Bump the all-julia-packages group across 7 directories with 12 updates

[shravanngoswamii]

Supersedes #1433.

Same dependency bump dependabot proposed there, opened from a plain branch instead of dependabot/julia/* so julia-runtest's force_latest_compatible_version (which auto-forces true purely by matching the branch name against dependabot/julia*/compathelper/*) doesn't force-pin every compat entry to the top of its own range independently of what's actually installable.

The CI failures on #1433 were exactly that: force_latest_compatible_version=true demands OrderedCollections sit at literally 2.0.1, but MCMCChains caps at OrderedCollections = "1.4" across every version DynamicPPL allows (checked 6.0.0 through 7.7.0 directly against MCMCChains' own Project.toml -- all of them). Under normal resolution this isn't a conflict at all: OrderedCollections just resolves to 1.x, same as before the bump, since MCMCChains' ceiling never moves regardless of what DynamicPPL's own compat permits. This is a genuine false positive, not a real issue -- confirmed by resolving cleanly on Julia 1.10, 1.11, and 1.12 (all three CI matrix versions) with no forcing involved.

Verified locally before pushing:

• Pkg.resolve() succeeds on test/Project.toml against Julia 1.10, 1.11, and 1.12
• Full test suite (GROUP=Group1 and GROUP=Group2) passes on Julia 1.10 (the min job, furthest from what's already validated): 61959 + further tests, all pass (8 pre-existing @test_broken markers, not failures)
• The enzyme, floattypes, marginallogdensities, and reversediff jobs already passed on the original PR (those don't hit force_latest_compatible_version the same way), so their new exact-version compat pins (ADTypes, AbstractPPL, Bijectors, DifferentiationInterface, Enzyme, ForwardDiff, Distributions, LogDensityProblems) needed no further changes

No other changes were needed -- unlike the AbstractPPL.jl and AdvancedHMC.jl dependabot PRs bumped around the same time, this one didn't have a real version-floor mismatch (like a package needing a newer Julia than what's tested) hiding behind the force-latest false positive.

---

Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.64%. Comparing base (5989b33) to head (0823bc6).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1435   +/-   ##
=======================================
  Coverage   81.64%   81.64%           
=======================================
  Files          50       50           
  Lines        3579     3579           
=======================================
  Hits         2922     2922           
  Misses        657      657           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

DynamicPPL.jl documentation for PR #1435 is available at:
https://TuringLang.github.io/DynamicPPL.jl/previews/PR1435/

[github-actions]

Benchmarks @ b2c4923

Performance Ratio: gradient time divided by log-density time.

For very small models these ratios are noisy across runs and machines; raw primal and gradient timings are more reliable. The benchmarks are aimed at DynamicPPL developers and mainly catch obvious allocation or type-stability regressions. See benchmark notes for details.

===================================================================================================
                                               eval                       gradient                 
                                            ----------  -------------------------------------------
Model                        dim    linked      primal     FwdDiff    RvsDiff    Mooncake    Enzyme
---------------------------------------------------------------------------------------------------
Simple assume observe*         1     false     4.63 ns       12.55    1464.83       39.85     12.81
Simple assume observe*         1      true     4.63 ns       12.85    1641.32       40.33     12.91
Smorgasbord                  201     false     5.91 μs       71.23     134.72        6.98      9.72
Smorgasbord                  201      true     7.63 μs       74.74     139.79        6.50      7.02
Loop univariate 1k          1000     false     17.8 μs      965.17     301.76        8.43      6.77
Loop univariate 1k          1000      true     19.1 μs     1412.05     288.43        7.66      6.34
Multivariate 1k             1000     false     23.2 μs      360.03      75.47        9.03      3.07
Multivariate 1k             1000      true     29.3 μs      268.63      60.21        6.80      3.17
Loop univariate 10k        10000     false    172.0 μs    12044.13     331.50        8.44      6.72
Loop univariate 10k        10000      true    188.0 μs    12255.90     309.11        7.87      6.21
Multivariate 10k           10000     false    197.0 μs     5067.26      90.76       11.43      2.31
Multivariate 10k           10000      true    198.0 μs     5115.79      90.84       11.40      2.29
Dynamic                       15     false     1.42 μs         err      46.42       15.97     11.37
Dynamic                       10      true     1.94 μs        1.93      58.50       14.51     21.27
Submodel*                      1     false     4.64 ns       12.68    1751.53       26.35     12.97
Submodel*                      1      true     4.64 ns       12.58    1923.77       40.64     13.11
LDA                           12      true     24.0 μs        0.46       1.95       33.41       err
===================================================================================================

Main @ a49099c

===================================================================================================
                                               eval                       gradient                 
                                            ----------  -------------------------------------------
Model                        dim    linked      primal     FwdDiff    RvsDiff    Mooncake    Enzyme
---------------------------------------------------------------------------------------------------
Simple assume observe*         1     false     4.23 ns       14.24    1493.25       46.43     19.86
Simple assume observe*         1      true      4.2 ns       14.68    1738.30       48.48     18.89
Smorgasbord                  201     false     5.77 μs       77.40     135.28        7.09      7.66
Smorgasbord                  201      true     7.42 μs       82.78     141.00        6.24      5.52
Loop univariate 1k          1000     false     16.3 μs     1163.74     324.33        9.84      8.01
Loop univariate 1k          1000      true     19.8 μs     1587.32     270.03        8.52      6.40
Multivariate 1k             1000     false     24.5 μs      342.53      68.19        8.35      1.97
Multivariate 1k             1000      true     22.6 μs      310.30      65.98       10.08      2.19
Loop univariate 10k        10000     false    158.0 μs    16760.14     369.88       10.29      8.07
Loop univariate 10k        10000      true    194.0 μs    15744.91     296.36        8.34      6.53
Multivariate 10k           10000     false    217.0 μs     5613.47      78.67       10.15      1.86
Multivariate 10k           10000      true    215.0 μs     5633.93      79.35       10.11      1.89
Dynamic                       15     false     1.39 μs         err      42.39       13.70     11.51
Dynamic                       10      true     1.98 μs        2.02      54.20       11.89     26.50
Submodel*                      1     false     4.23 ns       14.46    1788.80       47.99     19.91
Submodel*                      1      true     4.21 ns       14.46    2099.72       47.78     19.95
LDA                           12      true     23.7 μs        0.60       2.02       29.17       err
===================================================================================================

Environment

Julia Version 1.11.9
Commit 53a02c0720c (2026-02-06 00:27 UTC)
Build Info:
  Official https://julialang.org/ release
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 4 × AMD EPYC 7763 64-Core Processor
  WORD_SIZE: 64
  LLVM: libLLVM-16.0.6 (ORCJIT, znver3)
Threads: 1 default, 0 interactive, 1 GC (on 4 virtual cores)

[yebai]

Thanks, @shravanngoswamii!