This makes it easier to immediately start using dotfiles and other
config settings on first boot. It may even be useful to add an option
to run the entire bootstrap script in the installation media. Note that
if this route is taken, some assumptions regarding installation will
have to be changed to adjust for the chroot environment.
This commit makes it so that downloading the entire repository to run
the install scripts is no longer necessary.
It assumes that you have an active internet connection, which should be
a given since you need an internet connection to run pacstrap anyway.
Now that I use borderless and gapless windows by default, the window
management feature of kitty has become even more useful to me. This
change makes it easy to determine the active kitty window.
Despite having used tmux exclusively before, I have grown accustomed to
the benefits of using kitty as the window manager. tmux may still be
useful, for example, over ssh, but kitty is arguably the way to go for
local user sessions.
Note that this will affect the output of `env` since color codes are
used. Using a function to invoke man is not ideal since fish uses its
own function for its man pages.
It turns out that merge commit messages are used to keep track of
branch information since the commits themselves only reference other
commits and the actual branches may be deleted at a future date (even
`git clone` by default will not pull in any other branches).
Personally, I find a meaningful commit message more useful than a merge
commit message, so I'll be using that instead. Later when looking at
the commit history, "Merge branch <branch_name>" will not be useful
since the mentioned branches will have already been deleted.
Some projects rebase all changes into one commit, then push that
single commit to master. This seems like a very convenient approach
for small fixes, although if done for larger changes, each individual
commit and its respective explanation is lost.
In particular, this merge commit adds a bunch of config changes that
make polybar (and the system itself) more practical to use. Borders and
gaps are now disabled by default, and can be manually set through bspc
for your ricing needs.
It turns out that the left padding from the bspwm labels are used in
addition to the left padding from polybar, so in order for padding to be
consistent across both the left and right sections, the section with the
bspwm module needs to have its padding subtracted by the amount used by
the bspwm labels.
Time will tell whether or not I keep this, but one less color makes
it easier to seamlessly integrate the bar with different color schemes.
With this change, the interface is also arguably cleaner.
It turns out that showing the window title is very useful for actually
using the window manager, especially with no visual indicator through
window borders to determine which one is selected.
If I stayed on zsh instead of switch to fish, I would've probably never
known how easy it is to add certain information to the prompt. Since
the prompt itself is just a function, you can run any commands you want
inside it to get information, including git commands.
This commit adds the current branch you're on only when inside a git
repository, and only when you're not in a tty.
Realistically, you use a window manager to take up all the available
space on a screen. Borders and gaps are counter-intuitive in this
regard.
Since polybar has a module that shows the title of the focused window,
using a border width even in gapless mode is no longer necessary. This
also works conveniently well with bspwm's monocle desktop layout, which
will also inherit the no-border no-gaps methodology and take up all the
available screen space.
It turns out that polybar is able to provide very useful information for
window management in bspwm, so much so that disabling the bar completely
wouldn't make sense.
Since there's already a way to hide the bar (with a sxhkd keybind), and
since a "no bar" option would break this functionality even though the
hidden bar keybind produces the same result, it makes sense to remove
this.
This removes the minimize, maximize, and close buttons from GTK
applications, since we don't use them under bspwm anyway.
Note that Qt applications should already do this, no configuration
necessary.
The cursor theme size is set based on the X cursor size, which itself is
based on the X DPI.
Note that (as far as I know) there is no easy way to change the X cursor
size without restarting the X session with the new DPI.
Lots of changes here with the goal of streamlining the installation
process. The install scripts work very well now, although there are
still some final changes that I would like to make (notably the removal
of the unzip requirement and the inclusion of more options to automate
the process even further).
The polkit package is used to let unprivileged users run certain
commands. With it, it is possible to run system commands such as
`poweroff` and `reboot` with an unprivileged user.
Traditionally, only privileged users have access to power management,
but for many desktop setups, you may want to let unprivileged users
perform operations such as shutdown the machine if no other users
are logged in.
It turns out that the README for the xeventbind PKGBUILD actually works
very well here. On another note, I've yet to figure out how exactly I
want to document PKGBUILDs.
This is an X-specific script that changes (mostly) everything that needs
to be changed on resolution change, with the exception of X cursor size.
Note that any programs already running are not affected. For this kind
of functionality, you probably want to use Wayland instead.
It turns out that replacing the contents of locale.gen with the locales
you want is not a good idea, since then it is impossible to reference
the file at a later date to make changes as needed.
It turns out the font that I was setting before didn't actually do
anything. This is the real way to double the font size in your virtual
console, particularly useful for HiDPI displays, since your tty does not
know about your X DPI.
Since the username appears before the hostname (user@host), and since
anyone using these install scripts will probably want to log in with the
local user account and not the root account, this change makes sense.
Since mkinitcpio is already ran on the installation of the linux
package with pacstrap, and since we don't change any config settings,
recreating the initramfs is not necessary.
This should work, although I haven't tested it yet.
In the future it may be easier to simply download the other install
scripts from this install script. That way we solve the problem of
having to unzip this entire repository in the installation image.
This should ensure that a user knows what they're doing before running
the install scripts. Note that parted will also ask for a confirmation
before overwriting existing data.
Since fonts and other desktop-related packages are now handled by
tari-desktop, this is no longer necessary. Note that I also created
tari-web before I figured out HiDPI support for Qt and GTK applications.