Kristoffer Carlsson

Stack overflow on `pow` with mixed float and complex args

2

``` julia> NaNMath.pow(1.0, 1.0+im) ERROR: StackOverflowError: Stacktrace: [1] pow(x::ComplexF64, y::ComplexF64) (repeats 2 times) @ NaNMath ~/.julia/packages/NaNMath/ceWIc/src/NaNMath.jl:26 ```

bump the version

1

This is required to get rid of the dependency on Test that comes from the pre Project.toml era.

https://github.com/JuliaRegistries/General/pull/93498 is buggy but passed CI. The missing check is that for every version, each compat entry needs to have a corresponding entry with that name in `Deps.toml`.

Just like normal dependencies

DelimitedFiles should now require compat entries

3

In https://github.com/JuliaRegistries/General/pull/66039, compat for DelimtiedFiles was added so that when people start using 1.9 (where DelimitedFiles is not an stdlib) the resolver should not start picking old versions without DelimitedFiles...

Ambiguity on nightly:

3

``` 1 ambiguities found Ambiguity #1 polygamma(a::Number, b::AbstractNumbers.AbstractNumber) in AbstractNumbers at /home/pkgeval/.julia/packages/AbstractNumbers/ONC07/src/overloads.jl:1251 polygamma(m::Integer, z::Number) in SpecialFunctions at /home/pkgeval/.julia/packages/SpecialFunctions/LC8dm/src/gamma.jl:562 ``` Seems like a real ambiguity, or?

Allow to make a comparison between two commit SHAs from an issue.

2

Sometimes more convenient when having a discussion about a regression than having to go to the actual commits.

add some latency benchmarks

3

Fixes #251

Runtime of functions is nice, but it is also nice to track things like time to first plot. Incorporating a set of compilation time benchmarks is a bit tricky because...

https://github.com/JuliaCI/BaseBenchmarks.jl/blob/9306a2797514518052189bcb4a70fbc73cf3009e/src/problem/Laplacian.jl#L41-L51 It is just benchmarking the setup of the sparse matrix (effectively testing sparse matmat)