| Age | Commit message (Collapse) | Author |
|---|
|
Introduce `ScriptCommand` class that parses a command line into a script name
and optional args, using bash-like quoting and escaping. `ScriptCommand`
searches for a file with that script name on a passed list of directories; the
directories may be specified relative to a particular base directory.
`ScriptCommand` supports the special case of a script name containing unescaped
spaces. It guarantees that either the returned script file exists, or its
`error()` string is non-empty.
Replace `LLLeap::create()` logic, from which `ScriptCommand` was partly
derived, with a `ScriptCommand` instance.
Make `LLLUAmanager::runScriptFile()` use a `ScriptCommand` instance to parse
the passed command line.
Subsume `LLAppViewer::init()` script-path-searching logic for `--luafile`
into `ScriptCommand`. In fact that lambda now simply calls
`LLLUAmanager::runScriptFile()`.
Make `lluau::dostring()` accept an optional vector of script argument strings.
Following PUC-Rio Lua convention, pass these arguments into a Lua script as
the predefined global `arg`, and also as the script's `...` argument.
`LuaState::expr()` also accepts and passes through script argument strings.
Change the log tag for the Lua script interruption message: if we want it, we
can still enable it, but we don't necessarily want it along with all other
"Lua" DEBUG messages.
Remove `LuaState::script_finished_fn`, which isn't used any more. Also remove
the corresponding `LLLUAmanager::script_finished_fn`. This allows us to
simplify `~LuaState()` slightly, as well as the parameter signatures for
`LLLUAmanager::runScriptFile()` and `runScriptLine()`.
|
|
|
|
If the C++ runtime is already handling an exception, don't try to launch more
Lua operations.
|
|
Consensus seems to be that (a) string_view is, in effect, already a reference,
(b) it's small enough to make pass-by-value reasonable and (c) the optimizer
can reason about values way better than it can about references.
|
|
Make central Lua engine functionality conditional on that flag.
|
|
|
|
|
|
`setdtor('description', object, function)` returns a proxy userdata object
referencing object and function. When the proxy is garbage-collected, or at
the end of the script, its destructor calls `function(object)`.
The original object may be retrieved as `proxy._target`, e.g. to pass it to
the `table` library. The proxy also has a metatable with metamethods
supporting arithmetic operations, string concatenation, length and table
indexing. For other operations, retrieve `proxy._target`. (But don't assign to
`proxy._target`. It will appear to work, in that subsequent references to
`proxy._target` will retrieve the replacement object -- however, the
destructor will still call `function(original object)`.)
Fix bugs in `lua_setfieldv()`, `lua_rawgetfield()` and `lua_rawsetfield()`.
Add C++ functions `lua_destroyuserdata()` to explicitly destroy a
`lua_emplace<T>()` userdata object, plus `lua_destroybounduserdata()`. The
latter can bind such a userdata object as an upvalue to pass to `LL.atexit()`.
Make `LL.help()` and `LL.leaphelp()` help text include the `LL.` prefix.
|
|
Use a static unordered_map to allow a function receiving (lua_State* L) to
look up the LuaState instance managing that lua_State. We've thought about
this from time to time already. LuaState's constructor creates the map entry;
its destructor removes it; the new static getParent(lua_State* L) method
performs the lookup.
Migrate lluau::set_interrupts_counter() and check_interrupts_counter() into
LuaState member functions. Add a new mInterrupts counter for them.
Importantly, LuaState::check_interrupts_counter(), which is indirectly called
by a lua_callbacks().interrupt function, no longer performs any Lua stack
operations. Empirically, it seems the Lua engine is capable of interrupting
itself at a moment when re-entry confuses it.
Change previous lluau::set_interrupts_counter(L, 0) calls to
LuaState::getParent(L).set_interrupts_counter(0).
Also add LuaStackDelta class, and a lua_checkdelta() helper macro, to verify
that the Lua data stack depth on exit from a block differs from the depth on
entry by exactly the expected amount. Sprinkle lua_checkdelta() macros in
likely places.
|
|
luaL_checkstack() accepts a third parameter which is included in the stack
overflow error message. We've been passing nullptr, leading to messages of the
form "stack overflow ((null))". lluau_checkstack() implicitly passes
__FUNCTION__, so we can distinguish which underlying luaL_checkstack() call
encountered the stack overflow condition.
Also, when calling each atexit() function, pass Luau's debug.traceback()
function as the lua_pcall() error handler. This should help diagnose errors in
atexit() functions.
|
|
Leverage C++ overloads to allow use of generic function names disambiguated by
argument type.
This allows using templates for certain common operation sequences.
|
|
Remove LLLUAmanager::mumbleScriptLine() LuaState& parameters. Make
startScriptLine(), waitScriptLine() and runScriptLine() exactly parallel to
startScriptFile(), waitScriptFile() and runScriptFile(). That means that
runScriptLine()'s C++ coroutine instantiates and destroys its own LuaState,
which means that LL.atexit() functions will run on the Lua-specific C++
coroutine rather than (say) the viewer's main coroutine.
Introduce LLLUAmanager::script_result typedef for std::pair<int, LLSD> and use
in method returns.
Remove LuaState::initLuaState(); move its logic back into the constructor.
Remove initLuaState() calls in the expr() error cases: they're moot now that
we won't get subsequent expr() calls on the same LuaState instance.
Remove LLFloaterLUADebug "Use clean lua_State" checkbox and the cleanLuaState()
method. Remove mState member.
Remove explicit LuaState declarations from LLLUAmanager tests. Adapt one test
for implicit LuaState: it was directly calling LuaState::obtainListener() to
discover the LuaListener's reply-pump name. But since that test also captures
two leap.request() calls from the Lua script, it can just look at the "reply"
key in either of those requests.
|
|
|
|
This replaces type_tag<T>(), which searched and possibly extended the type_tags
unordered_map at runtime. If we called lua_emplace<T>() from different threads,
that would require locking type_tags.
In contrast, the compiler must instantiate a distinct TypeTag<T> for every
distinct T passed to lua_emplace<T>(), so each gets a distinct value at static
initialization time. No locking is required; no lookup; no allocations.
Add a test to llluamanager_test.cpp to verify that each distinct T passed to
lua_emplace<T>() gets its own TypeTag<T>::value, and that each gets its own
destructor -- but that different lua_emplace<T>() calls with the same T share
the same TypeTag<T>::value and the same destructor.
|
|
It turns out that Luau does not honor PUC-Rio Lua's __gc metafunction, so
despite elaborate measures, the previous lua_emplace<T>() implementation would
not have destroyed the contained C++ T object when the resulting userdata
object was garbage-collected.
Moreover, using LL.atexit() as the mechanism to destroy lua_emplace<T>()
userdata objects (e.g. LuaListener) would have been slightly fragile because
we also want to use LL.atexit() to make the final fiber.run() call, when
appropriate. Introducing an order dependency between fiber.run() and the
LuaListener destructor would not be robust.
Both of those problems are addressed by leveraging one of Luau's extensions
over PUC-Rio Lua. A Luau userdata object can have an int tag; and a tag can
have an associated C++ destructor function. When any userdata object bearing
that tag is garbage-collected, Luau will call that destructor; and Luau's
lua_close() function destroys all userdata objects.
The resulting lua_emplace<T>() and lua_toclass<T>() code is far simpler.
It only remains to generate a distinct int tag value for each different C++
type passed to the lua_emplace<T>() template.
unordered_map<std::type_index, int> addresses that need.
|
|
We might decide to leave some of them in place.
|
|
lua_emplace<T>() was passing LL.atexit() a closure binding the new userdata
with a cleanup function. The trouble with that was that a strong reference to
the new userdata would prevent it ever being garbage collected, even if that
was the only remaining reference.
Instead, create a new weak table referencing the userdata, and bind that into
the cleanup function's closure. Then if the only remaining reference to the
userdata is from the weak table, the userdata can be collected.
Make lua_emplace_call_gc<T>() check the bound weak table in case the userdata
has in fact been collected.
Also, in lua_toclass<T>(), use luaL_checkudata() to synopsize comparing the
putative userdata's metatable against the one synthesized by lua_emplace<T>().
This saves several explicit steps.
|
|
Instead of deriving LuaListener from LLInstanceTracker with an int key,
generating a unique int key and storing that key in the Registry, use new
lua_emplace<LuaState>() to store the LuaListener directly in a Lua userdata
object in the Lua Registry.
Because lua_emplace<T>() uses LL.atexit() to guarantee that ~LuaState will
destroy the T object, we no longer need ~LuaState() to make a special call
specifically to destroy the LuaListener, if any. So we no longer need
LuaState::getListener() separate from obtainListener().
Since LuaListener is no longer an LLInstanceTracker subclass, make
LuaState::obtainListener() return LuaListener& rather than LuaListener::ptr_t.
|
|
Publish new LL.atexit() function that accepts a Lua function (or C++ closure)
and saves it (in Registry["atexit"] table) to call later.
Make ~LuaState() walk the Registry["atexit"] table, if it exists, calling each
function appended to that table.
(Consider using that mechanism to clean up a LuaListener, if one was
instantiated. Possibly also use for p.s. leap.run()? But that's run after
every expr() call, instead of only at ~LuaState() time. Pragmatically, though,
the distinction only matters for a LUA Debug Console LUA string with "clean
lua_State" unchecked.)
For use by future lua_function() entry points, lua_emplace<T>(ctor args...)
pushes a Lua userdata object containing a newly-constructed T instance --
actually a std::optional<T> to avoid double destruction. lua_emplace<T>() is
specifically intended to be usable even for T with a nontrivial destructor: it
gives the userdata a metatable with a __gc function that destroys the
contained T instance when the userdata is garbage collected. But since garbage
collection doesn't guarantee to clean up global variables with __gc methods,
lua_emplace<T>() also uses LL.atexit() to ensure that ~T() will run when the
LuaState is destroyed.
The companion to lua_emplace<T>() is lua_toclass<T>(), which returns a
non-nullptr T* if the referenced index is in fact a userdata created by
lua_emplace<T>() for the same T, that has not yet been destroyed. This lets
C++ code access a T previously embedded in Lua userdata.
|
|
Use in LuaState::expr() so we can catch a runaway in-memory Lua chunk as well
as a script read from a file.
|
|
Our std::strings are UTF-8 encoded, so conversion from std::string to
std::filesystem::path must use UTF-8 decoding. The native Windows
std::filesystem::path constructor and assignment operator accepting
std::string use "native narrow encoding," which mangles path strings
containing UTF-8 encoded non-ASCII characters.
fsyspath's std::string constructor and assignment operator explicitly engage
std::filesystem::u8path() to handle encoding. u8path() is deprecated in C++20,
but once we adapt fsyspath's conversion to C++20 conventions, consuming code
need not be modified.
|
|
This helps a Lua script log its own identity, or find associated files
relative to its location in the filesystem.
Add more comprehensive logging around the start and end of a given Lua script,
or its "p.s." fiber.run() call.
|
|
|
|
Push throwing Lua errors down into LLRequireResolver::findModule() and
findModuleImpl() so their callers don't have to handle the error case. That
eliminates finishrequire().
require() itself now only retrieves (and pops) the passed module name and
calls LLRequireResolver::resolveRequire() to do the actual work.
resolveRequire() is now void. It only instantiates LLRequireResolver and calls
its findModule().
findModule() is now also void. It's guaranteed to either push the loaded Lua
module or throw a Lua error. In particular, when findPathImpl() cannot find
the specified module, findModule() throws an error. That replaces
ModuleStatus::NotFound.
Since std::filesystem::path::append() aka operator/() detects when its right
operand is absolute and, in that case, discards the left operand, we no longer
need resolveAndStoreDefaultPaths(): we can just invoke that operation inline.
When findModule() pushes _MODULES on the Lua stack, it uses LuaRemover (below)
to ensure that _MODULES is removed again no matter how findModules() exits.
findModuleImpl() now accepts the candidate pathname as its argument. That
eliminates mAbsolutePath.
findModuleImpl() now returns only bool: true means the module was found and
loaded and pushed on the Lua stack, false means not found and nothing was
pushed; no return means an error was reported.
Push running a newly found module's source file down into findModuleImpl().
That eliminates the distinction between Cached and FileRead, which obviates
ModuleStatus: a bool return means either "previously cached" or "we read it,
compiled it, loaded it and ran it." That also eliminates the need to store the
module's textual content in mSourceCode.
Similarly, once loading the module succeeds, findModuleImpl() caches it in
_MODULES right away. That eliminates ResolvedRequire since we need not pass
the full pathname of the found module (or its contents) back up through the
call chain.
Move require() code that runs the new module into private runModule() method,
called by findModuleImpl() in the not-cached case. runModule() is the only
remaining method that can push either a string error message or the desired
module, because of its funny stack manipulations. That means the check for a
string error message on the stack top can move down to findModuleImpl().
Add LuaRemover class to ensure that on exit from some particular C++ block,
the specified Lua stack entry will definitely be removed. This is different
from LuaPopper in that it engages lua_remove() rather than lua_pop().
Also ditch obsolete await_event() Lua entry point.
|
|
|
|
|
|
|
|
|
|
|
|
leaphelp() (no argument) shows a list of all LEAP APIs.
leaphelp(API) shows further help for a specific API.
Both forms query LuaListener's LeapListener and report its responses. In
future we might reimplement leaphelp() as a Lua function.
Add LuaState::getListener() method, which checks whether there's a LuaListener
associated with this LuaState and returns a pointer if so.
Add LuaState::obtainListener() method, which finds or creates a LuaListener
for this LuaState and returns its pointer.
Both the above use logic migrated from the Lua listen_events() entry point,
which now calls obtainListener() instead.
|
|
help() with no argument lists all our viewer builtins.
help(function, function, ...) shows help text for each named function. Each
argument can be either a string or the function in question (e.g. help(help)).
To support Lua-related text containing line breaks, make LLTextEditor::
pasteTextWithLinebreaks() a public template method. Change the existing
implementation, which specifically accepts (const LLWString&), into its
LLWString specialization. The generic template passes llconvert(arg) to that
specialization, the one real implementation.
Make LLFloaterLUADebug methods call pasteTextWithLinebreaks() instead of
insertText(), which ignores newline characters.
To allow help() to accept an actual function as well as a string name, add a
lookup-by-function-pointer map to LuaFunction. (A Lua function does not store
a name.) Make the constructor store an entry in the new lookup map as well as
in the original registry map.
Change LuaFunction::getRegistry() and getRegistered() to getState() and
getRState(), respectively. Each returns a std::pair, but the first binds
non-const references while the second binds const references.
|
|
|
|
We add a suffix to let us publish a Lua foo() function that wraps a C++ foo()
function. Of course the lua_CFunction must accept lua_State* and extract its
parameters from the Lua stack, so it must invoke different C++ code than the
C++ foo() function it's trying to reach. So the lua_CFunction is a method of
the LuaFunction subclass instance named foo_lua.
The suffix was _luadecl, but since the class name shows up in log messages,
make it the more streamlined _lua instead.
|
|
Extend the LuaFunction::Registry map to store helptext as well as the function
pointer.
Add help text to every existing lua_function() invocation.
|
|
macOS clang produces fatal warnings when trying to pass a const char*
parameter to luaL_error() (-Wformat-security). Temporarily suppressing that
requires #pragma clang directives which, in turn, produce fatal warnings in
VS.
Moreover, VS recognizes that luaL_error() never returns, and so diagnoses the
following return statement as unreachable code. But that return statement is
the whole reason for lluau::error()'s existence...
|
|
Add LuaState::expr() that evaluates a Lua snippet and reports back any result
(or error) left on the stack.
Add LLLUAmanager::runScriptFile() and runScriptLine() overloads that accept a
callback with an (int count, LLSD result) signature. The count disambiguates
(error, no result, one result, array of results). Also add overloads that accept
an existing LuaState instance. Also add waitScriptFile() and waitScriptLine()
methods that pause the calling coroutine until the Lua script completes, and
return its results.
Instead of giving LuaState a description to use for all subsequent checkLua()
calls, remove description from its constructor and data members. Move to
expr() and checkLua() parameters: we want a description specific to each
operation, rather than for the LuaState as a whole. This prepares for
persistent LuaState instances.
For now, the existing script_finished_fn semantics remain: the callback will
be called only when the LuaState is destroyed. This may need to change as we
migrate towards longer-lasting LuaState instances.
Make lua_function(name) macro append suffixes to the name for both the
LuaFunction subclass declaration and the instance declaration. This allows
publishing a lua_function() name such as sleep(), which already has a
different C++ declaration.
Move the Lua sleep() entry point to a standalone lua_function(sleep), instead
of a lambda in the body of runScriptFile().
|
|
The intention is to decentralize Luau entry points into our C++ code,
permitting a given entry point to be added to the .cpp file that already deals
with that class or functional area. Continuing to add every such entry point
to llluamanager.cpp doesn't scale well.
Extract LuaListener class from llluamanager.cpp to its own header and .cpp
file.
Extract from llluamanager into lua_function.h (and .cpp) declarations useful
for adding a lua_function Luau entry point, e.g.:
lua_register()
lua_rawlen()
lua_tostdstring()
lua_pushstdstring()
lua_tollsd()
lua_pushllsd()
LuaPopper
lua_function() and LuaFunction class
LuaState
lua_what
lua_stack
DebugExit
|
