Possible problem on `contype=L`; verify these problems against SIF

[abelsiqueira]

The classification tests compare contype=L against decoding and counting the linear constraints.

• SIF says these problems have linear constraints, but deconding gives otherwise:
  "RDW2D52U", "RDW2D52F", "RDW2D52B", "RDW2D51U", "RDW2D51F"
• Decoding gives linear constraints, but SIF disagrees:
  "STEENBRA", "DALLASM", "GRIDNETC", "GRIDNETH", "GRIDNETG", "STEENBRD", "BIGBANK", "GRIDNETI", "LINSPANH", "STEENBRE", "SPANHYD", "GRIDNETE", "GRIDNETF", "STEENBRF", "GRIDNETD", "SMBANK", "LSNNODOC", "DALLASS", "GRIDNETA", "STEENBRG", "GRIDNETB", "STEENBRC", "DALLASL", "WATER", "STEENBRB"

[amontoison]

@nimgould
This issue is quite old but I don't know if you are aware of this.

[nimgould]

"RDW2D52U", "RDW2D52F", "RDW2D52B", "RDW2D51U", "RDW2D51F" only have linear constraints, but the decoder can't tell as the constraints are encoded via (linear) element functions; the decoder is unable to look inside these, and so treats them as nonlinear.

The other problems you mention only involve network constraints; these are linear, but special and thus classified separately, as some software is specifically designed for this class

[nimgould]

(no, @amontoison , I was not aware of this, I have not been notified of julia issues. Thanks for letting me know)