Observed :
g=a.cloneNode();ac=g.getContext("2d")for(i in ;ac[c.at](innerWidth/)c[i[2]+[i[7]]]=i2,0);ac[c.a](innerHeight/128,innerHeight/128);
Reason : there are several 2D contexts in the code, and some of them are used before the last one is declared. RegPack always insert the hashing loop at the declaration of the last context, yet replaces methods earlier in the code (which in turns changes the offset).
A strategy is needed for cases where a context's methods are called before all contexts are used, such as :
c = a.getContext("2d");
c.fillStyle = "#fff";
c.fillRect(0,0,innerWidth,innerHeight);
na = a.cloneNode();
nc = na.getContext("2d");
One cannot iterate on nc before it is defined. But at this point methods from c have already been used.
Observed :
g=a.cloneNode();ac=g.getContext("2d")for(i in ;ac[c.at](innerWidth/)c[i[2]+[i[7]]]=i2,0);ac[c.a](innerHeight/128,innerHeight/128);
Reason : there are several 2D contexts in the code, and some of them are used before the last one is declared. RegPack always insert the hashing loop at the declaration of the last context, yet replaces methods earlier in the code (which in turns changes the offset).
A strategy is needed for cases where a context's methods are called before all contexts are used, such as :
One cannot iterate on
nc
before it is defined. But at this point methods fromc
have already been used.