_p9k_taskwarrior_init_data:32: bad floating point constant

[igbanam]

I recently upgraded taskwarrior to 3.0.1 and I get this error

_p9k_taskwarrior_init_data:32: bad floating point constant

This looks similar to #1365.

Commenting out this line "fixes" it — i.e. allows me to get the prompt back — but I'm unsure if it's the right fix.

The output of the command string in "ts" is 3.0.1

$ task +PENDING -OVERDUE list rc.verbose=nothing rc.color=0 rc._forcecolor=0 rc.report.list.labels= rc.report.list.columns=due.epoch </dev/null 2>/dev/null
3.0.1

…thus…

$ ($(command task +PENDING -OVERDUE list rc.verbose=nothing rc.color=0 rc._forcecolor=0 rc.report.list.labels= rc.report.list.columns=due.epoch </dev/null 2>/dev/null))
zsh: command not found: 3.0.1

I think this is the error here? …but my zsh-fu is not even at beginner level

powerlevel10k/internal/p10k.zsh

Line 5268 in 178fcda

_p9k_taskwarrior_next_due=${${(on)ts}[1]}

[igbanam]

…and I think I may have figured this one out. PR in a bit

[romkatv]

The output of "3.0.1" does not look like what we are trying to get. I suppose there was a backward-incompatible change in the command line API of task, so we now need to invoke a different command if the version of task is 3.x.

The PR you sent is on the wrong track, I'm afraid. I suggest that you disable taskwarrior prompt segment in your ~/.p10k.zsh and reverd the change you've made to powerlevel10k.

[igbanam]

That's okay. I'll close the PR.

Yes, 3.x has some breaking changes in them.

Now, writing this comment, I'm realizing ts may have meant "timestamp" where p10k is trying to set the timestamp of the next due task, yes? If so, the command needs to be reworked. I sadly don't know much TaskWarrior to help with the command either.

I'll leave the issue open for someone who hopefully knows how to get that.

[igbanam]

Also, how does one disable the task bit for p10k?

I searched and can't point to one shell variable I can turn off to disable this

[romkatv]

Also, how does one disable the task bit for p10k?

Open ~/.p10k.zsh and search for "timewarrior". You'll know what to do.

[romkatv]

I'm realizing ts may have meant "timestamp"

It's an array of pending tasks.

[cridemichel]

Hi,
I do not know if this can help but after having had same issue, I started reusing my own segment function, i.e.

function prompt_mytaskwarrior() {
  emulate -L zsh
  setopt pipefail
  (( $+commands[task] ))          || return
  local line IFS
  task +PENDING count 2>/dev/null | read -r lineP || return
  task +OVERDUE count 2>/dev/null | read -r lineO || return
  [[ "${lineP}" == "0" ]] && return
  lineP="⇡${lineP}"
  if [[ "${lineO}" == "0" ]]
  then
    lineO=""
  else
    lineO="!${lineO}"
  fi
  local tags="${lineP}${lineO}"
  [[ -n $tags ]] || return
  p10k segment -f 208 -i $'' -t $tags
}

and, although it is slow, it works flawlessly. To use it add this function to .zshrc before sourcing powerlevel10k and replace "taskwarrior" with" mytaskwarrior" in ".p10k.zsh"

best C.

[Bckempa]

This is a regression in taskwarrior that will be fixed in a 3.0.2 patch release

[romkatv]

Thanks! That's great know.

[Bckempa]

I no longer see this behavior after updating to Taskwarrior 3.0.2

[romkatv]

Thanks for the confirmation. Closing.

[igbanam]

Looks like the problem was with task. I can confirm this is fixed for me as well. Thanks @Bckempa!