Front page | perl.perl5.porters |
Postings from June 2013
So this is going to make me look like a total fool. But I'm going to ask nonetheless. Possibly this is even covered by some documentation somewhere I didn't manage to be creative enough to discover. What's the reason for having op_ppaddr in the OP itself? I can imagine two off the top of my head and I'm not sure on either of them that they're any good: - It's an optimization to have the ptr to chase within the structure which is likely already in the L1 cache. If we had to look it up by OP type from PL_ppaddr, then we'd have to look at a distant bit of memory to find the function pointer. - It's necessary because some nasty, nasty, nasty, horrible evilness is replacing the actual function pointers in OPs secretly for nefarious purposes such as nulling OPs without running afoul of assumptions in further bits of compilation. Wouldn't surprise me. I've had to resort to that in my prototype numeric JIT to hack around my poor understanding of the compile/optimize/linklist phase. Is there another reason? Which one is it? --SteffenThread Next