| 00:00.42 | *** join/#devuan Ryushin (user@windwalker.chrisdos.com) |
| 00:05.45 | *** join/#devuan luke-jr (~luke-jr@unaffiliated/luke-jr) |
| 00:05.51 | luke-jr | "Select and install software" fails :/ |
| 00:06.04 | luke-jr | and I can't find a log to tell me why |
| 00:07.37 | *** join/#devuan clopez (~tau@neutrino.es) |
| 00:14.03 | luke-jr | got past that manually and completed install.. after booting, network isn't brought up because the device is named enp0s4 but the config looks for eth0 |
| 00:17.16 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 00:38.40 | DocScrutinizer05 | OUCH, *that* mess |
| 00:38.56 | DocScrutinizer05 | another systemd/udev fsckup |
| 00:39.15 | DocScrutinizer05 | interface names too long to even get handled by standard tools |
| 00:59.33 | *** join/#devuan Inocuous (~Inocuous@44-117-181-66.dsl.sacoriver.net) |
| 01:00.18 | *** join/#devuan luke-jr (~luke-jr@unaffiliated/luke-jr) |
| 01:10.17 | *** join/#devuan luke-jr (~luke-jr@unaffiliated/luke-jr) |
| 01:32.13 | *** join/#devuan Hoshpak (~Hoshpak@p2003005B4B21F000503823E1104833C3.dip0.t-ipconnect.de) |
| 01:39.36 | *** join/#devuan Defiant (erik@x50abd710.dyn.telefonica.de) |
| 01:56.09 | *** join/#devuan golinux (~golinux@2605:6000:ebc6:d200:96de:80ff:febc:9ea8) |
| 01:56.09 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 02:18.00 | *** join/#devuan cdanderson42 (~nuhrin@67.170.89.70) |
| 02:27.17 | *** join/#devuan dormito (~quassel@cpe-24-243-2-136.satx.res.rr.com) |
| 03:11.00 | DocScrutinizer05 | luke-jr: what did you do?? |
| 03:11.38 | DocScrutinizer05 | ...has left this server (Excess Flood). |
| 03:11.51 | luke-jr | DocScrutinizer05: I only have crap ISP available |
| 03:12.07 | luke-jr | TCP packet lost = IRC server eventually gets a bunch of data at once |
| 03:12.25 | DocScrutinizer05 | aaah yep, that happens |
| 03:14.39 | ksx4system | luke-jr: get ZNC :) |
| 03:24.57 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 03:55.28 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 03:57.59 | DocScrutinizer05 | well, that *could* even happen with a bouncer, when the link between bouncer server and IRC server is crap |
| 04:00.17 | *** join/#devuan DocScrutinizer05 (~saturn@openmoko/engineers/joerg) |
| 04:05.29 | *** join/#devuan da80 (~user2@67.58.164.15) |
| 04:14.05 | luke-jr | in fact, I am using ZNC locally ;) |
| 04:28.28 | *** join/#devuan d413 (~user@67.58.164.15) |
| 05:02.12 | *** join/#devuan towo^work (~towo@unaffiliated/towo/x-4064351) |
| 05:13.44 | *** join/#devuan Chanku (~Chanku@c-68-52-188-31.hsd1.tn.comcast.net) |
| 05:21.39 | *** join/#devuan pav5088 (~markp@58-7-210-12.dyn.iinet.net.au) |
| 05:30.23 | *** join/#devuan cdanderson42 (~nuhrin@c-73-221-62-247.hsd1.wa.comcast.net) |
| 05:48.39 | *** join/#devuan detha (~detha@unaffiliated/detha) |
| 06:53.49 | *** join/#devuan Deys (~deys@182.48.232.113) |
| 06:59.21 | *** join/#devuan joered (~pingus@93-58-161-247.ip159.fastwebnet.it) |
| 07:11.18 | *** join/#devuan egon0 (~quassel@2001:4dd0:ff00:8d2c:854e:edb6:3f97:3ad2) |
| 07:25.24 | *** join/#devuan TwistedFate (~TwistedFa@unaffiliated/twistedfate) |
| 07:46.38 | *** join/#devuan dave24 (~dave24@92.207.128.213) |
| 07:53.13 | *** join/#devuan b0stik (~b0stik@host166-40-static.225-95-b.business.telecomitalia.it) |
| 08:19.01 | *** join/#devuan cocoadaemon (~foo@x53.octopuce.fr) |
| 08:32.36 | *** join/#devuan kelsoo (~kelsoo@86.53.50.178) |
| 08:44.04 | *** join/#devuan kelsoo1 (~kelsoo@86.53.50.178) |
| 08:45.13 | *** join/#devuan rleigh (~rleigh@unaffiliated/rleigh) |
| 08:50.19 | *** join/#devuan kelsoo (~kelsoo@86.53.50.178) |
| 08:58.28 | *** join/#devuan jotun (~marvin@2001:41d0:e:275::1) |
| 09:18.29 | *** join/#devuan Deys (~deys@182.48.232.113) |
| 09:19.15 | *** join/#devuan antofox (4e0c384a@gateway/web/freenode/ip.78.12.56.74) |
| 09:19.36 | antofox | morning| |
| 09:23.06 | *** join/#devuan usr1 (~user@78-69-19-70-no188.tbcn.telia.com) |
| 09:48.11 | jaromil | luke-jr: hi!! |
| 09:50.04 | jaromil | luke-jr: reading backlog about weird device name. is that happening with devuan?! we need to sort that then |
| 09:50.36 | jaromil | ethernet devices should just be ethN with progressive numbering from 0. period. |
| 10:07.26 | *** join/#devuan rleigh (~rleigh@unaffiliated/rleigh) |
| 10:12.08 | *** join/#devuan penelopa (~penelopa@unstable.nl) |
| 10:30.09 | *** join/#devuan debdog (~debdog@HSI-KBW-091-089-090-004.hsi2.kabelbw.de) |
| 10:31.55 | *** join/#devuan clopez (~tau@neutrino.es) |
| 10:41.30 | Centurion_Dan | jaromil: It's a known issue with squeeze, and stems from here... https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ |
| 10:41.31 | sat0shi | Title: PredictableNetworkInterfaceNames (at www.freedesktop.org) |
| 10:42.02 | Centurion_Dan | slaps sat0shi |
| 10:42.24 | Centurion_Dan | sat0shi? |
| 10:46.55 | Centurion_Dan | !sat0shi |
| 10:46.56 | sat0shi | Centurion_Dan: Error: "sat0shi" is not a valid command. |
| 10:47.07 | Centurion_Dan | !owner |
| 10:47.08 | sat0shi | Centurion_Dan: Error: You don't have the owner capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified. |
| 10:47.33 | Centurion_Dan | slaps sat0shi again. |
| 10:48.38 | *** part/#devuan sat0shi (~avocado@2a03:b0c0:0:1010::20a:e001) |
| 10:52.47 | detha | Suppose one wants to run a local devuan mirror repo, and pull the debian part from an already existing local debian mirror, does that mean one has to run one's own amprolla setup? |
| 10:53.46 | hellekin | detha: sounds like it. |
| 10:54.33 | detha | hmm. any pointers to doc on how to go about that? |
| 10:57.17 | *** part/#devuan penelopa (~penelopa@unstable.nl) |
| 10:57.31 | Centurion_Dan | detha: no. You can just mirror the /devuan and /merged part and then setup a web redirect to your local debian mirror. |
| 10:58.39 | hellekin | oh, that's interesting |
| 10:58.51 | detha | Centurion_Dan: thanks, cool. Which mirror doesn't mind rsync? |
| 10:59.12 | hellekin | detha https://devuan.org/os/mirrors |
| 11:00.24 | Centurion_Dan | No hard documentation on the process yet, but once we have a few more mirrors in place we'll likely push a package out that makes it dead easy for either private local mirrors or official mirrors. |
| 11:00.53 | Centurion_Dan | we don't allow public rsync at this stage from our mirrors. |
| 11:01.13 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 11:02.05 | hellekin | Centurion_Dan: since when? |
| 11:02.37 | Centurion_Dan | there is a mirror sync script here tho: https://git.devuan.org/snippets/6 |
| 11:03.30 | Centurion_Dan | hellekin: it's always been that way as far as I can tell. We don't currently support anonymous rsync. |
| 11:03.53 | detha | I see. That gives the release ISOs, but not the repository. |
| 11:04.47 | hellekin | Centurion_Dan, detha this script was updated (but I don't remember why) at https://git.devuan.org/snippets/8 |
| 11:06.15 | hellekin | detha: mirroring the release ISOs is documented at https://files.devuan.org/MIRRORS.txt and that's a different thing |
| 11:06.42 | hellekin | Centurion_Dan: I confused with files.do that accepts rsync. |
| 11:07.53 | detha | Background to all this: servers here do not have internet access. They do however have access to a local mirror server. Just trying to figure out how to populate that mirror server for devuan |
| 11:08.18 | Centurion_Dan | hellekin: your snippets link is dead. Packages mirroring script is https://git.devuan.org/snippets/6 |
| 11:08.56 | jaromil | hellekin: Centurion_Dan is talking about amprolla, not files |
| 11:09.10 | hellekin | https://git.devuan.org/snippets/8 ? I was just comparing it with 6, and the variables are cleaner and chmod snae. |
| 11:09.12 | hellekin | sane |
| 11:10.53 | jaromil | detha: if servers don't have internet access, then that's tricky atm. nextime has well clear in his todo the part that makes amprolla able to sync mirrors, but it is not there yet |
| 11:11.10 | jaromil | that's another good reason why we call this beta |
| 11:11.45 | jaromil | however, the best people to inform you about the state of amprolla are Centurion_Dan and nextime |
| 11:12.19 | Centurion_Dan | hellekin: I get a 404 when looking at snippets 8 |
| 11:12.45 | jaromil | detha: to stay up to date keep an eye on https://git.devuan.org/devuan-infrastructure/amprolla/activity and https://git.devuan.org/devuan-infrastructure/amprolla/issues |
| 11:13.16 | jaromil | Centurion_Dan: me too 404 on the snippet 8 |
| 11:13.23 | detha | jaromil: thanks, will do |
| 11:13.24 | hellekin | Centurion_Dan: weird... At worst it should give you 401. |
| 11:15.18 | Centurion_Dan | is it a private snippet?? |
| 11:16.33 | hellekin | wtf. Now it's public :P |
| 11:17.05 | Centurion_Dan | detha: I can walk you through the url rewriting side of things if you like. IT |
| 11:17.14 | hellekin | Would snippets be private by default? In another person's project? Least surprise :P |
| 11:18.30 | detha | Centurion_Dan: I would very much like to know a bit more about it yes. Let me do some reading up first though, just started looking at this |
| 11:19.29 | Centurion_Dan | the rewrites pretty straight forward - I can share the nginx configs, and the plan is to put them on the repo mirrors to make propagation and mirroring easier - especially for leveraging CC mirrors. |
| 11:24.17 | *** join/#devuan pedro_ (~pedro@217.111.112.186) |
| 11:24.27 | detha | Configs would be nice, I'll have to convert them to apache for $reasons, but that shouldn't be hard |
| 11:26.25 | Centurion_Dan | pm me your email address and I'll send them to you. |
| 11:26.34 | *** join/#devuan kelsoo (~kelsoo@86.53.50.178) |
| 11:26.49 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 11:27.24 | detha | Centurion_Dan: thanks |
| 11:29.14 | *** join/#devuan rleigh (~rleigh@unaffiliated/rleigh) |
| 11:30.04 | pedro_ | Hi there, I know devuan is in an early stage, but is there anybody that build an ami for amazon ec2? |
| 11:30.05 | *** join/#devuan Deys (~deys@1.186.37.113) |
| 11:31.52 | Centurion_Dan | detha: if you can send me back any hints on converting that to apache style redirects, then that would be appreciated - or better still open an issue in https://git.devuan.org/devuan-infrastructure/amprolla/issues with that info and we will add that functionality into amprolla ;-) |
| 11:32.41 | detha | Centurion_Dan: I will do, once I've got it working properly. |
| 11:33.03 | jaromil | :^) |
| 11:42.20 | *** join/#devuan oscarandrea (~oscar@net-37-117-248-215.cust.vodafonedsl.it) |
| 11:42.37 | *** join/#devuan oscarandrea (~oscar@unaffiliated/oscarandrea) |
| 11:42.40 | oscarandrea | hi |
| 11:43.16 | hellekin | pedro_: what's an ami? I don't think Devuan will ever produce anything for Amazon. Maybe the community will. But it sounds way too far from a base OS functionality to be in Devuan. My guess. |
| 11:43.24 | oscarandrea | hi, you can do on devuan pinning the testing version as on debian without any problems ? |
| 11:45.04 | hellekin | oscarandrea: pinning works the same |
| 11:45.48 | oscarandrea | thanks :D |
| 11:46.23 | pedro_ | hellekin: An AMI is an image of the OS you can run in amazon EC2. AFAIK the official Debian EC2 AMI's are built with http://bootstrap-vz.readthedocs.io/en/master/ |
| 11:46.59 | pedro_ | So it should not be that difficult (I guess) to build Devuan images |
| 11:47.50 | pedro_ | Amazon has a big slice of the cloud and right now there are not that much non-systemd options there. The amazon linux ami and freebsd. |
| 11:50.07 | oscarandrea | in " Pin : release n = stretch" stretch insert right ? |
| 11:56.07 | hellekin | pedro_: if Amazon clients want Devuan, I guess Amazon will create the image, won't they? |
| 11:56.41 | hellekin | I tend to think that free labor for union-buster corporation is wasted work. |
| 11:57.04 | KatolaZ | hellekin: seconded |
| 11:59.48 | Centurion_Dan | hellekin: but having an AMI image may bear much fruit for us... |
| 12:04.59 | hellekin | Centurion_Dan: I'll be glad when we have that image. I just don't think it's our role to make it. It sounds a bit like asking tmux to implement linux-specific code to replace a 30 years old generic interface. That just sounds wrong. |
| 12:05.35 | KatolaZ | hellekin, Centurion_Dan: it might not be terribly hard to do that, after all |
| 12:05.48 | KatolaZ | since it should be not that much different from a vm-image |
| 12:05.57 | hellekin | yes |
| 12:06.19 | hellekin | I'd also be happy if competitors to Amazon had a similar image first :P |
| 12:06.23 | KatolaZ | it might be just another flavour in the vm-sdk |
| 12:06.32 | hellekin | possibly |
| 12:06.52 | KatolaZ | but I agree that it is probably not a pressing issue :) |
| 12:08.33 | oscarandrea | boys are the "testing " repository of devuan (in / merged / dists / |
| 12:08.33 | oscarandrea | ) |
| 12:08.33 | oscarandrea | but are not present testing the security update , what you tell me to do ? |
| 12:08.33 | oscarandrea | I'm updating to devuan testing |
| 12:11.36 | Centurion_Dan | we're not generating the security repo for testing yet... |
| 12:12.11 | oscarandrea | ok thanks |
| 12:13.42 | *** join/#devuan jsfarinet (~chatzilla@p2003008C4F26D3A952B7C3FFFE13D4E6.dip0.t-ipconnect.de) |
| 12:15.07 | pedro_ | emmmm, seriously, I do not think that creating an AMI that works an AWS EC2 is helping union-buster company. That said, almost every single company that produces technology is more or less on that group |
| 12:15.29 | pedro_ | I assume everybody is aware of the working conditions in china and the like |
| 12:16.19 | oscarandrea | excuse ignorance then I can use the security packages for debian jessie and other packages use the testing repo ? :) |
| 12:17.12 | pedro_ | That said, this is the thing: when people run a server in the cloud they have choices and most of their choices right now are privative or systemd choices |
| 12:20.19 | *** join/#devuan thms (~thms@unaffiliated/thms) |
| 12:20.21 | thms | Hi smart people |
| 12:21.52 | oscarandrea | excuse ignorance then I can use the security packages for debian jessie and other packages use the testing repo ? :) |
| 12:24.18 | KatolaZ | oscarandrea: I doubt security updates from jessie would be any useful for testing |
| 12:24.28 | KatolaZ | mainly due to package version mismatch |
| 12:26.44 | oscarandrea | thanks :) |
| 12:27.08 | oscarandrea | however, you do a great job with devuan :) |
| 12:27.16 | oscarandrea | now exit, bye :) |
| 12:37.38 | *** join/#devuan debdog (~debdog@HSI-KBW-091-089-090-004.hsi2.kabelbw.de) |
| 12:45.42 | openfbtd | Heh. Jokes aside, at some point the question of a code of conduct *will* be raised seriously in the scope of Devuan. |
| 12:46.08 | openfbtd | Not by some troll, probably by someone inside the project |
| 12:46.35 | openfbtd | Will probably be as much of a shitshow as the rest of these proposals :( |
| 12:47.12 | djph | openfbtd: "these proposals" being "I propose a code of conduct"? or in-general proposals that've come up for Devuan? |
| 12:47.23 | openfbtd | Proposals of a CoC. |
| 12:47.59 | openfbtd | From what I see, they quickly get brigaded by crybullies who want something like the contrubutor covenant. |
| 12:48.07 | djph | yeah, those are typically a trainwreck |
| 12:48.27 | openfbtd | Even if the original proposal had nothing to do with such a CoC :) |
| 12:48.34 | Centurion_Dan | I proposed a Code of Conduct recently for the NZ Open Source Society, because all the others were far too wordy to be useful and effective. |
| 12:48.46 | djph | ... it's too bad "do unto others..." and/or "seriously, don't be a dick" aren't good enough. |
| 12:49.05 | djph | or rather, aren't seen as "good enough" |
| 12:49.18 | openfbtd | djph, but they are. âThere are no rules. There are no rules about moderation either, enjoy your ban.â :) |
| 12:50.22 | openfbtd | Conduct rules hurt communities more than they help, because when you have rules, you have a whole other class of trolls who game the system while technically not breaking any. |
| 12:50.23 | Centurion_Dan | It basically boiled down to that simple don't be a dick, and if you are a dick then we will deal appropriately with you if somebody complains about it. |
| 12:51.01 | djph | openfbtd: I meant that every SJW will be on their high horse about how terribad Devuan is because the CoC doesn't cover every eventuality for every person's fragile "I'm a special snowflake" complex. |
| 12:51.47 | openfbtd | And it ends up in a childish âYou broke a ruleâ âNuh-uh I did not, check the wording!â âOk you're bannedâ *writes an op-ed for some bullshit internet magazine about Devuan being bullies who ignore their own rules* |
| 12:52.40 | Centurion_Dan | openfbtd: I disagree that they hurt communities, but it's usually that they are introduced after the shit has it the fan and the result is a fracturing of the community. |
| 12:53.32 | openfbtd | Centurion_Dan, I've seen that stupid game of ânuh-uh I didn't technically break any rulesâ so many times even in communities that didn't have a massive drama about that. |
| 12:53.52 | Centurion_Dan | djph: Most CoC's don't define their terms which makes them ineffective and open for abuse - usually by the SJW's in the first place. |
| 12:54.10 | openfbtd | It's extremely limiting. You end up putting yourself into a position when you either have to ignore your own rules or amend them every time someone finds a hole |
| 12:54.15 | openfbtd | And it's a losing game |
| 12:54.51 | openfbtd | I think the best way to deal with it is not to have rules but guidelines that are stated to be such |
| 12:54.54 | djph | yeah, I've seen that as well |
| 12:55.15 | openfbtd | And in those guidelines you also say that moderators can kick you off for whatever reason. |
| 12:55.23 | *** join/#devuan figlfdev (183e6d83@gateway/web/freenode/ip.24.62.109.131) |
| 12:55.52 | fsmithred | benevolent dictatorship is very useful and appropriate in some situations |
| 12:56.05 | djph | (1) "Don't Panic" (2) "Don't be a dick" (3) "If you're deemed to being a dick, mods will ban you" |
| 12:56.07 | openfbtd | Basically it's the âspirit of the lawâ vs âword of the lawâ |
| 12:56.49 | Centurion_Dan | Not really neccesarily, but in the NZOSS case we had to shut down the CoC discussion because it was turning into the sort of flamewar that would require a CoC to clean up the casualties... |
| 12:57.21 | openfbtd | Oh btw yeah, I love the tactic of all these CoC proposals. |
| 12:57.33 | openfbtd | â You guys are awful, you need a code of conduct |
| 12:57.38 | openfbtd | â No we're not, fuck off |
| 12:57.39 | figlfdev | was just looking at the botbot logs and found part of this conversation there |
| 12:57.45 | Centurion_Dan | :-[ |
| 12:57.56 | openfbtd | â Hey, see how rude that guy is? You need a CoC! |
| 12:58.08 | figlfdev | when jaromil suggested it, i nearly shat myself. im glad this community at least isnt gung ho about such a problematic "solution" |
| 12:58.08 | openfbtd | *facepalm* |
| 12:58.58 | fsmithred | if your parents didn't teach you a code of conduct, it's probably too late. |
| 12:59.02 | figlfdev | has anyone considered the "no code of conduct?" granted its got a couple clauses that are too childish for serious use, but its modifiable. |
| 12:59.19 | Centurion_Dan | we have a CoC... it's called the banhammer... |
| 12:59.41 | figlfdev | Centurion_Dan: yeah that works actually. |
| 12:59.50 | openfbtd | In any case, I hope the devuan community won't make the mistake of adopting a rigid ruleset. |
| 12:59.56 | figlfdev | for fun or inspiration: https://github.com/domgetter/NCoC |
| 13:00.26 | figlfdev | openfbtd: yeah i was concerned they might. and that it would ultimately be "for show" like the same rules used to kick whats now (us) off debian ml |
| 13:01.23 | figlfdev | "lets just use the rules from the debian mailing list, those are never abused" >:D |
| 13:02.34 | Centurion_Dan | too wordy... |
| 13:02.45 | figlfdev | definitely |
| 13:02.50 | openfbtd | We have an IRC network that was basically forced into adopting *some kind of a ruleset* |
| 13:02.57 | openfbtd | This is what I came up with: https://www.hypeirc.net/pages/rules.html |
| 13:03.22 | openfbtd | The people who wanted rules solidified were not happy :) |
| 13:03.30 | figlfdev | nice |
| 13:03.38 | fsmithred | thanks, figlfdev. lmao at "Your name is offensive. " |
| 13:04.58 | openfbtd | One of the things some of those people wanted is to make channels obey their own rules |
| 13:05.16 | openfbtd | Fuck that. It's their rules, they can break them :D |
| 13:07.05 | openfbtd | AFAIR even freenode doesn't enforce that btw. |
| 13:07.06 | figlfdev | now if you want *wordy*, here is the most intelligent article ive ever read on geeks, rules and communities: (then no more spam, hopefully) https://medium.com/@maradydd/when-nerds-collide-31895b01e68c (long article is looooooong) |
| 13:07.23 | openfbtd | And freenode is the most fascist of any IRC network I've seen so far |
| 13:07.51 | openfbtd | >While we enÂcourÂage all chanÂnel opÂerÂaÂtors to adopt our chanÂnel guideÂlines, if a project or comÂmuÂniÂty deÂcides to opÂerÂate othÂerÂwise then we reÂspect their deÂciÂsions. |
| 13:07.58 | openfbtd | What the hell |
| 13:08.11 | openfbtd | https://freenode.net/kb/answer/channels then |
| 13:09.07 | djph | that looks like a script for a text-to-speech parser |
| 13:09.10 | openfbtd | https://8fw.me/hell/8d86f7a0093a3be4e223790a4487609af661b582.png D: |
| 13:10.03 | openfbtd | Apparently it's this: https://en.wikipedia.org/wiki/Soft_hyphen |
| 13:10.46 | openfbtd | Oh, it's for breaking words |
| 13:10.59 | openfbtd | Just use text-align: justify >_> |
| 13:11.13 | *** join/#devuan Levure (~quassel@143.186-65-87.adsl-dyn.isp.belgacom.be) |
| 13:13.10 | djph | Q: Why don't you care about my feelings? |
| 13:13.13 | djph | We are not a support group for human emotion. We are a community that strives to focus around our topics. Anything else takes away from that. |
| 13:13.16 | djph | hehehe |
| 13:13.45 | djph | I've either had too little coffee in my whisky, or too much whiskey in my coffee ... |
| 13:24.05 | Centurion_Dan | alrighty then, on that note I'm off to bed... |
| 13:28.07 | hellekin | bad note to go to bed to |
| 13:28.57 | hellekin | Centurion_Dan: we're humans, we do care about your emotions, and there's no topic that is off-topic (except systemd) |
| 13:30.22 | figlfdev | caring about emotions is a wonderful and important quality, but its dangerous as a policy |
| 13:30.33 | figlfdev | hi hellekin |
| 13:32.21 | figlfdev | the threat of codification of caring is emotional blackmail holding up an entire project. its really too weird to believe, unless youve seen it happen more than once |
| 13:32.43 | figlfdev | (and anyone escaping systemd should be able to tell that lennart and co are brilliant at it) |
| 13:33.39 | figlfdev | "i dont always hate disabled people, but when i do, i use sysvinit" |
| 13:35.13 | hellekin | codification of caring ? |
| 13:35.25 | figlfdev | making it official policy / mandatory |
| 13:35.47 | figlfdev | caring works better when it is voluntary |
| 13:35.55 | figlfdev | (imo. always imo) |
| 13:36.30 | figlfdev | but just to talk about devuan for a minute, ive been running fsmithreds unofficial live since last night. good stuff |
| 13:37.11 | hellekin | oh, no way, no CoC. CoCs are terrible. I agree with < fsmithred> if your parents didn't teach you a code of conduct, it's probably too late. |
| 13:38.27 | figlfdev | yeah-- and we probably WILL have some kind of CoC at some point. only with the greatest reluctance and nose-holding i hope. |
| 13:38.38 | hellekin | there's still the https://talk.devuan.org/guidelines but they're not satisfactory either. |
| 13:38.51 | figlfdev | possibly involving throwing something into a volcano. |
| 13:39.36 | hellekin | In any case, I don't see a CoC gaining any authority with that crowd. People who are idiotic enough to not abide by reasonable social rules won't follow the rules of a CoC. |
| 13:39.58 | hellekin | So it's like a faith issue I guess. |
| 13:40.25 | *** join/#devuan Chanku (~Chanku@c-68-52-188-31.hsd1.tn.comcast.net) |
| 13:40.43 | figlfdev | hellekin: are you familiar with the merry pranksters? have you ever heard of them? |
| 13:42.04 | jaromil | I think our NCoC is gonna be: act quick to trolling, silence is consensus, if conflict is risen... golinux decides. lol |
| 13:42.15 | figlfdev | oh sweet lord |
| 13:42.24 | figlfdev | must remember to stay on golinuxs good side |
| 13:42.36 | jaromil | its very hard, that's why :^D |
| 13:42.44 | figlfdev | :D |
| 13:43.09 | figlfdev | jaromil: the merry pranksters were a band of hipsters/hippies decades ago, when they were more common (this sort of has a point.) |
| 13:43.41 | figlfdev | they told some pentagon officials they wanted to levitate the pentagon building 30 feet in the air. |
| 13:43.57 | figlfdev | the officials said "well no, you cant because if it fell down people could get hurt." |
| 13:44.13 | figlfdev | so they asked if they could levitate it just 3 or 5 feet or so |
| 13:44.21 | figlfdev | the officials said they guessed that would be ok. |
| 13:44.38 | *** join/#devuan Ryushin (~Ryushin@50-203-182-130-static.hfc.comcastbusiness.net) |
| 13:44.57 | figlfdev | and thats what "community" politics makes me think of :) |
| 13:45.07 | hellekin | I like jaromil's CoC :) |
| 13:45.31 | *** join/#devuan Akuli (~Akuli@mobile-access-5d6a5c-116.dhcp.inet.fi) |
| 13:47.37 | hellekin | lol at the Pentagon flight. How can someone come up with the idea that levitating a building could hurt people, and then think it wouldn't hurt people *if it was detached froom the ground anyway* (without mentioning the integrity of the structure). Oh well, if you give the choice between levitating 30 fett or levitating 3 feet, obviously nobody is gonna ask about "not levitating". |
| 13:49.26 | figlfdev | at least they had a policy, even if it was ad-hoc :) |
| 13:49.54 | figlfdev | and 3 is definitely better than 30 for this. |
| 13:56.13 | *** join/#devuan dardevelin (~dardeveli@unaffiliated/dardevelin) |
| 13:58.37 | *** join/#devuan golinux (~golinux@2605:6000:ebc6:d200:96de:80ff:febc:9ea8) |
| 13:58.37 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 13:58.50 | figlfdev | the possible integration of plain window managers with devuan style is pretty impressive |
| 13:58.55 | figlfdev | speak of the devil :) |
| 13:59.22 | figlfdev | is reading the ml about jwm |
| 14:01.03 | *** join/#devuan premoboss (~utente@62.212.13.223) |
| 14:03.09 | figlfdev | im mostly inclined toward icewm, but ive stuck the devuan logo in the "start" button before http://i.imgur.com/BbvVYBY.png |
| 14:03.49 | *** join/#devuan Deys (~deys@1.186.37.113) |
| 14:07.08 | hellekin | I'm totally inclined towards letting the user choose what their desktop will be. And I'd love to see great desktop configurations pop up, and eventually pick from those for default Devuan desktop. But personally I don't see myself using a desktop, so I'm ust suggesting how Devuan could do it better :) |
| 14:08.10 | figlfdev | i love the vt, but im always using a desktop because gimp and firefox dont play nice without it. |
| 14:08.43 | figlfdev | for those reasons im always using a tabbed term instead of say, tmux (though i still feel for what theyre going through) |
| 14:09.27 | djph | I like the minimalist look that I got with whatever DE I chose to install (MATE?) ... |
| 14:09.53 | figlfdev | gnuinos has openbox and tint2, and theyre really nice. |
| 14:10.21 | figlfdev | though tint2 configuration is a disaster if you want to change it. maybe if it only had one profile instead of 2. |
| 14:11.10 | *** join/#devuan cht0 (~cht0@2001:981:487a:1:14eb:278e:756e:e3b) |
| 14:12.29 | figlfdev | tint2s deceptively simple-looking config gui is completely unintuitive. i havent even bothered trying to get it to work after seeing what it does every time i try to move the taskbar |
| 14:12.34 | djph | about the only thing I don't like is that ~everything~ is white on black, and not different colors based on file "type" |
| 14:13.03 | figlfdev | what do you normally use, thunar? pcmanfm? |
| 14:13.03 | djph | although, that might've been me breaking my alias file |
| 14:13.05 | hellekin | djph: that's a shell configuration |
| 14:13.09 | djph | *aliases |
| 14:13.19 | djph | yeah, I know |
| 14:13.23 | figlfdev | ahh |
| 14:13.43 | figlfdev | you mean the term |
| 14:14.47 | jsfarinet | Just wait a couple of days, i'll try to put on devuan.talk our outlines for jwm - so you can have a look at them ;) |
| 14:14.59 | djph | yup, that was it, broke .bashrc |
| 14:18.16 | hellekin | figlfdev: what, you can't read minds? ;o) |
| 14:18.58 | figlfdev | well i just pictured icons first :) he was talking about mate so the term wasnt the first thing i imagined |
| 14:19.03 | golinux | I haven't read all the scrollback . . . just catching up. |
| 14:19.16 | golinux | But CoC are a waste of time IMO. |
| 14:19.34 | hellekin | golinux: \o/ |
| 14:19.34 | figlfdev | and thought, mostly |
| 14:19.38 | golinux | A CoC will not change human behavior |
| 14:20.09 | golinux | Hi hellekin! |
| 14:20.13 | hellekin | golinux: the one line you need to read from the backlog: < jaromil> I think our NCoC is gonna be: act quick to trolling, silence is consensus, if conflict is risen... golinux decides. lol |
| 14:20.38 | golinux | Thyat's funny!\ |
| 14:20.43 | figlfdev | terrifying. |
| 14:20.54 | Chanku | What about CoC's |
| 14:20.58 | golinux | :D |
| 14:21.06 | figlfdev | it reminds me of the schoolbus driver that says "quiet down or the bunny gets it!" |
| 14:21.08 | hellekin | Chanku: you missedit, sorry :] |
| 14:21.11 | golinux | Chanku: read the scrollback |
| 14:21.27 | figlfdev | but it could work. |
| 14:21.30 | Chanku | golinux, my scrollback doesn't have anything about a CoC |
| 14:21.32 | hellekin | figlfdev: no clue what that means but sounds funny |
| 14:21.43 | golinux | Look on botbot |
| 14:21.53 | Chanku | botbot? |
| 14:22.05 | golinux | https://botbot.me/freenode/devuan/ |
| 14:22.07 | figlfdev | hellekin: in south park theres a schoolbus driver that when the kids get loud, she puts a gun to a bunnys head and says "you kids shut up or the bunny gets it!" |
| 14:22.14 | hellekin | figlfdev: I heard about a school teacher who managed to get a silent classroom by threatening to write on the blackboard the names of the next victims in Games of Thrones. |
| 14:22.24 | figlfdev | LOL |
| 14:22.46 | figlfdev | thats brilliant |
| 14:23.33 | golinux | Anyway, CoC are a waste of time and energy. There is no cure for stupidity |
| 14:23.33 | figlfdev | agreed |
| 14:23.33 | hellekin | +1 |
| 14:23.36 | Chanku | CoC's also end up being used to silence dissent or people that disagree on non-code things :/ |
| 14:23.52 | Chanku | does not use Github, or Atom and the like, because of CoC's and the like |
| 14:25.02 | figlfdev | only when they wont accept the decision of the debian TC, Chanku ;) |
| 14:25.47 | golinux | Anyway back to catching up. |
| 14:26.37 | *** join/#devuan phylophyl_ (~phylophyl@126.red-88-20-12.staticip.rima-tde.net) |
| 14:27.07 | Chanku | I only kinda sorta like FreeBSD's CoC |
| 14:28.15 | figlfdev | one of the side effects that rarely gets talked about (except by the people leaning on the CoC that want to silence people as quickly as possible) |
| 14:28.50 | Chanku | hm? |
| 14:28.50 | figlfdev | is how much time gets spent Talking about the policy once it has to be "used" -- it can distract people for days (im totally guilty here) |
| 14:28.58 | Chanku | Yeah |
| 14:29.13 | golinux | <Centurion_Dan> we have a CoC... it's called the banhammer... LOL! |
| 14:29.33 | golinux | golinux misses all the fun while sleeping. |
| 14:30.27 | Chanku | I'll be honest...I've only seen CoC's used to push an agenda by SJW's |
| 14:30.44 | figlfdev | thats not (technically) all theyre good for, but its what they do best |
| 14:30.50 | figlfdev | and most often |
| 14:31.21 | Chanku | has a bit of a dislike of SJW's and that religion... |
| 14:31.29 | figlfdev | cult |
| 14:31.36 | Chanku | Yeah... |
| 14:31.43 | Chanku | is still recovering from that a bit |
| 14:32.17 | figlfdev | i was an sjw for a few months. i dont know how that happened, tbh (never again) |
| 14:32.48 | Chanku | I was actually an SJW (or if not that, damn close to it) for a bit... |
| 14:33.05 | Chanku | thankfully I was able to realize the BS and broke away |
| 14:33.07 | figlfdev | met a few too many racists-- never liked fascism though i started thinking being sjw-like would help (ha! never helps) |
| 14:33.25 | Chanku | Honestly SJW's just prove Horseshoe theory :P |
| 14:33.35 | figlfdev | horseshoe? |
| 14:34.38 | Chanku | Basically the extremes of either side of the political spectrum wrap around |
| 14:35.00 | Chanku | so the extreme left and extreme right are closer than the moderate-left and moderate-right |
| 14:35.01 | figlfdev | oh yeah |
| 14:35.37 | Chanku | I actually only got out of that due to GG and the Amazing Atheist...albeit I no longer watch him :/ |
| 14:36.05 | golinux | On caring and emotions . . . Emotional responses are in your own head. Only you and no one else is responsible for them |
| 14:36.35 | Chanku | Emotions are weird |
| 14:36.56 | golinux | Tiptoeing around other people buttons is an unhealthy co-dependency |
| 14:37.02 | luke-jr | jaromil: yes |
| 14:37.28 | Chanku | I naturally tiptoe around other people to avoid conflict at times |
| 14:37.57 | figlfdev | golinux: how long do you think we can keep up this wild-west philosophy? (as long as possible i would hope) |
| 14:38.24 | figlfdev | or has bsd managed it for decades?(in that case i suppose its possible) |
| 14:39.01 | golinux | It's not a matter of duration but one of clarity on the dynamics of interactions |
| 14:39.14 | *** join/#devuan jathan (~jathan@132.248.174.116) |
| 14:39.14 | figlfdev | same question, your wording |
| 14:39.37 | figlfdev | how well do you think we can maintain that level of clarity? |
| 14:39.39 | golinux | Mine was a statement not a question |
| 14:39.54 | figlfdev | either way you worded it fine |
| 14:40.20 | golinux | :) |
| 14:40.55 | golinux | Back to catching up . . . |
| 14:42.40 | dave24 | I have less political question: is devuan actually going to stick with sysvinit? or will there be several "flavours" of devuan with different inits, or just pick a decent one. |
| 14:43.24 | dave24 | because, tbh, sysvinit is not much better than systemd |
| 14:43.25 | figlfdev | imagines that sysvinit will be around for a while, but it will become increasingly easy to swap inits out |
| 14:43.46 | figlfdev | sysvinit ITSELF is not much better than systemd. there is maybe one way in which it is far superior |
| 14:43.59 | figlfdev | it doesnt try to take over. its there, if you replace it nothing bad happens |
| 14:44.16 | Chanku | actually I would argue sysvinit itself is better by systemd more than you realize |
| 14:44.36 | KatolaZ | dave24: it seems that the priority for the moment is to get rid of systemd, but it looks like the long-term plan is to provide support for several init systems |
| 14:44.47 | figlfdev | you could use a cement block as init, and if you want to use a different one just remove the block. |
| 14:45.07 | figlfdev | a cement block wont try to subsume hotplugging or desktop functionality |
| 14:45.17 | figlfdev | so its better than systemd, even if its a lousy init. |
| 14:45.41 | figlfdev | i think the only reason sysvinit is being used is because it has most of what devuan needs, and it was already there. |
| 14:45.48 | dave24 | yeah, i agree, systemd is vomit. point is sysvinit is by no means the way forward. imho daemontools-style supervisors are the way forward. |
| 14:45.53 | Chanku | sysvinit is also battle-tested |
| 14:46.01 | figlfdev | ive never met a hardcore sysvinit fan. it almost takes systemd to make sysvinit look good :) |
| 14:46.18 | KatolaZ | dave24: everybody has his/her own views on which init is the best |
| 14:46.19 | figlfdev | yeah, if it aint broke... but obviously people have leaned a little too hard on it for too long |
| 14:46.34 | figlfdev | i think devuan will improve that, but i cant imagine scrambling to do so |
| 14:46.45 | dave24 | KatolaZ: (plan for other inits): ah ok, good. |
| 14:46.55 | KatolaZ | and whetever my take or yours on the matter, you will always find thousands more people who disagree |
| 14:46.57 | algernon | figlfdev: fwiw, lots of stuff broke horribly when I tried to replace sysvinit with file-rc on Debian a good few years ago. |
| 14:46.58 | KatolaZ | :) |
| 14:47.25 | dave24 | KatolaZ: yes, hence several "flavours" like you currently have with DEs |
| 14:47.26 | figlfdev | algernon: only because file-rc isnt really setup to do what you need, i figure |
| 14:47.40 | figlfdev | and by "setup" i mean "integrated with/configured for debian" |
| 14:47.50 | KatolaZ | my understanding is that the priority is to have a functional low-lever userland that does not depend on systemd |
| 14:47.53 | figlfdev | more the latter: configuration than the former: "integration" |
| 14:48.46 | figlfdev | systemd creates too much um, what did golinux say? "codependency" |
| 14:48.54 | algernon | figlfdev: yep. I could make it work, but that wasn't a simple "replace it and nothing bad happens" situation. |
| 14:48.57 | figlfdev | at least in debian. and sysvinit did that, but not to the same degree |
| 14:49.15 | figlfdev | the goal of devuan (afaict) is to reduce, but at least not increase that level of "integration" |
| 14:49.50 | KatolaZ | I would probably prefer a very basic rc-style init anyway, but I understand that other users might disagree on this point |
| 14:50.01 | algernon | alternative inits alwayso required at least some adjustment, and meant some kind of pain. depending on what you swithed from and to, the pain and effort varied, but neither was ever zero. |
| 14:50.07 | figlfdev | well the init was never truly modular as in "put anything else here and it works." but the level of connectedness (dependency) is a runaway train, now. |
| 14:50.11 | KatolaZ | especially for personal workstations/laptop sysvinit is too much more than what is really needed |
| 14:50.52 | KatolaZ | algernon: there was a discussion on the DNG ML a while ago about possible strategies to make init systems really pluggable |
| 14:51.06 | figlfdev | i mean you could (if you were insane) do roughly the same thing with sysvinit that systemd is doing by design-- but no one would set out to do that and it wouldnt be easy as creating a monolith |
| 14:51.07 | KatolaZ | but I don't think there is a consensus on that as yet |
| 14:51.15 | algernon | KatolaZ: futile effort, tbh. |
| 14:51.49 | figlfdev | no one set out to make sysvinit a dependency hell. it only became as integrated as it is over decades of minor abuses-- not a plan for world domination or "unification" |
| 14:52.03 | figlfdev | and decades later, it is still something you could (relatively) easily replace |
| 14:52.31 | *** join/#devuan rleigh (~rleigh@unaffiliated/rleigh) |
| 14:53.09 | figlfdev | worlds over, its a better place to start for truly modular/free init. thats probably the nicest thing you can say other than "it mostly works." |
| 14:53.59 | Chanku | I'm still looking forward to the guy who made uselessd's init system |
| 14:55.03 | algernon | thinks inventing yet another init (or YA syslogd, etc) is a terrible idea, but whatever people want to spend their time on.. :P |
| 14:55.44 | Chanku | Yet Another Init already exists |
| 14:56.01 | Chanku | Yes there is an init with that name! |
| 14:56.04 | *** join/#devuan furrywolf (~randyg@184-194-18-78.pools.spcsdns.net) |
| 14:57.22 | Chanku | (It's my init :P) |
| 14:57.32 | openfbtd | algernon, competition is great and reinventing the wheel should be done whenever you think you can do better |
| 14:57.53 | Chanku | My init exists because I wanted to see if I could do it in Python |
| 14:57.54 | Chanku | I could |
| 14:57.55 | Chanku | and I did |
| 14:57.56 | openfbtd | Even the phrase âdon't reinvent the wheelâ is actually pretty stupid |
| 14:58.06 | openfbtd | People reinvented the wheel a lot of times |
| 14:58.12 | KatolaZ | algernon: ? |
| 14:58.31 | Chanku | Although from what I can tell...the init is going to be something we really haven't seen before |
| 14:59.08 | algernon | openfbtd: like I said, its their time. I still think it's a terrible idea, but I'm not going to stop anyone to pursue their interest. |
| 14:59.36 | openfbtd | Chanku, is there any actual info on that btw? |
| 14:59.39 | algernon | just boots into emacs and is done with it :P |
| 14:59.51 | Chanku | On what exactly...the Uselessd guy's init system? |
| 15:00.03 | openfbtd | Yeah |
| 15:00.06 | Chanku | Ah |
| 15:00.11 | Chanku | Nothing really public |
| 15:00.14 | KatolaZ | boots into emacs too, and he does not have systemd hanging around.... |
| 15:00.29 | Chanku | I do have access to some private information though |
| 15:12.42 | nat0 | Does Devuan package runit? |
| 15:12.59 | dave24 | nat0: debian does, so devuan picks it up |
| 15:13.47 | dave24 | nat0: but it'll be a bit of a pain to set up i think |
| 15:15.25 | nat0 | Hmm I want to try that in a VM for shits and giggles. |
| 15:15.33 | nat0 | Devuan's just using SysV right? |
| 15:15.44 | dave24 | nat0: yes |
| 15:15.55 | dave24 | let me know how it goes |
| 15:18.02 | djph | ugh, how does sysv jump from 'default runlevel 2' in /etc/inittab to 'runlevel 5' |
| 15:20.59 | KatolaZ | djph: ? |
| 15:25.03 | djph | trying to follow the init process - inittab "says" it defaults to runlevel2 as default. Trying to figure out where it swings up to runlevel 5 (i.e. started X) |
| 15:26.30 | fsmithred | debian doesn't use runlevels 3-5 |
| 15:26.41 | fsmithred | 2 is multi-user graphical |
| 15:27.03 | fsmithred | it can seem harsh for someone coming from redhat or suse |
| 15:27.44 | djph | well, seems I have to mess with that then |
| 15:28.00 | *** join/#devuan SylvieLorxu (~TheLastPr@dhcp-077-251-165-191.chello.nl) |
| 15:28.36 | *** join/#devuan Pali (~pali@Maemo/community/contributor/Pali) |
| 15:29.01 | fsmithred | some distros reconfigure runlevel 3 to be multi-user console |
| 15:29.34 | fsmithred | sidux? I think. Refracta for sure. probably a few others, because I've run into people who expect it to be that way. |
| 15:30.07 | fsmithred | update-rc.d is the command for reconfiguring runlevels |
| 15:30.20 | fsmithred | so you don't have to mess with the links manually |
| 15:30.21 | figlfdev | is really glad not *everyone* in devuan hates python |
| 15:30.33 | fsmithred | sysv-rc-conf is even easier |
| 15:30.34 | figlfdev | im not denying its bloated. its a kitchen-sink language. |
| 15:30.45 | djph | well, runlevel 1 is "single-user", 2 is "user defined", 3 is "multi-user w/networking", 4 is "user defined", 5 is "multi-user + X" |
| 15:30.56 | furrywolf | I hate python. :P |
| 15:31.05 | figlfdev | lots of people here hate python |
| 15:31.11 | figlfdev | i wouldnt try to change their mind, either |
| 15:31.21 | figlfdev | because for their own criteria theyre probably right. |
| 15:31.33 | figlfdev | but why do *you* hate it? (just curious) |
| 15:31.42 | fsmithred | djph, that's true for most major linux distros, but not debian |
| 15:31.53 | fsmithred | or its derivatives |
| 15:32.07 | furrywolf | C is by far my favorite language, and that's sad, as in 40+ years something better should have been developed. |
| 15:32.50 | figlfdev | go? swift? rust? c#? (excuse me. i would never use c# but it must have some advantages for some people.) |
| 15:33.02 | Chanku | likes Python |
| 15:33.17 | figlfdev | me too. though i hope someone picks up python 2 when the foundation drops it any year now |
| 15:33.21 | Chanku | C++ :P |
| 15:33.25 | figlfdev | they could fork and call it "boa" |
| 15:33.25 | Chanku | uses Python 3 |
| 15:33.44 | figlfdev | c++ is a myth, albeit a persistent one |
| 15:33.50 | djph | ah ha, no, it's not in runlevel2 |
| 15:34.00 | djph | Debian threw it into runlevel S |
| 15:34.10 | fsmithred | what's in runlevel 3? |
| 15:34.14 | djph | ... or ... whatever 'S' is meant to signify |
| 15:34.29 | fsmithred | S is single user |
| 15:34.44 | figlfdev | the thing about python 3 is that i was happy with bytes by default. now i have to specify everything that was default before. and it doesnt even work as reliably. |
| 15:34.58 | dave24 | anyone whos ever used the python C api will probably hate it. |
| 15:35.10 | Chanku | figlfdev oh? What do you mean by that? |
| 15:35.19 | figlfdev | i mean "explicit is better than implicit." oh alright then, so i hope python 4 requires me to do c-level memory management too! |
| 15:35.20 | djph | goddammit, back to digging through manpages |
| 15:35.22 | djph | .... |
| 15:35.48 | fsmithred | djph, what are you trying to do? I missed the beginning of the conversation. |
| 15:35.55 | figlfdev | because what i really want out of python is to constantly tell it how to massage a mix of ascii and utf, when python2 did every single thing i needed and python 3 does none of it. |
| 15:36.22 | figlfdev | ffs, just give me from __past__ import strings # (im sure that is impossible.) |
| 15:36.33 | furrywolf | I started designing my own language once, but never finished it. I took some of the best parts of C, perl, c++, java, lisp, and even php... |
| 15:36.51 | figlfdev | but all that aside-- there are plenty of professionals that would still rather python 2. including one i know that used to work for nokia. |
| 15:36.56 | Chanku | isn't utf-8 compatible with ascii? |
| 15:37.02 | figlfdev | Chanku: not at all. |
| 15:37.13 | dave24 | figlfdev: it is |
| 15:37.16 | figlfdev | it can REPRESENT ascii. but its not compatible |
| 15:37.22 | djph | okay, so it goes from single-user mode (which starts X) to runlevel 2 |
| 15:37.26 | dave24 | figlfdev: its backward compatible |
| 15:37.28 | figlfdev | dave24: if you mean the 7 bit standard yes |
| 15:37.37 | djph | ... |
| 15:37.39 | dave24 | not extended |
| 15:37.40 | figlfdev | dave24: so load a binary file and save it using strings. |
| 15:37.51 | figlfdev | cant do it. the encoding will corrupt it. |
| 15:38.37 | furrywolf | C syntax, OO, proper strings, all types bounds checked (I love pointer arithmetic and raw arrays, but history has proven letting people have access to them creates security issues), a security model for running foreign code, and even a built-in rdbms... but designing something like that is a LOT of work. |
| 15:38.55 | Chanku | Actually UTF-8 was specifically designed for backwards compatibility with ASCII |
| 15:39.13 | fsmithred | oh, I never noticed that x11-common starts in single-user. No idea what it does. |
| 15:39.14 | dave24 | figlfdev: binary is not ascii |
| 15:39.15 | figlfdev | Chanku: its only backwards compatible in a way that doesnt help at all in this case. |
| 15:39.23 | Chanku | Well what are you using it for? |
| 15:39.30 | figlfdev | dave24: youre technically right but in a way that totally misses my point |
| 15:39.37 | figlfdev | and i understand the difference fwiw. |
| 15:39.58 | figlfdev | this is about something suiting a purpose, which it doesnt-- not about whether it meets a spec- which it does. |
| 15:40.31 | figlfdev | and python 2 does. but not 3. switching to 3 has hurt the project i work on |
| 15:40.46 | figlfdev | and i cant find any way around it except to sacrifice quality (which is what i did) |
| 15:40.53 | figlfdev | i cant reproduce the functionality of 2. |
| 15:41.16 | Chanku | Well I mean I don't even know what you are even going for, I could probably explain to you how to do it if I knew what you were going for specifically :/ |
| 15:41.38 | figlfdev | Chanku: im pretty sure mine is an "edge case" thats WONTFIX. |
| 15:42.06 | figlfdev | but i dont care. i hope the pros that still prefer python 2 will fork it between now and 2020. because i will only use 3 long-term if i have no choice |
| 15:42.08 | Chanku | Well we won't know until you say. |
| 15:42.11 | figlfdev | short term, im using 3 just in case. |
| 15:42.30 | DocScrutinizer05 | !owner |
| 15:43.04 | Chanku | So what exactly are you trying to do with python 2? |
| 15:43.08 | figlfdev | chanku: i have a programming language that uses python2 strings for everything. loading binary files, displaying text in a term, storing urllib contents in an array, etc |
| 15:43.25 | Chanku | okay |
| 15:43.48 | figlfdev | now ive done my best to make it use utf8 in python 3, but if that fails (which it does) it falls back to whatever encoding is compatible with 8-bit bytestrings. |
| 15:43.48 | Chanku | well you can use .encode(encoding) for moving between utf-8 and ascii :/ |
| 15:44.15 | figlfdev | oh i do... its a nightmare. try # use utf8 except # use (other encoding) |
| 15:44.26 | Chanku | okay? |
| 15:44.37 | figlfdev | no it doesnt work. it is a hack. |
| 15:44.50 | figlfdev | it doesnt reproduce the flexibility of python 2 strings, because that is impossible. |
| 15:45.26 | figlfdev | now if i write a program in my language to load a binary file, it will save and load but it will be corrupted. |
| 15:45.35 | figlfdev | useless. but at least it runs. |
| 15:46.32 | figlfdev | ive read EVERYTHING i can about encoding. but all i want to do with encoding is simulate/reproduce python 2 strings. otherwise versions 3.x of my language are not 100% compatible with 2.x of my language. |
| 15:47.11 | Chanku | Honestly I would argue it's for the better as Python 2 strings were a cludge of sorts. |
| 15:47.13 | figlfdev | when the python foundation drops 2, i will already be using 3 (i do now.) but if some wonderful group forks it a la mate, trinity, devuan... then i will go back to python "2" (i think they should call the fork "boa") |
| 15:47.45 | figlfdev | Chanku: thats propaganda. it may be a kludge to python 3 users (maybe thats "everyone") but you can use the same argument as to why devuan is unnecessary. |
| 15:47.59 | figlfdev | no offense. just saying, its not a fair argument. |
| 15:48.18 | Chanku | I'd argue the underlying code would be a cludge...just like most of systemd :/ |
| 15:48.25 | figlfdev | touche. |
| 15:48.35 | figlfdev | but fact is, some professionals would rather use python 2. so i hope there are enough for a fork. |
| 15:48.43 | Leander256 | I've been told that python 2 is nowhere close from being dropped because companies like google have a massive amount of code running in python 2, and they're not about to rewrite it |
| 15:48.45 | *** join/#devuan giricz81 (~giricz81@ip4-83-240-38-90.cust.nbox.cz) |
| 15:48.55 | figlfdev | either way, im reluctantly using python 3 for now-- with my "kludge" to make it handle encoding. |
| 15:49.09 | figlfdev | just as i used debian jessie with sysvinit for a while. |
| 15:49.11 | djph | fsmithred: yeah, no idea what x11-common does either. Reading it now to try making sense of it |
| 15:49.45 | figlfdev | i have no delusions about people getting together just to cater to what i want. i only hope enough python coders care enough to fork when the time comes. |
| 15:49.55 | figlfdev | they might not. we shall find out. |
| 15:50.44 | figlfdev | Leander256: i think the chances of someone other than the python foundation supporting/forking 2 in 2020 (or before) are 50/50, but i have my fingers crossed |
| 15:51.26 | figlfdev | i do not think the python foundation will support it past 2020. they broke it because they thought it was right, they cant do both forever. |
| 15:51.56 | figlfdev | i use python for teaching programming. 2 is far more friendly for beginners than 3, imo. |
| 15:52.07 | figlfdev | but you can make an argument in favor of either |
| 15:52.58 | figlfdev | furrywolf: "started designing my own language once, but never finished it. I took some of the best parts of C, perl, c++, java, lisp, and even php..." id love to see the docs |
| 15:53.19 | figlfdev | mine is basic + logo + python with a little inspiration from bash and/or forth |
| 15:54.20 | furrywolf | this was abotu a decade ago, and no trace of it still exists. plus there weren't many traces outside of my head to begin with. |
| 15:54.24 | *** join/#devuan rleigh (~rleigh@unaffiliated/rleigh) |
| 15:55.10 | figlfdev | too bad. my thesis is that everyone should learn to code (something, anything) and every coder should make a programming language (even if it only has 5 functions and is a complete toy) |
| 15:55.25 | furrywolf | it originally started from a very unusual source... someone asked me about writing a scripting language for a muck. so it started out with being inherently parallel, event-driven, and with a built-in database... |
| 15:55.36 | furrywolf | I wrote a C compiler, does that count? lol |
| 15:55.38 | figlfdev | you can never totally appreciate programming unless you make a little toy language-- some demo that could be done in a week or less. |
| 15:55.46 | figlfdev | sure it counts |
| 15:55.52 | figlfdev | muck? |
| 15:56.01 | figlfdev | like for a lark? or is a muck another thing? |
| 15:56.10 | furrywolf | even output working sparc asm. |
| 15:56.59 | furrywolf | mucks are interactive text-based environments, similar to muds and such. I've never been a mucker, no interest in them at all, but many people I've known have been. |
| 15:57.15 | figlfdev | i wondered if it was related to muds |
| 15:58.00 | furrywolf | flex+bison lets you write a toy language pretty quickly. |
| 15:58.14 | figlfdev | oh one other thing about my language: i started devloping it a couple weeks or so before i switched to devuan |
| 15:58.34 | figlfdev | so i spent about a year developing it, and i did it in using devuan for that time |
| 15:58.50 | furrywolf | many mucks use MUF, a varient of forth... and generally regarded as horrid. |
| 15:58.59 | figlfdev | the thing about flex+bison is that if they do what you need, they will save you trouble |
| 15:59.17 | figlfdev | but if they do more than you need... then they just add another hassle that makes creating a language more complex |
| 15:59.49 | figlfdev | if youre doing a language in c with a recursive descent parser, why not use them? not that its the easiest way to create a language though. |
| 16:00.42 | figlfdev | a lot of people write their own parsers, which i think has equal merit (use a generator if it really makes it easier for you-- but only then) |
| 16:01.41 | furrywolf | I wrote a basic BASIC interpreter a very long time ago, with my own parser. I went overboard and allowed user-defined operators. it had a ?: operator by default. I added it because I like it, and because it is a very good test of a parser. :) |
| 16:02.11 | figlfdev | is genuinely impressed |
| 16:02.43 | furrywolf | I don't do much programming anymore. been about a decade since I wrote anything on-trivial. now I drive a delivery van. |
| 16:02.49 | furrywolf | s/on-/non- |
| 16:03.42 | figlfdev | i cant stand using a machine that doesnt have some kind of code involved. bash is ok, but it sure gets tedious to do some things that are a cinch in python |
| 16:04.25 | figlfdev | i started with dos and basic, and im not anti-gui but i would like gimp a lot more if it had a (simple) command line as an additional feature. |
| 16:04.33 | figlfdev | i know it is scriptable. |
| 16:04.35 | Chanku | programs often :P |
| 16:04.57 | furrywolf | I wrote a bash script once that did convolution on a dataset, among other things. it was one of those scripts that just kept growing from its original 20-line simple task. |
| 16:05.22 | furrywolf | were I to have started writing it knowing what it would have ended up as, it would not have been in bash. |
| 16:05.49 | figlfdev | longest bash script i did made a page with thumbnails from youtube videos |
| 16:06.18 | figlfdev | there were so many calls to things like ffmpeg to make thumbnails and stuff like that, bash was easier for 30% of it so i did the other 70% in bash too |
| 16:06.54 | figlfdev | it turned into the perfect example of why bash isnt ideal for all tasks :) especially larger ones. but some people can use it that way. |
| 16:07.27 | furrywolf | heh, this was a request from someone I knew. some magazine was having an online poll for the hottest actor or some stupid shit like that. it turned out the supporters of one of the people had written a windows macro to vote, clear cookies, vote, etc, 24/7. needless to say their person had a zillion times as many votes. she asked if I could write something to counter it... |
| 16:07.47 | furrywolf | it started as a simple script that wget'ed the vote url repeatedly. |
| 16:07.57 | Chanku | I know someone that wrote an IRC bot in Bash (at least supposedly did so) |
| 16:08.05 | figlfdev | heh |
| 16:08.26 | figlfdev | when i first used irc it was in dos |
| 16:08.29 | *** join/#devuan weems (~hagrid@unaffiliated/weems) |
| 16:08.31 | *** join/#devuan clopez (~tau@neutrino.es) |
| 16:08.48 | figlfdev | arachne browser / dialer + some dos-based irc client |
| 16:09.37 | zdzichu | oh, arachne |
| 16:09.46 | furrywolf | then they added a 1-pixel image that loaded a cookie, so it had to keep track of cookies. then it had to keep track of multiple false user agents with subtle differences, and use a proxy. then it had to perform analysis of the data. then they added a captcha. I wrote the captcha solver in C. but the same bash script had to interact with it. then the bash script started outputting graphs of voting rates vs time to identify every time |
| 16:09.51 | furrywolf | and it ended up devolving into a programming contest. |
| 16:09.52 | furrywolf | lol |
| 16:09.55 | zdzichu | the memories |
| 16:09.59 | figlfdev | nice! |
| 16:10.05 | zdzichu | I've used it some 20 years ago |
| 16:10.31 | furrywolf | and the bash script went from 20 lines to many hundred, plus a C support program. |
| 16:10.32 | furrywolf | I wasn't going to do image processing in bash. that would just be wrong. |
| 16:11.09 | figlfdev | learned today that furrywolf isnt only a troublemaker (knew that already,) but also a better coder than he will ever be |
| 16:11.30 | figlfdev | "that would just be wrong." it would be epic. but youd have to be nuts. |
| 16:11.34 | furrywolf | it was rather humorous. at some point the magazine gave up. I think the two people with software ballot-stuffers had more votes than the earth's population, while every other canidate had a couple dozen. |
| 16:11.50 | figlfdev | ha |
| 16:12.15 | furrywolf | writing the captcha solver was fun. I _like_ problems like that. |
| 16:12.18 | figlfdev | i was involved with a student union that caught the other person running for student council hacking the voting system. |
| 16:12.20 | furrywolf | I got about a 97% success rate. |
| 16:12.31 | figlfdev | i was like "well, this WILL prepare them for 'real' elections." |
| 16:13.03 | furrywolf | (hint: any captcha that will stop a robot will also stop people. because robots are better than many people.) |
| 16:13.31 | figlfdev | yeah, i like the ones where you have to turn your head to guess what the squiggles are saying. |
| 16:13.43 | figlfdev | or rather to help guess. |
| 16:13.47 | furrywolf | I made mine monitor the vote results and always vote ever so slightly faster than they were. it was very amusing. |
| 16:14.17 | figlfdev | "is that a 9 or a G? i dont know! its smeared halfway across the screen" |
| 16:16.03 | furrywolf | mine did some rudimentary learning... first it'd process the image to remove the background (a random grid, easy to remove as it was regular). then it'd separate the letters (they never overlapped much, so was reasonably easy). then it'd rotate them to a consistent angle. at this point they were still distorted and of random sizes and fonts. I used brute-force comparrisons to the database here, to generate a letter probability. |
| 16:16.30 | figlfdev | incredible |
| 16:16.37 | furrywolf | it then made its most likely guess. if the vote was successful, the distorted image was added to the database as an example of what that letter can look like, for comparing to in the future. |
| 16:16.59 | furrywolf | at some point I capped the db size because it was using excessive cpu to compare. |
| 16:17.12 | figlfdev | so you basically made an ai, just to make a friends favorite celebrity slightly more popular. |
| 16:17.34 | furrywolf | I've been very tempted lately to write one to solve google's click-the-images captcha. that shit fucking pisses me off. |
| 16:17.36 | figlfdev | id rather drive a delivery van :) |
| 16:17.41 | furrywolf | it was fun. :P |
| 16:17.58 | figlfdev | i bet |
| 16:18.21 | furrywolf | I'm willing to bet I can solve google's captcha with a rather high success rate. |
| 16:18.29 | figlfdev | i wouldnt bet against you |
| 16:19.29 | furrywolf | and I'm not that good of a coder. which probably explains why spam is still rampant on forums no matter how obnoxious the captchas are, because spammers hire better coders. |
| 16:19.31 | figlfdev | but could you put a neural net tutorial on the talk.devuan site? (or any other devuan-related site/forum/wiki) |
| 16:19.42 | furrywolf | I know very very little about neural nets. |
| 16:20.15 | figlfdev | "probably explains why spam is still rampant" spam is not going away because the desire to spam will always outweigh the effort to eliminate spam |
| 16:21.02 | furrywolf | I think that captchas are futile, as any captcha that a person can reliably solve, a robot can too. |
| 16:21.04 | furrywolf | and all they do is piss people off. |
| 16:22.20 | hellekin | furrywolf: even if some CAPTCHAs are robot-proof, they're not slave-proof, and spammers know that very well. |
| 16:22.43 | figlfdev | also most efforts to remove spam *have to be reasonable*, and spammers never need be |
| 16:22.50 | furrywolf | in particular, they piss ME off. I'm not going to spent three fucking minutes helping google identify which images have street numbers for free just because I want to respond to someone's post. |
| 16:23.00 | figlfdev | "they're not slave-proof" sad but true. |
| 16:23.00 | furrywolf | s/spent/spend |
| 16:23.52 | figlfdev | "three fucking minutes helping google identify which images have street numbers" exactly. bad enough we are training their ai to distinguish every kind of image so they can build skynet |
| 16:24.08 | figlfdev | but i draw the line (or intend to) at helping stalkview |
| 16:24.21 | figlfdev | skip that one and go to the next... |
| 16:24.46 | furrywolf | my general policy is not to use any site that uses google's captchas. |
| 16:25.59 | furrywolf | I liked their old one where they were trying to get you to OCR words for them. you could figure out which word they already knew and which word they were trying to get you to transcribe for them, and always answer it wrong. they'd take it, because they only actually knew one of the two words. |
| 16:26.55 | Chanku | How could you tell? |
| 16:26.57 | furrywolf | I do hope somewhere in google books there's a bunch of words mistranscribed as fuckgoogle, because I put that in every captcha. :) |
| 16:27.17 | furrywolf | Chanku: the distortions were actual scan errors, not software distortion of an easily recognizable word. |
| 16:27.37 | Chanku | Alright |
| 16:28.14 | furrywolf | they'd look like misprints, dust, page curl, etc. |
| 16:29.14 | Chanku | okay |
| 16:38.05 | fsmithred | AHA! |
| 16:39.11 | flrn | :) |
| 16:39.19 | fsmithred | Centurion_Dan, golinux, KatolaZ: I solved the mystery of why I get a grub splash when I make a uefi iso on one system but not on the other system. Had to remove grub-pc (or -bin?) and now it works. |
| 16:39.52 | furrywolf | I never solved the mystery of why grub always failed to render properly on one of my laptops. |
| 16:40.26 | flrn | furrywolf: the downside is, that you tell it about you self... |
| 16:40.45 | furrywolf | http://fw.bushytails.net/grubvideowtf2.jpg |
| 16:41.18 | furrywolf | I feel like I need to do a bad dracula impression when reading it. |
| 16:41.43 | furrywolf | eeeedit the commmmands before booooting! |
| 16:43.42 | MinceR | :) |
| 16:49.16 | golinux | fsmithred: Congrats! |
| 16:54.42 | fsmithred | I'm re-testing now, using the script that wasn't working. |
| 16:54.46 | *** join/#devuan wildlander (~wild@unaffiliated/wildlander) |
| 16:55.11 | fsmithred | gotta go out now. It'll be ready when I get back. |
| 16:56.53 | nat0 | falls asleep |
| 17:05.57 | *** join/#devuan ItSANgo__ (~itsango@zaq7ac458cb.zaq.ne.jp) |
| 17:07.48 | *** join/#devuan chr[] (~chr@chris.uscreen.net) |
| 17:13.03 | *** join/#devuan Markus_KMi (uid95611@gateway/web/irccloud.com/x-subztvrtpaxbxupv) |
| 17:19.39 | *** join/#devuan oldlaptop (~quassel@50.36.249.13) |
| 17:21.13 | *** join/#devuan egon0 (~quassel@2a01:1e8:e100:8035:90be:4b72:e309:649) |
| 17:33.36 | *** join/#devuan jsfarinet (~chatzilla@p2003008C4F26D3A952B7C3FFFE13D4E6.dip0.t-ipconnect.de) |
| 17:34.05 | *** join/#devuan cdanderson42 (~nuhrin@67.170.89.70) |
| 17:37.33 | *** join/#devuan cocoadaemon (~foo@2a01:e35:8a99:e90:f83d:d5d9:d496:9ddf) |
| 17:38.37 | *** part/#devuan jsfarinet (~chatzilla@p2003008C4F26D3A952B7C3FFFE13D4E6.dip0.t-ipconnect.de) |
| 17:39.17 | *** join/#devuan jsfarinet (~chatzilla@p2003008C4F26D3A952B7C3FFFE13D4E6.dip0.t-ipconnect.de) |
| 17:43.22 | *** join/#devuan debdog (~debdog@HSI-KBW-091-089-090-004.hsi2.kabelbw.de) |
| 17:51.11 | gci_admin | furrywolf, on a forum I manage we use a two page auth. There is a long string that has to be copied from one page and pasted as an answer to a query in another page. So far, no bot has gotten past that for over a year now. |
| 17:53.38 | gci_admin | CAPTCHA never worked that well. |
| 17:55.05 | *** join/#devuan luke-jr (~luke-jr@unaffiliated/luke-jr) |
| 17:57.20 | *** join/#devuan luke-jr (~luke-jr@unaffiliated/luke-jr) |
| 18:12.29 | *** join/#devuan oldlaptop (~quassel@50.36.249.13) |
| 18:20.23 | *** join/#devuan reinob (~reinob@Maemo/community/council/reinob) |
| 18:20.49 | DocScrutinizer05 | well, a decent Turing test is never simple to design |
| 18:21.21 | DocScrutinizer05 | and in the end that's exactly what all those Captchas etc are all about |
| 18:22.25 | DocScrutinizer05 | a *precise* and quick&convenient Turing test... is a real challenge |
| 18:38.51 | *** join/#devuan cdanderson42 (~nuhrin@67.170.89.70) |
| 18:40.19 | *** join/#devuan cdanderson42 (~nuhrin@67.170.89.70) |
| 18:53.28 | *** join/#devuan TwistedFate (~TwistedFa@unaffiliated/twistedfate) |
| 19:13.25 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 19:30.40 | *** join/#devuan jsfarinet (~chatzilla@p2003008C4F26D3A952B7C3FFFE13D4E6.dip0.t-ipconnect.de) |
| 19:33.44 | fsmithred | golinux, I was wrong. I didn't solve it. |
| 19:35.13 | *** join/#devuan cdanderson42 (~nuhrin@67.170.89.70) |
| 19:35.53 | *** join/#devuan Inocuous (~Inocuous@44-117-181-66.dsl.sacoriver.net) |
| 19:38.20 | golinux | fsmithred: :( Sorry to hear that. |
| 19:39.01 | *** join/#devuan minnesotags (~herbgarci@c-50-137-242-97.hsd1.mn.comcast.net) |
| 19:39.11 | DocScrutinizer51 | UEFI meh! |
| 19:39.36 | *** join/#devuan minnesotags (~herbgarci@c-50-137-242-97.hsd1.mn.comcast.net) |
| 19:39.54 | DocScrutinizer51 | define it as a known accepted flaw |
| 19:40.16 | DocScrutinizer51 | ley others solve it |
| 19:40.26 | DocScrutinizer51 | let* |
| 19:40.48 | fsmithred | lol |
| 19:41.06 | fsmithred | it does work correct with the cli script and manually copying the efi files |
| 19:41.32 | DocScrutinizer51 | aiui it's just the bootscreen wallpaper? |
| 19:41.47 | fsmithred | so it's not outside the realm of possibilities that only those who have the desire and skills needed can have that |
| 19:41.55 | fsmithred | yes, bootscreen picture |
| 19:42.02 | fsmithred | pretty graphics |
| 19:42.13 | DocScrutinizer51 | nevermind then |
| 19:42.17 | *** join/#devuan minnesotags (~herbgarci@c-50-137-242-97.hsd1.mn.comcast.net) |
| 19:42.39 | fsmithred | bugs the shit out of me that I can't see the problem, and that's the main motivation for contuing |
| 19:42.46 | DocScrutinizer51 | those using uefi don't deserve nice gfx ;P |
| 19:43.08 | fsmithred | well, that and the fact that Refracta has a really cool wallpaper |
| 19:43.37 | DocScrutinizer51 | make 'cp' a script that prints UID, EUID, ossix permissions etc pp |
| 19:44.12 | fsmithred | why? |
| 19:44.24 | DocScrutinizer51 | tofind the reasn |
| 19:44.40 | DocScrutinizer51 | the diff between script and manual |
| 19:44.42 | fsmithred | I don't follow |
| 19:45.23 | fsmithred | I am trying to break the problem down |
| 19:45.32 | DocScrutinizer51 | when a manual call to cp does what a call to cp inside script fails... run cp under strace |
| 19:46.14 | fsmithred | it works with a separate script script for making the efi files - I'm calling that the manual method - have to copy the efi files over to the iso manually |
| 19:46.31 | fsmithred | it doesn't work with the big script that has the separate script added to it as a function |
| 19:46.57 | fsmithred | so I'm running it now without that function, and I'll use the manual method, then let the rest of the big script run |
| 19:48.23 | fsmithred | if it still doesn't work, I know it's something outside of that function |
| 19:49.38 | fsmithred | I've never used strace - can I run the whole script with it and get more inforamtion that set -x gives me? |
| 19:49.51 | fsmithred | s/that/than/ |
| 19:52.36 | jsfarinet | Hi all. I've a question re devuan talk ... |
| 19:53.33 | golinux | jsfarinet: Yes? |
| 19:54.12 | jsfarinet | i created an account (to put somewhere there,jaromil and hellekin suggested study) the rudimentary outlines and examples |
| 19:54.59 | jsfarinet | we put together for a jwm desktop on devuan. But unfortunately, apparentely i'm unable to create a new topic (?) |
| 19:56.19 | jsfarinet | Do i need an authorization or something like that? |
| 19:56.25 | golinux | That's the way that software works. You can only post in certain areas until you are 'trusted' or some BS close to that. |
| 19:57.05 | jsfarinet | Ok. So, how could i "gain" trust? |
| 19:57.16 | jsfarinet | If at all ... of course ;) |
| 19:57.51 | golinux | Maybe put it on the http://wiki.friendsofdevuan.org/doku.php/community_artwork |
| 19:57.52 | *** join/#devuan curvian__ (~curvv@unaffiliated/curvv) |
| 19:58.01 | golinux | It's a bean counter |
| 19:58.54 | golinux | Ka-chink ka-chink |
| 19:59.57 | golinux | Maybe hellekin can fix it for you. |
| 20:01.03 | jsfarinet | @golinux That does not match so much - when i'm finished with doing an lxdm greeter for devuan may be it would be the right place :) |
| 20:03.24 | golinux | I just upped your trust level. See if you can now. |
| 20:03.58 | golinux | is reluctant to do admin things but . . . |
| 20:07.12 | golinux | I just rtfm a bit and I think it needs to be higher to create a new topic. (where is hellekin?) |
| 20:09.43 | jsfarinet | Thanks golinux, i'll try ... |
| 20:09.47 | Chanku | I wonder if there are any decent lightweight spreadsheet programs... |
| 20:09.57 | jsfarinet | But no, i still cannot :-( |
| 20:10.43 | jsfarinet | Anyway, thanks for your patience, hey :) |
| 20:10.48 | golinux | This is how it works. Pretty convoluted . . . https://meta.discourse.org/t/what-do-user-trust-levels-do/4924/2 |
| 20:11.21 | jsfarinet | sys admin os ;) ;) |
| 20:13.58 | jsfarinet | Don't worry ... Is hellekin orsomeone else highleveled able to read *ALL* wikis? If so he could have a look at the wiki we created |
| 20:14.17 | jsfarinet | at git.devuan and decide if it's worth to be comunicated ... |
| 20:15.16 | DocScrutinizer05 | fsmithred: prolly strace is overkill, but yes you could run the while script under strace and it gives you a 5 magnituides more output than set -x |
| 20:15.59 | DocScrutinizer05 | I'd try running the copy script in a separate subshell called from the main script, to start with |
| 20:16.29 | DocScrutinizer05 | dunno how that changes things, but... |
| 20:18.28 | DocScrutinizer05 | also adding a sleep 10 before the cp might help for other previous commands to finish async - dunno if there's a mount or sync or whatever that is not finished so the cp fails |
| 20:18.50 | DocScrutinizer05 | just for test of course |
| 20:19.40 | fsmithred | I just installed strace, and I'm going to try it |
| 20:19.53 | fsmithred | I already run the script from a terminal, so I see lots of output |
| 20:20.00 | fsmithred | and there's an error log |
| 20:20.39 | fsmithred | I was wondering about adding a sleep somewhere, but I figured that was on my mind 'cause I've done some stuff for audio recently. |
| 20:21.28 | fsmithred | everything had to start in the right order |
| 20:31.34 | DocScrutinizer51 | fsmithred: for strace you'll need that 'follow forks' flag |
| 20:36.01 | fsmithred | ok, I just ran it without any options |
| 20:36.07 | fsmithred | and yeah, it might be overkill |
| 20:37.16 | DocScrutinizer51 | use the fork option and filter to maybe only see file functions |
| 20:38.26 | *** join/#devuan radsy (~irc@unaffiliated/radsy) |
| 20:38.26 | fsmithred | filter, like with grep? |
| 20:38.32 | DocScrutinizer51 | no strace crack here either |
| 20:38.51 | DocScrutinizer51 | no there are filter options in strace |
| 20:39.17 | DocScrutinizer51 | debug:file or whatever |
| 20:40.42 | DocScrutinizer51 | I only use strace in a logfile which I view in less |
| 20:41.37 | DocScrutinizer51 | since it's still a lot of grep_like searching |
| 20:41.56 | fsmithred | reading all the options now |
| 20:42.50 | *** join/#devuan JackWinter (~jack@vodsl-8402.vo.lu) |
| 20:43.42 | DocScrutinizer51 | read() on filehandle3 returns error -> search backward for the open() that tells which file is on FH3 |
| 20:44.24 | DocScrutinizer51 | I want less for that |
| 20:47.01 | fsmithred | ok, we have a winner here |
| 20:47.19 | fsmithred | it's in the function I added to make the efi files |
| 20:47.36 | fsmithred | I'm gonna start with adding some sleeps in the places where I do stuff manually |
| 20:48.04 | fsmithred | but not right now... I need a break. DocScrutinizer51 thanks for your help |
| 20:49.05 | DocScrutinizer51 | which help? I'm just a noob with many ideas |
| 20:49.35 | fsmithred | you gave me some ideas |
| 20:50.06 | DocScrutinizer51 | yw for that |
| 21:02.01 | KatolaZ | fsmithred: congrats! :) |
| 21:03.35 | DocScrutinizer51 | ok, the URL to the script source is still in my log. I hope it's not an expired pastebin |
| 21:04.46 | DocScrutinizer51 | will have one more look when I'm back home |
| 21:06.34 | *** join/#devuan n4dir (~user@pD9EF16B9.dip0.t-ipconnect.de) |
| 21:24.16 | golinux | KatolaZ: Read on |
| 21:24.54 | *** join/#devuan minnesotags (~herbgarci@c-50-137-242-97.hsd1.mn.comcast.net) |
| 21:25.10 | golinux | Back to the drawing board. Hopefully it will get sorted. |
| 21:25.35 | minnesotags | Howdy |
| 21:26.19 | KatolaZ | golinux: ? |
| 21:26.21 | *** join/#devuan zxz (~zz@c-76-23-25-230.hsd1.ut.comcast.net) |
| 21:26.33 | golinux | the efi stuff |
| 21:26.59 | golinux | and your shoutout to fsmithred |
| 21:28.17 | KatolaZ | oh I see |
| 21:28.25 | KatolaZ | it seems he didn't solve it |
| 21:28.26 | KatolaZ | or yes |
| 21:28.32 | KatolaZ | I am unsure |
| 21:28.33 | KatolaZ | :D |
| 21:29.07 | *** join/#devuan Markus_KMi (uid95611@gateway/web/irccloud.com/x-dtklleprbburpmas) |
| 21:29.34 | KatolaZ | ok |
| 21:29.42 | KatolaZ | it seems that is has not been solved, so far |
| 21:30.08 | KatolaZ | I don't withdraw the congrats, anyway |
| 21:30.18 | KatolaZ | I know it will be solved, sooner or later :) |
| 21:30.41 | KatolaZ | golinux: a few nights ago I was working on an ascii-art graphics for the Devuan Logo |
| 21:30.54 | KatolaZ | I don't know if anybody has done anything in that direction already |
| 21:31.10 | golinux | Have a screenie? |
| 21:31.33 | KatolaZ | nothing complete |
| 21:31.42 | KatolaZ | I worked on the "D" so far |
| 21:31.53 | KatolaZ | which required a bit of an effort |
| 21:32.02 | KatolaZ | (it's too "curvy"...) |
| 21:32.12 | golinux | That would be so cool for ascii!! |
| 21:32.21 | KatolaZ | <PROTECTED> |
| 21:32.22 | KatolaZ | <PROTECTED> |
| 21:32.22 | KatolaZ | <PROTECTED> |
| 21:32.22 | KatolaZ | <PROTECTED> |
| 21:32.22 | KatolaZ | <PROTECTED> |
| 21:32.24 | KatolaZ | <PROTECTED> |
| 21:32.26 | KatolaZ | <PROTECTED> |
| 21:32.29 | KatolaZ | <PROTECTED> |
| 21:32.31 | KatolaZ | <PROTECTED> |
| 21:32.34 | KatolaZ | <PROTECTED> |
| 21:32.36 | KatolaZ | <PROTECTED> |
| 21:32.39 | KatolaZ | <PROTECTED> |
| 21:32.41 | KatolaZ | <PROTECTED> |
| 21:32.44 | KatolaZ | you see |
| 21:32.46 | KatolaZ | nothing much |
| 21:32.49 | KatolaZ | but I tried to remain as close as possible to the svg logo |
| 21:33.04 | KatolaZ | I can easily add the rest |
| 21:33.13 | KatolaZ | which is much easier, indeed |
| 21:33.21 | KatolaZ | just want to be sure it's not repeated work :) |
| 21:33.22 | golinux | The tail is too long and thick |
| 21:33.32 | golinux | The proportions are off. |
| 21:33.42 | KatolaZ | well, it's ascii art :D |
| 21:33.49 | KatolaZ | and not a good one at that |
| 21:33.51 | DocScrutinizer05 | please! |
| 21:34.15 | KatolaZ | and I didn't want to go beyond 12 lines |
| 21:34.29 | DocScrutinizer05 | better stay <=3 |
| 21:34.30 | KatolaZ | so that we could include it in a issue file :) |
| 21:34.47 | DocScrutinizer05 | for chan posts |
| 21:34.59 | DocScrutinizer05 | üastebin for the larger |
| 21:35.14 | golinux | I would do it in an image editor on top of the logo |
| 21:35.46 | KatolaZ | ok |
| 21:35.49 | KatolaZ | I pass then |
| 21:35.57 | KatolaZ | it was just a nighty experiment :) |
| 21:36.25 | golinux | hellekin and I tweaked the logo proportions for many days to get it just right. That should be preserved |
| 21:37.04 | KatolaZ | then we might need an ascii-art wizard... :) |
| 21:37.13 | golinux | Oh don't pass. It's a cool idea. |
| 21:37.37 | KatolaZ | I know it's cool |
| 21:37.57 | golinux | I've never done ascii art. :( |
| 21:38.00 | KatolaZ | only I don't know whether I could get any closer to the "official" proportions |
| 21:38.16 | KatolaZ | the problem is only the D, obviously |
| 21:38.22 | KatolaZ | which is pretty "curvy" |
| 21:38.43 | KatolaZ | the upper-left size will be almost impossible to reproduce exactly in ascii-art... |
| 21:38.49 | DocScrutinizer05 | ((CoC <fsmithred> benevolent dictatorship is very useful and appropriate in some situations)) indeed, the only working approach |
| 21:39.00 | golinux | I might play with it a bit |
| 21:39.06 | KatolaZ | I will do the same |
| 21:39.12 | KatolaZ | and let you know |
| 21:39.19 | golinux | OK |
| 21:39.29 | *** join/#devuan Artemis3 (~artemis3@190.37.30.10) |
| 21:40.42 | golinux | Time may be an issue before long though. Have received notice of a proposed 2000 acre sludge site in the county. Application is ~400 pages. |
| 21:41.13 | golinux | I'm gonna be up to my eyeballs in shit. |
| 21:41.44 | golinux | They have the nerve to call this toxic stuff 'beneficial land use application'. |
| 21:43.03 | golinux | DocScrutinizer05: fsmithred possesses great wisdom. |
| 21:44.29 | DocScrutinizer05 | yes |
| 21:46.14 | fsmithred | KatolaZ, I like the ascii swoosh |
| 21:46.16 | DocScrutinizer05 | btw you got that on every soccer match |
| 21:47.35 | DocScrutinizer05 | referee might be wrong but his decision counts. When he's performing poor then next time he won't get hired for the job |
| 21:48.14 | DocScrutinizer05 | so he's a benevolent dictator for a limited time |
| 21:49.01 | DocScrutinizer05 | there are rulesets that he's supposed to obey but you can't really call him out on that |
| 21:50.37 | DocScrutinizer05 | this works since dunno 100 years now? |
| 21:51.34 | DocScrutinizer05 | figures two football teams trying to agree on a CoC before the match |
| 21:58.40 | fsmithred | two rugby teams would be easier |
| 21:59.05 | fsmithred | "Don't eat your own team members." |
| 22:00.57 | fsmithred | I'll be back in about an hour to test the latest snapshot |
| 22:03.40 | DocScrutinizer05 | do you have the pastebin URL for me? |
| 22:04.01 | DocScrutinizer05 | it's a tad hard to find in the backscroll |
| 22:05.07 | DocScrutinizer05 | fsmithred: ^^^ |
| 22:05.24 | DocScrutinizer05 | I'm curious |
| 22:05.55 | DocScrutinizer05 | waaaait, I recall a searchstring ;-) |
| 22:07.37 | DocScrutinizer05 | https://gist.github.com/fsmithred/149637f013d7da3fc9c94de7b333d7c4 right? |
| 22:08.18 | DocScrutinizer05 | and the infringing code was between :<IGNORe and IGNORE? |
| 22:10.30 | DocScrutinizer05 | https://gist.github.com/fsmithred/149637f013d7da3fc9c94de7b333d7c4#file-snapshot_uefi_bad-L782 ? |
| 22:10.54 | DocScrutinizer05 | even 787? |
| 22:13.51 | DocScrutinizer05 | fsmithred: tbh I never seen pushd, popd in any script. Dunno if it's smart, non-standard (Bourne) shell, or dangerous |
| 22:20.05 | KatolaZ | DocScrutinizer05: popd/pushd is just a way to deal with the directory stack |
| 22:20.17 | DocScrutinizer05 | I know pushd popd |
| 22:20.34 | DocScrutinizer05 | I just wonder if it's supposed to get used in scripts |
| 22:20.46 | DocScrutinizer05 | is it POSIX? ;-) |
| 22:21.40 | DocScrutinizer05 | pushd is a BASH-builtin? |
| 22:21.45 | DocScrutinizer05 | so... |
| 22:22.29 | DocScrutinizer05 | shebang #!/usr/bin/env bash so prolly safe |
| 22:23.39 | DocScrutinizer05 | anyway 787 is outside the pushd popd sequence |
| 22:28.56 | *** join/#devuan SuicideJunkie (~nick@CPE000f6638c2b8-CM0c473de90d90.cpe.net.cable.rogers.com) |
| 22:30.00 | DocScrutinizer05 | glares at dd if=/dev/zero of=boot/grub/efiboot.img bs=1K count=1440 |
| 22:30.42 | DocScrutinizer05 | pulls out a 50⬠bill to place on the table |
| 22:31.25 | DocScrutinizer05 | fsmithred: you ever watched the drive LED of the of= drive of dd, after dd has long finished? |
| 22:31.32 | DocScrutinizer05 | try adding a sync |
| 22:32.51 | *** join/#devuan weems (~hagrid@unaffiliated/weems) |
| 22:33.04 | DocScrutinizer05 | it's still a tad strange when all former commands between dd and the cp would succeed while cp fails, but... |
| 22:33.37 | DocScrutinizer05 | **fiiiishy* |
| 22:34.55 | DocScrutinizer05 | either add sync(8) or use the sync option to dd |
| 22:36.33 | DocScrutinizer05 | <PROTECTED> |
| 22:36.34 | DocScrutinizer05 | <PROTECTED> |
| 22:41.33 | DocScrutinizer05 | ooh you're not cp'ing to the freshly created volume... it gets umounted in 775 |
| 22:41.55 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 22:42.43 | KatolaZ | DocScrutinizer05: I guess the problem (whatever it is ) is not in the use of pushd/popd |
| 22:42.54 | DocScrutinizer05 | no, prolly not |
| 22:43.08 | DocScrutinizer05 | since shebang #!/usr/bin/env bash so prolly safe |
| 22:43.11 | KatolaZ | but I have lost you guys |
| 22:43.49 | DocScrutinizer05 | I still frown at dd without sync |
| 22:44.12 | DocScrutinizer05 | flushing cache of a massive dd can take minutes |
| 22:44.24 | DocScrutinizer05 | after dd already terminated |
| 22:45.14 | KatolaZ | DocScrutinizer05: |
| 22:45.20 | KatolaZ | it is not a "masive" dd... |
| 22:45.25 | KatolaZ | it's just 1.44 MB |
| 22:45.29 | DocScrutinizer05 | I dunno, I don't see how it's related to a cp that happens even after umount of the volume created by dd&&mkfs&&mount -loop, but... |
| 22:45.44 | KatolaZ | I don't know what's the problem though |
| 22:45.51 | KatolaZ | and what are the symptoms |
| 22:46.01 | DocScrutinizer05 | I think it's the cp in line 787 failing |
| 22:46.10 | DocScrutinizer05 | or the rsyncs in the lines before |
| 22:46.19 | KatolaZ | I don't even have the current version of the script at hand.... |
| 22:46.42 | DocScrutinizer05 | fsmithred said the splashscreen (wallpaper) works when he cp manually, but not when done in script |
| 22:46.50 | DocScrutinizer05 | https://gist.github.com/fsmithred/149637f013d7da3fc9c94de7b333d7c4#file-snapshot_uefi_bad-L782 |
| 22:47.10 | DocScrutinizer05 | I dunno if that's the _curent_ version |
| 22:47.41 | DocScrutinizer05 | the code is commented out there, by :<<IGNORE |
| 22:48.04 | DocScrutinizer05 | and the script fails as soon as fsmithred undoes the commenting |
| 22:48.24 | DocScrutinizer05 | aah wait that was a syntax error |
| 22:48.43 | DocScrutinizer05 | nevertheless I think that's the location where the failing cp happens |
| 22:50.57 | *** join/#devuan kelsoo (~kelsoo@86.53.50.178) |
| 22:52.18 | DocScrutinizer05 | ((<KatolaZ> it is not a "masive" dd...)) may be. But then it's also not minutes but milliseconds until the script executes commands that are meant to work on the dd'ed file |
| 22:52.33 | KatolaZ | DocScrutinizer05: there is no "&" |
| 22:52.46 | KatolaZ | so dd exit only when it's done |
| 22:52.53 | DocScrutinizer05 | yes, I know, but flushing cache happens async |
| 22:52.57 | DocScrutinizer05 | no |
| 22:53.13 | DocScrutinizer05 | dd exits when it's done writing to RAM cache |
| 22:53.20 | KatolaZ | well, if you read the file |
| 22:53.25 | KatolaZ | flushing happens anyway |
| 22:53.29 | DocScrutinizer05 | yes |
| 22:53.33 | DocScrutinizer05 | prolly |
| 22:53.33 | KatolaZ | so... |
| 22:53.54 | DocScrutinizer05 | I'd use a sync anyway, if just for best practice |
| 22:54.16 | DocScrutinizer05 | even in dd as option |
| 22:55.19 | DocScrutinizer05 | I wouldn't dare to run a mkfs on a loopmounted file that just been created by dd without sync |
| 22:56.18 | DocScrutinizer05 | that's down into the gzts of IO up to the ellbows |
| 22:56.23 | DocScrutinizer05 | guts |
| 22:57.22 | DocScrutinizer05 | *that's _hands_ down into... |
| 22:59.12 | DocScrutinizer05 | prolly the filesystem/IO handlers have separate cache for direct block access to the loopmounted "volume" (which is actually a file), and other ones for the fs where the file is just a file and no volume |
| 23:00.40 | DocScrutinizer05 | so the fs cache could write cached blocks _after_ the loopmounted volume handler already accessed them via direct IO |
| 23:01.11 | DocScrutinizer05 | dunno, i'm talkimng outa my rear here |
| 23:01.31 | DocScrutinizer05 | at least *I* wouldn't dare |
| 23:03.19 | KatolaZ | DocScrutinizer05: I guess the dd is not the problem, IMHO |
| 23:03.33 | KatolaZ | too much complicated as an explanation |
| 23:03.57 | DocScrutinizer05 | still, the dd is before mount and the cp is after umount and consequently to another destination most likely |
| 23:04.15 | DocScrutinizer05 | so yes, I don't see how it's related |
| 23:04.20 | DocScrutinizer05 | but it stinks |
| 23:04.50 | DocScrutinizer05 | well, smells funny |
| 23:04.52 | KatolaZ | DocScrutinizer05: I still have not got what is the symptom |
| 23:05.00 | KatolaZ | i.e., what goes wrong |
| 23:05.34 | DocScrutinizer05 | the sympton according to fsmithred is that the copy of the splashscreen fails when done in script, while it works when he does it manually |
| 23:06.09 | DocScrutinizer05 | he obviously needs to stop the script to do that manually |
| 23:06.26 | KatolaZ | does he does the same sequence of commands, pasted from the script? |
| 23:06.31 | *** join/#devuan freemangordon (~ivo@46.249.74.23) |
| 23:06.38 | DocScrutinizer05 | aiui yes |
| 23:07.27 | DocScrutinizer05 | timing problem aka race condition would be a good explanation |
| 23:07.37 | DocScrutinizer05 | thus sync(8) |
| 23:07.54 | DocScrutinizer05 | or just add a sleep 30 to test, right before the cp |
| 23:08.30 | KatolaZ | 30 seconds to write 1.44 MB?!? |
| 23:08.50 | KatolaZ | that does not make any sense to me, to be honest |
| 23:08.55 | DocScrutinizer05 | meh, make it 180s so you can get a coffee, and then see if the test worked |
| 23:09.10 | DocScrutinizer05 | it's just to "bisect" |
| 23:09.36 | DocScrutinizer05 | no need to optimize |
| 23:10.14 | DocScrutinizer05 | or use a "read foo" and make the walk to the kitchen your timer ;-) |
| 23:10.42 | KatolaZ | still don't get |
| 23:10.52 | KatolaZ | the splash is not copied in the img |
| 23:10.56 | KatolaZ | at all |
| 23:11.09 | KatolaZ | the cp is at line 787 |
| 23:11.12 | DocScrutinizer05 | when it's *really* a race, a long sleep (or break) will change the test result |
| 23:11.16 | DocScrutinizer05 | yep |
| 23:11.21 | KatolaZ | after the loopback has been removed |
| 23:11.30 | DocScrutinizer05 | yes, I noticed that |
| 23:11.33 | KatolaZ | unomuntd and removed |
| 23:11.36 | DocScrutinizer05 | yes |
| 23:11.49 | DocScrutinizer05 | thus [2016-06-02 Thu 01:04:15] <DocScrutinizer05> so yes, I don't see how it's related |
| 23:11.53 | KatolaZ | so how the splash coud appear in the loop-mounted image??? |
| 23:12.05 | DocScrutinizer05 | it maybe doesn't |
| 23:12.36 | DocScrutinizer05 | maybe the image appears in the same location like the splash file, I dunno and I don't want to analyze the whole script |
| 23:13.04 | KatolaZ | DocScrutinizer05: maybe it's just pointless to look for the problem without fsmithred |
| 23:13.05 | DocScrutinizer05 | I just detect something fishy about dd and missing sync |
| 23:13.29 | KatolaZ | and I don't think the problem is there anyway |
| 23:13.44 | KatolaZ | if the splash has to be copied in the loop-mounted image |
| 23:13.54 | DocScrutinizer05 | fsmithred is supposed to be back shortly |
| 23:13.58 | KatolaZ | it should be copied before the loop-mount is unmounted |
| 23:14.03 | DocScrutinizer05 | yep |
| 23:14.05 | KatolaZ | not after |
| 23:14.11 | KatolaZ | as it is copied atm |
| 23:14.16 | DocScrutinizer05 | I don't really understand what's being done there, at all |
| 23:14.51 | DocScrutinizer05 | I would have *thought* you want the splash file *inside* the .iso |
| 23:15.07 | DocScrutinizer05 | so maybe that is the bug |
| 23:15.19 | KatolaZ | dunno |
| 23:15.20 | DocScrutinizer05 | premature umount |
| 23:15.26 | DocScrutinizer05 | me neither |
| 23:15.31 | KatolaZ | without fsmithred those are just speculations |
| 23:15.32 | KatolaZ | :) |
| 23:15.38 | DocScrutinizer05 | yep :-) |
| 23:15.38 | KatolaZ | at least mine |
| 23:15.45 | DocScrutinizer05 | mine too |
| 23:16.20 | DocScrutinizer05 | now for something different: it's late here |
| 23:16.57 | DocScrutinizer05 | and I want a coffee |
| 23:19.08 | fsmithred | back. |
| 23:19.15 | DocScrutinizer05 | \o/ |
| 23:19.29 | DocScrutinizer05 | fsmithred: where to do you copy that splashfile? |
| 23:19.54 | fsmithred | to iso/boot/grub/ |
| 23:20.08 | DocScrutinizer05 | fsmithred: where to do you copy that splashfile?https://gist.github.com/fsmithred/149637f013d7da3fc9c94de7b333d7c4#file-snapshot_uefi_bad-L787 that's the problem? |
| 23:20.11 | fsmithred | and then everything inside iso/ goes to the root of the iso file |
| 23:20.47 | fsmithred | oh god, that one was several days ago. Problem is 'if if' instead of 'if fi' |
| 23:21.08 | DocScrutinizer05 | fsmithred: you're aware you umount the loopmounted thing before you do the cp in line 787? |
| 23:21.33 | DocScrutinizer05 | yes, I know it's not of today |
| 23:22.19 | DocScrutinizer05 | do you have a link to a recent one? |
| 23:22.23 | fsmithred | yeah, that umount should be ok there |
| 23:22.51 | KatolaZ | fsmithred: does the splash need to go in the loop-mounted img file, by any chance? |
| 23:23.09 | KatolaZ | (my comment may well be completely bogus though) |
| 23:23.10 | fsmithred | gimme a minute to look at it and sort it in my mind |
| 23:23.27 | DocScrutinizer05 | link please :-) |
| 23:24.02 | GeneralStupid | is it possible to use armbian sources with devuan / debian? |
| 23:24.47 | fsmithred | ok, last build works! |
| 23:24.55 | fsmithred | I figured it out (partly) |
| 23:25.04 | fsmithred | I was not using the same image file in both tests |
| 23:25.09 | KatolaZ | :) |
| 23:25.12 | fsmithred | or... |
| 23:25.14 | KatolaZ | happens |
| 23:25.24 | fsmithred | I was not naming the destination file in one |
| 23:25.42 | fsmithred | cp "$iso_dir"/isolinux/splash.png boot/grub/ |
| 23:25.59 | fsmithred | cp /home/snapshot/milky_640x480.png boot/grub/splash.png |
| 23:26.40 | fsmithred | cp /home/snapshot/milky_640x480.png boot/grub/splash.png |
| 23:26.45 | fsmithred | oops |
| 23:26.50 | DocScrutinizer05 | yeah, I see |
| 23:26.52 | fsmithred | milky_640x480.png: PNG image data, 640 x 480, 8-bit/color RGB, non-interlaced |
| 23:27.12 | fsmithred | splash.png: PNG image data, 640 x 480, 8-bit/color RGB, non-interlaced |
| 23:27.16 | DocScrutinizer05 | fsmithred: anyway pretty please add a sync after dd, or use the sync option of dd |
| 23:27.25 | fsmithred | those look the same to me |
| 23:27.37 | fsmithred | ok, I can add sync |
| 23:27.39 | KatolaZ | DocScrutinizer05: the sync is useless, my friend |
| 23:27.44 | fsmithred | oh? |
| 23:27.51 | DocScrutinizer05 | so it doesn't hurt either |
| 23:27.56 | fsmithred | oh right - dd has to finish before next thing will happen |
| 23:28.06 | KatolaZ | then why don't you use a sync after each cp? |
| 23:28.13 | KatolaZ | :) |
| 23:28.16 | DocScrutinizer05 | yes, and per default dd is 'async' since it uses cache |
| 23:28.46 | fsmithred | so, is it the fact that I didn't name the file in the one command? |
| 23:28.50 | KatolaZ | again, when you *read* the cache is synced |
| 23:28.55 | KatolaZ | and you read in the next line |
| 23:29.01 | KatolaZ | so the sync is useless, IMHO |
| 23:29.02 | DocScrutinizer05 | KatolaZ: a cp is just creating a simple file, while this dd creates a file used for loopmounting right after |
| 23:29.19 | KatolaZ | you read the just-created file in the following line |
| 23:29.21 | DocScrutinizer05 | but yes, you got a point there |
| 23:29.30 | KatolaZ | mkdosfs does that |
| 23:29.37 | KatolaZ | in the following line |
| 23:29.40 | KatolaZ | and anyway |
| 23:29.53 | KatolaZ | since you are creatin an empty dosfs into the img |
| 23:30.05 | DocScrutinizer05 | before mkdosfs comes mount -loop and I don't trust it to not bypass the cache of the parent filesystem |
| 23:30.09 | KatolaZ | whatever you put in the img file is irrelevant |
| 23:30.25 | KatolaZ | mnt is after mkdosfs |
| 23:30.38 | fsmithred | oh yeah, I already checked md5sum is same on both image files |
| 23:30.40 | DocScrutinizer05 | umm, ok, my bad then |
| 23:30.52 | KatolaZ | dd if=/dev/zero of=boot/grub/efiboot.img bs=1K count=1440 |
| 23:30.52 | KatolaZ | /sbin/mkdosfs -F 12 boot/grub/efiboot.img |
| 23:30.52 | KatolaZ | |
| 23:30.52 | KatolaZ | mkdir img-mnt |
| 23:30.52 | KatolaZ | |
| 23:30.54 | KatolaZ | mount -o loop boot/grub/efiboot.img img-mnt |
| 23:31.03 | DocScrutinizer05 | ack |
| 23:31.23 | KatolaZ | with mkdosfs you are actually writing on that file |
| 23:31.31 | KatolaZ | so the cache is synced in any case |
| 23:31.36 | KatolaZ | 101% sure |
| 23:31.41 | DocScrutinizer05 | more importantly, you need to read it |
| 23:32.07 | DocScrutinizer05 | even most imnportant, it's on parent fs level |
| 23:32.30 | DocScrutinizer05 | so the parent fs cache gets flushed |
| 23:32.38 | DocScrutinizer05 | no sync needed :-) |
| 23:32.49 | *** join/#devuan nighttime (watching@freenode/staff/uptime) |
| 23:32.55 | KatolaZ | DocScrutinizer05: :) |
| 23:33.30 | KatolaZ | ok I will call it a day then |
| 23:33.32 | KatolaZ | o/ |
| 23:35.34 | fsmithred | I'm making one more snapshot to prove that there's no difference between the two png files. |
| 23:35.34 | DocScrutinizer05 | considering stuff like journaling filesystems, I even strongly doubt a mount -loop could do a direct block IO shortcut |
| 23:38.07 | DocScrutinizer05 | hmmm waht been that issue with journalling fs and a fsync() after each write to 'fix' the problem? |
| 23:38.30 | DocScrutinizer05 | can't remember |
| 23:38.52 | DocScrutinizer05 | was that something about databases? |
