Run formatter, start using always_use_return = true

[mhauru]

JuliaFormatter has a new major release, and our GitHub actions are already using it. To catch up, I ran the formatter on all code locally. While I was at it, I enabled always_use_return = true, which according to the Blue style guide should be done anyway, but JuliaFormatter doesn't include it in its BlueStyle config (see domluna/JuliaFormatter.jl#906).

The first commit runs JuliaFormatter v2.1.0 with old settings and does nothing else. The second one removes trailing whitespace. The third adds always_use_return = true and its consequences.

[github-actions]

Benchmark Report for Commit e366079

Computer Information

Julia Version 1.11.4
Commit 8561cc3d68d (2025-03-10 11:36 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)

Benchmark Results

|                 Model | Dimension |  AD Backend |      VarInfo Type | Linked | Eval Time / Ref Time | AD Time / Eval Time |
|-----------------------|-----------|-------------|-------------------|--------|----------------------|---------------------|
| Simple assume observe |         1 | forwarddiff |             typed |  false |                  9.7 |                 1.6 |
|           Smorgasbord |       201 | forwarddiff |             typed |  false |                613.9 |                43.3 |
|           Smorgasbord |       201 | forwarddiff | simple_namedtuple |   true |                436.9 |                44.7 |
|           Smorgasbord |       201 | forwarddiff |           untyped |   true |               1262.6 |                26.8 |
|           Smorgasbord |       201 | forwarddiff |       simple_dict |   true |               4243.6 |                18.8 |
|           Smorgasbord |       201 | reversediff |             typed |   true |               1498.2 |                29.4 |
|           Smorgasbord |       201 |    mooncake |             typed |   true |                967.6 |                 5.3 |
|    Loop univariate 1k |      1000 |    mooncake |             typed |   true |               5655.9 |                 4.1 |
|       Multivariate 1k |      1000 |    mooncake |             typed |   true |               1162.0 |                 8.2 |
|   Loop univariate 10k |     10000 |    mooncake |             typed |   true |              62951.3 |                 3.7 |
|      Multivariate 10k |     10000 |    mooncake |             typed |   true |               9775.3 |                 9.2 |
|               Dynamic |        10 |    mooncake |             typed |   true |                142.2 |                12.4 |
|              Submodel |         1 |    mooncake |             typed |   true |                 26.4 |                 8.6 |
|                   LDA |        12 | reversediff |             typed |   true |                490.4 |                 5.7 |

[codecov]

Codecov Report

Attention: Patch coverage is 87.30159% with 8 lines in your changes missing coverage. Please review.

Project coverage is 84.89%. Comparing base (019e41b) to head (e366079).

Files with missing lines	Patch %	Lines
src/varinfo.jl	82.85%	6 Missing ⚠️
src/model_utils.jl	0.00%	1 Missing ⚠️
src/simple_varinfo.jl	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #886      +/-   ##
==========================================
+ Coverage   84.87%   84.89%   +0.01%     
==========================================
  Files          34       34              
  Lines        3815     3820       +5     
==========================================
+ Hits         3238     3243       +5     
  Misses        577      577              

☔ View full report in Codecov by Sentry.
📢 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.

[coveralls]

Pull Request Test Coverage Report for Build 14377580780

Details

• 27 of 63 (42.86%) changed or added relevant lines in 12 files are covered.
• 1187 unchanged lines in 26 files lost coverage.
• Overall coverage increased (+0.02%) to 84.984%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/model_utils.jl	0	1	0.0%
src/debug_utils.jl	0	2	0.0%
src/extract_priors.jl	0	2	0.0%
src/simple_varinfo.jl	1	3	33.33%
src/varinfo.jl	6	35	17.14%

Files with Coverage Reduction	New Missed Lines	%
src/DynamicPPL.jl	1	90.91%
src/submodel_macro.jl	2	82.35%
src/extract_priors.jl	4	51.72%
src/test_utils/model_interface.jl	5	22.22%
src/test_utils/varinfo.jl	5	78.26%
src/values_as_in_model.jl	6	43.9%
src/logdensityfunction.jl	7	65.22%
src/model_utils.jl	7	7.14%
src/distribution_wrappers.jl	13	9.76%
src/model.jl	13	80.33%

Totals
Change from base Build 14321484767:	0.02%
Covered Lines:	3243
Relevant Lines:	3816

---

💛 - Coveralls

[penelopeysm]

Last time I checked, JuliaFormatter v2 had some bugs. (For additional context, the initial release of v2 was yanked because there were various bugs: JuliaRegistries/General#117438.) The issue tracker suggests to me that there are still quite a few outstanding issues. Are we sure that it's production ready?

Should we be fixing it to v1 instead? https://github.com/TuringLang/actions/blob/e94733241a8cb6a44159eff8eb347c1cb037a53c/Format/action.yml#L19-L25

In an ideal world, I'd love to use v2 and report issues upstream but I fear it might be too annoying to run and compare two different versions of JuliaFormatter.

[mhauru]

That's a good question. v2 was fine on this repo, but I couldn't make a similar PR for Turing.jl because of a Formatter bug. Depends a lot on how quickly the bugs get fixed. I'm happy with fixing to v1 too.

[penelopeysm]

Something like this is a bit scary: domluna/JuliaFormatter.jl#909

[penelopeysm]

We've agreed to stick to v1 for now!