Ian McInerney comments

Results 230 comments of


                                            Ian McInerney

CI: Run all tests compiled for a configuration

> Looks like in our case we need to do a plain make before a make test will succeed (as we're doing here in the CI). Just pointing it out...

Skip polishing when there is no active set detected

@goulart-paul any further comments on this, or is it good to merge?

Null pointer in CUDA backend when using polish with no active set

The relevant backtrace for this error is ``` #6 0x0000000000522ee9 in check (result=CUSPARSE_STATUS_INVALID_VALUE, func=0x5bfe08 "cusparseSpMV( CUDA_handle->cusparseHandle, CUSPARSE_OPERATION_NON_TRANSPOSE, &alpha, A->SpMatDescr, vecx, &beta, vecy, CUDA_FLOAT, CUSPARSE_SPMV_ALG_DEFAULT, A->SpMatBuffer)", file=0x5bf710 "/home/b63738im/dev/optimization/osqp/c_code/algebra/cuda/src/cuda_lin_alg.cu", line=946) at /home/b63738im/dev/optimization/osqp/c_code/algebra/cuda/include/helper_cuda.h:589...

NULL pointer in CUDA backend when using polish on an unconstrained problem

Relevant backtrace: ``` #6 0x0000000000522ee9 in check (result=CUSPARSE_STATUS_INVALID_VALUE, func=0x5bfe08 "cusparseSpMV( CUDA_handle->cusparseHandle, CUSPARSE_OPERATION_NON_TRANSPOSE, &alpha, A->SpMatDescr, vecx, &beta, vecy, CUDA_FLOAT, CUSPARSE_SPMV_ALG_DEFAULT, A->SpMatBuffer)", file=0x5bf710 "/home/b63738im/dev/optimization/osqp/c_code/algebra/cuda/src/cuda_lin_alg.cu", line=946) at /home/b63738im/dev/optimization/osqp/c_code/algebra/cuda/include/helper_cuda.h:589 #7 0x000000000052f380 in cuda_mat_Axpy (A=0x1a846f50,...

Rust Static Binary Licenses / Static Binary Licensing Policy

> 2. Other packaging ecosystems have inconsistent takes on the issue >> Fedora pays lip service to the topic here but does not seem to consistently incorporate dependency licenses into...

Custom timing functions to allow profiling in code generation

Looking at this a bit more, I am now thinking that it would be best to simply allow the user to provide their own implementation of `osqp_tic` and `osqp_toc` if...

Custom timing functions to allow profiling in code generation

@goulart-paul, I think consistency should be an important factor when building these style of interfaces. If someone has to use two different flows for adding in custom override functions it...

Custom timing functions to allow profiling in code generation

A version of this has been merged into the `develop-1.0` branch and will be included in the v1.0 release. All that is left is to document it, but that is...

Rename error.h to osqp_error.h to avoid OS collisions

Thank you for proposing this. Please note that our next planned release is being developed on the `develop-1.0` branch, not the master branch, so please retarget the PR against that...

Enable user-defined functions for interrupt and timing interfacing

@vineetbansal thanks for the quick look. I have merged this now, so you can start modifying the Python interface to work with the new systems. Let me know if you...