#archlinux-ports | Logs for 2018-02-05

Back
[00:39:19] <buildmaster> firefox-developer-edition is broken (says buildknecht3).
[01:37:31] -!- isacdaavid has quit [Ping timeout: 256 seconds]
[02:50:23] -!- |ast| has quit [Quit: Lost terminal]
[04:13:28] -!- bill-auger_ has joined #archlinux-ports
[04:13:40] -!- bill-auger has quit [Ping timeout: 260 seconds]
[05:43:28] -!- isacdaavid has joined #archlinux-ports
[05:54:44] -!- hringriin_ has joined #archlinux-ports
[05:57:37] -!- hringriin has quit [Ping timeout: 248 seconds]
[05:57:37] hringriin_ is now known as hringriin
[07:31:12] -!- MrBIOS has quit [Quit: MrBIOS]
[07:45:32] -!- MrBIOS has joined #archlinux-ports
[07:53:21] -!- MrBIOS has quit [Quit: MrBIOS]
[08:02:56] -!- oaken-source has joined #archlinux-ports
[08:52:34] -!- deep42thought has joined #archlinux-ports
[11:42:41] -!- kerberizer has quit [Ping timeout: 248 seconds]
[11:47:24] -!- kerberizer has joined #archlinux-ports
[12:38:21] -!- yans has joined #archlinux-ports
[13:28:39] -!- guys has quit [Ping timeout: 268 seconds]
[13:39:26] <deep42thought> lddtree gives me strange output on /usr/lib/libfakeroot/libfakeroot.so
[13:39:34] <deep42thought> that is, on archlinux32
[14:00:05] -!- guys has joined #archlinux-ports
[14:22:57] -!- guys has quit [Ping timeout: 240 seconds]
[15:34:35] -!- migul has joined #archlinux-ports
[15:46:53] -!- guys has joined #archlinux-ports
[16:45:43] <buildmaster> firefox-developer-edition is broken (says buildknecht).
[16:47:42] insecta_ is now known as insecta
[16:49:55] <tyzoid> deep42thought: install 16.04 on it, then complain to canonical that ubuntu is crashing your machine :P
[16:50:06] <deep42thought> :-D
[16:50:21] <deep42thought> I'll do that right after the new power supply
[16:51:07] <tyzoid> btw, my preferred stress-test is boinc seti@home, since it'll stress my GPU and CPU simultaneously
[16:51:22] <deep42thought> this machine runs boinc permanently
[16:51:33] <deep42thought> I'm not sure, that it's running seti, though
[16:52:30] <deep42thought> but it's only a 12W cpu, so I suppose the stress is marginal ;-)
[16:56:35] <deep42thought> oh, it's late, gotta go ...
[16:56:36] <deep42thought> cu
[16:56:38] -!- deep42thought has quit [Quit: Leaving.]
[17:02:00] -!- MrBIOS has joined #archlinux-ports
[17:02:46] -!- MrBIOS has quit [Client Quit]
[17:18:15] -!- MrBIOS has joined #archlinux-ports
[18:00:31] -!- bill-auger_ has quit [Read error: Connection reset by peer]
[18:00:53] -!- bill-auger has joined #archlinux-ports
[18:45:44] -!- isacdaavid has quit [Quit: Leaving.]
[19:04:17] -!- deep42thought has joined #archlinux-ports
[19:42:55] -!- migul has quit [Quit: Leaving]
[19:54:46] -!- yans has quit [Quit: chaos is the only true answer]
[21:01:37] <deep42thought> I compiled a list of to-be-deleted packages and put it on https://buildmaster.archlinux32.org
[21:01:50] <deep42thought> These are packages which the build master believes should be deleted
[21:02:00] <deep42thought> if anyone has any objections, let me know :-)
[21:02:54] <deep42thought> note, that for lazyness I always appended the epoch and sub_pkgrel
[21:05:39] -!- migul has joined #archlinux-ports
[21:08:07] -!- abaumann has joined #archlinux-ports
[21:08:13] <abaumann> they can always be readded.. :-)
[21:09:05] <abaumann> most packages are from community, none from core, some from extra. seems reasonable.
[21:10:43] <deep42thought> yeah, actually, I could check that all packages are in tyzoid's archive :-)
[21:13:23] -!- MrBIOS has quit [Quit: MrBIOS]
[21:14:31] <tyzoid> deep42thought: are these old versions of stuff?
[21:14:56] <deep42thought> no, these are packages which won't be built for one or another reason
[21:15:12] <deep42thought> but they are in our repositories, because they were built in the past
[21:15:14] -!- MrBIOS has joined #archlinux-ports
[21:15:14] <tyzoid> community/atom-0:1.23.3-1.0-i686.pkg.tar.xz <-- I don't think we should delete this package
[21:15:56] <tyzoid> I don't think we really have a choice with CUDA
[21:16:00] <tyzoid> but it'd be nice not to remove it
[21:16:19] <deep42thought> buildmaster: why dont you keep atom
[21:16:33] <deep42thought> tyzoid: is there a point in keeping packages which we cannot update?
[21:16:42] <deep42thought> e.g. cuda
[21:16:50] <buildmaster> Package "atom" has dependencies on the deletion list:
[21:16:50] <buildmaster> electron builds electron
[21:17:01] <abaumann> :-)
[21:17:25] <abaumann> physics in the buildmaster.. ;-)
[21:17:34] <tyzoid> deep42thought: If it doesn't pose issues, then why not? If it's currently posing issues, then I don't see a way around removing it.
[21:17:53] <tyzoid> CUDA in particular can be quite useful
[21:18:18] <abaumann> newest cuda is simply not supported anymore by upstream.
[21:18:42] <abaumann> but we can put it into the AUR, like for opera -> AUR: opera32
[21:18:55] <tyzoid> I think AUR is probably a good way to go.
[21:19:04] <deep42thought> yeah
[21:19:18] <tyzoid> is electron unsavable?
[21:19:37] <tyzoid> buildmaster: why dont you keep electron
[21:19:58] <deep42thought> this function is yet to be optimized
[21:19:59] <abaumann> it had a Javascript engine issue which is not 32-bit anymore..
[21:20:01] <deep42thought> (accellerated)
[21:20:09] <buildmaster> Package "electron" is explicitely blacklisted.
[21:20:37] <tyzoid> abaumann: Is it possible to patch out? Electron is kinda required for things like slack, discord, atom, etc.
[21:21:07] <abaumann> I didn't try.. maybe it's just used for scripting, yes.
[21:22:53] <tyzoid> ok. deep42thought: I didn't see much else. Perhaps we should post this to the ML?
[21:23:08] <abaumann> good idea.
[21:23:10] <deep42thought> yeah, good idea
[21:23:52] <tyzoid> For me, I think electron/atom are two that we should keep, even if we can't update further.
[21:24:09] <tyzoid> since they're used by a bunch of other stuff
[21:25:54] <bill-auger> regarding electron that is said to be a hopeless cause
[21:27:11] <tyzoid> bill-auger: True, but so much future development requires it, and I don't see that trend slowing down.
[21:27:25] <bill-auger> there is an ongoing discussion about it's licensing issues (and all chromium derrived programs) on the linux-libre mailing list - this post specifically what i leared about electron https://lists.gnu.org
[21:27:26] <phrik> Title: Re: [directory-discuss] FSF opinion on chromium, QtWebEngine, electron (at lists.gnu.org)
[21:28:25] <abaumann> that's an interesting link. thanks.
[21:28:27] <bill-auger> literally then what you are saying is - the "trend" that is not slowing down is a a trend to package programs in a way that violates their licenses
[21:28:57] <abaumann> from this point of view, electron belongs into the AUR anyway.
[21:29:05] <abaumann> no matter what depends on it.
[21:29:31] <bill-auger> i do agree though it is a popular trend - also the same with qt-webengine vs qt-webkit
[21:30:17] <tyzoid> So, if we remove electron from our repos, would npm install -g electron still work? Or does that have the same issue of no-i686 support?
[21:30:24] <bill-auger> nearly once a month some some program in parabola needs to be blacklisted merely for their decision to change from qt-webkit to qt-webengine
[21:30:53] <abaumann> yeah. this is a sad thing.
[21:30:58] <bill-auger> surely you can use an external package manager - but that is never really advisable anyways
[21:31:08] <abaumann> tyzoid: I would guess, it's the same issue..
[21:31:14] <guys> bill-auger: uhhhh... electron was built from source fro beginning to end, IIRC
[21:31:42] <tyzoid> bill-auger: Unfortunately, devs will often do easier thing, rather than right thing.
[21:31:50] <guys> https://git.archlinux.org
[21:31:51] <bill-auger> built by the upstream from source - and then distribute in mainly binary form without sources
[21:31:51] <phrik> Title: PKGBUILD\trunk - svntogit/community.git - Git clone of the 'community' repository (at git.archlinux.org)
[21:31:55] <abaumann> elecron, a new java. cool. :-)
[21:32:34] <guys> We extract chromium sources and patch them to use system ffmpeg et al, then use that in our electron build
[21:32:44] <bill-auger> i was only reporting what is fedora's decision for blacklisting it - ive never looked at it myself
[21:32:58] <guys> And Arch32 is simply using the latest version of archlinux.org's electron that actually built for i686
[21:33:35] <guys> You'll forgive me for not being a huge fan of libre distributions, then?
[21:33:38] <guys> :p
[21:34:23] <guys> Though I do agree that electron is horrible as a toolkit
[21:34:39] <guys> It is simply ewwww
[21:34:59] <tyzoid> As a user, I dislike it because of the insane resource requirements.
[21:35:19] <tyzoid> but that aside, people seem to like it, or at least tolerate it enough to use it quite a bit
[21:35:26] <guys> atom, is the most ironic text editor in the history of the human race.
[21:35:34] <tyzoid> lol
[21:35:35] <tyzoid> true
[21:35:47] <abaumann> :-)
[21:35:52] <deep42thought> !grab guys
[21:35:52] <phrik> deep42thought: Tada!
[21:36:04] <guys> fair enough :D
[21:36:17] <tyzoid> lol, quotegrab in #archlinux-ports. A rare occurrence indeed.
[21:36:27] <guys> OTOH you do absolutely need web engines for e.g. reading ebooks
[21:36:56] <guys> I don't think anyone was insane enough to do that with electron... it is usually the qt/gtk bindings that are used for ereaders.
[21:36:56] <tyzoid> guys: I don't see why?
[21:37:06] <guys> tyzoid: because ebooks are written in html5
[21:37:13] <abaumann> unzip + lynx?
[21:37:13] <guys> it is literally the spec
[21:37:19] <tyzoid> for which format?
[21:37:22] <tyzoid> .epub?
[21:37:23] <abaumann> epub
[21:37:29] <guys> I dare you to read a fixed-format EPUB3 in lynx
[21:37:51] <abaumann> aeh.. maybe not. :-)
[21:37:54] <tyzoid> couldn't the gnome devs just expand evince to support epub?
[21:38:07] <guys> Maybe they did already.
[21:38:15] <guys> I know xreader should support it via webkit
[21:38:34] <tyzoid> webkit at least beats electron in resource management.
[21:38:56] <tyzoid> I don't need 200MB of ram to load a single blank page
[21:39:36] <bill-auger> i agree with that too
[21:39:54] <guys> tyzoid: nah, AFAICT evince still does not support EPUB, while cinnamon's evince fork, xreader, does through webkit2gtk
[21:39:57] -!- oaken-source has quit [Ping timeout: 240 seconds]
[21:40:02] <guys> I should know, as I am the xreader packager :D
[21:40:04] <bill-auger> QT people have a very unhealthy habit of piling on the fattest new things onto places where is vastly simpler thing will do
[21:40:53] <bill-auger> i think their motto should be "well my PC has 32GB of RAM - why doesnt yours - you suck"
[21:41:32] <guys> I like Qt and found it worked quite well on my 2GB RAM machine.
[21:41:52] <guys> Including calibre's qt5-webkit ebook reader.
[21:42:55] <tyzoid> bill-auger: s/QT/node/
[21:45:28] <tyzoid> guys: That is interesting though, I might need to switch from evince to xreader at some point.
[21:46:20] -!- abaumann has quit [Quit: leaving]
[21:46:33] <bill-auger> oh javascript people are much worse - luckily their stuff does uses on the order of KB RAM not MB or most websites would be utterly unusable by the majrity of PCs
[21:47:13] <tyzoid> bill-auger: I find the opposite to be true. That's the reason I'm forced to run adblock
[21:47:40] <tyzoid> It's not the javascript itself, but all the objects it loads in. Audio, Video, Autoplay, etc.
[21:48:02] <deep42thought> slow unfolding menus
[21:48:07] <guys> tyzoid: it's a shame the gtk/gnome devs are so insane that cinnamon ended up forking everything in order to provide a stable and decent-looking desktop
[21:48:15] <tyzoid> But esp. given the new nature of things like reactjs, people package huge libraries into objects and shove it all at a browser
[21:48:26] <guys> otherwise a unified evince package might have served everyone...
[21:49:44] <bill-auger> that is a similar philosophy but more like "well my internet connection is 20MB /sec - what's yours - you suck too"
[21:51:29] <bill-auger> broadly speaking developers (especially gamers) are not always aware that their dev boxes are in the top 98% performance of PCs - but most of the computers in the world can not run the programs they put out
[21:53:02] <bill-auger> i had a discussion about 5 years ago with someone who was making a website entirely in webGL and he was very distrubed when i told him to realize that 90% of the people in the world would not be actually able to use his website for 5-10 years
[21:53:50] * bill-auger 5-10 years is like multiple lifetimes to a gamer
[21:55:29] -!- deep42thought has quit [Quit: Leaving.]
[21:55:34] <bill-auger> designing for the lowest common denominator may be a good design practice but it sounds like blastphemy to gamers and other bleeding edgers
[22:45:41] -!- migul has quit [Quit: Leaving]
[22:57:09] -!- jemadux has joined #archlinux-ports