From: Stefan Sperling Subject: Re: merge chokes, creates bogus conflicts To: Mark Jamsek Cc: Christian Weisgerber , gameoftrees@openbsd.org Date: Thu, 16 Feb 2023 13:27:16 +0100 On Thu, Feb 16, 2023 at 11:14:54PM +1100, Mark Jamsek wrote: > On 23-02-16 12:45PM, Stefan Sperling wrote: > > On Thu, Feb 16, 2023 at 02:11:54AM +0100, Stefan Sperling wrote: > > > If we want to allow users to commit such files in spite of the > > > obvious downsides, here is a patch which would make it possible: > > I agree that, while such conflict markers should not be embedded > verbatim into versioned files, users should be able to override this > check and commit unresolved conflicts for such cases as demonstrated by > this thread. IIRC, git just lets you commit conflicts, but we don't in > Fossil, so a flag is provided to allow committing with unresolved > conflicts. However, I might even suggest we only allow such commits > interactively, by prompting the user when the -C flag is used and > conflicts indeed do exist in the commitables to confirm they want to > commit despite the unresolved conflicts. > > In any case, I'll have a go at implementing your below proposal. Great, thanks! Please feel free to use the diff I sent as a starting point. > It'll > be trivial to add a confirmation prompt after that's done if we think an > extra layer is needed. It' just a lot easier to resolve conflicts as > they arise rather than sometime later and perhaps even by someone other > than the conflict creator. I agree that such a prompt would be nice to have. And yes, this should be added afterwards because this will create churn as there would need to be a new callback which gets passed around and implements the prompt. There should again be a way to disable the interactive check (such as via -N) because there might be cases where merges get scripted and may need to deal with such weird files.