Help generator: Add option to check HTML links.

We look up all "a href" and "img src" link in the project's HTML
files and give a warning if the link target does not exist or is
not listed in the project's filter sections.

Task-number: QTBUG-2485
Reviewed-by: kh1

5 files changed, 87 insertions, 4 deletions

diff --git a/tools/assistant/lib/qhelpgenerator.cpp b/tools/assistant/lib/qhelpgenerator.cpp
index 48d73aa..1bb4cc8 100644
--- a/tools/assistant/lib/qhelpgenerator.cpp
+++ b/tools/assistant/lib/qhelpgenerator.cpp
@@ -47,6 +47,7 @@
 #include <QtCore/QFileInfo>
 #include <QtCore/QDir>
 #include <QtCore/QDebug>
+#include <QtCore/QSet>
 #include <QtCore/QVariant>
 #include <QtCore/QDateTime>
 #include <QtCore/QTextCodec>
@@ -824,4 +825,68 @@ bool QHelpGenerator::insertMetaData(const QMap<QString, QVariant> &metaData)
     return true;
 }
 
+bool QHelpGenerator::checkLinks(const QHelpDataInterface &helpData)
+{
+    /*
+     * Step 1: Gather the canoncal file paths of all files in the project.
+     *         We use a set, because there will be a lot of look-ups.
+     */
+    QSet<QString> files;
+    foreach (const QHelpDataFilterSection &filterSection, helpData.filterSections()) {
+        foreach (const QString &file, filterSection.files()) {
+            QFileInfo fileInfo(helpData.rootPath() + QDir::separator() + file);
+            const QString &canonicalFileName = fileInfo.canonicalFilePath();
+            if (!fileInfo.exists())
+                emit warning(tr("File '%1' does not exist.").arg(file));
+            else
+                files.insert(canonicalFileName);
+        }
+    }
+
+    /*
+     * Step 2: Check the hypertext and image references of all HTML files.
+     *         Note that we don't parse the files, but simply grep for the
+     *         respective HTML elements. Therefore. contents that are e.g.
+     *         commented out can cause false warning.
+     */
+    bool allLinksOk = true;
+    foreach (const QString &fileName, files) {
+        if (!fileName.endsWith(QLatin1String("html"))
+            && !fileName.endsWith(QLatin1String("htm")))
+            continue;
+        QFile htmlFile(fileName);
+        if (!htmlFile.open(QIODevice::ReadOnly)) {
+            emit warning(tr("File '%1' cannot be opened.").arg(fileName));
+            continue;
+        }
+        const QRegExp linkPattern(QLatin1String("<(?:a href|img src)=\"?([^#\">]+)[#\">]"));
+        QTextStream stream(&htmlFile);
+        const QString codec = QHelpGlobal::codecFromData(htmlFile.read(1000));
+        stream.setCodec(QTextCodec::codecForName(codec.toLatin1().constData()));
+        const QString &content = stream.readAll();
+        QStringList invalidLinks;
+        for (int pos = linkPattern.indexIn(content); pos != -1;
+             pos = linkPattern.indexIn(content, pos + 1)) {
+            const QString& linkedFileName = linkPattern.cap(1);
+            if (linkedFileName.contains(QLatin1String("://")))
+                continue;
+            const QString curDir = QFileInfo(fileName).dir().path();
+            const QString &canonicalLinkedFileName =
+                QFileInfo(curDir + QDir::separator() + linkedFileName).canonicalFilePath();
+            if (!files.contains(canonicalLinkedFileName)
+                && !invalidLinks.contains(canonicalLinkedFileName)) {
+                emit warning(tr("File '%1' contains an invalid link to file '%2'").
+                         arg(fileName).arg(linkedFileName));
+                allLinksOk = false;
+                invalidLinks.append(canonicalLinkedFileName);
+            }
+        }
+    }
+
+    if (!allLinksOk)
+        d->error = tr("Invalid links in HTML files.");
+    return allLinksOk;
+}
+
 QT_END_NAMESPACE
+
diff --git a/tools/assistant/lib/qhelpgenerator_p.h b/tools/assistant/lib/qhelpgenerator_p.h
index 849e724..77189b0 100644
--- a/tools/assistant/lib/qhelpgenerator_p.h
+++ b/tools/assistant/lib/qhelpgenerator_p.h
@@ -74,6 +74,7 @@ public:
 
     bool generate(QHelpDataInterface *helpData,
         const QString &outputFileName);
+    bool checkLinks(const QHelpDataInterface &helpData);
     QString error() const;
 
 Q_SIGNALS:
diff --git a/tools/assistant/tools/qhelpgenerator/main.cpp b/tools/assistant/tools/qhelpgenerator/main.cpp
index 367d895..5866bc8 100644
--- a/tools/assistant/tools/qhelpgenerator/main.cpp
+++ b/tools/assistant/tools/qhelpgenerator/main.cpp
@@ -58,6 +58,7 @@ int main(int argc, char *argv[])
     QString basePath;
     bool showHelp = false;
     bool showVersion = false;
+    bool checkLinks = false;
 
     for (int i = 1; i < argc; ++i) {
         arg = QString::fromLocal8Bit(argv[i]);
@@ -72,6 +73,8 @@ int main(int argc, char *argv[])
             showVersion = true;
         } else if (arg == QLatin1String("-h")) {
             showHelp = true;
+        } else if (arg == QLatin1String("-c")) {
+            checkLinks = true;
         } else {
             QFileInfo fi(arg);
             projectFile = fi.absoluteFilePath();
@@ -93,6 +96,8 @@ int main(int argc, char *argv[])
         "                         file called <compressed-file>.\n"
         "                         If this option is not specified\n"
         "                         a default name will be used.\n"
+        "  -c                     Checks whether all links in HTML files\n"
+        "                         point to files in this help project.\n"
         "  -v                     Displays the version of \n"
         "                         qhelpgenerator.\n\n");
 
@@ -111,9 +116,11 @@ int main(int argc, char *argv[])
     }
 
     if (compressedFile.isEmpty()) {
-        QFileInfo fi(projectFile);
-        compressedFile = basePath + QDir::separator()
-            + fi.baseName() + QLatin1String(".qch");
+        if (!checkLinks) {
+            QFileInfo fi(projectFile);
+            compressedFile = basePath + QDir::separator()
+                             + fi.baseName() + QLatin1String(".qch");
+        }
     } else {
         // check if the output dir exists -- create if it doesn't
         QFileInfo fi(compressedFile);
@@ -134,7 +141,11 @@ int main(int argc, char *argv[])
 
     QCoreApplication app(argc, argv);
     HelpGenerator generator;
-    bool success = generator.generate(helpData, compressedFile);
+    bool success = true;
+    if (checkLinks)
+        success = generator.checkLinks(*helpData);
+    if (success && !compressedFile.isEmpty())
+        success = generator.generate(helpData, compressedFile);
     delete helpData;
     if (!success) {
         fprintf(stderr, "%s\n", qPrintable(generator.error()));
diff --git a/tools/assistant/tools/shared/helpgenerator.cpp b/tools/assistant/tools/shared/helpgenerator.cpp
index 956a22a..68b714e 100644
--- a/tools/assistant/tools/shared/helpgenerator.cpp
+++ b/tools/assistant/tools/shared/helpgenerator.cpp
@@ -61,6 +61,11 @@ bool HelpGenerator::generate(QHelpDataInterface *helpData,
     return generator->generate(helpData, outputFileName);
 }
 
+bool HelpGenerator::checkLinks(const QHelpDataInterface &helpData)
+{
+    return generator->checkLinks(helpData);
+}
+
 QString HelpGenerator::error() const
 {
     return generator->error();
diff --git a/tools/assistant/tools/shared/helpgenerator.h b/tools/assistant/tools/shared/helpgenerator.h
index ffb31bc..83a8195 100644
--- a/tools/assistant/tools/shared/helpgenerator.h
+++ b/tools/assistant/tools/shared/helpgenerator.h
@@ -57,6 +57,7 @@ public:
     HelpGenerator();
     bool generate(QHelpDataInterface *helpData,
         const QString &outputFileName);
+    bool checkLinks(const QHelpDataInterface &helpData);
     QString error() const;
 
 private slots: