| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
|
|
|
|
|
| |
`grow_memory` -> `memory.grow`
`current_memory` -> `memory.size`
This could have been a smaller change, but I took the opportunity to
rename the Token types, Expr types, and callback functions too. Many of
these are sorted alphabetically, so I resorted based on their new names.
|
| | |
|
| |
|
|
| |
* Fix edge case elem segment bounds checking
* Fix bounds checking when importing spectest table
|
| |
|
|
|
| |
The spectest module's functions and globals are no longer overloaded, so
their implementations in `spectest-interp` and `spec-wasm2c-prefix.c`
have to be changed as well.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The `call_indirect` instruction now allows the type to be specified
inline or via the common "type use" syntax (or both):
```
call_indirect (type $t1) ...
call_indirect (param i32 f32) (result f64) ...
call_indirect (type $t2) (param i32) ...
```
This means that `CallIndirectExpr` changes from storing a `Var` (the
referenced func type) to a `FuncDeclaration`, which can store both a
type use and a function signature. Most of the changes here are fallout
from that change.
The other major change is that function signature resolution and error
checking now needs to iterate over a functions expr list looking for
`call_indirect` instructions. I'm not sure if this is a significant
overhead in parsing/validation yet, but there are plenty of ways to
optimize it if it ends up in profiles.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
* Mostly just updating test expectations
* Function signatures need to be checked as part of parsing (see
assert_malformed test in `testsuite/func.wast`; added
`ValidateFuncSignatures` for this.
|
| |
|
|
|
|
|
| |
* Implicit function types are now appended to the module, not directly
after the function/import that uses them.
* Float and int literals can have an underscore separating digits.
|
| | |
|
| |
|
|
|
|
|
| |
* Return error when floats overflow (used to return inf)
* Significantly simplfy the significand parser in FloatParser::ParseHex
* Add some new parse tests to hexfloat.cc; the previous tests only
checked valid/canonical strings
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add support for quoted modules: `(module quote "...")`
* Binary modules must be annotated: `(module binary "...")`
* Multiple result blocks are no longer a parser error:
`(func (result i32) (result i32) ...)`
* Function types can specify unused bind variables:
`(type (func (param $foo)))`
* Rename `RawModule` -> `ScriptModule`. This encapsulates a module that
may not be parsed yet, whether binary or "quoted".
* Validate load/store offsets and alignment in the parser, not in the
validator. The spec tests assume that you can catch these errors with
`assert_malformed`.
* Parse wast files in `wasm-interp` when checking malformed/invalid/etc.
modules. This allows us to run all assertions at the same time, which
is nice. `wasm-interp` should probably be renamed, though.
* Two tests in `type.wast` fail because they use:
`(assert_invalid (module quote "..."))`. I'd prefer that we don't
support this, since it's unnecessary, and additional work. I'll fix in
a follow-up CL if we decide this is worth keeping.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Update testsuite; various lexing/parsing fixes
Lexer changes:
* Switch re2c parser to UTF-8 parser. This can almost be done "for
free" with a flag, but required a bit of work to allow us to catch
malformed UTF-8 as well.
* Change the re2c fill value to 0xff, since it's never a valid UTF-8 byte.
* Allow for more reserved tokens (basically any ascii aside from
parentheses, double-quote, and semi-colon)
* Remove "infinity" from lexer, only "inf" is allowed now.
* Change definition of EOF token, it was implemented incorrectly. The
correct way to handle it is to only return it from FILL when there is no
more data to fill.
* \r is a valid escape.
Parser changes:
* Changes to match the spec parser:
- block signatures use (result <type>) syntax
- func/global/table/memory can have multiple inline exports
- inline imports are handled in func definition instead of import
definition
- allow for inline modules (i.e. no "(module ...)" s-expr required)
* Remove FuncField. This was previously used for parsing
params/results/locals, but it's less code to just parse
right-recursive (i.e. backward) and insert everything at the front.
This requires reversing the indexes in the BindingHash too.
* Remove the nasty macros `APPEND_FIELD_TO_LIST`,
`APPEND_ITEM_TO_VECTOR`, `APPEND_INLINE_EXPORT`, and
`CHECK_IMPORT_ORDERING`. This behavior is all handled by
`append_module_fields` now.
* All inline imports/exports are handled by returning additional
ModuleFields in a list. This removes the need for `OptionalExport`,
`ExportedFunc`, `ExportedGlobal`, `ExportedTable`, and
`ExportedMemory`.
* Use "_opt" suffix instead of "non_empty_" prefix, e.g.:
- text_list => text_list_opt, non_empty_text_list => text_list
* The locations changed for some symbols, typically the use the name
following the LPAR now, e.g. (import
^^^^^^
* Add PPA for re2c 0.16
* add -y to skip confirmation on travis
|
| | |
|
| |
|
|
| |
This is enough to pass the testuite, but doesn't yet validate utf8
strings in the text format.
|
| |
|
|
| |
This exposed a small bug where an empty init expr would have type 0,
which would print as "(null)" since it didn't exist.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Positive NaN (+nan) is defined as being a NaN value where the sign bit
is clear, the exponent is all ones, and the tag has only the "quiet" bit
set. The quiet bit is the most-significant bit of the tag. For example,
for a 32-bit float, +nan is 0x7cf00000.
"Canonical NaN" is either +nan or -nan, where -nan is +nan with the sign
bit set.
"Arithmetic NaN" is defined as any other quiet NaN (i.e. the quiet bit
must be set, but any other bit can be either 0 or 1.)
This change doesn't update the interpreter because it is only a
loosening of the previous NaN behavior.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* Update wasm binary format label to 1
Rebaseline tests, but no other changes.
* Revert test/spec/globals.txt
* Update testsuite
* Update spec binary.wast test and testsuite repo
|
| |
|
|
|
|
| |
Fixed a few typechecker errors:
* br_table must have consistent type signature for all branch targets
* if without else cannot have a signature
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The typechecking is now shared between the validator and
binary-reader-interpreter as well.
* Shared the various LabelType enumerations -> WabtLabelType.
* Fixed the "invalid depth" errors, the max value was incorrect.
* Removed some tests that aren't useful anymore:
- interp/if-then-br hasn't been useful for a while, but now is even
less so because it doesn't validate without dropping the i32.const
from the true branch
- typecheck/bad-br-multi-type isn't possible because the block must
specify the signature for the br.
- typecheck/bad-label-multi-type used to test matching signature for
fallthrough and br, but isn't possible because of block signatures.
|
| | |
|
| |
|
|
|
|
|
| |
Also:
* Check all data/elem segments before updating memory/tables
* assert_exhaustion should check for call stack or value stack
exhaustion.
* Don't allow out-of-bound size 0 data segments
|
| |
|
|
| |
Also, update assert_malformed to initialize the environment so the
"spectest" module is available.
|
| | |
|
| |
|
|
|
|
|
| |
* Non-imported globals can not be used in initializer expressions
* block/loop/if labels can be repeated at the end
* get_global index in initializer expression should reference use module
index space, not "defined" global index space
|
| | |
|
| |
|
|
|
|
|
| |
* `offset=` and `align=` can now specify hex values
* data and elem segment offsets are now not checked; they can be out of
order or overlapping, and even out-of-bounds if the size is zero
* added the new tests `skip-stack-guard-page` and `unwind`
|
| |
|
|
|
|
| |
This updates third_party/testsuite to the current head of the `binary_0xc` branch
(https://github.com/WebAssembly/testsuite/commit/58c1bf67cc65a1219d49114c41805bafc75ac0c6),
which is the collection of tests from the current head of the spec repo's `binary_0xc` branch
(https://github.com/WebAssembly/spec/commit/0fa1531c2deebb767b263d405dbce849c9663f55).
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
It was an interesting experiment, but it is not maintained or tested.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Change 7c482b1a of the spec repo now requires that implicit function
types are created when creating a new func or import, but only when the
type is not explicitly specified, and the have not been specified. For
example:
```
(func (param i32))
(import "foo" "bar" (result i32)
```
This was previously done in sexpr-wasm when writing the binary, but now
these types can be referenced by `call_import`, so it has to be handled
earlier.
Since all signatures are now defined in the module, I made it so the
function signature is shared with the module function types. The tricky
part here is when a function type and an explicit signature are both
specified:
```
(func (type $t) (param f32) ...)
```
In this case, we need to keep the explicit signature around long enough
to be checked against the function type, so it must be owned by the
function. The WASM_FUNC_DECLARATION_FLAG_SHARED_SIGNATURE flag specifies
this case.
To simplify code that uses the AST, I reverted many of the
changes introduced in 73e5bc7d. Now the function declaration's signature
is always valid to access, and will have been resolved if it was
specified with a func type instead of an explicit signature.
Some other changes:
* There was a bug in the interpreter's `f32.demote/f64` rounding when
given a value that is very close to positive or negative F32_MAX
* Added `update-spec-tests.py` to more easily update the spec tests in
`test/interp/spec/*.txt` from the files in
`third_party/testsuite/*.wast`. Previously I would just do it
manually.
|
| |
|
|
|
| |
`(func (br 0))` is valid, because the function introduces an implicit
block scope. This change also updates the testsuite (which has tests for
this behavior).
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* Modify AST parser to disallow any other order than `param`, `result`,
`local`.
* Allow parsing all var integers as int64 (the lookups will fail later.)
* Fix interpreter logic determining whether to adjust the stack; we were
correctly handling the case where `unreachable` was being used in an
expression (the value stack is adjusted "up" as if a value was
actually produced), but the same behavior is required for `br`,
`br_table` and `return`.
* Add NAT token to lexer/parser, which is an integer value without a
sign.
|
| |
|
|
|
| |
It's not really the appropriate place to handle it. Now that we have
wasm-wast and wasm-interp, we can at least verify that the tools are
internally consistent.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
* Parse memory pages as uint64, as we need to check whether it's invalid
in assert_invalid, and one of the spec tests uses the value 2**32.
* Update WASM_MAX_PAGES to 65535 (not 65536)
* Add the new spec tests: binary.wast, break-drop.wast and
typecheck.wast. A couple of these were added before, but weren't added
as .txt files in the test directory.
|
| |
|
|
|
| |
You can't git clone to a specific commit SHA, so we just download the
source archive from Github.
|
| |
|
|
|
| |
This makes it easier to build for users who don't run
`git submodule update --init`.
|
| | |
|
| |
|
|
|
|
|
| |
* current_memory and grow_memory return page size, not byte size
* also update testsuite, which fixes a couple of spec tests
* disable d8/spec tests (in a hacky way), because the binary_0xb v8
support isn't landed yet, so we can't easily test it
|
| |
|
|
| |
These will be useful for testing the binary reader.
|
| | |
|
| |
|
|
| |
It's pretty clunky though
|
| | |
|
| | |
|
| | |
|
