Kristoffer Carlsson
Kristoffer Carlsson
> The part which removes the fast path in which we check if (n_threads_marking == 0) outside of the lock is relevant to 1.10. Is this, or can it be...
> At the leaves of the tree, the nodes are tiny (2-by-2 can occur) so of course only a single thread should be used inside a call to the BLAS....
> But note that Pkg.REPLExt is still broken, as I inserted a call to the wrong insert_extension_triggers function, and using locate_package_env as an example is not helpful to derive the...
Would also be useful to see if there is anyway to show a regression from this PR on a non-synthetic benchmark. That was done in the first post in this...
Yeah, at that point you might as well just hide the pointer value and show it as `Ptr{Int64}(🖕)`
Without doing any analyzing it seems that the results are consistent on 1.7 at least.
@nanosoldier `runtests(vs =":release-1.10")`
Regarding `annotate!` in InMemoryDataSets.jl, probably fixed in https://github.com/sl-solution/InMemoryDatasets.jl/pull/117/. Not sure what that code is really doing though. @PallHaraldsson, please try to keep on topic for this PR, if you want...
@tecosaur, can you push your commits you want backported to here. I get merge conflicts for some and it is easier if you just do it.
I think it was https://github.com/JuliaLang/julia/pull/53715 as well