-
-
Notifications
You must be signed in to change notification settings - Fork 5.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
bpart-related IR verification regression #57329
Comments
The Line 1391 in c269329
There is another instance where this is called with julia/Compiler/src/ssair/verify.jl Line 115 in c269329
No idea if an |
Bisected to #57213 (which isn't in the diff I was looking at; I probably mixed something up). |
MWE, reduced from FileTrees->Dagger->ArnoldiMethod->Graphs
module FileTrees
export FileTree, File, maketree
struct FileTree
children::Vector
end
function setparent(x::FileTree, parent=parent)
copy!(p.children, setparent.(x.children, 0))
end
maketree(node) = setparent(_maketree0)
macro spawn(exs...)
ex = exs[end]
return _par(__module__, ex)
end
function _par(mod, ex::Expr; lazy=false)
if false
elseif lazy
return Expr(ex.head, _par_inner.(Ref, ex.args, lazy=lazy, recur=recur, opts=opts)...)
end
end
function mutable()
return fetch(@spawn scope=scope _mutable_inner())
end
struct StaticArrayStyle{N} <: Base.Broadcast.AbstractArrayStyle{N} end
Broadcast._axes(bc::Base.Broadcast.Broadcasted{<:StaticArrayStyle}, ::Nothing) = static_combine_axes(bc.args...)
end
|
This check is wrong now, since Base does exist early, while it is trying to check if this verify call is part of the Compiler or "user" code. It is now getting always-true, rather than the expected desirable result (which seems to be always-false by the comment there): julia/Compiler/src/ssair/verify.jl Line 113 in c269329
|
Should fix the error part of #57329. I also cannot reproduce the underlying assertion error on master, so that was likely fixed in one of the other PRs. Closes #57329 as a result, but of course there could be other issues with the same symptoms, which can get their own issues. (cherry picked from commit 88b292d)
Many packages on PkgEval started to fail with:
Example: FileTrees.jl
I haven't bisected this yet; the whole changeset includes e485be8...79ce168, out of which #57230 looks the most suspicious (cc @Keno).
The text was updated successfully, but these errors were encountered: