@LinearConstraint with syntiax as @addConstraint

[FabrizioLacalandra]

@LinearConstraint(expr) (as well as @QuadConstraint(expr))
does not enable the full expressive power of
@addConstraint(m::Model, con)

or that of
@defExpr(ref, expr)

so that the following does not work
MyLCns = @LinearConstraint([i=1:2:4], x[i]+x[i+1] <=1)

It could be useful to have the same expressive power also for @LinearConstraint to define a (linear/quadratic) constraint without addint it to the model. Later one could get the coefficient of the linear constraints and manipulate them

Fabrizio

[mlubin]

It shouldn't be hard to get this working:

@LinearConstraint(MyLCns[i=1:2:4], x[i]+x[i+1] <=1)

My only concern is that it could be confusing to someone reading the model how this differs from @addConstraint. What about @defLinearConstraint for consistency with @defExpr?

[FabrizioLacalandra]

Totally agree, in fact it is a def of a cns. My final goal is to have this things working
MyLCns.terms.vars

[mlubin]

See @build_constraint. We don't have the one-line version that constructs a JuMP container given index sets. If this is needed, please open a new issue.