| 23 Nov 2025 |
K900 | But has some nice fixes and I don't want to build it twice for 25.11 and unstable | 15:53:12 |
hexa | I also still have a protobuf situation, I think 🫠 | 15:55:12 |
hexa | the squashing of protobuf versions didn't happen after all as I hoped | 15:55:39 |
Vladimír Čunát |
We plan to do branchoff tomorrow or Tuesday evening (CET).
| 15:55:46 |
hexa | * the squashing of protobuf versions didn't happen after all as I hoped it would | 15:55:50 |
hexa | is staging-25.05 next? | 15:56:34 |
Vladimír Čunát | If we need something for the 25.11 release, I'd say that takes priority. | 15:57:30 |
Vladimír Čunát | But yes, otherwise I thought of 25.05. | 15:57:39 |
K900 | Protobuf will be a significant rebuild | 15:58:01 |
K900 | And will drag most of Qt along | 15:58:06 |
K900 | So if we're doing that I'd like to have Qt as well | 15:58:12 |
Vladimír Čunát | python3.pkgs.protobuf might suffice, but that would still not be a minor rebuild. | 16:01:03 |
Vladimír Čunát | * python3.pkgs.protobuf might suffice maybe, but that would still not be a minor rebuild. | 16:01:07 |
hexa | maybe I can just revert system protobuf in openvino for now, then it will be small-ish | 16:01:54 |
hexa | still bisecting | 16:02:09 |
hexa | e1ece2c5218fa3141ee7bef86f16710640c10d4c is the first bad commit
commit e1ece2c5218fa3141ee7bef86f16710640c10d4c (HEAD)
Author: Martin Weinelt <hexa@darmstadt.ccc.de>
Date: Mon Nov 17 20:31:00 2025 +0100
openvino: use system protobuf
pkgs/by-name/op/openvino/package.nix | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
| 16:12:34 |
hexa | 👏 | 16:12:43 |
Vladimír Čunát | No rebuild?
Rebuild: linux 4, darwin 0
| 16:13:33 |
hexa | little to none, yeah | 16:16:17 |
hexa | will confirm this is sufficient now | 16:16:27 |
hexa | I still think the cause it that this mixes protobuf versions | 16:16:44 |
hexa | still seeing protobuf_30 via apache-orc | 16:17:18 |
hexa | and protobuf_33 via opencv | 16:17:27 |
hexa | and protobuf_31 via arrow-cpp | 16:17:45 |
hexa | and system protobuf is 32 | 16:17:55 |
hexa | * and default protobuf is 32 | 16:18:04 |
hexa | the upcoming staging-next will start recursing into python 3.14 and stop recursing into 3.12 | 16:41:55 |
Vladimír Čunát | So, feel free to suggest what's Hydra's next mass-rebuild task. I expect it will run out of work in 12-24 hours. | 18:34:07 |
K900 | It would be nice to have Qt 6.10.1 after all | 18:39:02 |
K900 | I'm half a qtwebengine and a migraine away from validating it all locally | 18:39:14 |