[Webkit-unassigned] [Bug 219427] [WASM-References] Add support for table.copy
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Wed Dec 2 11:05:43 PST 2020
https://bugs.webkit.org/show_bug.cgi?id=219427
--- Comment #4 from Saam Barati <sbarati at apple.com> ---
Comment on attachment 415201
--> https://bugs.webkit.org/attachment.cgi?id=415201
Patch
View in context: https://bugs.webkit.org/attachment.cgi?id=415201&action=review
> Source/JavaScriptCore/ChangeLog:8
> + Add support for table.copy from reference types proposal:
might be worth describing what it does
> Source/JavaScriptCore/wasm/WasmFunctionParser.h:670
> + TypedExpression lenght;
lenght => length
> Source/JavaScriptCore/wasm/WasmInstance.cpp:134
> + for (uint32_t idx = len; idx > 0; --idx)
> + fn(dstTable, srcTable, dstOffset + (idx - 1), srcOffset + (idx - 1));
I prefer to write this loop as:
```
for (uint32_t idx = len; idx--; )
```
Then inside the body of the loop, you can just refer to idx instead of (idx - 1)
>> Source/JavaScriptCore/wasm/WasmOperations.cpp:729
>> +}
>
> I think you can use Checked<uint32_t> for this.
just make sure to not use CrashOnOverflow and use RecordOverflow instead
> Source/JavaScriptCore/wasm/WasmOperations.cpp:739
> + if (dstTable->isExternrefTable() != srcTable->isExternrefTable())
> + return false;
the error message you throw in the JIT code seems different than this type of error
> Source/JavaScriptCore/wasm/WasmSectionParser.cpp:424
> + // We don't have the table.init instruction to use passive element segments,
> + // but we need to parse them for the table.copy spec test, so we just skip them.
This seems very fishy. We're parsing something we don't support just for a test?
> Source/JavaScriptCore/wasm/WasmSectionParser.cpp:501
> + // We don't have the table.init instruction to use passive element segments,
> + // but we need to parse them for the table.copy spec test, so we just skip them.
ditto
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20201202/6ba904fd/attachment.htm>
More information about the webkit-unassigned
mailing list