Although the old version of GEP_tup_like was incorrect in some
cases, I do not believe we ever used it in an incorrect fashion.
In particular, it could go wrong with extended index sequences
like [0, 1, 3], but as near as I can tell we only ever use it
with short sequences like [0, i].
This commit allows patterns like:
alt x { some(_) { ... } none { } }
without the '.' after none. The parser suspends judgment about
whether a bare ident is a tag or a new bound variable; instead,
the resolver disambiguates.
This means that any code after resolution that pattern-matches on
patterns needs to call pat_util::normalize_pat, which consults
an environment to do this disambiguation.
In addition, local variables are no longer allowed to shadow
tag names, so this required changing some code (e.g. renaming
variables named "mut", and renaming ast::sub to subtract).
The parser currently accepts patterns with and without the '.'.
Once the compiler and libraries are changed, it will no longer
accept the '.'.
There is now only one path doing crate expanding and typechecking,
which should make it less likely for the pretty-printing code to be
broken by changes to the compilation pipeline.
Closes#1536
This fixes issues #843 and #1546. The cost of an upcall is
unfortunate, though. I assume there must be a way to simply manually
compute the pointer or size, using something akin to the formula in
`align_to` in `rust_util.h`. I could not get this to work,
unfortunately.
Remove disr_val from ast::variant_ and always use ty::variant_info
when the value is needed. Move what was done during parsing into
other passes, primary typeck.rs. This move also correctly type checks
the disr. value expression; thus, fixing rustc --pretty=typed when
disr. values are used.
Before, literal printing would basically get derailed completely when
a literal was encountered that did not end up being printed. This
caused the strangeness seen in #1532.
Also cleans up pretty-printing of discriminants a little.
Closes#1510Closes#1532
Since we are no longer logging to the console it's possible for us to
hit a plain-old-fail statement and not output anything.
This adds a defensive mechanism that will monitor the emitted errors
and compare them to the result of the compiler task. If the compiler
fails without emitting an error it results in an ICE.
