From 725826231fda0f736157f087355501fbf61fcc4e Mon Sep 17 00:00:00 2001
From: "Dirk-Jan C. Binnema" <djcb@djcbsoftware.nl>
Date: Tue, 16 Mar 2021 16:51:01 +0200
Subject: [PATCH] utils: Add remove_ctrl

Add a helper function to remove control characters / multi-spaces, and a
test.
---
 lib/utils/mu-utils.cc   | 16 ++++++++++++++++
 lib/utils/test-utils.cc | 16 ++++++++++++++++
 2 files changed, 32 insertions(+)

diff --git a/lib/utils/mu-utils.cc b/lib/utils/mu-utils.cc
index 59ece80b..d80dfb2f 100644
--- a/lib/utils/mu-utils.cc
+++ b/lib/utils/mu-utils.cc
@@ -147,6 +147,22 @@ Mu::utf8_clean (const std::string& dirty)
 	clean.erase (clean.find_last_not_of(" ") + 1); // remove trailing space
 
 	return clean;
+std::string
+Mu::remove_ctrl (const std::string& str)
+{
+        char prev{'\0'};
+        std::string result;
+        result.reserve(str.length());
+
+        for (auto&& c: str) {
+                if (::iscntrl(c) || c == ' ') {
+                        if (prev != ' ')
+                                result += prev = ' ';
+                } else
+                        result += prev = c;
+        }
+
+        return result;
 }
 
 std::vector<std::string>
diff --git a/lib/utils/test-utils.cc b/lib/utils/test-utils.cc
index b1f666e5..020135f6 100644
--- a/lib/utils/test-utils.cc
+++ b/lib/utils/test-utils.cc
@@ -136,6 +136,22 @@ test_flatten ()
 	test_cases (cases, [](auto s, auto f){ return utf8_flatten(s); });
 }
 
+static void
+test_remove_ctrl ()
+{
+	CaseVec cases = {
+		{ "Foo\n\nbar", true,  "Foo bar" },
+		{ "",    false, "" },
+		{ "   ",    false, " " },
+		{ "Hello   World   ",    false, "Hello World " },
+		{ "Ångström", false,  "Ångström" },
+	};
+
+	test_cases (cases, [](auto s, auto f){ return remove_ctrl(s); });
+}
+
+
+
 static void
 test_clean ()
 {