Commit `1d6a2cda53`

1d6a2cda5366195c467051dc38b26571d54f4b9a

Verified · cmc

cmc <hello@cleberg.net> · 2026-10-09 06:50 UTC

Match Emacs word syntax for \<, \>, \b, \B, \w and \W in agenda and sparse-tree regexps

Layout: unified · split

Sources/OrgCore/Agenda/AgendaSearch.swift +9 −3

		@@ -480,6 +480,9 @@ public struct TagsMatcher: Sendable {
480	480	/// Emacs regular expressions run through ICU, case-insensitively as the matcher binds
481	481	/// `case-fold-search`.
482	482	enum EmacsRegex {
	483	/// Word syntax in an Org buffer: letters, digits, marks, `$` and `%`, but not `_`, `-` or `'`.
	484	static let word = "[\\p{L}\\p{N}\\p{M}$%\\u00b7]"
	485
483	486	static func translate(_ pattern: String) -> String {
484	487	var out = ""
485	488	var chars = Array(pattern)
		@@ -522,11 +525,14 @@ enum EmacsRegex {
522	525	out += n == "s" ? "[\\t\\n\\f\\r \\u00a0]" : "[^\\t\\n\\f\\r \\u00a0]"
523	526	i += 3
524	527	continue
525		case "<", ">": out += "\\b"
	528	case "<": out += "(?<!\(word))(?=\(word))"
	529	case ">": out += "(?<=\(word))(?!\(word))"
	530	case "b": out += "(?:(?<!\(word))(?=\(word))\|(?<=\(word))(?!\(word)))"
	531	case "B": out += "(?:(?<=\(word))(?=\(word))\|(?<!\(word))(?!\(word)))"
526	532	case "`": out += "^"
527	533	case "'": out += "$"
528		case "w": out += "[\\p{L}\\p{N}]"
529		case "W": out += "[^\\p{L}\\p{N}]"
	534	case "w": out += word
	535	case "W": out += "[^" + word.dropFirst()
530	536	default: out += "\\" + String(n)
531	537	}
532	538	i += 2

Tests/OrgCoreTests/AgendaSearchTests.swift +4

		@@ -28,6 +28,7 @@ struct AgendaSearchTests {
28	28	* Two words
29	29	The phrase two words
30	30	spans here; foo_bar and foo-bar.
	31	Costs $5 or 50% (don't) for the café.
31	32	* The last
32	33	word ends this entry with last
33	34	* word starts the next heading
		@@ -54,6 +55,9 @@ struct AgendaSearchTests {
54	55	"+alpha", "+alpha -beta", "+alpha +gamma", "-alpha", "+beta -nested", "{alph?a}", "+{^Gam}", "-{b[eo]}",
55	56	"+{\\<rent\\>}", "+\"two words\"", "+\"two words\"", "+\"the line\"", "{boiler\\b}", "+alpha +\"in the body\"",
56	57	"+{foo\\s-bar}", "+boiler -plate", "+{a b}", "+{alpha} -{x y}",
	58	// Emacs word syntax in Org buffers: `_`, `-` and `'` are not word characters; `$` and `%` are.
	59	"+{\\<bar\\>}", "+{foo\\b}", "+{o\\B_}", "+{\\<5\\>}", "+{\\b50\\b}", "+{don\\>}", "+{\\w+_\\w+}", "+{\\W5}",
	60	"+{caf\\w\\>}", "+{\\<\\$5}",
57	61	// Headlines only.
58	62	"rent", "alpha", "+alpha -beta", "+section", "-alpha", "pay rent", "*with alpha and",
59	63	// TODO only and whole words.

Commit 1d6a2cda53