Changes in version 1.0.2.9000 Changes in version 1.0.2 (2023-09-09) - Fixes a bug in how the package documentation was specified. - ompr.roi now requires at least R 3.5.0 because ompr requires it. Changes in version 1.0.1 (2022-09-05) - ompr.roi does not use a deprecated {Matrix} 1.4-2 function anymore. Changes in version 1.0.0 (2022-01-27) - ompr.roi is now licensed MIT. - LazyData was removed from the DESCRIPTION file as is produced a CRAN warning. - ROI's status and message information are now part of the returned solution and can be accessed calling additional_solver_output() on the returned value. Breaking changes - In line with ROI's way of encoding a general termination status, the solver_status(result) is now "success" if ROI returned code = 0, otherwise it is "error". Changes in version 0.8.0 (2018-06-11) Breaking changes - The package exports column and row duals of LPs. Extracting duals might be solver dependent. Only tested with GLPK. This requires version 0.8 of ompr. - Minimum supported R version is 3.4.0 since slam also requires >= 3.4 New features - ompr.roi works with other ompr backends as well. - ompr.roi always returns a solution, even if the solution status is not optimal. It extracts the ROI solution with force=TRUE. Changes in version 0.7.0 (2017-11-18) Major breaking changes - The package uses the new sparse matrix interface of ompr and ROI. - Minimum supported R version is now 3.3.0 Minor changes - as_ROI_model exports an 'ompr' model to ROI::OP Github Issue #9 Changes in version 0.6.1 (2017-10-23) Bugfix - Fixed a bug where the ROI::V_bound object was incorrectly initialised. This causes an error with ROI >= 0.3.0. Changes in version 0.6.0 (2017-04-19) - First release on CRAN