On 02/26/2018 05:14 PM, Luke Shumaker wrote: > On Mon, 26 Feb 2018 00:46:48 -0500, > Eli Schwartz wrote: >>> +++ b/db-functions >>> @@ -450,7 +450,7 @@ arch_repo_add() { >>> # package files might be relative to repo dir >>> pushd "${FTP_BASE}/${repo}/os/${arch}" >/dev/null >>> /usr/bin/repo-add -q "${repo}${DBEXT}" ${pkgs[@]} \ >>> - || error "repo-add ${repo}${DBEXT} ${pkgs[@]}" >>> + || error 'repo-add %q %s' "${repo}${DBEXT}" "${pkgs[*]@Q}" >>> popd >/dev/null >>> set_repo_permission "${repo}" "${arch}" >>> >>> @@ -468,7 +468,7 @@ arch_repo_remove() { >>> return 1 >>> fi >>> /usr/bin/repo-remove -q "${dbfile}" ${pkgs[@]} \ >>> - || error "repo-remove ${dbfile} ${pkgs[@]}" >>> + || error 'repo-remove %q %s' "$dbfile" "${pkgs[*]@Q}" >>> set_repo_permission "${repo}" "${arch}" >> >> I think for consistency we should use the same style which means using >> "${dbfile@Q}" > > I was going for consistency with the repo-add version, which doesn't > have a single dbfile variable to @Q. Would you have me introduce a > dbfile variable in arch_repo_add?
Well, we basically use it hardcoded three times, so why not. :p Just like the repo-add version, except with pushd "${dbfile%/*}" Actually, this decreases the difference between the two enough that we may want to just have one call the other... or do this: https://lists.archlinux.org/pipermail/arch-projects/2018-February/004832.html -- Eli Schwartz Bug Wrangler and Trusted User
signature.asc
Description: OpenPGP digital signature