Tim Holy
From a security perspective, how much of the risk is "oops, forgot to initialize this!" vs. "I just intercepted your credit card number because it was still left over in...
Out of curiosity, why do we add backedges from `ssair/inlining.jl`? They already seem to be added by ordinary inference, but for this PR I had to make sure they were...
I think all the niggling issues have been fixed, and it even fixes a genuine invalidation bug that's been lurking in 1.8 (see f59125872deae7e5b53ba82d1f62ec839657a865). Assuming CI passes, this seems ready...
Doctest failure should have been fixed in #46021. I think the rest are unrelated.
Don't merge yet, we still need other changes. In particular, (1) we need `precompile(mi::MethodInstance)` and likely also `precompile(m::Method, invokesig)` because otherwise we can't force precompilation of an `invoke`d method that...
I rebased & squashed, and now the second tiny commit fixes (2), i.e., this now seems to solve the use cases for which it was intended. The third implements `precompile`...
As far as I can tell none of these failures are related.
I've also run some benchmarks to see if there are any load-time regressions. This represents a pair of runs (shown is centroid ± half-width) on three different Julia versions. tl/dr:...
I think this can be merged if all tests pass (they do locally). This has become quite the intrusive change, and I no longer think this is appropriate for backporting,...
- https://github.com/FluxML/MacroTools.jl/pull/166 - https://github.com/JuliaLang/julia/pull/41885 - https://github.com/JuliaLang/Pkg.jl/pull/2699 - https://github.com/JuliaAlgebra/MultivariatePolynomials.jl/pull/172 - https://github.com/FluxML/MacroTools.jl/pull/167 - https://github.com/JuliaPlots/RecipesBase.jl/pull/82 - https://github.com/JuliaArrays/StaticArrays.jl/pull/950 - https://github.com/JuliaLang/julia/pull/41917 - https://github.com/JuliaSIMD/CPUSummary.jl/issues/3 (originally the CPUSummary issue identified in https://github.com/JuliaLang/julia/pull/41913)