Closed richardotis closed 4 years ago
Hi! This is the friendly automated conda-forge-linting service.
I just wanted to let you know that I linted all conda-recipes in your PR (recipe
) and found it was in an excellent condition.
@conda-forge-admin, please rerender
Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do.
Thanks for figuring that out. We can merge this as is, but shouldn't ipopt be using the conda-forge lapack (if it isn't)? Maybe this doesn't matter with lapack, but it does with blas, hopefully the vendored lapack si built with the conda-forge blas.
I don't have the background to answer that question, but I can confirm that my downstream test suite passes with roughly the same performance characteristics as when using ipopt 3.12.13. I'm not saying that it's comprehensive, but I was getting a few test failures with the other MSVC-based ipopt I was trying to build on Windows, so it is sensitive to numerical precision issues.
Also, are there no consequences that cyipopt is built with conda forge's lapack and ipopt uses its vendored lapack?
I'm not sure if cyipopt is actually using lapack here, or if it is just overlinking because it's what pkg-config
returns.
I guess, let's merge and see if issues arise. Can you add the note in the file though?
LGTM!
Please merge when ready.
Checklist
conda-smithy
(Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)Fixes #18
Now that ipopt 3.13 is vendoring some dependencies, it appears at least one implicit dependency (lapack) that was getting pulled in by upstream conda-forge recipes is now missing. This fix adds lapack as an explicit build-time dependency, which seems to fix the build errors and allows the test to pass.