[gentoo-user] New bashrc

[gentoo-user] New bashrc

[Peter Humphrey]

Hello list,

Would someone please either post the new bashrc file here or send me a copy, 
in which genfun_has_readline is defined? I seem to have lost mine altogether. I 
think portage may be being too clever in remembering config files that I've not 
allowed it to update previously, and so not offering to update it now.

I mean app-shells/bash-5.2_p26-r6.

Many thanks.

-- 
Regards,
Peter.

Re: [gentoo-user] New bashrc

[Dale]

[-- Attachment #1: Type: text/plain, Size: 1248 bytes --]

Peter Humphrey wrote:
> Hello list,
>
> Would someone please either post the new bashrc file here or send me a copy, 
> in which genfun_has_readline is defined? I seem to have lost mine altogether. I 
> think portage may be being too clever in remembering config files that I've not 
> allowed it to update previously, and so not offering to update it now.
>
> I mean app-shells/bash-5.2_p26-r6.
>
> Many thanks.
>


I think this gives the new config files.  I save the binaries that
emerge creates.  I found the new bash and extracted the config files. 
The ones that start with a number are in the bashrc.d directory.  They
are attached. 

I might add, after the update, I had issues with my PS1 and alias
settings.  My main rig went rather well given it has been through a few
updates.  The new rig however was like pulling teeth.  I ended up
copying the /etc/bash directory from my main rig.  Then it worked. 
Using dispatch-conf and me rarely being able to use it correctly is
causing problems.  I need to find a better tool.  Thing is, I don't
think one exists.  I'd encourage you to use the bashrc.d directory when
possible.  That way your settings aren't mixed in with defaults. 

Hope those files help. 

Dale

:-)  :-) 

[-- Attachment #2: 10-gentoo-color.bash --]
[-- Type: text/plain, Size: 2209 bytes --]

# /etc/bash/bashrc.d/10-gentoo-color.bash

if [[ ${NO_COLOR} ]]; then
	# Respect the user's wish not to use color. See https://no-color.org/.
	gentoo_color=0
elif [[ ${COLORTERM@a} == *x* && ${COLORTERM} == @(24bit|truecolor) ]]; then
	# The COLORTERM environment variable can reasonably be trusted here.
	# See https://github.com/termstandard/colors for further information.
	gentoo_color=1
elif unset -v COLORTERM; ! gentoo_color=$(tput colors 2>/dev/null); then
	# Either ncurses is not installed or no terminfo database could be
	# found. Fall back to a whitelist which covers the majority of terminal
	# emulators and virtual console implementations known to support color
	# and which remain (somewhat) popular. This will rarely happen, so the
	# list need not be exhaustive.
	case ${TERM} in
		*color*    |\
		*direct*   |\
		[Ekx]term* |\
		alacritty  |\
		aterm      |\
		dtterm     |\
		foot*      |\
		jfbterm    |\
		linux      |\
		mlterm     |\
		rxvt*      |\
		screen*    |\
		tmux*      |\
		wsvt25*    ) gentoo_color=1
	esac
elif (( gentoo_color == 16777216 )); then
	# Truecolor support is available. Advertise it.
	export COLORTERM=truecolor
fi

# For direxpand to be missing indicates that bash is lacking readline support.
if (( gentoo_color <= 0 )) || [[ ! $(shopt -p direxpand 2>/dev/null) ]]; then
	# Define a prompt without color.
	PS1='\u@\h \w \$ '
elif (( EUID == 0 )); then
	# If root, omit the username and print the hostname in red.
	PS1='\[\e[01;31m\]\h\[\e[01;34m\] \w \$\[\e[00m\] '
else
	# Otherwise, print the username and hostname in green.
	PS1='\[\e[01;32m\]\u@\h\[\e[01;34m\] \w \$\[\e[00m\] '
fi

if (( gentoo_color > 0 )); then
	# Colorize the output of diff(1), grep(1) and a few coreutils utilities.
	for _ in diff dir grep ls vdir; do
		alias "$_=$_ --color=auto"
	done

	# Enable colors for ls(1) and some other utilities that respect the
	# LS_COLORS variable. Prefer ~/.dir_colors, per bug #64489.
	if hash dircolors 2>/dev/null; then
		if [[ -f ~/.dir_colors ]]; then
			eval "$(dircolors -b -- ~/.dir_colors)"
		elif [[ -f /etc/DIR_COLORS ]]; then
			eval "$(dircolors -b /etc/DIR_COLORS)"
		else
			eval "$(dircolors -b)"
		fi
	fi
fi

unset -v gentoo_color

[-- Attachment #3: 10-gentoo-title.bash --]
[-- Type: text/plain, Size: 1851 bytes --]

# /etc/bash/bashrc.d/10-gentoo-title.bash

genfun_set_win_title() {
	# Assigns the basename of the current working directory, having
	# sanitised it with @Q parameter expansion. Useful for paths containing
	# newlines and such. As a special case, names consisting entirely of
	# graphemes shall not undergo the expansion, for reasons of cleanliness.
	genfun_sanitise_cwd() {
		_cwd=${PWD##*/}
		if [[ ! ${_cwd} ]]; then
			_cwd=${PWD}
		elif [[ ${_cwd} == *[![:graph:]]* ]]; then
			_cwd=${_cwd@Q}
		fi
	}

	# Sets the window title with the Set Text Parameters sequence. For
	# screen, the sequence defines the hardstatus (%h) and for tmux, the
	# pane_title (#T). For graphical terminal emulators, it is normal for
	# the title bar to be affected.
	genfun_set_win_title() {
		genfun_sanitise_cwd
		printf '\033]2;%s@%s - %s\007' "${USER}" "${HOSTNAME%%.*}" "${_cwd}"
	}

	genfun_set_win_title
}

# If the TTY is that of sshd(8) then proceed no further. Alas, there exist many
# operating environments in which the window title would otherwise not be
# restored upon ssh(1) exiting. Users wishing to coerce the historical
# behaviour have the option of setting PROMPT_COMMAND=(genfun_set_win_title).
if [[ ${SSH_TTY} && ${SSH_TTY} == "$(tty)" ]]; then
	return
fi

# Determine whether the terminal can handle the Set Text Parameters sequence.
# The only terminals permitted here are those for which there is empirical
# evidence that the sequence is supported and that the UTF-8 character encoding
# is handled correctly. Quite rightly, this precludes many vintage terminals.
# https://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h3-Operating-System-Commands
case ${TERM} in
	alacritty     |\
	foot*         |\
	rxvt-unicode* |\
	screen*       |\
	st-256color   |\
	tmux*         |\
	xterm*        )
		PROMPT_COMMAND+=('genfun_set_win_title')
esac

[-- Attachment #4: 15-gentoo-bashrc-check.bash --]
[-- Type: text/plain, Size: 1002 bytes --]

# /etc/bash/bashrc.d/15-gentoo-bashrc-check.bash

# Some users have ~/.bashrc as a copy of ${FILESDIR}/bashrc which either matches
# exactly or is only trivially modified. Such is an improper state of affairs
# and results in the bashrc.d drop-ins being sourced twice. Warn them that they
# should use the skeleton file instead. This drop-in should be removed no sooner
# than one year from the date of its introduction.

if [[ -e ${TMPDIR:-/tmp}/.gentoo-bashrc-check-${EUID} || ! -f ~/.bashrc ]]; then
	return
fi

{
	if grep -qxF 'for sh in /etc/bash/bashrc.d/* ; do' -- ~/.bashrc; then
		cat >&3 <<'EOF'
WARNING! Your ~/.bashrc file is based on an old copy of /etc/bash/bashrc, which
is not intended for use within a home directory. Please either delete ~/.bashrc
or replace it with a copy of /etc/skel/.bashrc before optionally customizing it
further. Neglecting to do so may result in bash behaving unexpectedly.

EOF
	fi
	touch -- "${TMPDIR:-/tmp}/.gentoo-bashrc-check-${EUID}"
} 3>&2 2>/dev/null

[-- Attachment #5: bash_logout --]
[-- Type: text/plain, Size: 205 bytes --]

# /etc/bash/bash_logout

# This file is sourced when a login shell terminates.

# You may wish to clear everyone's screen when they logout.
#clear

# Or maybe you want to leave a thoughtful note.
#fortune

[-- Attachment #6: bashrc --]
[-- Type: text/plain, Size: 894 bytes --]

# /etc/bash/bashrc

# Proceed no further in the case of a non-interactive shell.
if [[ $- != *i* ]]; then
	return
fi

# Disable errexit in case the user enabled it then chose to re-source this file.
shopt -u -o errexit

# Disable completion when the input buffer is empty. Mute STDERR because this
# option is only present in the case that bash was built with readline support.
shopt -s no_empty_cmd_completion 2>/dev/null &&

# Append to HISTFILE rather than overwrite upon exiting, per bug #139609. This
# option also requires for bash to have been built with readline support.
shopt -s histappend

# Initialise PROMPT_COMMAND as an array, which is permitted as of bash 5.1.
PROMPT_COMMAND=()

# Don't let the user influence the order of sourcing for bash 5.3 or greater.
unset -v GLOBSORT

for _ in /etc/bash/bashrc.d/*; do
	if [[ $_ == *.@(bash|sh) && -r $_ ]]; then
		source "$_"
	fi
done

Re: [gentoo-user] New bashrc

[Peter Humphrey]

On Saturday, 6 July 2024 17:31:08 BST Dale wrote:

> Hope those files help. 

Thank you Dale. In fact my problem turned out to be an invisible typo in a 
bash script. You know, the sort that isn't there until the umpteenth time you 
look, and there it is after all.  :)

-- 
Regards,
Peter.

Re: [gentoo-user] New bashrc

[Dale]

Peter Humphrey wrote:
> On Saturday, 6 July 2024 17:31:08 BST Dale wrote:
>
>> Hope those files help. 
> Thank you Dale. In fact my problem turned out to be an invisible typo in a 
> bash script. You know, the sort that isn't there until the umpteenth time you 
> look, and there it is after all.  :)
>


Maybe that was my problem as well.  My main rig update went fairly
well.  The new rig, which had a fresh new set of never altered files,
went sideways.  I really hated over writing the files that way but I
didn't know of a better way. 

Given that things are moving to a directory now, I wish someone would
write a really good wiki page on how the new way works and have some
examples.  One thing I don't know and haven't found a answer too, the
numbers.  Does it read from low to high?  That makes sense.  Does a
setting from a file with a higher number override settings for files
with lower numbers?  That's a biggie.  When I tried to set my PS1
variable, it clashed with a Gentoo file which set the same variable.  It
seems one doesn't override the other but turns into a wreck, with
injuries.  My bash kinda got weird.  I couldn't login at all. 

I suspect someone will write one at some point.  I just wish it was
sooner rather than later.  I'm not clear on a few things. 

Glad to help.  :-D

Dale

:-)  :-) 

P. S.  I might add, xorg.conf is doing the same thing and I have the
same questions about it.  It seems a lot of things are going the
directory route and it can be a good thing, just confusing at times.