| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Add a new function `magit-which-function' as a replacement
for the unreliable `which-function' function and use it as
the default of `magit-log-trace-definition-function'.
|
|
Re #3674.
|
|
`magit-rev-eq' used to call `magit-rev-verify' on each argument, which
returns the first line of `git rev-parse --verify'. In the case of a
range A..B, the standard output spans multiple lines, with first line
being B's hash and the second being "^A-hash". It also return with
exit status 128, but `magit-rev-verify' ignores that.
`magit-rev-eq' should only return t if A and B refer to the same
commit and a range never refers to any single commit, so we should
return nil in that case. Do that by using `magit-commit-p' instead.
This fixes a regression in `magit-diff-type', which (beginning with
296f8dc7) falsely concluded that a diff buffer showed unstaged changes
when diffing a range where the right side resolves to the same commit
as HEAD.
|
|
I am undecided whether to enable this by default and if so what style
to default to. Enable it for now on the development branch, so that
users notice and get a chance to comment on #3679 before the next
release.
|
|
|
|
In the rest of Magit we call [those] things "keywords". Do so here
too. Additionally we should avoid the term "note" because it means
something else in the context of editing text using `git-commit'.
A note is an additional text that is associated to a commit (similar
to a commit message, but unlike that not actually part of the commit
object) so the same word should not be used for a certain kind of
substring of a commit message.
|
|
Because many users put "~/.emacs.d/" under version-control but don't
customize `magit-repository-directory', that directory should not be
a member of `magit-repository-directories'. Otherwise `magit-status'
would not fall through to either offering to initialize the current
directory as a repository (`magit-init' would have to be used) or to
read a repository using regular file-name completion (except with two
prefix arguments), and users have grown accustomed to these fallback
behaviors.
Also update the documentation about `magit-repository-directories'
and `magit-status'. The documentation about the latter was rather
ambiguous until now. We trade that for making it a bit scary instead.
This reverts commit f77d1158b01923ca11d065f12aa0e69d6ce2415b, and
then also improves the documentation.
This fixes #3673.
|
|
In the log-select buffer shown by `magit-commit-instant-squash' and
`magit-commit-instant-fixup' point was no longer placed on the same
commit as was at point in the buffer from which the command was
invoked.
Closes #3674.
|
|
The doc-string of this command says nothing about performing other
cleanup, so we shouldn't do it. Users who want to drop empty commits
can use `--autosquash' (which is available in the popup) together with
`magit-rebase-interactive'.
Re #3670.
|
|
Closes #3670.
|
|
Previously it was possible to limit to changes shown in the status
buffer, but that was only a kludge because at the time this feature
was added it wasn't possible to easily get the diff type, what we
always wanted was a way to see whitespace errors while we can still
fix them but not when the errors have already been committed.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The value has to be a variable whose value has to be a keymap.
This enables packages other than `magit' and `forge' to use
keymap variables whose names use the packages symbol prefix.
|
|
When another window is already displaying the file buffer, then
this function did additionally display it in the current window.
Fixes #3662.
|
|
Closes #3659.
|
|
Use git-when-merged to help users ask "When was this commit merged
into this ref?" and "Assuming this commit was developed in a topic
branch, what were the other commits in that topic?".
|
|
|
|
Fixes #3654.
|
|
|
|
|
|
|
|
|
|
|
|
The old one is renamed magit-describe-section{=>-briefly}.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
