| 21 Jun 2025 |
 fgaz | it does pre-eval on the first search you perform. but even that first search shouldn't require all that memory, see the post I linked above | 20:51:03 |
| fgaz | or did you mean a downloadable db? | 20:51:39 |
 Alex | It could be downloaded.
That's just an idea of one way of implementing it. There are probably others.
And yeah, I can imagine how it could be possible to trim a lot of the memory usage if you only evaluate the metadata you need and not the whole derivation graph. | 21:05:24 |
 GenericNerdyUsername | like nix-index? | 21:15:22 |
| 22 Jun 2025 |
|  SomeoneSerge (hash-versioned python modules when) changed their display name from SomeoneSerge (UTC+U[-12,12]) to SomeoneSerge (Ever OOMed by Element). | 12:15:03 |
|  @steeringwheelrules:tchncs.de joined the room. | 13:27:25 |
 Tristan Ross | After a month, my VisionFive2 has finished building a graphical version of my config. It has COSMIC now and ssh should work. | 15:01:19 |
| Tristan Ross | It also went from 25.05 to 25.11 | 15:05:27 |
 dramforever | MONTH | 16:53:27 |
| Tristan Ross | In reply to @dramforever:matrix.org
MONTH Yes, it really took that long. It finished some time last night since it only had 30 derivations to finish. | 17:05:48 |
| 23 Jun 2025 |
| Tristan Ross | Has anyone gotten graphics to work on the VisionFive2? I'm using the nixos-hardware module. | 05:33:20 |
| dramforever | it doesn't work, the HDMI driver is not in mainline | 10:40:06 |
| dramforever | (yes, you've been using mainline linux on your vf2 this whole time) | 10:40:33 |
| Tristan Ross | In reply to @dramforever:matrix.org
it doesn't work, the HDMI driver is not in mainline What about the fork? Would that work? | 12:30:01 |
| Alex | In reply to @rosscomputerguy:matrix.org
What about the fork? Would that work? If you mean the vendor's branch, in my experience no.
(The "upstream" branch runs into a kernel panic on boot when HDMI is connected. The "dev" branch is a nightmare to compile.) | 13:23:44 |
| Tristan Ross | In reply to @alex:tunstall.xyz
If you mean the vendor's branch, in my experience no.
(The "upstream" branch runs into a kernel panic on boot when HDMI is connected. The "dev" branch is a nightmare to compile.) Oh, that's cursed | 13:50:10 |
| 25 Jun 2025 |
|  craigbro joined the room. | 02:14:57 |
| craigbro | greetings. I just install nix using the project installer on my orangepi rv2 running ubuntu. | 02:22:17 |
| craigbro | * greetings. I just installed nix using the project installer on my orangepi rv2 running ubuntu. | 02:22:25 |
| craigbro | thanks for all the work to get it working this far | 02:22:47 |
| dramforever | just have to go full nixos now | 02:23:19 |
| dramforever | (i don't know if there's anything ready for that) | 02:23:32 |
| craigbro | craig@orangepirv2:/etc/nix$ nix shell nixpkgs#cowsay -c cowsay "It Works!"
< It Works! >
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
| 02:26:47 |
| craigbro | so the binary caches listed on the wiki page run just fine. However, until I added them, I consistenty got a segfault patching libpcre while building bootstrap-tools (on every branch/tag/release I tried) | 02:27:44 |
| dramforever | whwat does cat /proc/cpuinfo say | 02:28:22 |
| dramforever | just, the first core | 02:28:30 |
| dramforever | * what does cat /proc/cpuinfo say | 02:28:37 |
| dramforever | it should be basically identical for the rest | 02:28:47 |
| craigbro | processor : 0
hart : 0
model name : Ky(R) X1
isa : rv64imafdcv_zicbom_zicboz_zicntr_zicond_zicsr_zifencei_zihintpause_zihpm_zfh_zfhmin_zca_zcd_zba_zbb_zbc_zbs_zkt_zve32f_zve32x_zve64d_zve64f_zve64x_zv
fh_zvfhmin_zvkt_sscofpmf_sstc_svinval_svnapot_svpbmt
mmu : sv39
uarch : ky,x60
mvendorid : 0x710
marchid : 0x8000000058000001
mimpid : 0x1000000049772200
| 02:28:59 |
| dramforever | hmm, not the pcre bug i was thinking then | 02:29:21 |
