This repository was archived by the owner on Nov 3, 2021. It is now read-only.
Should memory.init/drop of active segments be a validation error?
#107
Comments
|
There was a long discussion about this here: #27, starting from this comment. Luke explains the rationale in a following comment: "although taking into account what @conrad-watt's pov, the "active/passive" distinction goes away at runtime; there are simply "has the segment already been dropped", with active segments having been desugared into table.init;table.drop expressions executed before the start function" |
|
ok sgtm |
fitzgen
added a commit
to fitzgen/walrus
that referenced
this issue
Jul 24, 2019
…nit` be passive The validation rules do not require it: * https://github.com/WebAssembly/bulk-memory-operations/blob/eca021a3a9aaa3cca336ecdd5379e570023bbe3a/document/core/valid/instructions.rst#memoryinitx * https://github.com/WebAssembly/bulk-memory-operations/blob/eca021a3a9aaa3cca336ecdd5379e570023bbe3a/document/core/valid/instructions.rst#datadropx And the spec tests check that you can do this (and potentially get runtime errors for already implicitly dropped active segments): * https://github.com/WebAssembly/bulk-memory-operations/blob/07c57f16ff67b29e30719e1a692aad486f582021/test/core/bulk.wast#L148-L168 See also these discussions: * WebAssembly/bulk-memory-operations#107 * WebAssembly/bulk-memory-operations#27 (comment)
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Presumably they could be validation errors, but the DataCount section doesn't have enough information?
Was a bitfield considered rather than a single number for the DataCount section?
The text was updated successfully, but these errors were encountered: