It seems there's an edge case in EXWM where input events may occur
without keys to cause them (#8064), so these two keybind fixes need to
be ready to receive an empty vector from `this-single-command-raw-keys`.
Fix: #8064
This was a tricky regression to track down. 9753bfb tries to fix an
issue where the `default` face's :foreground changes to `#000000` in any
new frames created after the initial one (by calling `make-frame`),
because those frames' `background-color` and `foreground-color`
parameters default to "#000000" (possibly a bug with `disable-theme` too
eagerly defaulting them to black).
240493a replaces that with new, seemingly cleaner approach: setting
`frame-inherited-parameters`, which instructs `make-frame` to copy those
parameters from the last open frame, however, those parameters in the
initial daemon frame will be set to "unspecified-bg" or
"unspecified-fg" (see the docstring for `face-{back,fore}ground`), which
are invalid color strings. `make-frame` crashes tries to create a frame
with those color values, causing #8059.
Fix: #8059
Amend: 240493ae92
Amend: 9753bfb775
Those `face-*` calls sometimes returned nil, causing new frames spawned
from emacsclient to quietly crash sometimes. By instead relying on
`frame-inherited-parameters` we achieve the same but more stable result.
Amend: 9753bfb775
Now, uses can rebind ctrl+m by targeting [C-m] (or "<C-m>", same thing),
and it won't rebind RET. This will only work in GUI Emacs, however, and
there is no kkp support for this one.
Ref: 96d7e50f3e
Experienced users (the type to use symlinks in their projects) can add
-L back to `projectile-git-fd-args`, otherwise, its poorer UX for
beginners if Projectile indexes symlinked build artifacts (like 'result'
symlinks by Nix); i.e. things they didn't symlink themselves.
Fix: #7781
Instead of constantly checking $__DOOMSH, since I may eventually add
additional values for it (like in the cases where the doom script is
invoked by bin/doomscript, from a terminal emulator in Emacs, or future
wrappers).
c9acdb7 removes doom.cmd because it was broken in most cases. This adds
doom.ps1; an alternative script for Windows+Powershell users, which
properly initializes the state it needs. Naturally, it requires
Powershell 3+ be installed on your systems, but it can be invoked from
either cmd.exe or PowerShell.exe.
This is the first powershell script I've ever written, so I expect edge
cases (for one, shell commands passed to `exit!` will need to be guarded
against the environment).
This also requires emacs.exe be your $PATH, however, unless you set
$EMACS to its path first. E.g.
$env:EMACS = "C:\Program Files\Emacs\emacs-29.4\bin\emacs.exe"
That said, if you use WSL2, you're still far better off using the bash
script (bin/doom).
Ref: c9acdb72a4
New frames created after the first get initialized with #000000 as their
{back,fore}ground-color parameters, for some reason, making text
unreadable in dark themes.
`(de?)activate-mark-hook` is triggered a little too often in too many
edge cases where the user isn't interactively selecting a region. One
annoying edge case has non-evil motions used on evil operators
activating but not deactivating the mark (e.g. #8047).
This leaves non-evil users without the nicety of hl-line auto-disabling
itself when the region is active, but I'll revisit that in v3.0, when
most of these package defaults are moved out to modules.
Fix: #8047
