- Use vector instead of Array, reducing allocations.
- Use vectors instead of deques since we only push_back and index.
- Remove redundant typedefs of iterators.
- Comment out unused variable (used only in commented-out GDC code).
- A few whitespace changes.
- More error checking for pragma(intrinsic)
- Properly handle templating for real (use actual nr of bits, not .sizeof * 8)
- Template all .i* and .f* intrinsics
- The old names are deprecated aliases now
I also added a preliminary patch to make tango.math.Math use the new versions.
(I think it looks a lot nicer now)
Also moved the #defines for linkage types into a separate header instead of
mars.h so we can #include revisions.h without having to rebuild the entire
frontend every time we update.
(I'm using revisions.h to get the LLVM revision for use in preprocessor
conditionals. It should work with LLVM release 2.5, old trunk and new trunk)
I made sure to create a merge conflict with the code attached to ticket #229
(which comments this code out due to changes in the DMD 'Type' type hierarchy)
so that if it gets committed as-is there will be a human looking at it.
The switch allows the optimizer and inliner to run on all modules at once and opens the door for template instantiation improvements that should lower compile time and executable size.
- See through typedefs and enums in passByVal
- Don't depend on TypeFunction::parameters->dim being the actual number of
parameters; it contains unexpanded tuples as single elements.
- Implement x86-64 extern(C), hopefully correctly.
- Tried to be a bit smarter about extern(D) while I was there.
Interestingly, this code seems to be generating more efficient code than
gcc and llvm-gcc in some edge cases, like returning a `{ [7 x i8] }` loaded from
a stack slot from an extern(C) function. (gcc generates 7 1-byte loads, while
this code generates a 4-byte, a 2-byte and a 1-byte load)
I also added some changes to make sure structs being returned from functions or
passed in as parameters are stored in memory where the rest of the backend seems
to expect them to be. These should be removed when support for first-class
aggregates improves.
