From f8270369b688cb8bb69f10d4c772eeebaf23bc80 Mon Sep 17 00:00:00 2001
From: tastytea <tastytea@tastytea.de>
Date: Tue, 8 Jun 2021 17:30:29 +0200
Subject: [PATCH] Make whitespace-reduction a bit more efficient.

We now use 2 passes instead of 3.
---
 src/search.cpp                | 41 +++++++++++++++++++----------------
 tests/test_search_helpers.cpp |  2 +-
 2 files changed, 23 insertions(+), 20 deletions(-)

diff --git a/src/search.cpp b/src/search.cpp
index 2605e56..173d81f 100644
--- a/src/search.cpp
+++ b/src/search.cpp
@@ -133,26 +133,29 @@ file_in_epub cleanup_text(const std::string_view text)
                                            "(title|aria-label)"
                                            "=\"([[:alnum:]]+)\""};
 
-    // TODO: Make this more efficient, 3 → 1;
+    {
+        size_t pos{0};
+        while ((pos = output.find_first_of("\n\t\r", pos)) != string::npos)
+        {
+            if (output[pos] == '\r')
+            {
+                output.erase(pos, 1);
+            }
+            else
+            {
+                output.replace(pos, 1, " ");
+            }
+        }
+    }
+    {
+        size_t pos{0};
+        while ((pos = output.find("  ", pos)) != string::npos)
+        {
+            output.replace(pos, 2, " ");
+        }
+    }
+
     size_t pos{0};
-    while ((pos = output.find('\r', pos)) != string::npos)
-    {
-        output.erase(pos, 1);
-    }
-
-    pos = 0;
-    while ((pos = output.find_first_of("\n\t", pos)) != string::npos)
-    {
-        output.replace(pos, 1, " ");
-    }
-
-    pos = 0;
-    while ((pos = output.find("  ", pos)) != string::npos)
-    {
-        output.replace(pos, 2, " ");
-    }
-
-    pos = 0;
     file_in_epub file;
     size_t headline_start{string::npos};
     while ((pos = output.find('<', pos)) != string::npos)
diff --git a/tests/test_search_helpers.cpp b/tests/test_search_helpers.cpp
index d9026a6..cdb23bf 100644
--- a/tests/test_search_helpers.cpp
+++ b/tests/test_search_helpers.cpp
@@ -54,7 +54,7 @@ SCENARIO("Searching helpers work as intended")
                 }
 
                 THEN("No exception is thrown")
-                AND_THEN("The \\r are removed unchanged")
+                AND_THEN("The \\r are removed")
                 {
                     REQUIRE_FALSE(exception);
                     REQUIRE(text == "💖🦝");