Paul Eggert writes: > On 2025-02-15 02:42, Kévin Le Gouguec wrote: >> waiting for vc-dir-busy*after* invoking vc-dir seems to work fine, >> which makes complete sense > > Makes sense to me too. I didn't look at the patch in detail, but did notice a "TODO" comment. One of these "finishing touches" I mentioned, yup. Attaching for your consideration: 1. a patch that fixes the vc-dir synchronization, 2. a patch that ditches vc-dir entirely. Either fixes the issue, as far as I can validate. I still have a slight preference for the second one¹. FWIW, here the test takes ≈330ms with patch 1, and ≈68ms with patch 2². Happy to install whichever. ¹ tl;dr vc-dir is a middleman with no added value for the purposes of this test. ² Which is only worth consideration if you happen to need run that test repeatedly to reproduce a weird race condition 🥲