You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I believe the problem is more general than stdlibs, it looks like setting JULIA_CPU_TARGET always results in compiling code only for the generic target, if we can trust the parsing done by Base.parse_image_targets and Base.parse_cache_header. Consider this script
# Cleanup Example precompile dir
example_dir =joinpath(first(Base.DEPOT_PATH), "compiled", "v$(Base.thisminor(Base.VERSION))"[begin:end-2], "Example")
rm(example_dir; force=true, recursive=true)
# Install Exampleusing Pkg
Pkg.activate(; temp=true, io=devnull)
Pkg.add("Example"; io=stdout)
# Find compile cache
pkg = Base.identify_package("Example")
cachefiles = Base.find_all_in_cache_path(pkg)
isempty(cachefiles) &&error(pkg, " has not yet been precompiled for julia ", Base.VERSION, ". Reinstall the package with `Pkg.add(\"Example\"; io=stdout)` to see why")
pkgpath = Base.locate_package(pkg)
idx =findfirst(cachefiles) do cf
Base.stale_cachefile(pkgpath, cf) !==trueend
targets = Base.parse_image_targets(Base.parse_cache_header(cachefiles[idx])[7])
# Show target@show targets
Making this an issue so I can add it to the milestone:
Ref #54093
The text was updated successfully, but these errors were encountered: