build: bump compat and run JET in its own CI job - #151
Merged
Conversation
Collects the bumps from the four open CompatHelper PRs (#96, #99, #137, #138) into a single change, rebased onto current main: - JET: 0.9.18, 0.10 -> 0.9.18, 0.10, 0.11 - LinearSolve: 3.75 -> 3.75, 5 - Peaks: 0.5 -> 0.5, 0.6 PRs #137 and #138 are the same LinearSolve bump reported separately for [weakdeps] and [extras], so they collapse into one entry.
CI enables `force_latest_compatible_version` on `compathelper/*` branches, which pins OrderedCollections to 2.x. Latexify 0.16 does not support OrderedCollections 2, so the test environment no longer resolves. Restrict the bound to 1.8 so the forced-latest resolution picks 1.8.2 and Latexify 0.16.11 stays installable.
JET 0.10 and newer require Julia 1.12, so linting cannot run on the LTS runner. Follow the layout used in HarmonicBalance.jl: `runtests.jl` gates on `GROUP`, the matrix jobs run `GROUP=Core`, and a dedicated `jet` job runs `GROUP=JET` on Julia 1. Running JET on 1.12 for the first time surfaced 12 reports, none of which were visible before because the old `VERSION < v"1.12.0-beta"` guard meant JET only ever ran on the LTS at 0.9.x: - `LimitCycles` called `QuestBase.add_harmonic!` while only importing names from QuestBase, not the module itself. `add_pairs!` would have thrown an UndefVarError at runtime. - `bar` was assigned inside `if show_progress` and then read unconditionally in four functions. Bind it to `nothing` instead and guard on `isnothing`. - `idx_pairs` was assigned in an if/elseif with no fallback in `sort_2D`. Also disable `force_latest_compatible_version` on the LTS job. CompatHelper branches otherwise resolve every [compat] entry to its newest version, which the LTS cannot satisfy for dependencies that require Julia 1.12.
oameye
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Consolidates the four open CompatHelper PRs (#96, #99, #137, #138) into one change, rebased onto current main since they were all opened against 0.4.1. #137 and #138 are the same LinearSolve bump reported twice, once for
[weakdeps]and once for[extras].The JET bump then wouldn't resolve. JET 0.10 and up require Julia 1.12, and julia-runtest forces every
[compat]entry to its newest version oncompathelper/*branches, so the LTS runner was being asked to install a JET it can't have. JET now runs in its own job on Julia 1, gated byGROUPthe same way HarmonicBalance.jl does it, and the LTS job opts out offorce_latest_compatible_version, since that mode can never work there for a dependency that has moved past 1.10.That was also the first time JET ran on 1.12 at all. The old guard was
VERSION < v"1.12.0-beta", so in practice it only ever ran on the LTS at 0.9.x, and it came back with twelve reports. The real one:LimitCyclesdoesusing QuestBase: get_all_terms, ..., which binds those names but not the module itself, soadd_pairs!would have thrown anUndefVarErrorthe moment anyone called it. The rest arebarandidx_pairsgetting assigned inside a branch and then read outside it.OrderedCollections is narrowed to 1.8 rather than bumped. Latexify 0.16 doesn't work with OrderedCollections 2, which only surfaced once force_latest started picking the newest of everything. Worth widening again when Latexify catches up.