Previously it just gave up when it saw anything other than an ELF .o file. We
could have it ignore .a files, but we might as well do the right thing.
<rdar://problem/23045632>
Assertion failed: (NumUsePointsToFind > 0 && "There must be at least one
releasing instruction for an alloc"), function canPromoteAlloc
Revert "Fix comment for StackPromotion pass in SIL Passes"
Revert "Reapply the StackPromotion commit
0dd045ca04dcc10a33abf57f7e1b08260c4e3de1."
This reverts commit 3f4b1496bd and commit
199cfca13b.
It promotes allocations of native swift objects and array buffers to the stack if it is possible.
The SIL StackPromotion pass is the main part of the optimization. For details see comments there.
Unfortunately we need an additional LLVM pass to handle array buffers, which is not very nice.
I hope that we can get rid of it in future (again: for details see the comments in StackPromotion.cpp)
The optimization gives performance improvements in some benchmarks, mostly related to array literals:
ArrayLiteral: +12%
Combos: +16%
DictionaryLiteral: + 37%
RIPEMD: +10%
StringBuilder: +27%
StringInterpolation: +11%
And last but not least the new benchmark which is dedicated to test stack promotion:
ObjectAllocation: +52%
Conventionally, code completion results are matched with user input solely by
names. However, names are limited in expressiveness. From this comments, we start to
decorate code completion results with @keywords fields extracted from Clang doc comments.
These fields are added by API authors to comment the decl with information that
is not manifested clear enough through names. Code completion users' typing of the
keyword leads to the corresponding code completion results being selected as well.
Keywords can be arbitrarily long and can be multiple.
For instance, a function called "index()" has "@keyword find" in its doc comment.
Users' typing of "find" leads to "index()" being selected in the code completion list.
The code goes into its own sub-tree under 'tools' but tests go under 'test',
so that running 'check-swift' will also run all the SourceKit tests.
SourceKit is disabled on non-darwin platforms.
Now that one can not use analysis groups anymore on LLVM, we were given two
choices. Reinsert that dynamicism so we could via opt -load add in analyses /or/
just create our own version of opt. I decided to go with the later since it
enables us to simulate how IRGen sets up the LLVM pass pipeline exactly,
something we were unable to do before.
This also sets us up for whenever we need to respond to the new pass manager.
Swift SVN r32808
When users try to print the interface of a specific type (most often through cursor
infor query of SourceKit), we should simplify the original decls by replacing
archetypes with instantiated types, hiding extension details, and omitting
unfulfilled extension requirements. So the users can get the straight-to-the-point
"type interface". This commit builds the testing infrastructure for this feature,
and implements the first trick that wraps extension contents into the interface body.
This commit also moves some generic testing support from SourceKit to Swift.
Swift SVN r32630
- This allows 'swift <FOO>' to be used to invoke swift-<FOO>, so that we can
use 'swift' as a namespace for additional commands (like ones used for the
package manager.
- This patch just adds the basic subcommand functionality, but doesn't rework
either 'swift-autolink-extract' or 'swift -frontend' to follow this code path.
- <rdar://problem/22844530> Provide 'swift' support for subcommands
Swift SVN r32367
My temporary hackery around inferring default arguments from imported
APIs was too horrible. Make it slightly more sane by:
1) Actually marking these as default arguments in the type system,
rather than doing everything outside of the type system. This is a
step closer to what we would really do, if we go in this
direction. Put it behind the new -frontend flag
-enable-infer-default-arguments.
2) Only inferring a default argument from option sets and from
explicitly "nullable" parameters, as stated in the (Objective-)C API
or API notes. This eliminates a pile of spurious, non-sensical "=
nil"'s in the resulting output.
Note that there is one ugly tweak to the overloading rules to prefer
declarations with fewer defaulted arguments. This is a bad
implementation of what is probably a reasonable rule (prefer to bind
fewer default arguments), which intentionally only kicks in when we're
dealing with imported APIs that have default arguments.
Swift SVN r32078
Running sil-opt with the new call graph printing pass (which uses
llvm::outs) was resulting in some spew about "LLVM IO Error" because we
could potentially close stdout twice.
Swift SVN r32041
Till now, a SIL module would be only verified if an optimization has changed it. But if there were no changes, then no verification would happen and some SIL module format errors would stay unnoticed. This was happening in certain cases when reading a textual SIL module representation, which turned out to be broken, but SIL verifier wouldn't catch it.
Swift SVN r31863
To invoke the front-end on a SIL with whole-module optimizations enabled, execute:
swiftc -frontend myfile.sil
To invoke the front-end on a SIL without whole-module optimizations enabled, add a -primary-file option:
swiftc -frontend -primary-file myfile.sil
To invoke a sil-opt with whole-module optimizations enabled, use the -wmo option:
sil-opt myfile.sil -wmo
This change was need to be able to write SIL unit tests which should be compiled in the WMO mode.
Swift SVN r31862
