dissenting:
If this case were controlled by Oregon y. Kennedy, 456 U.S. 667, 102 S.Ct. 2083, 72 L.Ed.2d 416 (1982), retrial would not be barred, for the prosecution did not “intend[] to provoke [appellant] into moving for a mistrial.” To the contrary, the prosecutor intended, by concealing his misconduct, to preclude appellant from moving for a mistrial. Kennedy, however, does not control, for it is not to be *299applied retroactively. Commonwealth v. Beaver, 317 Pa. Super. 88, 102, 463 A.2d 1097, 1104 (1983). The case that does control is United States v. Dinitz, 424 U.S. 600, 96 S.Ct. 1075, 47 L.Ed.2d 267 (1976), which Kennedy overruled. Under Dinitz, retrial is barred by “prosecutorial misconduct undertaken in bad faith to prejudice or harass” the defendant. Since that is precisely the sort of prosecutorial misconduct that occurred here, we should hold that retrial is barred and reverse the order of the trial court permitting retrial.
In Oregon v. Kennedy, 456 U.S. 667, 102 S.Ct. 2083, 72 L.Ed.2d 416 (1982), the Court held
... that the circumstances under which such a defendant may invoke the bar of double jeopardy in a second effort to try him are limited to those cases in which the conduct giving rise to the successful motion for a mistrial was intended to provoke the defendant into moving for a mistrial.
Id. at 679, 102 S.Ct. at 2091.
I believe that Kennedy means what it says, and for that reason I cannot join Judge DEL SOLE, although I agree with much of what he says.1 Here, the prosecutor did not “intend[ ] to provoke [appellant] into moving for a mistrial.” His intention was just to the contrary. Usually, of course, intention must be inferred from conduct. The only inference to be drawn from the prosecutor’s conduct here is that he did not want appellant to know about the sentencing agreement with Ravenell. In other words, he hoped the trial would proceed smoothly without interruption to a guilty verdict: he intended, by concealing his misconduct, to preclude appellant from moving for a mistrial. Accordingly, if Kennedy controls this case, retrial is not barred.
*300However, Kennedy does not control this case. To explain this conclusion, it is necessary to consider United States v. Dinitz, 424 U.S. 600, 96 S.Ct. 1075, 47 L.Ed.2d 267 (1976).
Dinitz held:
The Double Jeopardy Clause does protect a defendant against governmental actions intended to provoke mistrial requests and thereby to subject defendants to the substantial burdens imposed by multiple prosecutions. It bars retrials where “bad-faith conduct by judge or prosecutor,” threatens the “[hjarassment of an accused by successive prosecutions or declaration of a mistrial so as to afford the prosecution a more favorable opportunity to convict” the defendant
424 U.S. at 611, 96 S.Ct. at 1079 (citations omitted).
See also Lee v. United States, 432 U.S. 23, 33, 97 S.Ct. 2141, 2147, 53 L.Ed.2d 80 (1977) (“Only if the underlying error was ‘motivated by bad faith or undertaken to harass or prejudice’ would there be any barrier to retrial.”) {citing United States v. Jorn, 400 U.S. 470, 485, 91 S.Ct. 547, 557, 27 L.Ed.2d 543 (1971)). Our Supreme Court adopted this analysis in Commonwealth v. Starks, 490 Pa. 336, 416 A.2d 498 (1980):
Generally the double jeopardy clause does not bar retrial of a defendant who obtains a new trial upon his request for a mistrial____
This restriction on the double jeopardy clause, however, is not without its own limitations. The Supreme Court has held that double jeopardy will bar retrial if the defendant sought the mistrial as a result of prosecutorial misconduct amounting to overreaching____
The United States Supreme Court has enunciated principally two types of prosecutorial overreaching. First there is the prosecutorial misconduct which is designed to provoke a mistrial in order to secure a second, perhaps more favorable, opportunity to convict the defendant. Second there is the prosecutorial misconduct undertaken in bad faith to prejudice or harass the defendant. In contrast to prosecutorial error, overreaching is not an *301inevitable part of the trial process and cannot be condoned. It signals the breakdown of the integrity of the judicial proceeding, and represents the type of prosecutorial tactic which the double jeopardy clause was designed to protect against.
Id., 490 Pa. at 340-41, 416 A.2d at 500 (citations omitted). And see Commonwealth v. Virtu, 495 Pa. 59, 65 n. 7, 432 A.2d 198, 201 n. 7 (1981) (In Starks, Supreme Court “set forth circumstances under which prosecutorial misconduct would be the basis for a claim of former jeopardy____”); Commonwealth v. Clark, 287 Pa.Super. 380, 390 n. 9, 430 A.2d 655, 660 n. 9 (1981) (Starks represents the Commonwealth’s “authoritative standard for overreaching cases____”)
When Kennedy’s holding is compared with Dinitz’s, it is apparent that Kennedy overruled Dinitz, for Kennedy reduced the bases upon which a defendant, after moving for a mistrial, could plead double jeopardy from the two provided by Dinitz to only one. As Justice STEVENS explained:
Today the Court once again recognizes that the exception properly encompasses the situation in which the prosecutor commits prejudicial error with the intent to provoke a mistrial. But the Court reaches out to limit the exception to that one situation, rejecting the previous recognition that prosecutorial overreaching or harassment is also within the exception.
Oregon v. Kennedy, supra, 456 U.S. at 687, 102 S.Ct. at 2095 (footnotes omitted).
The majority, Justice STEVENS said, had “gratuitously lop[ped] off a portion of the previously recognized exception.” Id. at 681, 102 S.Ct. at 2092.2
When one recognizes that Kennedy overruled Dinitz, the conclusion that Kennedy does not control this case becomes *302clear. Appellant was tried before Kennedy was decided. The decision whether to apply a new constitutional rule retroactively is a matter of judicial discretion, August v. Stasak, 492 Pa. 550, 424 A.2d 1328 (1981), and we have decided that because (among other reasons) Kennedy adopted a rule that “narrowed the prior [Dinitz ] standard of review applicable to mistrials, i.e., overruled cases holding to the contrary,” and because “injustice or hardship” would thereby be avoided, Kennedy should not be applied retroactively. Commonwealth v. Beaver, 317 Pa.Super. 88, 102, 463 A.2d 1097, 1104 (1983). See also Commonwealth v. Lark, 330 Pa.Super. 225, 230-231 n. 7, 479 A.2d 522, 524 n. 7 (1984) (“Our court has decided to apply Oregon v. Kennedy, supra, prospectively) (citations omitted).”
The question we must decide, therefore, is whether under Dinitz and the decisions by our Supreme Court adopting it, retrial is barred because the prosecutorial misconduct at issue here was “undertaken in bad faith to harass an accused by successive prosecutions or prejudice his prospects for an acquittal.” Commonwealth v. Clark, supra, 287 Pa.Super. at 390, 430 A.2d at 659 (citing Commonwealth v. Starks, supra). I have no doubt that it was; indeed, it would be difficult to imagine a closer fit with the Dinitz test.3 The prosecutor engaged in “misconduct”; he did so “in bad faith”; and his purpose was to “prejudice [appellant’s] prospects for an acquittal” — specifically, by concealing the sentencing agreement with Ravenell, to prevent the defense from damaging the credibility of the Commonwealth’s key witness.
The order of the trial court should be reversed.
. I note, however, that there is support for Judge DEL SOLE’s reading of Kennedy in Justice POWELL’s concurring opinion, which characterizes the holding of the majority as follows: "... the intention of a prosecutor determines whether his conduct, viewed by the defendant and the court as justifying a mistrial, bars a retrial of the defendant under the Double Jeopardy Clause.” 456 U.S. at 679, 102 S.Ct. at 2091.
. Judge WIEAND’s statement that Kennedy "did not represent a clear break with precedent” but only "clarified] prior decisional language,” concurring at 297. WIEAND, J., is not only contrary to Justice STEVENS’s view of Kennedy but to this court’s view. See discussion infra, citing our decisions that Kennedy is not to be applied retroactively.
. I cannot agree with the majority that there is a functional difference between concealed prosecutorial misconduct and overt misconduct, and in this I agree with Judge DEL SOLE. I also believe that the majority’s reliance on this distinction has the incongruous result of encouraging prosecutors to conceal misconduct, for if they are successful, upon the majority's analysis, double jeopardy will not bar a retrial. Thus, although the majority says that "[they] by no means suggest that concealed misconduct is somehow preferable to blatant misconduct," at 1126, they hold that it is preferable, for they do not apply a double jeopardy bar to concealed misconduct and would apply it to blatant misconduct.