The remaining ones should also be easy to remove with a
closer look at the situation.
Ideally, we would get rid of all of them at some point and
use safe wrapper functions for accessing the IrDsymbol
associated with a given declaration (which would emit the
declarations on the fly if not already present).
This commit fundamentally changes the way symbol emission in
LDC works: Previously, whenever a declaration was used in some
way, the compiler would check whether it actually needs to be
defined in the currently processed module, based only on the
symbol itself. This lack of contextual information proved to
be a major problem in correctly handling emission of templates
(see e.g. #454).
Now, the DtoResolve…() family of functions and similar only
ever declare the symbols, and definition is handled by doing
a single pass over Module::members for the root module. This
is the same strategy that DMD uses as well, which should
also reduce the maintainance burden down the road (which is
important as during the last few releases, there was pretty
much always a symbol emission related problem slowing us
down).
Our old approach might have been a bit better tuned w.r.t.
avoiding emission of unneeded template instances, but 2.064
will bring improvements here (DMD: FuncDeclaration::toObjFile).
Barring such issues, the change shoud also marginally improve
compile times because of declarations no longer being emitted
when they are not needed.
In the future, we should also consider refactoring the code
so that it no longer directly accesses Dsymbol::ir but uses
wrapper functions that ensure that the appropriate
DtoResolve…() function has been called.
GitHub: Fixes#454.
I am surprised and horrified at the same time that the test suite
(on platforms other than Windows, where the code path is triggered
in core.stdc.stdio) did not catch this at all.
The elemSize-dependent path probably doesn't make too much sense for
global variables, as it always refers to the total size of the global.
Should add a special case for arrays for clearer codegen in those cases.
GitHub: Fixes#477.
Instead of creating individual stores to the array elements an constant
array is created and assigned to the destination memory. This is much
less IR than before. With -O it is optimized to a memset.
This not only reduces code duplication, but the unification
also enables code a la StructLiteralExp to handle classes
(for CTFE class constant support in 2.063).
It might be worth considering to move the stride multiplication
down to the glue layer in the upstream sources. But assigning a
different meaning to AST nodes was a giant maintenance
liability, especially with regard to CTFE.
