[WIP] preserve offset when broadcasting IdOffsetRange #105
Conversation
| @@ -155,6 +155,14 @@ end | |||
| return IdOffsetRange(r.parent[s .- r.offset], r.offset) | |||
| end | |||
|
|
|||
| # offset-preserve broadcasting | |||
| Base.broadcasted(::Base.Broadcast.DefaultArrayStyle{1}, ::typeof(-), r::IdOffsetRange{T}, x::Integer) where T = | |||
There was a problem hiding this comment.
| Base.broadcasted(::Base.Broadcast.DefaultArrayStyle{1}, ::typeof(-), r::IdOffsetRange{T}, x::Integer) where T = | |
| Broadcast.broadcasted(::Base.Broadcast.DefaultArrayStyle{1}, ::typeof(-), r::IdOffsetRange{T}, x::Integer) where T = |
and similarly below.
|
Any update on this? What still needs to be done? |
|
Sorry! I was planning to find some time to check JuliaImages/ImageFiltering.jl#160 in detail but was filled with my PhD things (near the end of this semester); my experiments didn't go as smoothly as I expected. My limited time in julia was spent on setting up a Pkg mirror in BFSU for the Chinese community. For this PR itself, it's actually a bug fix so I believe some independent and reasonable tests here would be sufficient. Since this PR is proposed for JuliaImages/ImageFiltering.jl#160 and since that is basically approved by @jw3126, I think this one is good to go as long as the test there passes (you can check that by temporarily committing a Manifest.toml there with OffsetArrays of this branch). I'm truly sorry for the unavailability, I'll do a rough check tomorrow night. Feel free to open a new PR for this if you want to polish this by yourself. |
|
Thanks. I'll open a new PR for this one with an added test.
I already checked the tests there with a locally modified |
|
closed in favor of #114 |
This is a prototype patch to test JuliaImages/ImageFiltering.jl#160
fixes #104