Derrick Stolee <stolee@xxxxxxxxx> writes: > On 12/9/2020 7:13 PM, Junio C Hamano wrote: >> Josh Steadmon <steadmon@xxxxxxxxxx> writes: >> >>> # start registers the repo >>> - git config --get --global maintenance.repo "$(pwd)" && >>> + git config --get --global --fixed-value maintenance.repo "$(pwd)" && >> >> The rewrite makes it better than the original, but I wonder why the >> original did not do a more obvious > > maintenance.repo is a multi-valued config setting, so it is possible > that there are multiple existing values. Hence the reason for needing > the value filter. I do not quite get it. You mean as long as $(pwd) appears, you do not care what other value appear on the variable? Aren't we control of what repositories have been registered to the system at this point in the test sequence? It's not wrong per-se to use "does this value exist for the key?", especially with the --fixed-value option. It somehow just felt a bit unusual to me. In any case, thanks for the metacharacter fix. That is now on 'master' so the previous breakages are all gone with Josh's fix. Thanks, both.
