The thread started as a patch adding fetch.shallow, off by default. When enabled, a fetch or pull in a shallow repository that is not already scoped to specific refs fetches only the current branch's tracked upstream. The motivation is that a plain git pull in a shallow, sparsely checked-out clone of a repository with many branches can take minutes or hang.

Phillip Wood said the series he was reviewing has a diffstat of 12 files changed, 593 insertions and 33 deletions. That would take considerable time to review, yet the cover letter said only "Avoid fetching every branch of a new remote in a shallow repo." He could not tell how the patches relate to the discussion on separating which refs to fetch from where to store them. He said anyone who wants volunteers to spend time on review needs to make an effort in the cover letter, covering the design, trade-offs and discarded alternatives.

Junio C Hamano summarized the first three patches as he understood them. Patch 1 lets configuration stand in for git fetch --refmap=from:to remote refs... given on the command line. Patch 2 is a preparatory refactoring of a helper that gains a second call site. Patch 3 lets git fetch --refmap=from:to remote, without refspecs, default to fetching the refs that local branches build on. He said Phillip's comment was fair.

On a related point, Junio said he is not a heavy user of @{push}. If the chain of configuration that maps the current branch to its push destination always assumes remote.<remote>.fetch exists and does not consult remote.<remote>.fetchRefmap, then it could be called a bug introduced by patch 3.

Ben Knoble said he did not have the rest of the series fresh in mind, especially patches 4 and 5 on --limited-fetch. He asked whether the same problem would arise with the first three patches plus a remote that has a fetchRefmap and a limited or empty default fetch. If so, he said, it reads more like a bugfix for an earlier patch and might belong with it. He suggested reorganizing so the early part of the series is mostly about fetchRefmap. He also said he was unsure how overlapping fetch and fetchRefmap configuration should behave.

Harald Nordgren asked for "hopefully final" reviews. There is no sign yet of a restructured series or an expanded cover letter.