| 28 Mar 2026 |
 colemickens | I'm sort of a riscv noob, especially with regards to the ... revisions? Is TH1520 too old to use to build in nixpkgs? | 22:54:04 |
| 29 Mar 2026 |
 Alex | That seems to be RV64GC (with an old(?) version of RVV?) so unless you're deliberately targeting a higher profile like RVA23, it should be OK. | 01:58:44 |
 dramforever | th1520 should be fine | 03:41:26 |
|  faukah left the room. | 18:11:58 |
| colemickens | I have no idea what I was doing last night. I rebooted my builders and went back to a newer nixos-unstable and now thing seem to be building mostly fine. Really bizarre, but I was up late, maybe I was more tired than I thought. | 18:44:16 |
| Alex | I've been stuck for a few days on a problem with linking clangd-fuzzer getting OoM killed right at the end of buildPhase 18 hours into the clang build. Even -j1 isn't enough.
I tried useMoldLinker, but that's broken right now (due to something not putting NIX_CFLAGS_LINK under env?).
Then I tried updating the system again, but now useMoldLinker causes eval to overflow the stack.
Has anyone found a reasonable solution to building clang (LLVM 20 and 21) without needing ludicrous amounts of memory?
If push comes to shove I could perhaps suffer a larger swap file... | 18:51:51 |
| Alex | For reference:
[nixos@nixos:~/system]$ free -h
total used free shared buff/cache available
Mem: 7.7Gi 6.0Gi 105Mi 0B 1.6Gi 1.5Gi
Swap: 15Gi 152Mi 15Gi
(There is high memory usage right now because it's in the middle of retrying clang.) | 18:53:30 |
| 30 Mar 2026 |
 JamieMagee | I found an issue with nixpkgs configuration of gdb, which doesn't match upstream configuration. Specifically, gdb doesn't support in-process agent on risc-v: https://github.com/NixOS/nixpkgs/pull/504872 | 03:41:58 |
 misuzu | Try --cores 1 | 07:14:37 |
| Alex | I had a build almost ending already, and this is what the RSS looked like just within 10s of the OoM kill:
436704 bin/c-index-test
4203668 bin/clang-tidy
2798596 bin/clangd-fuzzer
So yeah I think --cores 1 is worth a try, though it'll probably take over 2 days to build a single clang version (and I need 2). 🫠| 10:16:04 |
| Alex | I have figured out a way to get LLVM linked with mold to evaluate (I think?) so I think I'll try that first. | 10:17:01 |
