Arch Linux User Repository

Thanks qbvt,

I've reported this issue under bash-preexec:

https://aur.archlinux.org/packages/bash-preexec/#comment-828581

Thanks for the tip!

Hello kuba-orlik,

This looks like an issue with bash-preexec [1][2], on which this package depends.

Unfortunately, I cannot reproduce your issue on my setup. This makes it hard for me to diagnose the issue you might have. Have you configured anything in your shell related to this HISTCONTROL variable ?

I think the best way to try and get this solved is:
1. Try to install bash-preexec [1] independantly of undistract-me. If this works, report back here.
2. If it does not work, manually download [2] and try to run the tests as described at the bottom of the README there. If the tests pass, but the bash-preexec AUR package fails to install, report back at [1] (I also take care of this package),
3. If the tests don't pass, report the issue on their side and/or try to fix it there,

[1] https://aur.archlinux.org/packages/bash-preexec/
[2] https://github.com/rcaloras/bash-preexec

Installation fails for me with errors:

==> Starting check()...
 ✓ __bp_install_after_session_init should exit with 1 if we're not using bash
 ✗ __bp_install should exit if it's already installed
   (from function `bp_install' in file ./bash-preexec.bats, line 13,
    in test file ./bash-preexec.bats, line 32)
     `bp_install' failed
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✗ __bp_install should remove trap logic and itself from PROMPT_COMMAND
   (in test file ./bash-preexec.bats, line 45)
     `eval "$PROMPT_COMMAND"' failed
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✗ __bp_install should preserve an existing DEBUG trap
   (in test file ./bash-preexec.bats, line 52)
     `trap_invoked_count=0' failed
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✓ __bp_sanitize_string should remove semicolons and trim space
 ✗ Appending to PROMPT_COMMAND should work after bp_install
   (from function `bp_install' in file ./bash-preexec.bats, line 13,
    in test file ./bash-preexec.bats, line 85)
     `bp_install' failed
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✗ Appending or prepending to PROMPT_COMMAND should work after bp_install_after_session_init
   (in test file ./bash-preexec.bats, line 102)
     `}' failed
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✗ Adding to PROMPT_COMMAND before and after initiating install
   (in test file ./bash-preexec.bats, line 115)
     `' failed
   after2
   before
   before2
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✗ Adding to PROMPT_COMMAND after with semicolon
   (in test file ./bash-preexec.bats, line 126)
     `' failed
   before
   /home/kuba/.cache/yay/bash-preexec/src/bash-preexec-0.4.1/test/./../bash-preexec.sh: line 73: HISTCONTROL: unbound variable
 ✗ during install PROMPT_COMMAND and precmd functions should be executed each once
   (in test file ./bash-preexec.bats, line 143)
     `[ "${lines[3]}" == "inside precmd" ]' failed

Hi, sorry for the late response,

In short, this points to a fork because:
1/ Upstream relies on an outdated or/and exotic version of bash-preexec,
2/ Using that version breaks compatibility with anything else that uses an up-to-date version of bash-preexec,
3/ Upstream seems abandonned,

To expand a little bit:
- regarding 2/, my fork updates bash-preexec with the "official", up-to-date version, which fixes compatibility issues, and also allows installing bash-preexec as a package, which I think is good,
- regarding 3/ I opened PR#68 on the upstream repo. That was 4 months ago, and I got no answer from the author.

In case you are/know the author, I'd actually prefer to merge my stuff in and delete my fork, but for now it's the only way for me to have the functionality.

Cheers,

stevenroose, if you take a look at the comparison between this package's version and the original, you'll note that an "Important Note" was added to the README explaining why.

A fork, particularly of a project apparently abandoned by upstream, is not necessarily malicious. I think you'll find the code comparison linked above reassuring.

Why is this package not using the original repo (https://github.com/jml/undistract-me), but a personal fork?