When calling 'luaK_storevar', the 'expdesc' for the variable must be
created before the one for the expression, to satisfy the assumptions
for register allocation. So, in a statement like 'global a = exp', where
'a' is actually '_ENV.a', this variable must be handled before the
initializing expression 'exp'.
When dumping a string, adding 2 to its size may overflow a size_t for
external strings, which may not have a header. (Adding 1 is Ok, because
all strings end with a '\0' not included in their size.) The new method
for saving NULL strings code them as a repeated string, using the
reserved index 0.
The check for limit of local variables is made after generating code to
initialize them. If there are too many local variables not initialized,
the coding of instruction OP_LOADNIL could overflow an argument.
A vararg table can be virtual. If the vararg table is used only as
a base in indexing expressions, the code does not need to create an
actual table for it. Instead, it compiles the indexing expressions
into direct accesses to the internal vararg data.
- LUAMOD_API defined as 'extern "C"' in C++.
- "ANSI C" is in fact "ISO C" (comments)
- Removed option -std from makefile in testes/libs. (Easier to change
to C++ for tests).
In C, we may have several "setjmp" nested, and the "longjmp" will go
to the one given by the corresponding "jmp_buf". In C++, a "throw"
will always go to the inner "catch". So, the "catch" must check
whether it is the recipient of the "throw" and, if not, rethrow
the exception to the outer level.
An external definition for LUA_32BITS can change the API, but libraries
check number-format compatibility when loading. So, any incompatible
modules will report a clear error.
These definitions were in luaconf.h only because the standard libraries
need them. Now that llimits.h is included by the libraries, it offers a
more private place for these definitions.
When both 'int' and 'l_obj' have 32 bits, an unsigned int needs a
cast to be assigned to 'l_obj'. (As long as 'l_obj' can count the
total memory used by the system, these casts should be safe.)
The cast of n (number of repetitions) to size_t may truncate its value,
causing a buffer overflow later. Better to check the buffer size
using lua_Integer, as all string lengths must fit in a lua_Integer and
n already is a lua_Integer. If everything fits in MAX_SIZE, then we can
safely convert n to size_t and compute the buffer size as a size_t.
As a corner case, n can be larger than size_t if the strings being
repeated have length zero, but in this case it will be multiplied by
zero, so an overflow in the cast is irrelevant.
A bad actor could fill only a few entries in a table (power of twos in
decreasing order, see tests) and produce a small table with a huge
length. If your program builds a table with external data and iterates
over its length, this behavior could be an issue.
That complicates a little object equality (and therefore table access
for long strings), but the old behavior was somewhat weird. (Short
strings, a concept otherwise absent from the manual, could not be
external.)
External strings created by DLLs may need the DLL code to be
deallocated. This implies that a DLL can only be unloaded after all
its strings were deallocated, which happen only after the run of all
finalizers. To ensure that order, we create a 'library string' to
represent each DLL and keep it locked. When this string is deallocated
(after the deallocation of any string created by the DLL) it closes its
corresponding DLL.
LUAI_MAXSTACK is limited to INT_MAX/2, so can use INT_MAX/2 to define
pseudo-indices (LUA_REGISTRYINDEX) in 'lua.h'. A change in the maximum
stack size does not need to change the Lua-C ABI.
All-weak tables are not being revisited after being visited during
propagation; if it gets a new metatable after that, the new metatable
may not be marked.
No thread started with pcall (instead of resume) can be closed,
because coroutine.close would not respect the expected number of
results from the protected call.
In a constructor, each field generates at least one opcode, and the
number of opcodes is limited by INT_MAX. Therefore, the counters for
number of fields cannot exceed this limit. (The current limit for
items in the hash part of a table has a limit smaller than INT_MAX.
However, as long as there are no overflows, the logic for table
resizing will handle that limit.)
A goto cannot jump into the scope of any variable declaration,
including 'global *'. To report the error, it needs a "name" for
the scope it is entering.
The parser uses "break" as fake label to compile "break" as "goto
break". To avoid producing this string at each use, it keeps it
available in its state.
In generational collection, objects marked as touched1 stay in gray
lists between collections. This commit fixes a bug introduced in
commit 808976bb59.
It can be a little slower, but only for quite large stacks and moreover
stack reallocation is not a common operation. With no strong contrary
reason, it is better to follow the standard.
When computing the Mersenne number, instead of spreading 1's a fixed
number of times (with shifts of 1, 2, 4, 8, 16, and 32), spread only
until the number becomes a Mersenne number.
To avoid complains from some tools, the addition when computing
math.random(n,m), which is computed as n + random(0, m - n), should
use unsigned integers.
New function 'resetCI' resets the CallInfo list of a thread, ensuring
a proper state when creating a new thread, closing a thread, or
closing a state, so that we can run code after that. (When closing a
thread, we need to run its __close metamethods; when closing a
state, we need to run its __close metamethods and its finalizers.)
If a metatable is a weak table, its __newindex field could be collected
by an emergency collection while being used in 'luaV_finishset'. (This
bug has similarities with bug 5.3.2-1, fixed in commit a272fa66.)
This function can be called unprotected, so it should not raise any
kind of errors. (It could raise a memory-allocation error when creating
a message).
Tests for memory-allocation errors moved from 'api.lua' to this new
file, as 'api.lua' was already too big. (Besides, these tests have
nothing to do with the API.)
After a weak table is traversed in the atomic phase, if it does not
have white values ('hasclears') it does not need to be retraversed
again. (Comments were correct, but code did not agree with them.)
Roberto I