!xmLtiCaAJxfhURjrXl:matrix.org

NixOS RISC-V

241 Members
NixOS on RISC-V https://wiki.nixos.org/wiki/RISC-V https://pad.lassul.us/NixOS-riscv64-linux 74 Servers

Load older messages

SenderMessageTime
18 Feb 2025
@rosscomputerguy:matrix.orgTristan Ross
In reply to @leons:is.currently.online
I've been tuned out of this for a long time, but I recall there being some discussions about trying to pool together some native riscv64 builders in the past, did that lead anywhere?
This issue seems to have gone dry some time ago: https://github.com/nix-community/infra/issues/715
Scrolling up Tristan Ross you seem to be doing native riscv64 compiles, is that using actual hardware or QEMU?
I do real hardware 		23:05:24
@leons:is.currently.onlineLeonOh wow, that's cool. Is that a small SBC or some larger system? Do you have some more details on your setup somewhere?23:06:12
@rosscomputerguy:matrix.orgTristan Ross VisionFive2 23:06:26
@rosscomputerguy:matrix.orgTristan Ross I'm still trying to get NixOS installed on it, the initrd can't find my NVME 23:06:49
@leons:is.currently.onlineLeonGotcha. Yeah, I have one of those, 2 BeagleV Ahead, a couple smaller Milk-V targets and just got the Framework RISC-V Mainboard. So it's not much, but was wondering whether I can throw these boards into a corner in my Uni's lab and have them do something more useful than collect dust until I or someone else finds out how to get NixOS running on the Framework mainboard haha23:08:31
@rosscomputerguy:matrix.orgTristan RossYeah, most systems use the JH7110 which isn't that powerful lol23:08:54
@leons:is.currently.onlineLeonRight right.23:09:01
@rosscomputerguy:matrix.orgTristan RossThere's the newer P550 but the price23:09:04
@leons:is.currently.onlineLeonFWIW, I'm running the pre-built Fedora 41 image on the Framework mainboard right now, and running GNOME + Firefox, listening to music, etc. is somewhat usable. Definitely better than what I remember from my VF223:10:41
@leons:is.currently.onlineLeon(I don't know how much frequency scaling this thing does, could thermals have an impact?)23:11:08
@rosscomputerguy:matrix.orgTristan Ross That silicon is probably more mature than the one I've got 23:11:19
@rosscomputerguy:matrix.orgTristan RossThe VF2 is older so I wouldn't doubt the Framework being more mature23:11:36
@rosscomputerguy:matrix.orgTristan Ross I've ran into some oddities in my VF2 like NIC DMA not working 23:11:59
@leons:is.currently.onlineLeonWell they both use the JH7110 SoC, so shouldn't be much of a difference I guess.23:12:27
@rosscomputerguy:matrix.orgTristan RossYeah but the rest of the board matters too23:12:45
@rosscomputerguy:matrix.orgTristan Ross And it's possible there's different revisions 23:12:55
@leons:is.currently.onlineLeonBut it sounds like there's not yet a Hydra or something I could hook this up to?23:13:49
@rosscomputerguy:matrix.orgTristan Ross Not currently 23:13:56
@rosscomputerguy:matrix.orgTristan Ross@matthewcroughan:defenestrate has floated the idea23:14:35
@leons:is.currently.onlineLeonCool, excited to maybe build something.23:16:19
19 Feb 2025
@0x4a6f:matrix.org[0x4A6F]I got a Milk-V Megrez waiting for me to have time working on it...00:01:19
21 Feb 2025
@rosscomputerguy:matrix.orgTristan Ross 
procmail-riscv64-linux> Whoeaaa!  This actually can't happen.
procmail-riscv64-linux> You have a look and see if you detect anything uncanny:
procmail-riscv64-linux> *******************************************************
procmail-riscv64-linux> _autotst.c: In function 'main':
procmail-riscv64-linux> _autotst.c:110:9: warning: ignoring return value of 'setgid' declared with attribute 'warn_unused_result' [-Wunused-result]
procmail-riscv64-linux>   110 |       { setgid(2);
procmail-riscv64-linux>       |         ^~~~~~~~~
procmail-riscv64-linux> _autotst.c:114:12: warning: ignoring return value of 'setuid' declared with attribute 'warn_unused_result' [-Wunused-result]
procmail-riscv64-linux>   114 |          { setuid(2);
procmail-riscv64-linux>       |            ^~~~~~~~~
procmail-riscv64-linux> _autotst.c:127:12: warning: ignoring return value of 'setegid' declared with attribute 'warn_unused_result' [-Wunused-result]
procmail-riscv64-linux>   127 |          { setegid(groups[i]);
procmail-riscv64-linux>       |            ^~~~~~~~~~~~~~~~~~
procmail-riscv64-linux> _autotst.c:145:6: warning: ignoring return value of 'chown' declared with attribute 'warn_unused_result' [-Wunused-result]
procmail-riscv64-linux>   145 |      chown(argv[1],2,getegid());setuid(2);     /* adjust the directory first */
procmail-riscv64-linux>       |      ^~~~~~~~~~~~~~~~~~~~~~~~~~
procmail-riscv64-linux> _autotst.c:145:33: warning: ignoring return value of 'setuid' declared with attribute 'warn_unused_result' [-Wunused-result]
procmail-riscv64-linux>   145 |      chown(argv[1],2,getegid());setuid(2);     /* adjust the directory first */
procmail-riscv64-linux>       |                                 ^~~~~~~~~
procmail-riscv64-linux> *******************************************************
procmail-riscv64-linux> I suggest you take a look at the definition of LDFLAGS*
procmail-riscv64-linux> in the Makefile before you try make again.
procmail-riscv64-linux> make[1]: *** [Makefile:124: ../autoconf.h] Error 1
procmail-riscv64-linux> make[1]: Leaving directory '/build/procmail-3.24/src'
procmail-riscv64-linux> make: *** [Makefile:155: autoconf.h] Error 2

procmail fails to build in a funny way

17:49:05
@rosscomputerguy:matrix.orgTristan RossLooks to be the same as https://github.com/NixOS/nixpkgs/issues/33491617:58:10
26 Feb 2025
@pouspous:matrix.orgpouspous removed their display name pouspous.07:36:18
@pouspous:matrix.orgpouspous left the room.07:36:21
@niclasoverby:beeper.com@niclasoverby:beeper.com joined the room.10:09:09
6 Mar 2025
@jellyterra:matrix.org@jellyterra:matrix.org changed their profile picture.13:04:46
7 Mar 2025
@SeTSeR:matrix.orgWhite Whale Hello! How can I override Python package `watchdog` from within overlay? I tried this:
```nix
python3Packages = super.python3Packages.overrideScope (self: super: {
  watchdog = super.watchdog.overrideAttrs (self: super: {
    doCheck = false;
  }
 });
```
However, `nix why-depends` still shows that dependency on old `watchdog` is still in place. It shows that `werkzeug` depends on this library. In Nixpkgs I found, that `werkzeug` has `watchdog` in `nativeBuildInputs`. However, neither putting `watchdog` override in `buildPackages` nor removing the library from `nativeBuildInputs` didn’t do a trick. What else could I try? 		19:05:07
@alex:tunstall.xyzAlex
In reply to @SeTSeR:matrix.org
Hello! How can I override Python package `watchdog` from within overlay? I tried this:
```nix
python3Packages = super.python3Packages.overrideScope (self: super: {
  watchdog = super.watchdog.overrideAttrs (self: super: {
    doCheck = false;
  }
 });
```
However, `nix why-depends` still shows that dependency on old `watchdog` is still in place. It shows that `werkzeug` depends on this library. In Nixpkgs I found, that `werkzeug` has `watchdog` in `nativeBuildInputs`. However, neither putting `watchdog` override in `buildPackages` nor removing the library from `nativeBuildInputs` didn’t do a trick. What else could I try?
It might be coming from one of the other Python package sets. Maybe look at the definition of python3Packages in Nixpkgs, because I suspect that it's actually an alias for one of the other sets. 		20:48:14
@SeTSeR:matrix.orgWhite Whale
In reply to @alex:tunstall.xyz
It might be coming from one of the other Python package sets. Maybe look at the definition of python3Packages in Nixpkgs, because I suspect that it's actually an alias for one of the other sets.
It’s strange, because I had to disable tests for some other packages and such override did the trick for them. But now I see the problem more clearly: werkzeug sees the updated watchdog, but for some reason this updated watchdog still runs tests. 		21:28:52

Show newer messages

Back to Room ListRoom Version: 10