| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Automatically refresh the following environment variables in all zsh
sessions inside tmux when we attach:
SSH_CONNECTION SSH_CLIENT SSH_TTY SSH_AUTH_SOCK DISPLAY
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Too many damn exploits.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Avoids errors like this one after every command:
warning: could not open directory 'pkg/': Permission denied
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Building in tmpfs is more trouble than it's worth, even with rather
large amounts of RAM many packages run out of space while building.
Combined with the fact that SSDs can actually sustain a really large
amounts of writes makes it unnecessary.
|
| |
|
|
|
|
| |
~/.zshrc:unalias:9 no such hash table element: run-help
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
10000 lines should suffice in general, no need to waste additional
memory. Additionally hist_expire_dups_first has been removed, because
this causes a lot of additional CPU load, probably loading $HISTFILE
over and over again.
If further history is required, it can be loaded with a function like
the following:
function hist {
local TEMPHISTFILE="$HISTFILE"
local TEMPSIZE=$SAVEHIST
fc -pa
HISTSIZE=$TEMPSIZE
fc -R "$TEMPHISTFILE"
fc -l 0
}
The problem with the above is that it will increase the memory footprint
of the zsh process, up to ~50MiB, that doesn't seem to be freed later.
Just grepping $HISTFILE produces a similar result, without this
drawback.
|
|
|
|
|
|
|
|
|
|
|
|
| |
i.e. the following clone fails:
$ git clone https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git
Cloning into 'linux-firmware'...
remote: Counting objects: 6843, done.
remote: Total 6843 (delta 0), reused 0 (delta 0)
error: object 95eb5456368d6f10f59bcc3f0cd9a8bffaf7243e: zeroPaddedFilemode: contains zero-padded file modes
fatal: fsck error in packed object
fatal: index-pack failed
|
|
|
|
|
|
| |
oh-my-zsh introduces *a lot* of bloat, but I've barely been using any of
it these past years. Just picking the options I use makes zsh easier to
work with and debug (and hopefully faster).
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
CtrlP is nice when there are not a lot of files in the directory, but
becomes really slow on larger codebases (i.e. the linux kernel). While
it's possible to work around this to some extend, there are much better
alternatives now.
The most general (and fastest) of which is fzf. Combined with fd, a
faster `find` it handles even absurd amounts of files practically
instantly.
The only real downside of this setup is that it requires both fzf and fd
to be installed.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Took many settings from tpope's vim-sensible[0] and the new vim
defaults.vim[1] and incorporated them into my vimrc. See also [2] for an
explanation of those common settings.
Everything is now properly divided into sections now as well, and
folding by markers (`{{{`) is enabled.
[0] https://github.com/tpope/vim-sensible
[1] https://github.com/vim/vim/blob/master/runtime/defaults.vim
[2] https://github.com/sheerun/vimrc
|
|
|
|
|
|
| |
Hopefully every major distro updated their isync version by now.
This reverts commit e1c90590177d3431dd118a105656b793ee010483.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
weechat unfortunately changes the salt each time it writes to sec.conf,
even if the password itself hasn't changed. From the documentation of
sec.crypt.salt:
description: use salt when generating key used in encryption
(recommended for maximum security); when enabled, the content of crypted
data in file sec.conf will be different on each write of the file; if
you put the file sec.conf in a version control system, then you can turn
off this option to have always same content in file
Since we're using a long and random password anyway, that isn't being
reused, this is safe to disable.
|
| |
|
| |
|
|
|
|
| |
i.e.: Channel created on So, 15 Feb 2004 06:08:31
|
| |
|
| |
|
| |
|
| |
|