On Thu, Apr 25, 2019 at 5:08 PM Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> wrote: > >> Solving (1) without (2) feels like a bit of a missed opportunity to > >> me. Ideally, what I would like is > >> > >> i. A central registry of trustworthy Git hooks that can be upgraded > >> using the system package manager to address (2). Perhaps just > >> git-hook-* commands on the $PATH. > >> > >> ii. Instead of putting hooks in .git/hooks, put a list of hooks to > >> run for each event in .git/config. > > > > The problem I had with this when discussing it was that our > > configuration system lacks a good way to control inheritance from outer > > files. I recently was working with a system-wide gitconfig file that > > referred to files I didn't have, and my Git installation was subtly > > broken in a variety of ways. > > > > If I have a system-wide hook to run for company code, but I have a > > checkout for my personal dotfiles on my machine where I don't want to > > run that hook, our configuration lacks a way for me to disable that > > system-wide configuration. However, using our current system, I can > > override core.hooksPath in this case and everything works fine. > > > > I mentioned this for completeness, and because I hope that some of those > > people will get some time to chime in here, but I think without that > > feature, we end up with a worse experience than we have now. > > I sent a proposal for this last year "How to undo previously set > configuration?": > https://public-inbox.org/git/874lkq11ug.fsf@xxxxxxxxxxxxxxxxxxx/ While reading that mail, it occurs to me that perhaps we can reuse the .gitignore idea. Instead of having a list of untracked files, we have a list of config keys. Instead of having .gitignore files associated to different directories to apply the rules to those dirs only, we have ignore rules that should apply on certain config files (probably based on path). A few differences from your reject/accept/priority example: - we don't redefine priority, inheritance rules apply the same way - reject/accept is handled the same way as positive/negative ignore rules. If we're lucky, we could even reuse the exclude code. - instead of special section names like [config "section"] we have something more like [config "/this/path"] # (or pattern) this lets us handle even other config files included by [include] or [includeIf] So, some examples [exclude] # exclude from all inherited files key = core.* # exclude core.* key = !core.bar # but keep core.bar [excludeIf "path:/etc/config"] # rules apply for only this file key = ... [excludeIf "glob:/home/*"] # rules apply for these config paths key = ... [excludeIf "system"] # special names for convenience maybe key = ... > Obviously the main bottleneck is someone like me working on patching it, Yes, manpower is always the problem. > but in this case it would be very useful if those who are interested in > this could look that proposal over and bikeshed it / point out issues I > may have missed, i.e. "no, this categorically won't work with this > proposed syntax due to XYZ you haven't thought of...". -- Duy
