Now with *almost* working EH codegen. Does not compile Phobos yet
because we run into the "instruction does not dominate all uses"
issue when an r-value result of toElemDtor is used and we need to
run cleanups in between. Should easily be fixed by promoting those
values to allocas.
Most of the changes outside of ir/irfunction.{h, cpp} are just
because CreateCallOrInvoke moved locations. I took the
opportunity to also make use of the different arg count
overloads where possible.
Never generates any landing pads or invoke instructions right now
for simplicity. The code for emitting them will be added back in
the next step.
The "after..." blocks without any precedessors remain for now, as
we need a clean way to suppress any codegen for that block (but
not new blocks, which might resolve labels) before tackling that
one.
Builds druntime/Phobos on OS X x86_64 (albeit without EH, of course).
Specifying the basic block before which to insert the new one
is not mandatory when calling llvm::BasicBlock::Create. This
was the only use of the tracked "end" block. The concept was
phony anyway because there is no single "end" to a scope with
unwinding and so on.
For prettying up the IR, it is possible to change the order
of basic blocks using move{Before, After}().
This fixes structs with members of structs with real fields. Did not
fail the test suite unless building with assertions before, but
crashed building real-world applications (and obviously led to wrong
struct layouts).
The front-end uses a single bool variable as gate for all destructors
(by replacing each dtor expression by `gate || dtor()`).
This flag is set at the end of the argprefix chain of CommaExp.
So when no exception occurs in argprefix, the dtors aren't invoked in
the inserted finally block.
Right after the finally block we have the call to NewExp's ctor, which
will take care of destructing its parameters.
_d_newarraymT and _d_newarraymiT are now named _d_newarraymTX and
_d_newarraymiTX. There is also a change in the signature: instead
of a variable length argumentlist the functions now require an array
of dimensions.
This should fix test runnable/test28.d
This is required for the GC to be able to call their dtors, if need
be (by appending a TypeInfo pointer after the actual struct in case
the struct has a dtor).
_d_newitemT() / _d_newitemiT() also perform the initialization, so
we don't need to take care of that in toir.cpp anymore.
Runtime function _d_delarray_t() expects the TypeInfo_Struct argument
to be null if the struct has no dtor.
This fixes runnable/hospital.d, which deletes a dynamic array of class
handles. But in this case, we only delete the memory allocated for the
handles and don't destruct the actual class instances!
For hospital.d, it doesn't matter as the class has no dtor.
But I doubt that's intended behavior...
Thereby enabling nested assign/binAssign expressions in left-hand-sides
of assign/binAssign expressions as well as taking the address of assign/
binAssign expressions.
Includes all tagged with v2.067.0-b2. Does not includes latest druntime changes from ldc (head) branch.
This is known to be broken. I only merged the frontend stuff but did not try to compile something...
