe not. It's not that much to back up and sometimes it is nice to22:10
darfosee what state it was in22:10
*** onodera has quit IRC22:36
j_vdarfo: sorry was afk. yes, p5-time-period from alan repo. and yes, exclude dev/ is a my preferred, as well as stuff like run/ and tmp/22:39
j_vI will add those to the default 'master.conf.ex', thanks for pointing that out.22:42
frinnstbtrfs + btrbk <322:43
j_vyeah, been meaning to get btrfs going. will probably install 3.4 to btrfs... that's the plan, anyways. already added your btrbk to my mpup.lst so I don't forget about it.22:44
j_vwill be afk for a while22:45
frinnstI plan to add it to opt or something. needs a bit of love first22:46
*** v0latile has quit IRC23:24
*** v0latile has joined #crux23:25
*** john_cephalopoda has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!