<nojr>Hi, quick question: Why is the Django package stuck at v 1.25?
***jonsger1 is now known as jonsger
<apteryx>nckx: thank you, I will check it a bit later
<apteryx>has anyone else noticed how running QEMU on Guix resets the terminal and then you loose the scroll-back buffer? Especially annoying in tests, where something crashes before QEMU starts, yet it cannot be inspected.
<apteryx>someone from #coreboot told me it must be seaBIOS sending an escape code; someone from #qemu told me my terminal is misbehaving (could reproduce the problem in xterm but not in gnome-terminal and others).
<apteryx>I could reproduce the issue with sakura & xfce4-terminal too.
<apteryx>you can try "C-l" in your terminal to see if you have the issue.
<apteryx>apparently QEMU does more than "C-l", because in xfce4-terminal I can scroll back after pressing C-l, but running QEMU, the scroll back is still deleted. Humph.
<roptat>it happened to me outside of qemu, I couldn't scroll back on a terminal anymore, but that was on fedora
<guix-vits>ME please unite, please separate, i was meaned ;)
<zzappie>efraim: whoa :) I was expecting vim user to have done it.
<nckx>raingloom, guix-vits: Actually if the arch patches work without modification I prefer to keep them intact, each with a header noting ‘copied verbatim from Arch at <link>’. Makes them easier to track compared to creating another subtly different variation.
<nckx>See gnu/packages/patches/xf86-video-voodoo-pcitag.patch for how I like to do it (habit from another project), but choose whatever format you like.
<reepca>tor button in icecat doesn't seem to be working... it shows that I'm connected to the proxy but https://check.torproject.org/ indicates that traffic isn't being routed through it. Any idea what's up with that?
<raingloom>nckx: in this case Arch also just copied it and the header has some info on where it's from. that should be enough, right?
<zzappie>I got a question: I sometimes need to quickly run and test a random project that says something like just sudo pip/npm install loads of stuff that is not in guix packages. Do manifests provide means to do arbitrary scripting after profile with core packages has been initialized?
<zzappie>or better put, what do you think is the best way to handle situations when you dont have time to write package definitons and need quickly spawn isolated invironmmanen containing bunch of stuff installed by particular language's package manager on a guix system and make it persist between invovations
<guix-vits>zzappie: "Sometimes it is desirable to isolate the environment as much as possible, for maximal purity and reproducibility." --container at the link above? "Fixed" is wrong wording.
<zzappie>Im doing something like "guix environment --manifest=manifest.scm --user=bcdb-js --container --ad-hoc coreutils --link-profile". Than I could run some commands on top. But as if I get it right changes made won't persist between guix environment invocations
<zzappie>sirgazil: Oh i've missed --profile option for "guix package". Yeah I know that it is beter to "guix import" and package things for reproducibility reasons. But Im not yet proficient enough with guile and guix to do it quickly enough. So I'm mixing two worlds here. But im gradually incorporating guix into my projects
<sirgazil>zzappie: Yeah, when I wanted to create development environments, I thought "guix environment" was the thing to do.
<zimoun>civodul, janneke: the bug#39575 about time-machine is weird. janneke and I, we are not failing for the same reason when building (python vs guile ;-)). Moreover I have noticed other commits around the offending one failing too (build error of Guile 2.2.4 or Python 3.7.2).
<zzappie>Speaking of docs... I've came accros a nice article about documentation. https://www.divio.com/blog/documentation/ Is there any roadmap for documenting things? It seems like guix grows quickly and and using guix has an effect on many development activities thus might benefit from different more interactive types of docs.
<nckx>zzappie: No roadmap I know of. ‘Just volunteer’ 🙂 That article is part of what led to the Cookbook.
*guix-vits hates Cloudflare and its captchas. Some people just use tor to solve bad routing issues.
<zzappie>nckx: Nice :) meanwhile ROADMAP was last updated in 2015. What about guix blogging contest?
<zzappie>this way we also popularizing guix in different communities and platforms
<nckx>zzappie: Do you see this as an actual contest (with announcements, winners, maybe prizes) or just encouraging more people to blog more about more Guix?
<guix-vits>Funniest thing is, that anyone can popularize or promote [gi:ks]... But critique in speeches is somewhat dangerous without a careful specifying (OS, err.. SD.. khm... we mean that exact distro from gnu.org!), or else a lot of [gi:ks] may be angered.
<zzappie>nckx: I duno I just made it up :) Could be both
<guix-vits>nckx: Are the <strike>D'Artangnans of Free Software</strike> Geeks really not that quick to anger?
<nckx>guix-vits: Apart from the package manager/system split (which isn't really a split IMO, but some people see it that way), what do you feel one needs to be ‘careful’ about? There is only one GNU Guix and no forks that I know of.
<nckx>guix-vits: I don't know how much you're joking, but if this is really your impression (even in jest) I'm genuinely surprised.
*guix-vits performs JOKE (vocal): ... yes, just there is two real [gi:ks] distros (which is really only one). The <strike>Guardians of Cardinal</strike> Arch, Debian, Fedora, Gentoo ... Puppy ... are cool, but aren't [gi:ks] distributions, of course (m'sew)!
*zzappie guix-vits minds are going to be blown when someone will port guix to bsd.
<roptat>Examining my system, I found that libevent has a dependency on python, but that's only for a script. Putcing it in a separate output should reduce the closure size from 180 MB to about 80 maybe. I'll try to do that tonight :)
<roptat>Since it's required by ntp, it's about 80 MB that can go away from my system's (and everyone else's system) closure