mirror of
https://github.com/WerWolv/ImHex.git
synced 2026-03-28 07:47:03 -05:00
impr: Use faster way of finding matching pattern files for loaded file
This commit is contained in:
WerWolv
2
lib/external/pattern_language
vendored
2
lib/external/pattern_language
vendored
Submodule lib/external/pattern_language updated: 7e00014c35...aa2ae12638
@@ -1505,100 +1505,6 @@ namespace hex::plugin::builtin {
|
||||
bool foundCorrectType = false;
|
||||
|
||||
auto mimeType = magic::getMIMEType(provider, 0, 4_KiB, true);
|
||||
runtime.addPragma("MIME", [&mimeType, &foundCorrectType](const pl::PatternLanguage &runtime, const std::string &value) {
|
||||
std::ignore = runtime;
|
||||
|
||||
if (!magic::isValidMIMEType(value))
|
||||
return false;
|
||||
|
||||
if (value == mimeType) {
|
||||
foundCorrectType = true;
|
||||
return true;
|
||||
}
|
||||
return !std::ranges::all_of(value, isspace) && !value.ends_with('\n') && !value.ends_with('\r');
|
||||
});
|
||||
|
||||
// Format: [ AA BB CC DD ] @ 0x12345678
|
||||
runtime.addPragma("magic", [provider, &foundCorrectType](pl::PatternLanguage &, const std::string &value) -> bool {
|
||||
const auto pattern = [value = value]() mutable -> std::optional<BinaryPattern> {
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
if (value.empty())
|
||||
return std::nullopt;
|
||||
|
||||
if (!value.starts_with('['))
|
||||
return std::nullopt;
|
||||
|
||||
value = value.substr(1);
|
||||
|
||||
const auto end = value.find(']');
|
||||
if (end == std::string::npos)
|
||||
return std::nullopt;
|
||||
value.resize(end);
|
||||
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
return BinaryPattern(value);
|
||||
}();
|
||||
|
||||
const auto address = [value = value, provider]() mutable -> std::optional<u64> {
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
if (value.empty())
|
||||
return std::nullopt;
|
||||
|
||||
const auto start = value.find('@');
|
||||
if (start == std::string::npos)
|
||||
return std::nullopt;
|
||||
|
||||
value = value.substr(start + 1);
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
size_t end = 0;
|
||||
auto result = std::stoll(value, &end, 0);
|
||||
if (end != value.length())
|
||||
return std::nullopt;
|
||||
|
||||
if (result < 0) {
|
||||
const auto size = provider->getActualSize();
|
||||
if (u64(-result) > size) {
|
||||
return std::nullopt;
|
||||
}
|
||||
|
||||
return size + result;
|
||||
} else {
|
||||
return result;
|
||||
}
|
||||
}();
|
||||
|
||||
if (!address)
|
||||
return false;
|
||||
if (!pattern)
|
||||
return false;
|
||||
|
||||
std::vector<u8> bytes(pattern->getSize());
|
||||
if (bytes.empty())
|
||||
return false;
|
||||
|
||||
provider->read(*address, bytes.data(), bytes.size());
|
||||
|
||||
if (pattern->matches(bytes))
|
||||
foundCorrectType = true;
|
||||
|
||||
return true;
|
||||
});
|
||||
|
||||
std::string author;
|
||||
runtime.addPragma("author", [&author](pl::PatternLanguage &, const std::string &value) -> bool {
|
||||
author = value;
|
||||
return true;
|
||||
});
|
||||
|
||||
std::string description;
|
||||
runtime.addPragma("description", [&description](pl::PatternLanguage &, const std::string &value) -> bool {
|
||||
description = value;
|
||||
return true;
|
||||
});
|
||||
|
||||
m_possiblePatternFiles.get(provider).clear();
|
||||
|
||||
@@ -1616,12 +1522,81 @@ namespace hex::plugin::builtin {
|
||||
if (!file.isValid())
|
||||
continue;
|
||||
|
||||
author.clear();
|
||||
description.clear();
|
||||
std::string author, description;
|
||||
|
||||
auto result = runtime.preprocessString(file.readString(), pl::api::Source::DefaultSource);
|
||||
if (!result.has_value()) {
|
||||
log::warn("Failed to preprocess file {} during MIME analysis", entry.path().string());
|
||||
const auto pragmaValues = runtime.getPragmaValues(file.readString());
|
||||
if (auto it = pragmaValues.find("author"); it != pragmaValues.end())
|
||||
author = it->second;
|
||||
if (auto it = pragmaValues.find("description"); it != pragmaValues.end())
|
||||
description = it->second;
|
||||
|
||||
// Format: #pragma MIME type/subtype
|
||||
if (auto it = pragmaValues.find("MIME"); it != pragmaValues.end()) {
|
||||
if (magic::isValidMIMEType(it->second) && it->second == mimeType)
|
||||
foundCorrectType = true;
|
||||
}
|
||||
// Format: #pragma magic [ AA BB CC DD ] @ 0x12345678
|
||||
if (auto it = pragmaValues.find("magic"); it != pragmaValues.end()) {
|
||||
const auto pattern = [value = it->second]() mutable -> std::optional<BinaryPattern> {
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
if (value.empty())
|
||||
return std::nullopt;
|
||||
|
||||
if (!value.starts_with('['))
|
||||
return std::nullopt;
|
||||
|
||||
value = value.substr(1);
|
||||
|
||||
const auto end = value.find(']');
|
||||
if (end == std::string::npos)
|
||||
return std::nullopt;
|
||||
value.resize(end);
|
||||
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
return BinaryPattern(value);
|
||||
}();
|
||||
|
||||
const auto address = [value = it->second, provider]() mutable -> std::optional<u64> {
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
if (value.empty())
|
||||
return std::nullopt;
|
||||
|
||||
const auto start = value.find('@');
|
||||
if (start == std::string::npos)
|
||||
return std::nullopt;
|
||||
|
||||
value = value.substr(start + 1);
|
||||
value = wolv::util::trim(value);
|
||||
|
||||
size_t end = 0;
|
||||
auto result = std::stoll(value, &end, 0);
|
||||
if (end != value.length())
|
||||
return std::nullopt;
|
||||
|
||||
if (result < 0) {
|
||||
const auto size = provider->getActualSize();
|
||||
if (u64(-result) > size) {
|
||||
return std::nullopt;
|
||||
}
|
||||
|
||||
return size + result;
|
||||
} else {
|
||||
return result;
|
||||
}
|
||||
}();
|
||||
|
||||
if (address && pattern) {
|
||||
std::vector<u8> bytes(pattern->getSize());
|
||||
if (!bytes.empty()) {
|
||||
provider->read(*address, bytes.data(), bytes.size());
|
||||
|
||||
if (pattern->matches(bytes))
|
||||
foundCorrectType = true;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (foundCorrectType) {
|
||||
@@ -2760,35 +2735,14 @@ namespace hex::plugin::builtin {
|
||||
}
|
||||
|
||||
const auto fileName = wolv::util::toUTF8String(adjustedPath.filename());
|
||||
m_patternNames[path] = fileName;
|
||||
|
||||
wolv::io::File file(path, wolv::io::File::Mode::Read);
|
||||
pl::api::Source source(file.readString());
|
||||
|
||||
// Only run the lexer on the source file and manually extract the #pragma description to make this
|
||||
// process as fast as possible. Running the preprocessor directly takes too much time
|
||||
auto result = runtime->getInternals().lexer->lex(&source);
|
||||
if (result.isOk()) {
|
||||
const auto tokens = result.unwrap();
|
||||
for (auto it = tokens.begin(); it != tokens.end(); ++it) {
|
||||
if (it->type == pl::core::Token::Type::Directive && std::get<pl::core::Token::Directive>(it->value) == pl::core::Token::Directive::Pragma) {
|
||||
++it;
|
||||
if (it != tokens.end() && it->type == pl::core::Token::Type::String) {
|
||||
auto literal = std::get<pl::core::Token::Literal>(it->value);
|
||||
auto string = std::get_if<std::string>(&literal);
|
||||
if (string != nullptr && *string == "description") {
|
||||
++it;
|
||||
if (it != tokens.end() && it->type == pl::core::Token::Type::String) {
|
||||
literal = std::get<pl::core::Token::Literal>(it->value);
|
||||
string = std::get_if<std::string>(&literal);
|
||||
if (string != nullptr) {
|
||||
m_patternNames[path] = fmt::format("{} ({})", *string, fileName);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
const auto pragmaValues = runtime->getPragmaValues(file.readString());
|
||||
if (auto it = pragmaValues.find("description"); it != pragmaValues.end() && !it->second.empty()) {
|
||||
m_patternNames[path] = fmt::format("{} ({})", it->second, fileName);
|
||||
} else {
|
||||
m_patternNames[path] = fileName;
|
||||
}
|
||||
|
||||
return m_patternNames[path];
|
||||
|
||||
Reference in New Issue
Block a user