Transpose allocates 16 bytes

[klamike]

@tmigot @amontoison

See existing discussion in JuliaSmoothOptimizers/NLPModelsTest.jl#128, JuliaSmoothOptimizers/NLPModels.jl#499 (comment)

Stopped tracking it recently in NLPModels: JuliaSmoothOptimizers/NLPModelsTest.jl#129

I think it's due to making the {Adjoint/Transpose/Conjugate}LinearOperator, which has one field. Maybe we should pre-allocate those when making LinearOperator?

[amontoison]

@klamike Is it still an issue?

[klamike]

It does still allocate 16 bytes. Whether that is an issue/problem, I don't know. Happy to close this.