From 7a577df3e283ca1dae5e3e1f83d746937650ab6f Mon Sep 17 00:00:00 2001
From: Grzegorz Orczykowski <grzesiek2010@gmail.com>
Date: Sat, 13 Apr 2024 01:07:52 +0200
Subject: [PATCH] Improved handling markdown

---
 .../java/org/odk/collect/android/utilities/HtmlUtils.java   | 5 +----
 .../org/odk/collect/android/utilities/HtmlUtilsTest.java    | 6 ++++++
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/collect_app/src/main/java/org/odk/collect/android/utilities/HtmlUtils.java b/collect_app/src/main/java/org/odk/collect/android/utilities/HtmlUtils.java
index d911b601670..90d6ccfe053 100644
--- a/collect_app/src/main/java/org/odk/collect/android/utilities/HtmlUtils.java
+++ b/collect_app/src/main/java/org/odk/collect/android/utilities/HtmlUtils.java
@@ -87,10 +87,7 @@ static String markdownToHtml(String text) {
         text = text.replaceAll("(?s)\\*\\*(.*?)\\*\\*", "<strong>$1</strong>");
 
         // emphasis using underscore
-        text = text.replaceAll("\\s_([^\\s][^_\n]*)_\\s", " <em>$1</em> ");
-        text = text.replaceAll("^_([^\\s][^_\n]*)_$", "<em>$1</em>");
-        text = text.replaceAll("^_([^\\s][^_\n]*)_\\s", "<em>$1</em> ");
-        text = text.replaceAll("\\s_([^\\s][^_\n]*)_$", " <em>$1</em>");
+        text = text.replaceAll("(^|[\\s]|[^\\w])_([^\\s][^_\n]*)_($|[\\s]|[^\\w])", "$1<em>$2</em>$3");
 
         // emphasis using asterisk
         text = text.replaceAll("\\*([^\\s][^\\*\n]*)\\*", "<em>$1</em>");
diff --git a/collect_app/src/test/java/org/odk/collect/android/utilities/HtmlUtilsTest.java b/collect_app/src/test/java/org/odk/collect/android/utilities/HtmlUtilsTest.java
index 2c75128f42f..066c048c18b 100644
--- a/collect_app/src/test/java/org/odk/collect/android/utilities/HtmlUtilsTest.java
+++ b/collect_app/src/test/java/org/odk/collect/android/utilities/HtmlUtilsTest.java
@@ -35,7 +35,13 @@ public void markDownToHtmlEscapesBackslash() {
                 {"A _B_ C", "A <em>B</em> C"},
                 {"A_B_ C", "A_B_ C"},
                 {"A _B_C", "A _B_C"},
+                {"_A_5", "_A_5"},
                 {"_A_", "<em>A</em>"},
+                {"(_A_)", "(<em>A</em>)"},
+                {"_A_?", "<em>A</em>?"},
+                {"*_A_", "*<em>A</em>"},
+                {"blah _A_!", "blah <em>A</em>!"},
+                {" _A_! blah", " <em>A</em>! blah"},
                 {"\\_ _AB\\__", "_ <em>AB_</em>"},
                 {"\\\\ _AB\\_\\\\_", "\\ <em>AB_\\</em>"},
                 {"A\\*B\\*C", "A*B*C"},