| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Do not compare reference values across executions
Since we optimize assuming a closed world, optimizations can change the types
and structure of GC data even in externally-visible ways. Because differences
are expected, the fuzzer already did not compare reference-typed values from
before and after optimizations when running with nominal typing. Update it to
not compare these values under any type system.
* Unpin V8
Our WasmGC output is no longer compatible with the previously pinned version and
the issue that caused us to pin it in the first place has been resolved.
|
| |
|
|
|
|
| |
More recent version of V8 include a change from `dataref` to `structref` that
prevent WasmGC modules produced by the fuzzer from validating. Use the last
compatible v8 version if it is in the path until we can update Binaryen to be
compatible with newer v8.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
An overview of this is in the README in the diff here (conveniently, it is near the
top of the diff). Basically, we fix up nn locals after each pass, by default. This keeps
things easy to reason about - what validates is what is valid wasm - but there are
some minor nuances as mentioned there, in particular, we ignore nameless blocks
(which are commonly added by various passes; ignoring them means we can keep
more locals non-nullable).
The key addition here is LocalStructuralDominance which checks which local
indexes have the "structural dominance" property of 1a, that is, that each get has
a set in its block or an outer block that precedes it. I optimized that function quite
a lot to reduce the overhead of running that logic after each pass. The overhead
is something like 2% on J2Wasm and 0% on Dart (0%, because in this mode we
shrink code size, so there is less work actually, and it balances out).
Since we run fixups after each pass, this PR removes logic to manually call the
fixup code from various places we used to call it (like eh-utils and various passes).
Various passes are now marked as requiresNonNullableLocalFixups => false.
That lets us skip running the fixups after them, which we normally do automatically.
This helps avoid overhead. Most passes still need the fixups, though - any pass
that adds a local, or a named block, or moves code around, likely does.
This removes a hack in SimplifyLocals that is no longer needed. Before we
worked to avoid moving a set into a try, as it might not validate. Now, we just do it
and let fixups happen automatically if they need to: in the common code they
probably don't, so the extra complexity seems not worth it.
Also removes a hack from StackIR. That hack tried to avoid roundtrip adding a
nondefaultable local. But we have the logic to fix that up now, and opts will
likely keep it non-nullable as well.
Various tests end up updated here because now a local can be non-nullable -
previous fixups are no longer needed.
Note that this doesn't remove the gc-nn-locals feature. That has been useful for
testing, and may still be useful in the future - it basically just allows nn locals in
all positions (that can't read the null default value at the entry). We can consider
removing it separately.
Fixes #4824
|
| |
|
|
|
| |
This flag is needed now that we have testcases using extended-const
in our test suite, as the fuzzer will grab random testcases, modify them,
and run them through v8.
|
| |
|
| |
This sets the C++ standard variable in the build to C++17, and makes use of std::optional (a C++17 library feature) in one place, to test that it's working.
|
| |
|
|
|
|
|
|
|
| |
This is another attempt to address #4073. Instead of relying on the
timestamp, this examines git log to gather the list of test files added
or modified within some fixed number of days. The number of days is
currently set to 30 (= 1 month) but can be changed. This will be enabled
by `--auto-initial-contents`, which is now disabled by default.
Hopefully fixes #4073.
|
| |
|
|
| |
NFC (#4090)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
#3792 added support for module linking and (register command to
wasm-shell, but forgot about three problems:
- Splitting spec tests prevents linking test modules together.
- Registered modules may still be used in assertions or an invoke
- Modules may re-export imported objects
This PR appends transformed modules after binary checks to a spec.wast
file, plus assertion tests and register commands. Then runs wasm-shell
on the whole file. It also keeps both the module name and its registered
name available in wasm-shell for use in shell commands and linked
modules. Furthermore, it correctly finds the module where an object is
defined even if it is imported and re-exported several times.
The updated version of imports.wast spec test is enabled to verify the
fixes.
|
| |
|
|
|
|
| |
This removes feature flags that are now included in `--wasm-staging` and
adds new experimental flags. Does not change the fuzzer's behavior at
the moment because the fuzzer does not seem to be currently enabled for
GC or typed funcref yet.
|
| |
|
|
|
|
| |
This will allow .fromBinary tests be executed with the desired featurs
so there will be no difference between those tests and .from-wast tests.
Fixes #3545
|
| |
|
|
|
|
|
|
|
|
|
|
| |
minify_check checks that we can print and read minified wast. The test
also, however, assumed that we round-trip such things perfectly. That's
never been true, and only by chance did this go unnoticed until now,
in #3523
The specific issue happening there is that we create a block without a
name. Then we write that as text, then read it. When we read it, we give
all such blocks a name (and we rely on optimizations to remove it later
when possible - this avoids optimizing in the parser). The extra name
looks like a bug to minify_check.
|
| |
|
|
| |
And demonstrate its capabilities by porting all tests of the
optimize-instructions pass to use lit and FileCheck.
|
| |
|
|
|
|
|
| |
`node` is he name used by the upstream project. `nodejs` is a legacy
name used on older debian/ubunru systems.
Searching for `nodejs` first meant it was finding my local (old)
`nodejs` package even those I have a more recent `node` in my $PATH.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
lit and FileCheck are the tools used to run the majority of tests in LLVM. Each
lit test file contains the commands to be run for that test, so lit tests are
much more flexible and can be more precise than our current ad hoc testing
system. FileCheck reads expected test output from comments, so it allows test
output to be written alongside and interspersed with test input, making tests
more readable and precise than in our current system.
This PR adds a new suite to check.py that runs lit tests in the test/lit
directory. A few tests have been ported to demonstrate the features of the new
test runner.
This change is motivated by a need for greater flexibility in testing wasm-split.
See #3359.
|
| |
|
|
| |
This makes it easier to install libbinaryen.so into an alternative
locations. Fixes part of issue #2999 for me.
|
| | |
|
| |
|
|
|
| |
The spec interpreter is no longer used at all. Mozjs is still used optionally, but
while it was crucial in the past for test coverage, it is entirely optional now and
not run by default, so no need to warn.
|
| |
|
| |
Adds a new script `./third_party/setup.py` to conveniently install necessary dependencies for testing and fuzzing, including the SpiderMonkey JS shell (mozjs), the V8 JS shell and WABT. Other scripts now automatically pick these up when installed and fall back to look for the tools in PATH like before.
|
| |
|
|
|
| |
Switch us back to C++ standard support to 14 (for now), so we can easily upgrade
again once the autoroller issues are resolved (atm the chromium roller does not
have a libc++ with c++17 support).
|
| | |
|
| |
|
| |
Allows for using `*` wildcards and simplifies the code!
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
This lets us run most tests at least on that platform.
Add a new function for skipping those tests, skip_if_on_windows,
so that it's easy to find which tests are disabled on windows for later fixing
efforts.
This fixes a few minor issues for windows, like comparisons
should ignore \r in some cases.
Rename all passes tests that use --dwarfdump to contain "dwarf"
in their name, which makes it easy to skip those (and is clearer
anyhow).
|
| |
|
|
|
|
|
|
| |
This was useful back when we didn't have many VMs to test in, and
we weren't confident in our binaries being valid wasm. Today we have
lots of testing on VMs, and in particular, this test tends to fail when we
do things like reorder SIMD opcode constants. (This doesn't fail on CI
as we don't have v8 installed there, so this path is never reached, but
it does happen locally.)
|
| |
|
|
|
|
|
| |
anyref future semantics were changed to only represent opaque host values, and thus renamed to externref.
[Chromium](https://bugs.chromium.org/p/v8/issues/detail?id=7748#c360) was just updated to today (not yet released). I couldn't find a Mozilla bugzilla ticket mentioning externref so I don't immediately know if they've updated yet.
https://github.com/WebAssembly/reference-types/pull/87
|
| | |
|
| |
|
|
|
| |
It looks like these were only uses as part of the vanilla tests
which were removed back in #2482.
|
| |
|
|
| |
fail_if_not_identical_to_file() (#2649)
|
| |
|
|
|
| |
This reverts commit 132daae1e9154782bb1afa5df80dfe7ea35f0369.
This change is the same as before but the fix in #2619 should now make it safe.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Track the beginning and end of each function, both when reading
and writing.
We track expressions and functions separately, instead of having a single
big map of (oldAddr) => (newAddr) because of the potentially ambiguous case
of the final expression in a function: it's end might be identical in offset
to the end of the function. So we have two different things that map to the
same offset. However, if the context is "the end of the function" then the
updated address is the new end of the function, even if the function ends
with a different instruction now, as the old last instruction might have
moved or been optimized out. Concretely, we have getNewExprAddr
and getNewFuncAddr, so we can ask to update the location of either
an expression or a function, and use that contextual information.
This checks for the DIE tag in order to know what we are looking for.
To be safe, if we hit an unknown tag, we halt, so that we don't silently
miss things.
As the test updates show, the new things we can do thanks to this
PR are to update compile unit and subprogram low_pc locations.
Note btw that in the first test (dwarfdump_roundtrip_dwarfdump.bin.txt)
we change 5 to 0: that is correct since that test does not write out
DWARF (it intentionally has no -g), so we do not track binary
locations while writing, and so we have nothing to update to (the
other tests show actual updating).
Also fix the order in the python test runner code to show a diff
of expected to encountered, and not the reverse, which confused
me.
|
| |
|
|
|
| |
Binaryen.js now uses binaryen (was Binaryen) as its global
name to align with the npm package. Also fixes issues with
emitting and testing both the JS and Wasm builds.
|
| |
|
|
|
| |
(#2542)" (#2576)
This reverts commit f62e171c38bea14302f9b79f7941a248ea704425.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Reland "Fix renaming in FixInvokeFunctionNamesWalker (#2513)"
In the previous iteration of this change we were not calling
`renameFunctions` for each of the functions we removed.
The problem manifested itself when we rename the imported function to
`emscripten_longjmp_jmpbuf` to `emscripten_longjmp`. In this case the
import of `emscripten_longjmp` already exists so we remove the import of
`emscripten_longjmp_jmpbuf` but we were not correclty calling
renameFunctions to handle the rename of all the uses.
Add an additional test case to cover the failures that we saw on the
emscripten tree.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This PR enables compiling Binaryen to WebAssembly when building binaryen.js. Since WebAssembly is best compiled and instantiated asynchronously in browsers, it also adds a new mechanism to tell if respectively when the module is ready by means of one of the following:
// Using a promise
const binaryen = require("binaryen");
binaryen.ready.then(() => {
... use normally ...
});
// Using await
const binaryen = require("binaryen");
(async () => {
await binaryen.ready;
... use normally ...
})();
// Where top-level await is available
const binaryen = await require("binaryen").ready;
... use normally ...
One can also tell if Binaryen is already ready (for example when assuming it in follow-up code) by:
if (/* we already know that */ binaryen.isReady) {
... use normally ...
} else {
throw Error("Binaryen is supposed to be ready here but isn't");
}
The JS test cases have been updated accordingly by wrapping everything in a test function and invoking it once ready. Documentation will have to be updated as well to cover this of course. New file size is about 2.5mb, even though the Wasm becomes inlined into the JS file which makes distribution across different environments a lot easier.
Also makes building binaryen (to either js or wasm) emit binaryen.js, and not binaryen_js.js etc.
Supersedes and thus fixes #1381
With .ready it also fixes #2452
|
| |
|
| |
This reverts commit f0a2e2c75c7bb3008f10b6edbb8dc4cfd27b7d28.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes https://github.com/emscripten-core/emscripten/issues/9950.
The issue only shows up when debug names are not present so most of
the changes in CL come from disabling debug names in the lld tests.
We want to make sure that wasm-emscripten-finalize runs fine without
debug names so I think it makes most sense to test in this mode.
The actual bugfix is in wasm-emscripten.cpp as part of the
FixInvokeFunctionNamesWalker. The problem was the name of the function
rather than is import name was being added to importRenames. This means
that when debug names were present (and the two names were the same)
we didn't see the bug.
|
| |
|
|
| |
(#2508)
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This updates spec test suite to that of the current up-to-date version
of https://github.com/WebAssembly/spec repo.
- All failing tests are added in `BLACKLIST` in shared.py with reasons.
- For tests that already existed and was passing and started failing
after the update, we add the new test to the blacklist and preserve
the old file by renaming it to 'old_[FILENAME].wast' not to lose test
coverage. When the cause of the error is fixed or the unsupported
construct gets support so the new test passes, we can delete the
corresponding 'old_[FILENAME].wast' file.
- Adds support for `spectest.print_[type] style imports.
|
| |
|
|
|
|
|
|
| |
This makes auto_update_tests.py update spec test outputs (ones that are
printed with `spectest.print` import) and extracts spec tests blacklist
into shared.py with comments for reasons why each of them fails.
Also deletes if-label-scope.fail.wast.log because it does not seem to
match with any of existing tests.
|
| |
|
|
|
|
|
|
| |
This makes test scripts simpler by reducing loop depths and extracting
repeating code into methods or variables.
- `get_tests` returns a list of tests with specified extensions. This
includes files with a full path rather than just file names.
- Reduces loop depths by using early exits and `get_tests`.
|
| |
|
|
|
|
| |
renamed (#2382)
Fixes https://github.com/WebAssembly/binaryen/issues/2180
|
| |
|
| |
This allows fuzzing in parallel invocations.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Another round of trying to push upstream things from my fork.
This PR only adds support for anyref itself as an opaque type. It does NOT implement the full [reference types proposal](https://github.com/WebAssembly/reference-types/blob/master/proposals/reference-types/Overview.md)--so no table.get/set/grow/etc or ref.null, ref.func, etc.
Figured it was easier to review and merge as we go, especially if I did something fundamentally wrong.
***
I did put it under the `--enable-reference-types` flag as I imagine that even though this PR doesn't complete the full feature set, it probably is the right home. Lmk if not.
I'll also be adding a few github comments to places I want to point out/question.
|
| |
|
|
|
|
|
|
| |
pep8 specifies 4 space indentation. The use of 2 spaces is, I believe
a historical anomaly where certain large organizations such as google
chose 2 over 4 and have yet to make the switch.
Since there isn't too much code in binaryen today it seems reasonable to
make the switch.
|
| | |
|
| |
|
|
|
| |
I fixed flatten.bin.txt which seems to have just had some corrupted data, and I removed some fancy unicode from the spec comments tests, which I'm not sure it's important enough to figure out how to fix.
Fixes #1691
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
Update build-js.sh to output to `out` directory. This is district
from the `bin` directory which is used by the cmake build and may or
may not live in the source tree. The `out` directory currently always
lives in the source tree.
As a followup change I hope to additionally move all test outout into
this tree.
See #2104
|
