zsh/coredumpctl: Fix the completion

An output from coredumpctl list is like

> TIME                            PID   UID   GID SIG COREFILE  EXE
> Sun 2016-05-29 18:44:03 CEST  14578  1000  1000   6 none      /tmp/pacaurtmp-wieland/python33/src/Python-3.3.6/python

  ^1  ^2         ^3       ^4    ^5

, but the previous sub() command turns that into

> TIMEPID   UID   GID SIG COREFILE  EXE
> Sun2016-05-29 18:44:03 CEST  14578  1000  1000   6 none      /tmp/pacaurtmp-wieland/python33/src/Python-3.3.6/python
  ^1            ^2       ^3    ^4     ^5

so the whole pipeline generated entries like

$UID:$DESCRIPTION

but that's not useful and probably not what was supposed to happen.

This now generates entries like

$PID:$DESCRIPTION

which make everything work.

Note that with this commmit, the completions will be sorted by PID by
ZSH.

diff --git a/shell-completion/zsh/_coredumpctl b/shell-completion/zsh/_coredumpctl
index ea8cfb2decf12bf2508778cc8dfa61b42003b678..3445aa6f3d9684264aa06881c69ed2851cb9e603 100644 (file)
--- a/shell-completion/zsh/_coredumpctl
+++ b/shell-completion/zsh/_coredumpctl
@@ -17,7 +17,7 @@ _coredumpctl_command(){
         cmd="${${_coredumpctl_cmds[(r)$words[1]:*]%%:*}}"
         if (( $#cmd  )); then
             # user can set zstyle ':completion:*:*:coredumpctl:*' sort no for coredumps to be ordered by date, otherwise they get ordered by pid
-            _dumps=( "${(foa)$(coredumpctl list --no-legend | awk 'BEGIN{OFS=":"} {sub(/[[ \t]+/, ""); print $5,$0}' 2>/dev/null)}" )
+            _dumps=( "${(foa)$(coredumpctl list --no-legend | awk 'BEGIN{OFS=":"} {sub(/[[ \t]+/, ""); print $4,$0}' 2>/dev/null)}" )
             if [[ -n "$_dumps" ]]; then
                 _describe -t pids 'coredumps' _dumps
             else