| Age | Commit message (Collapse) | Author |
|---|
|
|
|
magit-diff-visit-worktree-file and magit-diff-visit-file
This allows us to get the old behavior of RET before the switch to defaulting to
visiting the diff but still allow consisten behavior.
Set this:
+(defcustom evil-collection-magit-visit-worktree-file-on-return t
to nil to match RET in magit without evil. This would be the original (but
updated) behavior.
Another way to configure the behavior (and more context):
From: `magit-diff-visit-file'
In the past \\`<return>' (this command) used to go to the file in the
worktree, if point is on an added or context line of a diff showing
staged changes. Set `magit-diff-visit-prefer-worktree' to t to restore
that behavior, but note that doing so makes the behavior inconsistent
and you would give up on the ability to visit the index blob. If you
already use \\[magit-diff-visit-worktree-file] to jump to the live \
file from committed changes,
it might be better to retrain muscle memory to do the same from staged
changes.
|
|
C-j/C-k are remapped from C-n/C-p for section navigation, but the
corresponding M-n/M-p (sibling navigation) had no M-j/M-k equivalents.
This is inconsistent, especially since git-rebase-mode already maps
M-j/M-k for line movement.
Add M-j/M-k as additional bindings for magit-section-forward-sibling
and magit-section-backward-sibling in magit-mode-map, and for
git-commit-next-message and git-commit-prev-message in
git-commit-mode-map.
|
|
|
|
|
|
Only rewrite the EXCLUDE-NEWLINE argument when
evil-collection-magit-in-visual-pre-command is t. Previously the code
always rewrote the argument to nil even when
evil-collection-magit-in-visual-pre-command was not bound.
|
|
Fix #878
|
|
|
|
|
|
|
|
Magit's git-rebase-mode uses a different mode and a different keymap from the
main Magit buffer. toggle-text-mode also needs to disable read-only-mode in
rebase-mode buffers.
|
|
|
|
|
|
|
|
25 was already deprecated a year ago:
https://github.com/emacs-evil/evil-collection/pull/508/files
|
|
|
|
|
|
|
|
|
|
|
|
See https://github.com/magit/transient/issues/192.
|
|
Escape usually means to:
1. Return to normal mode.
2. Abort the minibuffer.
By binding it to quit-window (or equivalent), hitting escape a few times
to exit a (potentially recursive) minibuffer, etc. can cause buffers to
vanish.
It's also:
1. Inconsistently applied.
2. Redundant with "q" (which is consistently applied).
|
|
Fixes #571
Co-authored-by: Micah Buckley-Farlee <micah@super.mx>
|
|
|
|
|
|
Keybinding for magit-mode-map has also changed, Z is now magit-worktree
instead of magit-stash. Remap it to magit-stash. Magit-worktree still
has its old mapping to %.
|
|
magit-worktree has been moved from "%" to "Z" in transient prefix
magit-dispatch in commit 9bec1c54ab0705eb68ae4aa6c8a0498999a018f6.
When mapping magit-stash to Z in magit-dispatch due to
evil-collection-magit-use-z-for-folds, first move magit-worktree back to
its original keybinding to avoid collision.
|
|
close #502
|
|
|
|
Without using `kbd', each key in the string "<escape>" is interpreted as a
prefix key.
|
|
fix #492
|
|
`magit-show-refs-popup` and `magit-dispatch-popup`
have been renamed without "-popup".
Define magit-show-refs as a transient command
https://github.com/magit/magit/commit/c3135cdb23e223ebf8f3c802616ebb81103fdbe9
Define magit-dispatch as a transient command
https://github.com/magit/magit/commit/dde9a5b77a483fa35d932aecd22c1a4c954ce77e
And make the docstring more readable,
with the related key and command on the same line.
|
|
This change adjusts the formatting of the instructions in the `git-rebase`
buffer to allow them to be properly formatted when git's `core.commentChar` is
set to `%`. In particular, lines will look something like this:
```
%-8s p
```
The root cause seems to be the how the lines are constructed: comment characters
are directly added to the format string. When it is `%`, `format` appears to
proceed until it finds the next `%` character, (corresponding to the `%` in
`%-8s`) above and prints it directly to the string rather than interpreting it
something to format.
This is fixed by constructing the format string with the comment character as an
argument to `format` rather than being added directly to the format string.
Fixes #468.
|
|
"O" is used to launch `magit-reset`, not `magit-revert`. "_" and "-"
should be used for the suffix commands to be consistent with the
prefix key "_".
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Later we use lookup-key in this way:
(lookup-key aux-map (kbd (car cmd)))
which was slightly different from how we were defining the keymap which resulted
in M-j & M-k not being bound or displayed properly.
|
|
Both should do the same thing, which is magit-section-{forward,backward}-sibling.
|
|
|
|
|
|
|
|
Fix test failure in when testing original section bindings:
+ `((,(copy-keymap magit-file-section-map) "\C-j" magit-diff-visit-worktree-file)
+ (,(copy-keymap magit-hunk-section-map) "\C-j" magit-diff-visit-worktree-file))
|
|
|
|
Move setup files into subdirectories. Move mode directories to modes subdir.
This will make it easy to specify a recipe in melpa that maintains the directory
structure. Add interactive functions to jump to .el and README files
See #170
|
