From fa5a3ef6be2faed2f77d3eea4cf6c0f9e8f176ce Mon Sep 17 00:00:00 2001
From: Claude <noreply@anthropic.com>
Date: Sat, 17 Oct 2026 23:25:27 +0000
Subject: [PATCH] Don't treat ASCII information separators as binary data

isTextOnly() treats every control character which Qt does not consider
white space as a sign of binary data, so text containing the ASCII
information separators CHAR(28) to CHAR(31) was displayed as a BLOB. These
characters are regularly used as field and record delimiters in text, so
exempt U+001C to U+001F from the check.

The new test rows fail without this change.

Fixes #4193
---
 src/Data.cpp            |  5 ++++-
 src/tests/TestRegex.cpp | 22 ++++++++++++++++++++++
 src/tests/TestRegex.h   |  2 ++
 3 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/src/Data.cpp b/src/Data.cpp
index 9be484347..0ebf9a4a5 100644
--- a/src/Data.cpp
+++ b/src/Data.cpp
@@ -51,7 +51,10 @@ bool isTextOnly(QByteArray data, const QString& encoding, bool quickTest)
         QString text (data);
         bool isPrintable = true;
         for(QChar character : text) {
-            if(character.isNonCharacter() || (character.category() == QChar::Other_Control && !character.isSpace())) {
+            // The ASCII information separators (U+001C to U+001F) are control characters which Qt does not consider
+            // white space. They are regularly used as field delimiters in text, though, so don't treat them as binary.
+            const bool isSeparator = character.unicode() >= 0x1C && character.unicode() <= 0x1F;
+            if(character.isNonCharacter() || (character.category() == QChar::Other_Control && !character.isSpace() && !isSeparator)) {
                 isPrintable = false;
                 break;
             }
diff --git a/src/tests/TestRegex.cpp b/src/tests/TestRegex.cpp
index b05d7a966..d725df5f3 100644
--- a/src/tests/TestRegex.cpp
+++ b/src/tests/TestRegex.cpp
@@ -42,6 +42,28 @@ void TestRegex::clipboardRows()
     QVERIFY(parseClipboard(text) == expected);
 }
 
+void TestRegex::isTextOnlyControlCharacters_data()
+{
+    QTest::addColumn<QByteArray>("data");
+    QTest::addColumn<bool>("expected");
+
+    QTest::newRow("plain") << QByteArray("John Mary") << true;
+    QTest::newRow("whitespace-controls") << QByteArray("a\tb\r\nc") << true;
+    QTest::newRow("unit-separator") << QByteArray("John\x1f" "Mary") << true;
+    QTest::newRow("record-separator") << QByteArray("a\x1e" "b") << true;
+    QTest::newRow("group-and-file-separator") << QByteArray("a\x1d" "b\x1c" "c") << true;
+    QTest::newRow("nul") << QByteArray("a\0b", 3) << false;
+    QTest::newRow("bell") << QByteArray("a\x07" "b") << false;
+    QTest::newRow("escape") << QByteArray("a\x1b" "b") << false;
+}
+
+void TestRegex::isTextOnlyControlCharacters()
+{
+    QFETCH(QByteArray, data);
+    QFETCH(bool, expected);
+    QCOMPARE(isTextOnly(data), expected);
+}
+
 void TestRegex::sqlQueryComments_data()
 {
     QTest::addColumn<QString>("dirtyQuery");
diff --git a/src/tests/TestRegex.h b/src/tests/TestRegex.h
index d96e71326..f2c908502 100644
--- a/src/tests/TestRegex.h
+++ b/src/tests/TestRegex.h
@@ -14,6 +14,8 @@ private slots:
     void sqlQueryComments_data();
     void clipboardRows();
     void clipboardRows_data();
+    void isTextOnlyControlCharacters();
+    void isTextOnlyControlCharacters_data();
 };
 
 #endif