I guess these ambiguities arise from TuringLang/DynamicPPL.jl#636 which has not been released yet (and therefore currently they show only up in DynamicPPL CI but not in Turing).

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 82.47%. Comparing base (803d2f5) to head (4ae44cb).

@@            Coverage Diff             @@
##           master    #2299      +/-   ##
==========================================
- Coverage   83.98%   82.47%   -1.51%     
==========================================
  Files          24       24              
  Lines        1617     1615       -2     
==========================================
- Hits         1358     1332      -26     
- Misses        259      283      +24     

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

Pull Request Test Coverage Report for Build 10317898020

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 2 unchanged lines in 1 file lost coverage.
• Overall coverage remained the same at 84.348%

Totals
Change from base Build 10281463329:	0.0%
Covered Lines:	1358
Relevant Lines:	1610

💛 - Coveralls

Thanks @devmotion, that makes sense.

We seem to have an unfortunate case here where a PR that passed tests (#2291) was merged, and then immediately after merging tests on master started to fail (https://github.com/TuringLang/Turing.jl/actions/runs/10281463329/job/28451089388). That's still the failure in this PR too. It's a concurrency problem of some kind, only appears on multithreaded. Haven't dug into it yet.

Merging this since the test failure is on master too and the loss of coverage seems unrelated, and this is holding up the DPPL PR. Need to get to the bottom of the test failure and how it slipped through in the first place though.