Mirrors/fish-shell
2
0
Fork 0
mirror of https://github.com/fish-shell/fish-shell synced 2024-12-25 20:33:08 +00:00
Code Issues Projects Releases Packages Wiki Activity

Optimize __fish_complete_suffix and fix duplicate listings

Browse source 
With a blank $suff (i.e. complete all files), __fish_complete_suffix
returned directories twice, once with the trailing `/` and once without.
This fixes that, and additionally speeds up the code by no longer
shelling out to `sort -u` as we no longer rely on brace expansion to
enumerate directories and files simultaneously.

In general, this behavior would occur when a directory exists that
matches the suffix search pattern (so a dir named 'foo.bar' with a
search pattern '.bar' would return 'foo.bar' twice).

Runtime has dropped from ~22ms to ~8ms on my machine, while also
returning more correct results.
This commit is contained in:
Mahmoud Al-Qudsi 2018-06-18 21:32:05 -05:00
parent 5061f1666b
commit 6e36b20e42
1 changed files with 25 additions and 14 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

39
share/functions/__fish_complete_suffix.fish
View file

@ -41,17 +41,33 @@ function __fish_complete_suffix -d "Complete using files"
# echo "base: $base" > /dev/tty # echo "base: $base" > /dev/tty
# echo "suffix: $suff" > /dev/tty # echo "suffix: $suff" > /dev/tty
set -l all
set -l dirs
# If $comp is "./ma" and the file is "main.py", we'll catch that case here, # If $comp is "./ma" and the file is "main.py", we'll catch that case here,
# but complete.cpp will not consider it a match, so we have to output the # but complete.cpp will not consider it a match, so we have to output the
# correct form. # correct form.
# Also do directory completion, since there might be files with the correct
# suffix in a subdirectory. `eval` is used since $suff may be passed in
# as {.foo,.bar} and we want to expand that.
eval "set all $base*$suff"
if not string match -qr '/$' -- $suff
eval "set dirs $base*/"
# The problem is that we now have each directory included twice in the output,
# once as `dir` and once as `dir/`. The runtime here is O(n) for n directories
# in the output, but hopefully since we have only one level (no nested results)
# it should be fast. The alternative is to shell out to `sort` and remove any
# duplicate results, but it would have to be a huge `n` to make up for the fork
# overhead.
for dir in $dirs
set all (string match -v (string match -r '(.*)/$' -- $dir)[2] -- $all)
end
end
set files $dirs $all
if string match -qr '^\\./' -- $comp if string match -qr '^\\./' -- $comp
# Also do directory completion, since there might be files set files ./$files
# with the correct suffix in a subdirectory
eval "set files ./$base*{$suff,/}"
else
# Also do directory completion, since there might be files
# with the correct suffix in a subdirectory
eval "set files $base*{$suff,/}"
end end
# Another problem is that expanded paths are not matched, either. # Another problem is that expanded paths are not matched, either.
@ -63,16 +79,11 @@ function __fish_complete_suffix -d "Complete using files"
set files (string replace -- $expanded $comp $files) set files (string replace -- $expanded $comp $files)
end end
# It's very unfortunate to do all this work in-process and have to shell out here, if set -q files[1]
# but unfortunately at this time expressions like "foo{t,te}*" applied against
# "footer" will result in "footer" being reported twice. Not sure if this can be
# term a "bug" per-se.
if test $files[1]
if not string match -q -- "$desc" "" if not string match -q -- "$desc" ""
set -l desc "\t$desc" set -l desc "\t$desc"
end end
printf "%s$desc\n" $files | sort -u printf "%s$desc\n" $files #| sort -u
end end
end end