fish-shell/tests/checks/wraps.fish
Mahmoud Al-Qudsi 4cb19e244b Sort and deduplicate output of complete -C
This addresses a long-standing TODO where `complete -C` output isn't
deduplicated.

With this patch, the same deduplication and sort procedure that is run on actual
pager completions is also executed for `complete -C` completions (with a `-C`
payload specified).

This makes it possible to use `complete -C` to test what completions will
actually be generated by the completions pager instead of it displaying
something completely divorced from reality, improving the productivity of fish
completions developers.

Note that completions that wouldn't be shown in the pager are also omitted from
the results, e.g. `test/buildroot/` and `test/fish_expand_test/` are omitted
from the check matches in `checks/complete_directories.fish` because even if
they were generated, the pager wouldn't have shown them. This again makes
reasoning about and debugging completions much easier and more sane.
2022-10-31 16:52:36 -05:00

91 lines
2.7 KiB
Fish

#RUN: %fish %s
# Validate some things about command wrapping.
set -g LANG C # For predictable error messages.
# This tests that we do not trigger a combinatorial explosion - see #5638.
# Ensure it completes successfully.
complete -c testcommand --wraps "testcommand x "
complete -c testcommand --wraps "testcommand y "
complete -c testcommand --no-files -a normal
complete -C'testcommand '
# CHECK: normal
# Test double wraps.
complete -c testcommand0 -x -a crosswalk
complete -c testcommand1 -x --wraps testcommand0
complete -c testcommand2 -x --wraps testcommand1
complete -C 'testcommand 0'
# CHECK: crosswalk
# This tests that a call to complete from within a completion doesn't trigger
# wrap chain explosion - #5638 again.
function testcommand2_complete
set -l tokens (commandline -opc) (commandline -ct)
set -e tokens[1]
echo $tokens 1>&2
end
complete -c testcommand2 -x -a "(testcommand2_complete)"
complete -c testcommand2 --wraps "testcommand2 from_wraps "
complete -C'testcommand2 explicit '
# CHECKERR: explicit
# CHECKERR: from_wraps explicit
# Test that prefixing with a variable assignment works - see #7344.
complete -c recvar --exclusive -a recvar_comp
complete -c recvar --wraps 'A=B recvar'
complete -C 'recvar '
# CHECK: recvar_comp
# Test that completions do not perform subcommands.
# That is, `FOO=(launch_missiles) command<tab>` does not launch any missiles.
set -g missile_count 0
function launch_missiles
set -g missile_count (math "$missile_count + 1")
end
# Ensure missile launching work.
launch_missiles
echo $missile_count
# CHECK: 1
set -g GOOD global_good
set -g BAD global_bad
function do_print_good_bad
echo "GOOD:$GOOD"
echo "BAD:$BAD"
end
complete -c print_good_bad -x -a '(do_print_good_bad)'
complete -C 'print_good_bad '
# CHECK: BAD:global_bad
# CHECK: GOOD:global_good
# Key check is completions should expand GOOD but not BAD,
# because GOOD is just a string but BAD contains a cmdsub
# which may do arbitrary things.
complete -C 'GOOD=local_good BAD=(launch_missiles) print_good_bad '
# CHECK: BAD:
# CHECK: GOOD:local_good
# Completion should not have launched any missiles.
echo $missile_count
# CHECK: 1
# Torture test with a bunch of variable assignments.
# Note this tests the existing behavior when there's conflicts
# (innermost wrapped command wins) but it's not clear that this is desirable.
function show_vars
echo "AAA:$AAA"
echo "BBB:$BBB"
echo "CCC:$CCC"
end
complete -c show_vars_cmd0 -x -a '(show_vars)'
complete -c show_vars_cmd1 -x --wraps 'AAA=aaa show_vars_cmd0'
complete -c show_vars_cmd2 -x --wraps 'AAA=nope BBB=bbb show_vars_cmd1'
complete -c show_vars_cmd3 -x --wraps 'BBB=nope show_vars_cmd2'
complete -C 'CCC=ccc show_vars_cmd3 '
# CHECK: AAA:aaa
# CHECK: BBB:bbb
# CHECK: CCC:ccc