Workster | jaeger, one little thing if any one renames or copies any ports collections from say compat-32.rsync.inactive to compat-32.rsync, rejmerge will fix compat-32.rsync.inactive but what will happen is compat-32 and contrib will end up still pointing to 3.2 and the .inactive will point to 3.3 can we add a sed line to the helper update to look for crux 3.2 and change it to 3.3 for compat-32 and contrib | 00:10 |
---|---|---|
Workster | or at least make a note of this in the handbook? | 00:10 |
Workster | i've seen it happen with previous releases where one will be on the older contrib repo | 00:11 |
jaeger | https://crux.nu/gitweb/?p=system/iso.git;a=blob;f=iso/bin/setup;h=1434c42932d7672a01df4e129bad849ffe47bbf5;hb=HEAD#l311 | 01:23 |
jaeger | something like that? | 01:23 |
*** onodera has quit IRC | 01:44 | |
*** mavrick61 has quit IRC | 03:29 | |
*** mavrick61 has joined #crux-devel | 03:30 | |
Workster | jaeger, yes but also for existing /etc/ports/contrib.rsync /etc/ports/compat-32.rsync files | 04:31 |
Workster | sed -e 's|3.2|3.3|g' /etc/ports/contrib.rsync > $ROOT/var/lib/pkg/rejected/etc/ports/contrib.rsync | 04:32 |
jaeger | rejmerge should take care of that | 04:32 |
Workster | something like that for them 2 files that are not the .inactive | 04:32 |
Workster | say you already enabled contrib and compat-32 those two files wont get bumped to 3.3 | 04:33 |
*** isidore has joined #crux-devel | 04:33 | |
Workster | only the .inactive will | 04:33 |
jaeger | take a closer look at the link I pasted | 04:37 |
jaeger | After an upgrade, setup will have installed contrib.rsync.inactive and you'll already have contrib.rsync. If setup finds both of them it will then move contrib.rsync.inactive to /var/lib/pkg/reject/etc/ports/contrib.rsync | 04:38 |
jaeger | then when you run rejmerge (which you do, right?) it will prompt you to update | 04:38 |
*** isidore has quit IRC | 04:38 | |
Workster | oh... sorry yes that will cover it. | 04:39 |
jaeger | yeah, that was exactly the issue for which those lines were added a while back :) | 04:39 |
Workster | i always rejmerge | 04:39 |
Workster | but in the past it never did fix that | 04:40 |
Workster | not sure when that was added. | 04:40 |
jaeger | 3.1 | 04:40 |
Workster | probably was then when i noticed this. | 04:40 |
Workster | # one-time fix for .inactive ports files | 04:41 |
Workster | one time fix for active ports files maybe | 04:41 |
Workster | i didn't look at it closer. sorry about that. also been in and out of the office | 04:42 |
Workster | what is left on the todo list? | 04:42 |
Workster | https://crux.nu/bugs/index.php?do=details&task_id=923 only that by the looks ofit | 04:45 |
jaeger | all good | 04:47 |
Workster | i have a 7 node distcc cluster i'll start testing most ports on 3.3 | 05:20 |
Romster | =======> ERROR: Signature mismatch found: | 08:35 |
Romster | MISMATCHsignify | 08:35 |
Romster | on anything in contrib utill you manually copy the contrib.pub file to /etc/ports/ | 08:36 |
Romster | it could do with a public key not found error? | 08:36 |
jue | Romster: will add the contrib pubkey to core/ports later | 09:07 |
jue | FYI, I've postponed the source-rename feature for now cause I think the whole thing needs more time and energy | 09:23 |
jue | erverything I've seen so far is not really promising | 09:24 |
jue | objections? | 09:25 |
jue | have to run now, bbl | 09:25 |
Romster | none | 09:25 |
Romster | other than making a compat-32.pub file and to put that and the contrib.pub in /etc/ports | 09:26 |
Romster | and possibly improve the pkgmk signify error for missing pub file | 09:26 |
Romster | source-rename isn't critical as we don't set the PKGMK_SOURCE_DIR | 09:27 |
frinnst2 | woooot | 10:26 |
frinnst2 | just signed for SIM1U-3B+ | 10:26 |
frinnst2 | :D | 10:26 |
frinnst2 | time to buy a 3C | 10:28 |
frinnst2 | 3D asdf | 10:29 |
Romster | 3E | 10:30 |
Romster | xD | 10:31 |
Romster | teK__, mercurial has moved to https://www.mercurial-scm.org/ | 10:32 |
Romster | and is at version 4.0 | 10:32 |
frinnst2 | :> | 10:33 |
frinnst2 | kvm for crux.nu looks like a reality now | 10:33 |
Romster | though you were going to run crux.nu on bare metal | 10:34 |
Romster | s/though/thought | 10:34 |
frinnst2 | we are | 10:36 |
frinnst2 | but with supermicro ipmi | 10:37 |
frinnst2 | finding the right harware has been a bit tricky. we need a SIM1U-3B+ and SIM1U-3D | 10:38 |
frinnst2 | SIM1U-3B+ was a bitch to find | 10:38 |
Romster | you only need impi if no one is on site to attach a screen and keyboard to it. | 11:55 |
Romster | on the rare occasions | 11:55 |
Romster | and i have been reading in here about not being able to get the right card. | 11:56 |
*** onodera has joined #crux-devel | 12:26 | |
Romster | teK__, your clang port is missing libLLVMSupport.so or it's missing out of llvm as revdep reports that libLLVMSupport.so is missing for clang | 13:05 |
Romster | and libLLVMX86CodeGen.so and many more that prt-get fsearch doens't even find in llvm or clang | 13:09 |
Romster | i think it's because of the now dyna library that llvm uses that the shared libraries option in cmake for clang has now broken it. | 13:11 |
*** onodera has quit IRC | 13:13 | |
frinnst2 | ipmi is for kvm over ip too | 13:51 |
frinnst2 | hence the need for a SIM1U-3B+ | 13:51 |
frinnst2 | SIM1U-3B didnt do kvm over ip | 13:52 |
frinnst2 | anyways. I've also ordered a SIM1U-3D | 13:58 |
frinnst2 | total cost ~100€ | 13:58 |
frinnst2 | what remains are disks | 13:58 |
*** onodera has joined #crux-devel | 15:15 | |
jaeger | cool | 16:03 |
jaeger | random git alias I like: alias.lg=log --color --graph --pretty=format:'%C(auto)%h%Creset -%C(auto)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit | 22:33 |
pedja | nice.I like to pass 'lcs = !git rev-parse --short HEAD' to git_create_archive.Nice short SHA of the last commit as a part of tarball name. | 22:42 |
pedja | it's trivial, but handy | 22:42 |
jaeger | cool | 22:55 |
pedja | you are using virtualization at work and in the home lab, right?opinions/gotchas on using hugepages on the host? | 23:05 |
pedja | I've been reading RH's Virtualization_Tuning_and_Optimization_Guide, and it has some interesting tips/practices | 23:07 |
pedja | not many of them applicable on my machine, but :) | 23:08 |
pedja | the more I learn, the more I realize the depths of my ignorance. | 23:10 |
jaeger | I used hugepages when running a gaming VM with VFIO/GPU passthrough but in general I'm not sure if it matters much | 23:11 |
jaeger | My work and home lab are vmware vsphere | 23:12 |
pedja | any plans to move to HyperV?They give out free licences :) | 23:12 |
pedja | funny thing about the video on 2k16/hyperV I watched is the hype about 'virtual switch'.I was like 'wtf, dude ?!' | 23:15 |
pedja | and, for 10Gb network and NVme SSD, it runs preeety slowly, which is interesting. | 23:17 |
pedja | TechinalPreview, granted, but still | 23:18 |
pedja | Techical | 23:18 |
pedja | In the 'English as a second language' category, my country is 17th (of 75).So why the hell I can't spell properly?National disgrace I am. | 23:23 |
pedja | The first is, what a surprise, Sweden. | 23:26 |
pedja | But they have fierce winters, too, so it balances out. | 23:27 |
jaeger | I don't plan to switch. I like vmware and a personal hypervisor license is free. For a bigger lab VMUG advantage isn't free but it's cheap | 23:28 |
jaeger | http://maevr.vertigomusic.com/ <-- this is neat. | 23:29 |
pedja | cooool. | 23:31 |
pedja | I imagine that is how the acid trip looks like :) | 23:32 |
jaeger | heh | 23:34 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!