IRC Logs for #crux-devel Monday, 2019-05-13

NomiusIt will01:43
TimB_pedja: are you planning to release the port on April 1st 2020? :))08:25
pedjaTimB_, I've learned over time not to plan that far ahead :)11:44
pedjalife is always 'you made plans? LOL, you fool'11:46
pedjasystemd port is the easy part, already have that somewhere. integrating it nicely is where the actual work is11:48
pedjaupstream default of KITCHEN_SINK isn't helping, either11:51
pedja2/3 of the Pkgfile is just disabling shit :)11:52
pedjawhat complicates things is that some parts are 'optional', but not really11:53
pedjasome stuff depends on it, and, as a surprise to no-one, no documentation mentions the fact11:54
pedjaso, if CRUX is, for some reason, forced to switch at some point, someone smarter then me(not that high of a bar, granted) will have to do it11:56
*** xor29ah has quit IRC12:15
*** xor29ah has joined #crux-devel12:16
pedjayay, polkit breakage. everything looks OK, no errors/warnings, and yet somehow it doesn't work with libvirt.13:18
pedjaI am tired of it, tbh13:18
*** stenur has joined #crux-devel14:11
*** onodera has joined #crux-devel17:00
*** dlcusa has quit IRC17:05
TimB_pedja: sorry, was at uni, then work :) uh, I hope the day never comes :D17:25
TimB_pedja: talking about a high bar. My rstudio port was working after all, I fucked up my r port by removing a file named INSTALL in the folder /usr/lib/R/bin :D17:27
TimB_what I've learned? not all files prtverify moans about are crap files :P https://crux.ninja/portdb/port/tb/r/ this now screws my perfect rating on jaegers portdb :P17:28
pedjaheh. the same happened to me with libguestfs-appliance. who knew it checks for README in the directory17:28
TimB_oh man, even a README? I mean, INSTALL in a folder named bin.. I should've checked. but README..17:29
pedjafun fact: that the README is required is actually part of the README. did I read it before? hell no :)17:32
TimB_ha :P17:32
TimB_well, I mean, I figured something was wrong when I noticed that I wasn't able to install any R packages from within R - said it's not made to compile packages and maybe I have installed it from an RPM? somebody in R went digging with me and found something that linked to the exact INSTALL file. I wish I had a picture of my face when I knew what's wrong17:35
TimB_I have been investigating that for weeks every now and then.. with that done, I have no idea what else to do with my life :P17:36
pedjait should've been named README.ffs instead of README.fixed :)17:37
TimB_I'd sign this petition :P17:37
pedjawhen in doubt, I check opensuse or fedora packages, if available17:39
pedjanice to have a reference by people that actually know what are they doing :)17:40
TimB_he :P I can't say I do that, but I try to look at other packagers work too ofc :)17:41
TimB_I took advice from j_v's openjdk8 port for the unpack_source() hook and utilized it in my rstudio port17:42
TimB_https://raw.githubusercontent.com/TimB87/crux-ports/master/rstudio/Pkgfile 8-)17:42
TimB_I guess if I wouldn't have found my error to be with R, I would still be suspecting rstudio to do something funny17:44
TimB_or qtwebengine or qtwebkit or whatever17:45
*** dlcusa has joined #crux-devel17:45
jaegerTimB_: perhaps prtverify could be modified to understand some exceptions17:47
TimB_jaeger: I would like that, I haven't checked the code for it yet tbh :) Just discovered that yesterday (finally)17:48
*** Workster has joined #crux-devel17:48
jaegerIf I remember right jue wrote it17:48
TimB_but I wasn't thinking when I put the find line in the Pkgfile, just wanting to please prtverify17:48
jaegerIn the case of R, is that a text file they've put into bin/ for some reason or is it actually a script or executable?17:49
TimB_it's a script17:49
TimB_http://sprunge.us/63O6ZU <- 75517:50
jaegerunusual for scripts to have all-caps names17:50
TimB_indeed. I haven't given it a second thought until I saw that there was indeed such a file needed17:52
*** stenur has quit IRC18:00
*** onodera has quit IRC18:09
*** onodera has joined #crux-devel18:10
TimB_/usr/lib/prtverify/30_junk_files.awk line 14 :)18:14
TimB_not sure what would be a better rule though, excluding anything in bin dirs?18:14
jaegerI was thinking more along the lines of a single file override for a single port... like an included override.d dir18:25
jaegerbecause I think it's not going to be very common that exceptions are needed18:25
jaegeranother silly example is https://crux.ninja/portdb/port/jaeger/iucode-tool/18:29
jaegerthough I'm going to drop the port since it's no longer useful18:29
TimB_oh, right. I agree that this is a very special case, so yeah. I know about it now, I can overlook prtverifys output for it, just a bit bummed for the the 100% perfect score on the unoffical portdb :D18:31
jaegervery few have perfect scores :) Besides, nobody's grading, I just put that in for information18:55
jaegerin the case of R it seems to be a valid false positive18:57
TimB_I still liked the 100% though :D19:00
jaeger:)19:00
TimB_and yeah, it wasn't easy to track down. I first suspected just rstudio to be broken. and I was bummed for quiet a time with it now. then found that in an r shell: http://dpaste.com/1WZ7PSX19:01
jaegersomewhat ambiguous message19:03
TimB_somebody over at #r made the hint that led to my conclusion (basically linked that: https://github.com/hadley/packman/blob/master/old/install.r to line 370)19:03
TimB_haha, yeah :)19:03
TimB_I'm glad it's sorted out19:04
jaegeryeah, good deal19:04
TimB_jaeger: btw, what is the rule of thumb for contrib, I linked a contestant list sometime ago and didn't get much feedback so far :)19:06
jaegerTo be honest it's probably just that since tek is mostly MIA we don't have anyone officially managing contrib users. I'll look back, was it via email?19:11
jaegernm, I see it on crux-devel from march19:12
TimB_oh, no, here on irc. but that's another rather good topic?19:12
TimB_no idea how far that has changed tbh, but I planed on moving biber and maybe the whole libreoffice repo as well19:12
jaegerI see that jue and frinnst both approved, so I can go ahead and add you I think19:13
jaegeractually, looks like frinnst already did that... so I'm missing something19:14
TimB_libreoffice is in a nice and shiny state as far as I can say. biber is just perl modules but a nice addition to texlive19:14
TimB_oh, I thought you meant my list about ports I would move to contrib. sorry. I have my keys :)19:14
jaegerI would say add whatever you think is useful and that you will maintain to contrib... if it's already there, coordinate with the current maintainer to take it over ot work with them, etc.19:14
jaegerAh, ok. Sorry, I misunderstood19:14
TimB_ok, great. I think I can't do much wrong with that :)19:16
jaeger:)19:16
TimB_regarding tek, I wrote him a mail about 2 of my fs issues open19:16
TimB_maybe you can take a look at opt/texlive. it's been broken for months and I posted a fixed version. would be nice to keep opt clean19:18
jaegertek_: care to weigh in? :) If you're too busy to work on crux stuff that's fine, just let us know if it's ok to do updates on your behalf or the like19:19
TimB_he is not reacting to my fs issue about libreoffice for weeks so I don't expect an early answer for that either19:19
TimB_on another note: I would like seeing the port health check on the official ports db. I think it encourages maintainers to have a look if they forget to check the health, but it can also warn a new user that the port they pull might be broken to start with19:31
jaegerWouldn't hurt my feelings, though the official portdb would need some work to support it19:33
TimB_I like the newer layout overall better really19:33
TimB_needs a README field if a port contains one to paste it :) but else19:34
jaegerwhat do you mean by 'to paste it'?19:38
TimB_https://crux.ninja/portdb/port/onodera/blender/ <- let's say onodera had a README for blender included in the ports dir19:39
onoderawow a custom port db19:40
TimB_just have an area on that page that displays the content? just a little idea to add some more useability19:40
onoderaI didnt know about this19:40
onoderait supports git repos?19:40
jaegerAh, just display it... I didn't know what you meant by paste in that context19:40
TimB_hehe, sorry, sometimes I just choose the wrong words hopping between german and english content19:41
jaegeronodera: it does, though not heavily tested19:41
onoderaneat19:41
jaegeronodera: yours is the only git repo in it at present19:41
onoderaim just here because of the mention of my name ;)19:41
onoderaah, that's a honor :)19:42
TimB_sorry :D I was just browsing through19:42
TimB_yanked the link I had at hand19:42
TimB_but isn't the mate repo also considered a git repo?19:42
jaegerthat would be pretty easy to add, probably19:42
jaegerIt could be used via git but it also has httpup support19:43
TimB_oh, git as a ports checkout driver, right19:43
juejaeger: why silly, we've decided sometime in the past that a Pkgfile should only contain ascii characters?19:53
jueTimB_: just checked your rstudio port, there is indeed a strange character at the end of line 46 ;919:55
jaegerJust a silly thing to worry about. I'm not asking to change it. It has no functional effect on the port or anything19:55
jaegerjue: we don't have 'ports' itself in git, right?19:55
TimB_oh, sorry about the character. will delete it, should be normal utf819:58
TimB_jue: but it was regarding the r Pkgfile :) not rstudio19:58
TimB_since r ruined my perfect score I thought unicode for rstudio was indeed ok :P19:59
TimB_https://github.com/TimB87/crux-ports/blob/master/r/Pkgfile line 69+7020:01
juejaeger: I think we should have, will look20:03
jaegerjue: I'd like to make a minor release of ports to remove the cvsup reference from ports(8), didn't see it in gitweb20:03
jaegerIt doesn't appear there's a tools/ports.git repo at all20:05
juehmm, right, looks like we lost it sometimes20:18
*** stenur has joined #crux-devel20:18
jaegerdoh20:19
jueI have a very old checkout here, last commit is from 201520:19
jaegerI wonder if I have any in my automated backups, will check20:20
jueTimB_: yeah, I see but would also say that INSTALL is very uncommon file name for a script20:23
jaegeryeah, that's why I suggested one-off overrides in an override.d folder or the like20:23
jueyeah, currently we have only one central whitelist that is normaly not really up-to-date20:25
jueand for sure only thought for core ports20:26
TimB_jue: agreed :)20:30
jueprtverify was mainly a project to test if it's possible to write something like that in awk, but, TBH, a rewrite in ruby or similar would be really appreciated ;)20:39
TimB_if I would show my ruby scripts I would probably need to hide under the stairs for a while :D20:45
jaegerjue: I have a ports.git in my backups, seems to be last touched 2017-07-2920:45
juefine, that's good news20:45
jaegerlatest changelog entry is: 1.5 - Released 2005-04-1320:46
jaegerseems to match the ports-1.5 tarball content20:47
jaeger$ diff -qr ports-test ports-1.520:47
jaegerOnly in ports-test: .git20:47
jaegerwant me to push that up to tools/ports.git on crux.nu?20:48
jueyeah, please do20:48
jaegerok, will do... and I'll add my small change after20:50
juejaeger: just saw that I've 4 patches for ports that newer got merged, ok if I send them to you for review?21:01
jueall are optimizations like remove of unneeded 'cat' calls etc.21:03
jaegersure, feel free to send them my way21:05
jueok21:05
*** onodera has quit IRC21:30
jaegerGot the patches, I'll review them after work21:54
*** deus_ex has joined #crux-devel23:12
*** pedja has quit IRC23:13
*** stenur has quit IRC23:58

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