Tiebreak parameter of length or end does not produce the expected results about fzf

[mattie20]

--scheme=path does exactly what I need. I haven't really had many problems with the default beyond this, so I wouldn't worry about it too much. Plus, it would hurt results if you were, for example, looking through man pages. Thank you!

[junegunn]

Does --scheme=path help?

fzf --tiebreak end --scheme path --query engine << EOF
Documents/c++code/pascal.wiki/engine.md
Documents/Archive/College/ME4201/AIAA EOP/Gas Turbine Engine Pictures/J79.JPG
Documents/Archive/College/ME4201/AIAA EOP/Gas Turbine Engine Pictures/CFM56 High Bypass Turbofan Engine.jpg
EOF

It's because in the default scheme, Engine after a space gets a higher bonus score than engine after /.

fzf/src/algo/algo.go

Lines 178 to 189 in 3ee1fc2

	case "default":
	bonusBoundaryWhite = bonusBoundary + 2
	bonusBoundaryDelimiter = bonusBoundary + 1
	case "path":
	bonusBoundaryWhite = bonusBoundary
	bonusBoundaryDelimiter = bonusBoundary + 1
	if os.PathSeparator == '/' {
	delimiterChars = "/"
	} else {
	delimiterChars = string([]rune{os.PathSeparator, '/'})
	}
	initialCharClass = charDelimiter

--scheme=path makes fzf use the same bonus point for both cases.

---

Come to think of it, it's a little strange that we don't use --scheme=path as the default when using fzf for files (i.e. using its default walker), let me see what I can do.

---

On second thought, I think it can be confusing to use different defaults on different contexts. While the default scheme gives a less desirable result in this case, it works okay in most cases where file names don't contain spaces.

[junegunn]

Hopefully, 64faaf61f3030c006392bce7e370200b912ad001 will help users better understand the options.