From cc5c0bf95c2bf14f516e745eafc6307e23609cca Mon Sep 17 00:00:00 2001
From: Humza Shahid <humzasaur@gmail.com>
Date: Mon, 6 Oct 2025 21:44:57 +0100
Subject: [PATCH] implement escape sequences for regex

---
 fcore/escape-string.sml                  | 63 ------------------------
 fcore/normal-mode/normal-search-mode.sml |  5 +-
 fcore/search-list/dfa-gen.sml            | 42 ++++++++++++++++
 shf.mlb                                  |  1 -
 4 files changed, 43 insertions(+), 68 deletions(-)
 delete mode 100644 fcore/escape-string.sml

diff --git a/fcore/escape-string.sml b/fcore/escape-string.sml
deleted file mode 100644
index ed6c687..0000000
--- a/fcore/escape-string.sml
+++ /dev/null
@@ -1,63 +0,0 @@
-structure EscapeString =
-struct
-  (* The specific escape sequences handled are in the 
-   * `rewriteAcc` function below.
-   * These escape sequences were decided on because
-   * they are the same escape sequences recognised in Standard ML.
-   * A reference is available here: 
-   * https://smlfamily.github.io/Basis/char.html#SIG:CHAR.fromCString:VAL
-   *
-   * However, thre omissions have been made:
-   * - \ddd denoting an integer in the range 0 - 255
-   * - \uxxxx denoting the character whose code is the integer xxxx
-   * - \f f\ denoting a sequence of characters to ignore
-   *
-   * In the first two cases, it is easier to type the character directly.
-   * In the third case, there is little use for it.
-   * *)
-  fun rewriteAcc (#"\\", acc) =
-        (case acc of
-           #"a" :: tl => #"\a" :: tl
-         | #"b" :: tl => #"\b" :: tl
-         | #"t" :: tl => #"\t" :: tl
-         | #"n" :: tl => #"\n" :: tl
-         | #"v" :: tl => #"\v" :: tl
-         | #"f" :: tl => #"\f" :: tl
-         | #"r" :: tl => #"\r" :: tl
-         | #"?" :: tl => #"?" :: tl
-         | #"\\" :: tl => #"\\" :: tl
-         | #"\"" :: tl => #"\"" :: tl
-         | #"'" :: tl => #"'" :: tl
-         | #"^" :: hd :: tl =>
-             (* handle control characters *)
-             let
-               val code = Char.ord hd
-             in
-               if code >= 64 andalso code <= 95 then
-                 let val chr = Char.chr (code - 64)
-                 in chr :: tl
-                 end
-               else
-                 (* invalid escape sequence: leave alone *)
-                 #"\\" :: acc
-             end
-         | _ =>
-             (* when there is no valid escape sequence, 
-              * just leave slash in output *)
-             #"\\" :: acc)
-    | rewriteAcc (chr, acc) = chr :: acc
-
-  fun help (pos, str, acc) =
-    if pos < 0 then
-      String.implode acc
-    else
-      let
-        val chr = String.sub (str, pos)
-        val acc = rewriteAcc (chr, acc)
-      in
-        help (pos - 1, str, acc)
-      end
-
-  fun unescape str =
-    help (String.size str - 1, str, [])
-end
diff --git a/fcore/normal-mode/normal-search-mode.sml b/fcore/normal-mode/normal-search-mode.sml
index 610705f..0a7d0b1 100644
--- a/fcore/normal-mode/normal-search-mode.sml
+++ b/fcore/normal-mode/normal-search-mode.sml
@@ -5,9 +5,7 @@ struct
   open MailboxType
 
   fun buildTempSearchList (searchString, buffer, cursorIdx) =
-    let val unescapedString = EscapeString.unescape searchString
-    in SearchList.buildRange (buffer, unescapedString, cursorIdx + 1111)
-    end
+    SearchList.buildRange (buffer, searchString, cursorIdx + 1111)
 
   fun addChr
     (app: app_type, searchString, searchCursorIdx, searchScrollColumn, chr) =
@@ -66,7 +64,6 @@ struct
         } = app
 
       val buffer = LineGap.goToStart buffer
-      val searchString = EscapeString.unescape searchString
       val initialMsg = [SEARCH (buffer, searchString, time)]
 
       (* move LineGap to first line displayed on screen *)
diff --git a/fcore/search-list/dfa-gen.sml b/fcore/search-list/dfa-gen.sml
index 5b62085..1ba181c 100644
--- a/fcore/search-list/dfa-gen.sml
+++ b/fcore/search-list/dfa-gen.sml
@@ -109,6 +109,30 @@ struct
       fun getRightParenIdx (pos, str) = loop (pos, str, 1, 0)
     end
 
+    (* assumes previous char is a backslash *)
+    fun isValidEscapeSequence chr =
+      case chr of
+      (* regex metacharacters *)
+        #"(" => true
+      | #")" => true
+      | #"[" => true
+      | #"]" => true
+      | #"+" => true
+      | #"*" => true
+      | #"|" => true
+      | #"?" => true
+      (* standard escape sequences *)
+      | #"\a" => true
+      | #"\b" => true
+      | #"\t" => true
+      | #"\n" => true
+      | #"\v" => true
+      | #"\f" => true
+      | #"\r" => true
+      | #"\\" => true
+      | #"\"" => true
+      | _ => false
+
     fun computeAtom (pos, str, stateNum) =
       if pos = String.size str then
         NONE
@@ -132,6 +156,24 @@ struct
         | #"*" => NONE
         | #"+" => NONE
         | #"." => SOME (pos + 1, WILDCARD (stateNum + 1), stateNum + 1)
+        | #"\\" =>
+            (* escape sequences *)
+            if pos + 1 = String.size str then
+              NONE
+            else
+              let
+                val chr = String.sub (str, pos + 1)
+              in
+                if isValidEscapeSequence chr then
+                  let
+                    val chr = CHAR_LITERAL {char = chr, position = stateNum + 1}
+                  in
+                    SOME (pos + 2, chr, stateNum + 1)
+                  end
+                else
+                  NONE
+              end
+        (* todo: [character classes] *)
         | chr =>
             let val chr = CHAR_LITERAL {char = chr, position = stateNum + 1}
             in SOME (pos + 1, chr, stateNum + 1)
diff --git a/shf.mlb b/shf.mlb
index af7e1e9..87b7585 100644
--- a/shf.mlb
+++ b/shf.mlb
@@ -6,7 +6,6 @@ lib/brolib-sml/src/gap_set.sml
 lib/cozette-sml/fonts-with-z-index/cozette-ascii.mlb
 
 (* FUNCTIONAL CORE *)
-fcore/escape-string.sml
 fcore/bin-search.sml
 
 ann