You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
During parsing of binary-encoded files, the input isn't fully validated against the specification.
For example, the specification states that "known sections may [must?] not appear out of order", however, this isn't currently being verified. Most likely, an assertion failure will occur during parsing, but this isn't guaranteed.
Likewise, the specification states that "each section is optional and may appear at most once", but since sections are parsed in a loop, the latest section of a specific type will take precedence, plus any side effects from previous occurrences of that section type.
The text was updated successfully, but these errors were encountered:
Additionally, the description of the names section states that "a validation error in this section does not cause validation for the whole module to fail, and is instead treated as if the section was absent". This seems to be imply that a rollback is necessary on encountering a validation error?
It also states that the count of function_names may be greater or less than the actual number of functions. Currently (post-#546), the existing implementation checks that the count is equal to the number of functions.
Yes, we should fix all these. I haven't made it a priority personally since it needs tests, and tests should arrive eventually in the spec repo, so not urgent to fix before that. However, if anyone wants to fix earlier, that's great of course.
During parsing of binary-encoded files, the input isn't fully validated against the specification.
For example, the specification states that "known sections may [must?] not appear out of order", however, this isn't currently being verified. Most likely, an assertion failure will occur during parsing, but this isn't guaranteed.
Likewise, the specification states that "each section is optional and may appear at most once", but since sections are parsed in a loop, the latest section of a specific type will take precedence, plus any side effects from previous occurrences of that section type.
The text was updated successfully, but these errors were encountered: