Add test coverage for warm-start target conversion paths
@chris-maes is already working on this.
Since Sep 8, 2026.
- Dominant language
- Cuda
- Stars
- 1k
- Forks
- 233
- Avg merge
- 4d 4h
- Merged PRs (30d)
- 95
Description
Description
#1803 split warm-start handling into three helpers but added no test coverage for the conversion paths themselves. CodeRabbit flagged this during review; it needs a GPU fixture with a populated warm start, so it was left out of that PR rather than expanding its scope.
Paths needing coverage:
- host-view (Cython spans) → CPU target
- host data → GPU target (H2D via
convert_to_gpu_warmstart) - device data → CPU target (D2H via
convert_to_cpu_warmstart)
The third is the one worth prioritising: it was silently dropped in an early revision of #1803 and only caught by review. The failure mode is a lost warm start with no error — a caller passing handle == nullptr from a context that does have a device (cython_solve.cu:181) simply gets degraded results.
Related: #1801, #1802, #1803, #1804.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.