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
I don't have a test case for this yet, but the fix in #5225 almost certainly also needs to be applied two the two spots in MatchFunctions that call GroupIdsAndMatch with Differ::GroupIdsHelperGetTypeId, which is never the right thing to do, since it ignores the established src/dst mappings for types.
I have a fix for this, and if you prefer, I can just fold it into #5225. But the fix doesn't have any effect on test/diff/diff_files, meaning that (assuming that there is indeed a bug here - it seems clear) it doesn't have any test coverage. Rather than expand the scope of #5218, I filed this separate issue.
The text was updated successfully, but these errors were encountered:
jimblandy
added a commit
to jimblandy/SPIRV-Tools
that referenced
this issue
May 22, 2023
Change `Differ::MatchFunctions` to use `GroupIdsAndMatchByMappedId`,
instead of calling `GroupIdsAndMatch` with
`Differ::GroupIdsHelperGetTypeId`, which is never correct: doing so
compares src and dst type ids directly, rather than consulting the
established mapping.
FixesKhronosGroup#5226.
Change `Differ::MatchFunctions` to use `GroupIdsAndMatchByMappedId`,
instead of calling `GroupIdsAndMatch` with
`Differ::GroupIdsHelperGetTypeId`, which is never correct: doing so
compares src and dst type ids directly, rather than consulting the
established mapping.
FixesKhronosGroup#5226.
I don't have a test case for this yet, but the fix in #5225 almost certainly also needs to be applied two the two spots in
MatchFunctions
that callGroupIdsAndMatch
withDiffer::GroupIdsHelperGetTypeId
, which is never the right thing to do, since it ignores the established src/dst mappings for types.I have a fix for this, and if you prefer, I can just fold it into #5225. But the fix doesn't have any effect on
test/diff/diff_files
, meaning that (assuming that there is indeed a bug here - it seems clear) it doesn't have any test coverage. Rather than expand the scope of #5218, I filed this separate issue.The text was updated successfully, but these errors were encountered: