QMimeDatabase: Fix MIME detection issues with magics in MIME hierarchies

Assume two MIME types A and B are registered, both with the same glob
pattern, A being parent of B, A with some magic rule, and B with another
magic rule. Given a file that matches the glob pattern and the magic rule
of A, the resulting MIME type depended on the order of registration of A
and B, because it would just check if some glob matching MIME type was
also a subclass of the magic matching MIME type.

The patch prefers the the MIME type that matches by magic if that
matches by glob pattern as well (i.e. A in our example).

The "recommended checking order" of the spec does handle that case.

Task-number: QTBUG-44846
Change-Id: I2af43f6199faf9a42cd9c35d3a045441afbd6217
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Reviewed-by: David Faure <david.faure@kdab.com>

6 files changed, 42 insertions, 2 deletions

diff --git a/src/corelib/mimetypes/qmimedatabase.cpp b/src/corelib/mimetypes/qmimedatabase.cpp
index 68e3c8f10d..a58698af53 100644
--- a/src/corelib/mimetypes/qmimedatabase.cpp
+++ b/src/corelib/mimetypes/qmimedatabase.cpp
@@ -382,9 +382,12 @@ QMimeType QMimeDatabasePrivate::mimeTypeForFileNameAndData(const QString &fileNa
 
         // Disambiguate conflicting extensions (if magic matching found something)
         if (candidateByData.isValid() && magicAccuracy > 0) {
-            // "for glob_match in glob_matches:"
-            // "if glob_match is subclass or equal to sniffed_type, use glob_match"
             const QString sniffedMime = candidateByData.name();
+            // If the sniffedMime matches a glob match, use it
+            if (candidatesByName.m_matchingMimeTypes.contains(sniffedMime)) {
+                *accuracyPtr = 100;
+                return candidateByData;
+            }
             for (const QString &m : qAsConst(candidatesByName.m_matchingMimeTypes)) {
                 if (inherits(m, sniffedMime)) {
                     // We have magic + pattern pointing to this, so it's a pretty good match
diff --git a/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy.foo b/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy.foo
new file mode 100644
index 0000000000..ed5e761417
--- /dev/null
+++ b/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy.foo
@@ -0,0 +1,3 @@
+<?foo>
+<blah>
+</blah>
diff --git a/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy.xml b/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy.xml
new file mode 100644
index 0000000000..27b5bd7e1f
--- /dev/null
+++ b/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy.xml
@@ -0,0 +1,21 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mime-info xmlns='http://www.freedesktop.org/standards/shared-mime-info'>
+    <!-- The real life example would be that this mime type is a sub class of application/xml
+         which has a magic that matches &lt;?xml .
+         XML files with an early appearing qnx tag are detected as
+         application/vnd.qnx.bar-descriptor. We want that XML files without the qnx tag to be
+         identified as application/xml, independent of the order in which the two are registered. -->
+    <mime-type type="application/vnd.qnx.bar-descriptor">
+      <sub-class-of type="application/foo"/>
+      <glob pattern="*.foo"/>
+      <magic><!-- higher priority than the parent magic -->
+        <match value="&lt;qnx&gt;" type="string" offset="0:200"/>
+      </magic>
+    </mime-type>
+    <mime-type type="application/foo">
+      <glob pattern="*.foo"/>
+      <magic priority="40">
+        <match value="&lt;?foo" type="string" offset="0"/>
+      </magic>
+    </mime-type>
+</mime-info>
diff --git a/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy2.foo b/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy2.foo
new file mode 100644
index 0000000000..fd90ed04b3
--- /dev/null
+++ b/tests/auto/corelib/mimetypes/qmimedatabase/magic-and-hierarchy2.foo
@@ -0,0 +1,3 @@
+<?foo>
+<qnx>
+</qnx>
diff --git a/tests/auto/corelib/mimetypes/qmimedatabase/testdata.qrc b/tests/auto/corelib/mimetypes/qmimedatabase/testdata.qrc
index 29666627a1..1002d0195d 100644
--- a/tests/auto/corelib/mimetypes/qmimedatabase/testdata.qrc
+++ b/tests/auto/corelib/mimetypes/qmimedatabase/testdata.qrc
@@ -7,5 +7,8 @@
         <file>invalid-magic1.xml</file>
         <file>invalid-magic2.xml</file>
         <file>invalid-magic3.xml</file>
+        <file>magic-and-hierarchy.xml</file>
+        <file>magic-and-hierarchy.foo</file>
+        <file>magic-and-hierarchy2.foo</file>
     </qresource>
 </RCC>
diff --git a/tests/auto/corelib/mimetypes/qmimedatabase/tst_qmimedatabase.cpp b/tests/auto/corelib/mimetypes/qmimedatabase/tst_qmimedatabase.cpp
index 3144c3071c..597d51e7e0 100644
--- a/tests/auto/corelib/mimetypes/qmimedatabase/tst_qmimedatabase.cpp
+++ b/tests/auto/corelib/mimetypes/qmimedatabase/tst_qmimedatabase.cpp
@@ -52,6 +52,7 @@ static const char *const additionalMimeFiles[] = {
     "invalid-magic1.xml",
     "invalid-magic2.xml",
     "invalid-magic3.xml",
+    "magic-and-hierarchy.xml",
     0
 };
 
@@ -985,6 +986,12 @@ void tst_QMimeDatabase::installNewGlobalMimeType()
         qDebug() << objcsrc.globPatterns();
     }
 
+    const QString fooTestFile = QLatin1String(RESOURCE_PREFIX "magic-and-hierarchy.foo");
+    QCOMPARE(db.mimeTypeForFile(fooTestFile).name(), QString::fromLatin1("application/foo"));
+
+    const QString fooTestFile2 = QLatin1String(RESOURCE_PREFIX "magic-and-hierarchy2.foo");
+    QCOMPARE(db.mimeTypeForFile(fooTestFile2).name(), QString::fromLatin1("application/vnd.qnx.bar-descriptor"));
+
     // Now test removing the mimetype definitions again
     for (int i = 0; i < m_additionalMimeFileNames.size(); ++i)
         QFile::remove(destDir + m_additionalMimeFileNames.at(i));