diff options
Diffstat (limited to 'indra/llcommon')
| -rw-r--r-- | indra/llcommon/CMakeLists.txt | 1 | ||||
| -rw-r--r-- | indra/llcommon/fsyspath.h | 74 | ||||
| -rw-r--r-- | indra/llcommon/lua_function.cpp | 6 | ||||
| -rw-r--r-- | indra/llcommon/lua_function.h | 4 |
4 files changed, 80 insertions, 5 deletions
diff --git a/indra/llcommon/CMakeLists.txt b/indra/llcommon/CMakeLists.txt index aed9ee080b..aa0b66f2f4 100644 --- a/indra/llcommon/CMakeLists.txt +++ b/indra/llcommon/CMakeLists.txt @@ -127,6 +127,7 @@ set(llcommon_HEADER_FILES commoncontrol.h ctype_workaround.h fix_macros.h + fsyspath.h function_types.h indra_constants.h lazyeventapi.h diff --git a/indra/llcommon/fsyspath.h b/indra/llcommon/fsyspath.h new file mode 100644 index 0000000000..aa4e0132bc --- /dev/null +++ b/indra/llcommon/fsyspath.h @@ -0,0 +1,74 @@ +/** + * @file fsyspath.h + * @author Nat Goodspeed + * @date 2024-04-03 + * @brief Adapt our UTF-8 std::strings for std::filesystem::path + * + * $LicenseInfo:firstyear=2024&license=viewerlgpl$ + * Copyright (c) 2024, Linden Research, Inc. + * $/LicenseInfo$ + */ + +#if ! defined(LL_FSYSPATH_H) +#define LL_FSYSPATH_H + +#include <filesystem> + +// While std::filesystem::path can be directly constructed from std::string on +// both Posix and Windows, that's not what we want on Windows. Per +// https://en.cppreference.com/w/cpp/filesystem/path/path: + +// ... the method of conversion to the native character set depends on the +// character type used by source. +// +// * If the source character type is char, the encoding of the source is +// assumed to be the native narrow encoding (so no conversion takes place on +// POSIX systems). +// * If the source character type is char8_t, conversion from UTF-8 to native +// filesystem encoding is used. (since C++20) +// * If the source character type is wchar_t, the input is assumed to be the +// native wide encoding (so no conversion takes places on Windows). + +// The trouble is that on Windows, from std::string ("source character type is +// char"), the "native narrow encoding" isn't UTF-8, so file paths containing +// non-ASCII characters get mangled. +// +// Once we're building with C++20, we could pass a UTF-8 std::string through a +// vector<char8_t> to engage std::filesystem::path's own UTF-8 conversion. But +// sigh, as of 2024-04-03 we're not yet there. +// +// Anyway, encapsulating the important UTF-8 conversions in our own subclass +// allows us to migrate forward to C++20 conventions without changing +// referencing code. + +class fsyspath: public std::filesystem::path +{ + using super = std::filesystem::path; + +public: + // default + fsyspath() {} + // construct from UTF-8 encoded std::string + fsyspath(const std::string& path): super(std::filesystem::u8path(path)) {} + // construct from UTF-8 encoded const char* + fsyspath(const char* path): super(std::filesystem::u8path(path)) {} + // construct from existing path + fsyspath(const super& path): super(path) {} + + fsyspath& operator=(const super& p) { super::operator=(p); return *this; } + fsyspath& operator=(const std::string& p) + { + super::operator=(std::filesystem::u8path(p)); + return *this; + } + fsyspath& operator=(const char* p) + { + super::operator=(std::filesystem::u8path(p)); + return *this; + } + + // shadow base-class string() method with UTF-8 aware method + std::string string() const { return super::u8string(); } +}; + +#endif /* ! defined(LL_FSYSPATH_H) */ diff --git a/indra/llcommon/lua_function.cpp b/indra/llcommon/lua_function.cpp index cd83f40e85..7a5668f384 100644 --- a/indra/llcommon/lua_function.cpp +++ b/indra/llcommon/lua_function.cpp @@ -17,13 +17,13 @@ // std headers #include <algorithm> #include <exception> -#include <filesystem> #include <iomanip> // std::quoted #include <map> #include <memory> // std::unique_ptr #include <typeinfo> // external library headers // other Linden headers +#include "fsyspath.h" #include "hexdump.h" #include "lleventcoro.h" #include "llsd.h" @@ -68,7 +68,7 @@ int lluau::loadstring(lua_State *L, const std::string &desc, const std::string & return luau_load(L, desc.data(), bytecode.get(), bytecodeSize, 0); } -std::filesystem::path lluau::source_path(lua_State* L) +fsyspath lluau::source_path(lua_State* L) { //Luau lua_Debug and lua_getinfo() are different compared to default Lua: //see https://github.com/luau-lang/luau/blob/80928acb92d1e4b6db16bada6d21b1fb6fa66265/VM/include/lua.h @@ -577,7 +577,7 @@ std::pair<int, LLSD> LuaState::expr(const std::string& desc, const std::string& // the next call to lua_next." // https://www.lua.org/manual/5.1/manual.html#lua_next if (lua_type(mState, -2) == LUA_TSTRING && - std::filesystem::path(lua_tostdstring(mState, -2)).stem() == "fiber") + fsyspath(lua_tostdstring(mState, -2)).stem() == "fiber") { found = true; break; diff --git a/indra/llcommon/lua_function.h b/indra/llcommon/lua_function.h index 785aadeb0c..ec1e6cdb10 100644 --- a/indra/llcommon/lua_function.h +++ b/indra/llcommon/lua_function.h @@ -16,9 +16,9 @@ #include "luau/lua.h" #include "luau/luaconf.h" #include "luau/lualib.h" +#include "fsyspath.h" #include "stringize.h" #include <exception> // std::uncaught_exceptions() -#include <filesystem> #include <memory> // std::shared_ptr #include <utility> // std::pair @@ -51,7 +51,7 @@ namespace lluau int dostring(lua_State* L, const std::string& desc, const std::string& text); int loadstring(lua_State* L, const std::string& desc, const std::string& text); - std::filesystem::path source_path(lua_State* L); + fsyspath source_path(lua_State* L); } // namespace lluau std::string lua_tostdstring(lua_State* L, int index); |