#archlinux32 | Logs for 2018-05-08

[00:37:11] -!- kadiro has joined #archlinux32
[00:38:51] <kadiro> I'm back for few minutes :) , a little question, How to update my /etc/os-release , because I have the old one that have antergos on it
[00:40:38] <kadiro> This is what I have inside it http://ix.io
[00:42:35] -!- buildmaster has quit [Ping timeout: 265 seconds]
[00:45:13] <kadiro> ok, may be i will be back tomorrow to ask again, have a good day you all
[00:47:33] -!- kadiro has quit [Quit: Leaving]
[01:03:02] -!- buildmaster has joined #archlinux32
[04:29:35] -!- davor has quit [Ping timeout: 276 seconds]
[04:33:56] -!- davor has joined #archlinux32
[05:56:06] <girls> kadiro: mine simply says https://ptpb.pw
[05:56:18] <girls> though, I think, we should add some "32"s there :-D
[05:58:40] -!- buildmaster has quit [Remote host closed the connection]
[05:58:45] -!- titus_livius has quit [Remote host closed the connection]
[05:59:59] -!- buildmaster has joined #archlinux32
[06:01:41] <tyzoid> girls: Yeah, probably
[06:01:42] -!- titus_livius has joined #archlinux32
[06:02:01] <tyzoid> oh, dang. Now we've got a message to patch xD
[06:02:04] * buildmaster goes insane.
[06:02:11] <girls> tyzoid: could you not wait another 2 seconds with your answer?
[06:02:12] <girls> :-/
[06:02:14] <tyzoid> just about sums it up xD
[06:02:16] <girls> :-D
[06:02:33] <tyzoid> also good morning
[06:02:40] <girls> good evening!
[06:02:57] <tyzoid> nope, it's morning here xD
[06:03:02] <tyzoid> just really really really early
[06:03:21] <tyzoid> anyway, here's what I was trying to do with the email system: https://dl.tyzoid.com
[06:03:44] <tyzoid> I diagrammed it, since the guys on #postfix didn't seem to understand what I was trying to do
[06:03:51] <tyzoid> nor did they seem to think it was going to work :/
[06:03:55] <girls> :-D
[06:03:57] <tyzoid> even though it seems like it should
[06:04:24] <girls> well, as far as I understood, the problem is the arrow you didn't draw
[06:04:36] <girls> namely from the lower right mail server to "inbound mail"
[06:05:02] <tyzoid> on page 2?
[06:05:09] <girls> oh
[06:05:14] <girls> haven't seen that :-/
[06:05:34] <tyzoid> page 1 is inbound from internet
[06:05:39] <tyzoid> page 2 is "inbound" from user
[06:06:24] <girls> !wtf os-release
[06:06:26] <phrik> girls: core/filesystem community/certbot
[06:06:40] <tyzoid> buildmaster: wtf os-release
[06:06:43] <girls> !wtf /etc/os-release
[06:06:43] <phrik> girls: What the hell is /etc/os-release?
[06:06:45] <buildmaster> [core] filesystem (2018.1-2.0): /usr/lib/os-release
[06:06:45] <buildmaster> [community] certbot (0.24.0-1.1): /usr/lib/python3.6/site-packages/certbot/tests/testdata/os-release
[06:06:58] <girls> hrrrm
[06:07:07] <girls> who installs that file???
[06:07:37] <tyzoid> post-install maybe?
[06:07:46] <girls> yeah
[06:07:49] <tyzoid> Polichronucci: Btw, that's what I was trying to achieve
[06:07:58] <girls> it was a rethorical question
[06:08:06] <girls> ... I don't like the post install hooks :-)
[06:08:19] <tyzoid> it seemed like I didn't explain myself well, so the diagram should hopefully help
[06:13:31] <girls> it looks, like core/filesystem has some sort of template in /usr/lib/os-release which will be used for /etc/os-release (by pacstrap?)
[06:18:14] <girls> how about that? https://ptpb.pw
[06:18:24] * buildmaster resumes sanity.
[06:21:55] <girls> cu later!
[06:35:54] -!- buildmaster has quit [Ping timeout: 265 seconds]
[07:12:41] -!- oaken-source has joined #archlinux32
[07:54:07] -!- buildmaster has joined #archlinux32
[08:09:09] -!- deep42thought has joined #archlinux32
[08:09:10] <buildmaster> Hi deep42thought!
[08:09:18] <deep42thought> hi buildmaster, what's up?
[08:09:19] <buildmaster> up? I'm up for 17 minutes, load average: 0.13, 0.19, 0.18
[08:56:37] -!- oaken-source has quit [Ping timeout: 268 seconds]
[10:06:58] -!- eduardoeae has quit [Ping timeout: 264 seconds]
[10:28:35] -!- oaken-source has joined #archlinux32
[11:15:21] -!- Vollzornbrot has quit [Ping timeout: 264 seconds]
[11:39:47] -!- deep42thought has quit [Quit: Leaving.]
[11:41:23] -!- eduardoeae has joined #archlinux32
[12:36:08] -!- dopsi has quit [Quit: ZNC - http://znc.in]
[12:41:41] -!- dopsi has joined #archlinux32
[13:19:56] <eschwartz> deep42thought: os-release is a packaged file... antergos updates this using the archlinux filesystem package, using a step in their calamares installer running sed on packaged files
[13:20:25] <eschwartz> etc/os-release is a symlink
[13:20:41] <eschwartz> $ file /etc/os-release
[13:20:41] <eschwartz> /etc/os-release: symbolic link to ../usr/lib/os-release
[13:20:50] <bill-auger> that is typuical yes
[13:22:38] <bill-auger> /usr/lib/os-release is the canonical vendor-installed version and /etc/os-release may be customized by the admin and is the one clients should read from
[13:23:23] <eschwartz> deep42thought: I debugged this on #antergos :p https://github.com
[13:23:25] <phrik> Title:Put correct distro info in /etc/os-release · Issue #853 · Antergos/Cnchi · GitHub (at github.com)
[13:23:26] <bill-auger> i had to research it a bit because it is also what os-prober loooks for
[13:24:05] <bill-auger> IIRC it will be looking for LIKE='arch'
[13:24:34] <bill-auger> LIKE='archlinux'
[13:27:18] <eschwartz> I don't see a "LIKE" field...
[13:27:57] <bill-auger> it is optional intended for forks such as antegros
[13:28:07] <bill-auger> archlinux itself would not have it
[13:28:46] <bill-auger> but parabola manjaro and other forks should add it to indicate that the system is "like archlinux"
[13:29:46] <eschwartz> I was looking at the man page, not the archlinux os-release :p
[13:29:50] <bill-auger> so clienta could derrive for example the package manager name and other characteristics like special file locations
[13:29:55] <eschwartz> It does specify ID_LIKE though
[13:30:04] <bill-auger> yes that it
[13:30:09] <bill-auger> i was only going from memory
[13:30:36] <eschwartz> would require packaging their own filesystem though...
[13:31:21] <eschwartz> hmm, I wonder if that should be a backup=() file in our package
[13:31:44] <bill-auger> ID_LIKE should be canonical - define by the upstream - and all forks should use the same one - i dont think arch ever officially defined it has but the most common one ive seen is 'archlinux'
[13:32:55] <eschwartz> I sort of assumed id_LIKE should contain the IDs used by the upstream distros
[13:33:12] <eschwartz> ID=arch
[13:33:12] <eschwartz> ID_LIKE=archlinux
[13:33:14] <eschwartz> ohoh
[13:33:36] <bill-auger> yes i think those should be the same
[13:34:37] <bill-auger> but ive seen ID_LIKE=arch and also archlinux - it can be a list so parabola has "arch archlinux" just to be safe
[13:34:59] <bill-auger> i assume that is because it was never actually officially defined
[13:44:41] <eschwartz> We define it by using it, I guess
[13:48:42] -!- deep42thought has joined #archlinux32
[13:48:43] <buildmaster> Hi deep42thought!
[13:49:39] <deep42thought> eschwartz, thanks for the clarification
[14:06:59] <buildmaster> filesystem is broken (says nlopc43).
[14:07:11] <deep42thought> hmm, apparently, I did something wrong
[14:14:17] <eschwartz> Did you not responding filesystem with PKGBUILD overrides?
[14:14:36] <deep42thought> I forgot to update the checksum
[14:14:41] <eschwartz> *respin
[14:14:44] <eschwartz> Auto-incorrect
[14:15:03] <eschwartz> No I was wondering why this needed clarification
[14:15:21] <deep42thought> I didn't notice, that the file I'm looking for is a symling
[14:15:24] <deep42thought> *symlink
[14:15:47] <deep42thought> I assumed, it would origin from /usr/lib/* (as anything else under /etc -especially /etc/systemd does)
[14:15:49] <eschwartz> But surely you did already respin os-release
[14:16:00] <deep42thought> no
[14:16:08] <deep42thought> I just recently changed it
[14:16:27] <deep42thought> my main desktop box doesn't even _have_ this file
[14:16:33] <deep42thought> so I never bothered about its existence
[14:19:06] <eschwartz> Why patch 98% of a file...
[14:19:21] <deep42thought> because it makes the override a lot easier
[14:19:22] <buildmaster> filesystem is broken (says nlopc43).
[14:19:43] <deep42thought> plus: we'll notice when you change something and can adopt
[14:20:19] <deep42thought> "File os-release is not a regular file -- refusing to patch"
[14:20:25] <deep42thought> apparently it's not easier :-/
[14:20:41] <bill-auger> and BTW i remember the spec says that the symlink should be relative ../usr/lib/ to account for chroots
[14:23:05] <buildmaster> glibc is broken (says nlopc46).
[14:23:08] <bill-auger> if you want to override i think the idea it to use /etc/arch-release and /etc/os-release is still a symlink to that
[14:23:36] <deep42thought> /etc/os-release is not owned by any package
[14:23:56] <bill-auger> in the case of a fork /etc/arch-release would be something else like /etc/antegros-release and /etc/os-release is still a symlink to that
[14:26:03] <bill-auger> that kinda makes sense deep42thought - i would make it part of a branding package
[14:26:23] <deep42thought> why not simply override /usr/lib/os-release?
[14:26:36] <bill-auger> because that is the vendor supplied file
[14:26:41] <deep42thought> we're not really branding archlinux
[14:26:48] <deep42thought> we're porting to a different arch
[14:27:07] <deep42thought> alarm also provides their own /usr/lib/os-release, IIRC
[14:27:11] <bill-auger> no but the problem seems to be that forks like antegros are using the upstream filesystem package
[14:27:32] <bill-auger> which is defining it to be identical with the upstream
[14:28:02] <deep42thought> but they do fork other stuff from archlinux32, don't they?
[14:28:16] <bill-auger> so there should be some way to encourage the downstreams to supply their own version of os-release even if they use the upsream filesystem package
[14:28:46] <deep42thought> why do they use upstream filesystem, but e.g. our gcc (upstream's does not compile on i686)
[14:29:57] <bill-auger> that i dunno - i just stepped into the conversation because i did some research on the os-release spec
[14:30:45] <deep42thought> branding would make sense to me, if we were "only" _adding_ something to archlinux - but we're (re)compiling everything, so it's not really added packages, but rather recompiled packages, with occassionally added patched
[14:30:48] <deep42thought> *patches
[14:31:19] <bill-auger> but there should be some way to encourage the downstreams to supply their own version of os-release - becuase if upstream filesystem package includes os-release then the forks will identify themselves *as* archlinux and not as derivatives with the ID_LIKE field
[14:32:06] <deep42thought> well, they can override the package or replace the symlink or define it in /etc/arch-release
[14:32:22] <deep42thought> https://ptpb.pw
[14:32:37] <deep42thought> this is what I'd put into /usr/lib/os-release of archlinux32
[14:32:41] <bill-auger> yes /etc/os-release is intended to be overridden - thats why it probably should not be included
[14:33:31] <bill-auger> yea see you put ID_LIKE="arch archlinux" - thats what parabola has too
[14:33:48] <deep42thought> I did it, because you mentioned it
[14:33:50] <deep42thought> :-)
[14:33:55] <bill-auger> but i think we just concluded that is wrong - it should match exactly what upstream arch has for ID=
[14:34:10] <deep42thought> oh
[14:34:27] <bill-auger> that what ID means after all - it must be canonical
[14:35:26] <bill-auger> i think parabola put that just becuase they did not know what it should be - or becuase some distros arbitrarily chose one or the other
[14:35:51] <deep42thought> so "ID=archlinux32" and "ID_LIKE=archlinux", then?
[14:36:07] <deep42thought> https://ptpb.pw
[14:36:10] <bill-auger> i think eschwartz sais arch upstream is ID='arch'
[14:36:37] <deep42thought> on my box it's "ID=arch" "ID_LIKE=archlinux"
[14:36:37] <bill-auger> it should be whatever is ID= in the official arch
[14:36:39] <deep42thought> :-/
[14:37:04] <bill-auger> you have that from official arch ?
[14:37:12] <deep42thought> yes
[14:37:33] <bill-auger> i think that is wrong - someone should raise a policy issue about that
[14:37:53] <bill-auger> that would totally explain the confusion why it is unclear what forks should use
[14:38:08] <deep42thought> https://git.archlinux.org
[14:38:08] <phrik> Title:os-release\trunk - svntogit/packages.git - Git clone of the 'packages' repository (at git.archlinux.org)
[14:38:11] <bill-auger> im quite sure arch should not haver any ID_LIKE
[14:38:23] <bill-auger> because arch is not "like" any other
[14:38:53] <bill-auger> ID_LIKE was intyeneded for forks to indicate what they derrive their lineage from
[14:50:28] -!- bill-auger has quit [Remote host closed the connection]
[14:52:57] -!- bill-auger has joined #archlinux32
[15:05:47] <eschwartz> hmm, what does create the os-release in etc
[15:08:13] <deep42thought> mine is from 2014-07-14 - whereas other files on my system are from 2012-05-10 (namely /etc/localtime)
[15:08:29] <deep42thought> maybe some pacman hook created it some time?
[15:25:13] <eschwartz> maybe systemd created it automatically
[15:25:39] <deep42thought> was the change to systemd around 2014?
[15:27:02] <deep42thought> nope, that was earlier
[15:44:29] <deep42thought> hmm, the glibc build fails with a segfault in ld ... but I don't see any linking issues with objdump - I guess, we built the toolchain in the wrong order (again) :-(
[15:48:46] -!- deep42thought has quit [Quit: Leaving.]
[16:37:05] -!- oaken-source has quit [Ping timeout: 240 seconds]
[16:49:39] -!- abaumann has joined #archlinux32
[16:49:40] <buildmaster> Hi abaumann!
[16:50:15] <abaumann> Just to chime into the os-release discussion: HOME_URL, SUPPORT_URL, BUG_REPORT_URL point to upstream in Archlinux32's os-release (at least on my machine). :-)
[16:50:51] <abaumann> * abaumann sees 71 updates, updates and reboots his machine
[16:50:53] -!- abaumann has quit [Client Quit]
[17:46:16] -!- oaken-source has joined #archlinux32
[18:26:06] -!- bill-auger has quit [Remote host closed the connection]
[18:28:54] -!- bill-auger has joined #archlinux32
[18:45:35] -!- deep42thought has joined #archlinux32
[18:45:35] <buildmaster> Hi deep42thought!
[18:46:07] <deep42thought> abaumann: the new filesystem in testing should have the correct values :-)
[18:52:47] <deep42thought> buildmaster: what's up?
[18:52:47] <buildmaster> up? I'm up for 11 hours, 1 minute, load average: 0.29, 0.17, 0.18
[18:53:54] <deep42thought> btw, abaumann: do you have any idea what's wrong with glibc - or what was wrong the last time?
[18:53:59] <deep42thought> I can't remember :-/
[18:54:56] -!- abaumann has joined #archlinux32
[18:54:57] <buildmaster> Hi abaumann!
[18:55:00] <deep42thought> Hi abaumann
[18:55:11] <abaumann> deep42thought: cool. :-)
[18:55:12] <abaumann> hi.
[19:23:57] -!- deep42thought has quit [Quit: Leaving.]
[19:31:25] -!- abaumann has quit [Quit: leaving]
[20:28:00] -!- Polichronucci has quit [Ping timeout: 264 seconds]
[20:29:35] -!- Polichronucci has joined #archlinux32
[21:34:52] -!- Polichronucci has quit [Ping timeout: 250 seconds]
[21:38:32] -!- buildmaster has quit [Ping timeout: 256 seconds]
[22:07:04] -!- buildmaster has joined #archlinux32
[22:24:16] -!- buildmaster has quit [Ping timeout: 250 seconds]
[22:51:27] -!- buildmaster has joined #archlinux32
[22:51:53] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.stdin.2018-05-08T22:51:57.LNvYTo"?.
[23:04:51] -!- oaken-source has quit [Ping timeout: 255 seconds]