El 22/5/25 a las 14:27, Colin Watson escribió:
I tried building this on both unstable and testing, and it seems to build cleanly now. Based on the errors, I suspect that it may have been fixed by something like a change in eckit, rather than having been a problem in magics++. Could you please check whether it still fails for you?
Yes, doing a debbisect right now. This one looks similar to #1105066 in pytorch-cuda, which I downgraded today because it does not seem to happen anymore. (Fortunately the build time for magics++ is a lot shorter so in this case there is a real chance that we know the BD to blame). Thanks.