Backports for 1.12.2 #59920

KristofferC · 2025-10-21T07:49:10Z

Backported PRs:

Need manual backport:

Contains multiple commits, manual intervention needed:

Non-merged PRs with backport label:

ViralBShah · 2025-10-21T16:10:19Z

Hopefully the linalg bump can make it in: #59927

KSepetanc · 2025-10-23T11:31:10Z

PR #59736 is marked here as needs manual backport, but has been marked as backported in 1.12.1. Why?

KristofferC · 2025-10-23T12:23:23Z

I guess I forgot to remove the label

KristofferC · 2025-10-24T15:41:53Z

Please cherry pick with -x when pushing them here.

vtjnash · 2025-10-24T16:05:27Z

I don't know who that is directed at, but note that #59921 is not a backport, but a v1.12-only bug

…#59849) (cherry picked from commit b26e7cc)

(cherry picked from commit a0a65e6)

…59878) Noticed during bug hunting of ericphanson/LicenseCheck.jl#11 (comment) and may cause performance degradation for interpreter-executed try blocks or scope entry. Co-authored-by: Dilum Aluthge <[email protected]> (cherry picked from commit c1353ac)

(cherry picked from commit 78ba3be)

@cjdoris

This prevents a bunch of invalidations, specifically these ones in Julia 1.12 with UnsafePointers.jl (CC @cjdoris): ```julia inserting convert(P::Type{<:Ptr}, p::UnsafePointers.UnsafePtr) @ UnsafePointers ~/git/UnsafePointers.jl/src/UnsafePointers.jl:66 invalidated: mt_backedges: 1: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_npointers(::Type{T} where T<:(NamedTuple{names, T} where {T<:Tuple, names})) (4 children) 2: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_npointers(::Type{T} where T<:AbstractString) (4 children) 3: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_npointers(::Type{T} where T<:Tuple{Symbol, Any}) (4 children) 4: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_npointers(::Type{T} where T<:(NamedTuple{(:message, :position, :hint), _A} where _A)) (4 children) 5: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_npointers(::Type{T} where T<:(SubString{_A} where _A)) (4 children) 6: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_alignment(::Type{T} where T<:(NamedTuple{(:message, :position, :hint), _A} where _A)) (15 children) 7: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_alignment(::Type{T} where T<:(SubString{_A} where _A)) (19 children) 8: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_alignment(::Type{T} where T<:Tuple{Any, Symbol, Symbol}) (21 children) 9: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_alignment(::Type{T} where T<:Tuple{Symbol, Any}) (23 children) 10: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_alignment(::Type{T} where T<:AbstractString) (52 children) 11: signature Tuple{typeof(convert), Type{Ptr{Base.DataTypeLayout}}, Any} triggered MethodInstance for Base.datatype_alignment(::Type{T} where T<:(NamedTuple{names, T} where {T<:Tuple, names})) (1739 children) ``` Using Cthulhu to look at the last invalidation with 1739 children showed this: ```julia datatype_alignment(dt::DataType) @ Base ~/.julia/juliaup/julia-1.12.1+0.x64.linux.gnu/share/julia/base/runtime_internals.jl:544 544 function datatype_alignment(dt::Type{T} where T<:(NamedTuple{names, T} where {T<:Tuple, names})::DataType)::Any 545 @_foldable_meta 546 (dt::Type{T} where T<:(NamedTuple{names, T} where {T<:Tuple, names}).layout::Any == C_NULL::Core.Const(Ptr{Nothing}(0x0000000000000000)))::Any && throw(UndefRefError()::Core.Const(UndefRefError())) 547 alignment::Any = unsafe_load(convert(Ptr{DataTypeLayout}::Type{Ptr{Base.DataTypeLayout}}, dt::Type{T} where T<:(NamedTuple{names, T} where {T<:Tuple, names}).layout::Any)::Any)::Any.alignment::Any 548 return Int::Type{Int64}(alignment::Any)::Any 549 end ``` The `.layout` field was getting inferred as `Any` which caused that method to be invalidated when UnsafePointers.jl defined a new `convert()` method. Would it be possible to backport this to 1.12? UnsafePointers is a dependency of PythonCall so it's quite widely used. --------- Co-authored-by: Shuhei Kadowaki <[email protected]> (cherry picked from commit 90f4289)

This change is not necessary for `NativeInterpreter` at all, but is particularly useful for JET. For error reporting, which is the purpose of `JETAnalyzer`, it is desirable to present error causes in a user-understandable form even when concrete evaluation presents errors. This requires analysis using regular abstract interpretation (constant propagation) in general. However, to reduce false positives, the precision that concrete evaluation brings is very important, and completely disabling concrete evaluation is not desirable either. Currently, `JETAnalyzer` aims to avoid this tradeoff by limiting concrete evaluation to only some functions, but this approach does not scale and occasionally causes problems like #59884. This commit enables ad-hoc cancellation of concrete evaluation based on the result of `concrete_eval_call`, allowing `JETAnalyzer` to fallback to abstract interpretation only when concrete evaluation causes errors, fundamentally avoiding such problems. (cherry picked from commit c6091de)

…iew` (#59915) When packages using non-Base compilers call `Base._which`, a method error occurs due to module context mismatch. Like other reflection functions, the module context of `mt::Compiler.MethodTableView` needs to be properly determined.

@ViralBShah

#59927) Stdlib: LinearAlgebra URL: https://github.com/JuliaLang/LinearAlgebra.jl.git Stdlib branch: release-1.12 Julia branch: backports-release-1.12 Old commit: 24f5e21 New commit: 5567504 Julia version: 1.12.1 LinearAlgebra version: 1.12.0(Does not match) Bump invoked by: @ViralBShah Powered by: [BumpStdlibs.jl](https://github.com/JuliaLang/BumpStdlibs.jl) Diff: JuliaLang/LinearAlgebra.jl@24f5e21...5567504 ``` $ git log --oneline 24f5e21..5567504 5567504 Backports to v1.12 (#1472) a2a4981 Overload array constructors for BunchKaufman (#1461) (#1466) 1ebc0cb Generic fallback for `fillstored!` (#1389) ``` Co-authored-by: ViralBShah <[email protected]>

…all_name` (#59921) This prevents invalidations of `isready()` from loading Distributed on 1.12: ```julia inserting isready(pool::Distributed.WorkerPool) @ Distributed ~/.julia/juliaup/julia-1.12.1+0.x64.linux.gnu/share/julia/stdlib/v1.12/Distributed/src/workerpool.jl:168 invalidated: mt_backedges: 1: signature Tuple{typeof(isready), Any} triggered MethodInstance for (::Compiler.var"#evalargs#abstract_eval_nonlinearized_foreigncall_name##0"{Expr, Compiler.StatementState, Compiler.Future{Compiler.CallMeta}, Vector{Any}, Int64})(::Compiler.AbstractInterpreter, ::Compiler.InferenceState) (0 children) 2: signature Tuple{typeof(isready), Any} triggered MethodInstance for (::Compiler.var"#evalargs#abstract_eval_nonlinearized_foreigncall_name##0"{Expr, Compiler.StatementState, Compiler.Future{Compiler.CallMeta}, Vector{Any}, Int64})(::Compiler.NativeInterpreter, ::Compiler.InferenceState) (0 children) 3: signature Tuple{typeof(isready), Any} triggered MethodInstance for (::Compiler.var"#evalargs#abstract_eval_nonlinearized_foreigncall_name##0"{Expr, Compiler.StatementState, Compiler.Future{Compiler.CallMeta}, Vector{Any}, Int64})(::Compiler.AbstractInterpreter, ::Compiler.InferenceState) (249 children) ``` I think ideally it wouldn't use a closure at all, but I'm not familiar enough with the code to refactor it that much. I'm making the PR against the `release-1.12` branch because this function is deleted entirely on master by #59165.

@JeffBezanson

Renaming happened in f9b7d27 / PR #59158 by @JeffBezanson

…59941) Co-authored-by: Erik Schnetter <[email protected]>

Closes #59901 (cherry picked from commit a1adecb)

Followup of #54609 (cherry picked from commit c2f42a6)

fix a bug in which parsing the digits after the decimal point of a (date-)time like "00:00:00.12" were interpreted as "12 milliseconds" instead of "0.12 seconds". Now, the fractional part is correctly parsed as a fraction of a second, then correctly converted to milliseconds. As before, only the first three digits after the decimal point are considered. Closes #59997 --- With a local build of this PR, the example from the above issue now evaluates to: ```julia julia> t2 = TOML.parse("""time = 00:00:00.2""") Dict{String, Any} with 1 entry: "time" => 00:00:00.2 julia> millisecond(t2["time"]) 200 ``` (cherry picked from commit 906d64e)

When encountering complicated cycles within cycles, make sure to revisit all applicable methods explicitly, since not all methods within the cycle will necessarily change return type after resolving the cycle. This shows up as the possibility of frames that don't get revisited before being cached, probably just caching `Union{}` as the call type instead. I always assumed this code was probably wrong, but didn't have any way to construct the counter-example to have confidence that fixing it would not cause some other side-effect. But still keep the backedge work lazy, since we don't want to allocate unnecessarily for a rarely used feature (recursion). Fix #59943 (cherry picked from commit f4847bf)

(cherry picked from commit a933cf4)

fixes #60009 (cherry picked from commit 97f880a)

(cherry picked from commit 6f2f7f5)

When errors occur, `waitall` may skip allocating Channel producers, leading to deadlock in the subsequent loop in the event that the user asked it to failfast (ironically). This is seen often in the failing of the threads_exec test ever since the test was added for this call. Simplify this to just use separate loops for the wait and the return computation. (cherry picked from commit e2f3178)

KristofferC · 2025-11-08T17:58:06Z

LoadError: IOError: could not spawn setenv(`7z x /private/var/tmp/agent-tempdirs/default-honeycrisp-XC9YQX9HH2.0/tmp/jl_l2pSjSpC3w-download.gz -so`)

on CI. Due to 7z bump? But why...

https://github.com/JuliaLang/Pkg.jl/blob/499e7d35565c0c8582d0d7b56bfe2109dcbac4c0/src/PlatformEngines.jl#L19-L21 should work. Any idea @nhz2?

nhz2 · 2025-11-08T19:17:53Z

Why is the 7z not getting the full path? Shouldn't it be something like /home/nathan/.julia/juliaup/julia-1.12.1+0.x64.linux.gnu/libexec/julia/7z

KristofferC · 2025-11-08T19:36:04Z

Yes, that's what I linked. Locally it seems ok...

julia> using p7zip_jll

julia> p7zip_jll.is_available()
true

julia> using Pkg

julia> Pkg.PlatformEngines.exe7z()
setenv(`/Users/kc/julia/usr/libexec/julia/7z

Even with this PR in juliaup it seems ok. I noticed the new p7zip_jll has a CSL dependency, could that be out of date on 1.12 or something? As an experiment, I will revert the 7z bump just to see what happens.

Edit: Actually, I don't have to revert because we have a CI run on that exact commit:

ba1baa8

KristofferC · 2025-11-08T19:48:23Z

It might be conflicting with 671cd64#diff-25078ad6c1e38d5b2a1a23baabf2e032f1f16d995cbdf47e95e2100e1d7786eb maybe.

nhz2 · 2025-11-08T19:54:32Z

Also, will the new CSL dep break existing Manifests for 1.12?

(cherry picked from commit fda084f)

In Markdown, links are written as `[desc](URL)`, and no space is allowed between `]` and `(`. However, Julia's Markdown parser explicitly accepted it (not clear why). This may break links in a few package manuals, but the fix will be easy in each case (just remove the extra spaces). But being more consistent with most (all??) other Markdown implementations seems more important to me; it is very surprising when one copies basic Markdown that works fine elsewhere into a Julia package manual and it suddenly is treated completely differently. Resolves JuliaDocs/Documenter.jl#2681 (cherry picked from commit fe30573)

KristofferC · 2025-11-08T20:59:02Z

I tried with some help from Claude with 87f5619 but didn't seem to help.

deps/p7zip.mk

KristofferC · 2025-11-09T12:48:23Z

Ah, it is not there. If I download the build artifact:

~/Downloads/julia-a81d1ec833
❯ cd libexec/julia   

~/Downloads/julia-a81d1ec833/libexec/julia
❯ ls
dsymutil lld

no 7z...

I backported some more from #58344, we'll see how that goes. I built the binary dist locally and it seems to be in there now at least.

KristofferC · 2025-11-09T13:31:13Z

Tangentially, the new 7-zip seems about 2x as fast as the old one, but still has nothing on zstd (timings for uncompressing the registry in gz and zstd format below)

7-zip 17.05:
  191.101 ms (7503 allocations: 239.80 MiB)
7-Zip (a) 25.01:
  95.626 ms (7019 allocations: 239.77 MiB)
zstd:
  24.648 ms (3568 allocations: 305.57 MiB)

Bring over the p7zip_jll handling improvements from f03e9c3 ([deps] enable zstd support): - Move p7zip binary from bindir to private_libexecdir - Add Windows DLL handling for binary builder installs - Simplify p7zip_jll.jl by removing LIBPATH handling - Add P7ZIP_BUILD_OPTS variable for consistent build flags These changes prepare for the p7zip 17.7.0 upgrade and ensure p7zip is installed in the correct location. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <[email protected]>

KristofferC · 2025-11-09T20:00:29Z

@nanosoldier runtests()

nanosoldier · 2025-11-11T08:19:51Z

The package evaluation job you requested has completed - possible new issues were detected.
The full report is available.

Report summary

❗ Packages that crashed

2 packages crashed only on the current version.

The process was aborted: 1 packages
Invalid LLVM IR was generated: 1 packages

46 packages crashed on the previous version too.

✖ Packages that failed

45 packages failed only on the current version.

Illegal method overwrites during precompilation: 1 packages
Package has test failures: 3 packages
Package tests unexpectedly errored: 20 packages
Tests became inactive: 1 packages
Test duration exceeded the time limit: 20 packages

3472 packages failed on the previous version too.

✔ Packages that passed tests

36 packages passed tests only on the current version.

Other: 36 packages

6145 packages passed tests on the previous version too.

➖ Packages that were skipped altogether

6 packages were skipped only on the current version.

Package could not be installed: 6 packages

1385 packages were skipped on the previous version too.

KristofferC · 2025-11-11T11:21:52Z

@nanosoldier runtests(["XAIBase", "Timestamps64", "JuliaInterpreter", "PiecewiseLinearFunctions", "FunctionOperators", "VanVleckRecursion", "C3D", "Polynomials", "BipartiteFactorGraphs", "SparseMatrixColorings", "ConstrainedShortestPaths", "Hecke", "VectorizationBase", "StatsFuns", "Andes", "SawyerEliassenSolver", "ImageMagick", "ExplainableAI", "RegressionTables", "MultiObjectiveAlgorithms", "SDiagonalizability", "CompressedSensing", "RegressionDynamicCausalModeling", "CellMetabolismBase", "CellMetabolism", "SatelliteAnalysis", "ProbabilisticParameterEstimators", "NonconvexMultistart", "SphericalFunctions", "CensoredDistributions", "SecondQuantizedAlgebra", "PowerAnalytics", "SimpleSDMPolygons", "HierarchicalEOM", "LongwaveModePropagator", "TenSolver", "SpiDy", "GeometricFlux", "GNSSReceiver", "PhysicsInformedRegression", "BoxCox", "EcoNetPostProcessing", "ReactionDiffusion"])

nanosoldier · 2025-11-12T11:32:51Z

The package evaluation job you requested has completed - possible new issues were detected.
The full report is available.

Report summary

✖ Packages that failed

18 packages failed only on the current version.

Package tests unexpectedly errored: 18 packages

7 packages failed on the previous version too.

✔ Packages that passed tests

18 packages passed tests on the previous version too.

KristofferC · 2025-11-12T21:03:04Z

@nanosoldier runtests(["XAIBase", "SawyerEliassenSolver", "SDiagonalizability", "BipartiteFactorGraphs", "PiecewiseLinearFunctions", "VanVleckRecursion", "RegressionDynamicCausalModeling", "JuliaInterpreter", "SecondQuantizedAlgebra", "StatsFuns", "C3D", "ProbabilisticParameterEstimators", "CellMetabolismBase", "CensoredDistributions", "ExplainableAI", "ConstrainedShortestPaths", "CellMetabolism", "HierarchicalEOM"])

JuliaRegistries/General#142382 should have fixed many of these

nanosoldier · 2025-11-12T22:28:53Z

The package evaluation job you requested has completed - possible new issues were detected.
The full report is available.

Report summary

✖ Packages that failed

7 packages failed only on the current version.

Package tests unexpectedly errored: 7 packages

✔ Packages that passed tests

11 packages passed tests on the previous version too.

KristofferC added the release Release management and versioning. label Oct 21, 2025

KristofferC changed the title ~~set VERSION to 1.12.1 (#59871)~~ Backports for 1.12.2 Oct 21, 2025

DilumAluthge added the don't squash Don't squash merge label Nov 2, 2025

KristofferC and others added 12 commits November 4, 2025 20:24

allow specifying projects in a workspace more than one level under it (…

166c99b

…#59849) (cherry picked from commit b26e7cc)

update llvmcall docs, opaque pointers are now supported (#59873)

4b01fcb

(cherry picked from commit a0a65e6)

Type-assert isfinite(::AbstractFloat) (#59888)

d526618

(cherry picked from commit 78ba3be)

Update left-over reference to Core.GlobalMethods (#59976)

ec3201f

Renaming happened in f9b7d27 / PR #59158 by @JeffBezanson

[backports-release-1.12] OpenSSL: Update to version 3.5.4 (#59710) (#…

e611d7f

…59941) Co-authored-by: Erik Schnetter <[email protected]>

[backports-release-1.12] Bump LibCURL_jll to 8.15.0 (#59983)

d0161e4

DilumAluthge force-pushed the backports-release-1.12 branch from 8d0ca62 to d0161e4 Compare November 5, 2025 01:24

jakobnissen and others added 8 commits November 7, 2025 19:09

Fix: waitall can throw when all tasks are done (#59905)

f9733bd

Closes #59901 (cherry picked from commit a1adecb)

Names docstring: compat & implicit using (#59937)

edf9fa7

Followup of #54609 (cherry picked from commit c2f42a6)

REPL: make interactive precompiles test more robust (#60006)

2fc0ebd

(cherry picked from commit a933cf4)

fix pointerarith_tfunc for Const ptr (#60011)

6286cfd

fixes #60009 (cherry picked from commit 97f880a)

Don't trigger full rebuild if git is dirty (#60023)

f8f009d

(cherry picked from commit 6f2f7f5)

KristofferC force-pushed the backports-release-1.12 branch from fe7f9d4 to c53224f Compare November 7, 2025 18:09

JamesWrigley and others added 2 commits November 8, 2025 21:00

Type-assert the return type of collect(...) in TOML (#59932)

fc0e312

(cherry picked from commit fda084f)

KristofferC force-pushed the backports-release-1.12 branch from 332e443 to a81d1ec Compare November 8, 2025 20:11

nhz2 reviewed Nov 8, 2025

View reviewed changes

deps/p7zip.mk Show resolved Hide resolved

nhz2 mentioned this pull request Nov 9, 2025

[p7zip] zipitem patch for 17.6 JuliaPackaging/Yggdrasil#12498

Merged

KristofferC force-pushed the backports-release-1.12 branch from a81d1ec to 2c8b3b5 Compare November 9, 2025 13:08

KristofferC and others added 4 commits November 9, 2025 20:59

bump Pkg to latest 1.12

4a79cf9

p7zip 17.7.0 (#60025)

84f859e

fix string completion with cursor in the middle of text

d9323c0

KristofferC force-pushed the backports-release-1.12 branch from 2c8b3b5 to d9323c0 Compare November 9, 2025 20:00

KristofferC merged commit bdc35b2 into release-1.12 Nov 13, 2025
8 checks passed

KristofferC deleted the backports-release-1.12 branch November 13, 2025 10:43

pankgeorg mentioned this pull request Nov 13, 2025

Julia 1.12 + PythonCall => Crash fonsp/Pluto.jl#3414

Open

Uh oh!

Backports for 1.12.2 #59920

Backports for 1.12.2 #59920

Conversation

KristofferC commented Oct 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ViralBShah commented Oct 21, 2025

Uh oh!

KSepetanc commented Oct 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

KristofferC commented Oct 23, 2025

Uh oh!

KristofferC commented Oct 24, 2025

Uh oh!

vtjnash commented Oct 24, 2025

Uh oh!

KristofferC commented Nov 8, 2025

Uh oh!

nhz2 commented Nov 8, 2025

Uh oh!

KristofferC commented Nov 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

KristofferC commented Nov 8, 2025

Uh oh!

nhz2 commented Nov 8, 2025

Uh oh!

KristofferC commented Nov 8, 2025

Uh oh!

Uh oh!

KristofferC commented Nov 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

KristofferC commented Nov 9, 2025

Uh oh!

KristofferC commented Nov 9, 2025

Uh oh!

nanosoldier commented Nov 11, 2025

❗ Packages that crashed

✖ Packages that failed

✔ Packages that passed tests

➖ Packages that were skipped altogether

Uh oh!

KristofferC commented Nov 11, 2025

Uh oh!

nanosoldier commented Nov 12, 2025

✖ Packages that failed

✔ Packages that passed tests

Uh oh!

KristofferC commented Nov 12, 2025

Uh oh!

nanosoldier commented Nov 12, 2025

✖ Packages that failed

✔ Packages that passed tests

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

19 participants

KristofferC commented Oct 21, 2025 •

edited

Loading

KSepetanc commented Oct 23, 2025 •

edited

Loading

KristofferC commented Nov 8, 2025 •

edited

Loading

KristofferC commented Nov 9, 2025 •

edited

Loading