Skip to content

Fix LTS extension test failures#1434

Merged
isaacsas merged 3 commits intoSciML:masterfrom
isaacsas:worktree-debug-lts-bifurcation
Mar 5, 2026
Merged

Fix LTS extension test failures#1434
isaacsas merged 3 commits intoSciML:masterfrom
isaacsas:worktree-debug-lts-bifurcation

Conversation

@isaacsas
Copy link
Member

@isaacsas isaacsas commented Mar 5, 2026

@TorkelE seems the issue was the mysterious collect_vars! not working on 1.10 issue again.

Does this fix look ok to you, if so this should get rid of the extension test failure.

isaacsas and others added 3 commits March 5, 2026 15:08
Adding println diagnostics to understand why @test_throws Exception
fails on Julia 1.10 LTS Linux x86_64 but passes everywhere else.
Will be removed once root cause is identified.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Fix `import ModelingToolkitBase as MT` crash (not in test deps) by using
`Catalyst.MT`. Add constantdefs content printing to both diagnostic
blocks to determine whether conslaw_species returns empty because
constantdefs is empty or because collect_vars! fails on Julia 1.10.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
…ompat

MT.collect_vars! returns an empty set on Julia 1.10 LTS (x86_64-linux)
despite valid conservation law expressions, causing conservationlaw_errorcheck
to silently pass when it should throw. Replace with Symbolics.get_variables!
which uses a simpler recursive traversal that works across all Julia versions.

Also removes temporary CI diagnostic code added for debugging.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@isaacsas isaacsas changed the title [DO NOT MERGE] debug: add temporary diagnostics for LTS BifurcationKit test failures Fix LTS extension test failures Mar 5, 2026
@isaacsas isaacsas requested a review from TorkelE March 5, 2026 22:00
@TorkelE
Copy link
Member

TorkelE commented Mar 5, 2026

I have slightly lost track of the various get_variables! and collect_vars! functions and what to use when, But tests seems to pass and if you think this choice of function is sensible then I am all for it

@isaacsas
Copy link
Member Author

isaacsas commented Mar 5, 2026

Conservation law equations shouldn't ever have anything but variables, so I think this is ok.

@isaacsas isaacsas merged commit ba4b5d9 into SciML:master Mar 5, 2026
19 checks passed
@isaacsas isaacsas deleted the worktree-debug-lts-bifurcation branch March 5, 2026 22:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants