From c4f360baf07107468667832be66e5e9c3f92d4b6 Mon Sep 17 00:00:00 2001
From: EuAndreh <eu@euandre.org>
Date: Tue, 15 Jul 2025 13:55:36 -0300
Subject: Support tokenizing character class expressions [a-z]

* src/paca.mjs (classStateStep): New function equivalent to
  `rangeStateStep()` for character class expressions.  For now it knowns
  how to handle escaping ([abc\-_]), simple ranges ([a-z]), negation
  ([^abc]) and the hyphen literal as the first char ([-a-z_]).
* tests.paca.mjs (test_classStateStep): New test entry has a test case
  each scenario described above.
---
 src/paca.mjs | 128 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 125 insertions(+), 3 deletions(-)

(limited to 'src/paca.mjs')

diff --git a/src/paca.mjs b/src/paca.mjs
index 0b64a87..b92bcdc 100644
--- a/src/paca.mjs
+++ b/src/paca.mjs
@@ -1,5 +1,6 @@
 import {
-	butlast, explode, isNumeric, last, mapValues, max, reduce, reduced,
+	butlast, dissoc, explode, isNumeric, last, mapValues, max, reduce,
+	reduced,
 } from "sjs";
 
 
@@ -11,6 +12,7 @@ const ConcatStep = {
 	ACCEPTING: "accepting",
 	ESCAPING:  "escaping",
 	RANGE:     "range",
+	CLASS:     "class",
 };
 
 const numFromDigits = digits =>
@@ -107,9 +109,130 @@ const rangeStateStep = ({ out, state, context }, char, _index, _next) => {
 	};
 };
 
+const classStateStep = ({ out, state, context }, char, _index, _next) => {
+	if (context.escaping) {
+		return {
+			out,
+			state,
+			context: dissoc({
+				...context,
+				set: context.set.concat(char),
+			}, "escaping"),
+		};
+	}
+
+	if (char === "]") {
+		if (context.range.where === "to") {
+			return reduced({
+				out,
+				state,
+				context,
+				error: new SyntaxError(
+					"unfinished character class range",
+				),
+			});
+		}
+
+		if (context.set.length === 0) {
+			return reduced({
+				out,
+				state,
+				context,
+				error: new ValueError("empty character class"),
+			});
+		}
+
+		return {
+			out: out.concat({
+				operator: "class",
+				set:      context.set,
+			}),
+			state:   ConcatStep.ACCEPTING,
+			context: null,
+		};
+	}
+
+	if (char === "\\") {
+		return {
+			out,
+			state,
+			context: {
+				...context,
+				escaping: true,
+			},
+		};
+	}
+
+	if (context.range.where === "to") {
+		const from = context.range.from;
+		const to   = char;
+
+		if (from.charCodeAt(0) > to.charCodeAt(0)) {
+			return reduced({
+				out,
+				state,
+				context,
+				error: new ValueError(
+					"bad class range values: " +
+						`[${from}-${to}]`,
+				),
+			});
+		}
+
+		return {
+			out,
+			state,
+			context: {
+				...context,
+				set:   context.set.concat({ from, to }),
+				range: {
+					from:  null,
+					where: "from",
+				},
+			},
+		};
+	}
+
+	if (char === "-" && context.set.length !== 0) {
+		return {
+			out,
+			state,
+			context: {
+				...context,
+				set: butlast(context.set),
+				range: {
+					from:  last(context.set),
+					where: "to",
+				},
+			},
+		};
+	}
+
+	if (char === "^" && context.set.length === 0) {
+		return {
+			out,
+			state,
+			context: {
+				...context,
+				caret: true,
+			},
+		};
+	}
+
+	return {
+		out,
+		state,
+		context: {
+			...context,
+			set: context.set.concat(char),
+		},
+	};
+};
+
 const STATE_FNS = {
 	[ConcatStep.ESCAPING]: escapingStateStep,
 	[ConcatStep.RANGE   ]:    rangeStateStep,
+	[ConcatStep.CLASS   ]:    classStateStep,
 };
 
 const TRANSITION_FNS = {
@@ -133,8 +256,7 @@ const TRANSITION_FNS = {
 		context: {
 			set:   [],
 			range: {
-				from:  [],
-				to:    [],
+				from:  null,
 				where: "from",
 			},
 		},
-- 
cgit v1.2.3