Backports for 1.12.5 #60612
Backports for 1.12.5 #60612
Conversation
…55 (#60505) Stdlib: NetworkOptions URL: https://github.com/JuliaLang/NetworkOptions.jl.git Stdlib branch: master Julia branch: backports-release-1.12 Old commit: 532992f New commit: 7034c55 Julia version: 1.12.3 NetworkOptions version: 1.3.0 (Does not match) Bump invoked by: @fingolfin Powered by: [BumpStdlibs.jl](https://github.com/JuliaLang/BumpStdlibs.jl) Diff: JuliaLang/NetworkOptions.jl@532992f...7034c55 ``` $ git log --oneline 532992f..7034c55 7034c55 Bump actions/checkout from 4 to 5 (#44) 46e14ef Merge pull request #43 from lgoettgens/lg/JET-ca_roots-path_convert ``` Co-authored-by: fingolfin <[email protected]>
(cherry picked from commit 7450aba)
(cherry picked from commit b6337dd)
libuv reports times in milliseconds, but the Julia docs and show() use units of 1/Sys.SC_CLK_TCK which is 100 on my laptop. The mismatch might be due to libuv changeset libuv/libuv@37a8aff. Fixes #53577 (cherry picked from commit 3d611fd)
Dot operators (e.g., .+, .-) are lowered to `BroadcastFunction` and do not have corresponding function definitions. Fix the regression by adding the needed wrapper. Test added. --------- Co-authored-by: Shuhei Kadowaki <[email protected]> (cherry picked from commit 7ba24a7)
`unpack_platform` might return nothing in case of an error. Together with PR #60506 and my previous changes, when applied to release-1.12, the JET report for GAP.jl is now free of reports. Well, if I filter out certain "unavoidable issues": ```julia using Revise, JET, AbstractAlgebra, GAP struct AnyFrameMethod <: ReportMatcher m::Union{Function,Method,Symbol} end function JET.match_report(matcher::AnyFrameMethod, @nospecialize(report::JET.InferenceErrorReport)) # check all VirtualFrames in the VirtualStackTrace for a match to the specified method m = matcher.m if m isa Symbol return any(vf -> vf.linfo.def.name === m, report.vst) elseif m isa Method return any(vf -> vf.linfo.def === m, report.vst) else # if m isa Function return any(vf -> vf.linfo.def in methods(m), report.vst) end end report_package(GAP; ignored_modules=[AnyFrameMethod(:is_loaded_directly)]) ``` (cherry picked from commit 8123a5c)
DilumAluthge
force-pushed
the
backports-release-1.12
branch
from
8354ea7 to
89dc101
Compare
|
Correct me if I am wrong. Continuing discourse discussion, PR #4444 should be backported. It had backport 1.12 label, has been tagged for 1.12.0 and then it has not been backported and its backport label has been removed. |
This isn't a direct backport of https://github.com/JuliaLang/julia/pull/60640/changes because because `llvm.powi.f16.i32` is very broken on LLVM 18 (at least for targets without hardware fp16 support like x86). This is only a performance backport, but it is a pretty important regression that was introduced in 1.12.
(cherry picked from commit 657273a)
This might cause double frees/use after frees. Found on the mimalloc PR (cherry picked from commit 06def89)
Since this task's stack or scope field could have been modified after it was marked by an incremental collection (and not just for copy stacks), move the barrier back unconditionally here. --------- Co-authored-by: Valentin Churavy <[email protected]> Co-authored-by: Jeff Bezanson <[email protected]> (cherry picked from commit 14ca1ab)
(cherry picked from commit 346b4e1)
Okay, I took a look at JuliaLang/Pkg.jl#4444, but unfortunately it doesn't backport cleanly, so it needs to be manually backported. I did create the |
…#60703) Stdlib: SparseArrays URL: https://github.com/JuliaSparse/SparseArrays.jl.git Stdlib branch: release-1.12 Julia branch: backports-release-1.12 Old commit: f81a30d New commit: 2376bf8 Julia version: 1.12.4 SparseArrays version: 1.12.0 (Does not match) Bump invoked by: @dkarrasch Powered by: [BumpStdlibs.jl](https://github.com/JuliaLang/BumpStdlibs.jl) Diff: JuliaSparse/SparseArrays.jl@f81a30d...2376bf8 ``` $ git log --oneline f81a30d..2376bf8 2376bf8 Backports to v1.12.5 (#669) e545b39 Update CI workflow for Julia and checkout action (#659) e129aea Separate out tests for the GPL build and do not run them when GPL libs are not present (#658) ``` Co-authored-by: dkarrasch <[email protected]>
(cherry picked from commit f9d461f)
DilumAluthge
force-pushed
the
backports-release-1.12
branch
from
cce3c45 to
6f29f9b
Compare
DilumAluthge
force-pushed
the
backports-release-1.12
branch
3 times, most recently
from
792f0f6 to
8d88074
Compare
(cherry picked from commit 14945c0)
…60740) This backports the part of #60645 that is applicable to 1.12. Furthermore, I added a type assertion two lines above the other changes that was taken from #58430. Without this, there are still JET warnings in 1.12 for the above query. cc @DilumAluthge
Should fix one of the GPUCompiler caching issues --------- Co-authored-by: Valentin Churavy <[email protected]> Co-authored-by: Dilum Aluthge <[email protected]>
The way this ABI optimization is implemented is a bit awkward, since it means that either the "bytes" have to start acting like "pointers" or vice-versa. This adds some minimal plumbing to make sure that we label these as roots appropriately (and zero-init them).
Inherit alignment from the original GC allocation with JL_SMALL_BYTE_ALIGNMENT as the minimum. Use alignment-sized integer chunks for the alloca type (matching emit_static_alloca) so SROA splits allocations into aligned pieces for better performance and vectorization. Also adds the missing setAlignment call in splitOnStack. Co-Authored-By: Claude Opus 4.5 <[email protected]> --------- Co-authored-by: Claude Opus 4.5 <[email protected]> (cherry picked from commit 54fde7e)
We were missing handling for the case where the binding that we're using is ambiguous. There are two possible behaviors: 1. The ambiguous binding gets ignored for the purpose of resolution 2. The ambiguity poisons and the imported binding is also ambiguous Current behavior between these two depends on resolution order (which is bad and part of what the assert was complaining about). This decides that case #2 is the correct behavior and fixes #60659. (cherry picked from commit 0f275e3)
(cherry picked from commit 23a98ff)
|
@nanosoldier |
|
@nanosoldier |
|
The package evaluation job you requested has completed - possible new issues were detected. Report summary❗ Packages that crashed3 packages crashed only on the current version.
51 packages crashed on the previous version too. ✖ Packages that failed40 packages failed only on the current version.
3557 packages failed on the previous version too. ✔ Packages that passed tests38 packages passed tests only on the current version.
6309 packages passed tests on the previous version too. ➖ Packages that were skipped altogether1394 packages were skipped on the previous version too. |
KristofferC
force-pushed
the
backports-release-1.12
branch
from
8b003c8 to
a1391e5
Compare
That is, if it is called twice (e.g. because precompilation is disabled) with identical arguments, then this should not trigger an assertion. This should resolve oscar-system/GAP.jl#1314 (cherry picked from commit 1708d99)
|
@nanosoldier |
|
I have pushed one commit (fc45339) that was pushed after my PkgEval invocation. fc45339 updates the checksums for Pkg; it only modifies four files under So I don't think fc45339 invalidates my PkgEval run, nor does it require a new PkgEval run. |
During incremental compaction a forwarded NewSSAValue with a negative id was treated like an SSAValue, producing SSAValue(-n) and tripping `renumber_ssa2` bounds checks. Only convert NewSSAValue to SSAValue when the id is positive, so new_new_nodes references remain valid. Fixes #57827.
Found by claude-code debugging a random zygote issue: SciML/NeuralPDE.jl#1020 (comment). `spvals` just doesn't exist within this function. This issue has apparently existed since Julia 1.6
|
We just ran a PkgEval? Why did you run another one? |
`MustAlias` has `has_extended_unionsplit(::MustAliasLattice) === true` so it is treated as a target for union split, but `uniontypes` does not handle `MustAlias` specially, causing union split to fail for `MustAlias`. This issue is addressed by always applying `uniontypes` to the result of `widenconst(x)`. External custom lattice implementations could previously provide their own `uniontypes`, but it would be cleaner interface design if they just overload `widenconst` for their custom lattice elements so that it returns `Union`-type.
|
The package evaluation job you requested has completed - possible new issues were detected. Report summary❗ Packages that crashed3 packages crashed only on the current version.
45 packages crashed on the previous version too. ✖ Packages that failed40 packages failed only on the current version.
3567 packages failed on the previous version too. ✔ Packages that passed tests54 packages passed tests only on the current version.
6288 packages passed tests on the previous version too. ➖ Packages that were skipped altogether6 packages were skipped only on the current version.
1394 packages were skipped on the previous version too. |
|
@nanosoldier |
|
The package evaluation job you requested has completed - possible new issues were detected. Report summary✖ Packages that failed5 packages failed only on the current version.
2 packages failed on the previous version too. ✔ Packages that passed tests2 packages passed tests only on the current version.
26 packages passed tests on the previous version too. |
20 participants
Backported PRs:
full_nameafter precompiling extension #60456cpu_infotimes are expressed in milliseconds #60480@allocated 1 .+ 1#60494getproperty(::DirEntry)#60506artifact_meta#60507@fastmath x^2inlining regression to 1.12 #60686usingan ambiguous binding #60804jl_reinit_foreign_typeidempotent even with asserts #60827Need manual backport:
parse_brackets#60403parse_brackets()#60476show(::IOBuffer, ::StackFrame)#60645emit_static_rootsfor "allroots" case #60656jl_method_lookup_by_ttfor custom method tables. #60718@invokelatest#60727Contains multiple commits, manual intervention needed:
Non-merged PRs with backport label:
jl_exit_threaded_regionexecuted whenthreading_runfinished or be interrupted #60822inline_cost_threshold == typemax(Int)#60121Core._apply_iterate#60062