09:13 <misterbzr> jaeger: Script worked great.

09:13 <misterbzr> since working on an archlinux laptop I changed to pkgadd to bsdtar.

09:30 <cruxbot> [xorg.git/3.6]: mesa: update to 22.0.4

09:31 <cruxbot> [opt.git/3.6]: lvm2: update to 2.03.16

09:33 <cruxbot> [core.git/3.6]: libdevmapper: update to 1.02.185

12:00 <farkuhar> the video that jaeger posted (10 most awesome Linux distros) has this gem of a quote: "it's slightly difficult to remove dependencies in crux after uninstalling a package". Upon skimming quickly through the comments, I didn't see anyone telling her about pkgfoster.

12:04 <farkuhar> One downside of pkgfoster, however, is that it only looks at hard dependencies. If for example you have texlive present when compiling zathura, then zathura will be built with support for synctex. If you later remove zathura, pkgfoster will NOT report the no-longer-needed soft dependency texlive.

12:11 <farkuhar> /usr/bin/finddeps does an exhaustive search to determine which packages own the shared libraries that a port depends on. pkgfoster only gives an approximate inverse of this relation, because it ignores soft dependencies.

12:17 <ppetrov^> farkuhar, sorry to interject, but this reminds me of an issue I had when using packages built on my main PC as a packages repo for my laptop. pkg-get would install dependencies extracted from the ports information, but will not of course install other stuff that was present on my build machine

12:18 <ppetrov^> i made my own, very simple script to report which packages contain libraries that are missing/required by installed stuff

12:18 <ppetrov^> https://github.com/slackalaxy/depsck

12:19 <ppetrov^> it's the revlibpkg script. I am open to comments, suggestions and degrading criticism

12:25 <SiFuh> SO agressive, such an unfriendly channel ;-)

12:28 <ppetrov^> me, being aggressive?

12:29 <ppetrov^> :P

12:39 <farkuhar> ppetrov^: no need to apologize for interjecting; we're here to have a conversation, not 50 separate monologues. I am fascinated by your experience with pkg-get, though. I was under the impression that distcc is preferred over pkg-get, among users who want to share the resources of their workstation with a less powerful laptop.

12:42 <ppetrov^> distcc is too complicated for me. I really wanted to make a simple packages repo, that's all

12:42 <SiFuh> Hahahaha, no I playing with that video Jaeger sent that CRUX is most popular and the CRUX guys are very nice and helpful (paraphrasing)

12:45 <farkuhar> ppetrov^: since revlibpkg is interpreted by bash, your custom function count_input() can be simplified to remove the redundant calls of echo and wc. Similarly, is_input_inst() can eliminate the unnecessary grep -q and just rely on the exit status of prt-get.

12:48 <ppetrov^> thanks, i will look into it

12:51 <farkuhar> also you can use the substring parameter expansion to test whether an argument starts with a dash: if [ "${t:0:1}" = "-" ]; then ... It might be faster than the glob you're using now.

12:53 <ppetrov^> thanks

12:54 <ppetrov^> in general, i have some problems to elegantly pipe between programs, so instead i would make something into a variable and then proceed from there. Which leads to redundancy

12:56 <farkuhar> You're in good company: https://github.com/matloff/TidyverseSkeptic (Scroll down to the section "Should we teach using pipes or functional composition? Neither!")

12:58 <ppetrov^> thanks, i'll check it

13:03 <ppetrov^> heh, i just watched the CRUX part of the video...

13:44 <Guest69> hello

13:45 <Guest69> is the release coming soon?

13:45 <Guest69> is a new release coming soon?

14:05 <SiFuh> Hehe, as Guest69 steps out of the saloon and looks left, the right and a tumbleweed rolls by.

14:05 <SiFuh> Guest69: it won't be rushed, so it will happen when it happens

14:09 <SiFuh> Moments before the release it could be halted, due to an issue. So it is always hard to say

14:17 <jaeger> misterbzr: great, glad to hear it

15:16 <farkuhar> ppetrov^: http://ix.io/3YdK is a slight reworking of your revlibpkg script, hopefully with less redundancy.

15:40 <farkuhar> correction (case-sensitive search): http://ix.io/3YdP

16:04 <SiFuh> Hmm lots of 2 vs 4 spaces in there

16:13 <ppetrov^> hey farkuhar , thank you very much.

16:14 <ppetrov^> ${t:0:1} <- this would have never occurred to me. I know what it does, simply i would have not thought of it

16:14 <SiFuh> ppetrov^: yeah it looks good

16:14 <ppetrov^> heh and feel free to check the other two scripts

16:27 <ppetrov^> {print gensub(/.*\/(.+):/,"\\1","g",$NF)}

16:28 <ppetrov^> ok, that's a mystery for me that i will spend some time to decipher

16:28 <ppetrov^> :)

18:18 <farkuhar> ppetrov^: the awk gensub command uses a back-reference \\1 to substitute the contents of the parenthesized group in the regex.

18:18 <ppetrov^> well, i'll need some time to 'get it'

18:18 <farkuhar> That way the substitution removes everything up to the last path delimiter, as well as the closing colon.

18:19 <ppetrov^> i appreciate all your help

18:19 <farkuhar> no problem.