Local values that are not mutated, don't need to be cleaned up, and
are immediate, don't need to be spilled. (All immediate args, and
non-pointer immediate let locals.)
Issue #667
Now, {a: {b: 10, c: 20}, d: @30} will simply write the values in the right
places, rather than creating intermediary records and then memmoving them.
Cuts about a megabyte off the unoptimized compiler size.
The builder functions in trans_build now look at an 'unreachable' flag
in the block context and don't generate code (returning undefined
placeholder values) when this flag is set. Threading the unreachable
flag through context still requires some care, but this seems a more
sane approach than re-checking for terminated blocks throughout the
compiler.
When creating a block, if you use its closest dominator as parent, the
flag will be automatically passed through. If you can't do that,
because the dominator is a scope block that you're trying to get out
of, you'll have to do something like this to explicitly pass on the
flag:
if bcx.unreachable { Unreachable(next_cx); }
Closes#949. Closes#946. Closes#942. Closes#895. Closes#894.
Closes#892. Closes#957. Closes#958.
Added the non_ty_var predicate (soon to be used)
Added a check in get_res_dtor (will be necessary for a future change
to type_of_fn)
Removed a gratuitous ret
Arguments that can't be safely referenced will be implicitly copied.
(Warnings for expensive copies will be forthcoming.)
This will allow us to get rid of most of the ampersands in function
signatures. See [1].
[1] https://mail.mozilla.org/pipermail/rust-dev/2011-September/000759.html
trans::type_of now has a constraint saying that its type argument
is statically sized. This eliminates the "impossible happened" case
in type_of. Yay!
I note that this change decreased translation time for stage2/rustc
from 16.1 s to 14.0 s. I also think many of the remaining checks
could be eliminated with some mildly clever use of constrained types
and further preconditions. Future work!
Experimenting with adding typestate constraints in the compiler.
Added a constraint to GEP_tag that says the variant index is in
bounds. Added necessary checks.
Having it in the alias pass was slightly more efficient (finding
expression roots has to be done in both passes), but further muddled
up the already complex alias checker.
Also factors out some duplication in the mutability-checking code.
The uniformity doesn't seem to be worth the extra noise and pointless
code being generated. If something doesn't produce a value, don't make
it return one. (For now, trans_[exprtype] things are left in the result-
returning form, even when they never return anything useful, since in
that case uniformity is arguably helpful.)
Vectors are now similar to our old, pre-internal vectors, except that
they are uniquely owned, not refcounted.
Their name should probably change too, then. I've renamed them to vec
in the runtime, will do so throughout the compiler later.
