RMG can't handle overall reactions with four products

[jwallen]

This occurs in some published mechanisms, e.g. USC-II, which have reactions like

C6H4O2 + O -> 2CO + C2H2 + CH2CO

Clearly this is an overall reaction, not an elementary reaction. In order to use these published mechanisms with RMG, we would need to add support for reactions with four products.

[jwallen]

Marileen and I have managed to get RMG to accept reactions of the form A + B + C -> W + X + Y + Z, at least with DASSL. However, I have not done DASPK yet, and I suspect that this commit breaks it. Since DASPK is not used much, I'm inclined to go ahead and push to master anyway, but if anyone would rather DASPK work, I can either leave this on a branch or make an attempt to fix DASPK.

Unfortunately, nothing I've done has changed the degree to which the maximum number of reactant or product species is hardcoded.