RFC: support mmap on Windows #3603
Conversation
|
I should mention that I deprecated the "easy" version of |
|
Thanks, Tim! @loladiro and @vtjnash should review, of course. |
|
|
||
| msync{T}(A::Array{T}) = msync(pointer(A), length(A)*sizeof(T)) | ||
|
|
||
| munmap{T}(A::Array{T}) = munmap(pointer(A), length(A)*sizeof(T)) |
There was a problem hiding this comment.
This is a bit of a foot gun, no? After munmap(A), doing anything at all with A will cause Julia to segfault.
There was a problem hiding this comment.
Agreed. It might be better to only munmap in the finalizer. Forcing a GC to munmap something is a bit awkward though.
There was a problem hiding this comment.
munmap does get called in the finalizer. Perhaps we shouldn't export it, though, or provide an interface beyond the pointer-based one. I'll change that.
|
I discovered a sneaky way to test it (sortof) on the latest Windows binary: I had to hack |
|
Thanks Tim, this is great to have. One concern that bothers me is that you'll get a segfault when trying to write to a mmap'ed region that is declared as read-only. Though I am not really sure what to do about that. @JeffBezanson any ideas? |
|
See also #3434 |
Stdlib: Pkg URL: https://github.com/JuliaLang/Pkg.jl.git Stdlib branch: master Julia branch: master Old commit: 047734e4c New commit: f570abd39 Julia version: 1.11.0-DEV Pkg version: 1.11.0 Bump invoked by: @KristofferC Powered by: [BumpStdlibs.jl](https://github.com/JuliaLang/BumpStdlibs.jl) Diff: JuliaLang/Pkg.jl@047734e...f570abd ``` $ git log --oneline 047734e4c..f570abd39 f570abd39 tweak how Pkg is loaded for precompiling when testing (#3606) d3bd38b90 sort compat entries in `pkg> compat` (#3605) 5e07cfed0 Ensure that `string(::VersionSpec)` is compatible with `semver_spec()` (#3580) 6bed7c41a Clarify handling of LOAD_PATH in docstring and REPL help (#3589) 5261b3be7 tweak test dep docs (#3574) 72b430d50 status: expand 2 symbol upgrade note to highlight *may* be upgradable (#3576) b32db473d precompile: show ext parent in output report (#3603) ``` Co-authored-by: Dilum Aluthge <dilum@aluthge.com>
I don't have a build environment set up for Windows, so I haven't tested this directly. However, I tested the "ingredients" successfully in a standalone file, so the odds are not entirely zero for this working. However, if it's not practical for anyone else to test this, I can (with some time) give that a try.
It does pass tests on Linux; I should mention that I have seen some test failures, but they don't seem to be consistent, and I got this to pass testall three times in a row before submitting.