Error when running commands with wildcards

[externl]

See example

❯ fish -v
fish, version 3.1.2

❯ fasd --version
1.0.1

~/workspace/fasd
❯ ls
a.log  b.log  c.log

~/workspace/fasd
❯ rm -rf *log
~/.config/fish/conf.d/__fasd_run.fish (line 1): No matches for wildcard '*log'. See `help expand`.
echo rm -rf *log
            ^
in command substitution
	called on line 1 of file ~/.config/fish/conf.d/__fasd_run.fish
in command substitution
	called on line 4 of file ~/.config/fish/conf.d/__fasd_run.fish
in function '__fasd_run' with arguments 'rm\ -rf\ \*log'
in event handler: handler for generic event 'fish_postexec'

~/workspace/fasd
❯ ls
(files were removed)

[ianhomer]

the __fasd_run command is being run after every command. If it should only be run after something like the following might help ...

function __fasd_run -e fish_postexec -d "fasd takes record of the directories changed into"
  set -lx RETVAL $status
  if test (string sub --length 3 $argv) = "cd " -a $RETVAL -eq 0 
    # if there was no error
    command fasd --proc (command fasd --sanitize (eval echo "$argv") | tr -s " " \n) > "/dev/null" 2>&1 &
  end
end

I'm surprised that this extension is using this postexec on all commands - which sounds a little execessive. I'd have thought setting up an alias for cd (which also passes argument to fasd) would be more appropriate. Not sure all the use cases for this and not entirely sure this right ... any how, food for thought.

I created an experiment for this at master...ianhomer:master - however this is a breaking change, so alternative suggestions welcome

[gretel]

thanks, happy to accept PR.

[ianhomer]

I've created PR - it probably needs peer review and iteration, since I wouldn't be surprised if there are side effects for some people - e.g. the opinionated override of cd and loss of implicit behaviour on fish postexec.

[gretel]

folks, have switched to elvish - looking for new maintainer(s) for all of https://github.com/fishgretel!