| 23 Jun 2026 |
 Vladimír Čunát | The error isn't shown anymore. | 17:36:42 |
| Vladimír Čunát | (The UI only shows error from the last attempt.) | 17:36:54 |
 Mic92 | There might be race condition in Nix creating the fetcher cache. Queue runner shouldn't have this. | 17:37:59 |
| Mic92 | https://hydra.nixos.org/build/332691361/nixlog/13 | 17:38:57 |
| Mic92 | This looks like a substitution error? | 17:39:14 |
 K900 | Restarting does not seem to help | 17:43:21 |
| Mic92 | Is it actually rebuilding indirect steps if you do this on the web? | 17:45:03 |
| K900 | It should | 17:45:34 |
| Mic92 | As in it did in the past? | 17:46:43 |
| K900 | Yes | 17:46:50 |
| Mic92 | Or it would be desirable | 17:46:53 |
| K900 | Also yes | 17:46:58 |
| K900 | (though ideally we would not have the concept of indirect steps, but that's way more work) | 17:47:14 |
| Mic92 | It might be actually more explicit now which is why restart now restarts less. But don't have this code mapped out in my head | 18:00:02 |
| K900 | https://github.com/NixOS/nixpkgs/pull/534682 | 18:01:20 |
| K900 | For the other thing | 18:01:30 |
|  citerism joined the room. | 18:18:18 |
|  Eli Saado joined the room. | 18:39:55 |
|  isabel changed their profile picture. | 19:54:31 |
| Mic92 | tcl-x86_64-w64-mingw32> In file included from /build/tcl8.6.16/generic/tclInt.h:36,
tcl-x86_64-w64-mingw32> from /build/tcl8.6.16/generic/tclIO.c:15:
tcl-x86_64-w64-mingw32> /build/tcl8.6.16/generic/tclPort.h:21:13: fatal error: tclWinPort.h: No such file or directory
tcl-x86_64-w64-mingw32> 21 | # include "tclWinPort.h"
tcl-x86_64-w64-mingw32> | ^~~~~~~~~~~~~~
tcl-x86_64-w64-mingw32> compilation terminated.
tcl-x86_64-w64-mingw32> make: *** [Makefile:1157: tclIndexObj.o] Error 1
tcl-x86_64-w64-mingw32> make: *** [Makefile:1154: tclHistory.o] Error 1
tcl-x86_64-w64-mingw32> make: *** [Makefile:1163: tclIO.o] Error 1
| 20:17:09 |
| Mic92 | I can reproduce this build error also locally | 20:17:17 |
| Mic92 | So this one is not a GC race. Would be also really weird to have the exact same one over and over again | 20:17:36 |
| Mic92 | This race condition is actually fixed in nix-eval-jobs master: https://github.com/NixOS/nix-eval-jobs/pull/402 | 20:25:56 |
 Sergei Zimmerman (xokdvium) | In reply to @joerg:thalheim.io
tcl-x86_64-w64-mingw32> In file included from /build/tcl8.6.16/generic/tclInt.h:36,
tcl-x86_64-w64-mingw32> from /build/tcl8.6.16/generic/tclIO.c:15:
tcl-x86_64-w64-mingw32> /build/tcl8.6.16/generic/tclPort.h:21:13: fatal error: tclWinPort.h: No such file or directory
tcl-x86_64-w64-mingw32> 21 | # include "tclWinPort.h"
tcl-x86_64-w64-mingw32> | ^~~~~~~~~~~~~~
tcl-x86_64-w64-mingw32> compilation terminated.
tcl-x86_64-w64-mingw32> make: *** [Makefile:1157: tclIndexObj.o] Error 1
tcl-x86_64-w64-mingw32> make: *** [Makefile:1154: tclHistory.o] Error 1
tcl-x86_64-w64-mingw32> make: *** [Makefile:1163: tclIO.o] Error 1
That’s not the error I was talking about | 20:31:31 |
| Sergei Zimmerman (xokdvium) | There were the other ones that seem to have succeeded after I restarted those | 20:31:50 |
| Sergei Zimmerman (xokdvium) | In reply to @joerg:thalheim.io
This race condition is actually fixed in nix-eval-jobs master: https://github.com/NixOS/nix-eval-jobs/pull/402 Hm that PR is about the local store, not the fetcher cache db | 20:34:32 |
| Sergei Zimmerman (xokdvium) | Either way looks like a nix/nej issue | 20:34:41 |
| Mic92 | Redacted or Malformed Event | 20:38:18 |
| Mic92 | Redacted or Malformed Event | 20:38:29 |
| Mic92 | I think I can just fix this the same way as the other PR | 20:41:18 |
