In an attempt to see if there is an interaction between cocaine and Provigil, 20mg or 40mg IV cocaine was given pre and post Provigil (400mg or 800mg) for 7 days. There was an interaction, but it turned out to be positive: Provigil reduced systemic cocaine exposure.

A safety study investigated (in 7 people) the interaction between cocaine (30mg IV) and Provigil (modafinil) 200mg or 400mg, or placebo, and found no synergistic effect on vital signs (T, BP, HR) or EKG.  Not only did it not augment cocaine euphoria, it blunted it in one person. 

In another study, 62 (mostly black) males addicted to cocaine were randomized to placebo, CBT, or Provigil 400mg.  Abstinence, the primary outcome, was measured by benzoylecgonine in the urine.  Patients on Provigil were abstinent longer, and produced fewer positive urines (i.e. fewer relapses.)  Importantly, no one got addicted to Provigil.

Unlike cocaine and Ritalin (methylphenidate) Provigil did not produce "cocaine like discriminitive stimulus" (i.e. didn't feel like cocaine; Ritalin and cocaine do feel like cocaine.)

That's all we know about Provigil vs. cocaine so far, which is pitiful but not inconsequential.  Given Provigil's near absence of terrible side effects, I say it's worth a try.

In the interest of completeness (and correctness) I have to correct the major paper, above (the 62  people with the urine tests) .  The authors of that paper propose the following potential mechanism:

 Its glutamate-enhancing action (Ferraro et al, 1998; 1999) might be clinically advantageous in cocaine dependence because the repeated administration of cocaine depletes extracellular glutamate levels

Except that the Ferraro paper doesn't actually say that. What it says is that it inhibits striatal and globus pallidus GABA, but doesn't directly affect glutamate.   In order for it to have any effect on  striatal glutamate, you needed 300mg/kg (i.e. 21,000mg.  See you on the other side.)  Given that GABA and glutamate are opposites (i.e. glutamate goes up because GABA goes down), it's probably a small point, but not an insignificant one: if it directly increases glutamate, it could antagonize Lamictal or even potentially cause seizures (and it does neither.)

The second  Ferraro reference finds essentially the same thing: inhibition of medial preoptic area and posterior hypothalamus GABA, and consequently glutamate increases.  And again, all of this occurs at preposterously high doses (100-300mg/kg.)

In interesting side finding of Ferraro's study is that the medial preoptic area and posterior hypothalamus are primarily controlled by tonic GABA inhibition; consequently modafinil's (or any drug's) effect of increasing glutamate in these areas can be blocked by giving a GABA-A antagonist.

So Provigil operates by  (probably) by antagonizing GABA, not specifically by enhancing glutamate (neither synthesis of or transport of). 

To further complicate this picture, it may be that the effects on GABA and glutamate are both indirect, and really the result of serotonin agonism. In an earlier study by the same guy, the decreases in GABA were partially prevented by a 5HT3 blocker (think Zofran, Remeron).  Does Provigil work through serotonin?  In a later study, the same guy finds that at 100mg/kg, Provigil does, after all,  increase serotonin in the medial preoptic area and posterior hypothalamus.   (At lower doses, 10-100mg/kg, it increases serotonin in the cortex, dorsal raphe and the amygdala.)  (And in another study, (yes, by that same guy again,) 3mg/kg Provigil, which in itself has no effect on serotonin, synergistically augmented serotonin increase to fluoxetine and imipramine.)

We already know that Provigil can reduce the sedation that comes from varying drugs, like SSRIs, general anesthesia,  haloperidol, and chlorpromazine. It would be interesting to see if Provigil was unable to improve sedation on Remeron, supporting the 5HT3 hypothesis.

Good luck out there.