Patch #7436 by Changlin Zhou (shzhc):
The patch submitter found a case where freed memory was being accessed again later. Fortunately (or unfortunately), this bug has not shown itself so far, and has therefore been easy to miss. In fact, somehow, everything still manages to work correctly without it.
This commit is contained in:
@@ -2702,6 +2702,7 @@ void addvert_Nurb(int mode)
|
||||
newbp->f1= 1;
|
||||
MEM_freeN(nu->bp);
|
||||
nu->bp= newbp;
|
||||
bp= newbp + 1;
|
||||
}
|
||||
else if(bp== (nu->bp+nu->pntsu-1)) { /* last */
|
||||
bp->f1= 0;
|
||||
@@ -2713,15 +2714,16 @@ void addvert_Nurb(int mode)
|
||||
nu->bp= newbp;
|
||||
newbp+= nu->pntsu;
|
||||
newbp->f1= 1;
|
||||
bp= newbp - 1;
|
||||
}
|
||||
else bp= 0;
|
||||
|
||||
if(bp) {
|
||||
nu->pntsu++;
|
||||
|
||||
|
||||
if(nu->resolu<3) nu->resolu++;
|
||||
makeknots(nu, 1, nu->flagu>>1);
|
||||
|
||||
|
||||
if(mode=='e') {
|
||||
VECCOPY(newbp->vec, bp->vec);
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user