Jonas Bernoulli <jo...@bernoul.li> writes: > > Which means that setting `magit-save-repository-buffers' to `nil', > *does* break `magit-blame'. I'll fix that soon. Patrick: I'll let you > know when I am done.
I have changed the behavior as follows: `magit-blame' now always saves the buffer which visits the file to be blamed before actually doing so. This is absolutely necessary, because `git blame' prints line numbers which we match against line numbers in the buffer. But if the line numbers in the buffer don't match the line numbers in the working tree file seen by Git, then they obviously won't match. But I also noticed that before this bugfix, `magit-blame' did *never* save any buffer or offer to save any buffer, regardless of the value of `magit-save-repository-buffer', because it binds `inhibit-magit-refresh' to `t` around the calls to `magit-run-git-async' which is what is used to run `git blame'. Since that doesn't match your experience, I suspect that you are using aa very old version, probably some v1.* release. Magit is currently at v2.2.2, and v2.1 was a major update. Anyway, `magit-blame' (from the `master' branch, and soon in v2.3.0) saves the appropriate buffer unconditionally, but never any other buffer. You might still not like being asked whether you want to safe buffers by *other* commands, in which case you should set `magit-save-repository-buffers' to `nil', but that won't affect the behavior of `magit-blame'. -- Jonas -- You received this message because you are subscribed to the Google Groups "magit" group. To unsubscribe from this group and stop receiving emails from it, send an email to magit+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
