scummvm/common/archive.cpp
2023-12-24 13:19:25 +01:00

580 lines
17 KiB
C++

/* ScummVM - Graphic Adventure Engine
*
* ScummVM is the legal property of its developers, whose names
* are too numerous to list here. Please refer to the COPYRIGHT
* file distributed with this source distribution.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
*/
#include "common/archive.h"
#include "common/file.h"
#include "common/fs.h"
#include "common/system.h"
#include "common/textconsole.h"
#include "common/memstream.h"
#include "common/punycode.h"
#include "common/debug.h"
namespace Common {
ArchiveMember::~ArchiveMember() {
}
U32String ArchiveMember::getDisplayName() const {
return getName();
}
bool ArchiveMember::isInMacArchive() const {
return false;
}
bool ArchiveMember::isDirectory() const {
return false;
}
void ArchiveMember::listChildren(ArchiveMemberList &childList, const char *pattern) const {
}
GenericArchiveMember::GenericArchiveMember(const String &pathStr, const Archive &parent)
: _parent(parent), _path(pathStr, parent.getPathSeparator()) {
}
GenericArchiveMember::GenericArchiveMember(const Path &path, const Archive &parent)
: _parent(parent), _path(path) {
}
String GenericArchiveMember::getName() const {
return _path.toString(_parent.getPathSeparator());
}
Path GenericArchiveMember::getPathInArchive() const {
return _path;
}
String GenericArchiveMember::getFileName() const {
return _path.getLastComponent().toString(_parent.getPathSeparator());
}
SeekableReadStream *GenericArchiveMember::createReadStream() const {
return _parent.createReadStreamForMember(_path);
}
SeekableReadStream *GenericArchiveMember::createReadStreamForAltStream(AltStreamType altStreamType) const {
return _parent.createReadStreamForMemberAltStream(_path, altStreamType);
}
bool GenericArchiveMember::isDirectory() const {
return _parent.isPathDirectory(_path);
}
void GenericArchiveMember::listChildren(ArchiveMemberList &childList, const char *pattern) const {
if (!pattern)
pattern = "*";
Common::Path searchPath = _path.appendComponent(pattern);
_parent.listMatchingMembers(childList, searchPath);
}
bool Archive::isPathDirectory(const Path &path) const {
return false;
}
int Archive::listMatchingMembers(ArchiveMemberList &list, const Path &pattern, bool matchPathComponents) const {
// Get all "names" (TODO: "files" ?)
ArchiveMemberList allNames;
listMembers(allNames);
String patternString = pattern.toString(getPathSeparator());
int matches = 0;
char pathSepString[2] = {getPathSeparator(), '\0'};
const char *wildcardExclusions = matchPathComponents ? NULL : pathSepString;
ArchiveMemberList::const_iterator it = allNames.begin();
for (; it != allNames.end(); ++it) {
// TODO: We match case-insenstivie for now, our API does not define whether that's ok or not though...
// For our use case case-insensitive is probably what we want to have though.
if ((*it)->getName().matchString(patternString, true, wildcardExclusions)) {
list.push_back(*it);
matches++;
}
}
return matches;
}
SeekableReadStream *Archive::createReadStreamForMemberAltStream(const Path &path, AltStreamType altStreamType) const {
return nullptr;
}
Common::Error Archive::dumpArchive(const Path &destPath) {
Common::ArchiveMemberList files;
listMembers(files);
byte *data = nullptr;
uint dataSize = 0;
for (auto &f : files) {
Common::Path filePath = f->getPathInArchive().punycodeEncode();
debug(1, "File: %s", filePath.toString().c_str());
// skip if f represents a directory
if (filePath.isSeparatorTerminated()) continue;
Common::SeekableReadStream *stream = f->createReadStream();
uint32 len = stream->size();
if (dataSize < len) {
free(data);
data = (byte *)malloc(stream->size());
dataSize = stream->size();
}
stream->read(data, len);
Common::DumpFile out;
Common::Path outPath = destPath.join(filePath);
if (!out.open(outPath, true)) {
return Common::Error(Common::kCreatingFileFailed, "Cannot open/create dump file " + outPath.toString(Common::Path::kNativeSeparator));
} else {
uint32 writtenBytes = out.write(data, len);
if (writtenBytes < len) {
// Not all data was written
out.close();
delete stream;
free(data);
return Common::Error(Common::kWritingFailed, "Not enough storage space! Please free up some storage and try again");
}
out.flush();
out.close();
}
delete stream;
}
free(data);
return Common::kNoError;
}
char Archive::getPathSeparator() const {
return '/';
}
SeekableReadStream *MemcachingCaseInsensitiveArchive::createReadStreamForMember(const Path &path) const {
return createReadStreamForMemberImpl(path, false, Common::AltStreamType::Invalid);
}
SeekableReadStream *MemcachingCaseInsensitiveArchive::createReadStreamForMemberAltStream(const Path &path, Common::AltStreamType altStreamType) const {
// There is no situation where an invalid alt stream should be returning anything unless the implementation
// of readContentsForPathAltStream is broken, and attempting that will break the cache keying since we used Invalid
// for keying the primary stream.
if (altStreamType == Common::AltStreamType::Invalid)
return nullptr;
return createReadStreamForMemberImpl(path, true, altStreamType);
}
SeekableReadStream *MemcachingCaseInsensitiveArchive::createReadStreamForMemberImpl(const Path &path, bool isAltStream, Common::AltStreamType altStreamType) const {
CacheKey cacheKey;
cacheKey.path = translatePath(path);
cacheKey.altStreamType = isAltStream ? altStreamType : AltStreamType::Invalid;
bool isNew = false;
if (!_cache.contains(cacheKey)) {
SharedArchiveContents readResult = isAltStream ? readContentsForPathAltStream(cacheKey.path, altStreamType) : readContentsForPath(cacheKey.path);
if (readResult._bypass)
return readResult._bypass;
_cache[cacheKey] = readResult;
isNew = true;
}
SharedArchiveContents* entry = &_cache[cacheKey];
// Errors and missing files. Just return nullptr,
// no need to create stream.
if (entry->isFileMissing())
return nullptr;
// Check whether the entry is still valid as WeakPtr might have expired.
if (!entry->makeStrong()) {
// If it's expired, recreate the entry.
SharedArchiveContents readResult = isAltStream ? readContentsForPathAltStream(cacheKey.path, altStreamType) : readContentsForPath(cacheKey.path);
if (readResult._bypass)
return readResult._bypass;
_cache[cacheKey] = readResult;
entry = &_cache[cacheKey];
isNew = true;
}
// It's possible that recreation failed in case of e.g. network
// share going offline.
if (entry->isFileMissing())
return nullptr;
// Now we have a valid contents reference. Make stream for it.
Common::MemoryReadStream *memStream = new Common::MemoryReadStream(entry->getContents(), entry->getSize());
// If the entry was just created and it's too big for strong caching,
// mark the copy in cache as weak
if (isNew && entry->getSize() > _maxStronglyCachedSize) {
entry->makeWeak();
}
return memStream;
}
SharedArchiveContents MemcachingCaseInsensitiveArchive::readContentsForPathAltStream(const Path &translatedPath, AltStreamType altStreamType) const {
return SharedArchiveContents();
}
MemcachingCaseInsensitiveArchive::CacheKey::CacheKey() : altStreamType(AltStreamType::Invalid) {
}
bool MemcachingCaseInsensitiveArchive::CacheKey_EqualTo::operator()(const CacheKey &x, const CacheKey &y) const {
return (x.altStreamType == y.altStreamType) && x.path.equalsIgnoreCase(y.path);
}
uint MemcachingCaseInsensitiveArchive::CacheKey_Hash::operator()(const CacheKey &x) const {
return static_cast<uint>(x.path.hashIgnoreCase() * 1000003u) ^ static_cast<uint>(x.altStreamType);
};
SearchSet::ArchiveNodeList::iterator SearchSet::find(const String &name) {
ArchiveNodeList::iterator it = _list.begin();
for (; it != _list.end(); ++it) {
if (it->_name == name)
break;
}
return it;
}
SearchSet::ArchiveNodeList::const_iterator SearchSet::find(const String &name) const {
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
if (it->_name == name)
break;
}
return it;
}
/*
Keep the nodes sorted according to descending priorities.
In case two or node nodes have the same priority, insertion
order prevails.
*/
void SearchSet::insert(const Node &node) {
ArchiveNodeList::iterator it = _list.begin();
for (; it != _list.end(); ++it) {
if (it->_priority < node._priority)
break;
}
_list.insert(it, node);
}
void SearchSet::add(const String &name, Archive *archive, int priority, bool autoFree) {
if (find(name) == _list.end()) {
Node node(priority, name, archive, autoFree);
insert(node);
} else {
if (autoFree)
delete archive;
warning("SearchSet::add: archive '%s' already present", name.c_str());
}
}
void SearchSet::addDirectory(const String &name, const Path &directory, int priority, int depth, bool flat) {
FSNode dir(directory);
addDirectory(name, dir, priority, depth, flat);
}
void SearchSet::addDirectory(const String &name, const FSNode &dir, int priority, int depth, bool flat) {
if (!dir.exists() || !dir.isDirectory())
return;
add(name, new FSDirectory(dir, depth, flat, _ignoreClashes), priority);
}
void SearchSet::addDirectory(const Path &directory, int priority, int depth, bool flat) {
addDirectory(directory.toString(), directory, priority, depth, flat);
}
void SearchSet::addDirectory(const FSNode &directory, int priority, int depth, bool flat) {
addDirectory(directory.getPath().toString(), directory, priority, depth, flat);
}
void SearchSet::addSubDirectoriesMatching(const FSNode &directory, String origPattern, bool ignoreCase, int priority, int depth, bool flat) {
FSList subDirs;
if (!directory.getChildren(subDirs))
return;
String nextPattern, pattern;
String::const_iterator sep = Common::find(origPattern.begin(), origPattern.end(), '/');
if (sep != origPattern.end()) {
pattern = String(origPattern.begin(), sep);
++sep;
if (sep != origPattern.end())
nextPattern = String(sep, origPattern.end());
} else {
pattern = origPattern;
}
// TODO: The code we have for displaying all matches, which vary only in case, might
// be a bit overhead, but as long as we want to display all useful information to the
// user we will need to keep track of all directory names added so far. We might
// want to reconsider this though.
typedef HashMap<String, bool, IgnoreCase_Hash, IgnoreCase_EqualTo> MatchList;
MatchList multipleMatches;
MatchList::iterator matchIter;
for (FSList::const_iterator i = subDirs.begin(); i != subDirs.end(); ++i) {
String name = i->getName();
if (matchString(name.c_str(), pattern.c_str(), ignoreCase)) {
matchIter = multipleMatches.find(name);
if (matchIter == multipleMatches.end()) {
multipleMatches[name] = true;
} else {
if (matchIter->_value) {
warning("Clash in case for match of pattern \"%s\" found in directory \"%s\": \"%s\"", pattern.c_str(),
directory.getPath().toString(Common::Path::kNativeSeparator).c_str(), matchIter->_key.c_str());
matchIter->_value = false;
}
warning("Clash in case for match of pattern \"%s\" found in directory \"%s\": \"%s\"", pattern.c_str(),
directory.getPath().toString(Common::Path::kNativeSeparator).c_str(), name.c_str());
}
if (nextPattern.empty())
addDirectory(name, *i, priority, depth, flat);
else
addSubDirectoriesMatching(*i, nextPattern, ignoreCase, priority, depth, flat);
}
}
}
void SearchSet::remove(const String &name) {
ArchiveNodeList::iterator it = find(name);
if (it != _list.end()) {
if (it->_autoFree)
delete it->_arc;
_list.erase(it);
}
}
bool SearchSet::hasArchive(const String &name) const {
return (find(name) != _list.end());
}
Archive *SearchSet::getArchive(const String &name) const {
auto arch = find(name);
if (arch == _list.end())
return nullptr;
return arch->_arc;
}
void SearchSet::clear() {
for (ArchiveNodeList::iterator i = _list.begin(); i != _list.end(); ++i) {
if (i->_autoFree)
delete i->_arc;
}
_list.clear();
}
void SearchSet::setPriority(const String &name, int priority) {
ArchiveNodeList::iterator it = find(name);
if (it == _list.end()) {
warning("SearchSet::setPriority: archive '%s' is not present", name.c_str());
return;
}
if (priority == it->_priority)
return;
Node node(*it);
_list.erase(it);
node._priority = priority;
insert(node);
}
bool SearchSet::hasFile(const Path &path) const {
if (path.empty())
return false;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
if (it->_arc->hasFile(path))
return true;
}
return false;
}
bool SearchSet::isPathDirectory(const Path &path) const {
if (path.empty())
return false;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
if (it->_arc->isPathDirectory(path)) {
// See if an earlier archive contains the same path as a non-directory file.
// If this is the case, then we want to return false here because getMember will return
// that file. This is a bit faster than hasFile for each archive first.
while (it != _list.begin()) {
--it;
if (it->_arc->hasFile(path))
return false;
}
return true;
}
}
return false;
}
int SearchSet::listMatchingMembers(ArchiveMemberList &list, const Path &pattern, bool matchPathComponents) const {
int matches = 0;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it)
matches += it->_arc->listMatchingMembers(list, pattern, matchPathComponents);
return matches;
}
int SearchSet::listMatchingMembers(ArchiveMemberDetailsList &list, const Path &pattern, bool matchPathComponents) const {
int matches = 0;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
List<ArchiveMemberPtr> matchingMembers;
matches += it->_arc->listMatchingMembers(matchingMembers, pattern, matchPathComponents);
for (ArchiveMemberPtr &member : matchingMembers)
list.push_back(ArchiveMemberDetails(member, it->_name));
}
return matches;
}
int SearchSet::listMembers(ArchiveMemberList &list) const {
int matches = 0;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it)
matches += it->_arc->listMembers(list);
return matches;
}
const ArchiveMemberPtr SearchSet::getMember(const Path &path, Archive **container) const {
if (path.empty())
return ArchiveMemberPtr();
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
if (it->_arc->hasFile(path)) {
if (container) {
*container = it->_arc;
}
return it->_arc->getMember(path);
}
}
return ArchiveMemberPtr();
}
const ArchiveMemberPtr SearchSet::getMember(const Path &path) const {
return getMember(path, nullptr);
}
SeekableReadStream *SearchSet::createReadStreamForMember(const Path &path) const {
if (path.empty())
return nullptr;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
SeekableReadStream *stream = it->_arc->createReadStreamForMember(path);
if (stream)
return stream;
}
return nullptr;
}
SeekableReadStream *SearchSet::createReadStreamForMemberAltStream(const Path &path, AltStreamType altStreamType) const {
if (path.empty())
return nullptr;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it) {
SeekableReadStream *stream = it->_arc->createReadStreamForMemberAltStream(path, altStreamType);
if (stream)
return stream;
}
return nullptr;
}
SeekableReadStream *SearchSet::createReadStreamForMemberNext(const Path &path, const Archive *starting) const {
if (path.empty())
return nullptr;
ArchiveNodeList::const_iterator it = _list.begin();
for (; it != _list.end(); ++it)
if (it->_arc == starting) {
++it;
break;
}
for (; it != _list.end(); ++it) {
SeekableReadStream *stream = it->_arc->createReadStreamForMember(path);
if (stream)
return stream;
}
return nullptr;
}
SearchManager::SearchManager() {
clear(); // Force a reset
}
void SearchManager::clear() {
SearchSet::clear();
// Always keep system specific archives in the SearchManager.
// But we give them a lower priority than the default priority (which is 0),
// so that archives added by client code are searched first.
if (g_system)
g_system->addSysArchivesToSearchSet(*this, -1);
#ifndef __ANDROID__
// Add the current dir as a very last resort.
// See also bug #3984.
// But don't do this for Android platform, since it may lead to crashes
addDirectory(".", ".", -2);
#endif
}
DECLARE_SINGLETON(SearchManager);
} // namespace Common