On 07.10.2020 16:16, Andrii Kolomoiets wrote: > Dmitry Gutov writes: > >> On 05.10.2020 09:02, Andrii Kolomoiets wrote: >>> Dmitry Gutov writes: >>> >>>> Where is that vc-backend called from, in our scenario? >>> It's called from 'vc-call'. The 'vc-revert-file' used it twice >>> to call the 'make-version-backups-p' and the 'revert' backend functions. >> >> Then a change in vc-revert-file could be sufficient. > > Can you please advice me what this change should look like? Get rid of > calling 'vc-call'? Yes. How about the attached patch? > In this case the function 'vc-version-backup-file' > must be changed as well. Does it actually make sense to use it on a directory? >>>> Could we make do with changing that code to use vc-responsible-backend >>>> instead of vc-backend instead? If it's not a function called >>>> frequently. >>> I went a little different way and made the 'vc-backend' return >>> correct >>> backend for directories. In case you missed it somehow here is the link >>> to the message: >>> https://lists.gnu.org/archive/html/bug-gnu-emacs/2020-09/msg02508.html >> >> Like I explained, it will create a cache entry that is never invalidated. > > In what way VC backend for directory could be changed? Like 'rm -rf .hg > && git init'? We can make the 'vc-backend' function to ignore cached > backend for directories. Though I think it's not efficient. Something like that. Or 'git init' inside a subdirectory. Not a frequent occurrence, but if we start using directories and files interchangeably in more places, we are likely to start caching other properties on them, too. So it's better to use a different function to detect which backend a directory belongs to. Also, your patch makes vc-registered work on directories. What does it mean to have a directory "registered"? It's not a well-defined notion, given that most of contemporary VC systems don't track directories, only the files inside them.