Subject: dua-cli: CI failure due to toml/toml-edit transition in unstable Hi Luiz,
I wanted to give you a detailed explanation of the current CI failure we are seeing in the dua-cli package on Salsa. The build is failing at the dependency installation stage with the following conflict: librust-toml-dev (1.1.2-3) Breaks: librust-toml-edit-dev (< 1) The version librust-toml-dev 1.1.2-3 was recently uploaded to unstable and introduced a new Breaks declaration against librust-toml-edit-dev versions older than 1. However, the current version of librust-toml-edit-dev in unstable is still 0.25.11-2, which satisfies that broken constraint and therefore causes the conflict. The problem is further complicated by an indirect dependency chain coming from librust-ratatui-dev: ratatui → compact-str → borsh → borsh-derive → proc-macro-crate → toml-edit (< 0.26) This means that both librust-toml-dev and librust-ratatui-dev, which are direct build dependencies of dua-cli, end up pulling in conflicting versions of librust-toml-edit-dev, making it impossible for the dependency resolver in the clean Salsa CI chroot to satisfy all constraints simultaneously. I confirmed this locally by running apt-get update in my build chroot: Installed: librust-toml-dev 1.1.2-2 (no conflict) Candidate: librust-toml-dev 1.1.2-3 (introduces the conflict) This is precisely why the package builds successfully in my local environment but fails in the Salsa CI, which always starts from a clean chroot and fetches the latest packages from the archive at build time. This is a transitional issue in Debian unstable and is not caused by anything in our packaging. The conflict will be resolved once the Rust team updates librust-toml-edit-dev to version >= 1 in unstable, which will make it compatible with the new librust-toml-dev 1.1.2-3. I will monitor the situation at: https://tracker.debian.org/pkg/rust-toml-edit https://tracker.debian.org/pkg/rust-toml There is nothing we need to change in our package for now. I will let you know as soon as the conflict is resolved upstream and the CI passes again. Best regards, Nilson F. Silva
