fix: gzip file parsing as a stream

Signed-off-by: Trial97 <alexandru.tripon97@gmail.com>
This commit is contained in:
Trial97 2025-04-19 00:41:24 +03:00
parent a39edb3b59
commit d1c7107575
No known key found for this signature in database
GPG Key ID: 55EF5DA53DB36318
3 changed files with 107 additions and 125 deletions

View File

@ -139,64 +139,80 @@ bool GZip::zip(const QByteArray& uncompressedBytes, QByteArray& compressedBytes)
return true;
}
GZipStream::GZipStream(const QString& filePath) : GZipStream(new QFile(filePath)) {}
GZipStream::GZipStream(QFile* file) : m_file(file) {}
bool GZipStream::initStream()
int inf(QFile* source, std::function<bool(const QByteArray&)> handleBlock)
{
memset(&m_strm, 0, sizeof(m_strm));
return (inflateInit2(&m_strm, 16 + MAX_WBITS) == Z_OK);
constexpr auto CHUNK = 16384;
int ret;
unsigned have;
z_stream strm;
memset(&strm, 0, sizeof(strm));
char in[CHUNK];
unsigned char out[CHUNK];
ret = inflateInit2(&strm, (16 + MAX_WBITS));
if (ret != Z_OK)
return ret;
/* decompress until deflate stream ends or end of file */
do {
strm.avail_in = source->read(in, CHUNK);
if (source->error()) {
(void)inflateEnd(&strm);
return Z_ERRNO;
}
if (strm.avail_in == 0)
break;
strm.next_in = reinterpret_cast<Bytef*>(in);
/* run inflate() on input until output buffer not full */
do {
strm.avail_out = CHUNK;
strm.next_out = out;
ret = inflate(&strm, Z_NO_FLUSH);
assert(ret != Z_STREAM_ERROR); /* state not clobbered */
switch (ret) {
case Z_NEED_DICT:
ret = Z_DATA_ERROR; /* and fall through */
case Z_DATA_ERROR:
case Z_MEM_ERROR:
(void)inflateEnd(&strm);
return ret;
}
have = CHUNK - strm.avail_out;
if (!handleBlock(QByteArray(reinterpret_cast<const char*>(out), have))) {
(void)inflateEnd(&strm);
return Z_OK;
}
bool GZipStream::unzipBlockByBlock(QByteArray& uncompressedBytes)
} while (strm.avail_out == 0);
/* done when inflate() says it's done */
} while (ret != Z_STREAM_END);
/* clean up and return */
(void)inflateEnd(&strm);
return ret == Z_STREAM_END ? Z_OK : Z_DATA_ERROR;
}
QString zerr(int ret)
{
uncompressedBytes.clear();
if (!m_file->isOpen()) {
if (!m_file->open(QIODevice::ReadOnly)) {
qWarning() << "Failed to open file:" << (m_file->fileName());
return false;
switch (ret) {
case Z_ERRNO:
return QObject::tr("error handling file");
case Z_STREAM_ERROR:
return QObject::tr("invalid compression level");
case Z_DATA_ERROR:
return QObject::tr("invalid or incomplete deflate data");
case Z_MEM_ERROR:
return QObject::tr("out of memory");
case Z_VERSION_ERROR:
return QObject::tr("zlib version mismatch!");
}
return {};
}
if (!m_strm.state && !initStream()) {
return false;
}
QByteArray compressedBlock;
unsigned int blockSize = 4096;
compressedBlock = m_file->read(blockSize);
if (compressedBlock.isEmpty()) {
return true; // End of file reached
}
bool done = processBlock(compressedBlock, uncompressedBytes);
if (inflateEnd(&m_strm) != Z_OK || !done) {
return false;
}
return done;
}
bool GZipStream::processBlock(const QByteArray& compressedBlock, QByteArray& uncompressedBytes)
QString GZip::readGzFileByBlocks(QFile* source, std::function<bool(const QByteArray&)> handleBlock)
{
m_strm.next_in = (Bytef*)compressedBlock.data();
m_strm.avail_in = compressedBlock.size();
unsigned int uncompLength = uncompressedBytes.size();
if (m_strm.total_out >= uncompLength) {
uncompressedBytes.resize(uncompLength * 2);
uncompLength *= 2;
}
m_strm.next_out = reinterpret_cast<Bytef*>(uncompressedBytes.data() + m_strm.total_out);
m_strm.avail_out = uncompLength - m_strm.total_out;
int err = inflate(&m_strm, Z_NO_FLUSH);
if (err != Z_OK && err != Z_STREAM_END) {
qWarning() << "Decompression failed with error code" << err;
return false;
}
return true;
auto ret = inf(source, handleBlock);
return zerr(ret);
}

View File

@ -1,28 +1,11 @@
#pragma once
#include <zlib.h>
#include <QByteArray>
#include <QFile>
class GZip {
public:
static bool unzip(const QByteArray& compressedBytes, QByteArray& uncompressedBytes);
static bool zip(const QByteArray& uncompressedBytes, QByteArray& compressedBytes);
};
namespace GZip {
class GZipStream {
public:
explicit GZipStream(const QString& filePath);
explicit GZipStream(QFile* file);
bool unzip(const QByteArray& compressedBytes, QByteArray& uncompressedBytes);
bool zip(const QByteArray& uncompressedBytes, QByteArray& compressedBytes);
QString readGzFileByBlocks(QFile* source, std::function<bool(const QByteArray&)> handleBlock);
// Decompress the next block and return the decompressed data
bool unzipBlockByBlock(QByteArray& uncompressedBytes);
private:
bool initStream();
bool processBlock(const QByteArray& compressedBlock, QByteArray& uncompressedBytes);
private:
QFile* m_file;
z_stream m_strm;
};
} // namespace GZip

View File

@ -151,54 +151,13 @@ void OtherLogsPage::on_selectLogBox_currentIndexChanged(const int index)
}
}
class ReadLineAbstract {
public:
ReadLineAbstract(QFile* file) : m_file(file)
{
if (file->fileName().endsWith(".gz"))
m_gz = new GZipStream(file);
}
~ReadLineAbstract() { delete m_gz; }
QString readLine()
{
if (!m_gz)
return QString::fromUtf8(m_file->readLine());
QString line;
for (;;) {
if (!m_decodedData.isEmpty()) {
int newlineIndex = m_decodedData.indexOf('\n');
if (newlineIndex != -1) {
line += QString::fromUtf8(m_decodedData).left(newlineIndex);
m_decodedData.remove(0, newlineIndex + 1);
return line;
}
line += QString::fromUtf8(m_decodedData);
m_decodedData.clear();
}
if (!m_gz->unzipBlockByBlock(m_decodedData)) { // If error occurs during unzipping
m_decodedData.clear();
return QObject::tr("The content of the file(%1) could not be decoded.").arg(m_file->fileName());
}
}
}
bool done() { return m_gz ? m_decodedData.isEmpty() : m_file->atEnd(); }
private:
QFile* m_file;
GZipStream* m_gz = nullptr;
QByteArray m_decodedData;
};
void OtherLogsPage::on_btnReload_clicked()
{
if (m_currentFile.isEmpty()) {
setControlsEnabled(false);
return;
}
QFile file(FS::PathCombine(m_path, m_currentFile));
if (!file.open(QFile::ReadOnly)) {
setControlsEnabled(false);
@ -220,15 +179,9 @@ void OtherLogsPage::on_btnReload_clicked()
showTooBig();
return;
}
ReadLineAbstract stream(&file);
// Try to determine a level for each line
ui->text->clear();
ui->text->setModel(nullptr);
m_model->clear();
auto line = stream.readLine();
while (!stream.done()) { // just read until the model is full or the file ended
auto handleLine = [this](QString line) {
if (line.isEmpty())
return false;
if (line.back() == '\n')
line = line.remove(line.size() - 1, 1);
MessageLevel::Enum level = MessageLevel::Unknown;
@ -245,10 +198,40 @@ void OtherLogsPage::on_btnReload_clicked()
}
m_model->append(level, line);
if (m_model->isOverFlow())
break;
return m_model->isOverFlow();
};
line = stream.readLine();
// Try to determine a level for each line
ui->text->clear();
ui->text->setModel(nullptr);
m_model->clear();
if (file.fileName().endsWith(".gz")) {
QString line;
auto error = GZip::readGzFileByBlocks(&file, [&line, handleLine](const QByteArray& d) {
auto block = d;
int newlineIndex = block.indexOf('\n');
while (newlineIndex != -1) {
line += QString::fromUtf8(block).left(newlineIndex);
block.remove(0, newlineIndex + 1);
if (handleLine(line)) {
line.clear();
return false;
}
line.clear();
newlineIndex = block.indexOf('\n');
}
line += QString::fromUtf8(block);
return true;
});
if (!error.isEmpty()) {
setPlainText(tr("The file (%1) encountered an error when reading: %2.").arg(file.fileName(), error));
return;
} else if (!line.isEmpty()) {
handleLine(line);
}
} else {
while (!file.atEnd() && !handleLine(QString::fromUtf8(file.readLine()))) {
}
}
ui->text->setModel(m_proxy);
ui->text->scrollToBottom();