Mathieu Besançon

@odow related to the 1.x label

@Wikunia also had some explanations on building bridges: https://opensourc.es/blog/constraint-solver-bridges/ https://opensourc.es/blog/generic-bridges/

agreed with @odow, if one adds additional branching for some reason on another problem and solves the current MIP as a subproblem, updating lower or upper bounds to fix the...

From the Gurobi [doc](https://www.gurobi.com/resource/functional-code-examples/): > A brief note for users of the Gurobi MATLAB® and R interfaces: our interfaces to these languages are built around the assumption that you will...

Not sure there is a way to know, I would say model generation time only.

> Is it just to gain model generation time or also solve time ? Another point where this can affect performance is callbacks, if constraints defined in callbacks require moving...

I wanted to get back to this issue and start thinking of possible solutions. Pinging @chriscoey now that Hypatia is released. Many solvers will take a pointer to an array...

An alternative to bridges would be to always use an interface when interacting with abstract linear operators, and both SAF and the pair-of-array type implementing this interface

The ambiguity list seems to have evolved with 1.5: ```julia julia> Test.detect_ambiguities(Base, Core) 44-element Array{Tuple{Method,Method},1}: (*(adjA::LinearAlgebra.Adjoint{var"#s826",var"#s825"} where var"#s825"

sorry didn't mean to sound pushy, it was to update the status of the issue if someone sees something