Delilah Ashley Wu's series fixes an inconsistency. When both $HOME/.gitconfig and $XDG_CONFIG_HOME/git/config exist, 'git config list --global' and 'git config get --global' read only the home file. The unscoped variants read both. Wu writes that this contradicts the documented behavior and has no good justification, and that bug reporters expected both files to be read.
The patches route the global scope through the config sequence code. The earlier RFC had a helper that counted successful reads of global files. Wu said that overcomplicated the error handling. Her revision has do_git_config_sequence() optionally fail when neither global file can be read, tracked with a simple boolean set when one is read successfully.
A preparatory patch adds tests recording that 'git config list' succeeds when no global configuration exists whereas 'git config list --global' fails. They check only the exit code. This guards against regressions in the following patch.
In a new cover letter Wu said she thinks she has addressed all the feedback from Junio's review. The thread has about 30 messages across versions. The excerpts do not show whether the series has been accepted.