*Re: [PRQ#80559] Deletion Request for codex-desktop-native* I am the maintainer of codex-desktop-native and I contest this deletion request.
The claim that my package duplicates codex-desktop-bin is technically incorrect. These are three architecturally distinct packages solving the same problem in fundamentally different ways: *codex-desktop-bin* distributes a pre-patched binary archive (codex-desktop-prepatched-...-x86_64.tar.gz) hosted on a third-party location. The patching happens outside the build process, before the user ever touches it. The user has no visibility into how that binary was produced. *openai-codex-desktop* is not a standalone desktop application at all — it is an Electron wrapper that depends on the openai-codex CLI package as a hard runtime dependency. It does not ship or build the desktop app itself. *codex-desktop-native* uses a fully automated, reproducible CI/CD pipeline: it downloads the official upstream macOS DMG, applies reverse-engineered compatibility patches, rebuilds native modules for Linux ABI, and produces a clean Linux-native binary. Every release includes checksums. The build process is fully auditable at https://github.com/mazixs/codex-desktop. The provides=('codex-desktop') and conflicts=('codex-desktop-bin') fields exist precisely to handle pacman conflict resolution at install time — this is standard practice and does not imply semantic duplication. Regarding the -bin suffix: I acknowledge this is a valid naming convention point and I am willing to rename the package to codex-desktop-native-bin to comply with AUR guidelines. This resolves the only legitimate technical objection in the request. I respectfully request that this deletion request be rejected or converted to a rename request. > > Четверг, 26 марта 2026, 04:45 +03:00 от [email protected]: > RubenKelevra [1] filed a deletion request for codex-desktop-native > [2]: > > Binary package without a "-bin" suffix. > > In addition, it's not fetching any official source, but a private > build binary stored in https://github.com/mazixs/codex-desktop > releases. > > In addition clearly a duplicate the author was aware of, as he/she > lists explicitly codex-desktop-bin as "provides" and "conflict". > > [1] https://aur.archlinux.org/account/RubenKelevra/ > [2] https://aur.archlinux.org/pkgbase/codex-desktop-native/ >
